Clinical Chemistry

Emmerich,, Joseph;Ridker, Paul, M

doi: 10.1373/clinchem.2007.100313pmid: 18310142

With rapid and dramatic success, the Genome-Wide Association Study (GWAS) has proven to be an effective tool for the discovery of unsuspected genetic determinants of common disorders and has opened a new armamentarium for the pathophysiologic exploration of numerous diseases. The best illustration of the feasibility and strength of the GWAS approach was demonstrated in June of 2007 by a consortium of more than 50 British research groups participating in the Wellcome Trust Case-Control Consortium (WTCCC). Working collaboratively, the WTCCC investigators studied 14 000 cases of 7 common diseases and 3000 shared controls and identified 24 independent association signals, 1 in bipolar disease, 1 in coronary disease, 9 in Crohn disease, 3 in rheumatoid arthritis, 7 in type 1 diabetes, and 3 in type 2 diabetes, each with a statistical effect approaching or exceeding genome–wide levels of significance (1). Remarkably, the key finding in this study for coronary heart disease—a clear association between vascular risk and common variation in the region of chromosome 9p21.3—was rapidly validated in a series of similar GWAS studies, including the Cardiogenics Consortium; the Ottawa, Dallas, and Framingham Heart Studies; and the DeCode Genetics program in Iceland (2)(3)(4)(5). The chromosome 9p21.3 region contains the coding sequences of genes for 2 cyclin-dependent kinase inhibitors known to play roles in tumor suppression, cell proliferation, and apoptosis. Thus, these validated GWAS findings for coronary disease not only raise the concept of a novel genetic determinant of disease, but also provide strong pathophysiologic support for prior work linking each of these processes directly to atherogenesis and plaque disruption. What is less clear from emerging GWAS studies is whether or not the discovery of new gene-disease associations will ultimately help identify persons at high risk, particularly for complex disorders such as atherothrombosis, for which major environmental determinants exist. As described by the Framingham Heart Study investigators as far back as 1961, older age, smoking, hypertension, hyperlipidemia, and diabetes are common determinants of coronary heart disease (6), and these “traditional” risk factors have been codified into global risk scores for the prediction of cardiovascular risk (7)(8). It is also widely recognized, however, that 15% to 20% of cases of incident myocardial infarction or ischemic stroke occur among individuals without these risk factors, and that nearly 50% occur in the absence of hyperlipidemia. Thus, there is considerable interest in finding new markers able to predict the occurrence of atherothrombosis more accurately and with improved risk classification. To date, work on novel atherothrombotic risk factors has largely focused on biomarkers of hemostasis and thrombosis (fibrinogen, von Willebrand factor, tissue plasminogen activator, plasminogen activator inhibitor, factor VII) or on biomarkers of inflammation [C-reactive protein (CRP), serum amyloid A, interleukin-6, sCD40L, intercellular adhesion molecule-1] (9). Unfortunately, with the exception of CRP, most of these novel biomarkers have failed to show additive value for the prediction of future vascular events. For example, in the development and validation of the Reynolds Risk Score, in which a panel of 35 factors was assessed among 24 558 initially healthy American women, the best simplified prediction model for future vascular events included the traditional Framingham risk determinants and only 2 other factors—high-sensitivity CRP (representing inflammatory risk) and parental history of myocardial infarction before age 60 (representing genetic risk) (10). These 2 new factors reclassified 40% to 50% of those at intermediate risk into higher or lower risk categories, an effect with important clinical implications for the targeting of preventive therapies (see www.reynoldsriskscore.org). In this issue of Clinical Chemistry, Talmud et al. directly address the clinical question of whether or not detection of polymorphism in the 9p21.3 region impacts on global risk prediction (11). In a carefully performed and thoughtful analysis, the authors first evaluated the relationship of rs10757274 (a previously defined tag single-nucleotide polymorphism for the 9p21.3 region) as a determinant of vascular risk in the prospective Northwick Park Heart Study-II (NPHS-II) in which 2742 men were followed over a 15-year period for incident coronary heart disease. As in prior studies, polymorphism at this locus was common (frequency of the G allele 0.48, 95% CI 0.47–0.50), and compared to men homozygous for the common A allele, men homozygous for the G allele had a 1.6 times greater hazard ratio for future vascular events (95% CI 1.12–2.28). These effects were largely independent of traditional risk markers and consistent for the individual endpoints of myocardial infarction and coronary artery bypass surgery. Somewhat surprisingly, the magnitude of effect on risk associated with polymorphism at rs10757274 was not materially altered in analyses after further adjustment for family history, despite the fact that family history is a major risk in the NPHS-II population (12). What makes the data from Talmud et al. novel is that the authors then take a second important step and ask whether polymorphism at rs10757274 substantively adds to risk prediction based on traditional markers such as those used in the Framingham Risk Score. Here, readers of the genetic-epidemiology literature will find data to satisfy a full range of opinions. For those who are skeptical that genetic information can improve risk prediction, ample data are supplied, in that the area under the ROC curve (as defined by the c-statistic) increased from 0.62 to only 0.64 when data on rs10757274 were added to a panel of traditional risk factors, a small and nonsignificant effect (P = 0.14). Thus, the authors conclude that, on its own, genetic variation near chromosome 9p21.3 does not add to the overall risk prediction n the NPHS-II cohort. On the other hand, statisticians have recently come to recognize that the c-statistic (an effective tool for evaluating diagnostic tests) may be ill suited for the evaluation of risk prediction models. For example, Cook has demonstrated that widely accepted risk factors such as LDL and HDL cholesterol, hypertension, and even smoking typically have multivariable relative risks <2.0 and thus on their own would also not substantively improve the c-statistic (13). Further, in risk prediction modeling, calibration and reclassification may play as great a role in defining clinical utility as does discrimination, the test characteristic summarized by the c statistic (14). Recognizing this issue, Talmud et al. also provide promising data for those who believe polymorphism data will ultimately impact on our ability to predict clinical risk. Specifically, despite marginal improvement in the c statistic, the authors also show that knowledge of rs10757274 genotype does improve model fit as described by the Bayes Information Criteria and that model calibration as assessed by the Hosmer-Lemeshow statistic improves when genotype information is included. Although the total number of individuals reclassified as a result of rs10757274 data was small in the NPHS-II cohort, those reclassified were often at intermediate risk, the category for which improved risk prediction is likely to be most helpful. To further address reclassification, Talmud et al. additionally model the effect of 10 hypothetical randomly assigned gene variants with similar allele frequencies and risk characteristics as rs10757274. Although such variants are not known to exist, the authors find that this multimarker approach could indeed effectively reclassify large proportions of patients in much the same way the multimarker Framingham or Reynolds Risk Scores combine data from several pathways to improve risk detection. Thus, the study from Talmud et al. nicely demonstrates the clinical complexity that will accompany attempts to treat specific single-nucleotide polymorphisms as risk factors for multifactorial disorders such as atherothrombosis. This complexity is not limited to genetic discovery, but will also impact on the development of other high-throughput technologies in molecular biology, including proteomics. In our opinion, whether or not fishing for new genes will ultimately catch heart disease far enough upstream for prevention to be successful remains an open question. In the meantime, considerable pathophysiologic insight will come from ongoing GWAS studies, and the potential to find new targets for therapy will remain a driving force for this arena of research. Grant/funding Support: Supported by research grants to J.E. and P.M.R. from the Leducq Foundation, Paris FR. Financial Disclosures: None declared. References 1 . Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature (Lond) 2007 ; 447 : 661 -678. Crossref Search ADS 2 Helgadottir A, Thorleifsson G, Manolescu A, Gretarsdottir S, Blondal T, Jonasdottir A, et al. A common variant on chromosome 9p21 affects the risk of myocardial infarction. Science (Wash DC) 2007 ; 316 : 1491 -1493. Crossref Search ADS 3 McPherson R, Pertsemlidis A, Kavaslar N, Stewart A, Roberts R, Cox DR, et al. A common allele on chromosome 9 associated with coronary heart disease. Science (Wash DC) 2007 ; 316 : 1488 -1491. Crossref Search ADS 4 Samani NJ, Erdmann J, Hall AS, Hengstenberg C, Mangino M, Mayer B, et al. Genomewide association analysis of coronary artery disease. N Engl J Med 2007 ; 357 : 443 -453. Crossref Search ADS PubMed 5 Larson MG, Atwood LD, Benjamin EJ, Cupples LA, D’Agostino RB, Fox CS, et al. Framingham Heart Study 100K project: genome-wide associations for cardiovascular disease outcomes. BMC Med Genet 2007 ; 8 (Suppl 1): S5 . Crossref Search ADS PubMed 6 Kannel WB, Dawber TR, Kagan A, Revotskie N, Stokes J, 3rd. Factors of risk in the development of coronary heart disease: six year follow-up experience. The Framingham Study. Ann Intern Med 1961 ; 55 : 33 -50. Crossref Search ADS PubMed 7 . Executive Summary of the Third Report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA 2001 ; 285 : 2486 -2497. Crossref Search ADS PubMed 8 Wood D, De Backer G, Faergeman O, Graham I, Mancia G, Pyorala K. Prevention of coronary heart disease in clinical practice: recommendations of the Second Joint Task Force of European and other Societies on Coronary Prevention. Atherosclerosis 1998 ; 140 : 199 -270. Crossref Search ADS PubMed 9 Emmerich J. Atherothrombosis risk factors: too many and too little. J Thromb Haemost 2007 ; 5 : 1793 -1794. Crossref Search ADS PubMed 10 Ridker PM, Buring JE, Rifai N, Cook NR. Development and validation of improved algorithms for the assessment of global cardiovascular risk in women: the Reynolds Risk Score. JAMA 2007 ; 297 : 611 -619. Crossref Search ADS PubMed 11 Talmud PJ, Cooper JA, Palmen J, Lovering R, Drenos F, Hingorani AD, Humphries SE. Chromosome 9p21.3 coronary heart disease locus genotype and prospective risk of CHD in healthy middle-aged men. Clin Chem 2008 ; 54 : 467 -474. Crossref Search ADS PubMed 12 Hawe E, Talmud PJ, Miller GJ, Humphries SE. Family history is a coronary heart disease risk factor in the Second Northwick Park Heart Study. Ann Hum Genet 2003 ; 67 : 97 -106. Crossref Search ADS PubMed 13 Cook NR. Use and misuse of the receiver operating characteristic curve in risk prediction. Circulation 2007 ; 115 : 928 -935. Crossref Search ADS PubMed 14 Cook NR, Buring JE, Ridker PM. The effect of including C-reactive protein in cardiovascular risk prediction models for women. Ann Intern Med 2006 ; 145 : 21 -29. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Oudejans, Cees B, M

doi: 10.1373/clinchem.2007.100123pmid: 18310143

The human genome contains a large layer of hidden biological information that is not accessible by proteomic or metabolic methods (1)(2). This information does not involve the typical (end)products of gene expression such as proteins. Instead, it involves genes that are transcribed but not translated (noncoding RNA), and DNA sequences that are neither transcribed nor translated (noncoding DNA) (1)(2). When this information, despite being noncoding in nature, is unique for the fetus or at least different from that in maternal blood cells (1), it can be used as a biomarker during pregnancy for the purpose of noninvasive prenatal diagnostics such as detection of Down syndrome. This application, along with a systematic approach to target this category of mostly unexplored information, is elegantly shown by Chim and coworkers in the current issue of Clinical Chemistry (2). Why is the reporting of these findings so timely? According to the Fantom cDNA3 database (2005 release), the percentage of biological information presented by noncoding RNA is at least one-third of the more than 100 000 transcripts expressed in humans (3). Noncoding RNAs include the rapidly expanding family of small (21–35 nucleotides in length), noncoding microRNAs (miRNA) (4)(5), including small interfering RNA (siRNA), repeat-associated RNA (rasiRNA) (6), piwi interacting RNA (piRNA) (7), and mirtrons (8). The second category consists of noncoding DNA carrying differential DNA markers (2). When these signatures involve (chemical) DNA modifications, such as methylation of CpG dinucleotides, in the absence of sequence variations, they are called epigenetic. Epigenetic signatures that differ from those of fetal placenta cells and maternal blood cells can be used for noninvasive prenatal diagnostics (1). As for noncoding RNA, the type, number, and nature of noncoding DNA sequences has gained increased interest, with strong reappraisal of their biological significance. The paper by Stephen Chim and coworkers in the current issue of Clinical Chemistry (2) is an excellent demonstration of how the hidden and largely unexplored class of biological information provided by noncoding DNA and carried by placental DNA can be retrieved. Retrieval was done systematically and reliably with near completeness for the chromosome region of interest (2). By exploring methylated DNA markers, which differ from those of placenta and maternal blood cells, only those biomarkers were targeted that allow reliable discrimination between fetal target DNA and maternal background DNA when used for prenatal diagnostics. By focusing on markers on the long arm of chromosome 21, which are directly related to and therefore informative for the disease of interest (trisomy 21, Down syndrome) (9)(10), and analyzing all suitable markers present (114 CpG islands), Chim et al. identified multiple chromosome 21–specific DNA markers. By a combined approach, executed thoroughly, a large and complete set (n = 22) of differentially methylated fetal- and chromosome 21q–specific DNA markers was identified. Two of these markers, both unmethylated in the placenta, U-PDE9A and U-CGI137, and selected for their greatest difference in median methylation index, were validated in maternal plasma samples. For the latter gene, the method of analysis was successfully adapted to overcome the low number of differentially methylated CpG sites. The results of this investigation clearly demonstrate that markers of this nature, in which noncoding sequences represent and complement the fetal genetic barcode, can be detected reliably and specifically in the maternal plasma (2). The consequences of these findings are manifold. Although plasma is usually the primary specimen for clinical testing and contains the largest version of the human proteome (11), maternal plasma has not been useful in this way with respect to the completeness of the biological information from the fetus. At least 19.3% of additional data, not accessible by proteomic or metabolic methods, is available by molecular targeting for this class of hidden information (2). Epigenetic differences between the placenta and maternal blood cells were found on 22 of 114 CpG islands studied on chromosome 21q (2). The same gain in information can be expected for other chromosomes and for noncoding RNA, including microRNA. Recent studies show that of 345 miRNAs analyzed, 275 (80%) are expressed in the human placenta, and 53 (15%) show preferential or exclusive expression in the placenta (12). Therefore, the percentage of additional biological information on fetal well-being that becomes available by targeting noncoding RNA and noncoding DNA likely ranges between 20% and 30%. By analogy with the exploration of chromosome 21q (2), large-scale searches for epigenetic and related noncoding markers of other chromosomes will increase by 20%–30% the number and type of epigenetic biomarkers for the majority of pregnancy-associated and pregnancy-induced conditions. Such conditions include those for which there is a need for genetic testing (e.g., trisomy 13 and 18) and those associated with placental origin and placental dysfunction (such as preeclampsia). In preeclampsia, maternal transmission of the familial forms of preeclampsia appears related to epigenetic regulation of susceptibility genes (13), and thus disease-specific epigenetic signatures, besides being of diagnostic relevance, could be informative for etiological pathways involved in the condition. The development of epigenetic tests does not mean that proteomic methods will become obsolete. For example, in addition to chemical modifications of CpG dinucleotides, epigenetic modifications of DNA also consist of specific protein sets bound to noncoding DNA, a characteristic that is the rule rather than an exception. For transcription factors that function as master control switches, the number of bound target sequences can be several thousands (14). Recently, high-throughput assays, such as the ChipSeq assay, combined with solid phase amplification and sequencing have become available to map protein-DNA interactions comprehensively across the mammalian genomes (14). By this approach, cell-specific interactome signatures can be identified in a cost-effective manner. Transcription factors important for placental function (GCM1) or involved in placental dysfunction (STOX1) are excellent candidates for this approach (13)(15) and should be considered in the context of prenatal diagnostics. In conclusion, if the systematic approach followed by Chim and coworkers is applied in a genome-wide manner, or preceded by chromatin immunoprecipitation as well as extended to chromosome-specific, cell-specific or genome-wide explorations of noncoding RNA, this approach will certainly identify novel networks of noncoding DNA and RNA sequences. Such information will not only increase the number of novel fetal biomarkers for molecular diagnostics in pregnancy, but also provide mechanistic insight regarding essential placental processes that regulate and control pregnancy. Systematic explorations as described in this issue of Clinical Chemistry (2) are highly recommended and suitable for this purpose. Grant/funding Support: Supported in part by the SAFE network (Project Number: LSHB-CT-2004-503243). Financial Disclosures: None declared. References 1 Poon LL, Leung TN, Lau TK, Chow KC, Lo YM. Differential DNA methylation between fetus and mother as a strategy for detecting fetal DNA in maternal plasma. Clin Chem 2002 ; 48 : 35 -41. Crossref Search ADS PubMed 2 Chim SSC, Jin S, Lee TYH, Lun FMF, Lee WS, Chan LYS, et al. Systematic search for placental DNA methylation markers on chromosome 21: towards a maternal plasma-based epigenetic test for fetal trisomy 21. Clin Chem 2008 ; 54 : 500 -511. Crossref Search ADS PubMed 3 . The FANTOM consortiumet al. The transcriptional landscape of the mammalian genome. Science (Wash DC) 2005 ; 309 : 1559 -1563. Crossref Search ADS 4 Lee RC, Feinbaum RL, Ambros V. The C. elegans heterochronic gene lin-4 encodes small RNAs with antisense complementarity to lin-14. Cell 1993 ; 75 : 843 -854. Crossref Search ADS PubMed 5 Bartel DP. MicroRNAs: genomics, biogenesis, mechanism and function. Cell 2004 ; 116 : 281 -297. Crossref Search ADS PubMed 6 Smalheiser NR, Torvik VI. Mammalian microRNAs derived from genomic repeats. Trends Genet 2005 ; 21 : 322 -326. Crossref Search ADS PubMed 7 Hartig JV, Tomary Y, Forsteman K. piRNAs: the ancient hunters of genome invaders. Genes Dev 2007 ; 21 : 1707 -1713. Crossref Search ADS PubMed 8 Berezikov W, Chung W-J, Willis J, Cuppen E, Lai EC. Mammalian mirtron genes. Molecular Cell 2007 ; 28 : 328 -336. Crossref Search ADS PubMed 9 Oudejans CB, Go AT, Visser A, Mulders MA, Westerman BA, Blankenstein MA, van Vugt JM. Detection of chromosome 21-encoded mRNA of placental origin in maternal plasma. Clin Chem 2003 ; 49 : 1445 -1449. Crossref Search ADS PubMed 10 Lo YMD, Tsui NB, Chiu RWK, Lau TK, Leung TN, Heung MM, et al. Plasma placental allelic ratio permits noninvasive prenatal chromosomal aneuploidy detection. Nat Med 2007 ; 13 : 218 -223. Crossref Search ADS PubMed 11 Anderson NL, Anderson NG. The human plasma proteome. Mol Cell Proteomics 2002 ; 1.11 : 845 -867. 12 Liang L, Ridzon D, Wong L, Chen C. Characterization of microRNA expression profiles in normal human tissues. BMC Genomics 2007 ; 8 : 166 . Crossref Search ADS PubMed 13 van Dijk M, Mulders J, Poutsma A, Könst AAM, Lachmeijer AMA, et al. Maternal segregation of the Dutch preeclampsia locus at 10q22 with a novel member of the winged helix gene family. Nat Genet 2005 ; 37 : 514 -519. Crossref Search ADS PubMed 14 Johnson DS, Mortazavi A, Myers RM, Wold B. Genome-wide mapping of in vivo protein-DNA interactions. Science (Wash DC) 2007 ; 316 : 1497 -1502. Crossref Search ADS 15 Anson-Cartwright L, Dawson K, Holmyard D, Fisher SJ, Lazarrini RA, Cross JC. The glial cells missing-1 protein is essential for branching morphogenesis in the chorioallantoic placenta. Nat Genet 2000 ; 25 : 248 -250. © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Swerdloff, Ronald, S;Wang,, Christina

doi: 10.1373/clinchem.2007.101303pmid: 18310144

Fritz and coworkers (1) have provided data arguing against the validity of a popular analog-based direct assay for free testosterone (T). Their rigorous dissection of characteristics of the assay and use of carefully defined solutions demonstrates that the assay correlates generally with total T but does not measure dialyzable T; this observation is consistent with prior reservations about this class of assay by reported by the Endocrine Society (2)(3) and of other investigators (4)(5)(6). why worry about free t? Testosterone circulates in the blood of men and women in several forms. It is bound tightly to sex-hormone binding globulin (SHBG), loosely to albumin, and unbound to proteins (free) (7). In most, but not all, clinical conditions, a measurement of total T is adequate for the evaluation of a patient. It is widely believed that the SHBG-bound T is not readily available to most tissues, whereas albumin-bound and free T are bioavailable. Because SHBG concentrations can be influenced by many factors (e.g., decreased by obesity, testosterone treatment, and hyperandrogenic female conditions such as polycystic ovary syndrome and increased by aging, pregnancy, and estrogen therapy), there are clinical situations in which measured concentrations of total T may not reflect the bioavailable concentrations or the clinical status of the patient (3)(7)(8). In these circumstances a supplemental test assessing bioavailable or free T will be helpful in clinical decision-making. how is free t or bioavailable t determined? Free T has been estimated for some time by dialyzing serum across a semipermeable membrane under selected conditions. The T in the serum that crosses the membrane is believed to be unbound or free. Because free T accounts for only about 2%–3% of the total T in men and even less in women, direct measurement of T in the dialysate has been reported (9) but is technically difficult with most assays. To overcome this limitation of T assay sensitivity, most dialysis assays for free T use tracer amounts of radiolabeled T added to the serum, and the isotope is measured in the dialysate. The percentage of dialyzable tracer is determined and multiplied by the total T to calculate the free T (7). This method, although sensitive and reproducible, is cumbersome and not easily adaptable to automated methods. An alternative method is to estimate free T by measuring total T and SHBG and calculate a free T with an algorithm based on the law of mass action or by empirical equations (7)(10). This calculated free T gives excellent correlations with the free T measured with the dialysis method (5). A third approach is to measure bioavailable T as a free T surrogate by precipitating SHBG with ammonium sulfate and measuring the albumin-bound and free T in the supernatant (11). Although the latter approach is technically simple, care must be taken to assure the correct concentration of ammonium sulfate for complete precipitation of the SHBG-T complex. The fourth method is that studied by Fritz et al. (1). Although this method is used widely, the analog-based assay described in their report is an automated black-box package with a proprietary radioactive buffer provided by the manufacturer; this class of assay has been criticized as having poor accuracy, sensitivity, and between-assay comparability and being influenced by the dilution of serum (2)(3)(4)(5)(6). It is of note that the reference interval for free T by analog-based assay is much lower (concentrations about one-fifth as high) than that for the equilibrium dialysis assay (2). This calibration difference between assays is a major problem. what did fritz and coworkers find in their studies? These investigators dissected one analog-based free T assay by measuring total T and free T by analog methods both in the retentate and dialysate fractions of serum after equilibrium dialysis. Free T was also measured using the standard equilibrium assay method on the dialysate. The analog free T assay correlated with total T in the retentate but neither the total T assay nor analog free T assay measured any T in the dialysate, although the concentrations were in the reference range by equilibrium dialysis. Fritz et al. then used the same method on the sample while covarying concentrations of SHBG, protein-bound T, and total T through dilution of serum. The results showed that the analog-based assay tracked total T more closely than free T by the dialysis method. In the third experiment they diluted the serum while adding increased amounts of protein-free T to keep the total T in the serum constant. The results showed that the total T and free T by the analog method remained constant while the free T by the dialysis method increased as expected. Thus their studies showed that free T measurement with the analog-based free T assay has no advantage over total T measurements and does not measure unbound T. Fritz et al. conclude that the assay is not properly calibrated, lacks specificity, and does not measure free T. what does the analog assay measure? It is not entirely clear what constituent(s) are being measured, but free T measured by the analog-based assay correlates with total T. Fritz et al. (1) speculated that the assay nonspecificity may be due to protein-T complexes binding to the T antibody, leading to a 3-way competition between free testosterone, testosterone complexes, and testosterone conjugates (analogs) for binding to the same antibody. This explanation should be testable using liquid chromatography–tandem mass spectrometry (12). do the findings of fritz and coworkers apply to other branded analog-based free t assays? The authors are cautious, but given the information provided, other free T analog assays probably show similar characteristics of tracking total T and not free T; thus providing no advantage over total T measurements. why are these analog-based free t assays so widely used? The wide use of these assays is based in part on their ease of performance, automated nature, and lower cost than free T measurement by dialysis. The popularity of these assays may also be a result of lack of specific knowledge of physicians regarding the limitations of the assay methodology. Because the free T by analog method tracks total T, and because most men with total T concentrations that are very low do not require a free T to confirm the diagnosis of male hypogonadism, the results seem to fit the clinical impression. Because the experts in the field recommend free T measurements in equivocal cases of hypogonadism, the practicing physician often orders the correct test but the wrong assay. should the analog-based free t assay be used in clinical practice? Because the analog-based assay does not measure free T, it makes no sense to label it as such. Because it tracks total T but has not been shown to add to the values of total T results, we do not recommend its use. are the current methods of total t assays adequate? It is important to note that free T measured by equilibrium dialysis requires a sensitive, specific, precise, and accurate assay for total T. Recent studies have shown that the current methods of measurement of total T lacked the required sensitivity for samples with very low concentrations, as occur in severely testosterone deficient men, normal or T-deficient women, and children; this sensitivity limitation can be corrected by newer methods based on mass spectrometry (12)(13)(14). Furthermore, the free T analog–based assay may be only one of a number of widely used but poorly assessed and quality controlled hormone assays (2)(15). The CDC has provided quality control services for several analytes, including recent efforts to do so for total T by liquid chromatography–tandem mass spectrometry. It is time that physicians, investigators, clinical chemists, pathologists, and editors of medical journals insist on better surveillance of assays used in research and patient care. References 1 Fritz KS, McKean AJS, Nelson JC, Wilcox RB. Analog-based free testosterone test results linked to total testosterone concentrations, not free testosterone concentrations. Clin Chem 2008 ; 54 : 512 -516. Crossref Search ADS PubMed 2 Rosner W, Auchus RJ, Azziz R, Sluss PM, Raff H. Utility, limitations, and pitfalls in measuring testosterone: an Endocrine Society Position Statement. J Clin Endocrinol Metab 2007 ; 92 : 405 -413. Crossref Search ADS PubMed 3 Bhasin S, Cunningham GR, Hayes FJ, Matsumoto AM, Snyder PJ, Swerdloff RS, Montori VM. Testosterone therapy in adult men with androgen deficiency syndromes: an Endocrine Society Clinical Practice Guideline. J Clin Endocrinol Metab 2006 ; 91 : 1995 -2010. Crossref Search ADS PubMed 4 Van UK, Stockl D, Kaufman JM, Fiers T, De LA, Thienpont LM. Validation of 5 routine assays for serum free testosterone with a candidate reference measurement procedure based on ultrafiltration and isotope dilution-gas chromatography-mass spectrometry. Clin Biochem 2005 ; 38 : 253 -261. Crossref Search ADS PubMed 5 Miller KK, Rosner W, Lee H, Hier J, Sesmilo G, Schoenfeld D, et al. Measurement of free testosterone in normal women and women with androgen deficiency: comparison of methods. J Clin Endocrinol Metab 2004 ; 89 : 525 -533. Crossref Search ADS PubMed 6 Winters SJ, Kelley DE, Goodpaster B. The analog free testosterone assay: are the results in men clinically useful?. Clin Chem 1998 ; 44 : 2178 -2182. Crossref Search ADS PubMed 7 Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 1999 ; 84 : 3666 -3672. Crossref Search ADS PubMed 8 Chang WY, Knochenhauer ES, Bartolucci AA, Azziz R. Phenotypic spectrum of polycystic ovary syndrome: clinical and biochemical characterization of the three major clinical subgroups. Fertil Steril 2005 ; 83 : 1717 -1723. Crossref Search ADS PubMed 9 Sinha-Hikim I, Arver S, Beall G, Shen R, Guerrero M, Sattler F, et al. The use of a sensitive equilibrium dialysis method for the measurement of free testosterone levels in healthy, cycling women and in human immunodeficiency virus-infected women. J Clin Endocrinol Metab 1998 ; 83 : 1312 -1318. PubMed 10 Sodergard R, Backstrom T, Shanbhag V, Carstensen H. Calculation of free and bound fractions of testosterone and estradiol-17 beta to human plasma proteins at body temperature. J Steroid Biochem 1982 ; 16 : 801 -810. Crossref Search ADS PubMed 11 Morley JE, Patrick P, Perry HM. III. Evaluation of assays available to measure free testosterone. Metabolism 2002 ; 51 : 554 -559. Crossref Search ADS PubMed 12 Wang C, Catlin DH, Demers LM, Starcevic B, Swerdloff RS. Measurement of total serum testosterone in adult men: comparison of current laboratory methods versus liquid chromatography-tandem mass spectrometry. J Clin Endocrinol Metab 2004 ; 89 : 534 -543. Crossref Search ADS PubMed 13 Sikaris K, McLachlan RI, Kazlauskas R, de KD, Holden CA, Handelsman DJ. Reproductive hormone reference intervals for healthy fertile young men: evaluation of automated platform assays. J Clin Endocrinol Metab 2005 ; 90 : 5928 -5936. Crossref Search ADS PubMed 14 Taieb J, Mathian B, Millot F, Patricot MC, Mathieu E, Queyrel N, et al. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography-mass spectrometry in sera from 116 men, women, and children. Clin Chem 2003 ; 49 : 1381 -1395. Crossref Search ADS PubMed 15 Matsumoto AM, Bremner WJ. Serum testosterone assays: accuracy matters. J Clin Endocrinol Metab 2004 ; 89 : 520 -524. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Dennis Lo, Y, M;Chiu, Rossa W, K

doi: 10.1373/clinchem.2007.100016pmid: 18202154

Abstract Background: The discovery of circulating cell-free fetal nucleic acids in maternal plasma has opened up new possibilities for noninvasive prenatal diagnosis. The potential application of this technology for the noninvasive prenatal detection of fetal chromosomal aneuploidies is an aspect of this field that is being actively investigated. The main challenge of work in this area is the fact that cell-free fetal nucleic acids represent only a minor fraction of the total nucleic acids in maternal plasma. Methods and Results: We performed a review of the literature, which revealed that investigators have applied methods based on the physical and molecular enrichment of fetal nucleic acid targets from maternal plasma. The former includes the use of size fractionation of plasma DNA and the use of the controversial formaldehyde treatment method. The latter has been achieved through the development of fetal epigenetic and fetal RNA markers. The aneuploidy status of the fetus has been explored through the use of allelic ratio analysis of plasma fetal epigenetic and RNA markers. Digital PCR has been shown to offer high precision for allelic ratio and relative chromosome dosage analyses. Conclusions: After a decade of work, the theoretical and practical feasibility of prenatal fetal chromosomal aneuploidy detection by plasma nucleic acid analysis has been demonstrated in studies using small sample sets. Larger scale independent studies will be needed to validate these initial observations. If these larger scale studies prove successful, it is expected that with further development of new fetal DNA/RNA markers and new analytical methods, molecular noninvasive prenatal diagnosis of the major chromosomal aneuploidies could become a routine practice in the near future. Prenatal diagnosis is now an established part of modern obstetrics care. The detection of fetal chromosomal aneuploidies is a major reason why many pregnant women go for prenatal diagnosis. Many conventional methods for prenatal diagnosis, however, require obtaining fetal materials for analysis through procedures such as amniocentesis and chorionic villus sampling. These methods are invasive and constitute a finite risk to the fetus. To stratify pregnant women according to their risk of carrying a fetus affected by chromosomal aneuploidy, several screening methods have been developed, including ultrasonography and maternal serum biochemistry (1). These methods, however, are targeted at epiphenomena associated with the chromosomal aneuploidies, rather than at the core molecular abnormalities, and have limited sensitivities and specificities, with strictly defined gestational age windows that must be used for specific tests. To circumvent such limitations, there is a need for the development of a new generation of noninvasive tests that target the core molecular pathology of such fetal chromosomal abnormalities. The discovery of circulating cell-free fetal DNA in maternal plasma in 1997 has offered a new approach for noninvasive prenatal diagnosis (1)(2). Although this method can be readily applied for the detection of unique fetal genetic targets, e.g., the Y chromosome from a male fetus (3) and the Rh blood group, D antigen (RHD)1 gene of an RhD-positive fetus (4), the development of tests enabling the use of this approach for fetal chromosomal aneuploidies has been challenging. One fundamental technical challenge is a consequence of the fact that between weeks 11 and 17 weeks of gestation, fetal DNA constitutes only a mean of approximately 3% of total cell-free DNA in maternal plasma (5), with the obvious implication that most of the DNA in the plasma is maternal in origin. Thus, most molecular methods for aneuploidy detection, if applied to maternal plasma, would essentially be measuring the chromosome status of the mother. In addition, because cell-free nucleic acids are circulating in an extracellular form in maternal plasma, all of the cell-based methods for aneuploidy detection, e.g., fluorescence in situ hybridization, would not be applicable. This review summarizes the current developments in the use of circulating cell-free nucleic acids in maternal plasma for the noninvasive prenatal detection of fetal chromosomal aneuploidies. approaches to solve the problems associated with the low fractional concentration of fetal nucleic acids The fractional concentration of cell-free fetal DNA in maternal plasma is determined by the ratio of the absolute concentration of cell-free fetal DNA to the absolute concentration of total (maternal and fetal) cell-free DNA. Thus, the fractional concentration of cell-free fetal DNA can be increased through the selective enrichment of fetal DNA, or through the suppression of the background maternal DNA. A schematic illustration of the major approaches is shown in Fig. 1 . Figure 1. Open in new tabDownload slide Schematic illustration of the major approaches to increasing the fractional concentration of cell-free fetal DNA. The fractional concentration of fetal DNA in maternal plasma is given by the ratio of the absolute concentration of cell-free fetal DNA to that of the total cell-free DNA in maternal plasma (A). Approaches to increase the fractional concentration of fetal nucleic acids in maternal plasma may involve selective enrichment of fetal DNA (B), suppression of maternal background DNA (C), or elimination of the maternal nucleic acid background by detecting nucleic acids that are virtually completely fetus specific, such as fetal epigenetic or fetal RNA markers (D). Figure 1. Open in new tabDownload slide Schematic illustration of the major approaches to increasing the fractional concentration of cell-free fetal DNA. The fractional concentration of fetal DNA in maternal plasma is given by the ratio of the absolute concentration of cell-free fetal DNA to that of the total cell-free DNA in maternal plasma (A). Approaches to increase the fractional concentration of fetal nucleic acids in maternal plasma may involve selective enrichment of fetal DNA (B), suppression of maternal background DNA (C), or elimination of the maternal nucleic acid background by detecting nucleic acids that are virtually completely fetus specific, such as fetal epigenetic or fetal RNA markers (D). size fractionation of plasma dna Selective enrichment of fetal DNA requires the targeting of fetal DNA characteristics that are different from those of maternal DNA in maternal plasma. Chan et al. have demonstrated that circulating fetal DNA molecules are generally shorter than the circulating maternal DNA molecules (6). Li et al. have shown that enrichment of fetal DNA can be achieved by the selective targeting of the shorter plasma DNA molecules (7). This method has been applied to enhance the detection of paternally inherited fetal mutations in maternal plasma (8). However, whether the degree of enrichment achieved might allow the direct detection of fetal chromosomal aneuploidies is currently unknown. Furthermore, the reported method for such size fractionation is based on the size separation of plasma DNA using agarose gel electrophoresis, followed by the extraction of DNA from manually cut agarose gel slices containing different size fractions (7). Apart from the labor-intensiveness of this procedure, it is prone to contamination. Thus, new and potentially automatable methods for the size fractionation of plasma DNA must be developed before this approach can become practical for noninvasive prenatal diagnosis. suppression of the maternal dna background With regard to the suppression of the maternal DNA background, Lui et al. have demonstrated in a sex-mismatched bone marrow transplantation model that most of the DNA in plasma is hematopoietic in origin (9). It has been postulated that maternal hematopoietic cells might also be the major cell types contributing to the maternal nucleic acid background in maternal plasma (10)(11). Dhallan et al. hypothesized that the addition of formaldehyde might result in the stabilization of the maternal leukocytes following venesection, thus reducing the liberation of maternal DNA by such cells into the plasma, a process that might result in the dilution of the fetal DNA in maternal plasma (12). Although the initial data presented by Dhallan et al. were impressive (12), these results have not been reproduced by a number of independent groups (13)(14). One possible explanation for these discrepant results is the use by Dhallan et al. of an imprecise analytical method that might overestimate the fractional fetal DNA concentration in a proportion of formaldehyde-treated samples (15). Nonetheless, it would be interesting to test preservatives other than formaldehyde for their ability to suppress the maternal DNA background. molecular enrichment strategies: fetal epigenetic markers and fetal rna markers Another approach to address the low fractional concentration of fetal DNA is the targeting of selected subsets of nucleic acids in maternal plasma that are virtually completely fetus specific. One approach is to identify loci that exhibit fetus-specific epigenetic signatures. Epigenetics is a field that studies molecular processes that affect gene expression without altering DNA sequences (16). One of the best studied epigenetic processes is DNA methylation. In 2002, Poon et al. postulated that loci exhibiting differential DNA methylation patterns between fetal and maternal tissues might be used to develop fetal epigenetic markers for detection in maternal plasma (17). Chim et al. have reported that the SERPINB5 gene, coding for maspin, is hypomethylated in placental tissues and hypermethylated in maternal blood cells (11). Because the placenta is likely to be the major source of fetal DNA in maternal plasma (18)(19), and as discussed above, the maternal hematopoietic cells are likely to be a major source of maternal DNA in maternal plasma, hypomethylated SERPINB5 sequences may serve as a marker for placental DNA in maternal plasma. The feasibility of this epigenetic approach for noninvasive prenatal diagnosis has been demonstrated by the good correlation between the concentrations of hypomethylated SERPINB5 sequences and SRY sequences from male fetuses in maternal plasma, and the clearance of SERPINB5 sequences from maternal plasma following delivery (11). The location of SERPINB5 on chromosome 18 has also provided a valuable opportunity to test the application of this epigenetic approach for the prenatal detection of fetal chromosomal aneuploidy, using trisomy 18 as a model system (see later sections) (20). Since the development of the SERPINB5 marker, many other fetal epigenetic markers suitable for detection in maternal plasma have been described, including the RASSF1A gene on chromosome 3 (21)(22) and numerous markers on chromosome 21 (23)(24). Through the targeting of such fetus-specific DNA methylation markers in maternal plasma, the detected signal is virtually completely fetal in origin. Thus, the number of fetal chromosomes on which the epigenetic marker is located can be ascertained (see later sections). From the systematic survey of chromosome 21 (24), it appears that markers informative for noninvasive prenatal diagnosis are relatively plentiful. The main limitation of this DNA methylation approach is that many of the commonly used methods for detecting DNA methylation markers, e.g., methylation-specific PCR (25), are based on bisulfite conversion. Bisulfite conversion has been shown to result in a massive degradation of the input DNA (26). This undesirable characteristic is detrimental for the detection of circulating fetal DNA markers in which a relatively limited number of target molecules are present in a particular sample. In this regard, markers that are hypermethylated in the placenta and hypomethylated in maternal blood cells are particularly valuable because methylation-sensitive restriction enzymes that cut hypomethylated sequences but leave hypermethylated sequences intact can be used for the selective destruction of the maternal sequences in maternal plasma (22). The extension of such a restriction enzyme–based approach to markers exhibiting a reverse pattern of differential DNA methylation, namely hypomethylated in the placenta and hypermethylated in maternal blood cells, will require methods that allow the selective detection of the restricted (i.e., hypomethylated) sequences. One such approach has recently been described, which is based on the use of a stem-loop primer (27). Another possible approach is through the development of new enzymes that would selectively restrict methylated (i.e., maternal) DNA, while leaving hypomethylated (i.e., fetal) DNA intact. RNA molecules represent another nucleic acid subset in maternal plasma that can be targeted for fetal-specific molecules. Fetal RNA was first detected in maternal plasma in 2000 (28). This finding was followed by work demonstrating the extraordinary stability of plasma RNA molecules (29), possibly through their association with particulate matter (30), a phenomenon that has been postulated to protect plasma RNA against degradation by plasma RNases (31). One important development is the demonstration that placental mRNA represents an important source of fetal RNA in maternal plasma (32). This latter discovery and the above-mentioned realization that maternal hematopoietic cells are likely to be the major contributor of maternal-derived nucleic acids in maternal plasma have led to the development of a systematic microarray-based approach for the identification of new placental mRNA markers suitable for maternal plasma detection (10). This series of developments has led to the demonstration that mRNA from genes located on chromosome 21 can be detected in maternal plasma (33)(34). One such gene for which mRNA in maternal plasma has been shown to be completely fetal-specific is placenta-specific 4 (PLAC4) (34). Thus, with the identification of fetus-specific mRNA markers in maternal plasma from a chromosome involved in a chromosomal aneuploidy, all that remains would be to develop a technology for obtaining chromosome copy number information from such a marker (see later sections). One key advantage of plasma RNA markers is that there is an intrinsic amplification process in which a gene is transcribed into multiple mRNA copies. Another advantage is that plasma RNA markers can be readily detected by reverse transcriptase PCR or other amplification technologies. The main disadvantage is that in many reported procedures TRIZOL-treated plasma was used to stabilize plasma RNA during storage (34), and thus archival plasma samples not been treated in this manner might not be suitable for plasma RNA analysis. approaches for determining fetal chromosome dose in cell-free nucleic acids in maternal plasma Although physical or molecular approaches for fetal nucleic acid enrichment may render the fetal proportion of nucleic acids to be more readily detectable in maternal plasma, methods are needed to allow assessment of the number of potentially aneuploid chromosomes in the fetal genome. Previous studies have shown that the fetal DNA concentration in maternal plasma is increased in pregnancies with certain fetal aneuploidies, such as trisomy 21 (35). Large interindividual variations in maternal plasma fetal DNA and PLAC4 mRNA concentrations, however, have precluded the use of their mere quantification as a robust means for identifying fetal aneuploidies (34). Thus, strategies that allow the objective determination of fetal chromosome dose are needed, and the main approaches are shown in Fig. 2 . Figure 2. Open in new tabDownload slide Approaches for determining fetal chromosome dosage in cell-free nucleic acids in maternal plasma using trisomy 21 as an example. (A), allelic ratio analysis involves the assessment of the ratio between alleles at a heterozygous locus located on chromosome 21. The allelic ratio for such a locus would be expected to be 1:1 for a euploid fetus but 2:1 or 1:2 for a trisomic fetus. (B), relative chromosome dosage analysis involves the assessment of the ratio between a chromosome 21 locus and a nonchromosome 21 reference locus. The ratio among fetal-derived nucleic acid molecules would be expected to be 2:2 for a euploid fetus but 3:2 for a trisomic fetus. Figure 2. Open in new tabDownload slide Approaches for determining fetal chromosome dosage in cell-free nucleic acids in maternal plasma using trisomy 21 as an example. (A), allelic ratio analysis involves the assessment of the ratio between alleles at a heterozygous locus located on chromosome 21. The allelic ratio for such a locus would be expected to be 1:1 for a euploid fetus but 2:1 or 1:2 for a trisomic fetus. (B), relative chromosome dosage analysis involves the assessment of the ratio between a chromosome 21 locus and a nonchromosome 21 reference locus. The ratio among fetal-derived nucleic acid molecules would be expected to be 2:2 for a euploid fetus but 3:2 for a trisomic fetus. allelic ratio analysis One approach for determining fetal chromosome dose is via the analysis of allelic ratio of genetic variations at the detected locus. This approach can be used only if the fetus is heterozygous at the detected locus. This approach can be applied in combination with the approaches described in the previous section, when the detected molecules are sufficiently enriched for fetal-derived targets. The simplest illustration of this concept is the molecular targeting of fetus-specific targets using DNA methylation markers or placental mRNA markers. For DNA methylation markers, the first demonstration is the use of hypomethylated SERPINB5 sequences as a fetus-specific target on chromosome 18 (11). Through the determination of the allelic ratio of a single nucleotide polymorphism (SNP) in the hypomethylated SERPINB5 promoter in heterozygous fetuses, trisomy 18 can be detected (20). This approach is called the epigenetic allelic ratio approach. With the recognition of fetal DNA methylation markers on chromosome 21 (24), this approach can potentially be applied for the noninvasive prenatal detection of trisomy 21. For RNA markers, the first demonstrated application of the allelic ratio approach was for an SNP in the expressed region of PLAC4 (34). This approach is called the RNA-SNP approach. This first series using the RNA-SNP approach established a sensitivity of 90% and a specificity of 96.5% for the detection of fetal trisomy 21 from maternal plasma (34). These results suggest that for informative (i.e., heterozygous) cases, the RNA-SNP approach is the most accurate single-marker approach for the noninvasive prenatal detection of fetal trisomy 21. Approaches for improving the diagnostic sensitivity and specificity of the RNA-SNP approach may be on the horizon. It has been shown that the robust measurement of the RNA-SNP allelic ratio requires at least 1000 molecules per reaction (34). It has recently been shown that through the use of digital PCR, it is possible to reduce the number of input molecules (36). Digital PCR is a method for molecular analysis in which multiple PCRs are carried out on a sample diluted to a concentration such that on average each reaction will contain ≤1 target molecule (37). In this manner, a proportion of the reactions will be negative, owing to the absence of the target molecule in a given reaction. For the subset of the reactions giving a positive detection result, most of the reactions will be positive because of the presence of a single target molecule. By use of this method, the number of target molecules can be accurately counted. In the context of the digital version of RNA-SNP allelic ratio analysis, the number of each RNA allele will be counted. Fetal aneuploidy can thus be detected through the use of appropriate statistical analysis to determine the probability that one of the alleles is overrepresented, i.e., when trisomy is present (36). With this digital PCR-based approach fetal aneuploidy status can be determined very accurately with just a few hundred molecules, opening the possibility for the application of this assay method even very early on in gestation, at a time when the circulating PLAC4 mRNA concentration in maternal plasma is relatively low (34). The main disadvantage of digital PCR-based techniques is their requirement for the performance of multiple PCR analyses per tested sample, a process that is tedious if carried out manually. However, with the development of automated strategies for digital PCR analysis, such as microfluidics (38), variants of emulsion PCR (39)(40), and various approaches for massively parallel sequence analysis (41), it is likely that such a system will eventually be practical for use in routine diagnosis. relative chromosome dosage analysis The main disadvantage of the allelic ratio–based approach is the requirement that the fetus be heterozygous for the analyzed genetic polymorphisms. To overcome the requirement for heterozygosity, methods that directly measure the relative dose of different chromosomes are required. However, conventional methods for direct measurement of chromosome dose [e.g., real-time PCR (42) and paralogous sequence quantification (43)] cannot be directly used, because of the low fractional concentration of circulating fetal DNA. One potential solution to this limitation is to combine such conventional methods for relative chromosome dose analysis with methods for the physical [e.g., size fractionation of plasma DNA (7) and suppression of maternal DNA background (44)] or molecular enrichment of fetal DNA [e.g., fetus-specific DNA methylation (23)(24)] as discussed earlier in this report. An alternative solution is the development of analytical strategies that can perform relative chromosome analysis despite the low fractional concentration of circulating fetal DNA in maternal plasma. In this regard, the digital PCR-based approach, which has previously been discussed in this review in the context of allelic ratio analysis, can also be applied for chromosome dose analysis, the so-called digital relative chromosome dose approach (36). In this method, digital PCR analysis is carried out for 1 or more targets located on a chromosome involved in a trisomy (e.g., chromosome 21) and for 1 or more targets located on a reference chromosome not involved in the trisomy. Through the use of appropriate statistical procedures for testing the presence of overrepresentation of reactions involving the potentially aneuploid chromosome, digital relative chromosome dose measurement has been demonstrated to possess the discrimination power to detect the presence of DNA from a trisomic fetus, even if the trisomic DNA constitutes only some 10% of the tested sample (36). It is envisioned that the digital relative chromosome dosage approach can also be combined with the physical (7)(12) and molecular methods (11)(23)(24) for fetal DNA enrichment described above. The superior discrimination power of digital PCR can be expected to allow an aneuploidy detection system to be based on a degree of fetal DNA enrichment that would otherwise be insufficient for conventional nondigital PCR systems. Conclusions Following a decade of development, it has been demonstrated that noninvasive prenatal detection of fetal chromosomal aneuploidies can be achieved by analysis of cell-free fetal nucleic acids in maternal plasma. Thus far, feasibility studies have been carried out using a relatively small sample set. Over the next few years, larger-scale independent studies are needed to validate these initial observations. Further development of new fetal DNA/RNA markers and new analytical methods is expected to allow molecular noninvasive prenatal diagnosis of the major chromosomal aneuploidies to become a routinely practiced reality in the near future. Such a development would ultimately make prenatal testing safer and less stressful for the millions of pregnant women each year who undergo such testing. Grant/funding Support: The authors are supported by an Area of Excellence Grant from the University Grants Committee of Hong Kong, the Hong Kong Research Grants Council, the Innovation and Technology Fund, and the Li Ka Shing Foundation. Financial Disclosures: The authors hold patents or have filed patent applications on aspects of prenatal diagnosis based on plasma nucleic acids. Y.M.D.L. is a consultant to Sequenom, Inc. Sequenom has licensed intellectual property in noninvasive prenatal diagnosis. 1 " Human genes: RHD, Rh blood group, D antigen; placenta-specific 4, PLAC4. References 1 Lo YMD, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, Wainscoat JS. Presence of fetal DNA in maternal plasma and serum. Lancet 1997 ; 350 : 485 -487. Crossref Search ADS PubMed 2 Lo YMD, Chiu RWK. Prenatal diagnosis: progress through plasma nucleic acids. Nat Rev Genet 2007 ; 8 : 71 -77. Crossref Search ADS PubMed 3 Costa JM, Benachi A, Gautier E. New strategy for prenatal diagnosis of X-linked disorders. N Engl J Med 2002 ; 346 : 1502 . Crossref Search ADS PubMed 4 Lo YMD, Hjelm NM, Fidler C, Sargent IL, Murphy MF, Chamberlain PF, et al. Prenatal diagnosis of fetal RhD status by molecular analysis of maternal plasma. N Engl J Med 1998 ; 339 : 1734 -1738. Crossref Search ADS PubMed 5 Lo YMD, Tein MS, Lau TK, Haines CJ, Leung TN, Poon PM, et al. Quantitative analysis of fetal DNA in maternal plasma and serum: implications for noninvasive prenatal diagnosis. Am J Hum Genet 1998 ; 62 : 768 -775. Crossref Search ADS PubMed 6 Chan KCA, Zhang J, Hui AB, Wong N, Lau TK, Leung TN, et al. Size distributions of maternal and fetal DNA in maternal plasma. Clin Chem 2004 ; 50 : 88 -92. Crossref Search ADS PubMed 7 Li Y, Zimmermann B, Rusterholz C, Kang A, Holzgreve W, Hahn S. Size separation of circulatory DNA in maternal plasma permits ready detection of fetal DNA polymorphisms. Clin Chem 2004 ; 50 : 1002 -1011. Crossref Search ADS PubMed 8 Li Y, Di Naro E, Vitucci A, Zimmermann B, Holzgreve W, Hahn S. Detection of paternally inherited fetal point mutations for beta-thalassemia using size-fractionated cell-free DNA in maternal plasma. JAMA 2005 ; 293 : 843 -849. Crossref Search ADS PubMed 9 Lui YYN, Chik KW, Chiu RWK, Ho CY, Lam CW, Lo YMD. Predominant hematopoietic origin of cell-free DNA in plasma and serum after sex-mismatched bone marrow transplantation. Clin Chem 2002 ; 48 : 421 -427. Crossref Search ADS PubMed 10 Tsui NBY, Chim SSC, Chiu RWK, Lau TK, Ng EKO, Leung TN, et al. Systematic microarray-based identification of placental mRNA in maternal plasma: towards non-invasive prenatal gene expression profiling. J Med Genet 2004 ; 41 : 461 -467. Crossref Search ADS PubMed 11 Chim SSC, Tong YK, Chiu RWK, Lau TK, Leung TN, Chan LYS, et al. Detection of the placental epigenetic signature of the maspin gene in maternal plasma. Proc Natl Acad Sci U S A 2005 ; 102 : 14753 -14758. Crossref Search ADS PubMed 12 Dhallan R, Au WC, Mattagajasingh S, Emche S, Bayliss P, Damewood M, et al. Methods to increase the percentage of free fetal DNA recovered from the maternal circulation. JAMA 2004 ; 291 : 1114 -1119. Crossref Search ADS PubMed 13 Chung GTY, Chiu RWK, Chan KCA, Lau TK, Leung TN, Lo YMD. Lack of dramatic enrichment of fetal DNA in maternal plasma by formaldehyde treatment. Clin Chem 2005 ; 51 : 655 -658. Crossref Search ADS PubMed 14 Chinnapapagari SK, Holzgreve W, Lapaire O, Zimmermann B, Hahn S. Treatment of maternal blood samples with formaldehyde does not alter the proportion of circulatory fetal nucleic acids (DNA and mRNA) in maternal plasma. Clin Chem 2005 ; 51 : 652 -655. Crossref Search ADS PubMed 15 Chung GTY, Chiu RWK, Chan KCA, Lau TK, Leung TN, Chan LW, Lo YMD. Detrimental effect of formaldehyde on plasma RNA detection. Clin Chem 2005 ; 51 : 1074 -1076. PubMed 16 Jones PA, Baylin SB. The fundamental role of epigenetic events in cancer. Nat Rev Genet 2002 ; 3 : 415 -428. Crossref Search ADS PubMed 17 Poon LLM, Leung TN, Lau TK, Chow KCK, Lo YMD. Differential DNA methylation between fetus and mother as a strategy for detecting fetal DNA in maternal plasma. Clin Chem 2002 ; 48 : 35 -41. Crossref Search ADS PubMed 18 Flori E, Doray B, Gautier E, Kohler M, Ernault P, Flori J, Costa JM. Circulating cell-free fetal DNA in maternal serum appears to originate from cyto- and syncytio-trophoblastic cells: case report. Hum Reprod 2004 ; 19 : 723 -724. Crossref Search ADS PubMed 19 Alberry M, Maddocks D, Jones M, Abdel Hadi M, Abdel-Fattah S, Avent N, Soothill PW. Free fetal DNA in maternal plasma in anembrynoic pregnancies: confirmation that the origin is the trophoblast. Prenat Diagn 2007 ; 27 : 415 -418. Crossref Search ADS PubMed 20 Tong YK, Ding C, Chiu RWK, Gerovassili A, Chim SSC, Leung TY, et al. Noninvasive prenatal detection of fetal trisomy 18 by epigenetic allelic ratio analysis in maternal plasma: theoretical and empirical considerations. Clin Chem 2006 ; 52 : 2194 -2202. Crossref Search ADS PubMed 21 Chiu RWK, Chim SSC, Wong IHN, Wong CS, Lee WS, To KF, et al. Hypermethylation of RASSF1A in human and rhesus placentas. Am J Pathol 2007 ; 170 : 941 -950. Crossref Search ADS PubMed 22 Chan KCA, Ding C, Gerovassili A, Yeung SW, Chiu RWK, Leung TN, et al. Hypermethylated RASSF1A in maternal plasma: a universal fetal DNA marker that improves the reliability of noninvasive prenatal diagnosis. Clin Chem 2006 ; 52 : 2211 -2218. Crossref Search ADS PubMed 23 Old RW, Crea F, Puszyk W, Hulten MA. Candidate epigenetic biomarkers for non-invasive prenatal diagnosis of Down syndrome. Reprod Biomed Online 2007 ; 15 : 227 -235. Crossref Search ADS PubMed 24 Chim SSC, Jin S, Lee TYH, Lun FMF, Lee WS, Chan LYS, et al. Systematic search for placental DNA-methylation markers on chromosome 21: toward a maternal plasma-based epigenetic test for fetal trisomy 21. Clin Chem 2008 ; 54 : 200 -211. 25 Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci U S A 1996 ; 93 : 9821 -9826. Crossref Search ADS PubMed 26 Grunau C, Clark SJ, Rosenthal A. Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucleic Acids Res 2001 ; 29 : E65 -E65. Crossref Search ADS PubMed 27 Tong YK, Chiu RWK, Leung TY, Ding C, Lau TK, Leung TN, Lo YMD. Detection of restriction enzyme-digested target DNA by PCR amplification using a stem-loop primer: application to the detection of hypomethylated fetal DNA in maternal plasma. Clin Chem 2007 ;: 53 . 28 Poon LLM, Leung TN, Lau TK, Lo YMD. Presence of fetal RNA in maternal plasma. Clin Chem 2000 ; 46 : 1832 -1834. Crossref Search ADS PubMed 29 Tsui NBY, Ng EKO, Lo YMD. Stability of endogenous and added RNA in blood specimens, serum, and plasma. Clin Chem 2002 ; 48 : 1647 -1653. Crossref Search ADS PubMed 30 Ng EKO, Tsui NBY, Lam NY, Chiu RWK, Yu SC, Wong SC, et al. Presence of filterable and nonfilterable mRNA in the plasma of cancer patients and healthy individuals. Clin Chem 2002 ; 48 : 1212 -1217. Crossref Search ADS PubMed 31 Hasselmann DO, Rappl G, Tilgen W, Reinhold U. Extracellular tyrosinase mRNA within apoptotic bodies is protected from degradation in human serum. Clin Chem 2001 ; 47 : 1488 -1489. Crossref Search ADS PubMed 32 Ng EKO, Tsui NBY, Lau TK, Leung TN, Chiu RWK, Panesar NS, et al. mRNA of placental origin is readily detectable in maternal plasma. Proc Natl Acad Sci USA 2003 ; 100 : 4748 -4753. Crossref Search ADS PubMed 33 Oudejans CB, Go AT, Visser A, Mulders MA, Westerman BA, Blankenstein MA, van Vugt JM. Detection of chromosome 21-encoded mRNA of placental origin in maternal plasma. Clin Chem 2003 ; 49 : 1445 -1449. Crossref Search ADS PubMed 34 Lo YMD, Tsui NBY, Chiu RWK, Lau TK, Leung TN, Heung MMS, et al. Plasma placental RNA allelic ratio permits noninvasive prenatal chromosomal aneuploidy detection. Nat Med 2007 ; 13 : 218 -223. Crossref Search ADS PubMed 35 Lo YMD, Lau TK, Zhang J, Leung TN, Chang AM, Hjelm NM, et al. Increased fetal DNA concentrations in the plasma of pregnant women carrying fetuses with trisomy 21. Clin Chem 1999 ; 45 : 1747 -1751. Crossref Search ADS PubMed 36 Lo YMD, Lun FMF, Chan KCA, Tsui NBY, Chong KC, Lau TK, et al. Digital PCR for the molecular detection of fetal chromosomal aneuploidy. Proc Natl Acad Sci U S A 2007 ; 104 : 13116 -13121. Crossref Search ADS PubMed 37 Vogelstein B, Kinzler KW. Digital PCR. Proc Natl Acad Sci U S A 1999 ; 96 : 9236 -9241. Crossref Search ADS PubMed 38 Warren L, Bryder D, Weissman IL, Quake SR. Transcription factor profiling in individual hematopoietic progenitors by digital RT-PCR. Proc Natl Acad Sci U S A 2006 ; 103 : 17807 -17812. Crossref Search ADS PubMed 39 Dressman D, Yan H, Traverso G, Kinzler KW, Vogelstein B. Transforming single DNA molecules into fluorescent magnetic particles for detection and enumeration of genetic variations. Proc Natl Acad Sci U S A 2003 ; 100 : 8817 -8822. Crossref Search ADS PubMed 40 Diehl F, Li M, Dressman D, He Y, Shen D, Szabo S, et al. Detection and quantification of mutations in the plasma of patients with colorectal tumors. Proc Natl Acad Sci U S A 2005 ; 102 : 16368 -16373. Crossref Search ADS PubMed 41 Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, et al. Genome sequencing in microfabricated high-density picolitre reactors. Nature (Lond) 2005 ; 437 : 376 -380. Crossref Search ADS 42 Zimmermann B, Holzgreve W, Wenzel F, Hahn S. Novel real-time quantitative PCR test for trisomy 21. Clin Chem 2002 ; 48 : 362 -363. Crossref Search ADS PubMed 43 Deutsch S, Choudhury U, Merla G, Howald C, Sylvan A, Antonarakis SE. Detection of aneuploidies by paralogous sequence quantification. J Med Genet 2004 ; 41 : 908 -915. Crossref Search ADS PubMed 44 Dhallan R, Guo X, Emche S, Damewood M, Bayliss P, Cronin M, et al. A non-invasive test for prenatal diagnosis based on fetal DNA present in maternal blood: a preliminary study. Lancet 2007 ; 369 : 474 -481. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Talmud, Philippa, J;Cooper, Jackie, A;Palmen,, Jutta;Lovering,, Ruth;Drenos,, Fotios;Hingorani, Aroon, D;Humphries, Steve, E

doi: 10.1373/clinchem.2007.095489pmid: 18250146

Abstract Background: We investigated whether chromosome 9p21.3 single-nucleotide polymorphisms (SNPs), identified in coronary heart disease (CHD) genome-wide association scans, added significantly to the predictive utility for CHD of conventional risk factors (CRF) in the Framingham risk score (FRS) algorithm. Methods: In the Northwick Park Heart Study II of 2742 men (270 CHD events occurring during a 15-year prospective study), rs10757274 A>G [mean frequency G = 0.48 (95% CI 0.47–0.50)] was genotyped. Using the area under the ROC curve (AROC) and the likelihood ratio (LR) statistic, we assessed the discriminatory performance of the FRS based on CRFs with and without genotype. Results: rs10757274 A>G was associated with incident CHD, with an effect size as reported previously [hazard ratio in GG vs AA men of 1.60 (95% CI 1.12–2.28)], independent of CRFs and family history of CHD. Although the AROC for CRFs alone [0.62 (95% CI 0.58–0.66)] did not increase significantly (P = 0.14) when rs10757274 A>G genotype was added [0.64 (95% CI 0.60–0.68)], including genotype gave better fit (LR P = 0.01) and including rs10757274 moved 369 men (13.5% of the total) into more accurate risk categories. To model polygenic effects, 10 hypothetical, randomly assigned gene variants, with similar effect size and frequencies were added. Two variants made significant AROC improvements to the FRS prediction (P = 0.01), whereas further variants had smaller incremental effects (final AROC = 0.71, P <0.001 vs CRFs; LR vs CRFs P <0.0001). Conclusions: Although overall, rs10757274 did not add substantially to the usefulness of the FRS for predicting future events, it did improve reclassification of CHD risk, and thus may have clinical utility. The genome-wide association scan, a direct outcome of the Human Genome Project and HapMap, has revolutionized the field of genetics of complex, common diseases by identifying novel loci and gene regions associated with disease risk. To overcome the problem of type I errors, all findings require replication. Recently, 3 genome-wide association scans all identified a single region on chromosome 9p21.3 associated with coronary heart disease (CHD) or myocardial infarction (MI)1 risk (1)(2)(3). Helgadottir et al. (3) identified a single-nucleotide polymorphism (SNP), rs10757278, that showed significant association with risk (P <10−20) and was replicated in a total of 4587 cases and 12 769 controls, with a population-attributable fraction (PAF) of 20% for MI. Resequencing of the region did not identify any obvious causative mutations. McPherson et al. (1) identified 2 SNPs in the same genomic region, rs10757274 and rs2383206, both associated with risk, in 6 independent studies comprising more than 23 000 individuals and with a PAF of 10% to 15%. The Wellcome Trust Case Control Consortium identified rs1333049 in this region, also showing strong association with CHD risk (P <10−14) (2). Finally, a study that combined the Wellcome Trust Case Control Consortium data with data from Germany also confirmed the chromosome 9 region and identified 6 additional novel candidate genes for CHD (4). 9p21.3 is a chromosomal region relatively replete of open reading frames, and the closest genes are a cluster consisting of CDKN2A-ARF-CDKN2B,2 which lie in a linkage disequilibrium (LD) block, adjacent to these SNPs. This locus has been associated with tumor suppression, but also plays a role in cell proliferation, senescence, and apoptosis (5), all features implicated in atherogenesis. To date, the potential mechanism by which variants in this chromosome 9 region increase risk of CHD is unclear, and none of the genotypes used in the 3 studies were associated with any of the classic CHD risk factors such as blood pressure and lipid concentrations. The identification of this locus as being associated with CHD risk thus identifies a novel and potentially highly important new causal pathway for investigation and the development of therapeutic approaches that will complement current risk-reducing modalities. While statistically robust, these studies have not addressed the issue of clinical utility. Case-control studies are efficient for gene discovery, yet they provide limited information on population allele frequency, attributable risk, effect of genes on other important risk factors for cardiovascular disease, or derivation of metrics essential to evaluate the predictive utility of genetic information. All of this requires prospective studies, which additionally allow assessment of effects on incident as opposed to prevalent CHD. Here, we investigated whether addition of chromosome 9p21.3 genotypes improves the prediction of CHD events of conventional risk factors (CRFs), such as cholesterol, triglycerides, blood pressure, age, and smoking, used in the Framingham risk algorithm (6). We followed 2742 healthy middle-aged men from the prospective Northwick Park Heart Study II (NPHS-II) for an average of 14 years, with 270 CHD events. We evaluated discrimination by use of the area under the ROC curve (AROC) based on combinations of CRFs and chromosome 9p21.3 region genotype, and we assessed the ability of the genotype to stratify individuals into risk categories by determining the number of men correctly reclassified. Model fit was assessed using the likelihood ratio (LR) and Bayes information criterion (BIC), which are global measures combining both discrimination and calibration, as suggested recently (7). The potential clinical benefit of such changes in risk stratification was examined by evaluating the proportion of subjects reclassified based on a 10-year CHD risk threshold for intervention of 20% as recommended by current guidelines. Materials and Methods nphs-ii NPHS-II is a prospective study of healthy middle-aged men (50–64 years old) recruited from 9 UK general practices (8). Full details have been reported (8)(9). Family history of CHD was assessed by questionnaire at baseline as described (10). In the 2742 white men with genotype data, by December 2005, there had been 270 CHD events comprising 175 acute CHD events (42 fatal), 72 coronary artery revascularization procedures, and 23 silent MIs. genotyping We genotyped rs10757274 and rs2383206 by use of TaqMan technology (Applied Biosciences). (See the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol54/issue3 for more details on methods.) statistical analysis We assessed associations with CHD risk by use of Cox proportional hazards models and derived hazard ratios (HRs). Analyses were stratified by general practice. For the conventional model, a score was derived based on age, triglycerides, cholesterol, smoking, and systolic blood pressure (9). (See the online Data Supplement for complete details on methods.) We then fitted a model including both conventional factors and rs10757274 genotype and obtained a second score by weighting according to the β-coefficients from the model (Table 1 ). We evaluated the models by use of BIC and the LR χ2 (see the online Data Supplement for details on methods). Table 1. Baseline characteristics and HRs of conventional risk factors in NPHS-II.1 . No CHD . With CHD . P Value . β-Coefficient (SE) . HR (95% CI) . P value . n 2472 270 Age, years 56.0 (3.4) 56.6 (3.5) 0.007 0.059 (0.019) 1.35 (1.12–1.62) 0.002 Current smoker, % 27.1 (671) 37.0 (100) 0.001 0.478 (0.129) 1.61 (1.25–2.08) <0.0001 BMI, kg/m22 26.2 (3.4) 26.7 (3.3) 0.03 Systolic blood pressure, mmHg 136.8 (18.6) 141.3 (19.3) 0.0002 0.011 (0.005) 1.24 (1.04–1.48) 0.02 Diastolic blood pressure, mmHg 84.4 (11.2) 86.9 (11.4) 0.0004 Cholesterol, mmol/L 5.70 (1.01) 6.06 (1.03) <0.0001 0.319 (0.076) 1.38 (1.19–1.61) <0.0001 HDL cholesterol, mmol/L 1.72 (0.59) 1.59 (0.55) 0.002 Triglycerides, mmol/L2 1.77 (0.94) 2.05 (1.06) <0.0001 0.299 (0.170) 1.17 (0.98–1.39) 0.08 rs10757274, n (%)  AA 680 (28.0) 53 (20.1) 0.03 AG + GG:AA  AG 1186 (48.8) 138 (52.3) 0.383 (0.155) 1.47 (1.08–1.99) 0.01  GG 564 (23.2) 73 (27.7) G allele, frequency (95% CI) 0.48 (0.46–0.49) 0.54 (0.49–0.58) 0.007 rs2383206, n (%)  AA 601 (24.7) 50 (18.7)  AG 1203 (49.5) 143 (53.6) 0.14  GG 626 (25.8) 74 (27.7) Gallele, frequency (95% CI) 0.51 (0.49–0.52) 0.54 (0.50–0.59) 0.08 . No CHD . With CHD . P Value . β-Coefficient (SE) . HR (95% CI) . P value . n 2472 270 Age, years 56.0 (3.4) 56.6 (3.5) 0.007 0.059 (0.019) 1.35 (1.12–1.62) 0.002 Current smoker, % 27.1 (671) 37.0 (100) 0.001 0.478 (0.129) 1.61 (1.25–2.08) <0.0001 BMI, kg/m22 26.2 (3.4) 26.7 (3.3) 0.03 Systolic blood pressure, mmHg 136.8 (18.6) 141.3 (19.3) 0.0002 0.011 (0.005) 1.24 (1.04–1.48) 0.02 Diastolic blood pressure, mmHg 84.4 (11.2) 86.9 (11.4) 0.0004 Cholesterol, mmol/L 5.70 (1.01) 6.06 (1.03) <0.0001 0.319 (0.076) 1.38 (1.19–1.61) <0.0001 HDL cholesterol, mmol/L 1.72 (0.59) 1.59 (0.55) 0.002 Triglycerides, mmol/L2 1.77 (0.94) 2.05 (1.06) <0.0001 0.299 (0.170) 1.17 (0.98–1.39) 0.08 rs10757274, n (%)  AA 680 (28.0) 53 (20.1) 0.03 AG + GG:AA  AG 1186 (48.8) 138 (52.3) 0.383 (0.155) 1.47 (1.08–1.99) 0.01  GG 564 (23.2) 73 (27.7) G allele, frequency (95% CI) 0.48 (0.46–0.49) 0.54 (0.49–0.58) 0.007 rs2383206, n (%)  AA 601 (24.7) 50 (18.7)  AG 1203 (49.5) 143 (53.6) 0.14  GG 626 (25.8) 74 (27.7) Gallele, frequency (95% CI) 0.51 (0.49–0.52) 0.54 (0.50–0.59) 0.08 1 Data are mean (SD) unless noted otherwise. 2 Geometric mean (approximate SD). Open in new tab Table 1. Baseline characteristics and HRs of conventional risk factors in NPHS-II.1 . No CHD . With CHD . P Value . β-Coefficient (SE) . HR (95% CI) . P value . n 2472 270 Age, years 56.0 (3.4) 56.6 (3.5) 0.007 0.059 (0.019) 1.35 (1.12–1.62) 0.002 Current smoker, % 27.1 (671) 37.0 (100) 0.001 0.478 (0.129) 1.61 (1.25–2.08) <0.0001 BMI, kg/m22 26.2 (3.4) 26.7 (3.3) 0.03 Systolic blood pressure, mmHg 136.8 (18.6) 141.3 (19.3) 0.0002 0.011 (0.005) 1.24 (1.04–1.48) 0.02 Diastolic blood pressure, mmHg 84.4 (11.2) 86.9 (11.4) 0.0004 Cholesterol, mmol/L 5.70 (1.01) 6.06 (1.03) <0.0001 0.319 (0.076) 1.38 (1.19–1.61) <0.0001 HDL cholesterol, mmol/L 1.72 (0.59) 1.59 (0.55) 0.002 Triglycerides, mmol/L2 1.77 (0.94) 2.05 (1.06) <0.0001 0.299 (0.170) 1.17 (0.98–1.39) 0.08 rs10757274, n (%)  AA 680 (28.0) 53 (20.1) 0.03 AG + GG:AA  AG 1186 (48.8) 138 (52.3) 0.383 (0.155) 1.47 (1.08–1.99) 0.01  GG 564 (23.2) 73 (27.7) G allele, frequency (95% CI) 0.48 (0.46–0.49) 0.54 (0.49–0.58) 0.007 rs2383206, n (%)  AA 601 (24.7) 50 (18.7)  AG 1203 (49.5) 143 (53.6) 0.14  GG 626 (25.8) 74 (27.7) Gallele, frequency (95% CI) 0.51 (0.49–0.52) 0.54 (0.50–0.59) 0.08 . No CHD . With CHD . P Value . β-Coefficient (SE) . HR (95% CI) . P value . n 2472 270 Age, years 56.0 (3.4) 56.6 (3.5) 0.007 0.059 (0.019) 1.35 (1.12–1.62) 0.002 Current smoker, % 27.1 (671) 37.0 (100) 0.001 0.478 (0.129) 1.61 (1.25–2.08) <0.0001 BMI, kg/m22 26.2 (3.4) 26.7 (3.3) 0.03 Systolic blood pressure, mmHg 136.8 (18.6) 141.3 (19.3) 0.0002 0.011 (0.005) 1.24 (1.04–1.48) 0.02 Diastolic blood pressure, mmHg 84.4 (11.2) 86.9 (11.4) 0.0004 Cholesterol, mmol/L 5.70 (1.01) 6.06 (1.03) <0.0001 0.319 (0.076) 1.38 (1.19–1.61) <0.0001 HDL cholesterol, mmol/L 1.72 (0.59) 1.59 (0.55) 0.002 Triglycerides, mmol/L2 1.77 (0.94) 2.05 (1.06) <0.0001 0.299 (0.170) 1.17 (0.98–1.39) 0.08 rs10757274, n (%)  AA 680 (28.0) 53 (20.1) 0.03 AG + GG:AA  AG 1186 (48.8) 138 (52.3) 0.383 (0.155) 1.47 (1.08–1.99) 0.01  GG 564 (23.2) 73 (27.7) G allele, frequency (95% CI) 0.48 (0.46–0.49) 0.54 (0.49–0.58) 0.007 rs2383206, n (%)  AA 601 (24.7) 50 (18.7)  AG 1203 (49.5) 143 (53.6) 0.14  GG 626 (25.8) 74 (27.7) Gallele, frequency (95% CI) 0.51 (0.49–0.52) 0.54 (0.50–0.59) 0.08 1 Data are mean (SD) unless noted otherwise. 2 Geometric mean (approximate SD). Open in new tab We used a simple model testing the changes of the ROC area with increasing genetic information to display the importance of combining a number of genes, with moderate effects, with the classic risk factors used by Framingham in the prediction of CHD events. (See the online Data Supplement for details on methods.) Results conventional risk factors and chd events The baseline characteristics of the Northwick Park Heart Study II, stratified by subsequent CHD event, are presented in Table 1 . The men who went on to develop CHD during follow-up (n = 270) were older and had higher plasma cholesterol, triglycerides, and blood pressure, lower HDL cholesterol, and higher prevalence of smoking than those who remained CHD free (n = 2472). The HRs associated with these traits are presented in Table 1 . Based on the measured variables included in the Framingham algorithm, the AROC given by this set of CRFs was 0.62 (95% CI 0.58–0.66), with a DR5 [detection rates (or sensitivies) for a 5% false-positive] of 13.5%. genotype effects on chd traits and chd The genotype distribution of rs10757274 A>G and rs283206 A>G were in Hardy-Weinberg equilibrium and strong LD (r2 = 0.89). The association of these SNPs with intermediate traits is shown in Supplementary Table 1 in the online Data Supplement. There was some evidence of a modest association with levels of apolipoprotein A-I and HDL cholesterol, with mean (SD) HDL levels being higher in those homozygous for the G allele [AA 1.68 (0.58) mmol/L, GG 1.76 (0.61) mmol/L, P = 0.03], whereas fibrinogen levels were lower in G carriers [AA 2.76 (0.52), GG + AG 2.69 (0.51), P = 0.002]. For rs10757274, the frequency of the G allele was significantly higher (P <0.007) in the CHD-positive men vs the CHD-free group [0.54 (95% CI 0.49–0.58) and 0.48 (95% CI 0.46–0.49), respectively]. As shown in Table 2 , compared with men homozygous for the common A allele, men homozygous for the G allele had HR of 1.60 (1.12–2.28), P = 0.03, an effect that was not materially changed after adjustment for CRF. The PAF for this SNP was 26.2% (95% CI 7.1–41.1). The Kaplan-Meier survival plot associated with rs10757274 is presented in Fig. 1 , showing the lower survival in those carrying one or more G alleles from the start and continuing thereafter. Figure 1. Open in new tabDownload slide Kaplan-Meier survival plot for rs10757274 A>G in NPHS-II men. Figure 1. Open in new tabDownload slide Kaplan-Meier survival plot for rs10757274 A>G in NPHS-II men. Table 2. HR associated with rs10757274 in NPHS-II men.123 Genotype . CHD rate per 1000 person-years . HR (95% CI) . . . . . Model 1 . Model 2 . Model 3 . AA 5.9 1.00 1.00 1.00 AG 8.5 1.40 (1.02–1.92) 1.38 (1.00–1.90) 1.58 (1.09–2.28) GG 9.5 1.60 (1.12–2.28) 1.57 (1.10–2.25) 1.96 (1.31–2.94) P value 0.03 0.04 0.003 AA 1.00 1.00 1.00 AG + GG 1.46 (1.08–1.97) 1.44 (1.06–1.96) 1.70 (1.19–2.41) P value 0.01 0.02 0.003 Genotype . CHD rate per 1000 person-years . HR (95% CI) . . . . . Model 1 . Model 2 . Model 3 . AA 5.9 1.00 1.00 1.00 AG 8.5 1.40 (1.02–1.92) 1.38 (1.00–1.90) 1.58 (1.09–2.28) GG 9.5 1.60 (1.12–2.28) 1.57 (1.10–2.25) 1.96 (1.31–2.94) P value 0.03 0.04 0.003 AA 1.00 1.00 1.00 AG + GG 1.46 (1.08–1.97) 1.44 (1.06–1.96) 1.70 (1.19–2.41) P value 0.01 0.02 0.003 1 Model 1: adjusted for age and general practice. 2 Model 2: adjusted for age, smoking, systolic blood pressure, cholesterol, triglycerides, and BMI. 3 Model 3: adjusted for age, smoking, systolic blood pressure, cholesterol, and calculated baseline HDL. Open in new tab Table 2. HR associated with rs10757274 in NPHS-II men.123 Genotype . CHD rate per 1000 person-years . HR (95% CI) . . . . . Model 1 . Model 2 . Model 3 . AA 5.9 1.00 1.00 1.00 AG 8.5 1.40 (1.02–1.92) 1.38 (1.00–1.90) 1.58 (1.09–2.28) GG 9.5 1.60 (1.12–2.28) 1.57 (1.10–2.25) 1.96 (1.31–2.94) P value 0.03 0.04 0.003 AA 1.00 1.00 1.00 AG + GG 1.46 (1.08–1.97) 1.44 (1.06–1.96) 1.70 (1.19–2.41) P value 0.01 0.02 0.003 Genotype . CHD rate per 1000 person-years . HR (95% CI) . . . . . Model 1 . Model 2 . Model 3 . AA 5.9 1.00 1.00 1.00 AG 8.5 1.40 (1.02–1.92) 1.38 (1.00–1.90) 1.58 (1.09–2.28) GG 9.5 1.60 (1.12–2.28) 1.57 (1.10–2.25) 1.96 (1.31–2.94) P value 0.03 0.04 0.003 AA 1.00 1.00 1.00 AG + GG 1.46 (1.08–1.97) 1.44 (1.06–1.96) 1.70 (1.19–2.41) P value 0.01 0.02 0.003 1 Model 1: adjusted for age and general practice. 2 Model 2: adjusted for age, smoking, systolic blood pressure, cholesterol, triglycerides, and BMI. 3 Model 3: adjusted for age, smoking, systolic blood pressure, cholesterol, and calculated baseline HDL. Open in new tab When baseline HDL levels were included, the HR for GG men was 1.96 (1.31–2.94), P <0.003, suggesting that the mechanism of risk was independent of HDL levels. Fibrinogen added to the model had no additional effect. The effect size was consistent in subjects with definite MI [HR 1.45 (95% CI 0.99–2.11)], coronary artery bypass graft [HR 1.52 (95% CI 0.85–2.72)], and silent MI [HR 1.40 (95% CI 0.52–3.77)]. Subjects with possible MIs showed the same effect size, HR 1.79 (95% CI 0.68–4.70), but were not included in the CHD analysis. There was no statistically significant evidence for the risk effect being different in those who smoked or had different levels of obesity [as assessed by body mass index (BMI)] or inflammation [as assessed by C-reactive protein (data not shown)]. We have previously reported that in the NPHS-II men, a family history of early CHD was strongly associated with higher future risk, independent of CRFs (10), but the HR for rs10757274 (G carriers vs AA men) was 1.42 (95% CI 1.05–1.93), P = 0.02, in a model of CRFs without family history and 1.41 (95% CI 1.04–1.92), P = 0.03, after adjustment for family history, suggesting that the SNP did not explain a significant part of the family history risk. The prevalence of a family history of early CHD was not different by rs10757274 genotype (see Supplementary Table 1a in the online Data Supplement). For rs283206, despite the strong LD with rs10757274, the effects on risk and trait association were generally smaller and did not reach statistical significance (see Supplementary Tables 1b and 2 in the online Data Supplement). Haplotype analysis of the combined SNPs did not add significantly to the risk estimates (data not shown). Neither SNP was associated with significantly increased risk of type 2 diabetes or cancers (data not shown). combined effect of crfs and rs10757274 in the risk algorithm When rs10757274 genotype data was added to the CRFs in the risk algorithm, for model 1 (adjusted for age and general practice) the AROC increased from 0.62 to 0.64 (P = 0.14) (Fig. 2 ) with a DR5 of 13%. Thus, on its own, genetic variation near CDKN2A did not add to the overall risk prediction of CRFs. However, since no single genotype is likely to add significantly to CRF risk prediction (11), we modeled the effect on CHD risk of up to 10 hypothetical, randomly assigned gene variants, with allele frequencies and risk similar to those of rs10757274. The risk associated with each additional SNP is shown in Fig. 3 . The addition of 1 further SNP with similar characteristics increases the AROC significantly (P <0.03), whereas the inclusion of 2 or more SNPs had a greater effect (P <0.001), with the addition of further SNPs having smaller incremental effect. As shown in Fig. 4 , the AROC for 10 SNPs was 0.76, with a DR5 of 24.2%. However, whether this improvement in the AROC is clinically useful remains to be seen, as the proportion of individuals with multiple independently segregating risk alleles in any population is likely to be small. Figure 2. Open in new tabDownload slide Hazard ratio for CHD for rs10757274 A>G in NPHS-II men. P value adjusted for age, cholesterol, triglycerides, BMI, systolic blood pressure, and smoking. Figure 2. Open in new tabDownload slide Hazard ratio for CHD for rs10757274 A>G in NPHS-II men. P value adjusted for age, cholesterol, triglycerides, BMI, systolic blood pressure, and smoking. Figure 3. Open in new tabDownload slide AROC for the algorithm based on CRFs only, CRFs plus rs10757274, and CRFs, rs10757274, and 10 additional genotypes of similar effect. Inset: AROC values for addition of 10 model SNPs. Figure 3. Open in new tabDownload slide AROC for the algorithm based on CRFs only, CRFs plus rs10757274, and CRFs, rs10757274, and 10 additional genotypes of similar effect. Inset: AROC values for addition of 10 model SNPs. Figure 4. Open in new tabDownload slide Comparative sequence analysis among human, monkey, dog, and mouse showing strong homology in the region where the SNPs rs10757274 and rs2383206 lie, suggesting conserved sequences of possible function. Taken from http://ecrbrowser.dcode.org/ (25). Figure 4. Open in new tabDownload slide Comparative sequence analysis among human, monkey, dog, and mouse showing strong homology in the region where the SNPs rs10757274 and rs2383206 lie, suggesting conserved sequences of possible function. Taken from http://ecrbrowser.dcode.org/ (25). clinical utility of genotype risk stratification The potential in ability of the SNP to improve risk stratification was examined, as shown in Supplementary Tables 3–4 in the online Data Supplement. Based on their CRF score, men were divided into those with a 10-year CHD risk of <5%, 5%–10%, 10%–20%, and >20% risk. After inclusion of the genotype, 585 (21.9%) men were reclassified, of whom 63% (369) moved into more accurate categories (defined by the observed risk corresponding better to the predicted risk in the new category). As shown in Supplementary Table 4 in the online Data Supplement, for the single SNP model the BIC value decreased, and the LR statistic increased significantly (P = 0.01). Both BIC and LR take into account the increase in the number of predictors. The nonsignificant P values for the Hosmer-Lemeshow statistic indicated no problems with calibration in any of the models, and the increase in the P values as terms were added indicated that predicted risks corresponded better to observed risks. The corresponding values for the model including the 10 hypothetical SNPs indicated the expected considerable improvements in calibration. We examined the potential clinical utility of this genotype for risk stratification in the NPHS-II men. The National Institute for Clinical Excellence, which sets standards of clinical treatment in the UK, recommends that subjects with a 10-year risk of cardiovascular disease >20% should be treated with statins. Using the variables in the Framingham algorithm, this equates to a risk score of 23. Of the 2670 men with a 9p21.3 genotype and complete trait data, 164 had scores >23, of whom 33 had an event (observed 10-year risk 23.3%). By adding rs10757274, 55 men who were originally below this cutoff using CRF alone were now above the cutoff; of these, 12 men went on to experience a definite CHD event (observed 10-year risk 24.0%). The mean cholesterol levels at baseline in these 2 groups were 6.73 and 6.78 mmol/L, respectively, and baseline LDL-C levels were 4.2 and 4.8 mmol/L. Based on the expected benefit of reducing their individual cholesterol levels to the Joint British Society Guidelines (JBS2) target of 4.0 mmol/L, the number of CHD events prevented per 100 treated would be 9.1 in the CRF-only group and 8.5 in the group identified using the genotype (P >0.5). Discussion In a prospective study of healthy middle-aged men, we confirm here the results reported from 4 large GWA studies that variation in the chromosome 9p21.3 region is strongly associated with CHD risk. The risk estimates associated with rs10757274 genotype in these UK men are very similar in magnitude to those reported (3), and because of the high frequency, the genotype has a high PAF of 26.2%. The second SNP examined showed similar but marginally smaller effects, but because of the strong LD between the two SNPs examined, their combined effect was not different from the single SNP effect. Because the risk allele showed association with higher plasma levels of apolipoprotein A1 and HDL cholesterol and lower levels of fibrinogen, adjustment for these factors increased the size of the associated risk, demonstrating that the risk effect is independent of these intermediate traits and confirming the earlier reports (1). The risk effect was similar in those with and without a family history of early CHD, and therefore was independent of this risk (10). Although a genotype may be strongly and robustly associated with CHD risk, if it has this effect through influencing the level of a CRF that is already included in the CHD risk-score algorithm (such as cholesterol, triglycerides, or blood pressure), it is unlikely to add significantly to the overall ability of the algorithm in risk prediction (11). The corollary of this is that genotypes whose risk mechanism is not working through the CRFs already included are more likely to have clinical utility, and this is the case with rs10757274 and rs2383206. Perhaps surprisingly, therefore, given its strong effect, addition of the rs10757274 genotype to the Framingham CRF risk variables did not add significantly to risk prediction in this group of middle-aged healthy men, with only a 3% improvement in the AROC value. However, it is now increasingly recognized that prediction can be improved significantly only by the inclusion of factors that both are common and have very large effects (12), and that a single genotype (or biomarker) associated with odds ratios in the region of 1.2–1.6 will not on its own significantly improve risk prediction for polygenic multifactorial CHD. The availability of several SNPs in combination has clear potential (13)(14), however. The addition of only one other model SNP (of similar allele frequency and effect as rs10757274) did significantly improve prediction, and the addition of a second or third SNP improved the AROC value by 8.4% and 13.3%, respectively; addition of each subsequent SNP had smaller incremental effects on improving the value. Were 10 such SNPs available, the predicted improvement would be 23%. This clearly demonstrates that genetic information has the potential to improve the overall ability to predict risk in the general population, with as few as 3 gene variants providing information that will have clinical utility. In the recent combined genome-wide association scan data (4), 6 novel genes determining CHD risk were reported with allele HRs in the range of 1.20 to 1.33 and risk allele frequencies between 0.22 and 0.77, strongly suggesting that this model is likely to be achievable in the near future. The addition of even strong risk markers such as CRP to the Framingham algorithm does not improve prediction, because they are highly correlated with other biomarkers already in the algorithm (such as BMI and smoking), so the potential utility of these independent genetic factors is clear. Although a risk marker may not itself significantly improve overall prediction in the population, it may still have clinical utility in risk stratification for the individual, as recently discussed (15). Genotyping for rs10757274 added significantly to the ability of the Framingham score to discriminate individuals who will suffer cardiac events, with 13.5% of the men moving into more accurate risk categories for their future events after inclusion of genotype. Specifically, it allowed reclassification of CHD risk in 3.3% of intermediate-risk men to the high-risk category, who could then be offered statin therapy to reduce their risk. The clinical utility of such genetic tests would be realized as additional similar risk variants were identified, as has been demonstrated by Samani et al. (4). In the UK, the current National Institute for Clinical Excellence guidelines propose that subjects with a Framingham 10-year CHD risk of >20% (equivalent to a CRF score >23) would qualify for statin therapy, aiming to reduce cholesterol to 4.0 mmol/L under JBS2 guidelines. One hundred sixty-four men (6.1% of the total) were identified as having a score >23 based on their baseline CRFs. Although treating subjects at low Framingham risk score will not be cost-effective because of the low event rate, the majority of CHD events occur in subjects with intermediate-risk scores, since this is the most common group. Using additional and independent factors to identify those in the <23 risk score group who are at high risk will have clinical benefit, and the data from the NPHS-II men supports this. There were 55 men (2.1% of the total group) who would not have qualified for statin treatment based on their CRF score alone (<23) but who would have a score >23 if rs10757274 genotype were added. In this group, individual statin treatment of their baseline LDL cholesterol levels to the JBS2 target would have prevented a similar number of future CHD events as in the 164 high-risk men. This strongly suggests that the use of this single genotype to identify individuals who have intermediate risk based on CRF is likely to have clinical utility. The reasons that this single SNP did not add significantly to overall AROC risk prediction are partly because the effect is relatively modest (although the CHD risk associated with any single SNP is always likely to be in the range 1.2–1.8 (16)) and partly because the number of individuals carrying the risk genotype is low. When other independent, confirmed CHD risk SNPs are identified, however, a greater proportion of men will be carrying such risk genotypes, and as suggested by the modeling, the clinical utility of adding SNP data to the algorithm will increase significantly. In addition, some individuals will, by chance, carry more than one risk allele (since they are inherited independently), and effects on risk are likely to be additive (17). Because many SNPs can be cheaply determined simultaneously in a single sample (for example, using DNA obtained from a buccal swab), this will not have major cost implications. The mechanism of the risk association of the chr9.21.3 locus is still unclear. Chromosome 9p21.3 is often deleted in malignant tumors, and attention has therefore focused on the role of the few genes within the region as potential cancer genes; cyclin-dependent kinase inhibitor 2A (melanoma, p16, inhibits CDK4) (CDKN2A) and cyclin-dependent kinase inhibitor 2B (p15, inhibits CDK4) (CDKN2B) encode p16INK4a and p16INK4b inhibitors of CDK4 kinase. Within the CDKN2A locus, an alternative reading frame specifies a protein (ARF) that is structurally unrelated to these p16s and functions as a stabilizer of the tumor suppressor protein p53 (reviewed in (5)). In spite of structural and functional differences, these 3 genes share a common functionality in cell cycle G1 control. Because CDKN2B is expressed in macrophages and is dramatically induced by transforming growth factor β, this gene may have a role in growth inhibition induced by transforming growth factor β (18). Thus the involvement of these genes in senescence (19) and apoptosis, both processes associated with atherosclerosis, suggests a potential CHD mechanism, either in plaque progression or rupture, due to changes in the senescence and apoptosis of endothelial or smooth muscle cells or monocyte-macrophage-foam cells in the lesion. Resequencing the coding regions, intron–exon boundaries, and regulatory regions in these genes, however, did not identify any functional variants (1)(2). This suggests that the functional changes may not be in the CDKN2A-ARF-CDKN2B gene loci themselves but in a regulatory region within this LD block. Recent evidence suggests that the entire locus might be coordinately suppressed by a cis-acting regulatory domain or by members of the Polycomb group of repressor complexes, which recognize histone modifications (20). Comparative sequence homology of 9p21.3 (http://ecrbrowser.dcode.org/) shows that the 58-kb LD block is highly conserved across several species (Fig. 4 ). There is also the possibility that this noncoding DNA is transcribed into RNA with distinct regulatory roles (21), and an antisense RNA to the gene cluster has been identified, lending support to this concept (22). The only other gene in the region, methylthioadenosine phosphorylase (MTAP), is ubiquitously expressed. It encodes an enzyme responsible for recycling 5′-methylthioadenosine (MTA) to S-adenosylmethionine (23). However, although variants may influence plasma levels of homocysteine and folate, which were measured in NPHS-II (24), there were no significant differences in levels of these factors by either SNP genotype (see Supplementary Table 1 in the online Data Supplement). Understanding the precise molecular mechanisms of this risk effect will not only identify new therapeutic targets but will also improve the accuracy with which these SNPs can be used for risk stratification. Finally, this analysis illustrates the value of studying genetic effects uncovered in case control studies in population-based cohort studies. This type of research is likely to help better understand the value of the exciting new genetic information for the betterment of public health. Grant/funding Support: NPHS-II was supported by the British Medical Research Council, the US National Institutes of Health (grant NHLBI 33014), and Du Pont Pharma, Wilmington, Delaware. J.A.C., J.P., R.L., F.D., A.D.H., and S.E.H. are supported by the British Heart Foundation (PG/2005/014 and SP/07/007/23671). Financial Disclosures: None declared. Acknowledgments: The authors dedicate this paper to the PI of the NPHS-II study, Professor George Miller, who died on August 14, 2006, after a long illness. 1 " Nonstandard abbreviations: MI, myocardial infarction; PAF, population-attributable fraction; LD, linkage disequilibrium; CRF, conventional risk factor; NPHS-II, Northwick Park Heart Study II; LR, likelihood ratio; BIC, Bayes information criterion; HR, hazard ratio. 2 " Human genes: CDKN2A, cyclin-dependent kinase inhibitor 2A (melanoma, p16, inhibits CDK4); CDKN2B, cyclin-dependent kinase inhibitor 2B (p15, inhibits CDK4); MTAP, methylthioadenosine phosphorylase. References 1 McPherson R, Pertsemlidis A, Kavaslar N, Stewart A, Roberts R, Cox DR, et al. A common allele on chromosome 9 associated with coronary heart disease. Science (Wash DC) 2007 ; 316 : 1488 -1491. Crossref Search ADS 2 . The Wellcome Trust Case Control Consortium. Genome-wide association study of 14,000 cases of seven common diseases and 3,000 shared controls. Nature (Lond) 2007 ; 447 : 661 -678. Crossref Search ADS 3 Helgadottir A, Thorleifsson G, Manolescu A, Gretarsdottir S, Blondal T, Jonasdottir A, et al. A common variant on chromosome 9p21 affects the risk of myocardial infarction. Science (Wash DC) 2007 ; 316 : 1491 -1493. Crossref Search ADS 4 Samani NJ, Erdmann J, Hall AS, Hengstenberg C, Mangino M, Mayer B, et al. Genomewide association analysis of coronary artery disease. N Engl J Med 2007 ; 357 : 443 -453. Crossref Search ADS PubMed 5 Gil J, Peters G. Regulation of the INK4b-ARF-INK4a tumour suppressor locus: all for one or one for all. Nat Rev Mol Cell Biol 2006 ; 7 : 667 -677. Crossref Search ADS PubMed 6 Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB. Prediction of coronary heart disease using risk factor categories. Circ 1998 ; 97 : 1837 -1847. Crossref Search ADS 7 Cook NR. Use and misuse of the receiver operating characteristic curve in risk prediction. Circ 2007 ; 115 : 928 -935. Crossref Search ADS 8 Miller GJ, Bauer KA, Barzegar S, Cooper JA, Rosenberg RD. Increased activation of the haemostatic system in men at high risk of fatal coronary heart disease. Thromb Haemost 1996 ; 75 : 767 -771. Crossref Search ADS PubMed 9 Cooper JA, Miller GJ, Humphries SE. A comparison of the PROCAM and Framingham point scoring systems for estimation of individual risk of coronary heart disease in the Second Northwick Park Heart Study. Atherosclerosis 2004 ; 181 : 93 -100. 10 Hawe E, Talmud PJ, Miller GJ, Humphries SE. Family history of coronary heart disease factor in the second Northwick Park Heart Study. Ann Hum Genet 2003 ; 67 : 97 -106. Crossref Search ADS PubMed 11 Humphries SE, Ridker PM, Talmud PJ. Genetic testing for cardiovascular disease susceptibility: a useful clinical management tool or possible misinformation?. Arterioscler Thromb Vasc Biol 2004 ; 24 : 628 -636. Crossref Search ADS PubMed 12 Wald NJ, Morris JK, Rish S. The efficacy of combining several risk factors as a screening test. J Med Screen 2005 ; 12 : 197 -201. Crossref Search ADS PubMed 13 Humphries SE, Cooper JA, Talmud PJ, Miller GJ. Candidate gene genotypes, along with conventional risk factor assessment, improve estimation of coronary heart disease risk in healthy UK men. Clin Chem 2007 ; 53 : 8 -16. Crossref Search ADS PubMed 14 Janssens AC, Aulchenko YS, Elefante S, Borsboom GJ, Steyerberg EW, van Duijn CM. Predictive testing for complex diseases using multiple genes: fact or fiction?. Genet Med 2006 ; 8 : 395 -400. Crossref Search ADS PubMed 15 Abuzeid AM, Hawe E, Humphries SE, Talmud PJ. Association between the Ala379Val variant of the lipoprotein associated phospholipase A2 and risk of myocardial infarction in the north and south of Europe. Atherosclerosis 2003 ; 168 : 283 -288. Crossref Search ADS PubMed 16 Drenos F, Miller GJ, Humphries SE. Increase of plasma fibrinogen levels and variability with age in a sample of middle aged healthy men. Ann Hum Genet 2007 ; 71 : 43 -53. Crossref Search ADS PubMed 17 Drenos F, Whittaker JC, Humphries SE. The use of meta-analysis risk estimates for candidate genes in combination to predict coronary heart disease risk. Ann Hum Genet 2007 ;. 18 Hannon GJ, Beach D. p15INK4B is a potential effector of TGF-beta-induced cell cycle arrest. Nature (Lond) 1994 ; 371 : 257 -261. Crossref Search ADS 19 Chen J, Huang X, Halicka D, Brodsky S, Avram A, Eskander J, et al. Contribution of p16INK4a and p21CIP1 pathways to induction of premature senescence of human endothelial cells: permissive role of p53. Am J Physiol Heart Circ Physiol 2006 ; 290 : H1575 -H1586. Crossref Search ADS PubMed 20 Jacobs JJ, Kieboom K, Marino S, DePinho RA, van Lohuizen M. The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature (Lond) 1999 ; 397 : 164 -168. Crossref Search ADS 21 Identification and analysis of functional elements in 1% of the human genome by the ENCODE pilot project. Nature (Lond) 2007 ; 447 : 799 -816. Crossref Search ADS 22 Pasmant E, Laurendeau I, Heron D, Vidaud M, Vidaud D, Bieche I. Characterization of a germ-line deletion, including the entire INK4/ARF locus, in a melanoma-neural system tumor family: identification of ANRIL, an antisense noncoding RNA whose expression coclusters with ARF. Cancer Res 2007 ; 67 : 3963 -3969. Crossref Search ADS PubMed 23 Schmid M, Sen M, Rosenbach MD, Carrera CJ, Friedman H, Carson DA. A methylthioadenosine phosphorylase (MTAP) fusion transcript identifies a new gene on chromosome 9p21 that is frequently deleted in cancer. Oncogene 2000 ; 19 : 5747 -5754. Crossref Search ADS PubMed 24 Dekou V, Gudnason V, Hawe E, Miller GJ, Stansbie D, Humphries SE. Gene-environment and gene-gene interaction in the determination of plasma homocysteine levels in healthy middle-aged men. Thromb Haemost 2001 ; 85 : 67 -74. Crossref Search ADS PubMed 25 Ovcharenko I, Nobrega MA, Loots GG, Stubbs L. ECR browser: a tool for visualizing and accessing data from comparisons of multiple vertebrate genomes. Nucleic Acids Res 2004 ; 32 : W280 -W286. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Lauretani,, Fulvio;Semba, Richard, D;Bandinelli,, Stefania;Miller, Edgar, R;Ruggiero,, Carmelinda;Cherubini,, Antonio;Guralnik, Jack, M;Ferrucci,, Luigi

doi: 10.1373/clinchem.2007.095521pmid: 18202159

Abstract Background: Recent studies suggest an association between polyunsaturated fatty acids (PUFAs) and the development of chronic kidney disease. The aim of this study was to examine the relationship between PUFAs and renal function in older adults. Methods: We performed a cross-sectional and prospective analysis of 931 adults, ≥65 years old, enrolled in the InCHIANTI study, a population-based cohort in Tuscany, Italy. Plasma PUFAs were measured at enrollment, and creatinine clearance was estimated by the Cockcroft-Gault equation at baseline and after 3-year follow-up. Results: At enrollment, participants with higher creatinine clearance had higher concentrations of HDL cholesterol, total plasma PUFAs, plasma n-3 fatty acid (FA), and plasma n-6 FA and lower triglycerides. From enrollment to the 3-year follow-up visit, creatinine clearance declined by 7.8 (12.2) mL/min (P <0.0001). Baseline total plasma PUFAs, n-3 FA, n-6 FA, and linoleic, linolenic, and arachidonic acids were strong independent predictors of less steep decline in creatinine clearance from baseline to follow-up (P <0.0001, after adjusting for baseline creatinine clearance). After adjusting for baseline creatinine, baseline total plasma PUFAs, n-3 FA, and linoleic, linolenic, and arachidonic acids were negatively associated with creatinine at 3-year follow-up. Participants with higher plasma PUFAs at enrollment had a lower risk of developing renal insufficiency, defined by a creatinine clearance <60 mL/min, during 3-year follow-up. Conclusion: High PUFA concentrations, both n-3 FA and n-6 FA, may attenuate the age-associated decline in renal function among older community-dwelling women and men. Chronic kidney disease is emerging as a major public health problem among older adults and can result in end-stage renal disease with need for dialysis or transplantation for kidney failure (1)(2). In the US, an estimated 19 million adults are in the early stage of disease (1). The Prevention of Renal and Vascular End-Stage Disease study in Europe showed that up to 12% of the adult population had some renal impairment (3). Creatinine clearance is widely used to assess chronic kidney disease in clinical practice and large epidemiologic studies (4)(5). The major risk factors for chronic kidney disease are increasing age, hypertension, diabetes, cardiovascular disease, and a family history of the disease (1). Abnormalities in lipids and atherogenic lipoprotein metabolism may contribute to glomerular and interstitial injury and progression of renal disease (4). Recent studies suggest that there may be an association between polyunsaturated fatty acids (PUFAs)1 and the development of chronic kidney disease (6). PUFA supplementation has been shown to reduce renal inflammation and fibrosis in animal models (6). PUFAs may protect kidney function by modulating the inflammatory response through downregulation of the production of proinflammatory cytokines, cyclooxygenase-2 activity, and expression of endothelial leukocyte adhesion molecules (7)(8). Accordingly, among older adults, high levels of plasma PUFAs were associated with lower levels of C-reactive protein (CRP) and proinflammatory cytokines such as interleukin (IL)-6 and tumor necrosis factor (TNF)-α and higher levels of antiinflammatory cytokines such as IL-10 and transforming growth factor (TGF)-β (9). We hypothesized that low total plasma PUFA levels were associated with an accelerated decline of kidney function in older adults. To test this hypothesis, we examined the relationship between total plasma PUFA levels and change in creatinine clearance over a 3-year follow-up in the older participants of the InCHIANTI study, a population-based epidemiology study conducted in Tuscany, Italy. Materials and Methods The study participants consisted of men and women ages 65 and older who participated in the InCHIANTI study (Aging in the Chianti Area) conducted in 2 small towns in Tuscany, Italy. The study protocol complied with the Declaration of Helsinki and was approved by the Italian National Institute of Research and Care on Aging Ethics Committee. The rationale, design, and data collection have been described (10). Briefly, in August 1998, 1270 people ages 65 years and older were randomly selected from the population registry of Greve in Chianti (population 11 709) and Bagno a Ripoli (population 4704); of 1256 eligible subjects, 1155 (90.1%) agreed to participate. Of the 1155 participants, 1055 (91.3%) donated a blood sample. The subjects who did not participate in the blood drawing were generally older and had greater comorbidity than those who participated in the blood drawing (11). Participants received an extensive description of the project and were enrolled in the study after formal consent. Participants were evaluated again for a 3-year follow-up visit that was conducted from 2001 to 2003 and included a new phlebotomy and laboratory testing. Demographic information and information on smoking and medication use were collected using standardized questionnaires. Average daily intakes of energy (kcal), carbohydrates, total protein, total lipids, etc. were estimated using the European Prospective Investigation into Cancer and Nutrition food frequency questionnaire, previously validated in the InCHIANTI population (12). All participants were examined by a trained geriatrician, and diseases (coronary heart disease, congestive heart failure, hypertension, diabetes, chronic obstructive pulmonary disease, osteoarthritis, and cancer) were ascertained according to standard, preestablished criteria and algorithms based on those used in the Women’s Health and Aging Study (13). Weight was measured using a high-precision mechanical scale. Standing height was measured to the nearest 0.1 cm. Body mass index (BMI) was calculated as weight in kg/(height in m)2. Mini-Mental Status Examination (MMSE) was administered at enrollment (14). Blood samples were collected in the morning after a 12-h fast. Aliquots of serum and plasma were immediately obtained and stored at −80 °C. Fatty acids (FAs) were measured using a fasting plasma sample. The rationale and methodology of the FA determination have been described (9). The intraassay and interassay CVs for all FAs were on average 1.6% and 3.3%, respectively. The following fatty acid–related variables were selected for analyses: (a) ω-6 fatty acids (n-6, mg/L), linoleic acid (18:2n-6, mg/L), and arachidonic acid (20:4n-6, mg/L); (b) ω-3 fatty acids (n-3, mg/L) and linolenic acid (18:3n-3, mg/L), and (c) total polyunsaturated fatty acids (total plasma PUFAs, mg/L) (9). Serum total cholesterol was assessed by commercial enzymatic tests (Roche Diagnostics, GmbH) and a Roche-Hitachi 917 analyzer. The lower detection limit was 30 mg/L. The intraassay and interassay CVs were 0.8% and 3.3%, respectively. Serum creatinine was measured by commercial enzymatic assay using a Modular P800 Hitachi Analyzer. The interassay CV was 2.3%, and the detection limit was 1 mg/L. Serum creatinine (Scr) measured by use of a modified Jaffe method was used to estimate creatinine clearance according to the Cockcroft-Gault formula (Ccr) = [(140 − age) × weight/(72 × Scr) × 0.85 (if the subject is female)] where Ccr is expressed in mL/min, age in years, weight in km, and Scr in mg/dL (15). We found a high correlation between the creatinine clearance estimated by the Cockcroft-Gault formula and that estimated by using the 4-variable Modification of Diet in Renal Disease (MDRD) equation (r = 0.71; P <0.0001) (16). Urinary protein was measured using a Modular P800 Hitachi Analyzer. Laboratory normal range for proteinuria was 0–150 mg/L. statistical analysis Variables are reported as means (SD) for normally distributed variables or as percentages. Anthropometric, biochemical, and clinical characteristics of the population at enrollment were stratified across creatinine clearance strata (<30, 30–59, ≥60 mL/min) and compared using ANOVA-based test for trend adjusting for age, sex, and BMI. Total plasma PUFAs were divided into quintiles (≤34.3, 34.4–36.9, 37.0–39.5, 39.6–42.11, ≥42.14 mg/L). We used test for a linear trend to compare changes in creatinine clearance between enrollment and 3-year follow-up across baseline total plasma PUFA quintiles. We used Pearson correlation to examine the relationship between creatinine clearance, all types of plasma PUFAs, and other variables. The relationship between changes in creatinine and changes in creatinine clearance from baseline to follow-up and all types of plasma PUFA was examined using multivariate linear models adjusted for multiple confounders. We obtained parsimonious models by starting with an initial models that included all variables associated with follow-up creatinine or creatinine clearance at a P level <0.10 (age, sex, BMI, education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL and HDL cholesterol, self-reported cancer, cardiovascular disease, and hypertension) plus baseline creatinine or creatinine clearance. We used a fully adjusted logistic regression analysis to test the hypothesis that lower plasma PUFA levels were associated with a significantly higher probability of developing renal insufficiency. In this analysis, participants with renal insufficiency at enrollment (defined by a creatinine clearance <60 mL/min) were excluded. The analysis was also repeated using a composite outcome consisting of either the development of renal insufficiency or death during the 3-year follow-up. All analyses were performed using SAS (v. 8.2, SAS Institute, Inc.) with a statistical significance level set at P <0.05. Results Of the 1055 participants who donated a blood sample at baseline, 931 (80.6%) had both plasma PUFAs and serum creatinine measurements available for this analysis. Of these 931 participants, 676 (72.7%) had creatinine clearance measurements available at the 3-year follow-up visit. Of the 255 subjects not included in the longitudinal analyses, 96 died between enrollment and follow-up, 147 refused to participate in the blood drawing and/or interview, and 12 moved out of the study area. Participants evaluated at baseline with missing creatinine clearance at the 3-year follow-up (n = 255) were significantly older (79.9 vs 73.2 years,P <0.0001), had lower baseline creatinine clearance (55.1 vs 67.1 mL/min, P <0.0001), and had lower total plasma PUFAs (36.9 vs 38.1 mg/L, P = 0.0005) than those included in the present study (n = 709). Of the 676 participants evaluated at the 3-year follow-up visit, 398 had creatinine clearance ≥60 mL/min at enrollment. The demographic and disease characteristics of the study participants at enrollment, according to creatinine clearance strata (<30, 30–59, ≥60 mL/min), are shown in Table 1 . Higher creatinine clearance was associated with younger age, male sex, and lower CRP level and BMI. Persons with higher creatinine clearance also had higher HDL cholesterol, total plasma PUFAs, plasma n-3 FA, and plasma n-6 FA and lower triglycerides. The proportion of persons with proteinuria >150 mg/L was lower among persons with higher creatinine clearance. The prevalence of congestive heart failure, stroke, and peripheral artery disease was significantly higher among persons with higher creatinine clearance. Table 1. Characteristics of the study population according to creatinine clearance strata. . Creatinine clearance . . . . . <30 mL/min . 30–59 mL/min . >60 mL/min . P1 . n 27 383 521 Age, years 87.50 (5.00) 79.65 (7.34) 71.87 (4.68) <0.0001 Female, % 70.4 65.8 47.22 <0.0001 Education, years 4 (3–5) 5 (3–5) 5 (5–6) 0.26 Current smokers, % 15 12.4 24.6 0.0002 Former smokers, % 29 24.3 37.9 0.0002 Smoking, pack-years 0.0 (0.0–25.4) 0.0 (0.0–8.2) 0.0 (0.0–28.0) 0.47 Alcohol, g/day 10.0 (7.14–28.57) 10.0 (4.29–20.0) 13.57 (5.71–22.86) 0.20 MMSE score 18.74 (8.20) 23.31 (5.15) 25.66 (3.59) 0.13 BMI, kg/m2 24.38 (4.00) 25.77 (3.68) 28.59 (3.76) <0.0001 Energy intake, kcal/day 1535 (344.29) 1813.65 (542.70) 2017.69 (562.32) 0.90 Carbohydrate intake, g/day 202.83 (44.84) 237.28 (76.08) 259.96 (82.28) 0.593 Protein intake, g/day 59.80 (15.29) 71.73 (20.50) 78.53 (19.76) 0.32 Total lipid intake, g/day 50.99 (14.46) 61.13 (20.26) 67.48 (19.31) 0.75 Saturated FA intake, g/day 17.68 (5.48) 21.15 (8.03) 22.47 (7.69) 0.44 Monounsatured FA intake, g/day 24.81 (7.91) 30.12 (10.55) 34.14 (10.77) 0.78 Polynsaturated FA intake, g/day 5.72 (1.70) 6.60 (2.11) 7.43 (2.21) 0.79 Total cholesterol, mg/L 2004.1 (277.0) 2161.2 (425.0) 2169.5 (390.8) 0.18 LDL cholesterol, mg/L 1168.9 (212.4) 1340.1 (358.0) 1369.7 (341.1) 0.07 HDL cholesterol, mg/L 487.0 (159.2) 584.3 (169.7) 544.7 (135.1) 0.02 Triglycerides, mg/L 1734.4 (989.2) 1185.1 (535.5) 1275.6 (685.0) 0.0004 Total plasma PUFAs, mg/L 32.26 (4.80) 37.34 (4.87) 37.76 (4.85) <0.0001 Plasma n-3 FA, mg/L 2.69 (0.58) 3.34 (0.97) 3.29 (0.96) 0.004 Plasma n-6 FA, mg*10/L 2.76 (0.45) 3.26 (0.44) 3.24 (0.45) <0.0001 CRP, mg/L 4.66 (1.76–15.20) 2.59 (1.30–5.56) 2.89 (1.36–5.85) <0.0001 Urine protein, mg/L 0 (0–5) 0.0 (0.0–0.0) 0.0 (0.0–0.0) 0.04 Coronary heart disease, % 10.00 5.93 5.21 0.43 Stroke, % 15.38 6.70 4.12 0.01 Diabetes, % 11.11 10.98 11.89 0.67 Hypertension, % 60.00 48.52 17.28 0.38 Peripheral artery disease, % 20.00 8.09 4.98 0.004 Congestive heart failure, % 44.44 8.56 4.86 <0.0001 COPD, % 26.92 7.18 8.65 0.35 Cancer, % 11.11 6.79 5.79 0.30 . Creatinine clearance . . . . . <30 mL/min . 30–59 mL/min . >60 mL/min . P1 . n 27 383 521 Age, years 87.50 (5.00) 79.65 (7.34) 71.87 (4.68) <0.0001 Female, % 70.4 65.8 47.22 <0.0001 Education, years 4 (3–5) 5 (3–5) 5 (5–6) 0.26 Current smokers, % 15 12.4 24.6 0.0002 Former smokers, % 29 24.3 37.9 0.0002 Smoking, pack-years 0.0 (0.0–25.4) 0.0 (0.0–8.2) 0.0 (0.0–28.0) 0.47 Alcohol, g/day 10.0 (7.14–28.57) 10.0 (4.29–20.0) 13.57 (5.71–22.86) 0.20 MMSE score 18.74 (8.20) 23.31 (5.15) 25.66 (3.59) 0.13 BMI, kg/m2 24.38 (4.00) 25.77 (3.68) 28.59 (3.76) <0.0001 Energy intake, kcal/day 1535 (344.29) 1813.65 (542.70) 2017.69 (562.32) 0.90 Carbohydrate intake, g/day 202.83 (44.84) 237.28 (76.08) 259.96 (82.28) 0.593 Protein intake, g/day 59.80 (15.29) 71.73 (20.50) 78.53 (19.76) 0.32 Total lipid intake, g/day 50.99 (14.46) 61.13 (20.26) 67.48 (19.31) 0.75 Saturated FA intake, g/day 17.68 (5.48) 21.15 (8.03) 22.47 (7.69) 0.44 Monounsatured FA intake, g/day 24.81 (7.91) 30.12 (10.55) 34.14 (10.77) 0.78 Polynsaturated FA intake, g/day 5.72 (1.70) 6.60 (2.11) 7.43 (2.21) 0.79 Total cholesterol, mg/L 2004.1 (277.0) 2161.2 (425.0) 2169.5 (390.8) 0.18 LDL cholesterol, mg/L 1168.9 (212.4) 1340.1 (358.0) 1369.7 (341.1) 0.07 HDL cholesterol, mg/L 487.0 (159.2) 584.3 (169.7) 544.7 (135.1) 0.02 Triglycerides, mg/L 1734.4 (989.2) 1185.1 (535.5) 1275.6 (685.0) 0.0004 Total plasma PUFAs, mg/L 32.26 (4.80) 37.34 (4.87) 37.76 (4.85) <0.0001 Plasma n-3 FA, mg/L 2.69 (0.58) 3.34 (0.97) 3.29 (0.96) 0.004 Plasma n-6 FA, mg*10/L 2.76 (0.45) 3.26 (0.44) 3.24 (0.45) <0.0001 CRP, mg/L 4.66 (1.76–15.20) 2.59 (1.30–5.56) 2.89 (1.36–5.85) <0.0001 Urine protein, mg/L 0 (0–5) 0.0 (0.0–0.0) 0.0 (0.0–0.0) 0.04 Coronary heart disease, % 10.00 5.93 5.21 0.43 Stroke, % 15.38 6.70 4.12 0.01 Diabetes, % 11.11 10.98 11.89 0.67 Hypertension, % 60.00 48.52 17.28 0.38 Peripheral artery disease, % 20.00 8.09 4.98 0.004 Congestive heart failure, % 44.44 8.56 4.86 <0.0001 COPD, % 26.92 7.18 8.65 0.35 Cancer, % 11.11 6.79 5.79 0.30 Continuous variables expressed as mean (SD) or as median (IQR). 1 From age-, sex-, and BMI-adjusted linear or multinomial logistic regression models as appropriate. Open in new tab Table 1. Characteristics of the study population according to creatinine clearance strata. . Creatinine clearance . . . . . <30 mL/min . 30–59 mL/min . >60 mL/min . P1 . n 27 383 521 Age, years 87.50 (5.00) 79.65 (7.34) 71.87 (4.68) <0.0001 Female, % 70.4 65.8 47.22 <0.0001 Education, years 4 (3–5) 5 (3–5) 5 (5–6) 0.26 Current smokers, % 15 12.4 24.6 0.0002 Former smokers, % 29 24.3 37.9 0.0002 Smoking, pack-years 0.0 (0.0–25.4) 0.0 (0.0–8.2) 0.0 (0.0–28.0) 0.47 Alcohol, g/day 10.0 (7.14–28.57) 10.0 (4.29–20.0) 13.57 (5.71–22.86) 0.20 MMSE score 18.74 (8.20) 23.31 (5.15) 25.66 (3.59) 0.13 BMI, kg/m2 24.38 (4.00) 25.77 (3.68) 28.59 (3.76) <0.0001 Energy intake, kcal/day 1535 (344.29) 1813.65 (542.70) 2017.69 (562.32) 0.90 Carbohydrate intake, g/day 202.83 (44.84) 237.28 (76.08) 259.96 (82.28) 0.593 Protein intake, g/day 59.80 (15.29) 71.73 (20.50) 78.53 (19.76) 0.32 Total lipid intake, g/day 50.99 (14.46) 61.13 (20.26) 67.48 (19.31) 0.75 Saturated FA intake, g/day 17.68 (5.48) 21.15 (8.03) 22.47 (7.69) 0.44 Monounsatured FA intake, g/day 24.81 (7.91) 30.12 (10.55) 34.14 (10.77) 0.78 Polynsaturated FA intake, g/day 5.72 (1.70) 6.60 (2.11) 7.43 (2.21) 0.79 Total cholesterol, mg/L 2004.1 (277.0) 2161.2 (425.0) 2169.5 (390.8) 0.18 LDL cholesterol, mg/L 1168.9 (212.4) 1340.1 (358.0) 1369.7 (341.1) 0.07 HDL cholesterol, mg/L 487.0 (159.2) 584.3 (169.7) 544.7 (135.1) 0.02 Triglycerides, mg/L 1734.4 (989.2) 1185.1 (535.5) 1275.6 (685.0) 0.0004 Total plasma PUFAs, mg/L 32.26 (4.80) 37.34 (4.87) 37.76 (4.85) <0.0001 Plasma n-3 FA, mg/L 2.69 (0.58) 3.34 (0.97) 3.29 (0.96) 0.004 Plasma n-6 FA, mg*10/L 2.76 (0.45) 3.26 (0.44) 3.24 (0.45) <0.0001 CRP, mg/L 4.66 (1.76–15.20) 2.59 (1.30–5.56) 2.89 (1.36–5.85) <0.0001 Urine protein, mg/L 0 (0–5) 0.0 (0.0–0.0) 0.0 (0.0–0.0) 0.04 Coronary heart disease, % 10.00 5.93 5.21 0.43 Stroke, % 15.38 6.70 4.12 0.01 Diabetes, % 11.11 10.98 11.89 0.67 Hypertension, % 60.00 48.52 17.28 0.38 Peripheral artery disease, % 20.00 8.09 4.98 0.004 Congestive heart failure, % 44.44 8.56 4.86 <0.0001 COPD, % 26.92 7.18 8.65 0.35 Cancer, % 11.11 6.79 5.79 0.30 . Creatinine clearance . . . . . <30 mL/min . 30–59 mL/min . >60 mL/min . P1 . n 27 383 521 Age, years 87.50 (5.00) 79.65 (7.34) 71.87 (4.68) <0.0001 Female, % 70.4 65.8 47.22 <0.0001 Education, years 4 (3–5) 5 (3–5) 5 (5–6) 0.26 Current smokers, % 15 12.4 24.6 0.0002 Former smokers, % 29 24.3 37.9 0.0002 Smoking, pack-years 0.0 (0.0–25.4) 0.0 (0.0–8.2) 0.0 (0.0–28.0) 0.47 Alcohol, g/day 10.0 (7.14–28.57) 10.0 (4.29–20.0) 13.57 (5.71–22.86) 0.20 MMSE score 18.74 (8.20) 23.31 (5.15) 25.66 (3.59) 0.13 BMI, kg/m2 24.38 (4.00) 25.77 (3.68) 28.59 (3.76) <0.0001 Energy intake, kcal/day 1535 (344.29) 1813.65 (542.70) 2017.69 (562.32) 0.90 Carbohydrate intake, g/day 202.83 (44.84) 237.28 (76.08) 259.96 (82.28) 0.593 Protein intake, g/day 59.80 (15.29) 71.73 (20.50) 78.53 (19.76) 0.32 Total lipid intake, g/day 50.99 (14.46) 61.13 (20.26) 67.48 (19.31) 0.75 Saturated FA intake, g/day 17.68 (5.48) 21.15 (8.03) 22.47 (7.69) 0.44 Monounsatured FA intake, g/day 24.81 (7.91) 30.12 (10.55) 34.14 (10.77) 0.78 Polynsaturated FA intake, g/day 5.72 (1.70) 6.60 (2.11) 7.43 (2.21) 0.79 Total cholesterol, mg/L 2004.1 (277.0) 2161.2 (425.0) 2169.5 (390.8) 0.18 LDL cholesterol, mg/L 1168.9 (212.4) 1340.1 (358.0) 1369.7 (341.1) 0.07 HDL cholesterol, mg/L 487.0 (159.2) 584.3 (169.7) 544.7 (135.1) 0.02 Triglycerides, mg/L 1734.4 (989.2) 1185.1 (535.5) 1275.6 (685.0) 0.0004 Total plasma PUFAs, mg/L 32.26 (4.80) 37.34 (4.87) 37.76 (4.85) <0.0001 Plasma n-3 FA, mg/L 2.69 (0.58) 3.34 (0.97) 3.29 (0.96) 0.004 Plasma n-6 FA, mg*10/L 2.76 (0.45) 3.26 (0.44) 3.24 (0.45) <0.0001 CRP, mg/L 4.66 (1.76–15.20) 2.59 (1.30–5.56) 2.89 (1.36–5.85) <0.0001 Urine protein, mg/L 0 (0–5) 0.0 (0.0–0.0) 0.0 (0.0–0.0) 0.04 Coronary heart disease, % 10.00 5.93 5.21 0.43 Stroke, % 15.38 6.70 4.12 0.01 Diabetes, % 11.11 10.98 11.89 0.67 Hypertension, % 60.00 48.52 17.28 0.38 Peripheral artery disease, % 20.00 8.09 4.98 0.004 Congestive heart failure, % 44.44 8.56 4.86 <0.0001 COPD, % 26.92 7.18 8.65 0.35 Cancer, % 11.11 6.79 5.79 0.30 Continuous variables expressed as mean (SD) or as median (IQR). 1 From age-, sex-, and BMI-adjusted linear or multinomial logistic regression models as appropriate. Open in new tab In age- and sex-adjusted analysis, plasma PUFAs were also inversely associated with urine protein excretion at enrollment (β = −0.15, SE = 0.06, P = 0.02) (data not shown). From enrollment to the 3-year follow-up visit, creatinine clearance declined by 7.8 (12.2) mL/min (P <0.0001). Creatinine clearance at enrollment and the 3-year follow-up were highly correlated (r = 0.76; P <0.0001). The relationship between total plasma PUFAs at enrollment and mean change in creatinine clearance between baseline and the 3-year visit is shown in Fig. 1 . The analysis is adjusted for baseline creatinine clearance. From the lowest to the highest quintile of total plasma PUFAs, the mean declines in creatinine clearance were 8.8, 5.0, 4.8, 3.2, and 1.4 mL/min, (P <0.0001). Figure 1. Open in new tabDownload slide Relationship of total plasma PUFAs (in quintiles; mg/L) at enrollment with change in creatinine clearance between enrollment and 3-year follow-up visit after adjusting for baseline creatinine clearance (*P <0.05 in relation to the lowest quintile). P for linear trend across quintiles <0.0001. Figure 1. Open in new tabDownload slide Relationship of total plasma PUFAs (in quintiles; mg/L) at enrollment with change in creatinine clearance between enrollment and 3-year follow-up visit after adjusting for baseline creatinine clearance (*P <0.05 in relation to the lowest quintile). P for linear trend across quintiles <0.0001. The relationships of total plasma PUFAs, n-3 FA, linolenic acid, n-6 FA, linoleic acid, and arachidonic acid with change in creatinine clearance from baseline to follow-up was examined using multivariate linear regression models adjusted for covariates, including education, cigarette smoking (pack-years), MMSE score, energy intake, alcohol intake, LDL cholesterol, HDL cholesterol, cancer, cardiovascular disease, and hypertension (Table 2 ). Adjusting for baseline creatinine clearance and other confounders, higher total plasma PUFAs, n-3 FA, linolenic acid, n-6 FA, linoleic acid, and arachidonic acid at enrollment were significantly associated with lower decline in creatinine clearance from baseline to follow-up; participants with higher plasma PUFAs at enrollment had a smaller decline in creatinine clearance from baseline to follow-up. Table 2. Backward regression models relating change in creatinine clearance over 3-year follow-up and total plasma PUFAs at baseline. . Dependent change in creatinine clearance . . . β (SE) . P . Total plasma PUFAs, mg/L 0.43 (0.11) <0.0001 Plasma n-3 FA, mg/L 6.28 (1.29) <0.0001 Plasma linolenic acid, mg/L 0.19 (0.05) 0.0004 Plasma n-6 FA, mg*10/L 3.62 (1.19) 0.03 Plasma linoleic acid, mg*10/L 0.08 (0.03) 0.004 Plasma arachidonic acid, mg*10/L 0.24 (0.06) <0.0001 . Dependent change in creatinine clearance . . . β (SE) . P . Total plasma PUFAs, mg/L 0.43 (0.11) <0.0001 Plasma n-3 FA, mg/L 6.28 (1.29) <0.0001 Plasma linolenic acid, mg/L 0.19 (0.05) 0.0004 Plasma n-6 FA, mg*10/L 3.62 (1.19) 0.03 Plasma linoleic acid, mg*10/L 0.08 (0.03) 0.004 Plasma arachidonic acid, mg*10/L 0.24 (0.06) <0.0001 Data adjusted for creatinine clearance at baseline, education, cigarette-smoking pack-years, MMSE score, energy intake, alcohol consumption, LDL cholesterol, HDL cholesterol, self-reported cancer, cardiovascular disease, and hypertension. Covariates were selected using a Pearson correlation coefficient <0.10. Open in new tab Table 2. Backward regression models relating change in creatinine clearance over 3-year follow-up and total plasma PUFAs at baseline. . Dependent change in creatinine clearance . . . β (SE) . P . Total plasma PUFAs, mg/L 0.43 (0.11) <0.0001 Plasma n-3 FA, mg/L 6.28 (1.29) <0.0001 Plasma linolenic acid, mg/L 0.19 (0.05) 0.0004 Plasma n-6 FA, mg*10/L 3.62 (1.19) 0.03 Plasma linoleic acid, mg*10/L 0.08 (0.03) 0.004 Plasma arachidonic acid, mg*10/L 0.24 (0.06) <0.0001 . Dependent change in creatinine clearance . . . β (SE) . P . Total plasma PUFAs, mg/L 0.43 (0.11) <0.0001 Plasma n-3 FA, mg/L 6.28 (1.29) <0.0001 Plasma linolenic acid, mg/L 0.19 (0.05) 0.0004 Plasma n-6 FA, mg*10/L 3.62 (1.19) 0.03 Plasma linoleic acid, mg*10/L 0.08 (0.03) 0.004 Plasma arachidonic acid, mg*10/L 0.24 (0.06) <0.0001 Data adjusted for creatinine clearance at baseline, education, cigarette-smoking pack-years, MMSE score, energy intake, alcohol consumption, LDL cholesterol, HDL cholesterol, self-reported cancer, cardiovascular disease, and hypertension. Covariates were selected using a Pearson correlation coefficient <0.10. Open in new tab The relationship of total plasma PUFAs, n-3 FA, linolenic acid, n-6 FA, linoleic acid, arachidonic acid, and creatinine at enrollment with change in creatinine from baseline to follow-up was examined using multivariate linear regression models adjusted for covariates, including age, sex, BMI, education, cigarette smoking (pack-years), MMSE score, energy intake, alcohol intake, LDL cholesterol, HDL cholesterol, cancer, cardiovascular disease, and hypertension (Table 3 ). Table 3. Backward regression models relating creatinine at baseline, creatinine at the 3-year follow-up study, and plasma PUFAs at baseline. . Creatinine at baseline . . Creatinine at 3-years follow-up study1 . . . β (SE) . P . β (SE) . P . Total plasma PUFAs, mg/L −0.005 (0.002) 0.06 −0.005 (0.002) 0.03 Plasma n-3 FA, mg/L −0.02 (0.02) 0.31 −0.12 (0.03) <0.0001 Plasma linolenic acid, mg/L −0.002 (0.001) 0.10 −0.003 (0.001) 0.002 Plasma n-6 FA, mg*10/L −0.05 (0.02) 0.03 −0.04 (0.02) 0.11 Plasma linoleic acid, mg*10/L −0.001 (0.0005) 0.07 −0.001 (0.0006) 0.01 Plasma arachidonic acid, mg*10/L −0.0001 (0.001) 0.95 −0.002 (0.0006) 0.003 . Creatinine at baseline . . Creatinine at 3-years follow-up study1 . . . β (SE) . P . β (SE) . P . Total plasma PUFAs, mg/L −0.005 (0.002) 0.06 −0.005 (0.002) 0.03 Plasma n-3 FA, mg/L −0.02 (0.02) 0.31 −0.12 (0.03) <0.0001 Plasma linolenic acid, mg/L −0.002 (0.001) 0.10 −0.003 (0.001) 0.002 Plasma n-6 FA, mg*10/L −0.05 (0.02) 0.03 −0.04 (0.02) 0.11 Plasma linoleic acid, mg*10/L −0.001 (0.0005) 0.07 −0.001 (0.0006) 0.01 Plasma arachidonic acid, mg*10/L −0.0001 (0.001) 0.95 −0.002 (0.0006) 0.003 Data adjusted for age, sex, BMI, education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL cholesterol, HDL cholesterol, self-reported cancer, cardiovascular disease, and hypertension. Covariates were selected using a Pearson correlation coefficient <0.10. 1 Also adjusted for creatinine at baseline. Open in new tab Table 3. Backward regression models relating creatinine at baseline, creatinine at the 3-year follow-up study, and plasma PUFAs at baseline. . Creatinine at baseline . . Creatinine at 3-years follow-up study1 . . . β (SE) . P . β (SE) . P . Total plasma PUFAs, mg/L −0.005 (0.002) 0.06 −0.005 (0.002) 0.03 Plasma n-3 FA, mg/L −0.02 (0.02) 0.31 −0.12 (0.03) <0.0001 Plasma linolenic acid, mg/L −0.002 (0.001) 0.10 −0.003 (0.001) 0.002 Plasma n-6 FA, mg*10/L −0.05 (0.02) 0.03 −0.04 (0.02) 0.11 Plasma linoleic acid, mg*10/L −0.001 (0.0005) 0.07 −0.001 (0.0006) 0.01 Plasma arachidonic acid, mg*10/L −0.0001 (0.001) 0.95 −0.002 (0.0006) 0.003 . Creatinine at baseline . . Creatinine at 3-years follow-up study1 . . . β (SE) . P . β (SE) . P . Total plasma PUFAs, mg/L −0.005 (0.002) 0.06 −0.005 (0.002) 0.03 Plasma n-3 FA, mg/L −0.02 (0.02) 0.31 −0.12 (0.03) <0.0001 Plasma linolenic acid, mg/L −0.002 (0.001) 0.10 −0.003 (0.001) 0.002 Plasma n-6 FA, mg*10/L −0.05 (0.02) 0.03 −0.04 (0.02) 0.11 Plasma linoleic acid, mg*10/L −0.001 (0.0005) 0.07 −0.001 (0.0006) 0.01 Plasma arachidonic acid, mg*10/L −0.0001 (0.001) 0.95 −0.002 (0.0006) 0.003 Data adjusted for age, sex, BMI, education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL cholesterol, HDL cholesterol, self-reported cancer, cardiovascular disease, and hypertension. Covariates were selected using a Pearson correlation coefficient <0.10. 1 Also adjusted for creatinine at baseline. Open in new tab At enrollment, total plasma PUFAs and n-6 FA were significantly associated with lower creatinine. Adjusting for baseline creatinine, higher total plasma PUFAs, n-3 FA, linolenic acid, linoleic acid, and arachidonic acid at enrollment were significantly associated with a smaller decline in creatinine from baseline to follow-up. Excluding participants with creatinine clearance <60 mL/min at baseline, and adjusting for multiple confounders, those with higher plasma PUFAs at enrollment had a lower risk of developing renal insufficiency, defined by a creatinine clearance <60 mL/min at the 3-year follow-up. Participants with higher plasma N-6 and N-3 fatty acids tended to have a lower risk of developing renal insufficiency (Table 4 ), although this trend was not statistically significant. N-3 FA concentrations were inversely associated with risk of developing renal insufficiency or dying over the 3-year follow-up (Table 5 ). Table 4. Odds of developing renal insufficiency (n = 134) (defined by a creatinine clearance <60 mL/min, among the 398 participants with creatinine clearance >60 mL/min at baseline) over 3-year follow-up according to plasma baseline PUFA level. . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.94 (0.89–0.99 0.015 Plasma n-3 FA, mg/L 0.60 (0.35–1.04) 0.06 0.68 (0.37–1.28) 0.23 Plasma linolenic acid, mg/L 0.98 (0.96–1.01) 0.10 0.99 (0.96–1.01) 0.30 Plasma n-6 FA, mg*10/L 0.68 (0.42–1.10) 0.11 0.96 (0.88–1.00) 0.09 Plasma linoleic acid, mg*10/L 0.99 (0.98–1.01) 0.59 1.00 (0.98–1.02) 0.98 Plasma arachidonic acid, mg*10/L 0.99 (0.96–1.02) 0.44 0.98 (0.96–1.02) 0.49 . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.94 (0.89–0.99 0.015 Plasma n-3 FA, mg/L 0.60 (0.35–1.04) 0.06 0.68 (0.37–1.28) 0.23 Plasma linolenic acid, mg/L 0.98 (0.96–1.01) 0.10 0.99 (0.96–1.01) 0.30 Plasma n-6 FA, mg*10/L 0.68 (0.42–1.10) 0.11 0.96 (0.88–1.00) 0.09 Plasma linoleic acid, mg*10/L 0.99 (0.98–1.01) 0.59 1.00 (0.98–1.02) 0.98 Plasma arachidonic acid, mg*10/L 0.99 (0.96–1.02) 0.44 0.98 (0.96–1.02) 0.49 1 Adjusted for education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL cholesterol, HDL cholesterol, cancer, congestive heart failure, cardiovascular disease, and hypertension. Open in new tab Table 4. Odds of developing renal insufficiency (n = 134) (defined by a creatinine clearance <60 mL/min, among the 398 participants with creatinine clearance >60 mL/min at baseline) over 3-year follow-up according to plasma baseline PUFA level. . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.94 (0.89–0.99 0.015 Plasma n-3 FA, mg/L 0.60 (0.35–1.04) 0.06 0.68 (0.37–1.28) 0.23 Plasma linolenic acid, mg/L 0.98 (0.96–1.01) 0.10 0.99 (0.96–1.01) 0.30 Plasma n-6 FA, mg*10/L 0.68 (0.42–1.10) 0.11 0.96 (0.88–1.00) 0.09 Plasma linoleic acid, mg*10/L 0.99 (0.98–1.01) 0.59 1.00 (0.98–1.02) 0.98 Plasma arachidonic acid, mg*10/L 0.99 (0.96–1.02) 0.44 0.98 (0.96–1.02) 0.49 . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.94 (0.89–0.99 0.015 Plasma n-3 FA, mg/L 0.60 (0.35–1.04) 0.06 0.68 (0.37–1.28) 0.23 Plasma linolenic acid, mg/L 0.98 (0.96–1.01) 0.10 0.99 (0.96–1.01) 0.30 Plasma n-6 FA, mg*10/L 0.68 (0.42–1.10) 0.11 0.96 (0.88–1.00) 0.09 Plasma linoleic acid, mg*10/L 0.99 (0.98–1.01) 0.59 1.00 (0.98–1.02) 0.98 Plasma arachidonic acid, mg*10/L 0.99 (0.96–1.02) 0.44 0.98 (0.96–1.02) 0.49 1 Adjusted for education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL cholesterol, HDL cholesterol, cancer, congestive heart failure, cardiovascular disease, and hypertension. Open in new tab Table 5. Odds of developing renal insufficiency (defined by a creatinine clearance <60 mL/min, among the 398 participants with creatinine clearance >60 mL/min at baseline) or dying over 3-year follow-up according to plasma baseline PUFA level. . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.98 (0.94–1.03) 0.36 Plasma n-3 FA, mg/L 0.47 (0.30–0.74) 0.001 0.53 (0.31–0.88) 0.02 Plasma linolenic acid, mg/L 0.96 (0.95–0.98) 0.002 0.97 (0.95–0.99) 0.008 Plasma n-6 FA, mg*10/L 0.71 (0.48–1.05) 0.08 0.93 (0.57–1.54) 0.19 Plasma linoleic acid, mg*10/L 0.98 (0.97–0.99) 0.01 0.99 (0.98–1.01) 0.30 Plasma arachidonic acid, mg*10/L 0.98 (0.96–1.01) 0.07 0.98 (0.95–1.01) 0.16 . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.98 (0.94–1.03) 0.36 Plasma n-3 FA, mg/L 0.47 (0.30–0.74) 0.001 0.53 (0.31–0.88) 0.02 Plasma linolenic acid, mg/L 0.96 (0.95–0.98) 0.002 0.97 (0.95–0.99) 0.008 Plasma n-6 FA, mg*10/L 0.71 (0.48–1.05) 0.08 0.93 (0.57–1.54) 0.19 Plasma linoleic acid, mg*10/L 0.98 (0.97–0.99) 0.01 0.99 (0.98–1.01) 0.30 Plasma arachidonic acid, mg*10/L 0.98 (0.96–1.01) 0.07 0.98 (0.95–1.01) 0.16 1 Adjusted for education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL cholesterol, HDL cholesterol, cancer, congestive heart failure, cardiovascular disease, and hypertension. Open in new tab Table 5. Odds of developing renal insufficiency (defined by a creatinine clearance <60 mL/min, among the 398 participants with creatinine clearance >60 mL/min at baseline) or dying over 3-year follow-up according to plasma baseline PUFA level. . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.98 (0.94–1.03) 0.36 Plasma n-3 FA, mg/L 0.47 (0.30–0.74) 0.001 0.53 (0.31–0.88) 0.02 Plasma linolenic acid, mg/L 0.96 (0.95–0.98) 0.002 0.97 (0.95–0.99) 0.008 Plasma n-6 FA, mg*10/L 0.71 (0.48–1.05) 0.08 0.93 (0.57–1.54) 0.19 Plasma linoleic acid, mg*10/L 0.98 (0.97–0.99) 0.01 0.99 (0.98–1.01) 0.30 Plasma arachidonic acid, mg*10/L 0.98 (0.96–1.01) 0.07 0.98 (0.95–1.01) 0.16 . Crude odds ratio (95% CI) . P . Fully adjusted odds ratio (95% CI)1 . P . Total plasma PUFAs, mg/L 0.95 (0.91–0.99) 0.02 0.98 (0.94–1.03) 0.36 Plasma n-3 FA, mg/L 0.47 (0.30–0.74) 0.001 0.53 (0.31–0.88) 0.02 Plasma linolenic acid, mg/L 0.96 (0.95–0.98) 0.002 0.97 (0.95–0.99) 0.008 Plasma n-6 FA, mg*10/L 0.71 (0.48–1.05) 0.08 0.93 (0.57–1.54) 0.19 Plasma linoleic acid, mg*10/L 0.98 (0.97–0.99) 0.01 0.99 (0.98–1.01) 0.30 Plasma arachidonic acid, mg*10/L 0.98 (0.96–1.01) 0.07 0.98 (0.95–1.01) 0.16 1 Adjusted for education, cigarette smoking pack-years, MMSE score, energy intake, alcohol, LDL cholesterol, HDL cholesterol, cancer, congestive heart failure, cardiovascular disease, and hypertension. Open in new tab Discussion This study shows that older adults with low total plasma PUFA concentrations have a greater decline in creatinine and creatinine clearance over 3 years of follow-up than those with higher concentrations of total plasma PUFAs. In addition, participants with lower baseline plasma PUFAs and free of renal insufficiency were significantly more likely to develop renal insufficiency at the 3-year follow-up than those with higher plasma PUFAs. These findings suggest that a higher dietary intake of PUFAs, both n-3 FA and n-6 FA, may be protective against progression to chronic kidney disease, and are consistent with observations from animal models that show that PUFA supplementation reduces progression of renal disease (6). The observation that total plasma PUFAs and also ω-3 fatty and ω-6 fatty acids separately appear to have a beneficial effect on renal function require consideration. In fact, ω-3 polyunsaturated fatty acids are generally considered more beneficial than ω-6 fatty acids (17). However, recent data showed that both ω-6 (18) and ω-3 (19) fatty acids have antiinflammatory properties. PUFAs are present in high concentrations not only in fish oil but also in vegetable oils. For example, large quantities of ω-6 fatty acids are present in sunflower oil, soybean/corn oil, and safflower oil, whereas large quantities of ω-3 fatty acids are present in flax oil and hemp oil. To our knowledge, this is the first human study that clearly observed a protective effect of PUFA on the age-associated decline of renal function. Strengths of this study are the relatively large sample size and the prospective, longitudinal analysis. The study is limited in that plasma PUFA levels were measured only at enrollment and not at the 3-year follow-up visit. Also, the loss of some respondents to follow-up may have influenced our findings. The mechanisms by which PUFAs may protect the kidneys from damage in older adults remain unknown. Data from the literature suggest that PUFAs may be antiinflammatory (6). The main histopathological changes associated with the progression of renal disease in older adults are fibrosis, glomerulonephritis, progressive tubulointerstitial injury, and renal fibrosis (6). In addition, experimental data have clearly shown that aging is associated with increased oxidative stress (19), enhanced tubular cell apoptosis (20), and exacerbation of glomerular inflammatory responses induced by glomerular fibrin deposition (21). Dietary fish oil supplementation has been shown to reduce progression of renal disease among patients with IgA nephropathy (22) and to suppress mesangial cell activation and proliferation in animal models (23). PUFAs may reduce inflammation through several possible pathways, such as reduction of nitric oxide (24), downregulation of TNF-α (25), and modulation of protein kinases (26). Plasma PUFA concentrations have previously been associated with lower levels of markers of inflammation in the InCHIANTI study (9); in the present analysis, markers of inflammation were not included as confounders in the analysis because proinflammatory cytokines are considered to be in the causal pathway between plasma PUFAs and progression of renal disease (6)(27). A recent double-blind, placebo-controlled trial in 103 middle-aged men and women showed that increased dietary intake of α-linolenic acid lowered CRP levels (28). Another trial involving 60 subjects with active rheumatoid arthritis showed that n-3 FA supplementation decreased CRP (29). Our findings prompt the hypothesis that a diet rich in PUFAs may be protective against the decline in renal function that is common with aging. A Mediterranean-style diet that is characterized by a relatively high consumption of fish and low consumption of saturated fats has been shown to be protective against cardiovascular disease (30)(31), markers of inflammation (32), and cancer (33). Further work is needed to confirm the association between plasma PUFAs and renal function in other cohorts of older persons and provide enough evidence to translate these findings into clinical trials. Grant/funding Support: This work was supported by National Institute on Aging Contracts N01-AG-916413, N01-AG-821336, N01-AG-5-0002, and NIA Grant R01 AG027012. The research was also supported in part by the Intramural Research Program, National Institute on Aging, NIH. Financial Disclosures: None declared. 1 " Nonstandard abbreviations: PUFA, polyunsaturated fatty acid; CRP, C-reactive protein; IL, interleukin; InCHIANTI, Aging in the Chianti Area; BMI, body mass index; MMSE, Mini-Mental Status Examination; FA, fatty acid; Ccr, Cockcroft-Gault formula; Scr, serum creatinine. References 1 Stevens LA, Coresh J, Greene T, Levey AS. Assessing kidney function: measured and estimated glomerular filtration rate. N Engl J Med 2006 ; 354 : 2473 -2483. Crossref Search ADS PubMed 2 Xue JL, Ma JZ, Louis TA, Collins AJ. Forecast of the number of patients with end-stage renal disease in the United States by the year 2010. J Am Soc Nephrol 2001 ; 12 : 2753 -2758. PubMed 3 de Zeeuw D, Hillege HL, de Jong PE. The kidney, a cardiovascular risk marker, and a new target for therapy. Kidney Int 2005 ; 68 (Suppl 98): S25 -S29. 4 Giannelli SV, Patel KV, Windham BG, Pizzarelli F, Ferrucci L, Guralnik JM. Magnitude of underascertainment of impaired kidney function in older adults with normal serum creatinine. J Am Geriatr Soc 2007 ; 55 : 816 -823. Crossref Search ADS PubMed 5 Fried LF, Orchard TJ, Kasiske BL. Effect of lipid reduction on the progression of renal disease: a meta-analysis. Kidney Int 2001 ; 59 : 260 -269. Crossref Search ADS PubMed 6 Baggio B, Musaachio E, Priante G. Polyunsaturated fatty acids and renal fibrosis: pathophysiologic link and potential clinical implications. J Nephrol 2005 ; 18 : 362 -367. PubMed 7 Calder PC. Dietary modification of inflammation with lipids. Proc Nutr Soc 2002 ; 61 : 345 -358. Crossref Search ADS PubMed 8 de Caterina R, Massaro M. Omega-3 fatty acids and the regulation of expression of endothelial pro-atherogenic and pro-inflammatory genes. J Membr Biol 2005 ; 206 : 103 -116. Crossref Search ADS PubMed 9 Ferrucci L, Cherubini A, Bandinelli S, Bartali B, Corsi A, Lauretani F, et al. Relationship of plasma polyunsaturated fatty acids to circulating inflammatory markers. J Clin Endocrinol Metab 2006 ; 91 : 439 -446. Crossref Search ADS PubMed 10 Ferrucci L, Bandinelli S, Benvenuti E, Di Iorio A, Macchi C, Harris TB, et al. Subsystems contributing to the decline in ability to walk: bridging the gap between epidemiology and geriatric practice in the InCHIANTI study. J Am Geriatr Soc 2000 ; 48 : 1618 -1625. Crossref Search ADS PubMed 11 Schrager MA, Metter EJ, Simonsick E, Ble A, Bandinelli S, Lauretani F, et al. Sarcopenic obesity and inflammation in the InCHIANTI study. J Appl Physiol 2007 ; 102 : 919 -925. Crossref Search ADS PubMed 12 Pisani P, Faggiano F, Krogh V, Palli D, Vineis P, Berrino F. Relative validity and reproducibility of a food frequency dietary questionnaire for use in the Italian EPIC centres. Int J Epidemiol 1997 ; 26 (Suppl 1): S152 -S160. Crossref Search ADS PubMed 13 Guralnik JM, Fried LP, Simonsick EM, Kasper D, Lafferty ME. The Women’s Health and Aging Study: health and social characteristics of older women with disability 1995 National Institute on Aging Bethesda, MD. NIH publ. no. 95-4009. 14 Folstein MF, Folstein SE, McHugh PR. Mini-Mental State: a practical method for grading the cognitive state of patients for the clinician. J Psychiatr Res 1975 ; 12 : 189 -198. Crossref Search ADS PubMed 15 Cockcroft DW, Gault MH. Prediction of creatinine clearance from serum creatinine. Nephron 1976 ; 16 : 31 -41. Crossref Search ADS PubMed 16 Levey AS, Coresh J, Greene T, Stevens LA, Zhang YL, Hendriksen S, et al. Using standardized serum creatinine values in the Modification of Diet in Renal Disease study equation for estimating glomerular filtration rate. Ann Intern Med 2006 ; 145 : 247 -254. Crossref Search ADS PubMed 17 Simopoulos AP, Leaf A, Salem N, Jr. Workshop statement on the essentiality of and recommended dietary intakes for omega-6 and omega-3 fatty acids. Prostaglandins Leukot Essent Fatty Acids 2000 ; 63 : 119 -121. Crossref Search ADS PubMed 18 Kapoor R, Huang YS. Gamma linolenic acid: an antiinflammatory omega-6 fatty acid. Curr Pharm Biotechnol 2006 ; 7 : 531 -534. Crossref Search ADS PubMed 19 Kim JW, Zou Y, Yoon S, Lee JH, Kim YK, Yu BP, et al. Vascular aging: molecular modulation of the prostanoid cascade by calorie restriction. J Gerontol A Biol Sci Med Sci 2004 ; 59 : B876 -B885. Crossref Search ADS PubMed 20 Qiao X, Chen X, Wu D, Ding R, Wang J, Hong Q, et al. Mitochondrial pathway is responsible for aging-related increase of tubular cell apoptosis in renal ischemia/reperfusion injury. J Gerontol A Biol Sci Med Sci 2005 ; 60 : 830 -839. Crossref Search ADS PubMed 21 Xi C, Chen X, Sun X, Shi S, Feng Z, Wang J, et al. Effects of alterations of glomerular fibrin deposition on renal inflammation in rats at different age stages. J Gerontol A Biol Sci Med Sci 2005 ; 60 : 1099 -1110. Crossref Search ADS PubMed 22 Donadio JV, Jr, Bergstralh EJ, Offord KP, Spencer DC, Holley KE. A controlled trial of fish oil in IgA nephropathy. N Engl J Med 1994 ; 331 : 1194 -1199. Crossref Search ADS PubMed 23 Grande JP, Walker HJ, Holub BJ, Warner GM, Keller DM, Haugen JD, et al. Suppressive effects of fish oil on mesangial cell proliferation in vitro and in vivo. Kidney Int 2000 ; 57 : 1027 -1040. Crossref Search ADS PubMed 24 Das UN. Long-chain polyunsaturated fatty acids interact with nitric oxide, superoxide anion, and transforming growth factor-beta to prevent human essential hypertension. Eur J Clin Nutr 2004 ; 58 : 195 -203. Crossref Search ADS PubMed 25 Kielar ML, Jeyarajah DR, Zhou XJ, Lu CY. Docosahexaenoic acid ameliorates urine ischemic acute renal failure and prevents increases in mRNA abundance for both TNF-alpha and inducible nitric oxide synthase. J Am Soc Nephrol 2003 ; 14 : 389 -396. Crossref Search ADS PubMed 26 de Jonge HW, Dekkers DH, Lamers JM. Polyunsaturated fatty acids and signalling via phospholipase C-beta and A2 in myocardium. Mol Cell Biochem 1996 ; 157 : 199 -210. Crossref Search ADS PubMed 27 Locatelli F, Canaud B, Eckardt KU, Stenvinkel P, Wanner C, Zoccali C. Oxidative stress in end-stage renal disease: an emerging threat to patient outcome. Nephrol Dial Transplant 2003 ; 18 : 1272 -1280. Crossref Search ADS PubMed 28 Bemelmans WJ, Lefrandt JD, Feskens EJ, van Haelst PL, Broer J, Meyboom-de Jon B, et al. Increased alpha-linolenic acid intake lowers C-reactive protein, but has no effect on markers of atherosclerosis. Eur J Clin Nutr 2004 ; 58 : 1083 -1089. Crossref Search ADS PubMed 29 Sundrarjun T, Komindr S, Archararit N, Dahlan W, Puchaiwatananon O, Angthararak S, et al. Effects of n-e fatty acids on serum interleukin-6, tumour necrosis factor-alpha and soluble tumour necrosis factor receptor p55 in active rheumatoid arthritis. J Int Med Res 2004 ; 32 : 443 -454. Crossref Search ADS PubMed 30 de Lorgeril M, Salen P. The Mediterranean-style diet for the prevention of cardiovascular diseases. Public Health Nutr 2006 ; 9 : 118 -123. Crossref Search ADS PubMed 31 Zatonski WA, Willett W. Changes in dietary fat and declining coronary heart disease in Poland: population based study. BMJ 2005 ; 331 : 187 -188. Crossref Search ADS PubMed 32 Zampelas A, Panagiotakos DB, Pitsavos C, Das UN, Chrysohoou C, Skoumas Y, et al. Fish consumption among healthy adults is associated with decreased levels of inflammatory markers related to cardiovascular disease: the ATTICA study. J Am Coll Cardiol 2005 ; 46 : 120 -124. Crossref Search ADS PubMed 33 Serra-Majem L, Roman B, Estruch R. Scientific evidence of interventions using the Mediterranean diet: a systematic review. Nutr Rev 2006 ; 64 : S27 -S47. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Chim, Stephen S, C;Shing, Tristan K, F;Hung, Emily C, W;Leung,, Tak-yeung;Lau,, Tze-kin;Chiu, Rossa W, K;Dennis Lo, Y, M

doi: 10.1373/clinchem.2007.097972pmid: 18218722

Abstract Background: The discovery of circulating fetal nucleic acids in maternal plasma has opened up new possibilities for noninvasive prenatal diagnosis. MicroRNAs (miRNAs), a class of small RNAs, have been intensely investigated recently because of their important regulatory role in gene expression. Because nucleic acids of placental origin are released into maternal plasma, we hypothesized that miRNAs produced by the placenta would also be released into maternal plasma. Methods: We systematically searched for placental miRNAs in maternal plasma to identify miRNAs that were at high concentrations in placentas compared with maternal blood cells and then investigated the stability and filterability of this novel class of pregnancy-associated markers in maternal plasma. Results: In a panel of TaqMan MicroRNA Assays available for 157 well-established miRNAs, 17 occurred at concentrations >10-fold higher in the placentas than in maternal blood cells and were undetectable in postdelivery maternal plasma. The 4 most abundant of these placental miRNAs (miR-141, miR-149, miR-299-5p, and miR-135b) were detectable in maternal plasma during pregnancy and showed reduced detection rates in postdelivery plasma. The plasma concentration of miR-141 increased as pregnancy progressed into the third trimester. Compared with mRNA encoded by CSH1 [chorionic somatomammotropin hormone 1 (placental lactogen)], miR-141 was even more stable in maternal plasma, and its concentration did not decrease after filtration. Conclusion: We have demonstrated the existence of placental miRNAs in maternal plasma and provide some information on their stability and physical nature. These findings open up a new class of molecular markers for pregnancy monitoring. The discovery of fetal nucleic acids in the plasma of pregnant women (1)(2)(3) has led to the development of a number of noninvasive prenatal diagnostic tests. Circulating fetal DNA in maternal plasma has been used for prenatal investigations of fetal rhesus D status (4)(5), sex-linked diseases (6), and β-thalassemia (7). Our group has used the epigenetic DNA differences that exist between placenta and maternal blood cells to develop universal fetal-DNA markers (8)(9)(10) that are sex- and polymorphism-independent. Furthermore, quantitative aberrations of fetus-derived mRNA transcripts have been shown in conditions such as preeclampsia (11) and fetal aneuploidies (12). The feasibility of detecting fetal chromosomal aneuploidies in maternal plasma has been demonstrated with a fetus-derived PLAC43 (placenta-specific 4) mRNA transcript (13). These findings suggest that the detection of circulating fetal nucleic acids holds much promise for noninvasive prenatal diagnosis. Recent studies on microRNAs (miRNAs)1 offer possibilities for developing yet another class of molecular markers. miRNAs are short (19–25 nucleotides), single-stranded, and nonprotein-coding RNAs (14)(15)(16) that regulate gene expression by binding to the 3′ untranslated region of the target mRNAs (17) and function in diverse biological processes, including development (18), differentiation (19), apoptosis (20), and oncogenesis (21)(22). Nucleic acids of placental origin were previously shown to be released into maternal plasma (2)(8); hence, it would be interesting to investigate whether miRNAs produced by the placenta are also released into maternal plasma. Because ribonuclease activity has been observed in blood (23), however, it has been unclear whether miRNA species in plasma are sufficiently stable to be detected. We describe our systematic search for placental miRNAs in maternal plasma and our investigation into some of the physical properties of the miRNAs we have discovered. Materials and Methods participant recruitment and sample collection This study was approved by the local institutional review board. Informed consent was obtained from women who had uncomplicated singleton pregnancies and who were patients in the Department of Obstetrics and Gynaecology at the Prince of Wales Hospital, Hong Kong. We recruited the first- and second-trimester participants in this study from women attending the antenatal clinic and recruited the third-trimester participants from women undergoing elective cesarean delivery. Samples of maternal peripheral blood (12 mL) were collected into tubes containing EDTA. For the third-trimester participants, we collected placental tissues and postdelivery maternal blood immediately and at 24 h after delivery, respectively. sample processing To harvest cell-free plasma, we centrifuged maternal blood samples twice at 4 °C. After the first centrifugation at 1600g for 10 min, we centrifuged the supernatant at 16 000g for 10 min to remove blood cells (24). We harvested maternal blood cells (including leukocytes and erythrocytes) by centrifuging the blood cells obtained in the first centrifugation at 2300g for 5 min to remove residual plasma. We then added Trizol LS reagent (Invitrogen) in volumetric ratios of 1:0.8 on 3:1 to the harvested maternal plasma and maternal blood cells, respectively. Placental tissues were preserved in RNAlater (Ambion) immediately following delivery. rna extraction We extracted total RNA containing small RNA molecules with the Trizol LS or Trizol reagent (Invitrogen) and the mirVana miRNA Isolation Kit (Ambion) (see Methods in the Data Supplement that accompanies the online version of this article athttp://www.clinchem.org/content/vol54/issue3). After the chloroform-addition steps and phase separation, we mixed the aqueous layer with 1.25 volumes of absolute ethanol, loaded the solution onto the cartridge provided in the mirVana miRNA Isolation Kit, and processed the sample (see Methods in the online Data Supplement). To minimize DNA contamination, we treated the eluted RNA preparation with DNase I (Invitrogen) (see Methods in the online Data Supplement). For miRNA profiling, we further diluted RNA preparations obtained from samples of placentas, maternal blood cells, or postdelivery maternal plasma to 1 mg/L, according to absorbance readings at 260 nm. For the other arms of our study, we did not dilute RNA preparations further. quantification of mirnas by real-time quantitative reverse transcription–pcr analysis We used the TaqMan MicroRNA Assay (Applied Biosystems), which has been shown to be highly specific for the intended miRNA but not for its longer preprocessed precursors or for other highly homologous miRNAs that differ in sequence by as little as 1 nucleotide (25). This assay entailed a 2-step quantitative reverse transcription–PCR (qRT-PCR)—reverse transcription of an miRNA of 19–25 nucleotides, priming with a stem-loop primer into a longer cDNA that is amenable to amplification, and quantification by a TaqMan-based qPCR. For each miRNA, we assessed the detection limits of the qRT-PCR assay and quantified the numbers of miRNA copies in samples with a calibration curve (see Methods in the online Data Supplement). miRNA concentrations were expressed as the number of copies per nanogram of RNA extracted from tissue or as the number of copies per liter of plasma. Substrate specificity and assay imprecision also were evaluated for selected miRNAs (see Methods in the online Data Supplement), although the data we obtained cannot be extrapolated directly to all of the other assays. mirna profiling of placentas, maternal blood cells, and postdelivery maternal plasma We quantified 157 well-established miRNAs in RNA extracts from 5 third-trimester placentas, 5 samples of maternal blood cells, and 5 samples of postdelivery maternal plasma. We used the TaqMan Array Human MicroRNA Panel v1.0 (Early Access) (Applied Biosystems), which contains 157 TaqMan MicroRNA Assays, including the respective reverse-transcription primers, PCR primers, and TaqMan probe. For each assay, we added 2.5 μL (2.5 ng) of the RNA extracted from each of the 15 samples for the reverse-transcription reaction. We used the TaqMan MicroRNA Reverse Transcription Kit (Applied Biosystems) for reverse transcription in a 25 μL of total reaction volume (see Methods in the online Data Supplement). We used the TaqMan Universal PCR Master Mix (Applied Biosystems) for the PCR (see Methods in the online Data Supplement). For each miRNA, we measured the median concentrations in the samples of placenta and maternal blood cells and evaluated the detection rates in the postdelivery samples of maternal plasma (Table 1 ; see Table in the online Data Supplement). We calculated the fold-change in concentration by dividing the median miRNA concentration in the placenta sample by that in the sample of maternal blood cells. Table 1. Placental miRNAs identified as candidates for pregnancy-associated markers in maternal plasma.1 miRNA . Concentration in placenta, copies/ng extracted RNA . Concentration in maternal blood cells, copies/ng extracted RNA . hsa-miR-141 760 000 (590 000–760 000) 6600 (6000–6800) hsa-miR-135b 35 000 (29 000–45 000) 0 (0–0) hsa-miR-149 15 000 (14 000–16 000) 0 (0–0) hsa-miR-299 11 000 (8300–12 000) 360 (350–710) hsa-miR-154* 11 000 (8000–11 000) 560 (500–960) hsa-miR-34c 9100 (6900–11 000) 0 (0–0) hsa-miR-200b 4100 (2500–5500) 390 (350–500) hsa-miR-139 3900 (3800–5700) 0 (0–0) hsa-miR-154 3400 (2900–3400) 210 (170–230) hsa-miR-368 1900 (1900–2000) 84 (76–170) hsa-miR-373 1900 (1000–3100) 73 (61–76) hsa-miR-137 1900 (1600–2400) 0 (0–0) hsa-miR-184 1800 (1800–2500) 110 (100–160) hsa-miR-372 960 (240–1700) 0 (0–0) hsa-miR-371 410 (180–580) 0 (0–0) hsa-miR-34b 330 (280–460) 30 (0–38) hsa-miR-337 84 (78–99) 0 (0–0) miRNA . Concentration in placenta, copies/ng extracted RNA . Concentration in maternal blood cells, copies/ng extracted RNA . hsa-miR-141 760 000 (590 000–760 000) 6600 (6000–6800) hsa-miR-135b 35 000 (29 000–45 000) 0 (0–0) hsa-miR-149 15 000 (14 000–16 000) 0 (0–0) hsa-miR-299 11 000 (8300–12 000) 360 (350–710) hsa-miR-154* 11 000 (8000–11 000) 560 (500–960) hsa-miR-34c 9100 (6900–11 000) 0 (0–0) hsa-miR-200b 4100 (2500–5500) 390 (350–500) hsa-miR-139 3900 (3800–5700) 0 (0–0) hsa-miR-154 3400 (2900–3400) 210 (170–230) hsa-miR-368 1900 (1900–2000) 84 (76–170) hsa-miR-373 1900 (1000–3100) 73 (61–76) hsa-miR-137 1900 (1600–2400) 0 (0–0) hsa-miR-184 1800 (1800–2500) 110 (100–160) hsa-miR-372 960 (240–1700) 0 (0–0) hsa-miR-371 410 (180–580) 0 (0–0) hsa-miR-34b 330 (280–460) 30 (0–38) hsa-miR-337 84 (78–99) 0 (0–0) 1 Data are presented as the median (IQR). The miRNAs in this table showed both (a) concentrations in the placenta >10-fold higher than in maternal blood cells and (b) lack of detection (detection rate = 0%) in 5 postdelivery samples of maternal plasma. Other miRNAs not fulfilling both of these criteria are shown in the Table in the online Data Supplement. Open in new tab Table 1. Placental miRNAs identified as candidates for pregnancy-associated markers in maternal plasma.1 miRNA . Concentration in placenta, copies/ng extracted RNA . Concentration in maternal blood cells, copies/ng extracted RNA . hsa-miR-141 760 000 (590 000–760 000) 6600 (6000–6800) hsa-miR-135b 35 000 (29 000–45 000) 0 (0–0) hsa-miR-149 15 000 (14 000–16 000) 0 (0–0) hsa-miR-299 11 000 (8300–12 000) 360 (350–710) hsa-miR-154* 11 000 (8000–11 000) 560 (500–960) hsa-miR-34c 9100 (6900–11 000) 0 (0–0) hsa-miR-200b 4100 (2500–5500) 390 (350–500) hsa-miR-139 3900 (3800–5700) 0 (0–0) hsa-miR-154 3400 (2900–3400) 210 (170–230) hsa-miR-368 1900 (1900–2000) 84 (76–170) hsa-miR-373 1900 (1000–3100) 73 (61–76) hsa-miR-137 1900 (1600–2400) 0 (0–0) hsa-miR-184 1800 (1800–2500) 110 (100–160) hsa-miR-372 960 (240–1700) 0 (0–0) hsa-miR-371 410 (180–580) 0 (0–0) hsa-miR-34b 330 (280–460) 30 (0–38) hsa-miR-337 84 (78–99) 0 (0–0) miRNA . Concentration in placenta, copies/ng extracted RNA . Concentration in maternal blood cells, copies/ng extracted RNA . hsa-miR-141 760 000 (590 000–760 000) 6600 (6000–6800) hsa-miR-135b 35 000 (29 000–45 000) 0 (0–0) hsa-miR-149 15 000 (14 000–16 000) 0 (0–0) hsa-miR-299 11 000 (8300–12 000) 360 (350–710) hsa-miR-154* 11 000 (8000–11 000) 560 (500–960) hsa-miR-34c 9100 (6900–11 000) 0 (0–0) hsa-miR-200b 4100 (2500–5500) 390 (350–500) hsa-miR-139 3900 (3800–5700) 0 (0–0) hsa-miR-154 3400 (2900–3400) 210 (170–230) hsa-miR-368 1900 (1900–2000) 84 (76–170) hsa-miR-373 1900 (1000–3100) 73 (61–76) hsa-miR-137 1900 (1600–2400) 0 (0–0) hsa-miR-184 1800 (1800–2500) 110 (100–160) hsa-miR-372 960 (240–1700) 0 (0–0) hsa-miR-371 410 (180–580) 0 (0–0) hsa-miR-34b 330 (280–460) 30 (0–38) hsa-miR-337 84 (78–99) 0 (0–0) 1 Data are presented as the median (IQR). The miRNAs in this table showed both (a) concentrations in the placenta >10-fold higher than in maternal blood cells and (b) lack of detection (detection rate = 0%) in 5 postdelivery samples of maternal plasma. Other miRNAs not fulfilling both of these criteria are shown in the Table in the online Data Supplement. Open in new tab detection of placental mirnas in maternal plasma We conducted TaqMan MicroRNA Assays for miR-16, miR-29a, miR-141, miR-149, miR-299-5p, and miR-135b (Applied Biosystems). To maximize the detection rates of these miRNAs in maternal plasma, we used more concentrated RNA preparations (2.5 μL, no dilution) in the reverse-transcription reaction. The other qRT-PCR steps were the same as those described in the 2 previous sections. filtration studies of placental mirna and mrna in maternal plasma To investigate whether the pregnancy-specific miRNA molecules in maternal plasma were associated with subcellular particles, as was previously demonstrated for placental mRNA (2), we filtered samples of maternal plasma. We divided each of 15 processed samples of third-trimester maternal plasma into 4 0.8-mL aliquots. We filtered 3 of the aliquots through a filter with a pore size of 5 μm, 0.45 μm, or 0.22 μm (Millex-GV; Millipore) and left the fourth aliquot unfiltered. We then extracted the RNA from the plasma samples with 1 mL of Trizol LS and quantified the miR-141 concentration in the plasma sample. stability of purified mirna and mrna in plasma We spiked 210 ng of purified RNA from placental tissues into 6 identical aliquots (0.8 mL each) of a plasma sample freshly collected from a randomly selected male individual and incubated the aliquots at room temperature for 0 s, 5 s, 15 s, 60 s, 1 h, and 2 h. At the end of the incubation period, we immediately added 1 mL Trizol LS reagent to stop any ribonuclease activity and processed the aliquots as described above. We quantified placenta-produced miR-141 and CSH1 mRNA transcripts in these plasma aliquots according to the methods described above and as previously reported (2). We analyzed another aliquot of this plasma sample with no added placental RNA as a control for the presence of any endogenous CSH1 transcript and miR-141. statistical analysis Statistical analyses were performed with SigmaStat 3.0 software (SPSS). Results presence of readily detectable quantities of mir-16 in maternal plasma miR-16 is ubiquitous in almost all somatic tissues (26), but its presence in plasma has not yet been explored. We detected miR-16 in maternal plasma collected from 6 third-trimester women at a median concentration of 3.4 × 1010 copies/L [interquartile range (IQR), 2.8–3.7 × 1010 copies/L], indicating that miRNA exists in maternal plasma in readily detectable concentrations and validating our protocol for the effective extraction and detection of short RNA species from plasma samples. The SDs for qRT-PCR assays of miR-16 including and excluding the RNA-extraction step were 0.26 and 0.16 threshold cycles (Ct), respectively. substrate specificity of the taqman microrna assay To rule out the possibility that the TaqMan MicroRNA Assay nonspecifically detects any contaminating genomic DNA in RNA preparations, we treated an RNA preparation with different combinations of DNase I and/or RNase A and assayed for miR-141, which was known to exist at a detectable concentration in this RNA preparation before any treatment. Before any treatment, we detected miR-141 at 2300 copies/ng of extracted RNA. After treatment with DNase I alone, we still detected this miRNA at 2000 copies/ng. After we treated the RNA preparation with RNase A alone or with DNase I plus RNase A, miR-141 decreased to nearly undetectable concentrations (0 copies/ng and 3 copies/ng, respectively). These results suggest that the TaqMan MicroRNA Assay for miR-141 detects RNA but not DNA. identification of placental mirna in maternal plasma Subsequent to detecting the ubiquitous miR-16 in plasma, we investigated the existence of other of miRNA species in maternal plasma that might be associated with pregnancy. Investigators have previously detected placental mRNA transcripts in maternal plasma, including mRNAs encoded by CSH1, CGB (chorionic gonadotropin, beta polypeptide), and CRH (corticotropin releasing hormone), and have described their rapid clearance from maternal plasma upon delivery of the fetus (2)(11). We hypothesized that placental miRNAs are detectable in maternal plasma and hence profiled the production of 157 miRNAs in 5 third-trimester placentas. By analogy with another previous report that circulating DNA in the plasma of nonpregnant individuals is predominately derived from hematopoietic cells (27), we further hypothesized that the majority of plasma miRNAs that are not associated with pregnancy also originate in the hematopoietic compartment. Because the aim of our study was to identify pregnancy-associated miRNAs in maternal plasma, we compared the miRNA profiles of the placental samples for these 5 pregnancies with the corresponding samples of maternal blood cells. We identified 34 miRNAs that were present in the placenta at concentrations >10-fold higher than in maternal blood cells (Table 1 ; see Table in the online Data Supplement). Ideally, pregnancy-associated markers should also disappear from the maternal plasma after delivery of the fetus. Hence, we considered only the 17 placental miRNAs that were not detected in the 24-h postdelivery maternal plasma as candidate markers in this phase of the study (Table 1 ). detection rates and clearance kinetics of placental mirnas in maternal plasma The detection rates of mRNA transcripts in maternal plasma are reportedly directly related to their concentrations in placental tissues (28). Hence, we reasoned that miRNAs present in high concentrations in the placenta would be more readily detectable in maternal plasma. We therefore investigated whether 4 miRNAs with the highest concentrations in the placenta (miR-141, miR-149, miR-299-5p, and miR-135b) are also present in maternal plasma. We measured detection rates and clearance kinetics in maternal plasma for the 4 selected miRNAs with a protocol that required a higher initial amount of total RNA. All 4 miRNAs were detected in postpartum maternal plasma at reduced median concentrations and reduced detection rates (Fig. 1 , A–D). In particular, the median postpartum concentrations of miR-141 and miR-149 decreased by ≥18-fold. In contrast, we found no systematic change in the concentration of miR-29a, which occurred at similar concentrations in the placenta and maternal blood cells (see Table in the online Data Supplement). miR-29a was used as a positive control for the successful extraction of RNA from all samples (Fig. 1E ). Figure 1. Open in new tabDownload slide Concentrations of miR-141 (A), miR-149 (B), miR-299-5p (C), miR-135b (D), and miR-29a (E) in maternal plasma before and at 24 h after delivery. Pairs of samples from the same pregnancy are depicted by identical symbols connected by a line. A zero value denotes a concentration lower than the respective detection limits: 4 × 105 copies/L for miR-141 and 4 × 104 copies/L for the 4 other miRNAs. The detection rates for the 10 predelivery and 10 postdelivery plasma samples were 100% and 50%, respectively, for miR-141, 80% and 0% for miR-149, 50% and 20% for miR-299-5p, and 20% and 0% for miR-135b. Wilcoxon signed rank tests revealed significant differences in plasma concentrations before and after delivery for miR-141 (P = 0.002) and miR-149 (P = 0.002), but not for miR-299-5p (P = 0.063) and miR-135b (P = 0.5). miR-29A was detected in all samples before and after delivery, with no significant change in concentration (P = 0.432). Figure 1. Open in new tabDownload slide Concentrations of miR-141 (A), miR-149 (B), miR-299-5p (C), miR-135b (D), and miR-29a (E) in maternal plasma before and at 24 h after delivery. Pairs of samples from the same pregnancy are depicted by identical symbols connected by a line. A zero value denotes a concentration lower than the respective detection limits: 4 × 105 copies/L for miR-141 and 4 × 104 copies/L for the 4 other miRNAs. The detection rates for the 10 predelivery and 10 postdelivery plasma samples were 100% and 50%, respectively, for miR-141, 80% and 0% for miR-149, 50% and 20% for miR-299-5p, and 20% and 0% for miR-135b. Wilcoxon signed rank tests revealed significant differences in plasma concentrations before and after delivery for miR-141 (P = 0.002) and miR-149 (P = 0.002), but not for miR-299-5p (P = 0.063) and miR-135b (P = 0.5). miR-29A was detected in all samples before and after delivery, with no significant change in concentration (P = 0.432). variation of placental mir-141 in maternal plasma with gestational age Because miR-141 was most readily detected in third-trimester maternal plasma, we investigated its occurrence in plasma during the first and second trimesters. The median gestational ages of the fetuses at the time of blood collection for the first, second, and third trimesters were 13.0 (IQR, 12.8–13.5) weeks, 17.4 (IQR, 17.3–17.6) weeks, and 38.5 (IQR, 38.3–38.6) weeks, respectively. Overall, we observed a trend of increasing miR-141 concentration with gestational age (Fig. 2 ). Figure 2. Open in new tabDownload slide Concentrations of miR-141 in first-, second-, and third-trimester samples of maternal plasma. The line within each box denotes the median. The horizontal borders of each box denote the 25th and 75th percentiles, and the limits of the vertical lines (“whiskers”) denote the 10th and 90th percentiles. Filled circles indicate data points outside the 10th and 90th percentiles. A zero value denotes a concentration lower than the miR-141 detection limit of 4 × 105 copies/L. The detection rates were 80%, 100%, and 100% for the 10 plasma samples each obtained at the first, second, and third trimesters, respectively. Significant differences in miR-141 concentrations were noted in first-, second-, and third-trimester plasma samples (P = 0.002, Kruskal–Wallis test). Posthoc pairwise comparisons revealed significant differences in plasma miR-141 concentrations between first- and third-trimester samples and between second- and third-trimester samples (P <0.05, Student–Newman–Keuls test). The plasma miR-141 concentration was observed to increase with gestational age (Spearman correlation coefficient, r, = 0.58; P <0.001). Figure 2. Open in new tabDownload slide Concentrations of miR-141 in first-, second-, and third-trimester samples of maternal plasma. The line within each box denotes the median. The horizontal borders of each box denote the 25th and 75th percentiles, and the limits of the vertical lines (“whiskers”) denote the 10th and 90th percentiles. Filled circles indicate data points outside the 10th and 90th percentiles. A zero value denotes a concentration lower than the miR-141 detection limit of 4 × 105 copies/L. The detection rates were 80%, 100%, and 100% for the 10 plasma samples each obtained at the first, second, and third trimesters, respectively. Significant differences in miR-141 concentrations were noted in first-, second-, and third-trimester plasma samples (P = 0.002, Kruskal–Wallis test). Posthoc pairwise comparisons revealed significant differences in plasma miR-141 concentrations between first- and third-trimester samples and between second- and third-trimester samples (P <0.05, Student–Newman–Keuls test). The plasma miR-141 concentration was observed to increase with gestational age (Spearman correlation coefficient, r, = 0.58; P <0.001). the effects of filtration of maternal plasma on circulating placental mirna and mrna species Because miR-141 was the most readily detectable miRNA in predelivery maternal plasma, we chose this miRNA for filtration studies to further elucidate the molecular characteristics of pregnancy-associated miRNA in maternal plasma. We detected miR-141 in 100% (15 of 15) of the plasma samples in all 4 filtration groups (no filtration or filtration through a 5-μm, 0.45-μm, or 0.22-μm filter) and observed no consistent change in its concentration in any group (P = 0.257, Friedman test; Fig. 3A ). Figure 3. Open in new tabDownload slide Concentrations of a placental miRNA (miR-141) (A) and a placental mRNA (CSH1 transcript) (B) in maternal plasma after no filtration or filtration with filters of different pore sizes. The horizontal borders of each box denote the 25th and 75th percentiles, and the limits of the vertical lines (“whiskers”) denote the 10th and 90th percentiles. Filled circles indicate data points outside the 10th and 90th percentiles. A zero value denotes a concentration lower than the detection limit of 4 × 105 copies/L for miR-141 or 4 × 104 copies/L for the CSH1 transcript. miR-141 was detected in 100% (15 of 15) of plasma samples in all 4 groups, whereas the CSH1 transcript was detected in 87%, 60%, 40%, and 0% of the 15 plasma samples in the no-filtration group and the 5-μm, 0.45-μm, and 0.22-μm filter groups, respectively. Figure 3. Open in new tabDownload slide Concentrations of a placental miRNA (miR-141) (A) and a placental mRNA (CSH1 transcript) (B) in maternal plasma after no filtration or filtration with filters of different pore sizes. The horizontal borders of each box denote the 25th and 75th percentiles, and the limits of the vertical lines (“whiskers”) denote the 10th and 90th percentiles. Filled circles indicate data points outside the 10th and 90th percentiles. A zero value denotes a concentration lower than the detection limit of 4 × 105 copies/L for miR-141 or 4 × 104 copies/L for the CSH1 transcript. miR-141 was detected in 100% (15 of 15) of plasma samples in all 4 groups, whereas the CSH1 transcript was detected in 87%, 60%, 40%, and 0% of the 15 plasma samples in the no-filtration group and the 5-μm, 0.45-μm, and 0.22-μm filter groups, respectively. In contrast, we detected the CSH1 transcript, a placental mRNA that reportedly is readily detected in predelivery maternal plasma (2), at reduced rates as the plasma was filtered through filters of increasingly smaller pore size (Fig. 3B ). A statistical analysis of the filtration groups showed a significant difference (P <0.001, Friedman test). A pairwise analysis confirmed a statistically significant difference between the no-filter and 0.45-μm filtration groups, and between the no-filter and 0.22-μm filtration groups (P <0.05, Dunn test). Overall, the comparisons of paired samples not filtered or filtered through a 0.45-μm filter showed that the concentration of the CSH1 transcript decreased by a median of 2.3-fold (IQR, 1.6- to 3.5-fold) in the samples with detectable concentrations. stability of purified placental mirna and mrna in plasma The different effects of filtration on the concentrations of miR-141 and the CSH1 transcript in maternal plasma prompted us to further investigate the stability of these 2 different classes of RNA molecules in their purified forms. We purified RNA containing both miR-141 and the CSH1 transcript from the placenta, spiked it into aliquots of plasma from a randomly chosen male individual, and incubated the aliquots for 0 s, 5 s, 15 s, 60 s, 1 h, and 2 h. We did not detect miR-141 and the CSH1 transcript in the plasma aliquot without the spiked placental RNA (Fig. 4 ). At the beginning of the incubation period (0 s), the concentrations of miR-141 and the CSH1 transcript were 1.3 × 109 copies/L and 5.5 × 108 copies/L, respectively; Fig. 4 presents the concentrations of detectable miR-141 and CSH1 transcript as percentages of these values. In the first 15 s of incubation, the miR-141 concentration decreased 24-fold, from 100.0% to 4.1%. Between 60 s and 2 h of incubation, the miR-141 concentration remained at about 1% (about 107 copies/L). In contrast, the concentration of the CSH1 transcript decreased in the first 15 s by >330-fold, from 100.0% to 0.3%. After 60 s, no CSH1 transcript could be amplified. We obtained similar results for both miR-141 and the CSH1 transcript when we repeated this experiment with 110 ng, 430 ng, and 830 ng of placental RNA (data not shown). Thus, the effect was independent of the amount of RNA added. Figure 4. Open in new tabDownload slide Percentages of exogenous purified miRNA (miR-141) and CSH1 mRNA remaining in plasma after incubation for different times. Concentrations of added miR-141 and CSH1 mRNA are presented as percentages of the starting concentrations at 0 s of incubation in plasma. Figure 4. Open in new tabDownload slide Percentages of exogenous purified miRNA (miR-141) and CSH1 mRNA remaining in plasma after incubation for different times. Concentrations of added miR-141 and CSH1 mRNA are presented as percentages of the starting concentrations at 0 s of incubation in plasma. Discussion Concordant with our hypothesis that the placenta releases nucleic acids into the maternal plasma, our present data have shown that placental miRNAs exist in maternal plasma in readily detectable quantities. By systematically searching a panel of 157 miRNA assays, we have identified 17 placental miRNAs as candidate markers for monitoring pregnancy in maternal plasma. Furthermore, we detected 4 placental miRNAs (miR-141, miR-149, miR-299-5p, and miR-135b) at higher rates in the maternal plasma before delivery than after delivery. Hence, we conclude that these miRNA species are associated with pregnancy. The highest detection rate and highest median concentration that we observed in predelivery maternal plasma were for miR-141, which was the most abundant placental miRNA detected in this study (Table 1 ). The reductions in the median concentrations of miR-141 and miR-149 in maternal plasma after delivery were statistically significant, whereas those for miR-299-5p and miR-135b were not. The detection rates for miR-299-5p and miR-135b were ≤50%. This result is consistent with the lower concentrations in the placenta observed for miR-299-5p and miR-135b, compared with miR-141. Thus, these miRNAs were not as readily detectable in maternal plasma as miR-141. We predict that when improved methods for extracting and quantifying miRNA become available, the detection rates for some of these placental miRNAs in maternal plasma will also increase. We also predict that as we expand our search to all of the 530 miRNAs that have been identified in humans to date (29), more pregnancy-associated miRNAs will be identified in maternal plasma. Our data also have shown that the plasma concentration of a placental miRNA, miR-141, increased as the pregnancy progressed into the third trimester. This increase in miR-141 in maternal plasma may reflect an increase in the size of the placenta or an increased concentration of miR-141 in the third-trimester placenta. The quantification of placental miRNAs in maternal plasma may offer a noninvasive means for monitoring gene regulation in the placenta. Recently, aberrant concentrations of miR-210 and miR-182 were found in preeclamptic placentas delivered at <37 weeks of gestation, compared with the concentrations for non-preeclamptic spontaneous preterm deliveries at matched gestation times (30). It would therefore be useful to investigate whether the aberrant concentrations of miRNAs in placentas involved in preeclampsia and other pathologic conditions are also reflected in maternal plasma. To develop this novel class of markers for clinical use, we explored the physical nature of a placental miRNA, miR-141, in maternal plasma and compared it with the properties of a placental mRNA, the CSH1 transcript, which have previously been established (2). The exceptional stability of cell-free mRNA in plasma is probably due to its association with subcellular particles (2)(31), e.g., syncytiotrophoblast microparticles (32). Much to our surprise, however, we were not able to filter out placental miR-141 in maternal plasma, even if 0.22-μm filters were used, in contrast to the placental CSH1 transcript. Hence, unlike the CSH1 transcript, miR-141 in maternal plasma is not predominantly associated with subcellular particles >0.22 μm in diameter. The question of whether miR-141 is significantly associated with particles <0.22 μm requires further exploration via ultracentrifugation, which can pellet particles the size of viruses. We further speculated about whether miRNAs themselves are intrinsically more stable in plasma than mRNAs. When we added a purified preparation of exogenous placental RNA to a sample of male plasma with no detectable endogenous miR-141 and CSH1 transcript, the added miR-141 demonstrated a slower rate of reduction and remained detectable for longer periods than CSH1 mRNA. Because these data were based on purified miRNA and mRNA in the absence of any protection (e.g., through association with particles) from nuclease activity in the plasma, the higher stability of the former offers an explanation for why miRNA species are readily detectable in plasma even if they are not associated with subcellular particles. In summary, we have shown that placental miRNAs represent a novel class of fetal nucleic acid markers in maternal plasma. We have also provided the first demonstration of the application of a search strategy for systematically discovering pregnancy-associated miRNAs in maternal plasma. Because miRNAs are exceptionally stable in plasma, they hold promise as markers in the clinical setting. The measurement of miRNAs in maternal plasma for prenatal monitoring and diagnosis would be an interesting future research direction. The biological significance of placental miRNAs in maternal plasma requires further elucidation, but an intriguing possibility is that these small molecules are taken up by cells exposed to the maternal circulation and may modulate gene expression of the maternal compartment. Furthermore, our work has opened up the possibilities that miRNA signatures specific to malignancies (21)(22) or viruses (33) are also released into the plasma and can be suitable for disease detection and monitoring. Grant/funding Support: Several of the authors are supported by the Research Grants Council of the Hong Kong Special Administrative Region Government, China (CERG Grant No. CUHK462907, CUHK Direct Grant Nos. 2005.1.027 and 2005.2.010). Y.M.D.L. is supported by the Chair Professorship Scheme of the Li Ka Shing Foundation. Financial Disclosures: Y.M.D.L., R.W.K.C., and S.S.C.C. hold patents for and have filed patent applications on aspects of the use of fetal nucleic acids in maternal plasma for noninvasive prenatal diagnosis, a proportion of which have been licensed to Sequenom, Inc. Y.M.D.L. is a consultant for Sequenom Inc. 1 " Nonstandard abbreviations: miRNA, microRNA; qRT-PCR, quantitative reverse transcription–PCR; IQR, interquartile range. 2 " These authors contributed equally to this study. 3 " Human genes: PLAC4, placenta-specific 4; CSH1, chorionic somatomammotropin hormone 1 (placental lactogen); CGB, chorionic gonadotropin, beta polypeptide; CRH, corticotropin releasing hormone. References 1 Lo YMD, Corbetta N, Chamberlain PF, Rai V, Sargent IL, Redman CW, et al. Presence of fetal DNA in maternal plasma and serum. Lancet 1997 ; 350 : 485 -487. Crossref Search ADS PubMed 2 Ng EKO, Tsui NBY, Lau TK, Leung TN, Chiu RWK, Panesar NS, et al. mRNA of placental origin is readily detectable in maternal plasma. Proc Natl Acad Sci U S A 2003 ; 100 : 4748 -4753. Crossref Search ADS PubMed 3 Lo YMD, Chiu RWK. Prenatal diagnosis: progress through plasma nucleic acids. Nat Rev Genet 2007 ; 8 : 71 -77. Crossref Search ADS PubMed 4 Lo YMD, Hjelm NM, Fidler C, Sargent IL, Murphy MF, Chamberlain PF, et al. Prenatal diagnosis of fetal RhD status by molecular analysis of maternal plasma. N Engl J Med 1998 ; 339 : 1734 -1738. Crossref Search ADS PubMed 5 Faas BH, Beuling EA, Christiaens GC, von dem Borne AE, van der Schoot CE. Detection of fetal RHD-specific sequences in maternal plasma [Letter]. Lancet 1998 ; 352 : 1196 . 6 Costa JM, Benachi A, Gautier E. New strategy for prenatal diagnosis of X-linked disorders [Letter]. N Engl J Med 2002 ; 346 : 1502 . Crossref Search ADS PubMed 7 Chiu RWK, Lau TK, Leung TN, Chow KCK, Chui DH, Lo YMD. Prenatal exclusion of beta thalassaemia major by examination of maternal plasma. Lancet 2002 ; 360 : 998 -1000. Crossref Search ADS PubMed 8 Chim SSC, Tong YK, Chiu RWK, Lau TK, Leung TN, Chan LYS, et al. Detection of the placental epigenetic signature of the maspin gene in maternal plasma. Proc Natl Acad Sci U S A 2005 ; 102 : 14753 -14758. Crossref Search ADS PubMed 9 Chiu RWK, Chim SSC, Wong IHN, Wong CSC, Lee WS, To KF, et al. Hypermethylation of RASSF1A in human and rhesus placentas. Am J Pathol 2007 ; 170 : 941 -950. Crossref Search ADS PubMed 10 Chan KCA, Ding C, Gerovassili A, Yeung SW, Chiu RWK, Leung TN, et al. Hypermethylated RASSF1A in maternal plasma: a universal fetal DNA marker that improves the reliability of noninvasive prenatal diagnosis. Clin Chem 2006 ; 52 : 2211 -2218. Crossref Search ADS PubMed 11 Ng EKO, Leung TN, Tsui NBY, Lau TK, Panesar NS, Chiu RWK, et al. The concentration of circulating corticotropin-releasing hormone mRNA in maternal plasma is increased in preeclampsia. Clin Chem 2003 ; 49 : 727 -731. Crossref Search ADS PubMed 12 Ng EKO, El-Sheikhah A, Chiu RWK, Chan KCA, Hogg M, Bindra R, et al. Evaluation of human chorionic gonadotropin beta-subunit mRNA concentrations in maternal serum in aneuploid pregnancies: a feasibility study. Clin Chem 2004 ; 50 : 1055 -1057. Crossref Search ADS PubMed 13 Lo YMD, Tsui NBY, Chiu RWK, Lau TK, Leung TN, Heung MMS, et al. Plasma placental RNA allelic ratio permits noninvasive prenatal chromosomal aneuploidy detection. Nat Med 2007 ; 13 : 218 -223. Crossref Search ADS PubMed 14 Lee RC, Ambros V. An extensive class of small RNAs in Caenorhabditis elegans. Science 2001 ; 294 : 862 -864. Crossref Search ADS PubMed 15 Lau NC, Lim LP, Weinstein EG, Bartel DP. An abundant class of tiny RNAs with probable regulatory roles in Caenorhabditis elegans. Science 2001 ; 294 : 858 -862. Crossref Search ADS PubMed 16 Lagos-Quintana M, Rauhut R, Lendeckel W, Tuschl T. Identification of novel genes coding for small expressed RNAs. Science 2001 ; 294 : 853 -858. Crossref Search ADS PubMed 17 Lai EC. Micro RNAs are complementary to 3′ UTR sequence motifs that mediate negative post-transcriptional regulation. Nat Genet 2002 ; 30 : 363 -364. Crossref Search ADS PubMed 18 Krichevsky AM, King KS, Donahue CP, Khrapko K, Kosik KS. A microRNA array reveals extensive regulation of microRNAs during brain development. RNA 2003 ; 9 : 1274 -1281. Crossref Search ADS PubMed 19 Esau C, Kang X, Peralta E, Hanson E, Marcusson EG, Ravichandran LV, et al. MicroRNA-143 regulates adipocyte differentiation. J Biol Chem 2004 ; 279 : 52361 -52365. Crossref Search ADS PubMed 20 Baehrecke EH. miRNAs: micro managers of programmed cell death. Curr Biol 2003 ; 13 : R473 -R475. Crossref Search ADS PubMed 21 Michael MZ, O’Connor SM, van Holst Pellekaan NG, Young GP, James RJ. Reduced accumulation of specific microRNAs in colorectal neoplasia. Mol Cancer Res 2003 ; 1 : 882 -891. PubMed 22 Calin GA, Ferracin M, Cimmino A, Di Leva G, Shimizu M, Wojcik SE, et al. A microRNA signature associated with prognosis and progression in chronic lymphocytic leukemia. N Engl J Med 2005 ; 353 : 1793 -1801. Crossref Search ADS PubMed 23 Reddi KK, Holland JF. Elevated serum ribonuclease in patients with pancreatic cancer. Proc Natl Acad Sci U S A 1976 ; 73 : 2308 -2310. Crossref Search ADS PubMed 24 Chiu RWK, Poon LLM, Lau TK, Leung TN, Wong EM, Lo YMD. Effects of blood-processing protocols on fetal and total DNA quantification in maternal plasma. Clin Chem 2001 ; 47 : 1607 -1613. Crossref Search ADS PubMed 25 Chen C, Ridzon DA, Broomer AJ, Zhou Z, Lee DH, Nguyen JT, et al. Real-time quantification of microRNAs by stem-loop RT-PCR. Nucleic Acids Res 2005 ; 33 : e179 . Crossref Search ADS PubMed 26 Lagos-Quintana M, Rauhut R, Meyer J, Borkhardt A, Tuschl T. New microRNAs from mouse and human. RNA 2003 ; 9 : 175 -179. Crossref Search ADS PubMed 27 Lui YY, Chik KW, Chiu RWK, Ho CY, Lam CWK, Lo YMD. Predominant hematopoietic origin of cell-free DNA in plasma and serum after sex-mismatched bone marrow transplantation. Clin Chem 2002 ; 48 : 421 -427. Crossref Search ADS PubMed 28 Tsui NBY, Chim SSC, Chiu RWK, Lau TK, Ng EKO, Leung TN, et al. Systematic micro-array based identification of placental mRNA in maternal plasma: towards non-invasive prenatal gene expression profiling. J Med Genet 2004 ; 41 : 461 -467. Crossref Search ADS PubMed 29 Griffiths-Jones S, Grocock RJ, van Dongen S, Bateman A, Enright AJ. miRBase: microRNA sequences, targets and gene nomenclature. Nucleic Acids Res 2006 ; 34 : D140 -D144. Crossref Search ADS PubMed 30 Pineles BL, Romero R, Montenegro D, Tarca AL, Han YM, Kim YM, et al. Distinct subsets of microRNAs are expressed differentially in the human placentas of patients with preeclampsia. Am J Obstet Gynecol 2007 ; 196 : 261.e1 -262.e6. Crossref Search ADS 31 Ng EKO, Tsui NBY, Lam NYL, Chiu RWK, Yu SC, Wong SC, et al. Presence of filterable and nonfilterable mRNA in the plasma of cancer patients and healthy individuals. Clin Chem 2002 ; 48 : 1212 -1217. Crossref Search ADS PubMed 32 Gupta AK, Holzgreve W, Huppertz B, Malek A, Schneider H, Hahn S. Detection of fetal DNA and RNA in placenta-derived syncytiotrophoblast microparticles generated in vitro. Clin Chem 2004 ; 50 : 2187 -2190. Crossref Search ADS PubMed 33 Omoto S, Ito M, Tsutsumi Y, Ichikawa Y, Okuyama H, Brisibe EA, et al. HIV-1 nef suppression by virally encoded microRNA. Retrovirology 2004 ; 1 : 44 . Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Bruce,, Sara;Hannula-Jouppi,, Katariina;Lindgren, Cecilia, M;Lipsanen-Nyman,, Marita;Kere,, Juha

doi: 10.1373/clinchem.2007.098491pmid: 18202157

Abstract Background: Epigenetic studies, such as the measurement of DNA methylation, are important in the investigation of syndromes influenced by imprinted genes. Quick and accurate quantification of methylation at such genes can be of appreciable diagnostic aid. Methods: We first digested genomic DNA with methylation-sensitive restriction enzymes and used DNA without digestion as a control and nonmethylated λ DNA as an internal control for digestion efficiency. We then performed quantitative real-time PCR analyses with 6 unique PCR assays to investigate 4 imprinting control regions on chromosomes 7 and 11 in individuals with uniparental disomy of chromosome 7 (UPD7) and in control individuals. Results: Our validation of the method demonstrated both quantitative recovery and low methodologic imprecision. The imprinted loci on chromosome 7 behaved as expected in maternal UPD7 (100% methylation) and paternal UPD7 (<10% methylation). In controls, the mean (SD) for percent methylation at 2 previously well-studied restriction sites were 46% (6%) for both H19 and KCNQ1OT1, a result consistent with the previously observed methylation rate of approximately 50%. The methylation percentages of all investigated imprinted loci were normally distributed, implying that the mean and SD can be used as a reference for screening methylation loss or gain. Conclusion: The investigated loci are of particular importance for investigating the congenital Silver–Russell and Beckwith–Wiedemann syndromes; however, the method can also be applied to other imprinted regions. This method is easy to set up, has no PCR bias, requires small amounts of DNA, and can easily be applied to large patient populations for screening the loss or gain of methylation. Given the notion that DNA methylation and histone modifications can cause heritable characteristics that occur without changing the DNA sequence, epigenetic studies have become increasingly important. In mammals, a methyl group can be attached to a cytosine base when it is located 5′ to a guanosine, a process referred to as CpG methylation. This type of methylation is of interest in the contexts of various cancers, imprinting, and regulatory effects involving the promoters of many genes (1). Imprinting refers to gene expression specific to the parent of origin and is maintained through mitosis and reset in the germ line. Imprinting control regions (ICRs)1 , which are typically located in CpG islands in proximity to imprinted genes, govern the nonexpression of one allele, which is often accompanied by parent-specific methylation patterns (2). The importance of these genes for human development and growth has been demonstrated in various syndromes, such as the Angelman, Prader–Willi, Beckwith–Wiedemann, and Silver–Russell syndromes (2). Genetic irregularities that influence the dosage of imprinted genes, such as uniparental disomy (UPD), duplications/deletions, and loss of the parent-specific methylation markers (loss/gain of methylation) are frequently found in these syndromes (3)(4). Methylation studies can be used to identify such genetic aberrations. DNA methylation can be studied in site-specific, large-scale, and global manners. When the loss or gain of methylation at imprinted genes is investigated, the study of DNA methylation at specific sites is often sufficient and is the focus of the present study. One can distinguish 2 main types of approaches, those based on digestion with methylation-sensitive restriction enzymes and those based on bisulfite treatment of DNA (5). Digestion with methylation-sensitive restriction enzymes is typically followed by Southern blotting, which is cumbersome and has limitations. It requires several micrograms of DNA and allows the study of only 1 restriction site in each experiment. The use of hot-stop PCR after digestion with methylation-sensitive restriction enzymes has also been described (6). Bisulfite treatment of DNA is typically followed by sequencing or methylation-specific PCR assays (both quantitative and conventional) to monitor methylation at specific sites (7)(8)(9)(10). Bisulfite-based methods often enable a detailed study of all CpGs in a region of interest and are thus important for exploratory studies in regions where ICRs have yet to be defined. The study of the loss or gain of methylation at specific sites should be quantitative, because cell populations may be a mosaic with respect to the methylation status at a given site. For the study of the loss or gain of methylation at imprinted genes, investigators have used Southern blotting, different PCR-based methods, and a matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) approach with bisulfite-treated DNA (6)(7)(11)(12)(13)(14). We evaluated a site-specific methylation analytical approach that combines simple digestion with methylation-sensitive restriction enzymes and subsequent quantitative real-time PCR (qRT-PCR) analysis of previously characterized ICRs (15)(16)(17)(18). This approach offers a number of practical advances in the study of imprinted genes in that all steps can be performed in a microtiter plate format, several loci can be studied from the same digest, small amounts of DNA are needed, and the qRT-PCR methodology is widely available. Materials and Methods study participants Forty Finnish individuals (21 females, 19 males) of typical height [mean (SD), 161 (6.7) cm and 175 (6.1) cm, respectively] were used as controls. Included in the study were 7 patients with maternal UPD for chromosome 7 (matUPD7) (1 patient with segmental matUPD7q31-qter and 6 patients with whole-chromosome matUPD7) and 1 patient with whole-chromosome paternal UPD for chromosome 7 (patUPD7), all of whom have previously been described (19)(20)(21)(22)(23). For analyses of H19 (H19, imprinted maternally expressed transcript) methylation, we included 20 patients with Silver–Russell syndrome (11 males and 9 females), all of whom had been diagnosed as previously described (22). The study was approved by the appropriate ethics review boards at the University of Helsinki, Finland, and the Karolinska Institutet, Sweden. sssi methylation of λ dna We methylated 1 μg genomic λ DNA with 4 U of the CpG methyltransferase SssI in 20-μL reaction volumes with 1× NEBuffer 2 and 160 μmol/L S-adenosylmethionine (all reagents from New England Biolabs) for 2 h at 37 °C and subsequently inactivated the enzyme at 65 °C for 20 min (for the formulation of NEBuffer 2, see the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol54/issue3). Four reactions were run in parallel to ensure optimal methylation efficiency and were subsequently pooled. Methylated λ DNA (100%) was diluted with nonmethylated (0%) λ DNA to create a dilution series of 75%, 50%, and 25% methylation. We then digested these DNA dilutions with HpaII (New England Biolabs) and measured the methylation percentage as described below. measurement of dna methylation DNA methylation was assayed in genomic DNA extracted from whole blood. We digested 130 ng of DNA with 2 U of the methylation-sensitive restriction enzyme HpaII (New England Biolabs) in a 50-μL volume and incubated another 130 ng of DNA with 1× NEBuffer 1 only (for the formulation of NEBuffer 1, see the online Data Supplement). We included an aliquot of an amplified λ DNA fragment with an HpaII restriction site in both reactions as an internal control for digestion efficiency. To reduce the variance, we first prepared the genomic DNA, NEBuffer 1, and the internal control together and then divided it before adding the HpaII enzyme in 1× NEBuffer 1. We incubated DNA (with or without enzyme) overnight at 37 °C in a PCR machine (Mastercycler; Eppendorf) to ensure complete digestion and then inactivated the enzyme at 62 °C for 20 min. To test for methylation at additional imprinted sites, we digested all samples similarly with the methylation-sensitive restriction enzyme NotI. The digested DNA and corresponding undigested DNA were then assayed in parallel with 7 different assays using the SYBR Green dye. Primers were designed with Primer3 software (24) and tested by electronic PCR for specificity (http://genome.ucsc.edu/cgi-bin/hgPcr). Table 1 presents the primer sequences and their annealing temperatures used for the assays. We carried out qRT-PCR analyses in 20-μL volumes with 1× Power SYBR Green PCR Master Mix (Applied Biosystems), 5 ng DNA, and 200 nmol/L of each primer. Each sample was quantified in duplicate or triplicate. Typical cycling conditions were as follows: 95 °C for 10 min and 40 cycles [95 °C for 30 s, the annealing temperature (see Table 1 ) for 30 s, and 72 °C for 45 s], followed by 72 °C for 10 min and dissociation analysis. SYBR Green fluorescence was measured during the extension step. If we observed a primer peak in the dissociation curves, which occurred for some amplicons, we confirmed the specificity by electrophoresing the product on a 1.5% agarose gel. The primers in the MEST2 [mesoderm specific transcript homolog (mouse)] ICR required a touch-down PCR profile consisting of 3 cycles each at annealing temperatures 65 °C, 63 °C, and 61 °C, followed by 31 cycles with annealing at 58 °C. All PCR runs were manually inspected with 7500 Fast System SDS software, version 1.3.1 (Applied Biosystems). Table 1. Primer sequences and annealing temperatures. Assay . Forward primer (5′–3′) . Reverse primer (5′–3′) . Annealing temperature, °C . Product size, bp . Electronic PCR1 . GRB10-NotI AAAGCCCTCCATGTCTACCC CCTTCCTGGTTCTTGCTCTG 58 185 chr7:50818278–50818462 MEST-HpaII GGCGAAAACTCTACCGACAG AATTCGCATCAGGGTGAGAC 58 227 chr7:129919520–129919746 H19-HpaII ACGCTTCCCCTTCTGTCTC GGAATGTTAATGTCTGGCCACT 60 228 chr11:1979616–1980099 KCNQ1OT1-HpaII-1 TCCCAACTTCCATCCCAATA CCATCTGCACCTTATGGACA 58 180 chr11:2676885–2677061 KCNQ1OT1-NotI CTCTGCGTGATGTGTTCACC GTGGGGGCTTCAGAACATC 60 253 chr11:2678162–2678414 KCNQ1OT1-HpaII-2 ATGCCAAGCATTGCCATAA CAGAGTCCTGCATTCCAACA 60 150 chr11:2678672–2678823 λ GCATCATACCTTCCGAGCA CCACAGATTCAAGTGGACGA 60 627 Assay . Forward primer (5′–3′) . Reverse primer (5′–3′) . Annealing temperature, °C . Product size, bp . Electronic PCR1 . GRB10-NotI AAAGCCCTCCATGTCTACCC CCTTCCTGGTTCTTGCTCTG 58 185 chr7:50818278–50818462 MEST-HpaII GGCGAAAACTCTACCGACAG AATTCGCATCAGGGTGAGAC 58 227 chr7:129919520–129919746 H19-HpaII ACGCTTCCCCTTCTGTCTC GGAATGTTAATGTCTGGCCACT 60 228 chr11:1979616–1980099 KCNQ1OT1-HpaII-1 TCCCAACTTCCATCCCAATA CCATCTGCACCTTATGGACA 58 180 chr11:2676885–2677061 KCNQ1OT1-NotI CTCTGCGTGATGTGTTCACC GTGGGGGCTTCAGAACATC 60 253 chr11:2678162–2678414 KCNQ1OT1-HpaII-2 ATGCCAAGCATTGCCATAA CAGAGTCCTGCATTCCAACA 60 150 chr11:2678672–2678823 λ GCATCATACCTTCCGAGCA CCACAGATTCAAGTGGACGA 60 627 1 See http://genome.ucsc.edu/cgi-bin/hgPcr. Open in new tab Table 1. Primer sequences and annealing temperatures. Assay . Forward primer (5′–3′) . Reverse primer (5′–3′) . Annealing temperature, °C . Product size, bp . Electronic PCR1 . GRB10-NotI AAAGCCCTCCATGTCTACCC CCTTCCTGGTTCTTGCTCTG 58 185 chr7:50818278–50818462 MEST-HpaII GGCGAAAACTCTACCGACAG AATTCGCATCAGGGTGAGAC 58 227 chr7:129919520–129919746 H19-HpaII ACGCTTCCCCTTCTGTCTC GGAATGTTAATGTCTGGCCACT 60 228 chr11:1979616–1980099 KCNQ1OT1-HpaII-1 TCCCAACTTCCATCCCAATA CCATCTGCACCTTATGGACA 58 180 chr11:2676885–2677061 KCNQ1OT1-NotI CTCTGCGTGATGTGTTCACC GTGGGGGCTTCAGAACATC 60 253 chr11:2678162–2678414 KCNQ1OT1-HpaII-2 ATGCCAAGCATTGCCATAA CAGAGTCCTGCATTCCAACA 60 150 chr11:2678672–2678823 λ GCATCATACCTTCCGAGCA CCACAGATTCAAGTGGACGA 60 627 Assay . Forward primer (5′–3′) . Reverse primer (5′–3′) . Annealing temperature, °C . Product size, bp . Electronic PCR1 . GRB10-NotI AAAGCCCTCCATGTCTACCC CCTTCCTGGTTCTTGCTCTG 58 185 chr7:50818278–50818462 MEST-HpaII GGCGAAAACTCTACCGACAG AATTCGCATCAGGGTGAGAC 58 227 chr7:129919520–129919746 H19-HpaII ACGCTTCCCCTTCTGTCTC GGAATGTTAATGTCTGGCCACT 60 228 chr11:1979616–1980099 KCNQ1OT1-HpaII-1 TCCCAACTTCCATCCCAATA CCATCTGCACCTTATGGACA 58 180 chr11:2676885–2677061 KCNQ1OT1-NotI CTCTGCGTGATGTGTTCACC GTGGGGGCTTCAGAACATC 60 253 chr11:2678162–2678414 KCNQ1OT1-HpaII-2 ATGCCAAGCATTGCCATAA CAGAGTCCTGCATTCCAACA 60 150 chr11:2678672–2678823 λ GCATCATACCTTCCGAGCA CCACAGATTCAAGTGGACGA 60 627 1 See http://genome.ucsc.edu/cgi-bin/hgPcr. Open in new tab sequencing reactions PCR reactions were carried out in 25-μL reactions containing 0.5 ng/μL of genomic DNA, 2.5 mmol/L MgCl2, 400 μmol/L of each deoxynucleoside triphosphate, 800 nmol/L of each primer, and 0.03 U/μL of HotStarTaq DNA polymerase (Qiagen). The PCR was performed with an initial denaturation step at 95 °C for 10 min, followed by 40 cycles of 30 s at 95 °C, 30 s at 60 °C, and 45 s at 72 °C, with a final extension of 10 min at 72 °C. PCR products were dephosphorylated with 400 U/L shrimp alkaline phosphatase (Amersham Biosciences/GE Healthcare) and 2 kU/L exonuclease I (New England Biolabs) and were sequenced with the DYEnamic ET Dye Terminator Kit (Amersham Biosciences/GE Healthcare) following the manufacturer’s instructions. Purified sequencing products were resolved on a MegaBACE 1000 instrument with MegaBACE Long Read Matrix (Amersham Biosciences/GE Healthcare) and visualized with Sequence Analyzer v3.0 software (Amersham Biosciences/GE Healthcare). Sequences were assembled and analyzed with Pregap and Gap4 software (www.cbi.pku.edu.cn/tools/staden). statistical analysis The relative degree of methylation at the investigated loci was calculated by subtracting the mean of the undigested DNA threshold cycle (Ct) values (controls for the starting DNA amount) from the mean of the digested DNA threshold cycle (Ct) values. We then calculated the amount of undigested DNA relative to the amount of digested DNA, PM, by exponentiating the obtained difference with a base of 2: PM = 2x, where x = (mean Ctdigest) − (mean Ctnondigest). The methylation percentage was calculated as 1/PM. In cases in which a nonmethylated restriction site was assayed, digestion efficiency was calculated as: 1 − 1/PM. The SE of the methylation percentage was calculated as previously described (25). We used the Kolmogorov–Smirnov test to evaluate whether relative amounts of methylation at the investigated loci were normally distributed in our sample by comparing the observed data to a randomized normal distribution with a mean and SD derived from the observed data. We used 2-sided Student t-tests to analyze differences between means. All statistical analyses were performed in the R statistical environment (26). Results quality assessments and validation We used a combination of digestion with methylation-sensitive restriction enzymes and subsequent qRT-PCR analysis to measure methylation (see Fig. 1 in the Data Supplement that accompanies the online version of this article at http://www.clinchem.org/content/vol54/issue3). To assess the quantitative recovery and imprecision of methylation measurements, we digested a dilution series of SssI-methylated λ DNA samples with HpaII and measured the degree of methylation in quadruplicate. A linear regression analysis of the data revealed quantitative recovery across the entire methylation range (0%–100%), with a slope of 0.983 and an adjusted R2 of 0.989 in the dilution series (Fig. 1 ). We repeated the methylation measurement in an identical manner and obtained similar results (slope, 0.987; adjusted R2, 0.970). The individual SDs for the Ct values of replicates were low (mean, 0.09; median, 0.08). Because the observed SDs were well below the difference expected for an imprinted locus (Ct = 1), we proceeded to measure methylation effects. Figure 1. Open in new tabDownload slide Quantitative recovery of methylation percentage measurement. The observed methylation percentage plotted against the expected methylation percentage (from the dilution series). Data are presented as the mean (SE). The solid line corresponds to the result from the linear regression of observed on the expected methylation percentage; the dashed line indicates a line of slope 1. Figure 1. Open in new tabDownload slide Quantitative recovery of methylation percentage measurement. The observed methylation percentage plotted against the expected methylation percentage (from the dilution series). Data are presented as the mean (SE). The solid line corresponds to the result from the linear regression of observed on the expected methylation percentage; the dashed line indicates a line of slope 1. We studied a total of 6 loci in the 4 ICRs of GRB10 (growth factor receptor-bound protein 10) (7p14), MEST (7q32), H19, and KCNQ1OT1 (11p15.5) via digestion with the methylation-sensitive 4-base cutter HpaII and the methylation-sensitive rare-cutter NotI (Table 1 ). We measured mean and median digestion efficiencies by cutting nonmethylated λ DNA, and they were both 96%. The success rate of the PCR assays was 99.7%. To assess the imprecision of the reactions, we measured the SDs of replicate Ct values. The mean (SD) values were 0.14 (0.13) for GRB10-NotI, 0.08 (0.07) for MEST-HpaII, 0.11 (0.09) for H19-HpaII, 0.15 (0.13) for KCNQ1OT1-HpaII-1, 0.14 (0.13) for KCNQ1OT1-NotI, and 0.10 (0.08) for KCNQ1OT1-HpaII-2. The distribution of the SDs for the replicates is shown for each assay in Fig. 2 . Figure 2. Open in new tabDownload slide PCR assay–specific imprecision of replicate measurements. Histograms describing the distributions of SDs of replicate measurements for all 6 PCR assays. Figure 2. Open in new tabDownload slide PCR assay–specific imprecision of replicate measurements. Histograms describing the distributions of SDs of replicate measurements for all 6 PCR assays. imprinted loci on chromosome 7 behave as expected in matupd7 and patupd7 We analyzed restriction sites in the ICRs of GRB10 and MEST in 7 patients with matUPD7, in 1 patient with patUPD7, and in the 40 control individuals. We found the ICR restriction sites of GRB10 and MEST to be hypermethylated (approximately 100% methylation) in matUPD7 and hypomethylated in the patUPD7 patient (<10% methylation). The controls showed intermediate but distinct methylation patterns (Table 2 ). A patient with segmental matUPD7 covering 7q31-qter showed normal methylation for GRB10-NotI (7p14) and hypermethylation at MEST-HpaII (7q32), results that are consistent with the localization of the genes with respect to the UPD segment (Table 2 ). In the controls, we observed a lower mean methylation percentage and SD for GRB10-NotI and a higher methylation percentage for MEST-HpaII. We also assayed the patUPD7 and matUPD7s for the imprinted loci on chromosome 11 (see below), and the results revealed the typical methylation patterns (Table 2 ), confirming the normality of these data apart from chromosome 7. Table 2. Methylation percentages observed at imprinted loci. Assay . Chromosome . Restriction site previously studied? . Controls (n = 40), % . . . Pnormality2 . matUPD7 (n = 7), %1 . . . patUPD7, % . . . . Mean . SD . Median . . Mean . SD . Median . . GRB10-NotI 7p14 No 24.1 6.4 22.9 0.426 100.3 16.0 102.1 2.9 MEST-HpaII 7q32 No 60.5 6.7 59.7 0.842 96.4 10.5 98.7 8.5 H19-HpaII 11p15.5 Yes 45.7 6.1 44.7 0.806 47.2 4.7 46.0 40.6 KCNQ1OT1-HpaII-1 11p15.5 No 37.6 6.5 37.8 0.355 44.2 4.2 44.0 42.8 KCNQ1OT1-NotI 11p15.5 Yes 45.6 6.4 46.1 0.855 51.5 7.2 53.3 36.5 KCNQ1OT1-HpaII-2 11p15.5 No 44.3 4.7 44.2 0.841 53.7 5.5 53.1 40.3 Assay . Chromosome . Restriction site previously studied? . Controls (n = 40), % . . . Pnormality2 . matUPD7 (n = 7), %1 . . . patUPD7, % . . . . Mean . SD . Median . . Mean . SD . Median . . GRB10-NotI 7p14 No 24.1 6.4 22.9 0.426 100.3 16.0 102.1 2.9 MEST-HpaII 7q32 No 60.5 6.7 59.7 0.842 96.4 10.5 98.7 8.5 H19-HpaII 11p15.5 Yes 45.7 6.1 44.7 0.806 47.2 4.7 46.0 40.6 KCNQ1OT1-HpaII-1 11p15.5 No 37.6 6.5 37.8 0.355 44.2 4.2 44.0 42.8 KCNQ1OT1-NotI 11p15.5 Yes 45.6 6.4 46.1 0.855 51.5 7.2 53.3 36.5 KCNQ1OT1-HpaII-2 11p15.5 No 44.3 4.7 44.2 0.841 53.7 5.5 53.1 40.3 1 For GRB10-Notl, n = 6; data for segmental matUPD7q31.33-qter are not included. 2 Testing the null hypothesis that the data are normally distributed in a Kolmogorov–Smirnov test. Open in new tab Table 2. Methylation percentages observed at imprinted loci. Assay . Chromosome . Restriction site previously studied? . Controls (n = 40), % . . . Pnormality2 . matUPD7 (n = 7), %1 . . . patUPD7, % . . . . Mean . SD . Median . . Mean . SD . Median . . GRB10-NotI 7p14 No 24.1 6.4 22.9 0.426 100.3 16.0 102.1 2.9 MEST-HpaII 7q32 No 60.5 6.7 59.7 0.842 96.4 10.5 98.7 8.5 H19-HpaII 11p15.5 Yes 45.7 6.1 44.7 0.806 47.2 4.7 46.0 40.6 KCNQ1OT1-HpaII-1 11p15.5 No 37.6 6.5 37.8 0.355 44.2 4.2 44.0 42.8 KCNQ1OT1-NotI 11p15.5 Yes 45.6 6.4 46.1 0.855 51.5 7.2 53.3 36.5 KCNQ1OT1-HpaII-2 11p15.5 No 44.3 4.7 44.2 0.841 53.7 5.5 53.1 40.3 Assay . Chromosome . Restriction site previously studied? . Controls (n = 40), % . . . Pnormality2 . matUPD7 (n = 7), %1 . . . patUPD7, % . . . . Mean . SD . Median . . Mean . SD . Median . . GRB10-NotI 7p14 No 24.1 6.4 22.9 0.426 100.3 16.0 102.1 2.9 MEST-HpaII 7q32 No 60.5 6.7 59.7 0.842 96.4 10.5 98.7 8.5 H19-HpaII 11p15.5 Yes 45.7 6.1 44.7 0.806 47.2 4.7 46.0 40.6 KCNQ1OT1-HpaII-1 11p15.5 No 37.6 6.5 37.8 0.355 44.2 4.2 44.0 42.8 KCNQ1OT1-NotI 11p15.5 Yes 45.6 6.4 46.1 0.855 51.5 7.2 53.3 36.5 KCNQ1OT1-HpaII-2 11p15.5 No 44.3 4.7 44.2 0.841 53.7 5.5 53.1 40.3 1 For GRB10-Notl, n = 6; data for segmental matUPD7q31.33-qter are not included. 2 Testing the null hypothesis that the data are normally distributed in a Kolmogorov–Smirnov test. Open in new tab comparison of methylation percentage and southern blotting results We then studied 2 ICR restriction sites that had previously been studied with Southern blotting: HpaII site 25 in the H19 ICR (H19-HpaII) and the NotI site in the KCNQ1OT1 ICR (KCNQ1OT1-NotI). In the 40 controls, the mean (SD) methylation percentage was 46% (6%) for both H19-HpaII and KCNQ1OT1-NotI (Table 2 ). These methylation percentages are similar to those of previous observations and are consistent with the 50% methylation level expected at an ICR (11)(13)(27)(28). kcnq1ot1 icr To study KCNQ1OT1 methylation with HpaII in addition to NotI, we designed 2 additional assays, one at 1 kb upstream (KCNQ1OT1-HpaII-1) and another at 0.5 kb downstream (KCNQ1OT1-HpaII-2) of the KCNQ1OT1-NotI site. The observed methylation percentage was 38% (7%) for the KCNQ1OT1-HpaII-1 assay and 44% (5%) for the KCNQ1OT1-HpaII-2 assay. These results correspond to a significant difference in the means (P <0.05). The closest restriction site to KCNQ1OT1-NotI, the KCNQ1OT1-HpaII-2 site, showed the most similar methylation percentage. In our analysis, one control individual appeared to be hypermethylated at the KCNQ1OT1-HpaII-2 site but not at the adjacent sites. Sequencing of this individual revealed a heterozygous C-to-T transition at the restriction site (ss74800507, to appear in Build 128 of the dbSNP database; see http://www.ncbi.nlm.nih.gov/projects/SNP/). Although the parental origin of the transition could not be determined, our failure to digest DNA at this site suggests that it originated from the nonmethylated paternal allele. distribution of methylation percentage at imprinted loci We tested whether the observed data followed a normal distribution to ascertain how methylation percentage varied among individuals. A normal approximation seemed to describe the data well for all loci, and the null hypothesis of a normal distribution could not be rejected in a Kolmogorov–Smirnov test (Fig. 3 , Table 2 ), implying that the mean and SD are meaningful as cutoffs for defining abnormal methylation. To evaluate this supposition, we tested 20 patients with Silver–Russell syndrome for methylation at the H19 ICR (H19-HpaII), a locus that has previously been shown to be hypomethylated in a substantial proportion of such patients (13)(28). The results in Fig. 4 show that hypomethylation at this locus is readily detected in a subset of Silver–Russell syndrome patients at a methylation percentage cutoff of −2 SDs (<33.5%). The wide range of methylation percentage values in the hypomethylated patients (3%–20%) also demonstrates the different extents of mosaicism for the methylation aberration. Figure 3. Open in new tabDownload slide Distribution of methylation percentages at ICRs in control individuals. All observed methylation percentages (circles) are plotted against a theoretical normal distribution with means and SDs derived from the observed data (black line). In the case of a perfectly normal distribution, the observed methylation percentages would completely overlap with the black line. For all assays, the theoretical normal distribution seems to describe the data reasonably well. Figure 3. Open in new tabDownload slide Distribution of methylation percentages at ICRs in control individuals. All observed methylation percentages (circles) are plotted against a theoretical normal distribution with means and SDs derived from the observed data (black line). In the case of a perfectly normal distribution, the observed methylation percentages would completely overlap with the black line. For all assays, the theoretical normal distribution seems to describe the data reasonably well. Figure 4. Open in new tabDownload slide Methylation at the H19 ICR (H19-HpaII) in patients with Silver–Russell syndrome (SRS) and in control individuals. A dashed line drawn at −2 SD was derived from the control population (33.5% methylation), and all relevant SDs and the mean are indicated on the y-axis. Ten of the 20 SRS patients are clearly below this cutoff value. Figure 4. Open in new tabDownload slide Methylation at the H19 ICR (H19-HpaII) in patients with Silver–Russell syndrome (SRS) and in control individuals. A dashed line drawn at −2 SD was derived from the control population (33.5% methylation), and all relevant SDs and the mean are indicated on the y-axis. Ten of the 20 SRS patients are clearly below this cutoff value. Discussion We have shown the feasibility of digestion with methylation-sensitive restriction enzymes followed by qRT-PCR analysis for quantifying methylation at imprinted loci. This approach offers an advantage over currently used methods in that it requires only 2 laboratory steps, the digestion and the quantification, both of which can be performed in a microtiter plate format (see Fig. 1 in the online Data Supplement). qRT-PCR is a frequently used method with a wide linear range, and we demonstrated quantitative proportionality across the entire methylation range with linear regression analysis. We ensured digestion efficiency by using excess amounts of enzyme and digesting overnight and monitored efficiency further by adding nonmethylated λ DNA in the digestion and control reactions to measure dosage differences. To demonstrate how the extremes of the methylation range behaved for imprinted loci, we studied matUPD7 and patUPD7 individuals at imprinted loci, which were distinct from the controls and consistent with the expected methylation patterns. These results suggest that the method can also be used to quickly screen for matUPD7 in patients with Silver–Russell syndrome and be easily confirmed with microsatellite markers. Studies of site-specific loss or gain of methylation have generally focused on imprinting disorders. Such investigations have been based on detailed studies of ICRs, which have shown all sites in defined regions in genomic DNA extracted from blood samples to be stably imprinted and to be representative of the methylation status of the entire region (16). Imprinted regions are unique in this respect, whereas other phenomena, such as cancers, often require a more general investigative approach, such as large-scale or global methylation studies (5). Such studies are further complicated by the presence of cell population–specific methylation profiles (8); however, in instances of the influence of imprinted genes on some cancers, such as IGF2 [insulin-like growth factor 2 (somatomedin A)] and colon cancer, the study of a single restriction site could suffice as a diagnostic marker (29). To assess the utility of our method and compare it with conventional Southern blotting (the most commonly used method for studying loss or gain of methylation at imprinted regions), we investigated 2 extensively studied restriction sites in ICRs of the chromosome 11p15.5 imprinted clusters (the H19 ICR and the KCNQ1OT1 ICR). The methylation percentages for all controls, matUPD7, and patUPD7 were similar to those reported in previous studies (11)(13)(28). The Southern-blotting approach is limited to a single locus at a time and requires a relatively large amount of DNA, whereas the qRT-PCR approach, which permits the study of several loci with the same digest, is more efficient, both in terms of laboratory work and the amounts of DNA required. Some of the shortcomings of our method are the frequency of restriction sites and the GC-rich nature of the DNA of ICRs, which limit the number of possible PCRs that can be designed. The latter limitation is also problematic for methylation-specific PCR, which is discussed below. Furthermore, polymorphisms and mutations can unexpectedly influence the results for the studied site. This consideration means that if the proposed method is to be used to screen for the loss or gain of methylation, the preferred approach is to use a restriction site that has previously been well studied with Southern blotting, bisulfite sequencing, or another method, to ensure that the mean methylation percentage is relatively invariable among individuals and that the restriction site is in the critical region of the ICR. Several good methods that are based on bisulfite treatment of DNA are available for studying DNA methylation. A general drawback with bisulfite treatment is the risk of degrading the DNA during the conversion step, which includes NaOH treatment and incubation at high temperatures (5). Degradation is less of a problem with restriction enzyme digestion. A general advantage of bisulfite-based methods is that they are not limited to the evaluation of methylation at restriction sites only; all types of CpG methylation can be studied. Numerous methylation-specific PCR methods are in use (5). These methods require careful design of primers and probes and extensive optimization of efficiency, because such methods are based on different assays that are designed to distinguish between bisulfite-converted and nonconverted cytosines (5). With our approach, methylation percentage is measured with the same PCR assay, and thus the same efficiency. Combined bisulfite restriction analysis (COBRA) and MALDI-TOF mass spectrometry are 2 other bisulfite-based methods; these are more labor intensive than our approach (see Fig. 1 in the online Data Supplement); furthermore, the MALDI-TOF method requires specialized equipment (12)(14). We found methylation percentage to be normally distributed at imprinted loci in the control individuals. The quantitative nature of the mean methylation percentage has been reported previously (30). The normal distribution of the data and the magnitudes of the SDs show that methylation can vary between individuals, but only to a certain extent, as is suggested by the range of the measurements. Individuals with a degree of methylation that deviates greatly from the mean are likely to exhibit some phenotypic consequence, such as in the Beckwith–Wiedemann and Silver–Russell syndromes for 11p15.5 (4). Such a finding was exemplified in our screening for the H19 ICR in a small set of patients with Silver–Russell syndromes, in which we demonstrated that a substantial proportion of the patients showed methylation well below −2 SDs of the mean. Even if our method is highly quantitative, the mosaicism that we observed between individuals for the methylation aberration suggests that some individuals may be hard to define as normally methylated or as hypomethylated. In such cases, additional tests at proximal loci or the evaluation of DNA extracted from other tissues might be of aid. Because the mean methylation percentage obtained differs somewhat across techniques, it is advisable to use empirically defined cutoffs for abnormality. Interestingly, we observed differences in mean methylation percentage both within and between ICRs. In the KCNQ1OT1 ICR, all 3 studied sites showed slightly different mean methylation percentages. The previously studied NotI site had a mean methylation of 46% and was the most centrally located site in the ICR. The other sites, located about 1 kb upstream and about 0.4 kb downstream of the NotI site, showed lower degrees of methylation. Our observation is also supported by the recent report of 30% mean methylation at a distinct locus in the H19 ICR (30), whereas the mean methylation percentage for the well-studied HpaII site 25 (also studied in the present work) has been repeatedly reported to be about 50% (13)(28). We further observed that the restriction sites in the ICRs of GRB10 and MEST had mean methylation percentages that were different from the expected 50%, even if they clearly behaved as imprinted in UPDs of chromosome 7. These results emphasize the difficulty in defining imprinting, because the sites would be acceptable for UPD screening whereas the unexpected mean methylation in controls remains to be explained. In summary, we have shown that our method of digestion with methylation-sensitive restriction enzymes followed by qRT-PCR analysis is a simple, efficient, and quantitative method for studying methylation, with the potential for parallel investigations of several imprinted loci. The small amounts of DNA required allow extensive analysis of precious sample collections; consequently, this approach may become an attractive alternative to Southern blotting in the future. Furthermore, this approach can easily be established in most genetic laboratories with commonly available instruments. Finally, we have reported the range of methylation percentages for control individuals for 6 restriction sites in 4 ICRs, results that could easily be applied to large patient collections for studies of the loss or gain of methylation. Grant/funding Support: This study was supported by the Magn Bergvalls Stiftelse, the Swedish Research Council, the Päivikki and Sakari Sohlberg Foundation, the Sigrid Jusélius Foundation, and the Academy of Finland. C.M.L. is supported by the Throne-Holst Foundation. Financial Disclosures: None declared. Acknowledgment: We thank Riitta Lehtinen for excellent technical support. 1 " Nonstandard abbreviations: ICR, imprinting control region; UPD, uniparental disomy; matUPD7, maternal UPD for chromosome 7; patUPD7, paternal UPD for chromosome 7; MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight; qRT-PCR, quantitative real-time PCR; Ct, threshold cycle. 2 " Human genes: MEST, mesoderm specific transcript homolog (mouse); H19, H19, imprinted maternally expressed transcript; KCNQ1OT1, KCNQ1 overlapping transcript 1; GRB10, growth factor receptor-bound protein 10; IGF2, insulin-like growth factor 2 (somatomedin A). References 1 Costello JF, Plass C. Methylation matters. J Med Genet 2001 ; 38 : 285 -303. Crossref Search ADS PubMed 2 Morison IM, Ramsay JP, Spencer HG. A census of mammalian imprinting. Trends Genet 2005 ; 21 : 457 -465. Crossref Search ADS PubMed 3 Horsthemke B, Buiting K. Imprinting defects on human chromosome 15. Cytogenet Genome Res 2006 ; 113 : 292 -299. Crossref Search ADS PubMed 4 Smith AC, Choufani S, Ferreira JC, Weksberg R. Growth regulation, imprinted genes, and chromosome 11p15.5. Pediatr Res 2007 ; 61 : 43R -47R. Crossref Search ADS PubMed 5 Fraga MF, Esteller M. DNA methylation: a profile of methods and applications. Biotechniques 2002 ; 33 : 632 . Crossref Search ADS PubMed 6 Sandovici I, Leppert M, Hawk PR, Suarez A, Linares Y, Sapienza C. Familial aggregation of abnormal methylation of parental alleles at the IGF2/H19 and IGF2R differentially methylated regions. Hum Mol Genet 2003 ; 12 : 1569 -1578. Crossref Search ADS PubMed 7 Cooper WN, Luharia A, Evans GA, Raza H, Haire AC, Grundy R, et al. Molecular subtypes and phenotypic expression of Beckwith-Wiedemann syndrome. Eur J Hum Genet 2005 ; 13 : 1025 -1032. Crossref Search ADS PubMed 8 Cottrell SE, Distler J, Goodman NS, Mooney SH, Kluth A, Olek A, et al. A real-time PCR assay for DNA-methylation using methylation-specific blockers. Nucleic Acids Res 2004 ; 32 : e10 . Crossref Search ADS PubMed 9 Thomassin H, Kress C, Grange T. MethylQuant: a sensitive method for quantifying methylation of specific cytosines within the genome. Nucleic Acids Res 2004 ; 32 : e168 . Crossref Search ADS PubMed 10 Trinh BN, Long TI, Laird PW. DNA methylation analysis by MethyLight technology. Methods 2001 ; 25 : 456 -462. Crossref Search ADS PubMed 11 Bliek J, Maas SM, Ruijter JM, Hennekam RC, Alders M, Westerveld A, Mannens MM. Increased tumour risk for BWS patients correlates with aberrant H19 and not KCNQ1OT1 methylation: occurrence of KCNQ1OT1 hypomethylation in familial cases of BWS. Hum Mol Genet 2001 ; 10 : 467 -476. Crossref Search ADS PubMed 12 Ehrich M, Nelson MR, Stanssens P, Zabeau M, Liloglou T, Xinarianos G, et al. Quantitative high-throughput analysis of DNA methylation patterns by base-specific cleavage and mass spectrometry. Proc Natl Acad Sci U S A 2005 ; 102 : 15785 -15790. Crossref Search ADS PubMed 13 Gicquel C, Rossignol S, Cabrol S, Houang M, Steunou V, Barbu V, et al. Epimutation of the telomeric imprinting center region on chromosome 11p15 in Silver-Russell syndrome. Nat Genet 2005 ; 37 : 1003 -1007. Crossref Search ADS PubMed 14 Xiong Z, Laird PW. COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res 1997 ; 25 : 2532 -2534. Crossref Search ADS PubMed 15 Arnaud P, Monk D, Hitchins M, Gordon E, Dean W, Beechey CV, et al. Conserved methylation imprints in the human and mouse GRB10 genes with divergent allelic expression suggests differential reading of the same mark. Hum Mol Genet 2003 ; 12 : 1005 -1019. Crossref Search ADS PubMed 16 Frevel MA, Sowerby SJ, Petersen GB, Reeve AE. Methylation sequencing analysis refines the region of H19 epimutation in Wilms tumor. J Biol Chem 1999 ; 274 : 29331 -29340. Crossref Search ADS PubMed 17 Kerjean A, Dupont JM, Vasseur C, Le Tessier D, Cuisset L, Paldi A, et al. Establishment of the paternal methylation imprint of the human H19 and MEST/PEG1 genes during spermatogenesis. Hum Mol Genet 2000 ; 9 : 2183 -2187. Crossref Search ADS PubMed 18 Smilinich NJ, Day CD, Fitzpatrick GV, Caldwell GM, Lossie AC, Cooper PR, et al. A maternally methylated CpG island in KvLQT1 is associated with an antisense paternal transcript and loss of imprinting in Beckwith-Wiedemann syndrome. Proc Natl Acad Sci U S A 1999 ; 96 : 8064 -8069. Crossref Search ADS PubMed 19 Bruce S, Leinonen R, Lindgren CM, Kivinen K, Dahlman-Wright K, Lipsanen-Nyman M, et al. Global analysis of uniparental disomy using high density genotyping arrays. J Med Genet 2005 ; 42 : 847 -851. Crossref Search ADS PubMed 20 Hannula K, Kere J, Pirinen S, Holmberg C, Lipsanen-Nyman M. Do patients with maternal uniparental disomy for chromosome 7 have a distinct mild Silver-Russell phenotype?. J Med Genet 2001 ; 38 : 273 -278. Crossref Search ADS PubMed 21 Hannula K, Lipsanen-Nyman M, Kontiokari T, Kere J. A narrow segment of maternal uniparental disomy of chromosome 7q31-qter in Silver-Russell syndrome delimits a candidate gene region. Am J Hum Genet 2001 ; 68 : 247 -253. Crossref Search ADS PubMed 22 Hannula K, Lipsanen-Nyman M, Kristo P, Kaitila I, Simola KO, Lenko HL, et al. Genetic screening for maternal uniparental disomy of chromosome 7 in prenatal and postnatal growth retardation of unknown cause. Pediatrics 2002 ; 109 : 441 -448. Crossref Search ADS PubMed 23 Hoglund P, Holmberg C, de la Chapelle A, Kere J. Paternal isodisomy for chromosome 7 is compatible with normal growth and development in a patient with congenital chloride diarrhea. Am J Hum Genet 1994 ; 55 : 747 -752. PubMed 24 Rozen S, Skaletsky H. Primer3 on the WWW for general users and for biologist programmers. Methods Mol Biol 2000 ; 132 : 365 -386. PubMed 25 Bookout AL, Mangelsdorf DJ. Quantitative real-time PCR protocol for analysis of nuclear receptor signaling pathways. Nucl Recept Signal 2003 ; 1 : e012 . PubMed 26 . R Development Core Team. R: A language and environment for statistical computing 2005 R Foundation for Statistical Computing Vienna, Austria. . 27 Reik W, Brown KW, Slatter RE, Sartori P, Elliott M, Maher ER. Allelic methylation of H19 and IGF2 in the Beckwith-Wiedemann syndrome. Hum Mol Genet 1994 ; 3 : 1297 -1301. Crossref Search ADS PubMed 28 Schonherr N, Meyer E, Eggermann K, Ranke MB, Wollmann HA, Eggermann T. (Epi)mutations in 11p15 significantly contribute to Silver-Russell syndrome: but are they generally involved in growth retardation?. Eur J Med Genet 2006 ; 49 : 414 -418. Crossref Search ADS PubMed 29 Cui H, Cruz-Correa M, Giardiello FM, Hutcheon DF, Kafonek DR, Brandenburg S, et al. Loss of IGF2 imprinting: a potential marker of colorectal cancer risk. Science 2003 ; 299 : 1753 -1755. Crossref Search ADS PubMed 30 Heijmans BT, Kremer D, Tobi EW, Boomsma DI, Slagboom PE. Heritable rather than age-related environmental and stochastic factors dominate variation in DNA methylation of the human IGF2/H19 locus. Hum Mol Genet 2007 ; 16 : 547 -554. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)

Chim, Stephen S, C;Jin,, Shengnan;Lee, Tracy Y, H;Lun, Fiona M, F;Lee, Wing, S;Chan, Lisa Y, S;Jin,, Yongjie;Yang,, Ningning;Tong, Yu, K;Leung, Tak, Y;Lau, Tze, K;Ding,, Chunming;Chiu, Rossa W, K;Lo, Y M, Dennis

Fritz, Kristofer, S;McKean, Alastair J, S;Nelson, Jerald, C;Wilcox, R, Bruce

doi: 10.1373/clinchem.2007.094870pmid: 18171714

Abstract Background: Analog-based free testosterone test results, sex hormone binding globulin (SHBG) concentrations, and total testosterone concentrations are somehow related. This study used new experiments to clarify these relationships. Methods: An analog-based free testosterone immunoassay and a total testosterone immunoassay were applied to well-defined fractions of serum testosterone. First, they were applied to the 2 fractions (retentate and dialysate) of normal male serum obtained by equilibrium dialysis. Second, they were applied to covaried concentrations of SHBG and total testosterone. Third, they were applied to decreasing concentrations of SHBG and protein-bound testosterone, offset by increasing concentrations of protein-free testosterone, while total testosterone was held constant. Results: The analog-based free testosterone assay and the total testosterone assay detected and reported serum testosterone test results from serum retentate, whereas neither assay detected the free testosterone in serum dialysate. Test results reported by the analog-based free testosterone assay followed varied concentrations of SHBG and total testosterone. When total testosterone was held constant, however, analog-based free testosterone test results did not follow varied concentrations of serum proteins or of free testosterone. Conclusion: An analog-based free testosterone immunoassay reported free testosterone test results that were related to total testosterone concentrations under varied experimental conditions. This alleged free testosterone assay did not detect serum free testosterone (the test results it reported were nonspecific) and should not be used for this purpose. Serum free testosterone concentrations in vivo are thought to be under the control of the hypothalamic-pituitary-gonadal negative feedback regulatory system. Thus the amount of serum luteinizing hormone (LH) secreted is determined by the negative feedback effects of the circulatory free (unbound) testosterone concentrations. The circulatory concentrations of LH directly affect the amount of testosterone secreted by the testes. Serum total testosterone concentrations vary in proportion to varied free testosterone concentrations and in proportion to varied sex hormone binding globulin (SHBG) concentrations. Analog-based free testosterone immunoassays are the most widely used free testosterone methods. They are the only automated free testosterone methods. Multiple studies report that analog-based free testosterone test results are proportional to SHBG concentrations (1)(2)(3)(4)(5)(6)(7). It has also been asserted that they are proportional to total testosterone concentrations (1)(7)(8). This led us to question whether these analog-based free testosterone test results are more closely related to the concentrations of total testosterone or SHBG. Materials and Methods testosterone immunoassays The analog-based free testosterone immunoassay studied (Coat-A-Count; Siemens Medical Solutions Diagnostics) is a manual method that uses a radiolabeled (125I) testosterone analog (a conjugated form of testosterone), immobilized testosterone antibody, and a single incubation, following a 21-fold dilution of the serum sample with a single aqueous reagent. It is critical to note that the calibration range of this analog-based free testosterone assay is 1.9 to 173 pmol/L; generally accepted reference ranges for free testosterone are approximately 170 to 730 pmol/L for adult males (9). We measured total testosterone using a total testosterone immunoassay (Coat-A-Count; Siemens). This manual method also uses radiolabeled (125I) free testosterone, immobilized testosterone antibody, and a single incubation, following a 21-fold dilution of the sample with a single aqueous reagent. We applied each assay to the same serum-based testosterone solutions (Fig. 1 ). We also used a well-documented tracer dialysis free testosterone assay to verify free testosterone concentrations in these same-serum based testosterone solutions (see below in Free Testosterone by Tracer Dialysis). Each assay was performed according to manufacturer’s instructions. Gamma radiation was detected and quantified using a Gamma 4000 multiwell automated γ counter (Beckman-Coulter). Each testosterone measurement reported is a mean of triplicate determinations, and each experiment was repeated for confirmation. Figure 1. Open in new tabDownload slide Illustration of experiments. A tracer dialysis free testosterone (Te) assay, an analog-based free Te assay, and a total Te assay were applied to 3 sets of solutions. (A), The solutions in the first experiment were serum retentate and serum dialysate. (B), In the second experiment, serum retentate was progressively diluted with serum dialysate, diluting total Te from 100% to 20%. (C), In the third experiment, an aliquot of serum dialysate was enriched with Te to match the total Te concentration in the serum retentate. The serum retentate was then progressively diluted with the Te-enriched dialysate. SHBG·Te, SHBG-bound Te. Figure 1. Open in new tabDownload slide Illustration of experiments. A tracer dialysis free testosterone (Te) assay, an analog-based free Te assay, and a total Te assay were applied to 3 sets of solutions. (A), The solutions in the first experiment were serum retentate and serum dialysate. (B), In the second experiment, serum retentate was progressively diluted with serum dialysate, diluting total Te from 100% to 20%. (C), In the third experiment, an aliquot of serum dialysate was enriched with Te to match the total Te concentration in the serum retentate. The serum retentate was then progressively diluted with the Te-enriched dialysate. SHBG·Te, SHBG-bound Te. normal human serum We obtained serum from 16 healthy male volunteers, age 21–55 years. Serum collection was approved by the institutional review board, and serum samples were given anonymous identifiers. These sera were pooled. In the pool, the total testosterone of 21.6 nmol/L (Coat-A-Count; in-house), free testosterone of 232 pmol/L (by dialysis), total protein of 7.4 g/dL, serum albumin of 4.4 g/dL, and SHBG of 26 nmol/L were within their respective reference intervals (Quest Diagnostics). testosterone Testosterone and ethanol were obtained from Sigma Aldrich. Testosterone was dissolved at room temperature in 95% ethanol to produce a stock solution containing 6.9 mmol/L. This stock solution was diluted with a well-characterized dialysate buffer (10) to a concentration of 0.35 μmol/L testosterone. preparative equilibrium dialysis Dialysis devices and dialysate buffer were obtained from Antech Diagnostics. The dialysis device uses 200 μL sample retentate and 2400 μL dialysis buffer. The chemical composition of this buffer has been reported (10). Serum samples were dialyzed for 18 h at 37 °C in an Isotemp Incubator, model 630D (Fisher Scientific) (11)(12)(13). A moisture-saturated atmosphere was maintained during dialysis by enclosing dialysis devices in closed containers with open water reservoirs. The pH values of serum dialysate and retentate were controlled to 7.4 (±0.1) during equilibrium dialysis at 37 °C by HEPES acid in the dialysate buffer (14). At equilibrium, the final HEPES ion concentration was calculated to be 54 mmol/L. free testosterone by tracer dialysis We determined free testosterone concentrations by use of a previously reported tracer dialysis method (11)(15) and total testosterone by use of the Coat-A-Count total testosterone immunoassay described above. A 1-μL aliquot (approximately 13 pmol) of stock 3H-testosterone (PerkinElmer) was added to 1 mL of each sample, incubated for 1 h at 37 °C, and dialyzed as described above. Radioactivity was determined in retentates and dialysates using a liquid scintillation counter (LS7500, Beckman). We calculated the fraction of free testosterone at equilibrium as 3H-testosterone in dialysate/3H-testosterone in serum. We then calculated serum free testosterone by multiplying the fraction of free testosterone by total testosterone concentration. matrix effects It is well documented that sample and assay reagent matrices can play an important role in the performance of free hormone immunoassays. As previously mentioned, the analog-based free testosterone and total testosterone assays each use a 21-fold sample dilution before incubation and quantification. Both of these assays add 1 mL of a “proprietary” radioactive buffer solution to 50 μL sample. Multiple attempts to determine the constitution of this proprietary radioactive buffer solution have been thwarted by the assay’s manufacturer. Therefore, for the purpose of this study, it is assumed that both the 21-fold sample dilution as well as the components of this radioactive buffer will minimize any differences in sample matrices between a serum retentate and serum dialysate. testosterone adsorption Testosterone can be adsorbed onto solid surfaces from protein-free aqueous solutions. We tested the borosilicate glassware used in this study (Fisher Scientific) for testosterone adsorption: <2% of testosterone in serum dialysate was adsorbed in the absence of serum proteins. experimental strategies Fig. 1 shows the experiment design. In the first experiment, preparative equilibrium dialysis was applied to the normal adult male serum as described above (200 μL retentate vs 2400 μL dialysate). The analog-based free testosterone assay and the total testosterone assay were applied to the resulting serum retentate and serum dialysate. The tracer dialysis method was also applied to the normal adult male serum to verify free testosterone concentrations. A total testosterone assay is expected to detect the concentrations of total testosterone in serum retentate, whereas a free testosterone assay is expected to detect the concentrations of free testosterone in serum dialysate. In the second experiment, concentrations of serum proteins, including SHBG, protein-bound testosterone, and total testosterone, were covaried. This was accomplished by progressively diluting serum retentate with serum dialysate (obtained using the equilibrium dialysis method described above). Normal serum retentate was diluted with normal serum dialysate to obtain serum protein and total testosterone concentrations of 100%, 90%, 80%, 70%, 60%, 50%, 40%, 30%, and 20% of normal levels. The analog-based free testosterone assay and the total testosterone assay were applied to these solutions. The tracer dialysis method was also applied to these solutions to verify dialyzable free testosterone concentrations. A total testosterone assay is expected to track the decreasing total testosterone concentrations, whereas a free testosterone assay is not expected to track serum protein or total testosterone concentrations. In the third experiment, decreasing concentrations of SHBG and protein-bound testosterone were offset by increasing concentrations of free testosterone, while total testosterone was held constant. This was accomplished by measuring the total testosterone in serum retentate and then adding testosterone to an aliquot of serum dialysate until the total testosterone concentration was equal in both retentate and dialysate (21.6 nmol/L). The serum retentate was then progressively diluted with the protein-free, testosterone-enriched dialysate from 1- to 1000-fold. This procedure varied free testosterone concentrations in the opposite direction of SHBG and protein-bound testosterone concentrations. The analog-based free testosterone assay and the total testosterone assay were applied to these solutions. The tracer dialysis method was also applied to these solutions to verify dialyzable free testosterone concentrations. A total testosterone assay is expected to track the constant total testosterone concentrations, whereas a free testosterone assay is expected to track the increasing free testosterone concentrations. Results In the first experiment, the analog-based free testosterone assay and the total testosterone assay detected the nondialyzable fraction of serum testosterone, that is, the testosterone in the serum retentate. The total testosterone assay reported mean 21.6 (SD 1.0) nmol/L in the serum retentate. The analog-based free testosterone assay reported 56 (4.5) pmol/L in the serum retentate. Neither assay detected dialyzable concentrations of serum testosterone. The tracer dialysis free testosterone assay reported 347 pmol/L of testosterone in serum dialysate. In the second experiment, when concentrations of serum protein, including SHBG, protein-bound testosterone, and total testosterone, were progressively decreased, analog-based free testosterone test results ranged from 56 to 6.2 pmol/L (Fig. 2 ). Total testosterone test results ranged from 21.2 to 6.6 nmol/L. (Fig. 2 ). Tracer dialysis free testosterone test results ranged from 304 to 169 pmol/L (Fig. 2 ). Total testosterone test results and analog-based free testosterone test results correlated more closely (r2 0.97; P <0.001) than analog-based free testosterone test results and tracer dialysis free testosterone test results; (r2 0.90; P <0.001). Figure 2. Open in new tabDownload slide Analog-based free testosterone assay tracks serum protein and total testosterone concentrations. Tracer dialysis free testosterone test results (A), analog-based free testosterone test results (B), and total testosterone test results (C) when these assays were applied to progressive dilutions of serum retentate with serum dialysate. Te, testosterone. Figure 2. Open in new tabDownload slide Analog-based free testosterone assay tracks serum protein and total testosterone concentrations. Tracer dialysis free testosterone test results (A), analog-based free testosterone test results (B), and total testosterone test results (C) when these assays were applied to progressive dilutions of serum retentate with serum dialysate. Te, testosterone. In the third experiment, when decreasing concentrations of serum protein, including SHBG and protein-bound testosterone, were offset by increasing concentrations of protein-free testosterone while total testosterone was held constant, analog-based free testosterone test results [51.8 (1.3) pmol/L] paralleled total testosterone test results [22.4 (0.5) nmol/L] (Fig. 3 ). Tracer dialysis test results, or free testosterone concentration, increased from 376 to 798 pmol/L (Fig. 3 ). Figure 3. Open in new tabDownload slide Analog-based free testosterone assay tracks total testosterone concentrations. Tracer dialysis free testosterone test results (A), analog-based free testosterone test results (B), and total testosterone test results (C) when these assays were applied to progressive dilutions of serum retentate with testosterone-enriched serum dialysate. Total testosterone was held constant. Te, testosterone. Figure 3. Open in new tabDownload slide Analog-based free testosterone assay tracks total testosterone concentrations. Tracer dialysis free testosterone test results (A), analog-based free testosterone test results (B), and total testosterone test results (C) when these assays were applied to progressive dilutions of serum retentate with testosterone-enriched serum dialysate. Total testosterone was held constant. Te, testosterone. Discussion The Endocrine Society recently reported a review of the evidence that analog-based free testosterone immunoassays should be avoided because of problems with accuracy and sensitivity (16). The data obtained in the present study document nonspecificity associated with insensitivity and incorrect calibration in one of these assays (Figs. 2 and 3 ). Based on our conclusions, the experimental strategies applied to this analog-based free testosterone assay should be applied to other, as yet untested, commercially available analog-based free testosterone assays. The design and results of this study differ from previous reports (1)(2)(3)(4)(5)(6)(7)(8). Previous studies have contributed to our understanding by applying these assays to patient samples. The present study used well-defined solutions prepared with the aim of elucidating to which form of serum testosterone an analog-based free testosterone assay would respond. Also, one of the more recent studies (1) found that SHBG was an important determinant in this analog-based free testosterone assay and that this assay appeared to measure a constant fraction of the total testosterone in adult male plasma, leading the authors to conclude that this analog-based free testosterone assay provides essentially the same information as a total testosterone assay when applied to healthy adult males. These previously published assertions are supported by the results of our second and third experiments (Figs. 2 and 3 ). There is a hypothetical explanation that might account for the nonspecificity observed in this assay. Nonspecificity would occur if serum protein testosterone complexes bind to testosterone antibody, leading to a 3-way competition between free testosterone, testosterone complexes, and testosterone conjugates (analogs) for binding to the same antibody. This competition would explain the lack of specificity and would confound calibration. The data from the second and third experiments in this study are consistent with this hypothesis (Figs. 2 and 3 ). Similar characteristics have recently been reported in an analog-based free thyroxine immunoassay (17). These data are now sufficient to warrant further testing of this hypothesis. Until the characteristics of assays such as this have been fully accounted for, they should not be confused with free hormone assays that are sensitive, specific, and gravimetrically calibrated with an analytical balance to a scientifically acceptable mass standard. There is no traceability when specificity is absent and there is no specificity when covariables have not been accounted for. Since the analog-based assay in this study does not detect or quantify free testosterone, it should not be used as a free testosterone assay. Grant/funding Support: This work was supported by the Loma Linda University School of Medicine and the Mortensen Chair. No extramural funds were used for this study. Financial Disclosures: J.C.N. is currently a consultant to Antech Diagnostics. He was formerly Senior Medical Director of Quest Diagnostics Nichols Institute, San Juan Capistrano, and has no current affiliation with Quest Diagnostics. Acknowledgments: We wish to acknowledge the expert technical assistance provided by Rene Weiss. References 1 Winters SJ, Kelley DE, Goodpaster B. The analog free testosterone assay: are the results in men clinically useful?. Clin Chem 1998 ; 44 : 2178 -2182. Crossref Search ADS PubMed 2 Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab 1999 ; 84 : 3666 -3672. Crossref Search ADS PubMed 3 Tibblin G, Adlerberth A, Lindstedt G, Bjorntorp P. The pituitary-gonadal axis and health in elderly men: a study of men born in 1913. Diabetes 1996 ; 45 : 1605 -1609. Crossref Search ADS PubMed 4 Jowett T, Chu F, Ekins R. Validity of current analog-based free hormone immunoassays. Steroids 1988 ; 52 : 365 -366. Crossref Search ADS PubMed 5 Haffner SM, Shaten J, Stern MP, Smith GD, Kuller L. Low levels of sex hormone-binding globulin and testosterone predict the development of non-insulin-dependent diabetes mellitus in men. Am J Epidemiol 1996 ; 143 : 889 -897. Crossref Search ADS PubMed 6 Slaats EH, Kennedy JC, Kruijswijk H. Interference of sex-hormone binding globulin in the “Coat-A-Count” testosterone no-extraction radioimmunoassay. Clin Chem 1987 ; 33 : 300 -302. Crossref Search ADS PubMed 7 Phillips GB, Jing T-Y, Resnick LM, Barbagallo M, Laragh JH, Sealey JE. Sex hormones and hemostatic risk factors for coronary heart disease in men with hypertension. J Hypertens 1993 ; 11 : 699 -702. Crossref Search ADS PubMed 8 Van Uytfanghe K, Stockl D, Kaufman JM, Fiers T, De Leenheer A, Thienpont LM. Validation of 5 routine assays for serum free testosterone with a candidate reference measurement procedure based on ultrafiltration and isotope dilution-gas chromatography-mass spectrometry. Clin Biochem 2005 ; 38 : 253 -261. Crossref Search ADS PubMed 9 Greenspan FS Gardner DG eds. Basic and Clinical Endocrinology 7th ed. 2004 McGraw-Hill New York. . 10 Nelson JC, Tomei RT. Direct determination of free thyroxin in undiluted serum by equilibrium dialysis/radioimmunoassay. Clin Chem 1988 ; 34 : 1737 -1744. Crossref Search ADS PubMed 11 Miller KK, Rosner W, Lee H, Hier J, Sesmilo G, Schoenfeld D, Neubauer G, Klibanski A. Measurement of free testosterone in normal women and women with androgen deficiency: comparison of methods. J Clin Endocrinol Metab 2004 ; 89 : 525 -533. Crossref Search ADS PubMed 12 Sinha-Hikim I, Arver S, Beall G, Shen R, Guerrero M, Sattler F, Shikuma C, Nelson JC, Landgren BM, Mazer NA, Bhasin S. The use of a sensitive equilibrium dialysis method for the measurement of free testosterone levels in healthy, cycling women and in human immunodeficiency virus-infected women. J Clin Endocrinol Metab 1998 ; 83 : 1312 -1318. PubMed 13 Umstot ES, Baxter JE, Andersen RN. A theoretically sound and practicable equilibrium dialysis method for measuring percentage of free testosterone. J Steroid Biochem 1985 ; 22 : 639 -648. Crossref Search ADS PubMed 14 Wilcox RB, Nelson JC. Time course of pH regulation in free thyroxin determinations in serum. Clin Chem 1991 ; 37 : 1298 -1300. Crossref Search ADS PubMed 15 Bhasin S, Swerdloff RS, Steiner B, Peterson MA, Meridores T, Galmirini M, Pandian MR, Goldberg R, Berman N. A biodegradable testosterone microcapsule formulation provides uniform eugonadal levels of testosterone for 10–11 weeks in hypogonadal men. J Clin Endocrinol Metab 1992 ; 74 : 75 -83. PubMed 16 Rosner W, Auchus RJ, Azziz R, Sluss PM, Raff H. Utility, limitations, and pitfalls in measuring testosterone: an Endocrine Society position statement. J Clin Endocrinol Metab 2007 ; 92 : 405 -413. Crossref Search ADS PubMed 17 Fritz KS, Wilcox RB, Nelson JC. A direct free T4 immunoassay with the characteristics of a total T4 immunoassay. Clin Chem 2007 ; 53 : 911 -915. Crossref Search ADS PubMed © 2008 The American Association for Clinical Chemistry This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/open_access/funder_policies/chorus/standard_publication_model)