TY - JOUR
AU - Swietach,, Pawel
AB - Abstract Background Mild hemolysis occurs physiologically in neonates, but more severe forms can lead to life-threatening anemia. Newborns in developing regions are particularly at-risk due to the higher incidence of triggers (protozoan infections, sepsis, certain genetic traits). In advanced healthcare facilities, hemolysis is monitored indirectly using resource-intensive methods that probe downstream ramifications. These approaches could potentially delay critical decisions in early-life care, and are not suitable for point-of-care testing. Rapid and cost-effective testing could be based on detecting red blood cell (RBC)-specific proteins, such as carbonic anhydrase I (CAI), in accessible fluids (e.g., urine). Methods Urine was collected from 26 full-term male neonates and analyzed for CAI using immunoassays (ELISA, western blot) and proteomics (mass spectrometry). The cohort included a range of hemolytic states, including admissions with infection, ABO incompatibility, and receiving phototherapy. Data were paired with hemoglobin, serum bilirubin (SBR), and C-reactive protein (CRP) measurements. Results Urine from a control cohort (CRP < 20 mg/L, SBR < 125µmol/L) had no detectable CAI, in line with results from healthy adults. CAI excretion was elevated in neonates with raised SBR (>125 µmol/L), including those qualifying for phototherapy. Newborns with low SBR (<125 µmol/L) but elevated CRP (>20 mg/L) produced urine with strong CAI immunoreactivity. Proteomics showed that CAI was the most abundant RBC-specific protein in CAI-immunopositive samples, and did not associate with other RBC-derived peptides, indicating an intravascular hemolytic source followed by CAI-selective excretion. Conclusions CAI is a direct biomarker of intravascular hemolysis that can be measured routinely in urine using non-invasive methods under minimal-laboratory conditions. IMPACT STATEMENT Hemolysis in the newborn can be life threatening, yet direct assays for rapid diagnosis are not available for regular monitoring. Instead, current clinical management relies on resource-intensive measurements of downstream ramifications, potentially delaying critical decisions in early-life care. Using a cohort of newborns manifesting various hemolytic states, we show that intravascular hemolysis can be detected by measuring CAI excretion in a small sample of urine using cost-effective immunoreactivity techniques. Since this biomarker reports cell rupture, it provides a direct readout of hemolysis. The method can improve resource allocation, identify ‘at-risk’ patients earlier, and may be implemented under minimal-laboratory conditions. Introduction Hemolysis is the rupturing of red blood cells (RBCs) that results in the release of their cytoplasmic contents. Various degrees of intravascular hemolysis take place in the first days of life. A mild, physiological form of hemolysis facilitates the process of replacing fetal hemoglobin with its adult form (1,). More severe forms, manifesting as jaundice, are associated with medical conditions, such as sepsis (e.g., Gram-positive bacteria), protozoan infection (e.g., malaria), alloimmunity (ABO and Rhesus incompatibility), genetic traits (e.g., sickle cell, glucose-6-phosphate dehydrogenase deficiency [G6PDD]), birth trauma, and prematurity. Severe neonatal jaundice represents a significant burden of morbidity worldwide, and is particularly problematic in developing regions, where hemolytic triggers are more prevalent (2–4,). Further, the incidence of certain hemolytic conditions (e.g., sickle cell trait, Rhesus disease, G6PDD) is related to ethnicity, resulting in at-risk groups (5–8). Since hemolysis can lead to life-threatening anemia, successful management must rely on early and accurate diagnosis. The timing, duration, and intensity of a hemolytic crisis cannot be readily predicted, and therefore at-risk patients should be monitored regularly, using assays that provide a direct, rapid, and linear readout of RBC rupture. Certain causes of hemolysis can be diagnosed with specific tests, such as Coombs test for autoimmune hemolytic anemia (9), but these are resource intensive and are unable to quantify the degree of ongoing hemolysis. The standard clinical tests for hemolysis involve taking blood for cell count, morphology, or biochemical assays (e.g., for unconjugated bilirubin, haptoglobin). Serum haptoglobin is commonly used as an inverse biochemical assay of hemolysis (i.e., levels fall during hemolysis), but its interpretation can be ambiguous if the baseline of circulating haptoglobin is not established with adequate precision. A linear and nonsaturating positive marker, such as serum bilirubin (SBR), overcomes this issue. However, this and similar diagnostic approaches do not provide a direct readout of hemolysis, in the sense that they do not probe for RBC-specific molecules released upon rupture. Instead, these interrogate the downstream cascades, which inadvertently introduces a delay in diagnosis. Moreover, babies diagnosed with jaundice on the basis of bilirubin measurements are often assumed to be hemolyzing, although a small proportion may have a severe nonhemolytic condition, such as Crigler–Najjar syndrome that requires urgent medical attention. Identification of such cases would be expedited by a direct assay for hemolysis. While blood-based tests are performed routinely in hospitals, they require trained personnel, laboratory equipment, and high standards of hygiene, and thus may not be feasible for ambulatory care, in developing countries, and outside advanced healthcare systems. These onerous requirements also preclude high-frequency, point-of-care testing in at-risk groups to accurately track hemolysis. Given that neonatal hemolysis is a global health concern, there is a medical and economic incentive for developing RBC-related biomarker tests that are fast, reliable, quantitative, and safe, yet technically simple and low cost. Proteins released from ruptured RBCs are good candidate-molecules for hemolysis biomarkers, provided they are (i) stable, (ii) distinguishable from potential non-RBC sources, and (iii) released in sufficient amount to produce a detectable signal. A desirable property of such a biomarker would be detection in readily accessible body fluids, such as urine, rather than in blood, for repeated testing. Carbonic anhydrase I (CAI) is a soluble protein that is highly abundant in the cytoplasm of RBCs (10, 11,) [∼4 g/L of adult blood, ∼1 g/L of neonatal blood (12, 13,) and rising over the first year of life (14,)], but not highly expressed in other tissues, where levels are typically orders of magnitude lower than inside red cells (15, 16,). The size of the CAI pool in RBCs would be sufficient to raise plasma CAI upon hemolysis, even after accounting for volume-dilution. Normally, only trace quantities of CAI are found in plasma (∼0.001 g/L) (13, 17,), thus providing very low background. Given its small size (29 kDa, Stokes radius 25 Ang) (12, 15, 18,), CAI is expected to cross the glomerular barrier and appear in urine, as has been suggested by a small-scale study of adults with renal disease (19,). However, the relationship between hemolysis and urinary CAI excretion is unknown. After glomerular filtration, CAI protein may, for example, be reclaimed by megalin/cubilin receptors in the kidney proximal tubule fluid (20,). One previous report has detected only trace amounts of CAI in the urine of apparently healthy adults (<5 µg/day) (17), but not explored its relationship with hemolysis. It has not yet been determined if CAI is detectable in the urine of hemolyzing neonates, and whether this information carries any diagnostic meaning. Here, we assayed urinary CAI excretion in a cohort of neonates that included nonhemolyzing controls and various conditions expected to produce a spectrum of hemolytic states. We show that the urinary CAI signal ranges from undetectable to strong, and relates to the underlying hemolytic condition. By adapting this method for lateral-flow immunochromatography ‘strip-tests,’ it would be possible to obtain rapid readouts of hemolysis for point-of-care testing in the newborn. Methods Patients Patient recruitment took place at Evelina London Children’s Hospital over a period of 12 months (starting August 2015), where admissions include prematurity, respiratory distress, sepsis, as well as cardiac, neurological, and surgical conditions (∼800 cases per year). For this study, 26 full-term babies were recruited; none were excluded. The cohort included newborns with jaundice that required treatment, mildly jaundiced, and those with no hemolytic condition (controls). In order to obtain a sufficient volume of urine for multiple types of analyses, male babies were recruited. Urine samples were obtained in the first 9 days of life and then anonymized. For this study, urine bags were used to eliminate possible fecal contamination of samples. Clinical information (date of birth, gestation, phototherapy, treatment with antibiotics) and standard clinical measurements of (i) blood hemoglobin, (ii) serum bilirubin, (iii) plasma [K+], (iv) plasma [creatinine], (v) C-reactive protein were taken but not made available to the researcher collecting or measuring from urine samples. Urine samples were frozen for storage (−20 °C) in 1.5 mL labelled tubes pretreated with 100 µL of protease inhibitor solution (Complete, Mini Protease Inhibitor Cocktail Tablets, Roche Diagnostics; 1 tablet/10 mL of final volume), and thawed when ready for measurements. To test the methodology with negative samples, urine was taken from a cohort of nonhemolyzing healthy adult volunteers (University of Oxford Central University Research Ethics Committee procedure 24). See supplement for power calculations. Measurements on Urine Samples Urine samples were centrifuged to remove solid deposits and aliquoted to provide at least 200 µL for immunotechniques (ELISA, immunoblots) as well as measurements of protein and creatinine. See Supplement for details. In the case of neonatal urine, 200–500 µL was reserved for mass spectrometry, which included a digestion step and measurement of total peptide. Any residual neonatal urine was destroyed before the end of the study period. Any residual urine remaining from adults was destroyed within 24 h of collection. ELISA Each urine sample was measured in triplicate. 50 μl aliquots of urine were added to each well of a 96-well high-binding microplate (Greiner Bio-One). As a positive control, lysates of RBCs were dissolved in adult urine. The plate was then incubated overnight at 37 °C and air-dried, washed with 0.05% phosphate-buffered saline (PBS) with Tween detergent (PBST), blocked with 10% fetal calf serum (FCS) in PBS for 2 h, and incubated with primary goat anti-human CAI polyclonal antibody (R&D Biosystems, Novus) diluted 1:250 in blocking solution (1 h at room temperature). Next, the plate was washed four times with PBST, and incubated with HRP-conjugated rabbit anti-goat IgG (H + L) secondary antibody, diluted 1:8000 in blocking buffer (1 h at room temperature), and then washed again. The signal was developed using OPD solution (Sigma/Merck) and absorbance was measured at 490 nm (Biotek ELx800 spectrophotometer or Cytation 5, Biotek). Background absorbance was measured in wells that contained PBS. Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) See supplement for details of protocol (21,) and analysis (MaxQuant v1.5.8.3) (22). Data were searched against a UniProt human database (v20170202). Statistics Significance of ELISA signal relative to background signal was tested by two-tail t-test. The relationship between urinary CAI status and clinical information was tested by the Freeman-Halton extension of Fisher’s exact probability test and Chi-squared test of a 2-by-3 contingency table. Significance of differences in the abundance of urinary proteins was tested by one-way ANOVA. Study Approval This prospective study was given ethical approval by the UK National Health Service Research Ethics Committee NRES (REC reference: 14/NS/0042, IRAS project ID: 170697). Written consent was obtained from the parent/legal guardian of the recruited newborn, or from adult donors themselves. Results RBC-Derived CAI Produces a Strong Signal for Detection by Immunoreactivity Assays The antibody against CAI was first tested for cross-reactivity with CAII, a related isoform, purified from bovine blood. Strong immunoreactivity was detected in lysates prepared from human RBCs but not with purified CAII, confirming isoform-specificity (Fig. 1, A and B). Human RBC-derived CAI could be detected in dilutions up to 105-fold in urine using ELISA (Fig. 1C). For this experiment, background absorbance was determined in wells containing PBS in place of urine. Urine samples from adult volunteers (N = 8; mean age: 34 years) showed no detectable CAI by ELISA (Fig. 1D), as expected from a minimal degree of hemolysis and confirming previous studies of only trace CAI excretion. This negative result was confirmed by Western blotting, a more sensitive method (Fig. 1E; note, prolonged film exposure revealed only trace CAI immunoreactivity in some samples; Fig. 1F). Fig. 1 Open in new tabDownload slide Testing CAI antibodies and immunoreactivity in adult urine. a) Specificity of the CAI antibody was tested by western blot on human RBC lysates diluted in urine (a source of CAI and CAII), and CAII purified from bovine blood. Dilutions were performed in urine obtained from a nonhemolyzing adult volunteer. b) Comparing antigen recognition by anti-CAI and anti-CAII antibodies by means of dot blot (5 ng bovine CAII per 2 µL/dot and 2 µL of 47-fold diluted human RBC lysate/dot). c) RBC lysates, prepared from an adult volunteer, were diluted serially in urine. ELISA using the CAI antibody was performed and quantified in terms of absorbance. d) CAI immunoreactivity measured by ELISA in urine samples from 8 adult male volunteers. Line shows background signal, determined in wells containing PBS only. e) Western blot of adult urine samples for short (30 s) and f) longer (120 s) film exposure time. Fig. 1 Open in new tabDownload slide Testing CAI antibodies and immunoreactivity in adult urine. a) Specificity of the CAI antibody was tested by western blot on human RBC lysates diluted in urine (a source of CAI and CAII), and CAII purified from bovine blood. Dilutions were performed in urine obtained from a nonhemolyzing adult volunteer. b) Comparing antigen recognition by anti-CAI and anti-CAII antibodies by means of dot blot (5 ng bovine CAII per 2 µL/dot and 2 µL of 47-fold diluted human RBC lysate/dot). c) RBC lysates, prepared from an adult volunteer, were diluted serially in urine. ELISA using the CAI antibody was performed and quantified in terms of absorbance. d) CAI immunoreactivity measured by ELISA in urine samples from 8 adult male volunteers. Line shows background signal, determined in wells containing PBS only. e) Western blot of adult urine samples for short (30 s) and f) longer (120 s) film exposure time. The linearity of CAI ELISA was demonstrated using various dilutions of human recombinant CAI (hrCAI) in adult urine (Supplemental Fig. 1). This calibration curve was used in subsequent analyses to calculate the amount of CAI (in ng/well). To test the reporting accuracy of the assay, a spike/recovery study was performed on urine samples from 4 adults (Supplemental Fig. 2). Samples were split into two groups, one of which was enriched in filtered (<100 kDa) hemolysate, prepared from human venous blood. Spiking was performed with known amounts of hrCAI. Absorbance measurements, performed before and after spiking (Supplemental Fig. 2A), were converted to an amount of CAI, which was then expressed relative to the known amount of hrCAI added (Supplemental Fig. 2B). The recovery for all spikes tested was close to 100%, confirming the assay’s accuracy. To determine the assay’s precision, repeated measurements on hrCAI-spiked adult urine samples were performed over a period of 8 days (Supplemental Fig. 3). The ELISA signal remained consistent, indicating a good level of precision. Urine Samples from Newborn Manifest a Range of CAI Immunoreactivity Urine was obtained from 26 full-term babies (gestation time: 37.0–42.4 weeks; mean 39.4 weeks) (Fig. 2A), collected between the 2nd and 9th day of life (mode 3 days) (Fig. 2B). Blood hemoglobin (Hb) ranged from 107 to 245 g/L (mean 174 g/L, SD 30 g/L); 22 babies had Hb in the normal range (134–199 g/L), two had Hb < 134 g/L, and two had Hb > 199 g/L (Fig. 2C). Plasma [K+] was 3.6–6.6 mM (mean 5.1 mM, SD 0.84 mM), with two samples classified as hyperkalemic (>6.0 mM). Mean plasma creatinine, urine creatinine, and the urine/plasma creatinine ratio were 54 µmol/L (SD 22 µmol/L), 25 µmol/L (SD 18 µmol/L), and 0.47 (SD 0.27), respectively. Renal function was normal (urinary protein/creatinine < 1.5 g/g) in all, but two newborns (23) (Fig. 2D). Fig. 2 Open in new tabDownload slide Detecting CAI immunoreactivity in urine samples from newborns. a) Gestation (mean indicated by continuous line). b) Age at urine collection (mean indicated by continuous line). c) Blood hemoglobin (mean indicated by continuous line); normal range in neonates is 133–199 g/L. (D) Ratio of urine protein to urine creatinine. Dashed line indicates 95th percentile expected for this age-group. e) Result of ELISA performed against CAI on 50 µL sample of undiluted urine. Dashed line shows background signal measured in wells with no urine (PBS-containing). Samples shaded grey produced detectable CAI immunoreactivity. Absorbance was converted to amount of CAI per well, based on a calibration curve shown in Supplemental Fig. 1. f) Western blot performed on 30 µL sample of undiluted urine. Note, urine samples 9, 13, and 16 had insufficient volume for western blotting. Fig. 2 Open in new tabDownload slide Detecting CAI immunoreactivity in urine samples from newborns. a) Gestation (mean indicated by continuous line). b) Age at urine collection (mean indicated by continuous line). c) Blood hemoglobin (mean indicated by continuous line); normal range in neonates is 133–199 g/L. (D) Ratio of urine protein to urine creatinine. Dashed line indicates 95th percentile expected for this age-group. e) Result of ELISA performed against CAI on 50 µL sample of undiluted urine. Dashed line shows background signal measured in wells with no urine (PBS-containing). Samples shaded grey produced detectable CAI immunoreactivity. Absorbance was converted to amount of CAI per well, based on a calibration curve shown in Supplemental Fig. 1. f) Western blot performed on 30 µL sample of undiluted urine. Note, urine samples 9, 13, and 16 had insufficient volume for western blotting. CAI immunoreactivity in neonatal urine was tested by ELISA. Background signal was determined in wells that contained PBS. Unlike in the case of healthy adults, where CAI levels in urine were consistently negligible (Fig. 1D), neonates produced urine with a wide range of CAI immunoreactivity determined by ELISA (Fig. 2E). This absorbance readout was converted into an amount of CAI using the calibration curve shown in Supplemental Fig. 1. Signals ranged from nil (i.e., at background level) in 9 samples, to a spectrum of positive signals in the remaining 17 samples, the strongest of which was equivalent to 15 ng CAI per well (50 µL). CAI status was confirmed by western blot (Fig. 2F). These findings provide the first evidence that significant levels of CAI-specific immunoreactivity can be measured in neonatal urine. The use of urine bags on male babies in this trial eliminates the possibility of fecal contamination. However, such contamination cannot be excluded if more crude methods are attempted to collect urine. Previous studies have demonstrated CAI expression in colorectal epithelium (16, 24), and this could potentially influence CAI readouts in urine contaminated with feces. To test if the presence of colorectal epithelium could meaningfully compromise the robustness of CAI measurements, ELISA measurements were performed on adult urine samples containing either filtered hemolysate (750×-diluted) or lysates prepared from human colorectal cell lines (Caco2, DLD1, HCT116 and HT29) (Supplemental Fig. 4A). Total protein concentration was measured to derive the CAI content per unit of total protein. A CAI signal equivalent to 1 ng required only ∼0.1 µg of hemolysate, but as much as 500 µg of colorectal epithelium lysate (Supplemental Fig. 4B). This difference is sufficiently large to indicate that fecal contamination is unlikely to influence urinary CAI levels. To confirm this experimentally, ELISA measurements were performed on adult urine dosed with human feces. Contaminated urine was produced by adding a sample of feces to urine (1:10 v/v), followed by vigorous mixing and two rounds of centrifugation (9000 rpm, 10 min, 4 °C). As a positive control, urine was spiked with hrCAI. As shown in Supplemental Fig. 5, fecal contamination as high as 10% did not affect background CAI signal, indicating that feces are highly unlikely to produce false-positive CAI immunoreactivity in urine samples, even if these were obtained in a crude way. Thus, any urinary CAI excretion is attributable to hemolysis. Urinary CAI Positivity Is Associated with Increased Levels of Bilirubin and C-Reactive Protein To investigate the relationship between urinary CAI excretion and clinical presentation, data were stratified according to phototherapy status. Twelve neonates receiving phototherapy had raised serum bilirubin (SBR), with a mean of 267 µmol/L (SD 44 µmol/L), on the day of urine collection. Among those not qualifying for phototherapy, 3 neonates had elevated C-reactive protein (CRP > 20 mg/L) (25,) and received antibiotics, indicating an underlying infection. The remaining neonates (11) manifested a range of SBR levels, from low to elevated, and were further subdivided into two groups, taking an SBR cut-off of 125 µmol/L. The cohort with low SBR (<125 µmol/L) was the ‘control’ group, and that with raised SBR (>125 µmol/L) but not meeting the criteria for phototherapy was referred to as the ‘subclinical’ group (Fig. 3A). Fig. 3 Open in new tabDownload slide Relating urinary CAI to clinical information. a) Stratification of neonates based on phototherapy status, C-reactive peptide (CRP) levels, and serum bilirubin (SBR). b) Scatter plot of urinary CAI signal versus SBR. Absorbance was converted to amount of CAI per well, based on a calibration curve shown in Supplemental Fig. 1. c) CAI signal (ELISA) categorized by neonate group. Statistical test: comparison to background signal (dashed line) by two-sided t-test of log-transformed data. d) CAI signal (ELISA) versus age (in days) at which urine sample was obtained. Fig. 3 Open in new tabDownload slide Relating urinary CAI to clinical information. a) Stratification of neonates based on phototherapy status, C-reactive peptide (CRP) levels, and serum bilirubin (SBR). b) Scatter plot of urinary CAI signal versus SBR. Absorbance was converted to amount of CAI per well, based on a calibration curve shown in Supplemental Fig. 1. c) CAI signal (ELISA) categorized by neonate group. Statistical test: comparison to background signal (dashed line) by two-sided t-test of log-transformed data. d) CAI signal (ELISA) versus age (in days) at which urine sample was obtained. The relationship between SBR and urinary CAI signal (ELISA) delineates these four groups (Fig. 3B). Absorbance was significantly above background (cut-off 0.05075 as measured in PBS) in the ‘infected,’ ‘phototherapy,’ and ‘subclinical’ groups (Fig. 3C). The ‘infected’ group had the highest CAI signals (3/3 CAI-immunopositive), but presented no evidence for renal failure (Fig. 2D), which argues for infection-related hemolysis. Strikingly, 10 of the 12 neonates receiving phototherapy, including 2 with a positive direct Coombs test, produced CAI-immunopositive urine. In contrast, none of the neonates in the control group excreted CAI in their urine. Differences in CAI immunoreactivity between the four groups were not explained by age at which urine was sampled (Fig. 3D;P = 0.62, one-way ANOVA). An association between urinary CAI-immunopositivity (a direct product of hemolysis) and elevated SBR (a downstream readout of hemolysis) was tested using a 2 × 3 contingency table for ‘control’, ‘subclinical’, and ‘phototherapy’ groups, representing increasing levels of SBR (mean±SD: 75 ± 36 µmol/L, 173 ± 35 µmol/L, and 267 ± 44 µmol/L). Fisher’s exact probability test and Chi-squared test indicated a significant association between CAI status and SBR (P = 0.005; Table 1). Table 1 2-by-3 contingency table for the Fisher Exact Probability Test. P = 0.0052. Chi-squared test (χ2 = 10.4, degrees of freedom = 2), P = 0.005 Urine CAI status (by ELISA): . Not receiving phototherapy . Receiving phototherapy . SBR<125 µmol/L . SBR>125 µmol/L . CAI-immunonegative 5 2 2 CAI-immunopositive 0 4 10 Urine CAI status (by ELISA): . Not receiving phototherapy . Receiving phototherapy . SBR<125 µmol/L . SBR>125 µmol/L . CAI-immunonegative 5 2 2 CAI-immunopositive 0 4 10 Open in new tab Table 1 2-by-3 contingency table for the Fisher Exact Probability Test. P = 0.0052. Chi-squared test (χ2 = 10.4, degrees of freedom = 2), P = 0.005 Urine CAI status (by ELISA): . Not receiving phototherapy . Receiving phototherapy . SBR<125 µmol/L . SBR>125 µmol/L . CAI-immunonegative 5 2 2 CAI-immunopositive 0 4 10 Urine CAI status (by ELISA): . Not receiving phototherapy . Receiving phototherapy . SBR<125 µmol/L . SBR>125 µmol/L . CAI-immunonegative 5 2 2 CAI-immunopositive 0 4 10 Open in new tab Protein Signature of Urine Indicates an Intravascular Hemolytic Source of Urinary CAI The detection of a CAI signal in urine implicates a large source of CAI, namely a population of ruptured RBCs, as no other tissue would be capable of releasing a sufficient amount of CAI. However, the results thus far cannot determine whether RBC rupture had taken place inside blood vessels or along the urinary tract. An extravascular hemolytic event would result in the urinary excretion of RBC proteins that do not normally cross the glomerular filter (e.g., membrane-bound or large soluble proteins), or those proteins that are normally reclaimed by the nephron. A related question is whether CAI is the most abundant RBC-specific protein detectable in urine. These two questions were addressed by proteomic analysis of urine samples. Mass spectrometry analysis of neonatal urine identified over 1000 proteins in one or more of the 25 samples (Supplemental Table 1; Supplemental Fig. 6A; note, there was insufficient volume in sample 27 to perform this analysis). To identify RBC-associated proteins, the urinary proteome was compared against a database of RBC proteins (Supplemental Fig. 6B) (10,). Among the 64 most abundant proteins in RBCs, collectively representing a third of all RBC peptides (10,), twenty-one were detected in at least one urine sample. Of these, CAI had the highest mean signal and showed greatest variation, i.e., potential dynamic range (Fig. 4A). Importantly, ankyrin-1 (ANK1), hemoglobin A (HBA), and hemoglobin B (HBB), which areRBC-specific, were absent or appeared at very low levels in urine. Ankyrin-1 (206 kDa) is too large to cross the glomerular filter (26,), and any filtered hemoglobin would be reclaimed by the megalin/cubilin receptor system (20), thus their absence from urine is consistent with an intravascular hemolytic event. Fig. 4 Open in new tabDownload slide Neonatal urine proteomics. a) List of 21 proteins that were present in urine samples and also in the top tertile of proteins in the RBC proteome. Among these, CAI had the highest mean level and sample-to-sample spread. Abundance was normalized to µg of total peptide. b) Volcano plot identifying proteins with significantly different levels between CAI-immunopositive (N = 16) and CAI-immunonegative (N = 9) urine samples. Proteins highlighted in red are in the top tertile of proteins found in RBCs. Shading of non-RBC proteins is proportional to the proteins’ abundance. The CAI-immunopositive group was enriched in 16 non-RBC specific proteins, including heme-binding protein 2, SPARC-like protein 1, zinc-α-1-glycoprotein, cathepsin Z, calreticulin, and L-selectin precursor. The CAI-immunonegative group was enriched in 23 non-RBC specific proteins, including calcyphosin, chondrolectin precursor, Lynx-1, platelet glycoprotein VI, tenascin X, and IGF-like family receptor 1. c) Label-free quantification (LFQ) of RBC-related proteins (normalized to urine volume), grouped by neonate category. Levels of CAI and PGK1 were significantly different (one-way ANOVA; *P < 0.05) but CAI manifested larger absolute changes. Fig. 4 Open in new tabDownload slide Neonatal urine proteomics. a) List of 21 proteins that were present in urine samples and also in the top tertile of proteins in the RBC proteome. Among these, CAI had the highest mean level and sample-to-sample spread. Abundance was normalized to µg of total peptide. b) Volcano plot identifying proteins with significantly different levels between CAI-immunopositive (N = 16) and CAI-immunonegative (N = 9) urine samples. Proteins highlighted in red are in the top tertile of proteins found in RBCs. Shading of non-RBC proteins is proportional to the proteins’ abundance. The CAI-immunopositive group was enriched in 16 non-RBC specific proteins, including heme-binding protein 2, SPARC-like protein 1, zinc-α-1-glycoprotein, cathepsin Z, calreticulin, and L-selectin precursor. The CAI-immunonegative group was enriched in 23 non-RBC specific proteins, including calcyphosin, chondrolectin precursor, Lynx-1, platelet glycoprotein VI, tenascin X, and IGF-like family receptor 1. c) Label-free quantification (LFQ) of RBC-related proteins (normalized to urine volume), grouped by neonate category. Levels of CAI and PGK1 were significantly different (one-way ANOVA; *P < 0.05) but CAI manifested larger absolute changes. A comparison of the proteomes of CAI-immunopositive and CAI-immunonegative urine confirms that CAI peptides were significantly enriched in the former (Fig. 4B). An analysis of urine by patient grouping confirmed CAI enrichment in the infected, phototherapy, and subclinical groups (Fig. 4C; one-way ANOVA, P = 0.021). Phosphoglycerate kinase 1 was also differentially abundant in the four groups (P = 0.008), but demonstrated lower intensity than CAI, making it a less suitable biomarker. Discussion This proof-of-principle study demonstrates that CAI can be detected in neonatal urine by immunoreactivity methods (Fig. 2) and mass spectrometry (Fig. 4, Supplemental Table 1). In the cohort of full-term neonates, two-thirds produced distinctly CAI-immunopositive urine (Fig. 2B). The remaining third of neonates produced urine with only trace CAI immunoreactivity, similar to that excreted by healthy adults (Fig. 1D) and consistent with previous reports that CAI is not normally excreted in urine. The source of CAI in urine can be explained in terms of intravascular hemolysis of RBCs, the largest pool of CAI in the body. Moreover, urinary CAI excretion correlated with serum bilirubin, an independent but indirect indicator of hemolysis (Fig. 3). Fecal contamination cannot account for urinary CAI immunoreactivity because measurements on feces-contaminated adult urine (Supplemental Fig. 5) and adult urine dosed with lysates of colorectal epithelial cells (Supplemental Fig. 4) produced negligible CAI signals. Although the appearance of CAI in urine may be predicted from the properties of the glomerular filter (26,), urinary CAI excretion has not been demonstrated in hemolyzing neonates. To explain the urinary CAI signal in immunopositive samples, CAI protein must have been released from a sizable pool of RBCs and then crossed the glomerulus, without subsequent degradation or reabsorption (20). The highest urinary CAI immunoreactivity was equal to a 100-fold dilution of RBC lysate, and the mean signal among CAI-immunopositive samples was equivalent to a 105-fold dilution (Fig. 2EcfFig. 1C). The appearance of CAI in urine samples collected for this study cannot be explained by an extravascular hemolysis along the urinary tract because CAI was not excreted with other RBC-specific proteins, such as ankyrin-1 and hemoglobin (Fig. 4B). This lack of association between CAI and other RBC-specific proteins indicates that the hemolysate underwent filtration and selective reabsorption that eliminated ankyrin-1 and hemoglobin, but allowed CAI to pass. Measurements of CAI in urine have excellent signal-to-noise ratio for detecting hemolytic events. The low background and positive correlation with hemolysis compare favorably to the marker haptoglobin, which falls during hemolysis and therefore requires accurate baseline information for the necessary subtraction. Also, CAI excretion is expected to increase proportionally with the degree of hemolysis, and therefore the detected signal is less prone to saturate. Under nonhemolyzing conditions, urinary CAI signal is close to zero, which compares favorably to markers such as reticulocyte count for which there is always a nonzero baseline level in blood and, consequently, a narrower dynamic range. Markers such as LDH are not specific to ruptured RBCs, and can yield false-positives. CAI testing is more directly traceable to hemolysis and therefore does not share the potential ambiguity of bilirubin measurements, which in rare cases such as Crigler–Najjar syndrome may lead to an erroneous diagnosis of hemolysis. Urine samples with the highest CAI immunoreactivity were obtained from three neonates with significantly raised CRP and in receipt of antibiotics to treat an underlying infection (Fig. 2B). Such CAI levels are indicative of a substantial degree of hemolysis. The source of CAI was not an extravascular hemolysis arising from a urinary tract infection [UTI, a rare event in the first days of life (27,)] because three major RBC-related proteins (ankyrin-1, hemoglobin alpha and beta) were detected at much lower intensities than CAI (400-, 15- and 100-fold, respectively). This discrepancy indicates that RBC proteins had been separated by the kidney. Furthermore, there was no proteomic evidence for markers of squamous epithelium nor neutrophil effector functions, which would be expected in a UTI (28). The squamous epithelium markers cytokeratins and periplakins were not detected in any sample, and there was no significant difference in markers serpin B3, cornulin, and desmoplakin between the high- and low-CRP cohorts (P = 0.35, 0.89, and 0.17, respectively). Neutrophil activity markers cathelicidins and calprotectins were not detected in any sample, and there was no significant difference in markers defensin-1, myeloperoxidase, cathepsin G, cathepsin L1, or pro-cathepsin H between the high- and low-CRP cohorts (P = 0.92, 0.69, 0.72, 0.63, 0.57, respectively). In accordance with routine clinical practice, based largely on SBR measurements, 12 of the recruited neonates were placed on phototherapy (National Institute for Health and Care Excellence guidelines), and among these, urinary CAI was detected in 10. The mean signal in this clinically jaundiced group was significantly above background but not as high as in the group with elevated CRP, reflective of a less severe but still significant hemolytic event. Among those neonates who did not qualify for phototherapy, half manifested raised SBR and had a significantly elevated mean CAI signal. The remaining neonates, characterized by low SBR (<125 µmol/L), produced urine with no detectable CAI (Fig. 3C). Data stratified by SBR level showed a significant association with CAI immunopositivity (Table 1), indicating that the presence of CAI in urine is meaningfully related to an established clinical index of hemolysis. Since the appearance of CAI in urine is expected to peak before the onset of downstream biochemical changes, we postulate that urinary CAI excretion is an early biomarker of hemolysis. To establish these dynamics, follow-up trials are needed to track the time course of urinary CAI excretion in a larger cohort of newborns and then correlate these data with hemolytic markers, including serum haptoglobin. The ELISA-based CAI detection method could be implemented in lateral-flow immunochromatography devices (strip-tests) for rapid readouts under minimal-laboratory conditions (i.e., similar to pregnancy kits). Testing could thus be performed regularly to track the onset and progression of hemolysis in early-life care with the necessary sampling frequency. This compares favorably to blood-based tests for LDH, haptoglobin, or SBR, which cannot achieve such high temporal resolution. A detectable rise in CAI may precede other markers of hemolysis and thus improve resource allocation in the clinic and identify ‘at-risk’ patients earlier. This methodology may be particularly useful in economically deprived regions with inadequate healthcare systems, in ambulatory care, and in developing countries where the incidence of hemolytic triggers (e.g., malaria, sickle cell disease) is higher. Author Contributions: All authors confirmed they have contributed to the intellectual content of this paper and have met the following 4 requirements: (a) significant contributions to the conception and design, acquisition of data, or analysis and interpretation of data; (b) drafting or revising the article for intellectual content; (c) final approval of the published article; and (d) agreement to be accountable for all aspects of the article thus ensuring that questions related to the accuracy or integrity of any part of the article are appropriately investigated and resolved. A. Hulikova, experimental design, measurements and analysis; H. Khan, clinical lead and provision of samples and clinical data; H. Kramer, urine proteomics and mass spectrometry analysis; P. Swietach, financial support, statistical analysis, administrative support, experimental design, analysis, and wrote the paper. Authors’ Disclosures or Potential Conflicts of Interest: Upon manuscript submission, all authors completed the author disclosure form. Authors' disclosures and/or potential conflicts of interest: Employment or Leadership: None declared. Consultant or Advisory Role: None declared. Stock Ownership: None declared. Honoraria: None declared. Research Funding: OUP-John Fell Fund (152/042). Expert Testimony: None declared. Patents: None declared. Human genes: Car1 Trial registration: CPMS 19576. Role of Sponsor: The funding organizations played no role in the design of study, choice of enrolled patients, review and interpretation of data, preparation of manuscript, or final approval of manuscript. Acknowledgments We thank Drs. Oliver Lomas and Killian Donovan (John Radcliffe Hospital, Oxford) for critically commenting on the manuscript. REFERENCES 1 O’Brien RT , Pearson HA. Physiologic anemia of the newborn infant . J Pediatr 1971 ; 79 : 132 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 2 Murray NA , Roberts IA. Haemolytic disease of the newborn . Arch Dis Child Fetal Neonatal Ed 2007 ; 92 : F83 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 3 Slusher TM , Zamora TG, Appiah D, Stanke JU, Strand MA, Lee BW, et al. Burden of severe neonatal jaundice: a systematic review and meta-analysis . BMJ 2017 ; 1 : e000105 . OpenURL Placeholder Text WorldCat 4 Mitra S , Rennie J. Neonatal jaundice: aetiology, diagnosis and treatment . Br J Hosp Med (Lond) 2017 ; 78 : 699 – 704 . Google Scholar Crossref Search ADS PubMed WorldCat 5 Dhaliwal G , Cornett PA, Tierney LM Jr. Hemolytic anemia . Am Fam Physician 2004 ; 69 : 2599 – 606 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 6 Moise KJ. Hemolytic disease of the fetus and newborn . Clin Adv Hematol Oncol 2013 ; 11 : 664 – 6 . Google Scholar PubMed OpenURL Placeholder Text WorldCat 7 Kumar S , Regan F. Management of pregnancies with RhD alloimmunisation . BMJ 2005 ; 330 : 1255 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 8 Moise KJ. Jr., Management of rhesus alloimmunization in pregnancy . Obstet Gynecol 2008 ; 112 : 164 – 76 . Google Scholar Crossref Search ADS PubMed WorldCat 9 Parker V , Tormey CA. The direct antiglobulin test: Indications, interpretation, and pitfalls . Arch Pathol Lab Med 2017 ; 141 : 305 – 10 . Google Scholar Crossref Search ADS PubMed WorldCat 10 Bryk AH , Wiśniewski JR. Quantitative analysis of human red blood cell proteome . J Proteome Res 2017 ; 16 : 2752 – 61 . Google Scholar Crossref Search ADS PubMed WorldCat 11 D’Alessandro A , Righetti PG, Zolla L. The red blood cell proteome and interactome: an update . J Proteome Res 2010 ; 9 : 144 – 63 . Google Scholar Crossref Search ADS PubMed WorldCat 12 Maren TH. Carbonic anhydrase: Chemistry, physiology, and inhibition . Physiol Rev 1967 ; 47 : 595 – 781 . Google Scholar Crossref Search ADS PubMed WorldCat 13 Wahlstrand T , Knuuttila KG, Wistrand PJ. A radioimmunosorbent technique for the assay of b- and c-types of carbonic anhydrase in human tissues . Scand J Clin Lab Invest 1979 ; 39 : 503 – 9 . Google Scholar Crossref Search ADS PubMed WorldCat 14 Brady HJ , Sowden JC, Edwards M, Lowe N, Butterworth PH. Multiple gf-1 binding sites flank the erythroid specific transcription unit of the human carbonic anhydrase i gene . FEBS Lett 1989 ; 257 : 451 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 15 Sly WS , Hu PY. Human carbonic anhydrases and carbonic anhydrase deficiencies . Annu Rev Biochem 1995 ; 64 : 375 – 401 . Google Scholar Crossref Search ADS PubMed WorldCat 16 Tashian RE. The carbonic anhydrases: Widening perspectives on their evolution, expression and function . Bioessays 1989 ; 10 : 186 – 92 . Google Scholar Crossref Search ADS PubMed WorldCat 17 Backman U , Danielsson B, Wistrand PJ. The excretion of carbonic anhydrase isozymes ca i and ca ii in the urine of apparently healthy subjects and in patients with kidney disease . Scand J Clin Lab Invest 1990 ; 50 : 627 – 33 . Google Scholar Crossref Search ADS PubMed WorldCat 18 Supuran CT. Carbonic anhydrases: Novel therapeutic applications for inhibitors and activators . Nat Rev Drug Discov 2008 ; 7 : 168 – 81 . Google Scholar Crossref Search ADS PubMed WorldCat 19 Robinson JR. Urinary excretion of carbonic anhydrase: a simple test for the detection of intravascular haemolysis . J Clin Pathol 1950 ; 3 : 142 – 5 . Google Scholar Crossref Search ADS PubMed WorldCat 20 Christensen EI , Birn H. Megalin and cubilin: Multifunctional endocytic receptors . Nat Rev Mol Cell Biol 2002 ; 3 : 258 – 66 . Google Scholar Crossref Search ADS WorldCat 21 Wisniewski JR , Zougman A, Nagaraj N, Mann M. Universal sample preparation method for proteome analysis . Nat Methods 2009 ; 6 : 359 – 62 . Google Scholar Crossref Search ADS PubMed WorldCat 22 Cox J , Hein MY, Luber CA, Paron I, Nagaraj N, Mann M. Accurate proteome-wide label-free quantification by delayed normalization and maximal peptide ratio extraction, termed maxlfq . Mol Cell Proteom 2014 ; 13 : 2513 – 26 . Google Scholar Crossref Search ADS WorldCat 23 El Hamel C , Chianea T, Thon S, Lepichoux A, Yardin C, Guigonis V. Normal values of urine total protein- and albumin-to-creatinine ratios in term newborns . Pediatr Nephrol 2017 ; 32 : 113 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 24 Sowden J , Leigh S, Talbot I, Delhanty J, Edwards Y. Expression from the proximal promoter of the carbonic anhydrase 1 gene as a marker for differentiation in colon epithelia . Differentiation 1993 ; 53 : 67 – 74 . Google Scholar Crossref Search ADS PubMed WorldCat 25 Perrone S , Lotti F, Longini M, Rossetti A, Bindi I, Bazzini F, et al. C reactive protein in healthy term newborns during the first 48 hours of life . Arch Dis Child Fetal Neonatal Ed 2018 ; 103 : F163 – 6 . Google Scholar Crossref Search ADS PubMed WorldCat 26 Chang RL , Deen WM, Robertson CR, Brenner BM. Permselectivity of the glomerular capillary wall: Iii. Restricted transport of polyanions . Kidney Int 1975 ; 8 : 212 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 27 Visser VE , Hall RT. Urine culture in the evaluation of suspected neonatal sepsis . J Pediatr 1979 ; 94 : 635 – 8 . Google Scholar Crossref Search ADS PubMed WorldCat 28 Yu Y , Sikorski P, Smith M, Bowman-Gholston C, Cacciabeve N, Nelson KE, Pieper R. Comprehensive metaproteomic analyses of urine in the presence and absence of neutrophil-associated inflammation in the urinary tract . Theranostics 2017 ; 7 : 238 – 52 . Google Scholar Crossref Search ADS PubMed WorldCat © American Association for Clinical Chemistry 2020. All rights reserved. For permissions, please email: journals.permissions@oup.com. 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)
TI - Detection of Intravascular Hemolysis in Newborn Infants Using Urinary Carbonic Anhydrase I Immunoreactivity
JO - The Journal of Applied Laboratory Medicine
DO - 10.1093/jalm/jfaa051
DA - 2020-09-01
UR - https://www.deepdyve.com/lp/oxford-university-press/detection-of-intravascular-hemolysis-in-newborn-infants-using-urinary-1wLChx097N
SP - 921
EP - 934
VL - 5
IS - 5
DP - DeepDyve
ER -