To investigate the impact of sampling methodology on gene expression data from primary estrogen receptor–positive (ERþ) breast cancer biopsies, global gene expression was measured in core-cut biopsies at baseline and surgery from patients randomly assigned to receive either two weeks of presurgical aromatase inhibitor (AI; n ¼ 157) or no presurgical treatment (n ¼ 56). Those genes most markedly altered in the AI group (eg, FOS, DUSP1, RGS1, FOSB) were similarly altered in the no treatment group; some widely investigated genes that were apparently unaffected in the AI group (eg, MYC) were counter- altered in the control group, masking actual AI-dependent changes. In the absence of a control group, these artefactual changes would likely lead to the most affected genes being the erroneous focus of research. The ﬁndings are likely relevant to all archival collections of ERþ breast cancer. Analysis of gene expression in biopsies taken before and after two weeks presurgery þ two weeks postsurgery, termed AI- treatment of primary breast cancer (BC) is frequently undertaken treated) or no perioperative treatment (termed control). to study mechanisms of response and resistance. We and others Core-cut biopsy samples in RNA later were analyzed from both have identified artefactual changes in gene expression that can the baseline and surgical sample from 213 patients (157 were result from the study procedures (1,2). Importantly, however, the all good-quality available AI-treated and 56 were randomly cho- degree of impact of those changes has not been evaluated in the sen controls). High-quality genome-wide expression data context of a specific therapy. The current study reveals the po- (GSE105777) were analyzed to identify statistically significant al- tential for profound errors in data interpretation that could occur tered gene expression and were compared between the AI- if such artefacts are not identified or are ignored. treated and control groups. Classical clinical factors were well The PeriOperative Endocrine Therapy-Individualising Care balanced between the two groups (Supplementary Table 1). (POETIC; CRUK/07/015) (3) trial randomly assigned 4486 post- A total of 3269 genes (n ¼ 1504 upregulated, n ¼ 1765 down- menopausal women with primary estrogen receptor–positive regulated) from treated tumors and 110 genes (n ¼ 70 upregu- (ERþ) BC 2:1 to receive perioperative aromatase inhibitor (AI; lated, n ¼ 40 downregulated) from control tumors were Received: December 19, 2017; Accepted: February 16, 2018 © The Author(s) 2018. Published by Oxford University Press. This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/4.0/), which permits non-commercial re-use, distribution, and reproduction in any medium, provided the original work is properly cited. For commercial re-use, please contact email@example.com Downloaded from https://academic.oup.com/jncics/article-abstract/2/2/pky005/5001448 by Ed 'DeepDyve' Gillespie user on 21 June 2018 2| JNCI J Natl Cancer Inst, 2018, Vol. 0, No. 0 differentially expressed between baseline and surgical samples, POETIC AI-treated and control cohorts (Figure 1) shows the near with P values of less than .001 by paired t test on the normalized identical quantitative levels of change for the genes most af- expression data (Supplementary Tables 2 and 3 and fected by each treatment. In essence, the changes of greatest Supplementary Figure 1, available online). In the control group, magnitude in the AI-treated tumors are completely artefactual. 25 of the top 30 differentially expressed genes ranked by an ab- In addition, there are widely investigated genes, such as MYC, solute fold-change (FC) were upregulated at surgery (Table 1), as that are statistically significantly altered in the control samples, previously reported (1). Many of these were early-response or but apparently unaffected in the treated group. Correction for stress-associated genes (FOS, RGS1, DUSP1, FOSB, JUN, and MYC), the degree of change in MYC expression within the control group as well as CD69 (FC ¼ 1.39, P ¼ 1.48E-4), an early T-cell activation reveals the highly statistically significant suppression of MYC as antigen associated with immune response (4), and TOB1 (FC ¼ a result of AI therapy (P ¼ .0003). 1.28, P ¼ 5.2E-06), an important mediator involved in T-cell acti- We examined gene expression data from an independent set vation and a critical determinant of estrogen-independent ERþ of core cuts taken at baseline and two weeks after starting AI BC survival (5). The five downregulated genes included those treatment in the FAIMoS study (6). The same markedly affected encoding hemoglobin (HBA and HBB). genes identified in the POETIC AI-treated arm were not ob- In the AI-treated cohort, in contrast to the mainly upregu- served in the top-ranked 30 genes identified in FAIMoS lated genes in the control group, 22 of the top 30 ranked genes (Supplementary Table 4). This discrepancy is likely to be (Table 1) were downregulated after two weeks of AI treatment. explained by the way the tissue core was taken and manipu- Most of these corresponded to proliferation, estrogen-respon- lated after therapy. The key difference between the POETIC and sive, or cycling genes. However, 11 of the 30 top-ranked differen- FAIMoS studies was that for the former the post-therapy core tially expressed genes in the AI-treated group (eight upregulated was from an excised tumor while in FAIMoS the core was taken and three downregulated) were among the 12 most altered genes from the breast with the tumor in situ. Thus, in FAIMoS, the in the control group, including FOS, DUSP1, and RGS1. These processes of sampling for the baseline and on-treatment core latter three genes were the most affected in both cohorts. A scat- biopsies were identical, while in POETIC, the second core biopsy terplot of the changes in individual gene expression between the was subject to a variable degree of ischemic conditions prior to Table 1. Top ranked genes in control and AI-treated samples. Top 30 regulated genes in control tumors and in AI-treated tumors. Rank in Parametric Rank in Rank in Parametric Rank in Symbol control† FC (S/B) ‡ p-value FDRˆ AI-treated† Symbol AI-treated† FC (S/B) ‡ p-value FDR control† -07 -07 -07 -07 FOS* 1 3.87 < 1x10 < 1x10 1 FOS* 1 4.14 < 1x10 < 1x10 1 -07 -07 -07 -07 RGS1* 2 3.22 < 1x10 < 1x10 3 DUSP1* 2 3.36 < 1x10 < 1x10 3 -07 -07 -07 -07 DUSP1* 3 3.06 < 1x10 < 1x10 2 RGS1* 3 3.33 < 1x10 < 1x10 2 -07 -04 -07 -07 HBA2* 4 -2.94 9.00x10 9.10x10 7 TFF1 4 -2.94 < 1x10 < 1x10 NA -06 -03 -07 -07 HBB* 5 -2.86 3.40x10 2.06x10 7 FOSB* 5 2.65 < 1x10 < 1x10 7 -06 -03 -07 -07 HBA1* 6 -2.50 6.30x10 3.21x10 12 TOP2A 6 -2.56 < 1x10 < 1x10 NA -07 -04 -07 -07 FOSB* 7 2.30 9.00x10 9.10x10 5 UBE2C 7 -2.44 < 1x10 < 1x10 NA -05 -02 -07 -07 RNY5 8 -2.13 5.04x10 1.39x10 83 HBB* 7 -2.44 < 1x10 < 1x10 5 -06 -03 -07 -07 CYR61* 9 2.05 1.70x10 1.43x10 10 HBA2* 7 -2.44 < 1x10 < 1x10 4 -06 -04 -07 -07 EGR1* 10 1.98 1.00x10 9.47x10 10 EGR1* 10 2.36 < 1x10 < 1x10 10 -07 -07 -07 -07 ZFP36* 11 1.97 < 1x10 < 1x10 18 CYR61* 10 2.36 < 1x10 < 1x10 9 -07 -04 -07 -07 SNORD3D* 12 1.86 4.00x10 4.66x10 30 CDC20 12 -2.13 < 1x10 < 1x10 NA -07 -07 -07 -07 TRK1 13 -1.72 < 1x10 < 1x10 49 NUSAP1 12 -2.13 < 1x10 < 1x10 NA -07 -07 -07 -07 JUN 14 1.67 < 1x10 < 1x10 65 HBA1* 12 -2.13 < 1x10 < 1x10 6 -07 -04 -07 -07 SGK1 15 1.62 < 1x10 1.68x10 32 SUSD3 15 -2.04 < 1x10 < 1x10 NA -05 -03 -07 -07 SNORD3A 15 1.62 1.61x10 6.10x10 79 FGFR3 16 -2.00 < 1x10 < 1x10 NA -05 -03 -07 -07 SNORD3C 17 1.61 1.30x10 5.05x10 97 NEK2 17 -1.96 < 1x10 < 1x10 NA -04 -02 -07 -07 LOC100132564 19 1.53 4.12x10 7.70x10 NA ZFP36* 18 1.94 < 1x10 < 1x10 11 -05 -03 -07 -07 RASD1 20 1.52 1.22x10 4.87x10 39 UHRF1 19 -1.92 < 1x10 < 1x10 NA -05 -02 -07 -07 RGS2 21 1.51 9.74x10 2.38x10 33 PRC1 19 -1.92 < 1x10 < 1x10 NA -05 -02 -07 -07 SNORD13 22 1.49 5.52x10 1.49x10 80 ASPM 19 -1.92 < 1x10 < 1x10 NA -06 -03 -07 -07 CCL3L3 23 1.48 6.80x10 3.21x10 80 AGR2 19 -1.92 < 1x10 < 1x10 NA -04 -02 -07 -07 KLF6 24 1.47 5.29x10 9.02x10 99 PDZK1 23 -1.85 < 1x10 < 1x10 NA -06 -03 -07 -07 APOLD1 25 1.45 2.00x10 1.60x10 71 PTTG1 24 -1.82 < 1x10 < 1x10 NA -05 -02 -07 -07 ATF3 26 1.43 3.71x10 1.12x10 138 ADCY1 24 -1.82 < 1x10 < 1x10 NA -05 -02 -07 -07 SPRY1 26 1.43 9.59x10 2.38x10 36 CDCA5 26 -1.79 < 1x10 < 1x10 NA -05 -03 -07 -07 MYC 28 1.41 2.28x10 8.23x10 NA CCNB2 26 -1.79 < 1x10 < 1x10 NA -06 -03 -07 -07 CEBPD 29 1.39 4.60x10 2.68x10 226 KIAA0101 26 -1.79 < 1x10 < 1x10 NA -06 -03 -07 -07 BTG2 29 1.39 9.20x10 3.77x10 138 STC2 26 -1.79 < 1x10 < 1x10 NA -04 -02 -07 -07 CD69 29 1.39 1.48x10 3.31x10 65 SNORD3D* 30 1.77 < 1x10 < 1x10 12 *Highlights the 11 genes regulated in AI-treated group that are among the 12 most regulated genes in control. AI ¼ aromatase inhibitor; B ¼ baseline; FC ¼ fold-change; FDR ¼ false discovery rate; S ¼ surgery. †Rank: by an absolute fold-change. ‡FC (S/B): fold-change of individual genes at surgery compared with baseline. Downloaded from https://academic.oup.com/jncics/article-abstract/2/2/pky005/5001448 by Ed 'DeepDyve' Gillespie user on 21 June 2018 Q. Gao et al. |3 Figure 1. Whole-genome gene expression changes between baseline and surgery in PeriOperative Endocrine Therapy-Individualising Care (POETIC): aromatase inhibi- tor (AI)–treated (n ¼ 157) vs control (n ¼ 56). The largest changes (above or below the dotted line) in gene expression seen with AI treatment are artefactual. PC ¼ POETIC control; PT ¼ POETIC treated. placement in RNA later. In many cases, the time of ischemia will have been subject to similar or perhaps greater ischemic will have been variably extended by the time taken to x-ray the conditions before fixation than the core-cut samples in POETIC. biopsy sample. Our earlier report described the changes that re- Investigators should establish that the collection process does sult from that delay in genes such as RGS1 and DUSP1, which not affect expression of the genes of interest prior to assuming are among the most affected genes in the AI-treated and control that the observed expression reflects the true expression in the arms in POETIC. Of note, in support of this explanation, the de- tumor in situ. crease seen in MYC expression in the POETIC AI-treated group after correction for the artefactual increase in the controls con- FUNDING curs with the statistically signifcant decrease seen in FAIMoS. We conclude that the majority of the most upregulated This work was supported by the Mary-Jean Mitchell Green genes (eg, FOS, DUSP1, RGS1, FOSB) and a small number of the Foundation, Breast Cancer Now, working in partnership most downregulated genes are identified as a result of pre- with Walk the Walk and the National Institutes of Health analytical sample processing. In addition, the true effect of AI Research Royal Marsden/Institute of Cancer Research treatment on other genes can be hidden by counteractive arte- Biomedical Research Centre. The POETIC trial (C1491/A8671/ factual change. In the absence of a control group, investigators CRUK/07/015, C1491/A15955, C406/A8962), from which sam- are likely to focus on the most extensive gene changes, yet ples were obtained for this study, was supported by Cancer these will include many ascribed wrongly to the effect of experi- Research UK, as is Institute of Cancer Research- Clinical mental intervention; some genes that would be the focus will be Trials and Statistical Unit through its core program grant. wrongly ignored because they are apparently unaffected by therapy. It is notable that our observations have been made in the context of withdrawal of estrogen stimulation, the strongest Notes transcriptional drive for ERþ BC. The artefacts are likely to be pronounced relative to true effects in the context of less impact- Affiliations of authors: Breast Cancer Now Research Centre (QG, ful therapy. Future presurgical studies should ensure that core ELK, RR, LAM, MD) and Clinical Trials and Statistics Unit (MCUC, cuts taken at surgery are either taken in an identical fashion to JM, JMB), The Institute of Cancer Research, London, UK; Ralph those at baseline or that a control group of patients is included Lauren Centre for Breast Cancer Research (ELK, KS, DE, VM, AD, to identify any process-related changes. MD) and Breast Unit, Royal Marsden Hospital, London, UK (IS); It should also be recognized that the majority of tissue- Royal Bournemouth Hospital, Bournemouth, UK (AS); Royal related studies in BC occur in archival excision specimens that Liverpool University Hospital, Liverpool, UK (CH); Queen Downloaded from https://academic.oup.com/jncics/article-abstract/2/2/pky005/5001448 by Ed 'DeepDyve' Gillespie user on 21 June 2018 4| JNCI J Natl Cancer Inst, 2018, Vol. 0, No. 0 Elizabeth University Hospital Glasgow, Govan, UK (EM); Poole References General Hospital, Dorset, UK (AE); University of Nottingham, 1. Lopez-Knowles E, Gao Q, Cheang MC, et al. Heterogeneity in global gene ex- Nottingham, UK (JR). pression proﬁles between biopsy specimens taken peri-surgically from pri- We thank all participating patients and staff at POETIC centers mary ER-positive breast carcinomas. Breast Cancer Res. 2016;18(1):39. 2. Pearce DA, Arthur LM, Turnbull AK, et al. Tumor sampling method can signiﬁ- and ICR-CTSU who contributed to trial management and data and cantly inﬂuence gene expression proﬁles derived from neoadjuvant window sample collection, and the independent data monitoring commit- studies. Sci Rep. 2016;6:29434. tee and trial steering committee, for their oversight of the trial. 3. Dowsett M, Smith I, Robertson J, et al. Endocrine therapy, new biologicals, and new study designs for presurgical studies in breast cancer. J Natl Cancer Inst QG analyzed the data and drafted the manuscript. ELK Monogr. 2011;2011(43):120–123. extracted RNA and drafted the manuscript. MC generated sub- 4. Shiow LR, Rosen DB, Brdickova N, et al. CD69 acts downstream of interferon- types. JM provided data and composed Supplementary Table 1. alpha/beta to inhibit S1P1 and lymphocyte egress from lymphoid organs. Nature. 2006;440(7083):540–544. KS and DE recorded the samples for the study. VM and AD sec- 5. Zhang YW, Nasto RE, Varghese R, et al. Acquisition of estrogen independence tioned and reviewed the histopathology of the samples. AS, HC, induces TOB1-related mechanisms supporting breast cancer cell proliferation. EM, EA, and JR were involved in sample acquisition. MD, IS, and Oncogene. 2016;35(13):1643–1656. 6. Smith IE, Walsh G, Skene A, et al. A phase II placebo-controlled trial of neoad- JB were involved in conception and design of POETIC. LAM, RR, juvant anastrozole alone or with geﬁtinib in early breast cancer. J Clin Oncol. and MD drafted the manuscript. All authors read and approved 2007;25(25):3816–3822. the final manuscript. Downloaded from https://academic.oup.com/jncics/article-abstract/2/2/pky005/5001448 by Ed 'DeepDyve' Gillespie user on 21 June 2018
JNCI Cancer Spectrum – Oxford University Press
Published: May 22, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.
All for just $49/month
Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly
Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.
Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera