Population Differences in the Rate of Proliferation of International HapMap Cell Lines
2010; Elsevier BV; Volume: 87; Issue: 6 Linguagem: Inglês
10.1016/j.ajhg.2010.10.018
ISSN1537-6605
AutoresAmy L. Stark, Wei Zhang, Tong Zhou, Peter H. O’Donnell, Christine M. Beiswanger, R. Stephanie Huang, Nancy J. Cox, M. Eileen Dolan,
Tópico(s)T-cell and B-cell Immunology
ResumoThe International HapMap Project is a resource for researchers containing genotype, sequencing, and expression information for EBV-transformed lymphoblastoid cell lines derived from populations across the world. The expansion of the HapMap beyond the four initial populations of Phase 2, referred to as Phase 3, has increased the sample number and ethnic diversity available for investigation. However, differences in the rate of cellular proliferation between the populations can serve as confounders in phenotype-genotype studies using these cell lines. Within the Phase 2 populations, the JPT and CHB cell lines grow faster (p < 0.0001) than the CEU or YRI cell lines. Phase 3 YRI cell lines grow significantly slower than Phase 2 YRI lines (p < 0.0001), with no widespread genetic differences based on common SNPs. In addition, we found significant growth differences between the cell lines in the Phase 2 ASN populations and the Han Chinese from the Denver metropolitan area panel in Phase 3 (p < 0.0001). Therefore, studies that separate HapMap panels into discovery and replication sets must take this into consideration. The International HapMap Project is a resource for researchers containing genotype, sequencing, and expression information for EBV-transformed lymphoblastoid cell lines derived from populations across the world. The expansion of the HapMap beyond the four initial populations of Phase 2, referred to as Phase 3, has increased the sample number and ethnic diversity available for investigation. However, differences in the rate of cellular proliferation between the populations can serve as confounders in phenotype-genotype studies using these cell lines. Within the Phase 2 populations, the JPT and CHB cell lines grow faster (p < 0.0001) than the CEU or YRI cell lines. Phase 3 YRI cell lines grow significantly slower than Phase 2 YRI lines (p < 0.0001), with no widespread genetic differences based on common SNPs. In addition, we found significant growth differences between the cell lines in the Phase 2 ASN populations and the Han Chinese from the Denver metropolitan area panel in Phase 3 (p < 0.0001). Therefore, studies that separate HapMap panels into discovery and replication sets must take this into consideration. The International HapMap Project was designed to characterize genetic variation of different individuals from multiple populations, and the samples generated for the project provide the research community with a unique resource for investigating genetics and cellular phenotypes that would be difficult to investigate in humans. Phases 1 and 2 (referred to as Phase 2 for this study) of the project focused on 270 Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (LCLs) from four populations: 90 Yoruba (YRI) individuals comprised of 30 trios collected from Ibadan, Nigeria; 45 unrelated Japanese individuals collected from Tokyo (JPT) and 45 unrelated Han Chinese individuals collected from Beijing (CHB), often considered together as 90 Asian (ASN) lines; and 90 Utah residents with ancestry from northern and western Europe (CEU), comprising 30 trios. Phase 3 of the HapMap project expanded the populations and number of cell lines available. The original populations were supplemented, and seven additional populations were added. With the exception of the CEU cell lines, which were collected and transformed approximately thirty years ago, all of the other cell lines were collected within the last ten years and transformed by Coriell Institute for Medical Research (Camden, NJ, USA). The HapMap cell lines provide extensive publicly available genotyping data,1International HapMap ConsortiumThe International HapMap Project.Nature. 2003; 426: 789-796Crossref PubMed Scopus (4975) Google Scholar, 2Frazer K.A. Ballinger D.G. Cox D.R. Hinds D.A. Stuve L.L. Gibbs R.A. Belmont J.W. Boudreau A. Hardenbol P. Leal S.M. et al.International HapMap ConsortiumA second generation human haplotype map of over 3.1 million SNPs.Nature. 2007; 449: 851-861Crossref PubMed Scopus (3599) Google Scholar allowing for studies of the contribution of genetics to baseline gene expression3Morley M. Molony C.M. Weber T.M. Devlin J.L. Ewens K.G. Spielman R.S. Cheung V.G. Genetic analysis of genome-wide variation in human gene expression.Nature. 2004; 430: 743-747Crossref PubMed Scopus (950) Google Scholar, 4Dixon A.L. Liang L. Moffatt M.F. Chen W. Heath S. Wong K.C. Taylor J. Burnett E. Gut I. Farrall M. et al.A genome-wide association study of global gene expression.Nat. Genet. 2007; 39: 1202-1207Crossref PubMed Scopus (790) Google Scholar, 5Spielman R.S. Bastone L.A. Burdick J.T. Morley M. Ewens W.J. Cheung V.G. Common genetic variants account for differences in gene expression among ethnic groups.Nat. Genet. 2007; 39: 226-231Crossref PubMed Scopus (395) Google Scholar, 6Storey J.D. Madeoy J. Strout J.L. Wurfel M. Ronald J. Akey J.M. Gene-expression variation within and among human populations.Am. J. Hum. Genet. 2007; 80: 502-509Abstract Full Text Full Text PDF PubMed Scopus (225) Google Scholar, 7Stranger B.E. Nica A.C. Forrest M.S. Dimas A. Bird C.P. Beazley C. Ingle C.E. Dunning M. Flicek P. Koller D. et al.Population genomics of human gene expression.Nat. Genet. 2007; 39: 1217-1224Crossref PubMed Scopus (907) Google Scholar, 8Zhang W. Duan S. Kistner E.O. Bleibel W.K. Huang R.S. Clark T.A. Chen T.X. Schweitzer A.C. Blume J.E. Cox N.J. Dolan M.E. Evaluation of genetic variation contributing to differences in gene expression between populations.Am. J. Hum. Genet. 2008; 82: 631-640Abstract Full Text Full Text PDF PubMed Scopus (158) Google Scholar and pharmacologic phenotypes.9Huang R.S. Kistner E.O. Bleibel W.K. Shukla S.J. Dolan M.E. Effect of population and gender on chemotherapeutic agent-induced cytotoxicity.Mol. Cancer Ther. 2007; 6: 31-36Crossref PubMed Scopus (66) Google Scholar, 10Jones T.S. Yang W. Evans W.E. Relling M.V. Using HapMap tools in pharmacogenomic discovery: the thiopurine methyltransferase polymorphism.Clin. Pharmacol. Ther. 2007; 81: 729-734Crossref PubMed Scopus (35) Google Scholar, 11Hartford C.M. Duan S. Delaney S.M. Mi S. Kistner E.O. Lamba J.K. Huang R.S. Dolan M.E. Population-specific genetic variants important in susceptibility to cytarabine arabinoside cytotoxicity.Blood. 2009; 113: 2145-2153Crossref PubMed Scopus (70) Google Scholar These lines have been utilized for other studies, including evaluation of copy-number variation,12Conrad D.F. Andrews T.D. Carter N.P. Hurles M.E. Pritchard J.K. A high-resolution survey of deletion polymorphism in the human genome.Nat. Genet. 2006; 38: 75-81Crossref PubMed Scopus (534) Google Scholar, 13Redon R. Ishikawa S. Fitch K.R. Feuk L. Perry G.H. Andrews T.D. Fiegler H. Shapero M.H. Carson A.R. Chen W. et al.Global variation in copy number in the human genome.Nature. 2006; 444: 444-454Crossref PubMed Scopus (3323) Google Scholar, 14Stranger B.E. Forrest M.S. Dunning M. Ingle C.E. Beazley C. Thorne N. Redon R. Bird C.P. de Grassi A. Lee C. et al.Relative impact of nucleotide and copy number variation on gene expression phenotypes.Science. 2007; 315: 848-853Crossref PubMed Scopus (1352) Google Scholar and are part of the 1000 Genomes Project,15Zhang W. Dolan M.E. Impact of the 1000 genomes project on the next wave of pharmacogenomic discovery.Pharmacogenomics. 2010; 11: 249-256Crossref PubMed Scopus (38) Google Scholar in which low-pass sequencing is used to identify most genetic variants that have frequencies of at least 1% in the populations studied. Thus, the cell lines are extraordinarily rich in genetic information, making them valuable for genotype-phenotype studies using the results of any cellular phenotype. A major advantage to these cell lines is that they offer an alternative to pharmacogenomic studies that would be considered challenging, if not impossible, to perform in human subjects.16Welsh M. Mangravite L. Medina M.W. Tantisira K. Zhang W. Huang R.S. McLeod H. Dolan M.E. Pharmacogenomic discovery using cell-based models.Pharmacol. Rev. 2009; 61: 413-429Crossref PubMed Scopus (101) Google Scholar The fundamental challenge in attempting to identify pharmacogenomic markers from patient trials is that such studies would require a homogenous population of patients treated with the same dosage regimen and minimal confounding variables. In oncology, the standard of care tends to change as new therapies are tested, and the vast majority of patients receive multiple drugs. In order to facilitate the use of cell lines for pharmacogenetic studies, investigators have undertaken a variety of studies designed to understand potential variability among cell lines, such as the EBV baseline copy-number, rate of cellular proliferation, and ATP levels.17Choy E. Yelensky R. Bonakdar S. Plenge R.M. Saxena R. De Jager P.L. Shaw S.Y. Wolfish C.S. Slavik J.M. Cotsapas C. et al.Genetic analysis of human traits in vitro: drug response and gene expression in lymphoblastoid cell lines.PLoS Genet. 2008; 4: e1000287Crossref PubMed Scopus (183) Google Scholar Furthermore, the effect of one confounding variable, cellular proliferation rate, on cellular susceptibility to chemotherapeutic-induced cytotoxicity using these LCLs was previously reported.18Stark A.L. Zhang W. Mi S. Duan S. O'Donnell P.H. Huang R.S. Dolan M.E. Heritable and non-genetic factors as variables of pharmacologic phenotypes in lymphoblastoid cell lines.Pharmacogenomics J. 2010; PubMed Google Scholar In this study, we have made a systematic examination of the rate of proliferation within and among populations, with the goal of recognizing and providing additional tools for future study involving these LCLs. The alamarBlue assay described previously was used to measure cellular growth rate in all cell lines.18Stark A.L. Zhang W. Mi S. Duan S. O'Donnell P.H. Huang R.S. Dolan M.E. Heritable and non-genetic factors as variables of pharmacologic phenotypes in lymphoblastoid cell lines.Pharmacogenomics J. 2010; PubMed Google Scholar Of the 11 populations included in the HapMap Project, only the CEU LCLs existed as previously established cell lines. The other ten populations were collected and established as cell lines specifically for the HapMap Project over the years 2002 through 2007. Peripheral-blood samples were collected from the participating populations according to strict institutional review board and community-engagement protocols. Mononuclear cells were isolated from the samples and transformed with EBV as described previously.19Beck J.C. Beiswanger C.M. John E.M. Satariano E. West D. Successful transformation of cryopreserved lymphocytes: a resource for epidemiological studies.Cancer Epidemiol. Biomarkers Prev. 2001; 10: 551-554PubMed Google Scholar Cell lines were minimally expanded (approximately 100-fold expansion) before inclusion in the HapMap population panels. In the original Phase 2 populations, there was no significant difference in the cellular proliferation rate between the YRI and CEU lines; however, the ASN (JPT and CHB) cell lines grew faster than the other two populations, as seen in Figure 1 (p < 0.0001). We also observed a significant difference in the rate of cellular proliferation within the cell lines derived from YRI samples: the second set of 90 YRI cell lines (Phase 3) grow significantly slower than the first set of 90 YRI lines (Phase 2), with p < 0.0001 (Figure 1). In contrast, there are no significant growth-rate differences between the two panels of the CEU cell lines. All population-growth comparisons were performed as conservative, nonparametric t tests. In addition, we evaluated growth differences between the cell lines in the Phase 2 ASN populations with an Asian population released in Phase 3, the Han Chinese from the Denver metropolitan area panel. The requirement of having at least three grandparents of Han Chinese ancestry was the same for both panels. Again, we observe a significantly (p = 0.0002) different rate of cellular proliferation (Figure 1). The significant difference remains even when comparing only the Chinese from Beijing with the Chinese from Denver (p < 0.0001, data not shown). These differences across populations with apparently similar genetic backgrounds provoked a more detailed investigation into whether the differences may be attributed to genetic or environmental causes. Principal component analysis (PCA) was performed with the non-linear iterative partial least-squares (NIPALS)20Wold S. Esbensen K. Geladi P. Principal components analysis.Chemom. Intell. Lab. Syst. 1987; 2: 37-52Crossref Scopus (8119) Google Scholar algorithm in the ade4 library21Dray S. Dufour A.B. The ade4 package: implementing the duality diagram for ecologists.J. Stat. Softw. 2007; 22: 1-20Crossref Scopus (3842) Google Scholar of the R package and the use of ∼650,000 common SNPs (minor allele frequency [MAF] > 0.05) between the two panels; the two populations did not separate into distinct clusters (Figure 2). The first two major principal components explain only 1% of the total data variance. This confirmed the expectation that the two panels of the YRI population are not genetically different. Because the panels did not separate, linked SNPs could not have been driving the separation. However, we repeated the analysis using only unlinked SNPs, with similar results (Figures S1–S4 available online). Similarly, we found no difference between the Han Chinese samples from Beijing in Phase 2 and those from the Denver metropolitan area in Phase 3 by using principal components (data not shown). We calculated Fst values, comparing the YRI Phase 2 and Phase 3 panels and found no evidence for widespread allele-frequency differences. The Fst values of the YRI populations demonstrated higher values (max 0.336) for males compared to females than for Phase 2 compared to Phase 3 (max 0.206). The YRI cell lines were collected from the same collection site (Ibadan, Nigeria) and are part of the same HapMap population (Yoruba), the only difference being what appeared to be a random division into two sets (90 cell lines/set) for the public release. To ensure that our observations were not due to the fact that the two panels were ordered, received, and grown at different times within our laboratory, we reevaluated 27 randomly selected, unrelated cell lines from the YRI population (14 YRI Phase 2 and 13 YRI Phase 3). In this subset, we used a one-tailed t test because we had a directional hypothesis. The YRI Phase 3 cell lines grew slower (p = 0.0599), suggesting that the difference in growth was not due to differences in the conditions and timing at which the cell lines were purchased and maintained (Figure 3). We investigated whether this difference may be related to the time required to establish an immortalized cell line after EBV transformation. We compared the time in days for the YRI Phase 2 and YRI Phase 3 cell lines to be frozen down (signifying successful immortalization and cell growth). Phase 3 cells required significantly more time compared to Phase 2 cells (p = 0.0002; with a significant difference, p = 0.0242, remaining even after removal of the ten outlier cell lines), as seen in Figure 4. The original cell lines transformed and then maintained at Coriell were selected for release with the first panel of HapMap primarily on the basis of how quickly they reached adequate cell density for harvesting of DNA. Therefore, cell lines that were growing more slowly were likely to become part of the second release of cell lines (Phase 3). Although all cell lines grew adequately to allow for eventual DNA harvest of comparable quality, trios with the least number of slow-growing cell lines were more likely to be included in Phase 2. Therefore, the difference in the cellular growth rate for the YRI panels is largely attributed to the selection and division at Coriell. The division of the population into two panels with respect to cellular-growth-rate variation has important ramifications for cellular-phenotype studies, particularly for studies that are designed to use one population as discovery and another as replication. Cellular growth rate has been shown to affect some pharmacological phenotypes, including drug sensitivity.9Huang R.S. Kistner E.O. Bleibel W.K. Shukla S.J. Dolan M.E. Effect of population and gender on chemotherapeutic agent-induced cytotoxicity.Mol. Cancer Ther. 2007; 6: 31-36Crossref PubMed Scopus (66) Google Scholar, 18Stark A.L. Zhang W. Mi S. Duan S. O'Donnell P.H. Huang R.S. Dolan M.E. Heritable and non-genetic factors as variables of pharmacologic phenotypes in lymphoblastoid cell lines.Pharmacogenomics J. 2010; PubMed Google Scholar Studies using non-HapMap, patient-derived LCLs have recognized the importance of looking at the cellular growth rate in pharmacological studies.22Morag A. Kirchheiner J. Rehavi M. Gurwitz D. Human lymphoblastoid cell line panels: novel tools for assessing shared drug pathways.Pharmacogenomics. 2010; 11: 327-340Crossref PubMed Scopus (39) Google Scholar For example, using the alamarBlue assay of cell-growth inhibition described previously,9Huang R.S. Kistner E.O. Bleibel W.K. Shukla S.J. Dolan M.E. Effect of population and gender on chemotherapeutic agent-induced cytotoxicity.Mol. Cancer Ther. 2007; 6: 31-36Crossref PubMed Scopus (66) Google Scholar, 18Stark A.L. Zhang W. Mi S. Duan S. O'Donnell P.H. Huang R.S. Dolan M.E. Heritable and non-genetic factors as variables of pharmacologic phenotypes in lymphoblastoid cell lines.Pharmacogenomics J. 2010; PubMed Google Scholar we found that the YRI Phase 2 population was significantly more sensitive to three chemotherapeutic agents—5′deoxyfluorouridine (5′-DFUR), pemetrexed, and carboplatin—compared to the YRI Phase 3 panel, with p ≤ 0.0001 (Figure 5). The slower-growing HapMap Phase 3 cell lines were more resistant to the drugs, consistent with previous results demonstrating an inverse relationship between slower-growing cells and drug sensitivity.18Stark A.L. Zhang W. Mi S. Duan S. O'Donnell P.H. Huang R.S. Dolan M.E. Heritable and non-genetic factors as variables of pharmacologic phenotypes in lymphoblastoid cell lines.Pharmacogenomics J. 2010; PubMed Google Scholar To address this problem when performing genotype-phenotype relationships in this population, the two YRI panels are best analyzed as a single cohort or, alternatively, randomly divided into two sets that are not significantly different in proliferation rate, thus allowing for a discovery and a replication population. There are ramifications beyond pharmacology for genotype-phenotype studies using panels of cell lines that grow at different rates, including gene expression. Many different studies of gene expression have utilized the HapMap cell lines.3Morley M. Molony C.M. Weber T.M. Devlin J.L. Ewens K.G. Spielman R.S. Cheung V.G. Genetic analysis of genome-wide variation in human gene expression.Nature. 2004; 430: 743-747Crossref PubMed Scopus (950) Google Scholar, 4Dixon A.L. Liang L. Moffatt M.F. Chen W. Heath S. Wong K.C. Taylor J. Burnett E. Gut I. Farrall M. et al.A genome-wide association study of global gene expression.Nat. Genet. 2007; 39: 1202-1207Crossref PubMed Scopus (790) Google Scholar, 5Spielman R.S. Bastone L.A. Burdick J.T. Morley M. Ewens W.J. Cheung V.G. Common genetic variants account for differences in gene expression among ethnic groups.Nat. Genet. 2007; 39: 226-231Crossref PubMed Scopus (395) Google Scholar, 6Storey J.D. Madeoy J. Strout J.L. Wurfel M. Ronald J. Akey J.M. Gene-expression variation within and among human populations.Am. J. Hum. Genet. 2007; 80: 502-509Abstract Full Text Full Text PDF PubMed Scopus (225) Google Scholar, 7Stranger B.E. Nica A.C. Forrest M.S. Dimas A. Bird C.P. Beazley C. Ingle C.E. Dunning M. Flicek P. Koller D. et al.Population genomics of human gene expression.Nat. Genet. 2007; 39: 1217-1224Crossref PubMed Scopus (907) Google Scholar, 8Zhang W. Duan S. Kistner E.O. Bleibel W.K. Huang R.S. Clark T.A. Chen T.X. Schweitzer A.C. Blume J.E. Cox N.J. Dolan M.E. Evaluation of genetic variation contributing to differences in gene expression between populations.Am. J. Hum. Genet. 2008; 82: 631-640Abstract Full Text Full Text PDF PubMed Scopus (158) Google Scholar Baseline gene expression for the Phase 2 CEU and YRI cell lines has been generated with the use of the Affymetrix GeneChip Human Exon 1.0 ST Array.8Zhang W. Duan S. Kistner E.O. Bleibel W.K. Huang R.S. Clark T.A. Chen T.X. Schweitzer A.C. Blume J.E. Cox N.J. Dolan M.E. Evaluation of genetic variation contributing to differences in gene expression between populations.Am. J. Hum. Genet. 2008; 82: 631-640Abstract Full Text Full Text PDF PubMed Scopus (158) Google Scholar Within the CEU population, we found no gene-expression signatures significantly correlated with growth rate; in contrast, there were 217 transcript clusters that had a Q value less than 0.10 in the YRI panel (Table S1). Of those, 135 genes had differential expression values between the CEU and YRI, with a p value less than 0.05 (Q value less than 0.037).8Zhang W. Duan S. Kistner E.O. Bleibel W.K. Huang R.S. Clark T.A. Chen T.X. Schweitzer A.C. Blume J.E. Cox N.J. Dolan M.E. Evaluation of genetic variation contributing to differences in gene expression between populations.Am. J. Hum. Genet. 2008; 82: 631-640Abstract Full Text Full Text PDF PubMed Scopus (158) Google Scholar The Phase 2 CEU and YRI panels do not grow at different rates; however, this gene list could be further evaluated in the Phase 3 YRI panel. Interpretation of results from gene-expression studies among populations should consider growth rate, particularly for panels that grow at different rates. Cellular growth rate was found to be 30% heritable;18Stark A.L. Zhang W. Mi S. Duan S. O'Donnell P.H. Huang R.S. Dolan M.E. Heritable and non-genetic factors as variables of pharmacologic phenotypes in lymphoblastoid cell lines.Pharmacogenomics J. 2010; PubMed Google Scholar there could be several rare alleles affecting the growth rate that are not randomly distributed between the two panels. These nonrandomly distributed alleles could also affect other phenotypes, not just cellular growth rate. The PCA merely confirms a similarity in overall genetic makeup but does not fully address the potential confounding factor of an artificial selection between the two populations. Furthermore, although estimates of heritability across large pedigrees provide a robust measure, it remains that ∼70% of cellular-growth-rate variability has environmental, and potentially technical causes that are not well understood or easily identifiable. Regardless of their origins, the importance of this finding is that these differences can have important ramifications for phenotypes of interest and must be considered in experimental design and analysis. In summary, we found in vitro cellular-growth-rate differences to be confounding and significant within the fully released HapMap YRI population and between the HapMap Han Chinese populations. Overall, we found important growth-rate differences between many of the HapMap populations. On the basis of the data reported here, we believe it is essential to consider growth rate for any studies that utilize different HapMap populations in genotype-phenotype studies. In particular, HapMap populations should be carefully evaluated when incorporated in genetic studies in which replication and validation panels are needed, or when comparisons within and among ethnic groups will be made. This work was supported by the Pharmacogenetics of Anticancer Agents Research (PAAR) Group, funded by the National Institutes of Health (NIH)/National Institute of General Medical Sciences grant GM61393, NIH/National Cancer Institute Breast SPORE P50 CA125183 and RO1 CA136765. The authors would also like to thank the PAAR cell line core at the University of Chicago for its assistance in ordering and receiving these cell lines. Christine M. Beiswanger is an employee of the Coriell Institute for Medical Research, a nonprofit 501(c)(3) biomedical research foundation. The Coriell Institute for Medical Research is contracted by the National Human Genome Research Institute (NHGRI) to establish, maintain, and distribute the HapMap cell lines to the research community. Fees collected for the distribution of the cell lines are returned to the NHGRI in a cost-reimbursement arrangement. Download .pdf (.5 MB) Help with pdf files Document S1. Four Figures and One Table The URLs for data presented herein are as follows:1000 Genomes Project, http://www.1000genomes.org/Coriell Cell Repositories, http://ccr.coriell.org/HapMap Project, http://www.HapMap.org/The R Project for Statistical Computing, http://www.r-project.org/
Referência(s)