Cytotechnology 39: 69–74, 2002. © 2003 Kluwer Academic Publishers. Printed in the Netherlands.
69
False cell lines: The problem and a solution John R. Masters Prostate Cancer Research Centre, UCL, London, U.K. (E-mail:
[email protected]; Fax: 0207 679 9366) Received 30 June 2002; accepted 13 August 2002
Key words: authentication, cell lines, cross-contamination, DNA profiling, HeLa
Abstract Hundreds of misleading reports are published every year containing data on human cancer cell lines that are derived from some other species, tissue or individual to that claimed. In consequence, millions of dollars provided for cancer research are being spent on the production of misleading data. This review describes how cross-contamination occurs, catalogues the use of false cell lines in leading biomedical journals, and suggests ways to resolve the problem.
History of cross-contamination It is probable that cross-contamination is almost as old as tissue culture. When animal and human cell lines were first established in the 1940s and 1950s, there were no definitive methods available to distinguish between cells derived from different species, let alone the same species. However, by the late 1950s there were problems and suspicions over the reproducibility of data, and by the 1960s there was abundant evidence of cross-contamination between species. But cross-contamination between species is only the tip of the iceberg – most of the problems are due to mixups between cells of the same species. It was not until 1967 that the first reliable method to distinguish between cells derived from different individuals of the same species was described by Stan Gartler (1967, 1968). The most infamous examples of cross-contamination relate to HeLa cells (Gold, 1986; Masters, 2002), the first human cancer cell line to be established in continuous culture (Gey et al., 1952). HeLa was derived from a rare glandular type of cervical cancer and continues to be one of the most widely used resources for biological studies needing human cells. In 1962, the ATCC was established to bring order into the world of tissue culture and provide a reliable source of authenticated cell lines. However, in 1967, Stan Gartler
showed that all the supposedly authentic human cell lines derived from different individuals and collected by the ATCC were in fact HeLa cells (Gartler, 1967, 1968). Stan Gartler’s findings have been confirmed and extended in many studies during the intervening 35 years, the best known of which are those of NelsonRees and colleagues (1974, 1976, 1981, 2001). The unequivocal evidence has been ignored by many scientists, and the HeLa cell derivatives described by Stan Gartler are still being sold by cell banks under their false names and misrepresented in high impact journals. The most exasperating aspect of this saga of scientific deception is that many journal editors have not confronted the problem and continue to publish misleading data.
The extent of the problem The DSMZ (German Collection of Microorganisms and Cell Cultures) has estimated the proportion of human cancer cell lines that are cross-contaminated. On the basis of the cell lines submitted to the DSMZ, nearly 20% are known cross-contaminants (MacLeod et al., 1999). There are strong reasons to believe that a figure of 20% underestimates the problem as, for the vast majority of cell lines, no evidence has ever
70 been provided that the cells were derived from the individual claimed. Consequently, it is probable that at least 20% of publications using one or more human cancer cell lines contains data that is misleading. Some of the most notorious cross-contaminants of human solid cancer cell lines are listed in Table 1, and a selection of the false human hematopoietic cell lines was described by Drexler et al. (1999).
How does cross-contamination happen? There are many ways in which two cell lines can get mixed up due to poor tissue culture practice combined with, in many laboratories, inadequate quality control. If the same labware (e.g., glass or plastic pipette) is accidentally re-used with a different cell line, there is a high probability that a few cells from the first cell line will be mixed with the second cell line. Media and other reagents shared among cell lines can easily lead to transfer of cells. It takes only one cell from an aggressive cell line such as HeLa to rapidly overtake a complete culture, so that within a few passages the slower growing cells have completely disappeared. Unless it is absolutely essential, two cell lines should never be in the same flow cabinet at the same time. Cells can survive in an aerosol in a flow cabinet and can contaminate another cell line when the culture vessel is opened. It is probable that many examples of cross-contamination occur as a result of clerical error or misunderstandings within a laboratory. A common cause of cross-contamination is poor labeling of frozen stocks. Confusion has been generated by the arbitrary changing of the names of cell lines. For example, the name of the human bladder cancer cell line EJ (the initials of the donor) was changed to MGH-U1 (Massachusetts General Hospital-Urology1). However, the EJ cell line had already been contaminated by another human bladder cancer cell line, T24. The same group later described another cell line that they had cross-contaminated with T24 cells, called MGH-U2. Most of the publications on false cell lines are not directly due to cross-contamination. The misleading data stems from a naive belief that because the cells were claimed to be from a particular tissue by a scientist or cell bank, there is no need to check the authenticity or origin of the cells. Most cell lines stored in cell banks are catalogued according to the claim made by the scientist placing the cells in the bank. Some cell line banks continue to sell cell lines
known to be cross-contaminants, albeit with an indication that these cell lines have the characteristics of HeLa or some other cell line. Much of the misleading data that is published is due to the bogus description of such cell lines under the false name and origin. Repeated requests have been made to cell banks to modify the names of the cross-contaminated cell lines to prevent unnecessary confusion (e.g., Stacey et al., 2000). Under such renaming cell lines such as KB would be known as HeLa-KB, i.e., a subline of HeLa. There may be a fear on the part of cell banks that this measure will reduce the number of samples that are sold.
Misconceptions about cross-contamination Some scientists believe that two cell lines can live in co-existence. This possibility is exceedingly remote when the two cell lines are in competition for the same space, such as the surface of a tissue culture vessel, and have different cell proliferation kinetics. Studies of population biology demonstrate that when two populations compete for the same niche, even a tiny survival advantage will result in the rapid extinction of the weaker population, and this is precisely what happens following cross-contamination. A cell line with a short doubling time will soon overgrow one with a longer doubling time. It may be possible that two cell lines could co-exist if the weaker cell line produced a factor essential for the survival of the fitter cell line, but I am unaware of any such example. In suspension cultures, there is less competition for space, and it is conceivable that two cell lines might co-exist for a longer period. The eventual outcome will be the same, with the faster growing cells displacing the slower growing cells, however small the growth advantage, as has been shown by the selection for a non-producer clone in hybridoma culture in a continuous reactor culture (Merten, 1989). There is consequently a misconception by some scientists (e.g., Li et al., 2000) that some crosscontaminated cell lines contain a mixture of two cell types. Except at the time of the original crosscontamination, this idea is incorrect, and the two cell lines were only together in the same culture vessel for a matter of days. Indeed, in many examples of crosscontamination, it is probable that the two cell lines were never in contact because the cross-contamination was due to mislabelling or some similar mix-up.
71 Table 1. Some of the more widely used cross-contaminants of HeLa and T24 cells that are used frequently under false descriptions Cell line origin
False description
References
HeLa, glandular cancer of the cervix
KB, oral cancer HEp-2 (H.Ep.-2), laryngeal cancer WISH, normal amnion Girardi heart, normal heart Chang liver, normal liver Intestine 407, embryonic intestinal cells
Gartler, 1967, 1968 Nelson-Rees et al., 1981 plus many others
T24, bladder cancer cells
EJ (also bladder cancer) MGH-U1 (EJ under another name) MGH-U2 (also bladder cancer) J82 (also bladder cancer). The original stocks are an authentic unique cell line HU456 (also bladder cancer) HU961T (spontaneously immortalized bladder cancer cells) ECV304 (spontaneously immortalized normal endothelial cells) TSU-Pr1 (prostate cancer cells) JCA-1 (prostate cancer cells)
O’Toole et al., 1983
Another misconception that has arisen is that the cross-contaminated cell line has somehow acquired one or more characteristics of the original cell line. HeLa cells are variously described as having the characteristics of skin or head and neck cancers (KB cells), normal epithelial cells (Hep-2), normal amnion cells (WISH, AV3 and FL), embryonic lung cells (L132), normal intestinal cells (Intestine 407) or normal liver cells (Chang liver). These are just a few of the many counterfeit descriptions of the cervical cancer cell line HeLa. So can HeLa cells express the characteristics of a variety of other cell types, and if so, how were these acquired? One possibility is that the two cell lines fused by somatic cell hybridisation. However, the evidence from genetic studies indicates that the cross-contaminants are pure HeLa cells and have not acquired any additional genetic information. How then, on numerous occasions, has each of these crosscontaminated cell lines gained the ability to express the specific characteristics of some other cell type, and is it just a coincidence that these are the characteristics of the tissue it was falsely claimed to be? Cell sublines such as these cross-contaminants can show minor phenotypic changes as a result of modified culture conditions. It is probable that objective examination
Dirks et al., 1999
Bokhoven et al., 2001
would reveal that the cross-contaminated cells have a similar phenotype and genotype to the cells from which they were derived.
Does peer review work? In a paper in Nature Medicine, ECV304 cells (originally thought to be normal endothelial cells) were mistakenly described to be of ‘mixed origin’ (Li et al., 2000), despite the fact that ECV304 cells are the human bladder cancer cell line T24 (Dirks et al., 1999). The authors, in common with most scientists using cross-contaminated cell lines, failed to indicate that ECV304 cells are not normal endothelial cells, but are bladder cancer cells. The failure to indicate the true origin of a resource is deliberately misleading. In cases where the cells are known to be cross-contaminants, the responsibility for the publication of false data lies with the journal editor. Many of the papers published in the leading journal Cancer Research use human cancer cell lines. Misrepresentation of cell lines such as KB cells is widespread, with claims that they are derived from a variety of sources. In recent publications, KB cells were claimed to be epidermoid carcinoma (Cabot et al.,
72 1999; Komatsu et al., 2000; Park et al., 2002) or orolaryngeal epidermoid carcinoma (Goan et al., 1999), despite the fact that these are HeLa cells, derived from a glandular cancer of the cervix. Whatever the application of these cells, it should be acknowledged that KB cells are in fact HeLa cells. Another leading journal that publishes many papers using human cancer cell lines is the Journal of Biological Chemistry. In recent years, the journal has included 4 papers using HEp-2 cells as a model of human epithelium (Tabuchi et al., 2000; Bayer et al., 2001; Gorman et al., 2001; Sinclair et al., 2002). It may be appropriate under certain defined circumstances to use a human glandular cancer of the cervix as a model of normal human epithelial cells, but it is never appropriate to deceive the reader by failing to indicate that HEp-2 cells are in fact HeLa cells masquerading under another name. Since 1999, about 300 publications have been indexed in Medline under a search for HEp-2 cells and a similar number for KB cells. Few of these publications acknowledge the fact that these are crosscontaminated cell lines disseminated under a false name. A Medline search suggested that KB cells have been used in 15 publications in Cancer Research, 9 in the Journal of Biological Chemistry, 5 in the Proceedings of the National Academy of Sciences of the U.S.A. and 1 in the Journal of Cell Biology. A search of the Nature website indicated that Nature has published 4 papers since 1997 using KB cells. One of these papers also used HEp-2 cells and implied that these differed from KB cells (Flatau et al., 1997). Another Nature paper used both KB and HeLa cells, but implied that they were different cell types (Yang et al., 1998). Journals specialising in cell biology might be expected to be aware of the origins of the cells described in their publications. A search of the Cell website, restricted to titles and abstracts, found 10 publications using KB cells and 2 using HEp-2. A search of the Journal of Cell Biology website, again restricted to titles and abstracts, found 13 publications using KB cells and 5 using HEp-2 cells. There is at least one example of a journal knowingly publishing false information. In September 2001, Cancer Research published a paper showing that the cell lines TSU-Pr1 and JCA-1 are the bladder cancer cell line T24, and not prostate cancer cell lines, as originally claimed (Van Bokhoven et al., 2001). However, Cancer Research then went on to publish three further publications later in 2001 that claimed
that TSU-Pr1 cells are of prostate origin (Chakraborty et al., 2001; Moffatt et al., 2001; Zha et al., 2001). Some journal editors use peer review as a shield to deflect their responsibility for the dissemination of false data. It is clear that peer review is failing to detect many examples of blatantly false data using human cancer cell lines. Authors and reviewers appear to be ignorant of the problem. As the use of crosscontaminated HeLa cell lines has been known since 1967 and the culprits have been named, it would be timely for journal editors to be aware of the literature on this subject. An understanding that many of the most frequently used names given to cells lines (such as KB, HEp-2 and ECV304) are false and misleading could alert journal editors and reviewers to the extent of the problem.
How is cross-contamination detected? Cross-contamination can happen in any laboratory, it takes only a momentary loss of concentration. Scientists with any care or understanding for their experiments check their cells on a daily basis and are familiar with their behaviour and appearance. This interaction with the cells being studied is of great benefit to any experiment and will prevent many problems, but it will not exclude cross-contamination, particularly where the cells are handled by a number of different people, some of whom may be less familiar with the cells. Many cell types look and behave in a similar manner in culture, but there are definitive tests that can be used to demonstrate the authenticity of a cell line, including HLA and isozyme typing, DNA fingerprinting and DNA profiling. DNA profiling is a means of quality control available to any laboratory, because it is cheap, simple and commercially available. The cost of checking the DNA profile of cells commercially is less than $200, a tiny fraction of the expense incurred for even the simplest experiment. DNA profiling is cheap and commercially available because it is the test used by the forensic service to identify crime suspects and resolve issues over parentage. A series of 7–10 polymorphic loci are analysed, providing an indication of the size of the two alleles at each locus, accurate to less than one base. This process yields a series of numbers that provide the equivalent of a bar code or credit card number for the cells (Masters, 2001). This series of numbers can be checked against those of other cell lines to exclude cross-contamination.
73 It has been claimed that DNA profiling could provide an internationally agreed laboratory standard for human cell lines (Masters et al., 2001; O’Brien 2001). The perfect scenario is when DNA from the tissue or donor of origin is available in addition to the cell line, because the cell line should give a DNA profile that is compatible with that of the donor, providing an almost absolute guarantee of authenticity. STR profiling is available for species in which parentage is an issue, including humans, dogs, horses and cattle. A similar test might be developed for mice if required.
The future If journal editors were aware of cross-contamination and were intolerant of the publication of cell lines under false descriptions, much of the problem would disappear. However, there are many unknown cross-contaminations. DNA profiling provides a simple, cheap method that would identify the majority of these in human cells, and can be extended to other species if needed. Scientists need to introduce this quality control measure as a part of routine practice and journal editors and referees should demand that such practices are an essential pre-requisite for publication. Complacency is perhaps the greatest problem. A past editor of Nature attacked scientists who raised the issue of cross-contamination, describing them as self-appointed vigilantes and said that it would be tragic if they corrupted the civilised habits of scientists (Nature, 1981). With attitudes such as these controlling the peer review system, scientific fraud as described in this review has increased and may continue unchecked.
References Bayer N, Schober D, Hutinger M, Blaas D and Fuchs R (2001) Inhibition of clathrin-dependent endocytosis has multiple effects on human rhinovirus serotype 2 cell entry. J Biol Chem 276: 3952–3962. Van Bokhoven A, Varella-Garcia M, Korch C and Miller GJ (2001) TSU-Pr1 and JCA-1 cells are derivatives of T24 bladder carcinoma cells and are not of prostatic origin. Cancer Res 61: 6340–6344. Cabot MC, Giuliano AE, Han TY and Liu YY (1999) SDZ PSC 833, the cyclosporine A analogue and multidrug resistance modulator, activates ceramide synthesis and increases vinblastine sensitivity in drug-sensitive and drug-resistant cancer cells. Cancer Res 59: 880–885.
Chakraborty M, Qiu SG, Vasudevan KM and Rangnekar VM (2001) Par-4 drives trafficking and activation of Fas and FasL to induce prostate cancer cell apoptosis and tumor regression. Cancer Res 61: 7235–7263. Dirks WG, MacLeod RA and Drexler HG (1999) ECV304 (endothelial) is really T24 (bladder carcinoma): Cell line crosscontamination at source. In Vitro Cell Devl Biol 35: 558–559. Drexler HG, Dirks WG and MacLeod RAF (1999) False human hematopoietic cell lines: cross-contamination and misinterpretations. Leukemia 13: 1601–1607. Editorial (1981) Responsibility for trust in research. Nature 289: 211–212. Flatau G, Lemichez E, Gauthier M, Chardin P, Paris S, Fiorentini C and Boquet P (1997) Toxin-induced activation of the G protein p21 Rho by deamidation of glutamine. Nature 387: 729–733. Gartler SM (1967) Genetic markers as tracers in cell culture. Natl Cancer Inst Monogr 26: 167–195. Gartler SM (1968) Apparent HeLa cell contamination of human heteroploid cell lines. Nature 217: 750–751. Gayraud B, Keene DR, Sakai LY and Ramirez F (2000) New insights into the assembly of extracellular microfibrils from the analysis of the fibrillin 1 mutation in the tight skin mouse. J Cell Biol 150: 667–680. Gey GO, Coffman WD and Kubicek MT (1952) Tissue culture studies of the proliferative capacity of cervical carcinoma and normal epithelium. Cancer Res 12: 264–265. Goan YG, Zhou B, Hu E, Mi S and Yen Y (1999) Overexpression of ribonucleotide reductase as a mechanism of resistance to 2,2difluorodeoxycytidine in the human KB cancer cell line. Cancer Res 59: 4204–4207. Gold M (1986) A conspiracy of cells. One woman’s immortal legacy and the scandal it caused. State University of New York Press. Gorman JJ, McKimm-Breschkin JL, Norton RS and Barnham KJ (2001) Antiviral activity and structural characteristics of the nonglycosylated central subdomain of human respiratory syncytial virus attachment (G) glycoprotein. J Biol Chem 276: 38988–39994. Komatsu M, Sumizawa T, Mutoh M, Chen ZS, Terada K, Furukawa T, Yang XL, Gao H, Miura N, Sugiyama T and Akiyama S (2000) Copper-transporting P-type adenosine triphosphatase (ATP7B) is associated with cisplatin resistance. Cancer Res 60: 1312–1316. Li J, Post M, Volk R, Gao Y, Li M, Metais C, Sato K, Tsai J, Aird W, Rosenberg RD, Hampton TG, Li J, Sellke F, Carmeliet P and Simons M (2000) PR39, a peptide regulator of angiogenesis. Nature Medicine 6: 49–55. MacLeod RAF, Dirks WG, Matsuo Y, Kaufmann M, Milch H and Drexler HG (1999) Widespread intraspecies cross-contamination of human tumour cell lines. Int J Cancer 83: 555–563. Masters JR (2000) Human cancer cell lines: Fact and fantasy. Nature Rev Molec Cell Biol 1: 233–236. Masters JRW, Thomson JA, Daly-Burns B, Reid YA, Dirks W, Packer P, Toji LH, Ohno T, Tanabe H, Arlett CF, Kelland LR, Harrison M, Virmani A, Ward TH, Ayres KG and Debenham PG (2001) STR profiling provides an international reference standard for human cell lines. Proc Natl Acad Sci USA 98: 8012–8017. Masters JR (2002) HeLa cells 50 years on: The good, the bad and the ugly. Nature Rev Cancer 2: 315–319. Merten O-W (1989) Culture of Hybridomas – A Survey. In: Miller AOA (ed.) Advanced Research on Animal Cell Technology, NATO ASI Series E. Applied Sciences, Vol. 156, (pp. 367–398) Kluwer Academic Publishers, Dordrecht. Moffatt KA, Johannes WU, Hedlund TE and Miller GJ (2001) Growth inhibitory effects of 1α, 25-Hidroxyvitamin D3 are me-
74 diated by increased levels of p21 in the prostatic carcinoma cell line ALVA-31. Cancer Res 61: 7122–7129. Nelson-Rees WA, Flandermeyer RR and Hawthorne PK (1974) Banded marker chromosomes as indicators of intraspecies cellular contamination. Science 184: 1093. Nelson-Rees WA and Flandermeyer RR (1976) HeLa cultures defined. Science 191: 96–98. Nelson-Rees WA, Daniels DW and Flandermeyer RR (1981) Crosscontamination of cells in culture. Science 212: 446–452. Nelson-Rees WA (2001) Responsibility for truth in research. Phil Trans R Soc Lond B 356: 849–851. O’Brien SJ (2001) Cell culture forensics. Proc Natl Acad Sci USA 98: 7656–7658. O’Toole CM, Povey S, Hepburn P and Franks LM (1983) Identity of some human bladder cancer cell lines. Nature 301: 429–430. Park SY, Lam W and Cheng Y (2002) X-ray repair crosscomplementing gene I protein plays an important role in camptothecin resistance. Cancer Res 62: 459–465.
Sinclair JF and O’Brien AD (2002) Cell surface-localized nucleolin is a eukaryotic receptor for the adhesin intimin-gamma of enterohemorrhagic Escherichia coli O157:H7. J Biol Chem 277: 2867–2885. Stacey GN, Masters JRW, Hay RJ, Drexler HG, MacLeod RAF and Freshney RI (2000) Cell contamination leads to inaccurate data: We must take action now. Nature 403: 356. Tabuchi M, Yoshimori T, Yamaguchi K, Yoshida T and Kishi F (2000) Human NRAMP2/DMT1, which mediates iron transport across endosomal membranes, is localized to late endosomes and lysosomes in HEp-2 cells. J Biol Chem 275: 22220–22228. Yang N, Higuchi O, Ohashi K, Nagata K, Wada A, Kangawa K, Nishida E and Mizuno K (1998) Cofilin phophorylation by LIMkinase 1 and its role in Rac-mediated actin reorganization. Nature 393: 809–812. Zha S, Gage WR, Sauvageot J, Saria EA, Putzi MJ, Ewing CM, Faith DA, Nelson WG, De Marzo AM and Isaacs WB (2001) Cyclooxygenase-2 is up-regulated in proliferative inflammatory atrophy of the prostate, but not in prostate carcinoma. Cancer Res 61: 8617–8623.
