Forty-seven years have passed since Stanley Gartler dropped a bombshell at the Second Decennial Review Conference on Cell, Tissue and Organ Culture (1). Reporting on the genetic properties of cultured cells before this prestigious international assembly, he announced that examination of 18 cell lines from independent sources revealed that they were identical to each other and to a celebrated and important cancer cell line known as HeLa . HeLa was established in culture in 1951 from a cancer biopsy removed from the cervix of an Afro-American woman named Henrietta Lacks (2). The celebrity and importance of HeLa derived from the fact that it was the first human cancer cell to be established in culture, thereby enriching the growing collection of cells available for biomedical research; also, it eventually became easy to culture and proved to be useful as a human cell model for cell physiology, morphology, and biochemistry studies. Hence, it is not surprising that the HeLa cell line found its way into laboratories throughout the world where experts and novices greeted the cells. No doubt, many recipients of the cell line had limited experience with the principles of cell culture quality control and with nuances of behavior and morphological and physiological characteristics of HeLa.
The major diagnostic tool used by Gartler to characterize and identify the 18 cell lines was based on the electrophoretic mobility of an isoenzyme, glucose-6-phosphate dehydrogenase (G6PD). Each of the 18 cell lines showed the A Type G6PD phenotype. The probability that all of the cell lines had a common origin in the distant past is untenable. Type A G6PD is found only in persons of African descent , and only at a frequency of 30%. Because the 18 cell cultures were from independent sources, Gartler concluded that each of the cultures became independently contaminated by HeLa cells, and the HeLa cells outgrew the original cells.
Among the original 18 cell cultures (before contamination) were 5 potentially useful human cell models: KB (an oropharyngeal cancer line), Hep-2 (a human laryngeal cancer cell line), Chang liver (a normal human liver cell line), Int-407 (a human intestinal epithelial cell line) and WISH (a human amnion-derived cell line). From 1966 to the present time, these cell lines, although demonstrated repeatedly to be HeLa, frequently surface as “cell models” of the tissue of origin. Quantitative information about such misleading publications will be described later. Publications of the 1970’s and 1980’s served to make some members of the scientific community aware that misidentification and cross-contamination were common and included inter-species as well as intra-species mix-ups The most significant contributions were made by Walter Nelson-Rees and his associates who maintained a repository under contract to the National Cancer Institute (3,4,5,6). Upon receiving cell lines for deposit, the team (including expert off-site collaborators) would perform cell authentication tests, such as isoenzyme analysis, karyotyping, and marker identification. Simple DNA methods were available but were used infrequently, probably because they were not sufficiently robust.. The paper by Nelson-Rees, Daniels, and Flandermeyer, published in Science in 1981, is a tour de force that provides the flavor of the extent of cross-contamination and misidentification. It also shows that cross-contamination did not spare many reputable laboratories (5). The many examples included a human breast cancer line that proved to be rat, and another that proved to be HeLa; a purported chicken line that proved to be rat, a mongoose cell line that proved to be human, a hamster cell line that proved to be marmoset, a purported embryonic rabbit cell line that proved to be African Green Monkey, and six cultures of independent origin from Russia, that proved to be HeLa. This is a small sample of a very large problem.
These reports spawned controversy and heated discussions leading to a contentious period, polarized camps and a climate that prevented the development of a profession-wide consensus regarding remedial action.
Although the data presented in the reports cited above, as well as many others, point to extensive cross-contamination and misidentification, they are limited with respect to two important questions: on a global level, how many cell lines are not what they are purported to be and what is the incidence of corruption of the cell culture-based biomedical literature? Have we uncovered the iceberg or only its tip? While we have only a limited amount of quantitative data regarding these questions, a cautious extrapolation of the data strongly suggests that only the tip has been uncovered. Major repositories possess thousands of cell lines and private collections many thousands of others. For example, there are about 10,000 cancer cell lines in existence, most in private holdings and not ordinarily available for release or other purposes. The American Type Culture Collection houses more than 3500 cell lines. The German repository (DSMZ) and the European Collection hold another few thousand cell lines. The cells in these holdings have been authenticated. However, there have been many misidentified/cross-contaminated cell lines submitted for deposit to the repositories and were rejected. The number of rejected cell lines rarely makes it into any quantification of the incidence of misidentification/cross-contamination, surfacing only during some informal discussions or until a public announcement of the error is made. Nevertheless, the DSMZ has published some quantitative data that can serve as a guide. In 1999, MacLeod et al. reported that 18% of the purportedly “new” cell lines submitted were “false” and 29% of all originators depositing cell lines had submitted at least one “false” culture (7). They also reported the results of another survey involving hematopoietic cell lines: 45 of 326 submissions (17.9%) were cross-contaminated including 42 intraspecific contaminants (8). A very important paper by Masters et al.(9 ), based on a cooperative study involving the ATCC, the DSMZ and JCRB (Japanese Cell Research Bank), estimated that cross-contamination based on DNA profiling was at a level of 36%. Such data encourages the opinion that the overall rate of cross-contamination and misidentification is probably greater than 20%. The very nature of many genetic studies with cultured cells require observation of the expression of specific, and sometimes unique genes. Hence, such researchers would quickly pick up a cross-contaminant or misidentified cell line resulting in a reduced frequency of “false” cell lines. Conversely, cancer cell lines present several unique challenges. These include the widespread use of cultured cells yet to be fully characterized and the risk of introducing host cells when nude mice are used as xenograft hosts(10). Anecdotal information by investigators close to the cancer research scene suggests that this field is beset with an above average frequency of “false” cell lines.
Corruption of the Scientific Literature
The extent of corruption of the cell culture literature can be estimated by inference from the data presented above as well as various surveys based on narrowly focused analyses. Illustrative of the latter are the following. Masters (11) conducted a Medline search for the years 2000-2004 with regard to the continued use of cell lines known for decades to be HeLa rather than what they were purported to be. His research uncovered 19 citations for the putative intestinal cell line, Int 407, 45 citations for WISH, the putative amnion cell line, 59 citations for the Chang liver cell line, 470 citations for Hep-2, the putative laryngeal cancer cell line, and 556 citations for KB, the putative oropharyngeal cancer cell line. Buehring et al.(12) uncovered 220 false papers as a result of an extensive PubMed search based on the use of well publicized cross-contaminated cultures. Their data did unearth a troublesome trend. The number of cell culture papers in the database increased 2-2.5-fold from 1969 through 2004 while the number of papers based on the use of cross-contaminated cultures increased by about 10-fold. A similar report of a putative endothelial model cell line, ECV-304, in reality was derived from the classic bladder epithelial cancer cell line, T-24 (7). Publicity did not discourage use of ECV-304 as its designation though some began to refer to it as an epithelial cell. In a draft of a letter to the editors of the International Journal of Cancer, MacLeod presents data showing that ECV-304 citations between 1999 and 2006 did not decline but may have increased during some years. A survey prepared by Buehring et al.and returned by 483 respondents can help give us some insight to the origin of the problem we face. Nine percent of those surveyed were unwittingly using HeLa contaminants, only 33% tested for authenticity, and 35% obtained their cells from other labs rather than from major repositories that authenticate the cells they distribute.
What can be concluded about the level of corruption of the cell culture-based literature? It is extensive and not abating! The frequency of use of misidentified cell lines probably is a good indication of the number of misleading reports in the literature. It appears that improvement of the picture will have to await a significant reduction in the use of false cell lines. This goal can be achieved through sustained, profession-wide education initiatives with compliance guaranteed by adoption of a zero tolerance policy that advocates no authentication /no grant and no authentication/no publication outlet (13).Progress to date regarding policy changes by journals may be summarized as follows. Cell, Biochemistry and Biophysics requires authentication while the journal BioTechniques, as of August 2008, requires that the author submit a statement indicating whether or not the cell lines have been authenticated.
- Gartler, Stanley . “ Genetic Markers as Tracers in Cell Culture.” Nat. Cancer Inst. Monograph 265:167-195. 1966.
- Gey, G.O, Coffman,W., and Kubicek, M. “Tissue Culture Studies of the Proliferative Capacity of Cervical Carcinoma and Normal Epithelium.” Cancer Res.12:264-265. 1952.
- Nelson-Rees, W., Flandermeyer, R., and Hawthorne, P. “Banded Marker Chromosomes as Indicators of Intraspecies Cellular Contamination.” Science 184:1093-1096. 1974.
- Nelson-Rees, W., and Flandermeyer, R. “Inter- and Intra-Species Contamination of Human Breast Tumor Cell Lines HBC and BrCa5 and Other Cell Cultures.” Science 195:1343-1344. 1977.
- Nelson-Rees, W., Daniels, D.W. and Flandermeyer, R. “Cross-Contamination of Cells in Culture.” Science 212:446-452. Review. 1981.
- Nelson-Rees, W. “Acceptance Speech. Lifetime Achievement Award. Annual Meeting, Soc. for inVitro Biology (personal communication). 2004.
- Macleod,R., Dirks,W., Matsuo, Y., Kaufmann, M., Milch, H, and Drexler, H. “Widespread Intraspecies Cross-Contamination of Human Tumor Cell Lines Arising at Source.” Int. J. Cancer 83:555-563. 1999.
- Drexler,H.,Dirks, W.,and MacLeod, R. “False Human Hematopoietic Cell Lines: Cross-Contaminations and Misrepresentations.” Leukemia 13:1601-1607. 1999.
- Masters,J.,Thomson,J., Daly-Burns B., Reid, Y., Dirks, W., Packer, P., Toji, L.,Ohno, T., Tanabe, H., Arlett, C., Kelland, L., Harrison, M., Vermani, A., Ward, T., Ayres,K., and Debenham, P. “Short Tandem Repeat Profiling Provides an International Reference Standard for Human Cell Lines.” Proc. Natl. Acad. Sci. USA, 98: 8012-8017. 2001.
- Pathak,S., Nemeth, M., and Multani, A. “Human Tumor Xenografts in Nude Mice Are Not Always of Human Origin: A Warning Signal.” Cancer, 83: 1891-1893. 1998.
- Masters, J. Report presented at the annual meeting of the Society for inVitro Biology. 2005.
- Buehring, G., Eby, A., and Eby, M. “Cell Line Cross-Contamination: How Aware Are Mammalian Cell Culturists of the Problem and How to Monitor It.” In Vitro Cell Dev. Biol. Anim. 40:211-215. 2004.
- Nardone, R. “Eradication of Cross-Contaminated Cell Lines: A Call for Action.” Cell Biology and Toxicology 23: 367-372. 2007.
- O’Brien, S., Kleiner, G., Olson, R., and Shannon, J. “Enzyme Polymorphisms as Genetic Signatures in Human Cell Cultures.” Science 195: 1345-1348. 1977.