Widespread intraspecies cross-contamination of human tumor cell lines arising at source

Abstract

We present a panoptic survey of cell line cross-contamination (CLCC) among original stocks of human cell lines, investigated using molecular genetic methods. The survey comprised 252 consecutive human cell lines, almost exclusively tumor-derived, submitted by their originators to the DSMZ and 5 additional cell repositories (CRs), using a combination of DNA profiling (4-locus minisatellite and multilocus microsatellite probes) and molecular cytogenetics, exploiting an interactive database (http://www.dsmz.de/). Widespread high levels of cross-contaminants (CCs) were uncovered, affecting 45 cell lines (18%) supplied by 27 of 93 originators (29%). Unlike previous reports, most CCs (42/45) occurred intraspecies, a discrepancy attributable to improved detection of the more insidious intraspecies CCs afforded by molecular methods. The most prolific CCs were classic tumor cell lines, the numbers of CCs they caused being as follows: HeLa (n ‫؍‬ 11), T-24 (n ‫؍‬ 4), SK-HEP-1 (n ‫؍‬ 4), U-937 (n ‫؍‬ 4) and HT-29 (n ‫؍‬ 3). All 5 supposed instances of spontaneous immortalization of normal cells were spurious, due to CLCC, including ECV304, the most cited human endothelial cell line. Although high, our figure for CCs at the source sets a lower limit only as (i) many older tumor cell lines were unavailable for comparison and (ii) circulating cell lines are often obtained indirectly, rather than via originators or CRs. The misidentified cell lines reported here have already been unwittingly used in several hundreds of potentially misleading reports, including use as inappropriate tumor models and subclones masquerading as independent replicates. We believe these findings indicate a grave and chronic problem demanding radical measures, to include extra controls over cell line authentication, provenance and availability. Int.

Karyotypic dissection of Hodgkin's disease cell lines reveals ectopic subtelomeres and ribosomal DNA at sites of multiple jumping translocations and genomic amplification

“…Cross-contamination is a major drawback to working with cell lines, 33 particularly in HD which lacks external identifiers in the form of recurrent translocations. The availability of three sister cell lines (unique among HD models) permits authentication based on convergent karyotypes and DNA profiles.…”

False human hematopoietic cell lines: cross-contaminations and misinterpretations

“…Our cell line bank experience shows that in a recent series from 1990 to 1998 a large percentage of allegedly new human cell lines (cell lines derived from solid tumors and hematopoietic malignancies) which were received from the original investigators (in contrast to cell cultures obtained from secondary sources) were cross-contaminated with other cell lines: 10 • 45/252 (18%) human cell lines were false; • 27/93 (29%) original donors supplied false cell lines.…”

“…Although routinely lower sensitivity can be achieved, an experienced cytogeneticist may detect a contamination of 1%. 9,10,28 However, chromosome analysis is a labor-intensive, time-consuming and rather expensive procedure; interpretation may require a high degree of skill. Still, cytogenetics without question is the method currently offering the highest versatility in the characterization of a cell line (not only specifically for the purpose of identification/ authentication).…”

Cytogenetic harvesting of commonly used tumor cell lines

“…In addition, chromosome rearrangements often indicate genes altered in cancer: Cancer Genome Project data show that 273/360 known cancer genes (76%) participate in chromosome rearrangements (http:// www.sanger.ac.uk/genetics/CGP/Census/chromosome.shtml) and are thus potentially ascertainable using cytogenetic methods. Unsurprisingly, cytogenetic data afford a uniquely informative global perspective on individual cancer cells and, by reference to previously published karyotypes, assist in detecting cross-contamination-a chronic unresolved problem 3 .…”

“…With G-banding alone, it is possible to detect, if not fully characterize, most chromosome rearrangements present in cell lines carrying up to approximately half a dozen separate changes; i.e., it is sufficient for confirming cell line identity. The utility of Gbanding in detecting cell line cross-contamination is illustrated in Figure 3, which shows CCRF-CEM derived from pediatric T-acute lymphoblastic leukemia (T-ALL) and its tetraploid derivate COLE, mistakenly published as a Hodgkin lymphoma cell line 3,14 .…”