| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
FJ
EXPRESS SUMMARY ARTICLE The Full-length version of this article is also available, published online January 6, 2005 as doi:10.1096/fj.04-3062fje. |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,¶,¶,1
* Institute of Legal Medicine,
Tyrolean Cancer Research Institute,
¶ Division of Molecular Pathophysiology, Department Biocenter, Medical University of Innsbruck, Innsbruck, Austria
1Correspondence: Tyrolean Cancer Research Institute, Innrain 66, A-6020 Innsbruck, Austria. E-mail: Reinhard.Kofler{at}uibk.ac.at
SPECIFIC AIMS
Cancer cell lines are used worldwide in biologic research, and DNA fingerprinting is the method of choice for proper cell line authentication. The stability of fingerprinting profiles of such cell lines during common experimental manipulations, however, remains ill defined. We therefore systematically determined short tandem repeat profiles during extended tissue culture, subcloning, and generation of drug-resistant subclones in four widely used leukemic cell lines.
PRINCICPAL FINDING
1. DNA fingerprinting profiles during long-term culture
To determine the effect of long-term culture on genetic stability of continuously growing cell lines, we maintained K562 (chronic myeloid leukemia), U937 (monocytic leukemia), Jurkat/FADD–/– (an anti-fas resistant derivative of Jurkat T cell acute lymphoblastic leukemia [T-ALL]), and CEM-C7H2 (a subclone of the T-ALL cell lines CCRF-CEM) in tissue culture for 1 year and determined their DNA fingerprints using probes for the gender-specific amelogenin and 15 short tandem repeats (STRs) at onset of the culture and at monthly intervals thereafter. For control purposes, one cell line (CEM-C7H2) was cultured and assayed in parallel and the two cultures were referred to as CEM-1 and -2. As shown in Fig. 1
, remarkable differences were observed: K562 and U937 revealed no changes in any of the 16 markers investigated throughout the 1-year observation period. In contrast, Jurkat-FADD–/– and both CEM-C7H2 cultures showed an increasing number of alterations in their fingerprinting profiles over time. Detailed analysis of these alterations revealed that the first changes in Jurkats (i.e., occurrence of new, or loss of existing, alleles) occurred by 3 months in culture, and amounted to a total of 11 changes over the 1-year observation period (Fig. 1
). In the two CEM-C7H2 cell lines cultured in parallel, changes in the profiles started 4-5 months later, and their number was lower in CEM-2 (6 changes) and higher in CEM-1 (20 changes). Since some of these changes were transient in Jurkats and CEM-1 (i.e., new markers appeared and subsequently disappeared), the cumulative profile after 1 year differed from that of the starting culture by fewer changes: a total of 5, 7, and 6 changes in Jurkat, CEM-1, and CEM-2, respectively, affecting 5 loci in each cell line. The above data revealed that cancer cell lines may differ dramatically in their genetic stability and that the DNA fingerprint of a given cell line may change considerably over time, and hence is not necessarily a stable identifier of a particular cell line.
|
To better understand the inception of these changes, we scrutinized fingerprinting patterns of the affected alleles in more detail. In some instances, a particular allele of the affected locus decreased over time and another allele of the same locus increased, thereby replacing each other (exemplified in Fig. 2
A). This suggested that a cell had acquired the respective new allele by mutation and outgrew the original clone it derived from. In other instances, an allele completely disappeared without being replaced (Fig. 2B
), which might be explained by loss of the clonotype(s) carrying this allele. In yet other instances, new alleles transiently appeared at low levels and disappeared again (Fig. 2C
), perhaps reflecting fluctuation in the relative abundance of the corresponding clonotype(s). Taken together, these observations suggested the presence of multiple clonotypes with distinct fingerprinting profiles and fluctuating abundance within the bulk culture.
|
2. Clonotypic heterogeneity in leukemia cell lines
To address whether cell lines investigated were indeed composed of distinct clonotypes and, if so, to what extent, we subcloned them at the end of the long-term culture by limiting dilution and subjected multiple subclones to DNA fingerprinting analysis. In the two cell lines (K562 and U937) without detectable STR changes during long-term bulk culture, 18 of 21 of the K562- and 23 of 26 of the U937 subclones showed no alterations in DNA fingerprint patterns compared with the parental cell lines. The three remaining subclones of each cell line revealed only a single altered STR (K562: Penta E, D8S1179, D21S11; U937: D5S818, D3S1358, D8S1179). Such alleles would not be expected to detectably alter the fingerprinting pattern if present in the bulk culture in similar abundance. This suggested the existence of a dominant clonotype even after a year in culture and a remarkable degree of genetic stability as defined by DNA fingerprinting in each of these two cancer cell lines. In sharp contrast, subclones generated from CEM-2 and Jurkat cells showed a high degree of clonotypic heterogeneity: 20 different clonotypes were observed among the 23 Jurkat and the 22 CEM-2 subclones, and profiles of the subclones differed from the parental culture by up to 6 changes in Jurkats and 13 in CEM-2. When the most divergent subclones were compared with the corresponding parental cell lines at the beginning of the long-term experiment, Jurkat and CCRF-CEM cells both had accumulated changes in 8 of the 15 STR loci studied.
3. DNA fingerprinting in cell lines generated under selection pressure
To assess how selection for certain phenotypes might affect cell line identification by DNA fingerprinting and whether this technique might be useful to illuminate the mechanisms leading to the selected phenotype, we analyzed glucocorticoid (GC) -resistant subclones of the GC-sensitive CCRF-CEM-C7H2 cell line generated by selection culture in GC-containing medium. For comparison, GC-sensitive subclones were used that derived from the same CCRF-CEM-C7H2 stock by limiting dilution in the absence of the hormone. Fingerprinting of 41 GC-resistant subclones revealed a degree of clonal heterogeneity and extent of changes from the parental cell line similar to that observed with 17 unselected CCRF-CEM-C7H2 subclones, arguing against selection of certain clonotypes. Closer observation of the ratios of the peak heights in individual loci, however, revealed a surprising difference in one of the STR markers: D5S818. Although both alleles (D5-12 and 13) were present in all subclones, a marked selection for an 
50% reduction in the heights of the D5-13 allele was seen in most GC-resistant subclones. Since this marker maps to a region on chromosome 5 (5q23.3-32), which includes the human GR gene (q31-33), it is tempting to speculate that clones were selected that had lost one allele of the target gene of this anti-leukemic drug (as detailed in the full on-line version, CCRF-CEM cells are tetraploid, hence, they have two D5-13 and two D5-12 alleles).
4. STR marker stability as parameter for identifying the parental origin of genetically drifted cell lines
As documented throughout this study, some cell lines dramatically alter their DNA fingerprinting profiles during in vitro manipulations and long-term culture, complicating subsequent assignment to their parental cell line. To identify an additional discriminatory parameter potentially useful for cell line assignment, we investigated the stability of the individual STR markers. We recorded whether a particular marker underwent alterations during long-term culture and counted its changes in the subclones generated by limiting dilution at the end of the long-term cultures of the four leukemic cell lines. This analysis revealed a dramatic difference in marker stability: four markers remained stable in all long-term culture systems and showed no (TH01) or only a few (TPOX, Penta E, Penta D) subclone mutations. The remaining showed changes during long-term culture and in a varying number of subclones. We tentatively ranked the 15 STRs using the number of altered subclones in a particular marker as parameter of stability. This ranking correlated reasonably well with a similar ranking based on statistical analyses of >50,000 paternity testings, which uncovered up to 30-fold differences in the mutation rate of the STRs used in this study. Thus, considering whether the changes between cell lines include alterations in "stable" loci may provide an additional argument in deciding the question of a cell line’s origin.
CONCLUSIONS AND SIGNIFICANCE
Tissue culture cell lines continue to be essential to research in the life sciences, and proper data interpretation depends on unambiguous definition of their parental origin, commonly performed by DNA fingerprinting. Even though this technique provides unambiguous results when comparing individuals (as in forensic medicine), it is not well understood whether cell lines maintain their fingerprinting profile throughout common experimental manipulations. We studied four widely used leukemia cell lines in this respect and observed unexpected and dramatic differences: two (U937 and K562) maintained their DNA fingerprinting profile unaltered throughout 12 months in culture and during limiting dilution subcloning. The two other cell lines (CCRF-CEM and Jurkat) showed marked alterations in DNA fingerprinting profiles during long-term culture and extensive changes in subclones generated by limiting dilution or during selection culture. This phenotype was associated with, and possibly caused by, defects in mismatch repair genes. Thus, depending on the cellular context, common tissue culture procedures may dramatically affect the DNA fingerprinting profiles, rendering definition of cell line derivation difficult (Fig. 3
). However, differences in STR marker stability and other considerations detailed in this study might offer assistance for assigning diverged clones to their parental origin. To what extent alterations in DNA fingerprinting profiles reflect functional differences in such cell lines is being investigated in our laboratory.
|
FOOTNOTES
To read the full text of this article, go to http://www.fasebj.org/cgi/doi/10.1096/fj.04-3062fje;
This article has been cited by other articles:
|
|
 |
|
  F. Demichelis, H. Greulich, J. A. Macoska, R. Beroukhim, W. R. Sellers, L. Garraway, and M. A. Rubin SNP panel identification assay (SPIA): a genetic-based assay for the identification of cell lines Nucleic Acids Res., April 1, 2008; 36(7): 2446 - 2456. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. J. Boonstra, A. W. van der Velden, E. C.W. Beerens, R. van Marion, Y. Morita-Fujimura, Y. Matsui, T. Nishihira, C. Tselepis, P. Hainaut, A. W. Lowe, et al. Mistaken Identity of Widely Used Esophageal Adenocarcinoma Cell Line TE-7 Cancer Res., September 1, 2007; 67(17): 7996 - 8001. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Presul, S. Schmidt, R. Kofler, and A. Helmberg Identification, tissue expression, and glucocorticoid responsiveness of alternative first exons of the human glucocorticoid receptor J. Mol. Endocrinol., January 1, 2007; 38(1): 79 - 90. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. K. Samraj, D. Sohn, K. Schulze-Osthoff, and I. Schmitz Loss of Caspase-9 Reveals Its Essential Role for Caspase-2 Activation and Mitochondrial Membrane Depolarization Mol. Biol. Cell, January 1, 2007; 18(1): 84 - 93. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Schmidt, J. A. E. Irving, L. Minto, E. Matheson, L. Nicholson, A. Ploner, W. Parson, A. Kofler, M. Amort, M. Erdel, et al. Glucocorticoid resistance in two key models of acute lymphoblastic leukemia occurs at the level of the glucocorticoid receptor FASEB J, December 1, 2006; 20(14): 2600 - 2602. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
