Heterogeneity of a human T-lymphoblastoid cell line

Experimental

Heterogeneity

Cell Research

of a Human K. SNOW’

171 (1987) 389-403

T-Lymphoblastoid

Cell Line

and W. JUDD

Department of Cell Biology, University of Auckland, Auckland, New Zealand

A human T-lymphoblastoid cell line (Jurkat) was cloned, and four resulting sublines were characterized in a variety of ways with the objective of gaining information on heterogeneity in cell lines. Within a few weeks of cloning, distinct cellular morphologies and growth patterns became apparent in the four sublines. Growth rate measurements made over 3 months did not show any significant differences between the sublines. Surface protein profiles obtained by radioimmunoprecipitation using antisera in conjunction with extracts from [‘5S]Met and 12’1-labeled cells revealed differences between the sublines. Analysis of total cell DNA showed that one of the sublines possessed only half the chromosome complement of the other sublines and the parental line. Karyotyping confmed this result and, in addition, demonstrated that chromosome numbers fluctuated around a mean value for each subline. Karyotypic variability became apparent within 2 months of cloning and tended to increase with time in culture. G-banding analysis showed that the analyzed cell populations contained distinctive cytogenetic aberrations. Properties of the cloned sublines were monitored over a 9-month period. One of the sublines that had shown heterogeneous morphology even after 6 weeks maintained the heterogeneity throughout this time. Another subline underwent a marked change in morphology (round to irregular) and growth habit (single cells to large clumps) with increasing time in culture. Interestingly, several alterations to surface proteins accompanied these growth changes. A third subline had relatively stable morphology and chromosome number throughout the 9month period. The modal chromosome number was hypotetraploid for three sublines and the parent line, but was diploid for another subline. However, it was interesting that progression toward tetraploidy in this subline was apparent after almost 2 years of culturing. The results showed that the original cell line consisted of a heterogeneous assemblage of cell types, some of which were quite unstable. Some implications for research using cultured cell lines are discussed. @ 1987 Academic press, IN.

Heterogeneity among tumor cells from an individual with a single malignancy is a widely recognized phenomenon (see [l, 21 for recent reviews). Although the pleoclonality of a few tumors has been shown [3, 41, most tumors are thought to be of monoclonal origin [5]. This means that generally the heterogeneity arises in clonal populations of tumor cells. Studies to date have demonstrated three points: first, that diversity can exist within a primary tumor; second, that primary tumors and secondary metastases may have different properties; and third, that heterogeneity may sometimes arise during in vitro culture of tumor cells. Using animal systems, subpopulations have been isolated from tumors induced by chemical, physical, or viral agents. Properties which demonstrate cellular heterogeneity include morphology, karyotype, growth rate, cell products, receptors, enzymes, immunologic characteristics, metastatic ability, and sensitivity to cytotoxic drugs [6, 71. ’ To whom reprint requests should be addressed. 389

Copyright Au rights of

0 1987 by Academic Press, Inc. any form reserved co14-4827/87 so3.00

reproductionin

390

Snow and Judd

The existence of heterogeneity within human tumors has also been established by a number of criteria: DNA content [8], enzyme markers [9], and antigen expression [lo, 111. Furthermore, phenotypic instability of a human leukemic Tcell line (CEM) has been shown [12]. In their study, Martin et al. [12] found that different subcultures of the CEM line had different surface antigen expression and cytogenetic features. This study extends the observation of Martin et al. by using another human Tlymphoblastoid cell line (Jurkat). Instead of looking at different samples of the cell line, one culture was cloned, and some of the resulting sublines were examined to provide information on possible subpopulations existing in the parent line. Growth characteristics, cell morphology, protein expression, DNA content, and karyotypes of each subline were examined. These properties showed that heterogeneity existed within the Jurkat cell line and demonstrated the emergence of heterogeneity within cloned populations of Jurkat cells. MATERIALS

AND

METHODS

Tissue culture. The Jurkat cell line (also known as the line JM) was derived in 1974 from the peripheral blood of a lCyear-old boy with acute lymphoblastic leukemia (ALL) [12, 131. Cells were grown as suspension cultures in RPM1 1640 medium supplemented with 10% FCS in polystyrene flasks (Nunc) at 37°C. Cloned sublines were developed by limiting dilution of Jurkat cells into %-well plates and recloning into RPM1 1640 containing 0.9% methylcellulose. A hemocytometer was used to determine cell density. Cultures were free of mycoplasma, as determined by Hoescht-33258 staining r141. [3’S]Methionine labeling of cells. Exponentially growing cells were [35S]-labeled at a density of 1~10~ cells/ml in L-methionine-free RPM1 1640 medium for 4 h at 37°C. For each 4 ml of cell suspension, 0.1 mCi of L-[3sS]methionine (NEN, 1125 Ci/mmol) was added. After labeling, cells were washed twice in Dulbecco’s PBS. “‘1 Zodination of cell surface proteins. A modification of the method by Morrison [15] was used to radioiodinate cells. Cells which were exponentially growing and had greater than 90% viability were washed three times with Dulbecco’s PBS containing Mg2+ and Ca” and then resuspended at 4~ 10’ cells/ml in PBS. To 500 ul of cell suspension was added 0.5 mCi of Na’? (Amersham, 15.9 mCi ‘?/ug iodine), 10 ul of lactoperoxidase (LPO; Sigma) solution (1.9 pg/rnl in PBS), and 40 ul of freshly diluted H20, (10-l M in PBS). The mixture was gently agitated and left for 5 min at room temperature, and then a further 10 ul of LPO solution and 20 ul of H202 were added. After a further 10 min at room temperature, the reaction was terminated by the addition of 5 ml of ice-cold PBS and the cells were washed three times in PBS. Cell extraction and radioimmunoprecipitation (RIP). Radiolabeled cells were resuspended at a density of 1.7~ lo6 cells/ml in ice-cold lysis buffer (0.05 M ‘Iris-HCl, 1% NP-40, 1 mM PMSF, pH 6.8) for 45 min. Nuclei were removed by centrifugation in an Eppendorf centrifuge for 5 min. RIP was performed according to the method of Kessler [16], using fixed Cowan I strain of Staphylococcus aureus (SACI) to adsorb immune complexes. Cell extracts containing approximately 10’ cpm were reacted with specific antisera after preincubation with normal rabbit serum (NRS) to remove any proteins bound nonspecifically by rabbit antibodies or by SACI. Anti-Jurkat and anti-peripheral blood T cell (PBLT) antisera were produced by immunization of New Zealand white rabbits with intact Jurkat cells and PHA-stimulated human T cells, respectively. SDS-PAGE. Discontinuous Laemmli gel systems were used [17], with 4% acrylamide in the stacking gel and a 5 to 15% acrylamide linear gradient in the separating gel. Fluorographs of polyacrylamide gels were prepared as described [ 181. Flow cytometry. Analyses were performed by Dr. B. C. Baguley, using an Ortho Instruments ICP 22A cytometer with a Model 2103 multichannel analyzer. Samples of exponentially growing cells were prepared as described by Taylor [19]. Chromosome analyzes. Chromosome spreads were prepared and Giemsa-stained as described by Worton and Duff [20]. G-banding was performed by digestion of chromosome spreads in 0.25% trypsin in phosphate-buffered saline for 30-40 s.

Tumor heterogeneity

391

RESULTS Growth Rates of Cultured

cells

Duplicate flasks of four clones (B2, Bl, Al, and Little) and the parent Jurkat line were subcultured every 2 to 5 days. Growth rate data were accumulated during 3 months of continuous growth (data not shown). Mean daily growth rates of the four sublines and parent line were not statistically significantly different over 3 months (using a one-way analysis of variance). However, continued observation of these cell cultures over a 2-year period has convinced us that the Little subline does grow more rapidly than the other sublines or Jurkat line. It is possible that a longer data collection period would support this opinion. Cell Morphology Figure 1 shows photographs of uncloned Jurkat cells and the four sublines shortly after cloning. Several differences among the sublines are apparent from Fig. 1 and from observation of the cultures. B2, Bl, Al, and Jurkat cells had an average cell diameter of 17-20 pm, whereas Little cells had a diameter of approximately 10 pm. B2 cells after 6 weeks in culture were quite uniform in size, had no cellular projections, and did not aggregate, even at high cell densities. Al cells after 6 weeks in culture had properties similar to those of B2 cells, but had some cellular projections. Bl cells showed a greater heterogeneity in cell size (10-40 pm in diameter) and had cellular projections, and some cells were elongated in shape while others were spherical. Bl cells readily aggregated, even at densities less than 5x ld cells/ml. Little subline after 11 weeks in culture contained approximately equal numbers of round cells which were devoid of cellular projections and cells which were elongated due to large cellular protrusions. Little cells did not readily aggregate. Cells of the uncloned Jurkat line displayed variable morphologies; however, the line appeared less heterogeneous than the B 1 subline. Approximately 70 % of uncloned Jurkat cells had cellular projections, although most of the cells were quite round in shape. Uncloned Jurkat cells had a 10-30 pm diameter range in cell size and showed aggregation properties which were intermediate between those of the Bl and B2 sublines. After 6 weeks in culture, B2 cells did not aggregate, even at densities approaching lo6 cells/ml (Figs. 2A and 2B). However, some aggregation did occur after 18 weeks in culture (Fig. 2 C) and clumping was very pronounced at 43 weeks in cell culture (Fig. 2D), even at densities less than 2.5~ lo5 cells/ml. Figures 2 C and 20 also show that with increasing time in culture, the B2 subline developed more cellular projections and became heterogeneous in size and shape. Thus over 43 weeks in culture the B2 subline dramatically changed both its morphology and mode of growth. Although the changes were not as obvious as for B2 cells, subline Al also developed more heterogeneity in cell size and shape and aggregated more readily

392 Snow and Judd

Fig. 1. Jurkat cells and four clones in culture. Cells as seen under phase contrast. Final magnification is 150X. B2, Bl, and Al sublines were photographed 6 weeks after cloning, and Little : cells were photographed after 11 weeks. (A) B2 subline; (B) Bl subline; (C) uncloned Jurkat tine; (0) ,41 subline: and Q Little subline.

with increasing time in culture. During 40 weeks of continuous subline remained quite homogeneous, whereas Bl maintained logical heterogeneity.

culture, 1:he Little its early morpho-

Tumor heterogeneity

Fig. 2. Morphology of the B2 subline. Cells of the B2 subline were photographed weeks, (C;? after 18 weeks, and (0) after 43 weeks in culture.

393

(A, B) after 6

3SS-Lab Neled Proteins RIP ’was performed using extracts of B2, Bl, and Al cells which were 35Slabeled 16 weeks after cloning and extracts of Little cells which were labeled 21

394 Snow and Judd 1

MW x 10-S

205-

116-

Q706-

29-

Fig. 3. Fluorograph of RIP or ?S-labeled proteins. After preincubation with NRS to remove nonspecific antigens, “S-labeled extracts were immunoprecipitated with 7 pl of NRS (lanes 1 to 4) 12 ul of anti-Jurkat antiserum (lanes 5 to 8) or 8 ul of anti-PBLT antiserum (Ianes 9 to 12). Each sample contained 7~ lo6 cpm at the start of the RIP. Lanes 1,5,9: Bl subline; lanes 2,6, JO: B2 subline; lanes 3, 7, II: Al Subline; lanes 4, 8, 12: Little subline.

weeks after cloning. Figure 3, lanes 1 to 4, shows that only a protein of approximately 50K MW was nonspecifically precipitated by NRS . Specific immunoprecipitations of cell extracts showed several differences between the sublines. Anti-Jurkat antiserum immunoprecipitated proteins of slightly different size from the clones in the 70K MW region (Fig. 3, lanes 5 to 8). A protein of 116K MW was immunoprecipitated by anti-Jurkat antiserum from Little cell extract only (Fig. 3, lane 8). Anti-PBLT antiserum shows that the sublines have different protein profiles between 160 and 200K MW. In this region, B2 (fig. 3, lane 10) and Al (lane II) have immunoprecipitated proteins of similar molecular weights, which are differ-

Tumor heterogeneity ent from those of Bl (lane 9) and Little reproducible. ‘251-labeled Cell &&ace

395

(lane 22) cells. These results were quite

Proteins

Cell surface proteins were of particular interest since plasma membrane properties affect tumor cell-immune system interactions, metastatic potential [21], and possibly drug resistance [22]. Cells of the Jurkat line and four sublines were radioiodinated using LPO and detergent extracts were made. Samples of extracts containing equivalent counts per minute were separated by PAGE under reducing (Fig. 4, lanes I to 6) and nonreducing (Fig. 4, lanes 7 to II) conditions. In the presence of 2-mercaptoethanol, major qualitative differences between clones were found in the regions HO-210K, 12&135K, and 6%70K MW. Quantitative differences between clones were seen at 20K, 35K, and 100 K MW. In addition, quantitative changes occurred in B2 subline proteins between 15 weeks (lane 3) and 35 weeks (lane 2) in culture: after a longer time in culture, the clone showed a relatively less intense band of 60K, a relatively more intense band of 9OK, but the same amount of 135K MW protein. In the absence of reducing agent, similar quantitative and qualitative differences were apparent, although at some different molecular weights. For example, the 65-70K MW bands seen in the presence of mercaptoethanol have decreased to approximately 6OK MW, but the sublines still show the same pattern of variation. The 65-70 K MW variations seen in Fig. 4 are similar to those seen in Fig. 3 of antiJurkat antiserum immunoprecipitated proteins. Further analysis of radioiodinated surface proteins by two-dimensional PAGE showed a series of spots from each detergent extract which spanned this molecular weight range and had different isoelectric points. The precise range of MW and pZ values for the series of spots varied for each subline (data not shown). Lectin binding studies have shown that the 65-70K MW protein is a glycoprotein (W. Judd, unpublished). Therefore, glycosylation differences could account for the observed molecular weight variation of the protein between clones 1301. Profiles of radioiodinated proteins near HO-210K MW were similar to those seen after RIP by anti-PBLT antiserum, suggesting that the 180-210K MW “Slabeled proteins of Fig. 3, lanes 9-12, were cell-surface proteins. At least some of these high-molecular-weight proteins probably belong to the human equivalent of the murine Ly5 antigen system [231. Differences observed only in Fig. 3 (i.e., not seen in Fig. 4) may involve intracellular proteins or surface proteins that are not accessible to radioiodination. Chromosome

Analyses of Jurkat

Sublines

Cytofluorometry of Jurkat cells and sublines B2, B 1, and Al after 9 to 25 weeks of culture showed no significant differences in DNA content between the various cell types and no significant change in DNA content with increasing time in

396 Snow and Judd

205-

11697-

45-

29-

Fig. 4. Radioiodinated proteins from Jurkat sublines. Lanes 1 to 6, 2-mercaptoethanol present; lanes 7 to II, 2-mercaptoethanol absent. Lanes 1 and 11, uncloned Jurkat; lane 2, B2 after 35 weeks in culture; lnnes 3 and 10, B2 after 15 weeks in culture; lanes 4 and 9, Bl after 15 weeks in culture; lanes 5 and 8, Al after 15 weeks in culture; lanes 6 and 7, Little after 20 weeks in culture. Lanes 4 and 9 were inserted from a shorter exposure fluorograph of the same gel.

culture (data not shown). Analyses of Little cells after 14 and 30 weeks of growth showed that these cells contained the same amount of DNA as normal human cells, which was approximately 50% of the DNA content of the other sublines and Jurkat cell line (Fig. 5). Giemsa-stained chromosomes were counted to yield information on the homogeneity/heterogeneity of DNA content within each clone, since cytofluorometry gave only an average result for a collection of cells (between 11,000 and 40,000 cells for each sample analyzed by cytofluorometry). Table 1 shows the number of

Tumor heterogeneity loo-

*

397

1

75PE

50-

25-

Channel

Number

x lo-*

Fig. 5. Cytograms of B2 and Little sublines. Relative DNA content of cells, which was measured from the intensity of fluorescence from each cell, is represented by channel number. Pigeon erythrocytes, used as internal standard, produced peak PE. (A) B2 subline after 46 weeks in culture. (If) Little subline after 14 weeks in culture.

chromosomes in Jurkat cells and sublines after various times in culture. Between 16 and 25 cells were analyzed for each result. The Jurkat line and sublines B2, Bl , and Al each showed a very wide range in the number of chromosomes per cell; however, the modal number was hypotetraploid in each case. In contrast, the modal number for the Little subline was diploid and the cell populations showed a much smaller range of chromosome numbers, even after 37 weeks in culture. A more detailed cytogenetic analysis was performed by G-banding cells from the Jurkat line, from the Bl subline after 90 weeks of culture, and from three different isolates of the Little subiine. The Little subline was analyzed after 23 weeks of growth, and after 53 weeks of growth was divided into two separate subclones (Little (1) and Little (2)), which were analyzed after 92 and 89 weeks of total culture, respectively. Table 2 lists the structural aberrations which were present in more than one cell per sample. The nomenclature used is according to the International System for Human Cytogenetic Nomenclature (ISCN) 1291. Identical karyotypes for cells within the same culture were seen only in the Little subline after 23 weeks of growth: 26478338

398

Snow and Judd

TABLE Chromosome Line subline

Time in culture (weeks)

Jurkat B2

9

B2

17

B2

37

Bl

9

Al

9

Little Little

14 37

numbers

1 of Jurkat

Number of chromosomes per cell 66, 74, 76, 78, 79, 83, 86(S), 87(5), 88(2), 94, 166, 167, 169 54, 70, 75, 77, 83, 84(2), 85, 86(2), 87(4), 88, 90(2), 172 48, 53, 66(2), 69, 70, 75(3), 76(2), 84, 85(2), 86(3), 87(2), 89, 90(2), 92(2), 172 67, 74, 77, 78, 79, 83, 84(2), 85(2), 86(5), 87(4), 88, 89, 250 65(2), 82, 83, 87, 88(2), 89, 90(2), 91(2), 145, 176, t83, 283 55, 71, 75, 76, 83(3), 84(2), 85(3), 86(2), 87(4), 88, 92, 153 46(12), 47(4), 48(2), 49 46(12), 47(3), 48(2), 49, 60(2), 95

sublines Mean number of chromosomes Standard per cell deviation

95

30

87

23

82

22

91

36

112

58

86 47 50

17 1 11

Nore. Results were obtained for between 16 and 25 cells of each linelsubline after various times in culture. Where more than one cell had the same number of-chromosomes, that number of cells is indicated in parentheses.

three cells had the karyotype 46 X, Y, 2p-, inv(3), 13qthree cells had the karyotype 46 X, Y, 2p-, inv(3) two cells had the karyotype 46 X, Y, 2p-, inv(3), 13p-q-. Several aberrations did appear fairly consistently. The aberration 2p- occurred at a very high frequency in all of the populations analyzed. Therefore, this deletion may have been present in the karyotype of the person from which the Jurkat line was derived or it may have occurred some time ago in vitro with subsequent selection during growth of the cell line. The aberration inv(3) is present at high frequency in the Bl and Little sublines but was not seen in the sample of uncloned Jurkat cells, although it may have been present at a higher frequency at the time of cloning the parent line. The aberration de1(6)(pter-+q22:) was found in Bl cells only; however, since it was present in only 8 out of 10 cells analyzed, the deletion may have occurred since the clone was isolated. A high number of chromosome 13 aberrations appeared as a feature of each Little cell population. This may reflect an inherent genetic stability of a chromosome 13 in the parent line at the time of cloning since an aberrant chromosome 13 was also seen in 5 out of 11 Jurkat cells. Table 3 lists the numerical aberrations which were present in more than one cell per population, based on the normal human karyotype and double the normal human karyotype for near diploid or near tetraploid cells, respectively. The parent line showed a striking variability in the specific chromosomes lost from each cell, although missing chromosomes 22 and Y were quite frequent. A small

Tumor heterogeneity

399

TABLE 2 Str~cturuZ ffberr~tions

IP2Pinv(3)(p2lqter) 4qw5q6qdel(6~(pter4qZ~) wllq13p13q33p-qW+ 17mmar 1 mar2 mar 3 mar4 mar 5 mar 6 mar 7 Number of cells karyotyped

seen after G-banding of Jurkat, described in the text

Jurkat

$90 weeks)

-

-

11 3 2 3 -

10 8 -

-

8

-

5 -

-

3

-

-

6 2

11

Little (at 23 weeks)

Little(l) (as 92 weeks)

Little(2) (at 89 weeks)

2 19 21 -

-

-

11 11 -

-

-

10 10 4 6 5

-

-

BI, and Little

-

5

-

-

4 4 6

-

2 7 2

-

3 3

-

10

21

11

-

9 2

cells as

-

-

8 1 2 8 3

10

range of chromosomes was involved in the numerical aberrations in the cloned cells: furthermore, chromosomes involved were different for the Bi and Little populations. The consistency seen in the Little and Bl clones suggests that the variability present in the parental Jurkat cell line was not due to random loss of chromosomes during sample preparation. DISCUSSION Results presented in this paper demonstrate that the Jurkat cell line consists of a heterogeneous population of cells. Four cfoned sublines each displayed a unique mo~hology and unique patterns of protein expression. Fu~he~ore, the Little subline had an average DNA content significantly different from those of the other sublines and parent line. The range of cytogenetic aberrations seen in the Jurkat line and Bl and Little sublines clearly demonstrates genetic instability of these cells. The results also show emergence of heterogeneity within cloned populations of

400 Snow and Judd

TABLE Numerical

aberrations

seen after G-banding of Jurkat, described in the text

Jurkat

-10 -11 -12 -15 -16 -17 -20 -21 -22 -Y -2Y -x +17 +19 +21 +Y Number of cells near diploid (range) Number of cells near tetraploid (range)

3

-

cells as

Little (at 23 weeks)

Little( 1) (as 92 weeks)

Little(2) (at 89 weeks)

-

-

-

-

-

-

2

-

10 2 9 4 -

-

0

0

1l(83-88)

and Little

Bl (at 90 weeks)

3

Bl,

3 2

7

-

-

4 2

10(89-100)

-

-

-

2

19(45,46) 2(92)

3 7 6

6 5

-

5

9(45-49)

3(45,46)

Wl)

7(8693)

human tumor cells. After only 6 weeks of growth, the Bl subline displayed marked morphological diversity, and after 9 weeks in culture had a variance of mean chromosome number greater than that of the Jurkat line (P~0.01). The Al and B2 sublines also contained wide ranges of chromosome numbers after 9 weeks in culture, although the variances were respectively smaller than and equal to the variance of the parent population. In the B2 subline, growth characteristics and expression of 60K and 90K MW proteins were altered, and an increase in chromosome number variance (P
Tumor heterogeneity

401

result in this sort of progressive increase in karyotypic diversity, particularly in the absence of selection pressure for a specific phenotype. A greater percentage of tetraploid cells is apparent in the older Little cultures compared with the 23-week culture (see Table 3). Selection of tetraploid cells may have occurred in the Little culture since new cytogenetic aberrations would have a greater chance of being lethal in diploid cells (due to a gene dosage effect). The production of lethal aberrations in diploid cells would also mean that the diversity in a diploid population would increase more slowly than in a tetraploid population having the same rate of aberration production. This was shown in Tables 2 and 3, where the Little (2) population had a wider range of aberrations than the Little (1) population. Since Little (1) and Little (2) subclones were grown under identical conditions, it seems likely that the difference in ploidy is stochastic, and continued growth of the Little (1) subclone would continue to select for an increasingly higher percentage of tetraploid cells. It is not clear why the Little (2) subclone has greater karyotopic variability than the Bl subline cultured for the same time period. Perhaps the cultures have different inherent genetic instabilities, or possibly the rapid production of variation in the Little (2) subclone is associated with the selection for tetraploidy. Further studies are required to establish how such cytogenetic aberrations are produced in cultured cell lines. Possibilities include defective repair of damage produced by unknown clastogenic substances, incomplete replication of DNA, and defective segregation at mitosis. Is one mechanism common to most established cell lines, and does the production of aberrations in vitro resemble the production of aberrations in uiuo? Aneuploidy has been correlated with a more aggressive tumor course and with the ability of ovarian and renal-cell tumors to form colonies in soft agar [31]. Is the production of cytogenetic aberrations an integral part of the uncontrolled growth of tumor cells in uiuo, and, if so, does this select for aneuploid tumor cell populations? Phenotypic instability of tumor cells in uiuo presents a major problem for effective cancer treatment due to the emergence of variants under the selection pressure of tumor therapy. Eradication of cells by the immune system or immunotoxins would be hindered by antigenic variation of any tumor-associated proteins, and rapid emergence of drug-resistant variants does pose a problem for chemotherapy [24]. Our results indicate that within a population of tumor cells, there exist very unstable populations as well as subpopulations which maintain fairly stable phenotypes. Subpopulations which maintain phenotypic diversity will be important to the survival of tumor cell populations which are subject to selection pressures. The rapid production of variants would ensure that some cells survived adverse conditions. During chemotherapy, the application of clastogenic drugs could actually enhance the production of resistant variants if the concentration of drug reaching the tumor cells is sublethal but high enough to cause cytogenetic aberrations and thereby increase the genetic diversity of the tumor cell population. The identification of subpopulation interactions which promote phenotypic stability would permit greater understanding of the control of heterogeneity.

402

Snow and Judd

Poste et al. [25, 261 found that development of heterogeneity in a uniform population of cells could be suppressed by the interaction of subpopulations with different characteristics. Some evidence [27, 281 suggests that this interaction does not occur by cell contact. Therefore it would be interesting to search for a diffusible cell product which suppresses the development of heterogeneity. For example, culture medium conditioned by the Jurkat cell line may contain a factor which suppresses the development of Bl heterogeneity. These findings, demonstrating variability in a cell line and rapid emergence of heterogeneity in clonal populations, have considerable implications for research using cell lines. The Jurkat cell line was studied simply because it was available, and there is no reason to believe that its general properties differ from those of other animal and human cell lines. Clearly a cell line is not necessarily a homogeneous population of cells that stays the same over long periods. Like Jurkat, most lines probably consist of a diverse range of subpopulations that can constantly change. Differences in culturing conditions probably exert selection pressures that select certain phenotypes. Such differences could include precise growth temperature, oxygen tension, components present (or absent) from particular batches of FCS (notoriously variable), and treatment of tissue culture flask surfaces performed by manufacturers to promote cell adhesion. It is conceivable that cell line heterogeneity and phenotypic instability may produce “experimental” results rather than the experimental protocols used. The same phenomena could also account for some differences in results between laboratories supposedly studying the same cell line. Other widely used human cell lines should also be examined to confirm that heterogeneity does extend beyond Jurkat cells. Assuming (as seems likely) that this is a widespread phenomenon, there is only a little that can be done to ameliorate the situation. Exchanges of cells between laboratories and not carrying lines as cultures for unnecessarily long periods are obvious measures. We thank the Cancer Society of New Zealand and its Auckland Division for financial support, as well as the University Grants Committee and the Auckland University Research Committee. We are also grateful to Dr. Bruce Baguley for assistance with DNA cytofluorometry, Dr. G. Finlay for the Jurkat cell line, Mr. M. Parslow for advice on cytogenetic analysis, and Dr. G. Seber for doing the statistical analysis.

REFERENCES 1. Dexter, D. L., and Calabresi, P. (1982) Biochim. Biophys. Acta 695, 97. 2. Heppner, G. H. (1984) Cancer Res. 44, 2259. 3. Reddy, A. L., and Fialkow, P. J. (1979) J. Exp. Med. 150, 878. 4. Woodruff, M. F. A. (1983) Brit. .I. Cancer 47, 589. 5. Nowell, P. C. (1976) Science 194, 23. 6. Dexter, D. L., Kowalski, H. M., Blazar, B. A., Fligiel, Z., Vogel, R., and Heppner, G. H. (1978) Cancer Res. 38, 3174. 7. Neri, A., and Nicolson, G. L. (1981) Znt. J. Cancer 28, 731. 8. Petersen, S. E., Bichel, P., and Lorentzen, M. (1978) Eur. .I. Cmcer 15, 393. 9. Baylin, S. B., Weisburger, W. R., Eggleston, J. C., Mendelsohn, G., Beaven, M. A., Abeloff, M. D., and Ettinger, D. S. (1978) N. En&. J. Med. 299, 105. 10. Albino, A. P., Lloyd, K. O., Houghton, A. N,, Oettgen, H. F., and Old, L. J. (1981) .I. Exp. Med. 154,1764.

Tumor heterogeneity 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

McGee, J. O’D., Woods, J. C., Ashall, F., Bramwell, M. E., and Harris, H. (1982) Lancet 2, 7. Martin, P. J., Giblett, E. R., and Hansen, J. A. (1982) Immunogenetics 15, 385. Schneider, U., Schwenk, H.-U., and Bomkamm, G. (1977) ht. .I. Cancer 19, 621. Chen, T. R. (1977) Exp. Cell Res. 104, 255. Morrison, M. (1980) in Methods in Enzymology (Van Vunakis, H., and Langone, J. J., Eds.), Vol. 70, p. 214, Academic Press, New York. Kessler, S. W. (1981) in Methods in Enzymology (Langone, J. J., and Van Vunakis, H., Eds.), Vol. 73, p. 442, Academic Press, New York. Laemmli, U. K. (1980) Nature (London) 227, 680. Bonner, W. M. (1983) in Methods in Enzymology (Fleischer, S., and Fleischer, B., Eds.), Vol. %, p. 215, Academic Press, New York. Taylor, I. W. (1980) J. Histochem. Cytochem 28, 1021. Worton, R. G., and Duff, D. (1979) in Methods in Enzymology (Jakoby, W. B., and Pastan, I. H., Eds.), Vol. 58, p. 322, Academic Press, New York. Poste, G., and Fidler, I. J. (1980) Nature (London) 283, 139. Juliano, R. L., and Ling, V. (1976) Biochim. Biophys. Acta 455, 152. Judd, W., Poodry, C. A., Broder, S., Friedman, S. M., Chess, L., and Strominger, J. L. (1980) Proc.

Natl.

Acad.

Sci.

USA

77, 6805.

24. Goldie, J. H., and Coldman, A. J. (1979) Cancer Treat. Rep. 63, 1727. 25. Poste, G., Doll, J., and Fidler, I. J. (1981) Proc. Natl. Acad. Sci. USA 78, 6226. 26. Poste, G., Tzeng, J., Doll. J., Greig, R., Rieman, D., and Zeidman, I. (1982) Proc. Natl. Sci. USA

27. 28. 29. 31. 31.

403

Acad.

79, 6574.

Miller, B. E., Miller, F. R., Leith, J., and Heppner, G. H. (1980) Cancer Res. 40, 3977. Miller, B. E., Miller, F. R., and Heppner, G. H. (1981) Cancer Res. 41, 4378. ISCN (1978) defined in Cytogen. Cell Genet. 21, 309. Latner, A. L., Marshall, T., and Gambi, M. (1980) Clin Chim. Acta 103, 51. Verheijen, R. H. M., Feitz, W. F. J., Beck, J. L. M., Debruyne, F. M. J., Vooys, G. P., Kenemans, P., and Herman, C. J. (1985) Znt. J. Cancer 35, 653.

Received October 3, 1986 Revised version received February 5, 1987

Printed

in Sweden