Degradation of specificity in cytolytic T lymphocyte clones: Mouse strain dependence and interstrain...

Eur. J. Immunol. 1984.14: 951-956 Loss of cytotoxic T lymphocyte clone specificity 951

Anne Wilson and Ken Shortman

Degradation of specificity in cytolytic T lymphocyte clones: mouse strain dependence and interstrain

The Walter and Eliza Hall Institute of Medical Research, Melbourne

transfer of nonspecific cyto~ysis*

Limiting-dilution culture of murine Ly-2' T cells with concanavalin A (Con A) and irradiated spleen filler cells produces, with high efficiency, cytolytic T lymphocyte (CTL) clones, With most mouse strains (inchding CBA and C57BLl6) the specificity of these CTL clones drops after day 6 of culture, so that by day 9 the majority of clones can lyse most murine target cells, whether syngeneic or allogeneic. The rate of specificity degradation and relative target cell preference varies with the mouse strain. Some strains (e.g. BALBlc) do not show this effect and CTL clones remain specific to day 9. Many low natural killer (NK) cell strains (e.g. C57BLl6J.bg) maintain CTL specificity in such cultures, but the correlation between CTL specificity and NK status is not absolute. Growth of BALBlc precursor cells on CBA filler cells leads to specificity degradation in the BALBlc CTL clones; however, the specificity of CBA- derived CTL clones is not maintained by growth on BALBlc fillers. The results suggest that specificity degradation is induced in the developing CTL-clone by factors in the culture environment, perhaps a soluble lymphokine (a differentiation factor) or by an infectious agent (an endogenous mouse virus). Although such CTL specificity loss may render many limiting dilution studies of the CTL specificity repertoire invalid, the problem may be bypassed by an appropriate choice of mouse strain.

1 Introduction

The lytic specificity of CTL is believed to reflect the specificity of the T cell receptor, and clones of CTL, obtained as perma- nent cell lines or as short-term limiting-dilution (LD) cultures, have been used to study the T cell specificity repertoire. How- ever, a caveat is raised by recent reports [l-71 which suggest that clones of CTL may also express another class of receptors mediating lysis. Thus, individual CTL clones sometimes dis- play patterns of target cell preferences which are either of the broad activated lymphocyte killer (ALK) [8] type, or of the more restricted natural killer (NK) cell type, in addition to the allospecific or H-2-restricted clonally distributed antigen- specific cytolysis. Culture conditions which prevent the development of such nonspecific cytolysis must be devised before LD analysis can provide reliable estimates of specific CTL precursor (CTL-P) frequencies.

We have described a high cloning efficiency, LD culture system, where most Ly-2' T cells produce CTL clones after stimulation with concanavalin A (Con A) in the presence of irradiated spleen filler cells [9, 101. However, the cells in such clones were large, granular lymphocytes which lysed most murine target cells, syngeneic or allogeneic [l]. We have demonstrated that these cytolytic cells were derived from the individual Ly-2' precursor cells and not from the irradiated fillers,

[I 45221

that the cytolytic cells were themselves Ly-2' Thy-l', and that individual clones showing nonspecific cytolysis at day 8-9 of culture were derived from clones which showed specific cytolysis at day 5-6 [2]. In the present study we report some observations which help to clarify the nature of this nonspecific killing. Different strains of mice gave CTL clones which varied in the degree of nonspecific cytolysis with different target cells, and CTL clones from some strains did not develop such anomolous killing at all. The ability to develop nonspecific cytolytic activity late in culture could be transferred in vitro from cultures of positive strains to cultures of negative strains.

2 Materials and methods

2.1 Mouse strains

Unless otherwise specified all mice were specific pathogen free, 5-6-week-old males bred at The Walter and Eliza Hall Institute animal facility. The strains used were CBAlCaH Wehi; BALBlc An Bradley Wehi; C57BLl6J Wehi; C57BLl6 Kap; AKWJ Wehi; DBA/IJ; A/J Wehi; SJWJ Wehi. The beige mutant mice (C57BLl6J.bg) were 5-8-week-old males reared under conventional conditions in a separate breeding facility at this Institute.

2.2 Mouse tumor cell lines * This work was supported by the National Institutes of Health,

U.S.A., Grant Number A1 17310; by the C. H. Warman Research Fund; and by the National Health and Medical Research Council, Australia,

Correspondence: Ken Shortman, The Walter and Eliza Hall Institute of Medical Research, Post Office, Royal Melbourne Hospital, Vic- toria 3050, Australia

Abbreviations: Con A Concanavalin A PHA Phytohemagglutinin CTL: Cytotoxic T lymphocyte NK: Natural killer (cell) ALK Acti- vated lymphocyte killer (cell) LD: Limiting dilution CTL-P CTL precursor SC: Spleen cell(s)

0 Verlag Chemie GmbH, D-6940 Weinheim, 1984

All cell lines were maintained in continuous culture in Dulbec- CO'S modified Eagle's medium containing 10% fetal calf serum. The lines were P815, an H-2d mastocytoma; EL4, an H-2b T lymphoma; R1, an H-2k thymoma and YAC-1, an H-2a lymphoma.

2.3 LD cultures

These are described in full detail elsewhere [9, 101. Spleen cells (SC) were cultured in flat-bottomed 96-well trays at levels

0014-2980/84/1010-0951$02.50/0

952 A. Wilson and K. Shortman Eur. J. Immunol. 1984.14: 951-956

between 1 and 20 cellsiwell with 0.5 x lo6 irradiated spleen filler cells in 0.2 ml HEPES-RPMI 1640 medium containing Con A , 2-mercaptoethanol, fetal calf serum and (except where otherwise stated) a supernatant from Con A-stimulated rat spleen cultures. Ninety-six channel replicators were used for setting up cultures so large numbers could readily be han- dled. After 7-9 days almost all responder T cells had grown to form clones averaging 60 000 cells. Control cultures containing irradiated filler cells alone were also set up; these consisted mainly of dead and damaged cells at day 7-9.

2.4 Assay for the cytolytic activity of LD cultures

The semiautomated radioautographic '"In-release assay used is described in full detail elsewhere [9, 111, where examples of the readout are also provided. Cultures were assayed either for direct cytolysis of labeled tumor target cells, or for lectin- enhanced cytolysis by the addition of phytohemagglutinin (PHA). Ninety-six-channel replicators were used for all trans- fers and additions of cell suspensions, as well as for supernatant sampling. Supernatants were removed from LD cultures and 'llIn-oxine labeled target cells (5000 cells, 4-5 cpmicell) added, followed where required by PHA-M (1 : loo), in a total volume of 0.17 ml medium. After mixing, the suspension of target and cultured cells was transferred to V-well trays. The trays were centrifuged to pellet the cells, then incubated for 4 h at 37°C. After again centrifuging, a 50-1.11 sample of the supernatant was removed from each well, transferred to soft U-well trays, and exposed overnight at -70 "C against X-ray film with enhancing screens. After development, a black spot represented a positive culture. The relative cytotoxic potential of cultures was quantitated when required by densitometry measurements on the film using an automatic coloriineter (Multiskan Titertek, Flow, Rockville, MD). The sensitivity of the assay was such that 200 mixed lymphocyte culture-gener- ated allospecific CTL were sufficient to register above the barely visible darkening caused by the 4 9 % spontaneous release

2.5 Frequency of precursors of cytolytic clones

The frequency among the responding SC of the precursors of cytolytic clones was calculated from the proportion of negative cultures as the slope of the line of best fit to the zero-order term of the Poisson probability distribution, as described elsewhere [2, 91. Each frequency calculation was based on two responder cell doses, with 192-,384 cultures per dose, as well as 192 filler cell background cultures. The precise fit of the data, for both direct and PHA-enhanced cytolytic cultures, to a single precursor model without helper or suppressor effects has been demonstrated repeatedly elsewhere [2, 9, lo]. In this study such results are, therefore, summarized as simple frequency estimates.

similar relative percent specific lysis values for the series of mouse strains tested, so only one representative series of values is presented.

3 Results 3.1 A simple index of relative CTL clone specificity

In a preceding report [2] we demonstrated that CTL clones obtained by LD culture of CBA or C57BLi6 SC or lymph node cells were target cell specific if tested early (day 5-6) but nonspecific if tested late (day 8-9). The formal proof was provided by split-culture assays on pairs of targets differing in H-2 haplotype: all clones sampled early lysed only one of the pair, whereas most clones sampled late lysed both [2]. A split-culture approach also allowed us to demonstrate that the early, apparently specific clones gave rise to late, nonspecific clones [2]. This change from specificity to nonspecificity was also monitored by comparing the lysis of one target cell (P815) in the presence of PHA (where all CTL-clones should register, regardless of specificity) with lysis of the same target cell in the absence of PHA. Over the period from day 6 to day 9 the level of lectin-enhanced clones did not change greatly, whereas the level of clones showing direct cytolysis of P815 changed from the low frequency expected of allospecific clones to a very high frequency of nonspecific clones [2]. Split-culture assays demonstrated that virtually all cultures showing direct cytolysis were also positive with the lectin-mediated readout [1], so it was possible to perform direct and lectin-enhanced CTL clone frequency estimates on separate unsplit trays of cultures, and thus maintain maximal sensitivity for clone detection.

We devised a simple index of relative specificity based on these earlier observations. Separate trays of unsplit LD cultures were assayed for cytolysis of a given target cell, either with or without PHA. The ratio of the frequency of all CTL clones (+ PHA) to the frequency of directly cytolytic clones (- PHA) served as a specificity index. Ratios around 1, where all CTL clones are capable of direct cytolysis of the target, indicated that almost all clones were nonspecific. Ratios between 10 and 20, the usual upper range, indicated a relatively high level of specific clones; if it is assumed that between 1 in 10 and 1 in 20 of all CTL-P have a pre-determined allo- specificity, such ratios could indicate the complete absence of nonspecific clones. Since when the level of directly lytic clones was low the error in their frequency estimation was propor- tionally high, specificity ratios above 10 were subject to con- siderable variation, and all values above 10 should probably be grouped as a single "high specificity" score. This + PHAi -PHA frequency index was used in the present study to monitor changes in CTL clone specificity in cultures of SC from different mouse strains (see Fig. 2), as discussed below.

3.2 The kinetics of CTL clone specificity degradation in cultures of CBA SC

2.6 NK activity of SC

Freshly prepared SC suspensions were assayed for direct cytolysis of YAC-1 tumor cells as a measure of NK activity. SC were incubated with "'In-labeled YAC-1 cells (104/well; 3-4 cpmkell) at 30 : 1 to 100 : 1 effector to target ratios, for either 6 h or 18 h. The background release averaged 3% and 11% for the latter two periods. All incubation conditions gave

To check if the kinetics of the change to nonspecific cytolysis was the same for all target cells, syngeneic and allogeneic, the previous studies [2] with CBA SC assayed on P815 targets were repeated with two allogeneic targets P815 (H-2d) and EL4 (H-2b), and the syngeneic target R1 (H-zk). The results are included in Fig. 1. The peak in the total frequency of detectable CTL clones (assayed in the presence of PHA) was

Loss of cytotoxic T lymphocyte clone specificity 953 Eur. J. Immunol. 1984.14: 951-956

CBA

D 4 --- - 5 6 7 8 9 5 6 7 8 9 5 6 7 0 9

MY5 OF CULTURE

from days 7 and 9, and was similar for all target cells. The efficiency of CTL clone development was a little lower in these experiments than usual; frequencies close to 0.1 are more usually obtained [l, 2 , 91. The frequency of clones showing direct cytolysis (assayed in the absence of PHA) was low initially, but rose rapidly to approach the total (+ PHA) CTL clone frequency by day 8 to 9. Direct cytolysis rose in parallel for all three target cells, including the syngeneic R1 target. Re-ex- pressing this data as a specificity ratio (Fig. 2) indicated that the degradation of specificity towards all three targets was continual from day 5 to day 8. Previous studies [2] had demonstrated that this decline in specificity was not a sensitivity problem, but was observed across a wide spectrum of effector to target cell ratios.

3.3 The kinetics of CTL clone specificity degradation in cultures of C57BL/6 SC

A comparable series of experiments was carried out culturing SC from C57BL/6 mice, and assaying CTL clone frequencies on the same three tumor cells, with EL4 now representing the syngeneic target (Figs. 1 and 2) . In these experiments the total frequency of CTL clones (assayed with PHA present) was high compared to CBA, although over a larger series of experiments the cloning efficiencies were not consistently different between two strains. Peak responses were again obtained from day 7-9. Direct cytolysis of all three targets arose late in the cultures, in general agreement with the CBA results. How- ever, the degradation of specificity appeared to start after day 6, later than for CBA cells, and even at day 9 the proportion of clones showing specificity degradation was lower than for CBA cells.

3.4 Failure of CTL clones in cultures of BALBIc SC to show specificity degradation

The comparable series of experiments on the kinetics of CTL clone development with SC from BALB/c mice, assaying on

Figure 1. The effect of mouse strain and tumor target on the frequency of clones showing direct cytolysis and lectin-enhanced cytolysis in LD cultures. Direct (- PHA) and lectin-enhanced (+ PHA) assays for the presence of cultures able to lyse the target cells were carried out on separate but parallel trays of cultures, and the frequency of positive clones (apparent CTL-P frequency) was calculated by LD analysis from the zero-order term of the Poisson distribution. Each graph represents the pooled data from two (EL4 and R1 targets) or four (P815 targets) separate experiments, with multiple assay trays for each point in each experiment. The mouse strain desig- nated served as the source of both the few responder and the many irradiated filler cells in each culture. Full details are given in Sect. 2.4.

the same three tumor cells with P815 now the syngeneic target, gave results quite different from CBA or C57BLi6 (Figs. 1 and 2). The frequency of total detectable CTL clones (assayed with PHA present) again peaked between days 7 and 9. This frequency was usually a little lower than with CBA or C57BL16,

50 s 5

1

P815 - EL4 0- - -o

A,..,.

1°k 5

__/-Q---_ ... b ....,,,. ----

-. ..... -O- -. -5 '... ....... ...,

..... .... ..._. ,,....." ..(

h. "

BALB/c

5 6 I 8 9 1 1 4 5 2 2 . 2 1 Days of culture

Figure 2. The degree of specificity degradation in LD cultures of SC from different mouse strains. The ratio of the frequencies of clones showing lectin-enhanced (+ PHA) cytolysis to those showing direct (- PHA) cytolysis is calculated from the data of Fig. 1. A ratio of 10 or above is indicative of a relatively high CTL clone specificity, whereas a ratio of 1 indicates that almost all clones are nonspecific.


although it should be emphasized that the degree of cytolysis (CTL clone size) was no less. However, the level of clones showing direct cytolysis remained low up to day 9 for all three targets. The specificity ratio index (Fig. 2) remained around 10 or above from day 6 to 9, and in many assays with BALB/c cells has never dropped below 9. By this criterion, CTL clones in BALB/c spleen cultures remain specific.

CBA CULNRES C57BUb CULTURES

~ -5HA P815

EL4 n

R' J-l '-

.005 91 9 2 .04 .08 46 32 .64 G€)5.01 .02 .04 .08 46 32 .64 114522.31 CYTOLYTIC ACTNITY [ABSORBANCE VAWES 1

Figure 3. The distribution of direct and lectin-enhanced cytolytic activity within positive clones in day-8 LD cultures. The degree of "'In release after incubation of labeled target cells with the cells in individual cultures in the presence (+ PHA) or absence (- PHA) of lectin was assessed by absorbance measurements on the exposed X-ray film of the radioautograph assay. Only values for cultures with readings above the spontaneous release background are shown. An absorbance value of 0.1 corresponds to about 35% specific lysis. Note the log, scale of absorbance readings used to compress the spread to very high absorbance readings. Only trays with 30% or less positive cultures were used, to reduce the coincidence rate; no correction is applied for residual coincidence of two clones in one culture well. Each distribution is based on 200-300 positive wells.

3.5 Extent of nonspecific cytolysis of various target cells by CTL clones

The data of Fig. 1 indicate the high frequency of CBA or C57BL/6-derived CTL clones which show detectable nonspecific cytolysis at day 8-9, but do not show the extent of this lysis. Previous studies with the CBA-P815 combination have shown that nonspecific lysis can be very effective, since bulk- harvested day 8-9 effector cells can give extensive '"In release at effector-to-target cell ratios of 3: 1 [l]. To compare the extent of direct cytolysis of the three target cells, densitometry measurements were made on the autoradiographs of "'In release from the targets by each individual clone, since the darkening of the film is linearly related to effector cell number [ll]. The level of responding cells/culture was kept low, so very few wells contained more than one clone. A comparison of the extent of direct cytolysislclone with the extent of PHA- mediated cytolysis for day 8 cultures is shown in Fig. 3.

As observed previously [l, 21, CTL clones harvested at day 8 from CBA SC cultures gave very extensive direct cytolysis of P815 target cells, almost as effective as the lysis obtained when PHA was added. In contrast, lysis of R1 or EL4 was less extensive, although still quite definite. However, the extent of cytolysis was not just due to inherent differences in lysability between target cells, since with C57BLiG-derived clones the target ranking changed, and more clones giving extensive lysis were seen with the syngeneic target EL4 than with P815.

3.6 Survey of mouse strains for CTL specificity degradation in culture and NK status

Since BALB/c mice appeared to maintain CTL specificity in LD culture, we surveyed the available strains in this Institute to see if any other such strains behaved in the same way. In view of the target cell preference differences in Fig. 3, this was carried out using both P815 and EL4 to monitor the extent of direct cytolysis at day 9 of culture. Fresh uncultured spleen suspensions from each strain were also tested for direct cytolysis of YAC-1 target cells, to see if a tendency to CTL specificity degradation in culture was correlated in any way with NK status. The results are summarized in Table 1.

Table 1. The NK status and CTL clone specificity maintenance of different mouse strains

Strain H-2 N K status"' Specificity o f CTL clones YAC target P815 target (H-2') EL4 target (H-2h)

(% specific lysis Frequency x 10' at day 9 Ratio Frequency x 10' at day 9 Ratio 30 : 1 E : T) + PHAI + PHA/

6 h 18 h no PHA with PHA -PHA no PHA with PHA -PHA

CBA AKR C57BLl6 C5lBU6 Kap BALBIc DBA NJ SJUJ C57BU6J. bg

k k b b d d

K ~ D ~ S

b

28 t 2 Of 1 2 t 1 3 k 1 2 t 2 0 2 0 Of 1

-I f 0 o + I

51 t 4 1 2 1

l o t 2 11 f 4 6 2 1 O f 1 0 t O o + 1 1 t 1

4 4 2 7 1 1 + 2 3 Y t 2 2 Y t 12 2 2 2 5 2 2 9 + 2

11+ 6 4 + 2

6 3 t 10 1.1 3 7 5 Y 2 2 t 6 2.0 1 5 t h 83 + 20 2.1 3 4 f . 5 81 + 15 2.8 5 3 t 38 31 + 10 13.8 3 f 1 32* 5 6.3 5 t 3 2 3 f 9 2.7 3 t 3 79 t 13 1.2 lf 5 59 f 24 13.3 2 f 1

60 f 13

8 2 t 3 91f 1 38f 6 23 f 10 2 2 t 5

22+ 9

3x+ 14

l X + 2 1 .

1.6 2.5 2.4

14.6 4.6 1 .3

11.1 13.8

1 .n

a) NK assays represent a single experiment with 5 replicate tubes per point. Other E : T ratios, and other experiments, confirmed the results. The frequency estimates of cytolytic clones represent pooled data from 2-5 separate experiments, with many trays of replicate cultures in each experiment. The approach is similar to Fig. 1 and 2 except that only day 9 of culture was assayed.

Eur. J. Immunol. 1984.14: 951-956 Loss of cytotoxic T lymphocyte clone specificity 955

Many mouse strains produced at day 9 of culture CTL clones showing cytolysis of P815 and EL4, and the “specificity ratios” suggested extensive or partial CTL clone specificity degradation. A few strains maintained relative specificity at this culture time. It was notable that the low NK, C57BL/6 beige mutant strain maintained high CTL specificity, whereas the highest NK strain, CBA, showed the most extensive CTL clone specificity loss. However, this correlation could not be made with all strains. Thus, BALB/c mice which maintained CTL clone specificity were NK positive (although low) whereas AKR mice, which showed CTL clone specificity degradation, had no detectable NK activity.

3.7 Genetic aspects of CTL clone specificity degradation

To determine if the tendency to lose or maintain CTL clone specificity in culture was determined by a simple dominant or recessive gene, the SC of the F1 offspring of CBA and BALB/c mice were cultured at LD along with SC of the parent strains. At 8 days the specificity index, assayed on P815 target cells, averaged 4.6 for the (CBA X BALB/c)FI, 1.3 for the CBA and 12.4 for the BALBlc. Thus, the F1 showed an intermediate level of specificity degradation. No backcross studies have been untertaken so far.

3.8 Contribution of responder and filler cells to CTL clone specificity degradation

One important issue was whether the tendency to CTL clone specificity degradation was determined entirely by the strain of origin of the single Ly-2’ precursor T cell itself, or by the strain of origin of the large numbers of irradiated filler cells responsible for the culture environment. This was tested by growing precursor cells on irradiated filler cells of a different strain, and testing CTL clone specificity at different times. It has been established that under such conditions the CTL are derived from the single precursor cell and not from the irradiated fillers [ 2 ] . A series of such tests for the CBA/BALB/c combination, assayed on P815, is summarized in Table 2.

As before, the CBA responders on CBA fillers combination showed low specificity at late culture times, whereas BALB/c responders on BALB/c fillers remained relatively specific. BALBic responders grown on CBA fillers gave a normal frequency of CTL clones, and these showed a definite specificity degradation, although less than CBA on CBA. The converse mix of CBA responders grown on BALB/c fillers also gave specificity degradation, again less than CBA on CBA fillers. Thus, all mixed cultures gave “intermediate” but definite specificity degradation similar to the genetic mix in the (CBA X BALB/c)F1 mice.

It could be argued that induction of specificity loss is due to certain soluble factors, and that the presence of the supernatants of Con A-stimulated rat cultures introduced another complication that might have confused the results. This supplement was added merely to maximize cloning efficiency [9] and CTL responses are obtained in its absence. Accordingly, the important experiments of Table 2 were repeated with cultures lacking any Con A supernatant supplement. The results, summarized in Table 3, were very similar. Clearly, CBA spleen irradiated filler cells or their products were able to induce loss of specificity in CTL clones of BALB/c origin.

Table 2. The contribution of responder cells and filler cells to the strain-related loss of specificity

Responder Filler Days of Frequency x 10’ of Ratio strain strain culture cytolytic clones”’ + PHAI

no PHA with PHA - PHA

CBA CBA 7 2 2 + 4 6 1 f 8 2.8 8 4 5 f 9 66f 8 1 .s 9 4 9 + 6 5 9 f s 1.2

BALBlc BALBIc 7 5 f l 4 6 f 1 1 9.8 8 2 + 3 38+ 6 15.8 9 4 + 1 34+ 5 9.0

CBA BALBIc 7 11?2 4 9 2 16 4.5 8 1 9 2 9 34+10 1.8 9 2 5 + 7 4 6 f 9 1.8

BALBIc CBA 7 9 f 2 3 5 5 9 3.9 8 1 8 f 5 3 0 f 2 1.7 9 1 7 f 2 2 9 f 4 1.7

The results are means of 9 experiments (CBA on CBA), or 5 experiments (BALB/c on BALBlc) or 4 experiments (all mixed cultures).

Table 3. Strain-related loss of specificity: contribution of responder and filler cells in the absence of Con A supernatant supplement”)

Responder Filler Frequency x 10’ of Ratio strain strain cytolytic clones + PHA/

no PHA with PHA -PHA

CBA CBA 42k 16 46+ 15 1 . 1 CBA BALBlc 28+ 16 57+25 2.0 BALBlc BALBlc 9 f S 1 1 8 f 7 2 12.4 BALBlc CBA 3 2 2 16 6 7 k 2 4 2.1

a) Cultures were set up as in Table 2 except the supplement of Con A-stimulated rat spleen supernatant was omitted. Results are the means of three separate experiments all assayed at day 8 of culture. Separate experiments comparing BALBic responders grown on either BALBIc or CBA fillers and assayed at day 9 of culture gave very similar results.

However, filler cells alone did not dictate the specificity degradation effect, since any culture containing CBA cells, as responders or fillers, showed a high level of nonspecific lysis at later times.

4 Discussion

These results should be considered in relation to two separate issues. One is the practical requirement to devise LD culture assays where CTL clone specificity is maintained, so that the frequency of precursor cells of a given antigenic and H-2 restriction specificity can be determined reliably. The other is the theoretical question of the cause of CTL clone specificity degradation and the nature of the CTL target cell interaction during nonspecific cytolysis.

It is evident that choice of BALB/c or C57BU6J.bg instead of CBA mice for LD precursor frequency studies could effec-


tively bypass the specificity degradation problem in our culture system. Even C57BLi6 cultures probably could be used if cytotoxic assays were performed around day 6 of culture with an appropriate choice of target cell. Other laboratories which have not encountered this problem may already have made a fortunate, if fortuitous, choice of mouse strain. However, since BALBlc-derived cytotoxic clones can be made capable of nonspecific cytolysis under some conditions (as our mixed cultures with CBA fillers demonstrate), there is no guarantee that other BALB/c mouse colonies (if, for example, not specific pathogen free), or longer-term BALB/c CTL clones, will maintain absolute specificity. This solution is, therefore, not entirely satisfactory, and it may be necessary to understand the basis of the effect in order to be sure of avoiding it.

The rules which determine whether CTL clone specificity will be maintained or lost are not clear from these results. The target cell used has some effect, but the mouse strain used to provide the cultured cells appears to be more important. There is no consistent correlation with H-2 haplotype, nor with cloning efficiency or the relative cytolytic activity of clones assayed in the presence of PHA. The correlation with NK status is striking with certain strain comparisons, but is not consistent .

The most important observation, which sets the direction of future work, is the transfer of the tendency to show CTL specificity degradation to BALB/c clones when they are grown on CBA spleen irradiated filler cells. An immediate explanation would be the production of a soluble lymphokine (such as a high level of a particular differentiation factor) by the irradiated, Con A-stimulated SC. Recent results by Brooks et al. [6, 71 on the reversible induction by high levels of growth factors of NK-like cytolysis in cloned CTL lines would be in line with this type of explanation. If this is the case, developing CBA CTL clones must also be able to make the factor, since they show specificity degradation when grown on the “defi- cient” BALB/c irradiated fillers. An alternative explanation is the transfer of an infectious agent endogenous to CBA cells,

such as a particular retrovirus. A viral etiology could also ex- plain the peculiar strain distribution of the phenomenon.

In any case the activation of the direct cytolysis response in BALB/c-derived LD CTL clones should provide a useful assay in attempts to define the causative or triggering influence. BALBk cells cultured on BALBic spleen fillers under these conditions, produce active, specific CTL clones, so any growth or differentiation factors required for normal CTL development are already provided. Induction of nonspecific lysis should, therefore, be a selective assay for just the factors caus- ing the switch to direct, nonspecific cytolysis. These studies are in progress.

Roland Scollay provided lively discussion and critical comment on the results and ideas presented in this study, and Rita Lang provided skilled assistance in several of the experiments.

Received April 18, 1984.

S References

1 Shortman, K., Wilson, A. , Scollay, R. and Chen, W.-F., Proc.

2 Shortman, K., Wilson, A. and Scollay, R., J. Immunol. 1984. 132:

3 Teh, H.-S., Bennink, J . and Von Boehmer, H., Eur. J. Immunol.

4 Teh, H.-S. and Yu, M. , J. Immunol., in press. 5 Moretta, A,, Panteleo, G., Moretta, L., Cerottini, J. C. and Min-

6 Brooks, C. G., Urdal, D. L. and Henney, C. S., lmmunol. Rev.

7 Brooks, C. G., Nature 1983. 305: 155. 8 Klein, E., Immunol. Today 1980. I : iv. 9 Wilson, A,, Chen, W.-F., Scollay, R. and Shortman, K., J.

10 Chen, W.-F., Wilson, A,, Scollay, R. and Shortman, K., J.

11 Shortman, K. and Wilson, A , , J . Immunol. Methods 1981.43: 135.

Natl. Acad. Sci. USA 1983. 80: 2728.

584.

1982. 12: 887.

gari, M. C. , J . Exp. Med. 1983. 157: 743.

1983. 72: 218.

Irnmunol. Methods 1982. 52: 283.

Irnmunol. Methods 1982. 52: 307.

Degradation of specificity in cytolytic T lymphocyte clones: Mouse strain dependence and interstrain...

Documents

Transcript of Degradation of specificity in cytolytic T lymphocyte clones: Mouse strain dependence and interstrain...