Vol. 18, No. 2 Printed in U.S.A.

INFECTION AND IMMUNITY, Nov. 1977, p. 377-385 Copyright © 1977 American Society for Microbiology

Mycoplasma-Dependent Activation of Normal Lymphocytes: Induction of a Lymphocyte-Mediated Cytotoxicity for Allogeneic and Syngeneic Mouse Target Cells K. E. ALDRIDGE,' B. C. COLE,1.2* AND J. R. WARD' Division ofArthritis, Department of Medicine,' and Department of Microbiology,2 University of Utah College of Medicine, Salt Lake City, Utah 84132

Received for publication 25 April 1977

Mycoplasma arthritidis, M. hominis, and M. arginini were tested for their ability to induce a cytotoxic response from normal CBA mouse lymphocytes against 5"Cr-labeled allogeneic and syngeneic target cells. In most cases, the mycoplasmas alone were not toxic for the target cells. Furthermore, the mycoplasmas did not result in decreased lymphocyte viability but, in fact, contributed to enhanced lymphocyte survival. In the absence of normal CBA lymphocytes, mycoplasmas alone did not induce a significant amount of cell damage in either the allogeneic or the syngeneic target cells. Strains of M. arthritidis and M. hominis, when added to the lymphocyte-target cell mixtures, induced statistically significant increases in 51Cr release from both target cell types at each assay period after 6 h. The release of 5"Cr was taken as a measure of cell death. M. arginini induced only low levels of cytotoxicity or none at all. Both arthritogenic and non-arthritogenic strains of M. arthritidis induced the cytotoxic response. The degree of cytotoxicity produced was directly related to the size of the initial inoculum. The presence or absence of serum in the culture medium did not contribute significantly to the cytotoxicity response.

Mycoplasmas are capable of causing arthritis in a wide variety of animal species, thus providing models for the study of both acute and chronic inflammation. Rats develop an acute suppurative arthritis after intravenous injection of Mycoplasma arthritidis (17, 33). Although cartilage and bone damage is seen with ankylosis, the disease resolves after 50 to 60 days and viable mycoplasmas can no longer be isolated from the tissues. The animals, which develop high levels of complement-fixing antibody, become completely immune to reinfection despite the absence of metabolic inhibiting and opsonizing antibodies, which usually contribute to host defense mechanisms (8, 12). Of particular interest is the ability of M. arthritidis to produce a polyarthritis of mice that progresses to a more chronic phase characterized histologically by massive synovial villus hyperplasia with mononuclear and plasma cell infiltration, fibroplasia, increased vascularity, and cartilage and bone destruction (15). Viable mycoplasmas could be recovered from some arthritic joints 217 days postinoculation (15). Other studies have indicated that the arthritis can persist for the life of the animal (unpublished results). Both Swiss and CBA mice produced substantial levels of persisting circulating complement-fix377

ing antibody after injection with M. arthritidis. Persisting metabolic inhibiting antibodies were detected in CBA mice (10), but only a low transient response was seen in Swiss mice (14). These observations may relate to the greater susceptibility of Swiss mice to M. arthritidis. Although the mechanisms of the inflammatory response are not known, a cell-mediated immune reaction has been detected in mice at all stages of the disease process. A correlation between severity of arthritis and the humoral or cell-mediated immune responses of individual mice has not been found. A correlation was seen in that mycoplasmas were recovered more frequently from those animals that exhibited the most severe arthritis (10). However, cell-mediated immune phenomena are believed to play a role in the pathogenesis of respiratory lesions induced by both M. pulmonis (19; G. H. Cassel and J. R. McGhee, Abstr. Annu. Meet. Am. Soc. Microbiol. 1976, D65, p.62) and M. pneumoniae (20, 31). Studies were thus initiated to provide more information on the role of cell-mediated immune phenomena in the inflammatory reactions induced in mice by M. arthritidis. The present report describes and documents the unexpected finding that M. arthritidis and M. hominis in-

378

ALDRIDGE, COLE, AND WARD

duce normal, unsensitized lymphocytes to exert cytotoxic effects in vitro on allogeneic and syngeneic mouse fibroblast target cells. MATERIALS AND METHODS Mycoplasma strains and mycoplasma culture procedures. The sources of M. arthritidis 158plOp9, 14124plO iv, PG-6, PN, Campo, and Sandoz L-4 were previously described (15, 16, 22). M. arthritidis strain 20-P, reportedly isolated from a patient with rheumatoid arthritis (24), was supplied by E. Jansson (University of Tampere, Finland). M. arthritidis M715011-084 (NIH), M. hominis 711-002-084 (NIH), and M. arginini 732-001-084 (NIH) were obtained from the National Institutes of Health, Bethesda, Md., courtesy of M. F. Barile. Each organism was grown in modified Hayflick medium consisting of PPLO broth or agar (Difco Laboratories, Detroit, Mich.) supplemented with final concentrations of 15% (vol/vol) inactivated horse serum, 5% (vol/vol) fresh yeast extract, and 1,000 U of penicillin G per ml (6, 23). Broth cultures were harvested in the log phase of growth 3 to 5 days postinoculation. The cultures were divided into 1- to 2-ml portions and frozen to -70°C until used. Random tubes of frozen cultures were thawed and titered by preparing serial 10-fold dilutions in mycoplasma broth and plating 0.01 ml in triplicate on mycoplasma agar. After 3 to 5 days of growth on agar, the colony-forming unit (CFU) titer per milliliter was determined by averaging the number of colonies in the highest dilution showing 10 to 50 colonies per 0.01 ml and multiplying by the dilution factor and by 100. For experimental use, the organisms were rapidly thawed at 37°C and diluted in the appropriate tissue culture medium. Target cells. Allogeneic L-cells (L-929, Microbiological Associates, Inc., Bethesda, Md.) and syngeneic CBA mouse embryo fibroblasts (CBA-MEF) were used as target cells for cytotoxicity experiments. Each culture was maintained in serial passage with medium RPMI 1640 (Flow Laboratories, Inc., Inglewood, Calif.) containing 10% (vol/vol) inactivated fetal calf serum (iFCS; Microbiological Associates) and 200 U of penicillin G per ml at a pH of 7.2 to 7.4 with ad libitum medium changes for maintenance. Continuous passage of L-cells was carried out without regard to the number of previous passages. CBA-MEF were aseptically prepared from 10- to 15-day pregnant female CBA/J mice by removing the fetuses, dissecting away the head, appendages, and viscera, and mincing the remaining tissue. Single-cell suspensions were prepared by incubating the minced tissue in 0.25% (wt/vol) trypsin (Difco) in calcium- and magnesiumfree phospate-buffered saline on a magnetic stirrer at 37°C for 15 to 30 min. The suspension was centrifuged at 200 x g for 5 min, washed once in 0.83% (wt/vol) NH4Cl to lyse erythrocytes, washed once in complete RPMI, and finally suspended in complete RPMI. Plastic culture flasks (75 cm2) were seeded with 5 x 107 viable cells in 20 ml of complete RPMI and were incubated at 37°C. To minimize any possibility of significant antigenic surface change, CBA-MEF were only carried through 12 passages and then discarded. New monolayers were then prepared from fetuses as described above.

INFECT. IMMUN. Radioactive labeling of cells for cytotoxicity experiments was done with sterile sodium chromate (Na5"CrO4) in isotonic saline (New England Nuclear Corp., Boston, Mass.). L-cells were labeled with 200 to 300 MlCi of Na51CrO4 per 107 cells, whereas CBAMEF were labeled with 300 to 400 ,Ci of Na5'CrO4 per 107 cells. Labeling was carried out by mixing trypsinized and washed monolayer cells in complete RPMI with 51Cr in a total volume of 1 to 1.5 ml and incubating at 37°C for 1 h. The cells were then washed three times in 4-ml portions of complete RPMI and suspended to give 5 x 104 cells per ml in complete RPMI. Cells labeled in this manner gave reproducible counts of 0.5 cpm per cell. Preparation of lymphocyte suspensions. Normal CBA/J female mice, 10 to 15 weeks old, were sacrificed by cervical dislocation. The spleen and mesenteric lymph nodes were aseptically removed, placed in a petri dish with 5 to 10 ml of complete RPMI, and gently teased apart with forceps and a scalpel. The suspensions were filtered through 80-mesh screen to remove tissue fragments and debris. The cells were then centrifuged at 200 x g, washed once with tris(hydroxymethyl)aminomethane-buffered 0.83% (wt/vol) NH4CI (pH 7.2) to lyse erythrocytes, washed once with complete RPMI, and resuspended to 107 viable lymphocytes per ml. The trypan blue method of determining cell count and viability was used (26). The lymphocytes from individual animals were treated separately and not pooled with other suspensions. Differential Wright staining of 12 lymphocyte suspensions yielded 87 to 100% lymphocytes, 0 to 3% monocytes, 0 to 5% polymorphonuclear leukocytes, and 0 to 5% blast cells. Lymphocyte survival. To determine the effect of viable and nonviable mycoplasmas on normal CBA lymphocytes, M. arthritidis 158p10p9 was used. Lymphocyte suspensions were made up to 107/ml in complete RPMI as described above. One milliliter of lymphocytes was added to 9 ml of complete RPMI containing 106 CFU of viable mycoplasmas per ml. For controls, lymphocytes were added to complete RPMI without mycoplasmas. At 24-h intervals thereafter, a portion of cells was removed from each flask and viability was determined as before. Cytotoxicity assay. As illustrated in Fig. 1, target cells labeled with 5"Cr (104 cells in 0.2 ml of RPMI medium) were added to sterile microtiter welLs (Falcon Plastics, Oxnard, Calif.) in groups of 12. The cells were incubated at 37°C in 10% CO2 and humidity for 20 to 24 h to facilitate attachment. The plates were then inverted on sterile absorbent pads, and the plating medium was drained off. All wells were then washed by adding 0.1 ml of complete RPMI medium to each well and again inverting the plates to allow drainage. Each test mycoplasma (106 CFU) suspended in 0.1 ml of complete RPMI was then added to 12 wells. Lymphocytes (106 in 0.1 ml of medium) prepared as described above were then added to each well. Control groups, of 12 wells each, consisted of: (i) target cells and lymphocytes, but no mycoplasmas (groups 13 to 15, Fig. 2); (ii) target cells and mycoplasmas, but no lymphocytes (groups 4, 8, 12, Fig. 2); and (iii) target cells alone (spontaneous release; group 16, Fig. 2). When all mixtures were complete, the plates were placed in 10% CO2 and humidity at 37°C. This point

MYCOPLASMA-MEDIATED LYMPHOCYTOTOXICITY

VOL. 18, 1977

Ir.. L

--i Harvest spleern and lymph nodes

Target cell monoloyer

jTrpsinize, wash and label with 5 Cr for hour.

l04 lobelled target cells in appropriate wells (filled circles). Incubate 18-24hrs. Wash I x. Add medium or medium with 106 mycoplasmas. Total release (TR) receives trypsin.

Lyse RBC's and Add 106 lympho:ytes to appropriate wells. ncr ubate at 37° C.

24

mycoplasma preparation. In all experiments, media, lymphocyte suspensions, target cells, and viable mycoplasma suspensions were cultured on mycoplasma agar to determine the presence or absence of mycoplasmas. Also, at each sample time in all experiments, one well from each triplicate was cultured after removing the supernatant, by adding 2 to 3 drops of mycoplasma broth, agitating and aspirating the broth with a Pasteur pipette, and placing it on mycoplasma agar. Positive mycoplasma cultures were obtained in every case that viable mycoplasmas were introduced into the test suspension but were never found in the accompanying controls.

wash.

4l

\(TtTR

51 Collect samples and measure Cr release in y-counter

FIG. 1. Stepwise representation of the cytotoxicity The numbers 6, 18, 24, and 48 represent the time (hours) of harvest of each group. RBC's, erythassay.

rocytes.

considered time 0. After 6, 18, 24, and 48 h of incubation, 3 of the 12 wells in each group were harvested by tilting the plates and withdrawing the supernatants with capillary pipettes drawn out to fine points. The supernatants were transferred to separate counting vials (Biovials; Beckman Instruments, Inc., Fullerton, Calif.). Determination of total releasable 5"Cr was performed by adding, at time 0, 0.2 ml of 0.5% (wt/vol) trypsin to five wells containing labeled target cells alone. After 48 h, the contents of each well were pooled with three washings from each well and transferred to five individual counting vials. All vials were assayed for 51Cr with a crystal scintillation spectrophotometer (Biogamma; Beckman Instruments), using 1-min counts per sample. The percentage of 51Cr release above background was determined using the following fonnula: percent 51Cr release = [(experimental release - spontaneous release)/(total release - spontaneous release)] x 100. All experimental results were then compared with controls with Student's t test (29). As illustrated in Fig. 2, a typical experiment used lymphocytes from three separate mice treated with three different mycoplasma preparations. Sixteen groups, plus a total-release control, are shown with four groups on each microtiter plate. To minimize cross-contamination, each plate contained only one was

379

RESU,LTS Effect of mycoplasmas on target celis. In experiments with 51Cr-labeled target cells, there was a gradual increase of 51Cr in the supernatant nonspecifically (spontaneous release) released from healthy cells, which constituted the background. During the first 18 to 24 h, spontaneous release was about 1%/h for L-cells and CBAMEF and thereafter decreased to about 0.5%/h. All of the mycoplasmas were tested for cytotoxicity against target cells alone. The amount of 61Cr release by L-cells in the presence of mycoplasmas was essentially the same as the background (spontaneous release). Viable mycoplasmas could be cultured at each harvest time. However, certain mycoplasmas did produce target cell damage on CBA-MEF (Fig. 3). M. arthritidis H606 was highly toxic for CBAMEF, resulting in a 49% 51Cr release after 24 h. The cytotoxicity assays with CBA-MEF and this mycoplasma are, therefore, not reported. M. arthritidis strains PG-6 and Sandoz L-4 were also toxic for CBA-MEF, as evidenced by a 14 and 13.2% 51Cr release after 48 h of incubation. HowPLATE I GROUP GROUP 2 GROUP 3 GROUP 4

-

-

TARGET TARGET TARGET TARGET

CELLS CELLS CELLS CELLS

LYMPHOCYTES MOUSE + MOUSE 2 LYMPHOCYTES + MOUSE 3 LYMPHOCYTES + MYCOPLASMA X +

+

MYCOPLASMA X

+

MYCOPLASMA X

+

MYCOPLASMA X

PLATE 2 GROUP GROUP GROUP GROUP

5

-

6 7 8

-

GROUP 9 GROUP 10GROUP 11GROUP 12-

GROUP GROUP GROUP GROUP GROUP a

b

TARGET CELLS + MOUSE I LYMPHOCYTES TARGET CELLS + MOUSE 2 LYMPHOCYTES TARGET CELLS + MOUSE 3 LYMPHOCYTES TARGET CELLS + MYCOPLASMA Y

TARGET TARGET TARGET TARGET

CELLS CELLS CELLS CELLS

PLATE 3 MOUSE I LYMPHOCYTES + MOUSE 2 LYMPHOCYTES + MOUSE 3 LYMPHOCYTES +

+

+

MYCOPLASMA Y

+

MYCOPLASMA Y MYCOPLASMA Y

+

+

+ +

MYCOPLASMA Z MYCOPLASMA Z MYCOPLASMA Z

MYCOPLASMA Z

PLATE 4 LYMPHOCYTES 13- TARGET CELLS + MOUSE 14- TARGET CELLS + MOUSE 2 LYMPHOCYTES 15- TARGET CELLS + MOUSE 3 LYMPHOCYTES 16a-TARGET CELLS ONLY 17b-TARGET CELLS + TRYPSIN

SPONTANEOUS RELEASE TOTAL RELEASE

FIG. 2. Representation of cytotoxicity

assay.

380

ALDRIDGE, COLE, AND WARD

INFECT. IMMUN.

CBA-MEF injected with: * M.arth. PG-6 * M.arth. 158p10p9 v M. arth. PN * M. arth. 14124plO I.V. o M.arth. Sandoz L-4 * M.arth M715-011o M. arth. 20-P 084 (NIH) v M. hominis 711-002o M.arth. Campo 084 (NIH) x M.orth. H606 A& M.orqinini 732-001084(NIH)

mycoplasmas, there was a significant progressive rise in 5"Cr release, indicating target cell death. Thus, after 48 h, 18.8% 51Cr release was observed. Since this is an allogeneic system, it would be expected that histocompatibility antigens would cause increasing lymphocyte-mediated cytotoxicity. However, the addition of mycoplasmas at a 1:1 ratio with lymphocytes caused a significant increase in the amount of 5"Cr released (Fig. 5, Table 1). M. hominis and all strains of M. arthritidis tested showed this effect dramatically (5"Cr release values ranged from 51.2 to 63.7% at 48 h), whereas M. arginini induced a weaker, but still significant, response (31.3% release at 48 h). Interestingly, at the 6-h test time, none of the mycoplasmas tested, except M. arthritidis Sandoz L-4, showed significant differences from controls (viable mycoplasmas). At all subsequent sampling times, each mycoplasma elicited a significant cytotoxicity response. Mycoplasma enhancement of cytotoxicity with CBA-MEF as targets. In contrast to the allogeneic system, the addition of normal CBA lymphocytes to 51Cr-labeled CBA-MEF resulted in no significant killing of cells (Fig. 6, Table 2). A maximum release of only 3.3% above background was seen in this syngeneic system. The addition of 106 CFU of the M. arthritidis strains or M. hominis produced a marked increase in 5Cr release above controls, which was first detected at 18 h and progressed to maximum values of 20.5 to 51.3% 51Cr release at 48 h. The P values obtained with these organisms were all