H-2 antigens incorporated into phospholipid vesicles elicit specific allogeneic cytotoxic T lymphocytes

CELLULAR

IMMUNOLOGY

55,

328-341 (1980)

H-2 Antigens Incorporated Specific Allogeneic

into Phospholipid Vesicles Cytotoxic T Lymphocytes

ARTHUR Department

H. HALE

of Microbiology and Immunology, Bowman Gray School of Medicine Wake Forest University, Winston-Salem, North Carolina 27103 Received

August

Elicit

30, 1979; Accepted

February

of

8, 1980

Different H-2 antigen-containing subcellular fractions were tested for their ability to elicit specific anti-H-2 cytotoxic T lymphocytes (CTLs). Intact cells and membrane vesicles were capable of eliciting strong anti-H-2 primary and secondary CTL responses. However, detergent-solubilized H-2 antigens partially purified with lentil lectin were greatly reduced in their capacity to elicit secondary anti-H-2 (CTLs) and at all amounts tested did not elicit a primary CTL response. Lentil lectin-purified H-2 antigens incorporated into egg lecithin plus cholesterol (30% w/w) vesicles elicited a strong secondary anti-H-2 CTL response and a low but significant primary anti-H-2 CTL response. The results also indicate that T-cell-defined specificities are closely associated with serologically defined private and public specificities.

INTRODUCTION The major histocompatibility antigens (H-2 in mice and HLA in humans) have been implicated as both stimulators and targets for syngeneic and allogeneic cytotoxic T lymphocytes (CTLs).’ More specifically, the H-2K and/or H-2D ends of the major histocompatibility complex (MHC) in mice are responsible for the recognition process which occurs between thymus-derived cells (T cells) and stimulator or target cells (1). The evidence to date is compelling that the H-2 antigens recognized by alloantisera are closely associated to those recognized by T cells (2-8). However, the evidence is by no means conclusive as to the identity of the H-2 antigens recognized by T cells (9-14). Recently a number of laboratories have shown that detergent-solubilized H-2 antigens and reconstituted membrane vesicles are capable of specific elicitation of secondary CTL responses (6, 15- 19). One limitation in the continuation of this work has been the observation that detergent-solubilized H-2 antigens elicit anti-H-2 secondary CTLs much less efficiently than intact cells or membrane vesicles (15). Also, the inability to elicit primary anti-H-2 CTLs with detergentsolubilized fractions has resulted in the speculation that the minimal molecular 1 Abbreviations used: Con A, concanavalin A; cpm, counts per minute; CTL, cytotoxic T lymphocytes; DOC, deoxycholate; E:T ratio, effector:target ratio; ip intraperitoneal; MHC, major histocompatibility complex; and PBS, phosphate-buffered saline (0.15 M NaCl, 0.01 M potassium phosphate, pH 7.4). 328 0008-8749/80/140328-14$02.00/O Copyright All rights

0 1980 by Academic Press, of reproduction in any form

Inc. reserved.

ELICITATION

OF CYTOTOXIC

T LYMPHOCYTES

329

requirements for primary CTL responses differ from secondary CTL responses. In this communication, we report the specific elicitation of anti-H-2 primary and secondary CTLs with partially purified H-2 antigens incorporated into egg lecithin plus cholesterol liposomes. MATERIALS

AND METHODS

Mice. Male and female mice of the following strains were used: BALB/c AnN (d, d), BALB.B (b, b), BALB.K (k, k), BALB.HTG (d, b), BlO.ASR (b, d), and C57/BL 6 (b, b). These were either DBA/2(d,d)C3H.OH(d,k),BIO.A(k,d), purchased from Cumberland View Farms (Clinton, Tenn.) or produced in our own breeding colony from breeding stock obtained from Herman Eisen (Massachusetts Institute of Technology). Letters in parentheses indicate the H-2K and H-2D alleles. Alloantiserum. Anti-H-2d serum was raised by multiple intraperitoneal (ip) injections of 2.0 x lo7 BlO.D2 spleen cells into BlO mice. Anti-H-2b alloantiserum was raised by multiple ip injections of 2.0 x lo7 B 10 spleen cells into BlO*D2 mice. Other antisera used in this work are anti-H-2.4 (private H-2Dd), anti-H-2.31 (private H-2Kd), anti-H-2.2 (private H-2Db), anti-H-2.33 (private H-2Kb), and anti-H-2.6, 27, 28, 29 (public H-2Kd, H-2Kb, H-2Dd, H-2Db). These antisera were produced and tested as described previously (9). Antisera for public and private specificities were generously provided by Herman Eisen. Anti-thy 1.2 serum and guinea pig complement was purchased from Litton Bionetics (Kensington, Md.). Media. RPM1 1640 (90 ml) was supplemented with 1.0 ml each of minimum Eagle’s medium nonessential amino acids (100X), sodium pyruvate (Gibco, 100X), L-glutamine (3%), and with 1.5 ml of 1 M N-2-hydroxyethyl-N’-2-ethanesulfonic acid, pH 7.4 (Sigma Chemical Co., St. Louis, MO.); penicillin, streptomycin, 2-mercaptoethanol, and heat-inactivated fetal calf serum (.56”C, 45 min, Flow Laboratories, Cockeyesville, Md.) were then added to final concentrations of 2.5 x lo4 units/liter, 25 mg/liter, 50 a, and 5% (v/v), respectively. Tumors. P815 (H-2d) and EL-4 (H-2b) were maintained by weekly ip transfers of 1.0 x 10fi ascites cells in mice of the strain of origin, DBA/2 and C57/ BL6, respectively. Production of CTLs. Secondary anti-H-2 CTLs were elicited using responder spleen cells primed 2-6 months earlier with 2.0 x lo7 allogeneic spleen cells introduced by intraperitoneal injections. Suspensions of viable primed or normal spleen cells (responder and stimulator cells) were treated with NH&l to remove red blood cells (13). Stimulator tumor cells (1.3 x 105) that had been irradiated with 10,000 R (X ray, Picker) were suspended in 1.0 ml of supplemented RPM1 1640 (13) and plated with 8.0 x lo6 responder cells, also in 1.0 ml of supplemented RPM1 1640, in 1.7 x 1.6-cm wells (Linbro, Hamden, Conn.). RPM1 1640 was supplemented with 10% fetal calf serum (heat inactivated, 56”C, 45 min), 0.03% glutamine, 2-mercaptoethanol (50 pm), and penicillin and streptomycin (Flow Laboratories, Inc., McLean, Va.) (13). When solubilized H-2 antigens, membrane vesicles, or H-2 antigens incorporated into liposomes were used to elicit an anti-H-2 response, the H-2 antigen-containing fractions were added to 8.0 x IO6 normal or primed spleen cells in 1.0 ml of

330

ARTHUR

H. HALE

supplemented RPM1 1640 (13) and incubated with the cells for 1.O hr with continual agitation. After 1.0 hr the cells were diluted to a final volume of 2.0 ml and plated in 1.7 x 1.6-cm wells (Linbro). After 5 days, in an atmosphere of 6% Co, and 94% air, the cells were harvested and washed by centrifugation (45Og), counted, and resuspended in supplemented RPM1 1640 at various cell concentrations for the cytotoxicity assay. Viability of all cells was determined by trypan blue exclusion. Target cells. BALB/c, BALB*B, BALB .K, BALB.HTG, and BlO*ASR spleen cells were stimulated with concanavalin A (Con A) by incubating 5.0 x lo7 spleen cells with 2.0 pg of Con A (Sigma) in 5 ml of supplemented RPM1 1640 (37°C 6% COZ, 94% air). After 2-3 days of incubation, cells were harvested, counted, and viability was determined by exclusion of trypan blue. Con A-stimulated cells were then incubated with 200 &i of Na2Wr0, (New England Nuclear, Boston, Mass.) in 0.5 ml of supplemented RPM1 1640 for 1.5 hr at 37°C in an atmosphere of 6% Co,, 94% air (13). Cytotoxic assays: Cytotoxic cell-mediated assays. 51Cr-labeled target cells (1.0 x 104) are mixed with serial dilutions of effector spleen cells in a total volume of 200 ~1 of supplement RPM1 1640. This mixture was incubated 4-6 hr at 37°C in an atmosphere of 6% CO2 and 94% air. The assay was stopped by diluting with cold phosphate-buffered saline and the amount of radioactivity in the supernatant and pellet was determined (13). Percentage specific release (Spec. Rel.) was calculated as 100 (E-C/l-C) where E is the fraction 51Cr released by antigen stimulated effector cells and C is the fraction of Yr released by a mockstimulated responder population. Antibody-mediated cytotoxic assays were carried out essentially as previously described (13). Inhibition of anti-H-2 antibody plus complement-mediated lysis by cells or subcellular fractions was performed as described previously (13). In order to evaluate the serologically defined H-2 content of different fractions relative to intact cells more quantitatively, the 50% level of inhibition was used as the end point of titration. For P815 the number of cells needed to inhibit 50% lysis varied between each experiment and ranged between 3.0 x lo5 and 5.0 x lo5 cells; EL-4 ranged from 3.2 x lo5 to 5.5 x lo5 cells. The inhibition by H-2 antigencontaining fractions was expressed in terms of cell equivalents per microgram of protein. The number of intact tumor cells which were required to give 50% inhibition of antibody plus complement mediated lysis was divided by the amount of protein (as determined by the method of Lowry (20)) of an H-2 antigencontaining fraction required to neutralize 50% of the antibody plus complementmediated lysis. Partial purification of H-2 antigens. Purified plasma membranes from P815 and EL-4 tumor cells were obtained as described by Esko et al. (21). For both cell types, the resultant plasma membranes were enriched between 15- and 20fold as indicated by the increase in specific activity of Na+/K+ dependent ATPase and serologically defined H-2 antigens (inhibition of anti-H-2 plus complementmediated lysis). Antigens were solubilized by incubating the plasma membranes at room temperature for 10 min with sodium deoxycholate (DOC) (Sigma) in phosphatebuffered saline (PBS (13)). The detergent extracts were then centrifuged 30 min at 100,OOOg. For maximum solubilization with deoxycholate, a detergent:membrane

ELICITATION

OF CYTOTOXIC

T LYMPHOCYTES

331

protein ratio (w/w) of 4 resulted in solubilization of >90% of the serologically defined H-2 antigens with a final protein concentration of 500 pg/ml. Sodium deoxycholate-solubilized H-2 antigens were further purified by passing the resulting 100,OOOg supematant over a Lens culinaris-conjugated Sepharose 4b column equilibrated in 0.2% deoxycholate in PBS. A final purification of the H-2 antigens of between 120- and 160-fold was obtained by these procedures. This material was dialyzed 36 hr against phosphate-buffered saline at 4°C before use in the elicitation of anti-H-2 CTLs and is referred to as high-density H-2 antigens (see below). Reconstitution of H-2 antigens into liposomes. Egg lecithin and cholesterol were purchased from Calbiochem (San Diego, Calif.). In a typical preparation, 200 pg of egg lecithin plus cholesterol (30% w/w) dissolved in chloroform was dried in the form of a film under N,. This film was then dissolved into a solution containing partially purified H-2 antigens (0.2% DOC) at a lipid to protein ratio of 1: 1. This solution was then dialyzed for 36-48 hr against PBS at 4°C. The dialysate was recovered as an opalescent solution with 280% of the protein associated with the vesicles. Characterization of the partially purified H-2 antigen containing fractions. Partially purified H-2 antigens and liposomes containing partially purified H-2 antigens were analyzed by flotation isopycnic centrifugation (5-40% w/w sucrose density gradients). H-2 antigen-containing fractions were mixed with 60% (w/w) sucrose in PBS to give a final sucrose concentration of 40% (w/w). A linear density gradient of sucrose (40-5%) was layered on top of this sample. After 14 hr of centrifugation at 200,OOOg (4”C), fractions were collected and analyzed for the presence of H-2 antigens (capacity of a fraction to inhibit anti-H-2 plus complement lysis), for protein content (Lowry et al. (20), and for radioactivity. In many of the experiments performed during this study, tumor cells (P815 and EL-4) were labeled with [3H]leucine (80 Ci 1 Ci = 3.7 x 10’” Bq)/mmol; New England Nuclear) for 8.0 hr at 37°C in leucine-free RPM1 1640 supplemented with heat-inactivated fetal calf serum dialyzed for 24 hr against PBS. To help characterize the lipid content of the H-2 antigen-containing fractions, [14C]phosphatidyl choline (1.5 Ci/mmol; New England Nuclear) was mixed with egg lecithin to a final specific activity of 0.5 &i/mg of lipid. This material was then used to form liposomes as described above. When liposomes were analyzed by flotation isopycnic centrifugation, it was evident that a small amount of protein was not incorporated into the liposomes. This protein fraction had a density > 1.12 g/cm3 and usually represented ~20% of the total protein. Because of this contamination of the liposome fraction with this high-density protein component, we further purified the H-2 antigen-containing liposomes by flotation isopycnic centrifugation as described above. High-density H-2 antigen and liposomes were collected from the 40-5% sucrose gradients and before analyzing the specificity and effectiveness of these fractions to elicit anti-H-2 CTLs, fractions were diluted 1:5 with PBS and centrifuged at 100,OOOg for 8.0 hr. The pellets were resuspended, protein determinations were done (Lowryet al. (20)) and appropriate dilutions were incubated with primed spleen cells as described above. No loss of H-2 antigen activity was observed by this method of concentrating the different fractions and removing the sucrose. Detergentsolubilized H-2 antigens are referred to as high-density H-2 antigens because, after dialysis against PBS, this protein fraction has a density greater than 1: 14 g/cm3.

ARTHUR

H. HALE

s

4

8

12

16

20

Fraction

FIG. 1. Flotation isopycnic centrifugation of dialyzed lentil lectin-purified H-2 antigens. Partially purified H-2d antigens from P8 15 cells labeled with [3H]leucine were dialyzed against PBS and suspended in sucrose (final concentration 40% sucrose, w/w). A 40-5% sucrose gradient was layered over the sample and spun at 200,OOOg for 14 hr. The gradients were fractionated and tested for H-2d cell equivalents (A) and radioactivity (B).

Liposomes (density between 1.11 and 1.06 g/cm3 containing H-2 antigens were defined by the presence of lipid ([14C]phosphatidylcholine), protein (presence of [3H]leucine or by direct measurement of protein, Lowry et al. (20)), and H-2 antigen content. Removal of serologically dejked H-2 specificities. To evaluate the relationship between serologically defined H-2 specificities and H-2 antigens which were responsible for elicitation of the secondary anti-H-2 CTL responses, we preincubated detergent-solubilized lentil lectin-purified H-2 antigen with either normal mouse serum, anti-H-2, anti-private H-2, or antipublic H-2 serum at a final dilution of 1:50 or 1:25 at 4°C for 30 min. Solubilized H-2 antigen fraction (0.1 ml) plus serum was incubated with 0.1 ml (packed volume) ofStaphylococcus A protein conjugated to Sepharose 4b beads (Pharmacia, Rahway, N.J.) for 10 min. The Sepharose 4b beads were removed by centrifugation (SOOg, 5 min) and the supernatant was mixed with egg lecithin plus cholesterol to form lipid vesicles as described above. Supernatants used in this study were tested for incomplete removal of the antigenic determinants by the different antisera. This was done by incubating each supernatant with fresh antisera (identical to the first antiserum used) and Staphylococcus A protein-Sepharose 4b beads. After removal of the

ELICITATION

4

OF CYTOTOXIC

6

12

T LYMPHOCYTES

16

20

Fraction

FIG. 2. Flotation isopycnic centrifugation of dialyzed egg lecithin plus cholesterol (A) and H-p-containing liposomes (B). Egg lecithin plus cholesterol (30% w/w) was mixed with trace amounts of [Wlphosphatidyl choline and dialyzed against PBS for 36 hr. This mixture was then suspended in sucrose (final concentration of 40%, w/w), a 40-5% sucrose gradient was layered over the sample, and the gradient was spun at 200,OOOg for 14 hr. The fractionated gradient was then tested for H-2d cell equivalents and radioactivity (A). A similar experiment was done with a mixture of lipid and H-2d-containing protein fraction resulting in formation of H-p-containing liposomes. The resultant gradient was fractionated and the H-2d cell equivalents and radioactivity ([3H]leucine and [Wlphosphatidyl choline) were determined (B).

beads by centrifugation, the resultant supernatant was incorporated into liposomes. The quantitative and qualitative patterns of lysis of target cells by CTLs elicited by either supernatant incorporated into liposomes were compared. In all experiments described in this paper the lysis of target cells by CTLs elicited with supernatants (H-2 antigens incorporated into liposomes) obtained after a second incubation with antisera, did not significantly differ from the lysis of target cells by CTLs elicited with H-2 antigens incorporated into liposomes after the first incubation with the same antisera (data not shown). RESULTS Characterization

of Partially

PuriJied H-2 Antigen-Containing

Fractions

Detergent-solubilized H-2d antigens from P8 15 cells labeled with [3H]leucine were dialyzed for 36 hr against PBS. The resulting opaque material was analyzed by flotation isopycnic centrifugation on 5-40% (w/w) sucrose gradients (Fig. 1A). The

334

ARTHUR

H. HALE

gradients were fractionated and the fractions monitored for radioactivity ([3H]leucine) and cell equivalents of H-2d antigens. The results indicate that after dialysis most of the H-2d containing material (Fig. 1A) and radioactivity (Fig. IB) remains at a density 2 1.12 g/cm3. Direct determination of protein content by the method of Lowry et al. (20) resulted in a similar fractionation profile (data not shown). Identical experiments have also been done with H-2d containing [3H]leucine and [*4C]phosphatidyl choline-labeled liposomes. (Figs. 2A, B). Formation of liposomes by dialysis in the absence of partially purified H-2d antigens followed by flotation isopycnic centrifugation, results in finding the majority of the [14C]phosphatidyl choline at a density K 1.02 g/cm3 (Fig. 2A). This fraction was shown not to inhibit anti-H-2d antibody-mediated cytolysis (Fig. 2A). Incorporation of H-2d antigen-containing protein into liposomes resulted in a significant shift in the density of the [14C]phosphatidyl choline label (1.05- 1.09 g/cm”) (Fig. 2B). Within this fraction was the majority of the [3H]leucine radioactivity (>85%) and most of the H-2d cell equivalents (>90%) (Fig. 2B). Direct determinations of protein by the method of Lowry et al. (20) confirmed the [3H]leucine data (data not shown). Similar experiments have been done with partially purified H-2b antigens. The results are in essence identical to those shown here for the H-2d gene products (data not shown). Electron microscopic examination of the liposome fractions indicated a unilamellar structure with heterogenous diameters of between 150 to 900 A (data not shown). The orientation of the H-2d antigens within the liposomes was investigated by inhibition of anti-H-2 antibody plus complement lysis. H-2 antigen-containing liposomes have greater than 80% of their antigenic sites available for binding anti-H-2 antibodies (Table 1). These results are consistent with reports of other investigators and lend support to the validity of liposomes as model-membrane systems (18, 19). Elicitation of Primary into Liposomes

and Secondary Anti-H-2

CTLs by H-2 Antigens Incorporated

We have previously determined that 1.3 x lo5 tumor cells (EL-4 or P815) represent the optimal number of stimulator cells under our culture conditions. Different fractions containing 1.3 x lo5 cell equivalents of H-2 antigens were then TABLE Relative Proportion

1

of H-2 Antigens Exposed to the Outside of Different Liposomes Liposomes

Percentage exposed to outsidea,*

H-2d-egg lecithin plus 30% cholesterol liposomes H-2d-egg lecithin plus 30% cholesterol liposomes

83.8 2 11.2 86.3 + 13.5

u Each determination represents six measurements with their standard deviations. b The percentage amount of H-2 antigens exposed to the outside of the liposomes was determined by the ability of the H-2-containing liposomes to inhibit anti-H-2 antibodies plus complement-dependent lysis. The total amount of H-2 antigens associated with liposomes was determined by solubilizing the H-2-containing liposomes with 0.2% Nonidet-P40 and analyzing the ability of this fraction to inhibit anti-H-2 antibodies plus complement-dependent lysis.

ELICITATION

OF CYTOTOXIC TABLE

Elicitation

335

T LYMPHOCYTES

2

of Secondary Anti-H-2 CTLs with H-2 Antigen-Containing

Liposomes”,”

Percentage specific Wr release (E:T Ratio 25: I) Con A-stimulated spleen cells Responder spleen cells BALB BALB BALB BALB BALB/c BALBlc BALB/c BALBlc

.B .B .B ,B

Stimulator fraction

BALB .B

BALBic

P815 tumor cells P815 plasma membrane PIIS-solubilized H-2” P815-H-2d in liposomes

-2.0 1.2 1.3 (1.4) 0.7

62.8 47.8 5.2 (27.8) 41.7

EL-4 tumor cells EL-4 plasma membranes EL-4-solubilized H-2” EL-4-H-2b in liposomes

48.7 27.8 3.1 (31.3) 31.2

-1.6 0.8 1.3 (0.7) 0.7

n Each determination represents the average of triplicate assays. The standard error for all determinations was less than 2.1%. * BALB .B and BALBic mice were injected intraperitoneally 8- 12 weeks earlier with 2.0 x 10’ BALB/c or BALB.B spleen cells, respectively. 7.0 x lo6 spleen cells from primed mice were incubated in virro with 1.3 x IO5 cell equivalents (optimal number of stimulator tumor cells) of each stimulator fraction for 5 days. I.3 x IO5 cell equivalents contain the following amounts of protein: P815 membranes, 1.2 pg protein; solubilized H-y, 1.08 pg protein; and, H-2d in liposomes, 1.08 pg protein-EL-4 membranes, 1.23 pg protein; solubilized H-2”, 0.98 pg protein; and, H-2b in liposomes, 0.98 pg protein. The resultant effector spleen cells were then tested for cytolytic activity on concanavalin A-stimulated spleen cells (BALB .B and BALB/c). The percentage specific release in parentheses is the lysis which resulted when responder populations were incubated with 10 times more protein of H-2 antigens. Incubation of responder cells with syngeneic stimulator fractions results in less than 2.0% specific lysis of either target cell. The percentage 51Cr release of BALB .B and BALBic Con A-stimulated spleen cells was 15.8 and 16.7, respectively.

tested for their capacity to elicit secondary anti-H-2 effector cells (Table 2). Either anti-H-2d or H-2b cytolytic effector cells were elicited by either intact tumor cells or membrane vesicles. However, detergent solubilization of purified plasma membranes followed by further purification over lentil lectin-Sepharose 4b columns resulted in a greatly reduced response (Table 2). By increasing the number of cell equivalents by a factor of 10 (1.3 x lo6 cell equivalents), we were able to demonstrate that lentil lectin-purified H-2 antigens (high-density H-2 antigens) were capable of eliciting specific anti-H-2 effector cells (Table 2, percentage specific release in parenthesis). The incorporation of 1.3 x lo5 cell equivalents of lentil lectin-purified H-2 antigens into egg lecithin plus cholesterol (30% w/w) liposomes increased the ability of the fraction to elicit anti-H-2 cytolytic effector cells (Table 2). As a control for all of the studies presented, responder cells were also incubated with H-2 antigens syngeneic to the responder cells. No significant lysis of either allogeneic or syngeneic target cells was observed (data not shown). The ability of H-2 antigens incorporated into liposomes to stimulate a specific primary anti-H-2 response is shown in Table 3. As with the secondary anti-H-2 cytolytic response, intact tumor cells are much more effective than purified plasma membrane vesicles or high-density H-2 antigens (Table 2). To date, we have not been able to demonstrate the ability of high-density H-2 antigens to elicit a primary

336

ARTHUR

H. HALE

anti-H-2 cytolytic response at any level tested (1.3 x 105- 1.3 x log cell equivalents). Lentil lectin-purified H-2 antigens incorporated into liposomes, however, elicited a low but significant level of cytolytic activity. The primary and secondary anti-H-2 responses elicited with all the subcellular fractions tested were abrogated by treatment of the cytolytic effector cells with anti-thy 1.2 serum plus complement (Table 4). These results clearly suggest that the cytolytic effector cells which are elicited by partially purified H-2 antigens incorporated into liposomes are T cells. Evaluation of the Relationship H-2K and H-2D Antigens

between Serologically

Dejined and T-Cell-Dejined

To evaluate the association of the serologically defined H-2 antigens incorporated into liposomes to those antigens required for stimulation of secondary anti-H-2 CTLs, we tested the ability of alloantisera to block the stimulation of anti-H-2 CTLs. Lentil lectin-purified H-2 antigens were incubated with alloantisera (final dilution 150) for 30 min at 4°C. The resultant antigen-antibody complexes were removed by the addition of Staphylococcus A protein-Sepharose 4b beads. The remaining detergent-solubilized protein fraction was mixed with egg lecithin plus cholesterol and dialyzed to form liposomes as described under Materials and Methods. The results (Table 5) clearly indicate that the anti-H-2 antisera were capable of complexing and specifically removing H-2 antigens responsible for TABLE Elicitation

3

of Primary Anti-H-2 CTLs with H-2 Antigen-Containing

Liposomes”,”

Percentage specific Yr release (E:T Ratio 25: 1) Con A-stimulated spleen cells Responder spleen cells

Stimulator fraction

BALB.B

BALBic

BALB .B BALB .B BALB.B BALB .B

P815 tumor cells P815 plasma membrane P815solubilized H-2d P815-H-2d in liposomes

1.3 0.8 -0.6 1.4

57.6 36.7 0.7 28.2

BALBlc BALBlc BALBlc BALBlc

EL-Ctumor cells EL-4 plasma membrane EL-Csolubilized H-2b EL-4-H-2b in liposomes

66.1 32.1 1.3 25.1

1.3 -2.1 0.2 -1.1

a Each determination represents the average of triplicate assays. The standard error for all determinations was less than 1.7%. b 7.0 x lo6 spleen cells from normal BALB .B or BALB/c mice were incubated in vi?ro with 1.3 x lo5 cell equivalents (optimal number of stimulator tumor cells) of tumor cells. All other stimulator fractions were added at 1.3 x 10” cell equivalents contain the following amounts of protein: P815 membranes, 12 pg protein; solubilized H-2d, 10.8 pg protein; H-2d in liposomes, 10.8 pg protein-EL-4 membranes, 12.3 pg protein; solubilized H-2b, 9.8 pg protein; H-2b in liposomes, 9.8 pg protein. The resultant effector spleen cells were then tested for cytolytic activity on concanavalin A-stimulated spleen cells (BALB .B and BALBk). Incubation of responder cells with syngeneic stimulator fractions results in less than 2.0% specific lysis of either target cell. The percentage Wr release of BALB .B and BALB/c Con A-stimulated spleen cells was 18.7 and 19.2, respectively.

ELICITATION

OF CYTOTOXIC TABLE

337

T LYMPHOCYTES

4

Sensitivity of Anti-H-2 Effector Cells Elicited With H-2 Antigen-Containing Anti-thy 1.2 Plus Complementfz~*

Liposomes to

Percentage specific 51Cr release (E:T Ratio 50: 1) Con A-stimulated spleen cells Treatment

Effector cells

BALB .B 1

BALBlc 2

1

2

-

22.1 21.2 22.7 2.1

43.7 44.6 42.6 1.3

BALB .B anti-H-2” NMS NMS plus c’ Anti-thy 1.2 Anti-thy 1.2 plus c’

-

BALBic

NMS NMS plus c’ Anti-thy 1.2 Anti-thy 1.2 plus c’

anti-H-2” 21.1 20.7 22.8 - 1.7

52 54 55 3.1

-

-

n Each determination represents the average of triplicate assays. The standard error for all determinations was less than 1.2%. b Primary (1) and secondary (2) anti-H-2 CTLs were elicited by incubating 7.0 x IO6 normal or anti-H-2 primed spleen cells with 1.3 x lo5 cell equivalents of H-2d or H-2b antigens incorporated into liposomes. 2.0 x lo6 resultant effector cells were incubated in 200 ~1 of a 1: 10 dilution of either normal mouse serum (NMS) or anti-thy 1.2 serum (diluted in RPM1 1640 medium) at room temperature. After a 20.min incubation period, the cells were washed by centrifugation (45Og, 5 min) and incubated in 200 ~1 of either RPM1 1640 or a 1:20 dilution of guinea pig complement for 20 min at 37°C. The percentage “‘Cr release of BALB.B and BALB/c Con A-stimulated spleen cells was 15.9 and 17.2, respectively.

elicitation of secondary anti-H-2 CTLs. Sequential addition of more anti-H-2 serum, followed by removal with Staphylococcus A protein-Sepharose 4b beads did not alter the results shown in Table 5. This control was done repeatedly for all experiments in this report; no difference in the lysis of target cells by CTLs elicited with the first or second supernatants was ever observed (data not shown). In an effort to evaluate more completely the specificity of the anti-H-2 CTL responses resulting from the coculturing of anti-H-2-primed spleen cells with lentil lectin-purified H-2 antigens, we tested anti-H-2-primed BALB .B and BALB/c spleen cells challenged with H-2 antigens incorporated into liposomes for their ability to recognize the H-2K and H-2D. ends of the MHC. Spleen cells from BALB .B mice primed 8-12 weeks earlier with BALB/c spleen cells when incubated with H-2d antigens, recognize both the H-2K and H-2D ends of the MHC (Table 6). This observation is strengthened by the fact that anti-H-2.31 (H-2Kd) serum has the capacity to remove from the partially purified H-2 antigen preparation an antigen required to elicit CTLs which recognize the H-2Kd end of the MHC (Table 6). Anti-H-2.4 (H-2Dd) serum eliminated the T-cell-defined antigens required to stimulate CTLs directed against the H-2Dd end of the MHC (Table 6). In

338

ARTHUR

H. HALE

summary, these results suggest that the T-cell-defined specificities are physically associated with the private serologically defined specificities at the H-2Kd and H-2Dd ends of the MHC. Similar specificities were observed for BALB/c spleen cell-primed BALB +B spleen cells (8-12 weeks earlier) and incubated with H-2b antigens incorporated into liposomes (Table 6). In order to analyze for the molecular relationship between the serologically defined public and private specificities, and those recognized by T cells, we have evaluated the lysis of different target cells by secondary anti-H-2b CTLs in which the responder and stimulator cell populations differ at the serologically defined public and private H-2 specificities (BALB . K anti-BALB -B) (25). As shown in Table 7, BALB *K anti-BALB 1B spleen cells incubated with H-2b antigens elicit CTLs which lyse Con A-stimulated BALB .B spleen cells effectively, however, little or no lysis was observed of Con A-stimulated BALB * K spleen cells (Table 7). We also tested these effector cells on target cells which were either H-2Kd (BlO*A) or H-2Dd (C3H * OH) to determine more precisely which specificities these effector cells were directed against. As shown in Table 7 significant cross-reactive lysis of both target cells (C3H. OH and B 10. A) occurred. Because of the lack of reactivity directed against the BALB . K target cells, we presume that the antigenic specificities being recognized are associated with the H-2d haplotype. Lysis of the H-2Kd target cell (C3HvOH) was prevented by the preincubation of the H-2b antigens with either anti-H-2.33 (H-2Kb) or anti-H-2.6, 22, 28, 29 (public specificities), suggesting that the T-cell-defined public specificities shared by the H2Kd and H-2Kb antigens are physically associated with the serologically defined HTABLE Inhibition

of Elicitation

5

of Secondary Anti-H-2 CTLs by Anti-H-2 Antibodies”*b Percentage specific Yr release (E:T Ratio 50: 1) Con A-stimulated spleen cells

Responder spleen cells

Stimulator fraction (H-2 antigen in Iiposomes)

Serum

BALB .B BALB.B BALB .B

H-2d H-2d H-2d

NMS Anti-H-2d Anti-H-2b

-

BALBlc BALBic BALBlc

H-2” H-2b H-2b

NMS Anti-H-2d Anti-H-2b

52.8 55.7 3.7

BALB .B

BALBlc 44.7 2.7 39.1 -

(t Each determination represents the average of quadruplicate assays. The standard error for all determinations was less than 1.1%. b BALB .B and BALB/c mice were injected intraperitoneally 8- 12 weeks earlier with 2.0 x 10’ BALB/c or BALB .B spleen cells, respectively. 7.0 x lo6 spleen cells from primed mice were incubated in virro with I .3 x IO5 cell equivalents of H-2 antigen incorporated in liposomes (2.1 pg protein) for 5 days. The solubilized H-2 antigens used to form the H-2 liposomes were incubated with a 1:50 dilution of the IgG fraction of either normal mouse serum (NMS), C57/BLlO anti-BlO.D2 serum (anti-H-2d), or BlO.D2 anti-C57BL/lO serum (anti-H-2b). Antigen-antibody complexes were removed with Staphylococcus A protein conjugated on to Sepharose 4b beads. The percentage Vr release of BALB .B and BALBic Con A-stimulated spleen cells was 21.2 and 13.8, respectively.

ELICITATION

OF CYTOTOXIC TABLE

339

T LYMPHOCYTES

6

Fine Specificity of Secondary Anti-H-2 CTLs Elicited With H-2 Antigens Incorporated into Liposome@ Percentage specific “‘Cr release (E:T Ratio 50: 1) Con A-stimulated spleen cells Responder

Stimulator H-2 antigens in liposomes

Serum

BALBlc (4 4

BALB .B BALB .B BALB.B BALB .B BALB .B BALB .B

H-2d H-2” H-2d H-2d H-2d H-2d

NMS Anti-2.4 Anti-2.31 Anti-2.2 Anti-2.33

62.7 63.4 38.9 27.7 64.2 60.7

0.7 1.1 0.7 1.2 1.3 0.9

37.8 38.9 39.7 1.8 37.9 36.5

25.7 26.7 0.9 27.8 26.7 28.7

BALBlc BALBlc BALB/c BALB/c BALB/c BALBlc

H-2b H-2b H-2b H-2b H-2b H-2b

NMS Anti-2.4 Anti-2.3 1 Anti-2.2 Anti-2.33

1.1 0.8 -0.7 1.3 1.4 -0.8

58.1 59.6 60.1 60.4 23.7 26.2

38.9 40.1 37.7 38.7 -0.8 35.7

31.2 30.7 28.9 32.4 31.9 1.2

BALB .B (b, b)

BALB .HTG Cd>b)

BIO.ASR (b, 4

a Each determination represents the average of triplicate assays. The standard error for all determinations was less than 2.7%. * BALB’B and BALBic mice were injected intraperitoneally 4-6 weeks earlier with 2.0 x 10’ BALBic or BALB.B spleen cells, respectively. 7.0 x IO6 spleen cells from primed mice were incubated in vitro with 1.3 x lo5 cell equivalents of H-2 antigen incorporated into liposomes (H-2d-l .7 pg of protein, H-2b-2, 1 pg of protein) for 5 days. The solubilized H-2 antigens used to form the H-2 liposomes were incubated with a 1:25 final dilution of the IgG fraction of either normal mouse serum (NMS), antiprivate sera. Antigen-antibody complexes were removed with Staphylococcus A protein-conjugated to Sepharose 4b beads. The percentage “‘Cr release of BALBic, BALB.B, BALB.HTG, and BlO.ASR was 16.4, 17.2, 18.9, and 15.3, respectively.

2Kb private and public specificities. Lysis of B10.A target cells was not prevented by anti-H-2.33 (H-2Kb) or H-2.2 (H-2IY), but was completely prevented by preincubation with anti-2.6,27,28,29. This result suggests strongly that the public specificities recognized by BALB . K anti-BALB .B CTLs are not physically associated with the serologically defined private H-2Db specificities, but are closely associated with the serologically defined public specificities at the H-2Db end of the MHC. Because of the lack of mouse recombinants which are either H-2KbDk or H2KkDb, we were unable to test BALB . K anti-BALB/c CTLs elicited with H-2d antigens. DISCUSSION The results reported here document the fact that a highly enriched H-2 antigen preparation incorporated into artificial vesicles of egg lecithin plus cholesterol 30% (w/w) is capable of eliciting both a primary and a secondary specific anti-H-2 CTL response in vitro. We also presented data to suggest that the serologically defined H-2 determinants are physically associated with those determinants recognized by T cells. The latter experiments done with antiprivate and antipublic sera suggest that both the serologically defined private and public specificities at the H-2D and

340

ARTHUR

H. HALE

H-2K ends of the MHC are physically associated with those determinants recognized by the T cell. Hansen et al. (9, 10) and Demant and Neauport-Sautes (11, 26) by sequential immune precipitation have shown that anti-H-2K private serum can coprecipitate H-2K public specificities, however, anti-H-2D private serum cannot precipitate all of the H-2D public specificities. As a result of this work the H-2L region has been identified and has been shown to be important for the elicitation of CTLs (22-24). Our results appear to correlate with these earlier observations and extend them by suggesting that the T-cell-defined specificities at the H-2D end of the MHC are physically associated with at least two serologically defined molecules coded by the H-2D and H-2L loci and one serologically defined molecule coded by the H-2K end of the MHC. This work extends other data which has shown that plasma membranes (6, 17), detergent solubilized (15), and reconstituted membranes (16, 18, 19) containing serologically defined H-2 or HLA antigens are capable of eliciting anti-MHC secondary CTL responses. A very recent observation that plasma membranes are capable of eliciting primary specific CTL responses (25) has been confirmed and extended by the observation that lentil lectin-purified H-2 antigens are capable, when incorporated into an artificial lipid vesicle (egg lecithin plus cholesterol 30% w/w), of eliciting a small but significant anti-H-2 primary CTL response. TABLE

7

Public Specificities of Secondary Anti-H-2 CTLs Elicited with H-2 Antigens Incorporated into Liposome@ Percentage specific Vr release (E:T Ratio 50: 1) Con A-stimulated spleen cells’ Responder BALB.K BALB.K BALB .K BALB .K BALB .K BALB .K BALB .K BALB .B antiBALB.Kd

Stimulator H-2 antigens in liposomes H-2b H-2b H-2b H-2” H-2” H-2” H-2b

-

Serum NMS Anti-H-2.4 Anti-H-2.3 1 Anti-H-2.2 Anti-H-2.33 Anti-H-2.6, (27-29) -

BALB .B (b> b)

BALB .K (k k)

C3H.OH (4 k)

73.7 72.8 70.7 71.2 41.2 39.7 1.2

0.1 1.1 -0.9 1.2 0.7 1.3 1.6

40.7 41.2 40.7 38.7 43.7 0.8 -1.3

0.7

62.4

-

B1O.A (k 4 36.8 39.8 32.6 31.7 32.6 33.7 1.0

-

” Each determination represents the average of triplicate assays. The standard error for all determinations was less than 2.1%. b BALB .K mice were injected intraperitoneally 8- 12 weeks earlier with 2.0 x 10’ BALB .B spleen cells. 7.0 x lo6 spleen cells from primed mice were incubated in vitro with 1.3 X lo5 cell equivalents of H-2 antigen incorporated in liposomes (2.0 pg protein) for 5 days. The solubilized H-2 antigens used to form the H-2 liposomes were incubated with a 1:25 dilution of the IgG fraction of either normal mouse serum (NMS), antiprivate or anti-public sera. Antigen-antibody complexes were removed with Staphylococcus A protein conjugated onto Sepharose 4b beads. c The percentage Yr release of BALB.B, BALB.K, C3H.OH, and Bl0.A was 11.7, 12.8, 15.9, and 13.9, respectively. d BALB .K target cells were shown to be susceptible to lysis by the use of BALB .B anti-BALB .K CTLs.

ELICITATION

OF CYTOTOXIC T LYMPHOCYTES

341

These results suggest a means by which the identification of T-cell-defined antigens can be made. The work also attempts to define the minimal molecular requirements for elicitation of CTLs. This includes not only the identification of the antigen responsible for elicitation, but also the requirements for presentation within the plane of a plasma membrane. The fact that lentil lectin-purified H-2 antigens were essentially lo-fold less efficient for elicitation of an anti-H-2 secondary CTL response and could not elicit a primary response, suggested that detergent solubilization or passage over a lentil-Sepharose column removed an active component required for elicitation. The reconstitution of the T-cell-defined activity by incorporation of the H-2 antigens into liposomes suggests that a stable membrane matrix is an important ingredient for antigen presentation to T cells. ACKNOWLEDGMENTS This work was supported by a research grant from the Forsyth Cancer Service, agrant to the Bowman Gray School of Medicine Cancer Center Support (Core) Grant (2-P30-CA 12197-08),National Institutes of Health, Bethesda, Maryland, The American Cancer Society MV-41, and by AI 15785 from the National Institutes of Health. A.H.H. is the recipient of a Research Career Development Award AI 00383.The author also wishes to thank Drs. Douglas Lyles and Donald Evans for helpful discussions.

REFERENCES 1. Bach, F. H., Bach, M. L., and Sondel, P. M., Nature (London) 259, 173, 1976. 2. Nabholz, M., Young, H., Meo, T., Miggiano, V., Rijnbeck, A., and Schreffler, D. C., Immunogenetics

3. 4. 5. 6.

1,456,

1975.

Forman, J., and Klein, J., Immunogenefics 1, 409, 1975. Forman, J., and Klein, J., J. Immunol. 115, 711, 1975. Brown, J. L., and Nathenson, S. G., .I. Immunol. 118, 98, 1977. Lemonnier, F., Burakoff, S. J., Mescher, M., Dorf, M. E., and Benacerraf, B., J. Immunol.

120,

1717, 1978.

7. Bevan, M. J., and Hyman, R., Immunogenetics 4,7, 1977. 8. Nabholz, M., Vives, J., Young, H. M., Meo, T., Miggiano, V., Rijnbeck, A., and Schreffler, D., Eur. J. Immunol. 4, 378, 1974. 9. Hansen, T. H., Cullen, S. E., Melvold, R., Kohn, H. I., and Flaherty, L., J. Exp. Med. 145, 1550, 1977.

10. Hansen, T. H., Cullen, S. E., and Sachs, D. H., J. Exp. Med. 145,438, 1977. 11. Neauport, Sautes C., Morello, D., Freed, J., Nathenson, S. G., and Demant, P., Eur. J. Immunol. 8, 511, 1977. 12. Matzinger, P., and Bevan, M., Cell Immunol. 29, 1, 1977. 13. Russell, J. R., Hale, A. H., Ginns, L. C., and Eisen, H. N.,Proc. Nat. Acad. Sci. USA 75,441,1978. 14. Russell, J. H., Hale, A. H., Inbar, D., and Eisen, H. N., Eur. .I. Immunol. 8, 640, 1978. 15. Mescher, M., Sherman, L., Lemonnier, F., and Burakoff, S., J. Exp. Med. 147, 946, 1978. 16. Fast, L. D., and Fan, D. P., J. Immunol. 120, 1092, 1978. 17. Engers, H. D., Thomas, K., Cerottini, J. C., and Brunner, K. T., J. Immunol. 115, 356, 1975. 18. Littman, D. R., Cullen, S. E., and Schwartz, B. D., Proc. Nat. Acad. Sci. USA 76, 902, 1979. 19. Engelhard, V. H., Strominger, J. L., Mescher, M., and Burakoff, S., Proc. Nat. Acad. Sci. USA 75, 5688, 1978. 20. Lowry, 0. H., Rosebrough, N. J., Farr, A. L., and Randall, R. J.,J. Biol. Chem. 193, 265, 1951. 21. Esko, J. D., Gilmore, J. R., and Glaser, M., Biochemistry 16, 1881, 1977. 22. Hansen, T. H., and Levy, R. B., J. Immunol. 120, 1836, 1978. 23. Biddeson, W. E., Hansen, T. H., Levy, R. B., and Doherty, P. C.,J. Exp. Med. 148, 1678, 1979. 24. Levy, R. B., Shearer, G. M., and Hansen, T. H., J. Immunol., in press, 1978. 25. Lemonnier, F., Mescher, M., Sherman, L., and Burakoff, S., J. Immunol. 120, 1114, 1978. 26. Demant, P., and Ntauport-Sautes, C., Immunogenetics 7, 295-311, 1978.