In previous studies, we have demonstrated that application of high hydrostatic pressure (P) to tumor cells in the presence of a slow-reacting membrane-impermeable cross-linker (CL), 2′-3′-adenosine dialdehyde, can rearrange cell surface proteins into immunogenic clusters. Here, we present evidence indicating that subsequent reduction of surface protein disulfides with N-acetyl-l-cysteine (NAC) further augments the immunogenic potential of PCL-modified tumor cells both in vitro and in vivo. Immunotherapy with PCL+NAC-modified 3LL-D122 Lewis lung carcinoma cells plus i.v. delivery of NAC in mice bearing established lung metastases provoked an antitumor response capable of eradicating the metastatic nodules as demonstrated by restoration of normal lung weight and histology. In addition, immunization with PCL+NAC-modified tumor cells gave rise to a strong delayed-type hypersensitivity recall response against parental D122 cells. We propose that this novel two-prong strategy, based on local immunization with autologous PCL+NAC-modified tumor cells and systemic boosting with NAC, could provide a practical, effective immunotherapeutic regimen for the treatment of human cancer.
Immunotherapy of cancer has been of limited clinical value to date, largely because of its failure to overcome the multiple escape strategies used by tumor cells (1). Renewed interest in cancer immunotherapy was stimulated in recent years by the identification of distinct tumor antigens that are capable of eliciting specific antitumor cytotoxic immune responses in vitro as well as in vivo (2, 3). However, the availability of synthetic or purified tumor antigens may not in itself be sufficient to offer an improved therapeutic modality because tumor antigens are numerous; metastatic tumor cells mutate frequently,thereby altering their antigenic nature; and the immunogenic hierarchy of various tumor antigens is largely unknown.
Current strategies to elicit specific antitumor T-cell responses include gene transfer into autologous tumor cells (4),vaccination with dendritic cells preloaded with tumor antigens (5, 6), systemic or paracrine cytokine therapy with Th1-type4cytokines (i.e., interleukin 2 and granulocyte-macrophage colony-stimulating factor; Refs. 7, 8), vaccination with T-cell epitopes derived from tumor proteins (2, 9),and vaccination with tumor-derived heat shock proteins (HSPs; Refs. 10, 11). However, tumor cells can evade immune killing either by down-regulation or total absence of immunogenic surface molecules, including MHC class I, or by synthesis and secretion of molecules capable of inactivating cytotoxic effectors, as well as by defective capillary neovasculature around the malignant tissue that slows down or even prevents cytotoxic cell egress into the tumor mass (1). Therefore, the strategy of an efficient immunotherapeutic attack on metastatic cancer must be based on combined local and systemic approaches capable of coping with such varied escape mechanisms simultaneously.
Over the last several years, our group has brought hydrostatic pressure to bear on the augmentation of the immunogenicity of tumor cells. We have developed a method in which cells are exposed to high hydrostatic pressure in the presence of a slow reacting biologically compatible impermeable cross-linker, AdA. This PCL procedure was shown to be effective and practical for increasing the immunogenicity of both murine (12, 13, 14, 15, 16) and human tumor cells (17). Hydrostatic pressure operates by transient depolymerization of cytoskeletal elements, thus allowing rearrangement of proteins on the cell surface (18, 19), as well as by induction of putative immunogenic HSPs (20). The net result of these alterations translates into formation of heterologous protein clusters on the surface of the treated cells, thereby converting such cells into potential targets for APCs. Despite being an effective and innocuous method, PCL by itself still suffers drawbacks similar to those of the other immunomodulating procedures mentioned above.
Rhodes et al. (21) observed that carbonyl groups artificially implanted on the surface of target cells can interact with free amine groups on the surface of CD4+ effectors to form stimulatory intercellular reversible Schiff bases that activate the Th1-like cytokine secretion pattern. An analogous intercellular bridge can be induced by infecting malignant cells with an innocuous strain of the Newcastle disease virus. The viral antigens that appear on the surface of the infected tumor cells deliver a specific recognition signal to the responding T cell via MHC/peptide-TCR interaction. Simultaneously, the Newcastle disease viral infection introduces a new class of cell-surface adhesion molecules, hemagglutinin-neuraminidases, which augment tumor cell-surface adhesiveness for responding effectors (22). The notion that reversible intercellular bridging between target and effector cells can strengthen specific immunological signals led us to investigate the effect of the production of reversible intercellular disulfide bonds in an antitumor immune response. We selected the biologically compatible reducing agent NAC, a mucolytic agent used in the treatment of respiratory disorders (23), to impose reactive sulfhydryl groups on the surface of tumor cells. In this study, we have tested the combination of PCL modification, NAC sulfhydryl reduction, and i.v. delivery of NAC (24, 25) in the induction of systemic antitumor cellular immunity. As our final goal, the capacity of this combined regimen to eradicate established metastasis was investigated.
MATERIALS AND METHODS
Female C57BL/6 (H-2b) and Balb/C(H-2d) mice, 8–12 weeks of age, were obtained from the animal breeding center, Weizmann Institute, Rehovot, Israel. Animals were maintained and treated according to NIH guidelines.
Tumor Cell Line
The D122 clone of the 3LL Lewis lung carcinoma (26), of C57BL/6 origin, was used. D122 cells were maintained in RPMI 1640 containing the following supplements, all from Beit HaEmek, Israel, unless otherwise noted: 10% FCS, glutamine,combined antibiotics, sodium pyruvate, nonessential amino acids, and 50μ m β-mercaptoethanol (Merck, Schuchardt, Germany). This nutrient-rich medium was termed complete medium. Cells were transferred twice a week and were free of Mycoplasma contamination as determined by a Mycoplasma ELISA test (Boehringer Mannheim GmbH, Mannheim, Germany) carried out every 3 months.
PB was composed of NaCl solution (saline) containing 8.0 g/l(150 mm) NaCl (pH 7.4). HBSS was composed of PB plus 1.0 g/l d-glucose (5.0 mm; pH 7.4).
Flow Cytometric Analysis
NAC Titration Assay.
Freshly removed C57BL splenocytes were treated with RBC-lysing solution(Sigma Chemical Co., St. Louis, MO) and washed twice in PB; aliquots of 3 × 106 cells were incubated with titrated concentrations of freshly prepared NAC in PB (pH 7.4) for 30 min at 37°C with occasional shaking. After cells were washed twice in PB, they were incubated with 10 μm F5M or 10μ m F5M-cys, i.e., F5M that was reacted with cysteine to block its active maleimide function. The blocked probe was prepared as follows. A solution containing 10μ m F5M and 10 mml-cysteine in PB (pH 7.4) was prepared and reacted at 25°C for 2 h in the dark to form F5M-cys. Conversion to F5M-cys was assessed by TLC on silica gel 60 F254 (Merck, Darmstadt, Germany) with a running solvent system of chloroform-methanol-0.1 msodium hydroxide (65:25:4, v/v) and was found to be >95% pure. Cells were incubated with the above probes for 1 h at 4°C, washed, and immediately analyzed on a FACSort instrument (Becton Dickinson,Franklin Lakes, NJ).
CD4 and CD8 Analysis.
Rat antimouse monoclonal antibodies were produced by hybridomas GK1.5(anti CD4) and 53-6 (anti CD8), which were obtained from ATCC, and mouse antirat-FITC F(ab′)2 was obtained from Jackson ImmunoResearch, West Grove, PA. Freshly removed C57BL splenocytes were treated with RBC-lysing solution and washed twice in PB; aliquots of 3 × 106cells were incubated with anti-CD4 or anti-CD8 antibodies for 30 min at 4°C. After the cells were washed, they were incubated with mouse antirat-FITC for an additional 30 min at 4°C. Cells were washed and immediately read on the FACSort.
Preparation of AdA
The synthesis of AdA was carried out by reacting adenosine(Fluka, Buchs, Switzerland) with sodium meta-periodate (BDH,Poole, United Kingdom) by a modification of the method described previously (13). One volume of a 90 mmsolution of adenosine was reacted with 0.1 volume of a 880 mm solution of sodium meta-periodate by dropwise addition of the sodium meta-periodate to the adenosine solution while mixing. The reaction was allowed to proceed for 15 min at room temperature, and then 0.05 volume of 2 m HCl were added. Quantitative removal of the iodate byproduct was carried out by the addition of 0.9 volume of a boiling solution of 0.1 mPbCl2 to the cooled AdA reaction mixture while stirring to form a precipitate of Pb(IO3)2. After 15 min on ice, the precipitate was filtered through three Whatman no. 1 filters(Whatman, Kent, United Kingdom) under vacuum. The slight excess of Pb2+ was removed by the addition of 0.02 volume of 1 m NaH2PO4and 0.004 volume of 1 mKH2PO4. A white colloidal precipitate was formed and was filtered after standing at 4°C overnight through three Whatman no. 1 filters under vacuum. The filtrate, composed of 40 mm AdA in an approximate ionic equivalent of PB, formed the stock solution. The standard working solution of 10 mm AdA was produced by diluting the stock solution 1:4 into HBSS. The colorimetric assay of Lappin and Clark (27) was used, with minor modifications, to determine the exact aldehyde concentration of the AdA preparation. The molecular structure of the AdA was verified by mass spectrometry (data not shown).
We have optimized the PCL methodology described previously (16). Briefly, a cell pellet containing up to 5 × 107 cells was suspended in 10 mm AdA solution in HBSS, transferred to a sterile liquipette (Elkay; Precision Laboratory Consumables, Galway, Ireland)and sealed in an air-free fashion with a sealing clip (Travenol Laboratories, Teva Medical, Ashdod, Israel). Immediately afterward, the sealed tube was placed into a custom-designed computer-controlled pressure device (Advanced Pressure Products, Ithaca, NY). The cells were then slowly brought to 1200 atm over the course of 15 min, locked in place for 15 min, and slowly decompressed to atmospheric pressure again over the course of 15 min such that the rates of compression and depression were equal. The cells were then washed twice immediately,resuspended in HBSS, and kept on ice until further use.
Untreated or PCL-treated cells were suspended in the treatment buffer, composed of 15 mm NAC (Sigma) and 1% FCS in HBSS(pH 7.4), and incubated for 30 min at 37°C. Treated cells were subsequently washed and resuspended in the maintenance buffer, composed of 15 mm NAC in PB, on ice until further manipulation. In experiments where the NAC-modified cells were injected into animals,200 μl of a 0.225 m NAC solution in PB was injected into the tail veins of mice ∼1 h before each inoculation with NAC-modified cells.
NAC Treatment of Stimulators.
Freshly removed C57BL splenocytes (stimulators) were treated with RBC-lysing solution, washed twice in PB, and incubated in HBSS alone or in HBSS containing NAC at 37°C for 30 min. Treated and untreated stimulators were washed three times and counted. Viability was >95%in all groups as assessed by trypan blue exclusion. Stimulators were irradiated with 20 Gy in all MLR assays. Freshly removed Balb/C splenocytes (responders) from animals previously immunized with∼5 × 107 C57BL splenocytes were isolated as above, washed twice in PB, and counted. Stimulators(1 × 105 to 1.25 × 104 cells) and responders (1 × 105 cells) were plated into 96-well microtiter plates in sets of six in a final volume of 200 μl per well. Plates were incubated for 4 days at 37°C and were pulsed with[3H]thymidine (6.7 Ci/mmol; ICN, Costa Mesa,CA) for the last 16 h of the assay. The plates were harvested(micromate 196 harvester; Packard, Merriden, CT) and scored on a 96-well plate reader (matrix 96; Packard). The stimulation index(SI) was defined as SI = [(R + S) − R − S]/R, where Rrepresents the scoring of responding splenocytes alone, and S represents the scoring of the stimulating splenocytes alone. A SI value >2 was considered a reliable positive response (28).
Exogenous Addition of NAC to MLR.
Stimulators and responders were prepared as above, and then 20 μl of 150 mm NAC in complete medium (pH 7.4) were added to the MLR cultures at the beginning of the incubation period. The cultures were incubated for 3–4 days and harvested as above.
Ten mice were immunized i.p. four times, 1 week apart, with 2 × 106 PCL+NAC-modified irradiated (100 Gy) D122 cells. Mice were then challenged with 2 × 105 viable tumor cells s.c.(100% of these animals remained tumor free); after 4 months, mice were boosted with 2 × 106UM-irradiated (100 Gy) D122 cells. One week later, the splenocytes from these mice were used as responders. Unmodified irradiated (100 Gy) D122 cells served as stimulators. Responders (1 × 105) were cultured with 1 × 105 to 2.5 × 104 stimulators per well in sets of six for 3 days. [3H]Thymidine (6.7 Ci/mmol; ICN) was added for the last 16 h of the assay. The plates were harvested(Packard micromate 196 harvester), and radioactivity was scored on a 96-well plate reader (Packard matrix 96). The SI was calculated as in the MLR assays.
In Vivo Tumor Assays
Groups of 9–10 age-matched mice were immunized i.p. four times, 1 week apart, with 2 × 106 UM or modified irradiated D122 cells. In all in vivo assays, tumor cells were irradiated with 100 Gy. Those mice receiving the NAC-modified cells also received i.v. injections of NAC as described above. A lethal challenge of 2 × 105 viable tumor cells was introduced s.c. 7 days after the last immunization, and the percentage of surviving animals was determined.
Experimental Metastasis Assay.
Groups of 10 age-matched mice were immunized i.p. three times, 1 week apart, with 2 × 106 UM or modified irradiated D122 cells, with or without the i.v. NAC, as described above. One week after the last immunization, 2 × 105 viable D122 cells were injected i.v. into the tail vein. Three weeks later, the mice were sacrificed, and the percentage of lung weights above normal (200 mg) was determined.
Metastasis Regression Assay.
Groups of 10 age-matched mice received i.v. injections of 5 × 105 viable D122 cells in the tail vein,and the cells were allowed to established lung metastases for 8 days (29). On day 9, mice received i.p. injections of 2 × 106 irradiated UM or modified D122 cells. Those mice receiving the NAC-modified cells also received NAC i.v. as described above. Two more immunizations, 7 days apart, were given, and 25 days after the last immunization, the mice were sacrificed and their lungs were weighed.
Lungs were removed and immersed in Bouin fixative for 24 h,washed twice, and resuspended in 70% ethanol. The tissue was mounted in paraffin blocks, sliced, stained with light green/hematoxylin, and examined microscopically.
Groups of five to six age-matched mice received i.p. injections of 2 × 106 UM ± NAC or PCL ± NAC-modified irradiated D122 cells. Those mice receiving the UM+NAC or PCL+NAC treatments were boosted 3–4 days later with a 200-μl injection of 15 mm NAC i.p. This immunization protocol was repeated twice more, at weekly intervals. One week after the last immunization, 2 × 105 UM irradiated D122 cells in 50 μl of HBSS were injected into the right ear and 50 μl of HBSS into the left ear of each mouse. After 24 h, mice received i.p. injections of a 200-μl solution of 125 μg/ml (1 mm)5-fluoro-2′-deoxyuridine (Sigma); 30 min later, they received i.p. injections of 2 μCi of 125I-Urd (2000 Ci/mmol;Amersham, Buckinghamshire, United Kingdom; Refs. 30, 31). After another 24 h, mice were sacrificed and the ears were removed. The radioactivity of each ear was then counted in a gamma counter (Riastar; Packard) and the R/L ratio determined. The upper limit for the R/L ratio was taken as 2, which is the maximal response obtained against sheep RBCs in the original assay (30).
Two-tailed P values were determined using either the nonparametric Wilcoxon test or Student’s t test to interpret the significance of the difference between the experimental groups.
Increased Immunogenicity of NAC-treated Cells.
Fresh C57BL splenocytes were incubated with increasing concentrations of NAC in PB for 30 min at 37°C and then labeled with the fluorescent probe F5M (see “Materials and Methods”). As a control, F5M was first reacted with l-cysteine and then used to estimate the nonspecific binding. The FACS analysis indicated that after treatment with 10 mm NAC, only a moderate-staining population was produced with an average mean fluorescence channel of 180 units. In contrast, at 15 mm NAC, there was a significant shift in the mean fluorescence channel (average of 790 units), with ∼14% of the population staining heavily, thereby demonstrating the ability of NAC at this concentration to efficiently expose surface thiols (Fig. 1). Viability was unaffected by treatment with up to 15 mmNAC (>95%) as determined by trypan blue exclusion.
The immunological effect of exposure of surface thiol groups on splenocytes was assessed using a MLR in which splenocytes of immunized female Balb/C (H-2d) mice acted as responders (R)and UM or NAC-reduced, irradiated splenocytes of female C57BL6/J(H-2b) mice, acted as stimulators (S). The SI of the allogeneic responders (R/S ratio of 1:1, see legend of Fig. 2) increased in a dose-dependent manner with increasing concentrations of NAC in the treatment buffer of the stimulators and reached a maximum of 21.7 SI units at 10 mm NAC (Fig. 2,A). Similar results were obtained by increasing the concentration of exogenous NAC added to the medium at the beginning of the MLR assay, with a maximal value of 33.7 SI units being observed again at 10 mm NAC (Fig. 2,B). In parallel, the control syngeneic MLR was performed in which responders from a Balb/C mouse immunized with Balb/C splenocytes were incubated with treated syngeneic stimulators under the same conditions of the allogeneic MLR. As shown in Fig. 2 C, syngeneic responders did not proliferate in the presence of untreated targets, NAC-reduced targets, or when NAC was added exogenously to the MLR cultures (all SI values <1). Given the fact that the reducing capacity of 15 mm NAC was significantly >10 mm and that the immunostimulatory capacity was similar to that of 10 mm, we chose 15 mm as the standard NAC concentration in subsequent experiments.
The in vivo immunostimulatory capacity of NAC was tested as follows. Ten C57BL6 mice were immunized with ∼5 × 107 Balb/C splenocytes. Eight days later, after the establishment of an active primary immune response, five mice received i.v. injections of NAC (367 mg/kg; final serum concentration,∼15 mm), whereas the remaining five mice received i.v. injections of PB; the splenic CD4 and CD8 expression was then analyzed by FACS. A significant increase in the relative expression of CD4 from 8 to 16% with no significant change in the expression of CD8 was observed in the allogeneic setting (Fig. 3). In a parallel experiment with naive mice, there was no difference in the expression of either CD4 or CD8 between the PB and NAC i.v groups(data not shown).
The immunotherapeutic potency of the NAC modality was tested with the nonimmunogenic D122 clone of the Lewis lung carcinoma, 3LL (32). Accordingly, an NAC treatment protocol composed of the following three stages was designed: (a) incubation of the cells in HBSS containing 15 mm NAC and 1%FCS (pH 7.4) at 37°C for 30 min; (b) washing and injection of the NAC-modified cells in PB containing 15 mmNAC (pH 7.4); (c) i.v. injection of NAC to achieve a final serum concentration of ∼15 mm (367 mg/kg) ∼1 h before each inoculation with NAC-treated tumor cells. In a survival assay, mice were inoculated i.p. four times, 1 week apart, with 2 × 106 irradiated UM or NAC-modified D122. One week after the last immunization, mice were challenged with 2 × 105 viable D122 cells s.c. The percentage of tumor-free animals in the treatment group was more than twice that of the group treated with UM cells (89%versus 42%; Fig. 4).
Following up on the survival experiment, we used the NAC protocol in an experimental metastasis assay. Here, mice were inoculated i.p. three times, 1 week apart, with 2 × 106irradiated UM or NAC-modified cells, and 1 week after the last injection, 2 × 105 viable D122 cells were administered i.v. into the tail vein. Twenty-five days later, mice were sacrificed, and their lungs were weighed and scored for metastatic load with respect to the average normal lung weight (200 mg). As presented in Fig. 5, there was only a slight difference between the UM-NACiv (80% above 200-mg level) versus the UM+NACiv (70% above 200-mg level)groups in lung weights. However, there was a significant difference between the NACmod-NACiv (70%) versus the NACmod+NACiv groups (25%). These experiments indicated that the combination of local immunization with NAC-treated tumor cells in addition to systemic NAC administration improved the antimetastatic activity considerably.
Combination of PCL and NAC Modifications in Antitumor Immune Response.
The capacity of splenocytes from mice immunized with PCL+NAC-modified D122 cells to proliferate in the presence of UM parental tumor cells in vitro was tested. Splenocytes from PCL+NAC-immunized mice proliferated (maximum SI, 5.4) in a dose-dependent manner against decreasing numbers of UM D122 targets in a MLTR (see “Materials and Methods”) as illustrated in Fig. 6. Splenocytes from naive mice did not proliferate in the presence of the tumor cells (SI <1; data not shown).
The efficacy of the combined PCL and NAC modifications in evoking regression of established metastases was then tested in vivo. Animals first received i.v. injections of 5 × 105 D122 cells, which were allowed to settle and to develop metastatic foci in the lungs over the course of 8 days. On day 9, immunotherapy was instituted by inoculation with 2 × 106 irradiated UM or PCL-modified D122 cells i.p., with or without NAC modification(including NAC i.v.). Overall, three inoculations were given 7 days apart, for three consecutive weeks. Mice were sacrificed 25 days after the last immunization, and the lungs were weighed. The results are presented in Fig. 7. As shown, in contrast to the mean lung weight of 420 ± 100 mg (mean ± SE) in mice treated with UM cells only,mice treated with PCL+NAC-modified cells demonstrated a normal lung weight of 200 ± 18 mg (Fig. 7,A). Examination of the histological sections of the lungs correlated with the metastatic load in each treatment group and clearly indicated the eradication of the metastatic nodules and recuperation of the alveolar spaces in the mice inoculated with PCL+NAC-treated cells (Fig. 7 B).
The efficacy of the antitumor T-cell response in vivodemonstrated in Fig. 7 was quantified by the radiometric DTH test of Vadas et al. (30). Mice were immunized three times, 1 week apart, with 2 × 106UM ± NAC or PCL ± NAC-modified irradiated D122 cells i.p.; those receiving UM+NAC- or PCL+NAC-modified cells were boosted between immunizations with NAC i.p. The higher ratio of the 125I-Urd uptake by the responding lymphocytes in the right ear (parental tumor challenge) versus the left ear (buffer) in the PCL+NAC immunization group lent further support to the contention that immunization with PCL+NAC-modified tumor cells triggered a specific T-cell immune response that can be recalled in a DTH assay (Fig. 8).
Successful application of immunotherapy as a modality in cancer treatment must be able to trigger an effective immune response against several tumor antigens simultaneously. Subjecting tumor cells to hydrostatic pressure in the presence of a biologically compatible cross-linker provides a simple, effective means for augmenting the expression of a battery of immunologically relevant surface molecules (16, 17). In this study, we have used a two-pronged strategy composed of immunization with PCL+NAC-modified cells and systemic administration of NAC as a general T-cell immunostimulant (24). The most notable result was obtained in a metastatic regression model in which mice received i.v. injections of highly metastatic D122 tumor cells and lung micrometastases were allowed to develop over an 8-day period (29). Subsequently, a series of immunotherapeutic regimens were administered, commencing on day 9. As shown in Fig. 7,A, there is a clear trend of reduction in mean lung weight progressing from the UM to the PCL+NAC immunization groups. This finding is supported by the histological findings in Fig. 7,B, which confirmed that the surface modification of the tumor cells by PCL+NAC in the presence of systemic immunostimulation with NAC induced a potent antitumor immune response capable of eradicating established metastases. Although the P value of PCL+NAC group versus PCL group in the regression assay was not statistically significant, the P value between PCL+NAC versus PCL inoculation groups in the DTH test was significant (P = 0.0475). Taken together, the emerging picture from Figs. 7 and 8 clearly indicates that the PCL+NAC regimen implemented in this study is superior to either PCL or NAC alone. The antitumor activity of this novel innocuous methodology compares well with other experimental immunotherapeutic approaches (33, 34, 35, 36).
In our previous work, immunization with PCL-modified B16.BL6 melanoma cells was shown to be effective in protecting against a subsequent lethal s.c. challenge of B16 cells (14) and in inducing a clear DTH recall response against EL4 and ARadLV 136 leukemia cells (13). However, at that point in the development of the method,immunization with PCL-modified cells was only partially effective in protecting against a lethal systemic challenge. The primary feature of the PCL+NAC methodology is its ability to lessen the metastatic load to an undetectable level. Further support for the potency of the PCL+NAC treatment emerged from the in vitro MLTR results where splenocytes from PCL+NAC-immunized mice responded strongly when cultured in the presence of UM D122 cells (Fig. 6).
The importance of thiols, (e.g., β-mercaptoethanol) in enhancing the growth of cells in tissue culture has been known for >20 years (37), but the exact mechanism has been a matter of long-standing debate (38). A number of studies have further shown that the presence of free thiols in the tissue culture medium is essential for lymphocyte proliferation and activation (24, 39, 40, 41). On the basis of findings that NAC can substitute for cysteine in tissue culture (42), we have tested in this study the ability of NAC as a costimulator in various in vitro assays. Our working hypothesis states that in the presence of NAC, surface intracellular disulfide bonds on the effector and/or target cell will be reduced to free sulfhydryls, thereby promoting the formation of transient intercellular disulfide bridges that could initiate a cascade of signal transduction inside the effector T cell, culminating in activation. In a parallel model,modification of target cells with carbonyl groups led to the formation of transient intercellular Schiff bases between the target carbonyl groups and amines constitutively expressed on presenting cell and T-cell surfaces, thus providing a strong costimulatory signal to CD4 Th-cells (21).
In view of the above, NAC was chosen as an innocuous reducing agent for exposing surface thiols groups on splenocytes as well as on tumor cells(Fig. 1). In the allogeneic MLR assay, NAC-treated targets, as well as NAC itself added exogenously to the culture medium, were able to significantly augment the proliferation of primed allogeneic effectors(Fig. 2, A and B) but had no effect in the parallel syngeneic MLR (Fig. 2,C). Furthermore, in a survival assay, NAC-modified D122 cells induced protection from a lethal challenge of parental D122 (Fig. 4).
As to the mechanism of the effector stimulation in the presence of thiol-enriched target surface, we showed that i.v. administration of NAC doubled the level of CD4+ splenocytes in mice immunized with allogeneic splenocytes, whereas the level of CD8+ cells remained unchanged 3 days after i.v. administration of NAC (Fig. 3). The origin of this effect may be in an increase of the half-life of the cell-APC coupling, which could increase the rate of transfer of an immunological signal. Lanzavecchia et al. (43)recently reviewed the kinetics of TCR activation. They pointed out that MHC/peptide complexes do not need to bind with high affinity to trigger the TCR. Strong TCR agonists are typically characterized by a Kd of 1–90 μmand half-lives of ∼10 s (44). In line with these values,it is suggested that the reactive thiols on the surface of target cells form transient intercellular bridges, which prolong the time of interaction between the effector and the stimulator and/or APC, thereby enhancing the rate of transfer of the specific immunological signal. This proposed mechanism could add another dimension to the kinetic proofreading model (45) by furnishing additional activation signals to T lymphocytes that could aid in breaking the nonresponsiveness of the immune system to disseminated metastatic foci.
Our results demonstrated that i.v. administration of NAC in addition to the NAC-modified tumor cells contributes significantly to the antitumor response. Taking into account the estimated 2-h mean half-life of the reduced form of NAC in the serum (23), we injected mice with a small volume of a concentrated NAC solution to reach a systemic concentration of reduced NAC of ∼15 mm in the serum ∼1 h before the NAC-treated tumor cells were injected. The rationale behind this was to present the host immune system with a specific immunogenic target and a systemic T-cell stimulant simultaneously. This combination synergized in the experimental metastasis model, where NAC surface modifications plus NAC i.v. (NACmod + NACiv group)proved to be the most effective in reducing the metastatic load, as seen both in the small (25%) percentage of lung weights above the normal 200-mg lung weight, as well as the gross appearance of the lungs themselves (Fig. 5).
In a preliminary set of experiments,5we detected an increase in the surface expression of HSP90 and HSP96 on the surface of PCL-treated B16.BL6 murine melanoma cells. Strong evidence is presented in the literature that these HSPs can serve as collectors and transporters of immunological peptides from the cytoplasm to the exterior plasma membrane. Indeed, vaccination with HSP70, HSP90, and GP96 was shown to elicit specific immunity against the tumors from which they were isolated (10, 46). On the basis of these converging lines of evidence, we hypothesize that PCL+NAC cells in the inoculum present immunogenic HSPs in addition to other immunogenic components on their surfaces and possibly leak HSP-peptide complexes into their immediate vicinity. APCs responding to these “danger signals” (47) pick up these immunological complexes, process them, and present them to both CD8+and CD4+ T cells. Circulating NAC stimulates these T cells in a positive feedback loop, thereby augmenting their capacity to respond to the remaining metastatic cells. Further clarification of the exact mechanism underlying the immunogenic potency of the PCL+NAC-treated cells is currently under investigation.
The antitumor potency of the combined PCL+NAC protocol stands out in its efficacy and applicability to the clinical setting. The advantage of this protocol manifested itself in the quantitative measurement of DTH to UM D122 cells after immunization with PCL+NAC-modified cells(Fig. 8). This immunization was sufficient to produce a significant DTH recall response to parental D122 cells in vivo. It is thus apparent that the PCL+NAC modification, combined with the NAC i.v. administration, is capable of breaking the barrier of nonresponsiveness of the immune system to the nonimmunogenic D122 tumor cells. The regimen of local immunization with PCL+NAC-modified tumor cells and systemic NAC administration meets the stringent criteria of a“multivalent” tumor vaccine that may eventually be applied in the clinic.
We thank Drs. Avi Eisenthal, Amnon Gonnene, Cohava Gelber, and David Dove for fruitful discussions, and Rachel Haimovitz for technical assistance.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
This work was partially supported by a grant from Immunotherapy Inc., New York.
The abbreviations used are: Th1,type 1 helper phenotype; HSP, heat shock protein; AdA, adenosine dialdehyde; PCL, pressure cross-linking; APC, antigen-presenting cell;TCR, T-cell receptor; NAC, N-acetyl-l-cytsteine; PB, phosphate buffer;F5M, fluorescein-5-maleimide; F5M-cys, F5M-cysteine-blocked probe;MLR, mixed lymphocyte reaction; MLTR, mixed lymphocyte-tumor reaction;SI, stimulation index; UM, unmodified; 125I-Urd,[125I]-iodo-2′-deoxyuridine; DTH, delayed-type hypersensitivity; R/L ratio, right/left ratio; FACS,fluorescence-activated cell sorting; NACiv, NAC i.v.; NACmod,NAC-modified.
Y. Goldman et al.Pressure-induced heat shock proteins promote the immunogenicity of tumor cells, manuscript in preparation.