CD40 ligation restores cytolytic T lymphocyte response and eliminates fibrosarcoma in the peritoneum of mice lacking CD4⁺ T cells

Abstract

Absence of CD4⁺ T cell help has been suggested as a mechanism for failed anti-tumor cytotoxic T lymphocytes (CTL) response. We examined the requirement for CD4⁺ T cells to eliminate an immunogenic murine fibrosarcoma (6132A) inoculated into the peritoneal cavity. Immunocompetent C3H mice eliminated both single and repeat intraperitoneal (IP) inoculums, and developed high frequency of 6132A-specific interferon-γ (IFNγ)-producing CTL in the peritoneal cavity. Adoptive transfer of peritoneal exudate cells (PEC) isolated from control mice, protected SCID mice from challenge with 6132A. In contrast, CD4 depleted mice had diminished ability to eliminate tumor and succumbed to repeat IP challenges. Mice depleted of CD4⁺ T cells lacked tumor-specific IFNγ producing CTL in the peritoneal cavity. Adoptive transfer of PEC from CD4 depleted mice failed to protect SCID mice from 6132A. However, splenocytes isolated from same CD4 depleted mice prevented tumor growth in SCID mice, suggesting that 6132A-specific CTL response was generated, but was not sustained in the peritoneum. Treating CD4 depleted mice with agonist anti-CD40 antibody, starting on days 3 or 8 after initiating tumor challenge, led to persistence of 6132A-specific IFNγ producing CTL in the peritoneum, and eliminated 6132A tumor. The findings suggest that CTL can be activated in the absence of CD4⁺ T cells, but CD4⁺ T cells are required for a persistent CTL response at the tumor site. Exogenous stimulation through CD40 can restore tumor-specific CTL activity to the peritoneum and promote tumor clearance in the absence of CD4⁺ T cells.

Introduction

The role of CD4⁺ T cells in tumor immunity has been controversial. CD4⁺ T cells do not recognize tumor cells directly, since tumor cells usually do not express MHC-II. Mice lacking CD4⁺ T cells have been shown to effectively eliminate implanted immunogenic tumor cells and tumor fragments [1–3]. In some experimental models, depletion of CD4⁺ T cells was shown to enhance tumor elimination, possibly due to the elimination of regulatory CD4⁺ T cells [4, 5]. However, a beneficial role of CD4⁺ T cells in tumor immunity has also been reported. Recent studies demonstrated that depletion of CD4⁺ T cells enhanced tumor growth and adoptive transfer of CD4⁺ T cells can improve tumor eradication [6–9]. The mechanisms by which CD4⁺ T cells contribute to tumor immunity are not fully defined.

Classic studies suggested that efficient CD8⁺ T cell priming requires CD4⁺ T cells [8, 10–12]. CD4⁺ T cells, expressing CD40 ligand, induce full maturation of dendritic cells (DC), rendering them capable of priming CD8⁺ T cells [11]. Activated CD8⁺ T cells are able to proliferate, migrate in response to inflammatory stimuli, develop memory phenotype, and eliminate cells presenting antigens [13–17]. However, recent data have challenged the classical role of CD4⁺ T cells in cytotoxic T lymphocyte (CTL) activation. These studies contend that CD4⁺ T cells may be dispensable for the activation of effector CD8⁺ T cells, but are crucially involved in the generation and maintenance of CD8⁺ memory T cells and secondary immune responses [7, 9, 18–20]. In the absence of CD4⁺ T cells even a robust primary CTL response, as seen during viral and bacterial infections, can wane quickly, and fail to eliminate pathogens or respond to secondary antigenic stimulation [13, 14, 21–23]. Failure of secondary responses can result from activation induced cell death (AICD), anergy, or failure to disseminate to the periphery [23–28]. Differences in the methods for assaying immune responses, the avidity of TCR transgenic T cells for antigen, the antigen dose, and the activation of toll like receptors, may account for the discrepancies between the reports [15, 29–32].

In the current study, we investigated the requirement for CD4⁺ T cells to mount an effective CTL response against 6132A fibrosarcoma in the peritoneal cavity. 6132A is an immunogenic tumor that was induced in a C3H mouse with UV irradiation [33, 34]. 6132A, injected as a cell suspension subcutaneously (SC) is readily eliminated [34]. Elimination of 6132A has been shown to require CD8⁺ T cells [33, 35]. The tumor cells do not express MHC-II, even after interferon-γ (IFNγ) treatment, and are not directly recognized by tumor-specific CD4⁺ T cells [36, 37]. Previous reports demonstrated that the peritoneal cavity is a good approximation of the tumor environment [28, 38–40]. The model permits the determination of anti-tumor cytolytic activity and enumeration of tumor-specific IFNγ producing cells, without resorting to TCR transgenic mice or to in vitro expansion of T cells. Furthermore, peritoneal tumor spread, which is commonly found in patients with visceral sarcomas as well as carcinomas originating from stomach, colon, ovaries or breast, is often resistant to conventional therapy [41, 42]. Gaining insight into the immune response to peritoneal tumor may permit the design of effective immunotherapies against intraperitoneal (IP) sarcomatosis and carcinomatosis.

We found that CD4⁺ T cells were not absolutely required to prime anti-tumor CTL, but were essential for mice to develop a persistent and effective anti-tumor CTL response in the peritoneal cavity. Mice treated with agonist anti-CD40 antibody, administered as late as 8–10 days after initiating tumor challenge, retained tumor-specific CTL reactivity in the peritoneum and eliminated the tumor.

Materials and methods

Mice

Female C3H/HeJ, C3.MRL (MRL) or C3.SCID (SCID), murine mammary tumor virus-negative, (MMTV^-) mice were purchased from Jackson Laboratories (Bar Harbor, ME), kept in a pathogen-free barrier facility, and treated in accordance with institutional animal research committee guidelines. Mice between 6 and 12 weeks of age were used in all experiments. The least number of mice to achieve statistically significant difference between the experimental and control groups was used.

Cell lines, antibodies, in vivo depletion and IL-2 treatment

The murine fibrosarcoma cell lines 6132A and 6139B were originally induced in C3H/HeN MMTV^- mice [33, 34]. YAC-1, YTS169 and GK1.5 cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA).

FGK-45 cells were a generous gift from Dr R.J. Noelle, Dartmouth School of Medicine. CD4⁺ and CD8⁺ T cells were depleted using IP injection of 300 μg of GK1.5 or YTS169 monoclonal antibodies (mAbs), respectively. Depletion of CD4⁺ and CD8⁺ T cells from the peritoneal cavity and spleen was confirmed by flow cytometry on days 0 and 14. NK cells were depleted using α-asialo-GM1 Ab (Wako Chemicals, Richmond, VA), 50 μl per injection, administered IP. Agonist anti-CD40 mAb (clone FGK-45) was injected IP (150 μg) daily. Recombinant human IL-2 (rhIL-2) (Chiron Corporation) was administered twice daily IP, 20,000 IU per injection.

Tumor challenge, isolation of peritoneal CD8⁺ T cells and adoptive transfer experiments

For single tumor inoculums 5×10⁶, 10⁷ or 5×10⁷ tumor cells were injected IP. For repeat tumor challenge, 6132A cells (10⁷ cells/injection) were injected IP into mice at 3-day intervals for five doses (Fig. 1). Two days after the last tumor injection, peritoneal exudate cells (PECs) and splenocytes were isolated by peritoneal lavage and mechanical dissociation, respectively. In some experiments, CD8⁺ PECs and splenocytes were enriched by negative T cell selection using MACS immunomagnetic bead sorting (Miltenyi, Auburn, CA). Greater than 95% purity of CD8⁺ T cells was achieved. For adoptive transfer experiments, either 2×10⁷ un-separated cells or 2×10⁶ purified CD8⁺ T cells were transferred into the peritoneal cavity of SCID mice. SCID mice were challenged with 10⁶ 6132A or control (6139B) cells injected SC.

Fig. 1.

Design of intraperitoneal (IP) tumor challenge experiments. Depletion of CD4⁺ T cells was achieved with IP injections of GK1.5 mAb on days −7, −1 and +7 relative to commencement of tumor injections (day 0). Mice were challenged with either a single injection of 5×10⁶, 10⁷ or 5×10⁷ 6132A tumor cells, or five injections of 10⁷ 6132A cells, given IP, every 3 days. Twice daily IP injections of recombinant human IL-2 (20,000 IU/injection) were administered on days 3–7. Agonistic αCD40 mAb was administered either between days 3 and 5 or between 8 and 10. Tumor-specific killing and IFNγ production by PECs was assessed 2 days after the last tumor injection (day 14). The peritoneal cavity was examined for tumor spread on day 14, 28 or 90

Cytotoxicity assay

Cytolytic activity by PECs and CD8⁺ T cells was determined using 4-h ⁵¹Cr release assay [43]. Briefly, PECs or CD8⁺ T cells were added to 5×10^{3 51}Cr (Amersham, Piscataway, NJ) labeled target cells (6132A, 6139B or YAC-1) at effector to target ratios between 1:1 and 1:100 in 96-well V-bottom plates. Following incubation, supernatants (30 μl) were transferred to 96 well Lumaplates (PerkinElmer, Wellesley, MA), dried overnight and radioactivity determined using a TopCount microplate scintillation counter (PerkinElmer). The percentage of cell lysis was calculated by the following formula:

Spontaneous release was <10% of maximum release in all cases. Lytic units were calculated per 10⁷ cells at the 10% lysis level [44]. In order to compare the results of multiple experiments, the lysis data in each experiment was expressed as a relative value, compared to lysis by control PECs.

ELISPOT assay

6132A-specific IFNγ producing T cells were enumerated using ELISPOT [45]. Briefly, ELISPOT plates (Millipore, Billerica MA) were coated overnight with 2.5 μg/ml mIFNγ capture antibody (BD Bioscience, San Jose CA) and blocked for 2 h with RPMI containing 10% FBS (BioWhittaker, Walkersville MD). T cells were added to plates at varying concentrations. Irradiated tumor cells were added at 2.5×10⁴ cells/well. T cells cultured in medium or stimulated with 50 ng/ml PMA and 1 μg/ml ionomycin (P/I) served as negative and positive controls, respectively. T cells were incubated at 37°C for 48 h, washed multiple times with H₂O followed by PBS containing 0.05% Tween 20, and incubated for 2 h with 1 μg/ml of biotinylated mIFNγ detection antibody (BD Bioscience). After further washing, plates were incubated for 1 h with ExtrAvidin^® linked to alkaline phosphatase (Sigma, St. Louis MO) and developed using Sigma Fast^™ BCIP/NBT substrate (Sigma). Plates were read using an Immunospot plate analyzer (CTL Analyzers, LLC, Cleveland, OH). In order to compare multiple experiments, the frequency of IFNγ⁺ responders in each experiment was presented as a relative value, compared to 6132A PECs or to T cells stimulated with P/I.

Statistical analysis

Comparison of tumor growth or elimination was carried out using Fishers exact test. Comparison of lytic units, frequency of IFN-γ secreting cells and tumor size at defined time points, was carried out using the Mann–Whitney U test. Results were considered significant at the 95% confidence level. All statistical analysis was carried out using SPSS v11.0.1 software (SPSS Inc, Chicago, IL).

Results

Tumor-specific CTL infiltrate the peritoneal cavity and eliminate multiple inoculums of 6132A tumor cells

Activation of anti-tumor CTL response is required to prevent the growth of 6132A tumor (not shown) [33, 34]. We developed a model to quantify the CTL response following peritoneal 6132A tumor inoculation. Mice were challenged with 10⁷ 6132A cells IP on days 0, 3, 6, 9 and 12 (Fig. 1). The experimental design was based on previous studies that demonstrated that PECs, from mice repeatedly inoculated with antigen into the peritoneal cavity, were antigen-specific and directly cytolytic [28]. The peritoneal cavity was examined for tumor growth on day 14, 28 or 90. Control mice eliminated repeat 6132A inoculums without fail. These findings are consistent with previous reports suggesting that immunocompetent C3H mice routinely eliminate 6132A tumor cells injected as a single cell suspension [33, 34].

We then examined the tumor-specific reactivity in the peritoneal cavity. Total PECs and CD8⁺ PECs specifically lysed 6132A tumor cells (Fig. 2a). Although killing of the NK-sensitive YAC-1 cells by PECs was also observed, cold target inhibition with unlabeled YAC-1 cells greatly reduced killing of ⁵¹Cr labeled YAC-1, but not 6132A cells. Greater killing of 6132A than control tumor cells was not due to a difference in class I MHC expression, as H-2K^k expression by the two fibrosarcomas is similar; or in susceptibility to perforin-mediated killing, as 6132A is less susceptible to perforin mediated LAK lysis than the control tumor (data not shown). The findings suggest that killing of 6132A cells was antigen specific. IFNγ production by PECs in response to tumor was determined using ELISPOT. We found significantly greater frequency of IFNγ producing T cells in response to 6132A than control tumor (Fig. 2b).

Fig. 2.

Tumor elimination is associated with accumulation of 6132A-specific cytolytic IFNγ producing PECs. PECs were isolated by lavage on day 14 after initial tumor challenge. a Cytotoxicity against 6132A, 6139B control fibrosarcoma, and, NK sensitive YAC-1 cells, was determined directly in a 4-h ⁵¹Cr-release assay without in vitro re-stimulation. Lysis of 6132A and YAC-1, but not 6139B fibrosarcoma cells was observed (*P<0.05). Cold target inhibition (CTI), with excess unlabeled YAC-1 cells, resulted in a significant decrease in killing of ⁵¹Cr labeled YAC-1 cells, but 6132A lysis remained high, suggesting that 6132A tumor cell killing was antigen specific. Shown are representative data from 13 separate experiments. b Frequency of IFNγ producing cells was determined by ELISPOT (Inset). Primary PECs were added to 25,000 irradiated 6132A or control 6139B fibrosarcoma cells. After 18 h, IFNγ positive spots were counted. A high frequency of IFNγ producing PECs was observed upon stimulation with 6132A, but not 6139B fibrosarcoma cells (*P<0.05). The frequency of PECs producing IFNγ in response to 6132A was nearly as high as to P/I, comprising approximately 1.5% of all PECs and 10–20% of CD8⁺ T cells, suggesting a high frequency of anti-tumor CTL in the peritoneal cavity (based on flow cytometry analysis; not shown). The figure shown is a representative of seven independent experiments

Depletion of NK cells with α-asialo GM1 did not prevent the elimination of tumor (Table 1), and PECs remained 6132A tumor-specific (Fig. 3a, d). There was a modest reduction that did not reach statistical significance in 6132A-specific cytotoxicity (NK depleted: 86±48 LU; control: 113±58 LU; P>0.05) and frequency of IFNγ producing PECs (NK depleted: 510±324 spots; control: 618±646 spots; P>0.05) in mice treated with α-asialo GM1, compared to control mice. The findings demonstrated that peritoneal T cells specifically secreted IFNγ and killed 6132A cells. Taken together, the data suggest that tumor-specific cytolytic IFNγ producing T cells, capable of eliminating 6132A cells, infiltrated the site of tumor inoculation, the peritoneal cavity.

Table 1.

Effect of CD4⁺ T cell depletion and treatment with CD40 agonist or rhIL-2 on survival in mice challenged IP with 6132A fibrosarcoma cells^a

Group no.	Treatment	Tumor free %(n/n)	Compared to group no.	P value
1	WT	100 (58/58)
2	αCD4	13 (8/60)	1	<0.001
3	αNK	100 (10/10)
4	αCD4 + αNK	0 (0/10)	3	<0.001
5	αCD4 + IL-2	70 (26/37)	2	<0.001
6	αCD4 + αNK + IL-2	15 (2/13)	5	<0.001
7	αCD4 + αCD40 (d3-5)	91 (19/21)	2	<0.001
8	αCD4 + αCD40 (d8-10)	95 (19/20)	2	<0.001

^aCombined data from multiple experiments (Fig. 1)

Fig. 3.

Treatment with agonistic anti-CD40 antibody restores tumor-specific reactivity to the peritoneal cavity of CD4 depleted mice. Cytolysis by PECs was directly tested in 4-h ⁵¹Cr release assay (a–c). Frequency of IFNγ producing PECs was tested using ELISPOT (d–f). Closed triangles represent activity against 6132A fibrosarcoma, while open diamonds represent activity against control (6139B) tumor. a, d Control and CD4 depleted mice were challenged with 6132A tumor cells as shown in Fig. 1, and PECs were isolated on day 14. Tumor-specific cytolysis (a) was significantly reduced in CD4 depleted mice (19.4±14%) compared to control mice (100%). Frequency of IFNγ producing cells (d) was also significantly reduced in CD4 depleted mice (13.8±12%) compared to control mice (100%). The data represent the mean and SD from 13 separate experiments. CD4 depleted mice were repeatedly inoculated with 6132A tumor cells into the peritoneal cavity, and treated daily (IP) with agonist anti-CD40 mAb on days 3–5 (early) or days 8–10 (late). Therapy with anti-CD40 restored a tumor-specific cytolysis to PECs when given either early (62.83±40%) or late (82.24±33%), and d IFNγ production when given late (70.93±28.1%). The data represent means and SD from five independent experiments. b, e Mice depleted of CD4⁺ T cells, and challenged with 6132A tumor cells were treated with rhIL-2, 20,000 IU, twice daily on days 3–7. Additionally, groups of mice were also depleted of NK cells using α-asialo GM1 mAb on days −7, −1 and +7. a, d NK cell depletion did not significantly alter 6132A-specific cytotoxicity (NK depleted: 86±48 LU; control: 113±58 LU; P≥0.05) or IFNγ production (NK depleted: 510±324 spots; control: 618±646 spots; P≥0.05) by PECs. The data represent means and SD from four experiments. Exogenous rhIL-2 had a minimal effect on b cytolysis (16.7±19.19%) or e frequency of IFNγ producing cells (11.73±8.38%). The data represent means and SD from four independent experiments. c, f MRL mice were depleted of CD4⁺ T cells and inoculated repeatedly with 6132A (Fig. 1). Depletion of CD4⁺ cells led to a reduction in c 6132A-specific cytolysis (62.2±32% vs. 15.8±18.57%) and f frequency of tumor-specific IFNγ producing T cells (51.22±12.28% vs. 10.68±8.71%). The data represent means and SD from five independent experiments. (P value <0.05 comparing to *control mice; **CD4 depleted mice; ***control MRL mice; “^†” represents P value <0.05 comparing reactivity against 6132A and 6139B fibrosarcomas.) All data are presented as percentage of control PEC activity

Mice depleted of CD4⁺ T cells have diminished capacity to eliminate peritoneal challenge with 6132A cells

We next examined the requirement for CD4⁺ T cells to eliminate 6132A cells inoculated into the peritoneal cavity. Mice were depleted of CD4⁺ T cells using GK1.5 mAb. Depletion of CD4⁺ T cells from the peritoneal cavity and spleen was confirmed by flow cytometry on days 0 and 14 (not shown). Depletion of CD4⁺ T cells did not significantly alter the percentage of CD8⁺ T cells in the spleen or peritoneal cavity (not shown). Mice were first challenged with a single IP inoculum of 5×10⁶, 1×10⁷ or 5×10⁷ 6132A cells. We found that 100, 100 and 90% of control mice were free of tumor. Examination of the peritoneal cavity in mice lacking CD4⁺ T cells, demonstrated that 100, 70 and 40% were free of tumor. Control and CD4 depleted mice were then challenged with five IP injections of 1×10⁷ 6132A cells (Fig. 1). Examination of the peritoneal cavity for the presence of tumor implants on day 14, 28 or 90, demonstrated that only 13% of CD4 depleted mice were free of tumor (Table 1). In contrast, all control mice were free of tumor. The findings demonstrated that CD4 depleted mice were able to mount an immune response against 6132A tumor in the peritoneal cavity and eliminate single inoculums. But compared to control mice, CD4 depleted mice had greatly reduced ability to eliminate large or multiple peritoneal 6132A tumor inoculums.

6132A-specific CTL activity does not persist in the peritoneal cavity of mice depleted of CD4⁺ T cells

We next examined the mechanism(s) leading to greater tumor growth in the peritoneum of CD4 depleted mice. CD4⁺ T cells do not recognize the MHC-II negative tumor cells directly, and 6132A cells do not express MHC-II on the cell surface even after IFNγ treatment [37]. We postulated that CD4⁺ T cells were required for effective CD8⁺ T cell response against 6132A in the peritoneal cavity. To examine their effect of on the CTL response, CD4⁺ T cells were depleted and mice were repeatedly inoculated IP with 6132A tumor cells (Fig. 1). PECs from control and CD4 depleted mice were isolated on day 14. 6132A-specific reactivity was immediately analyzed using 4-h ⁵¹Cr release and ELISPOT. 6132A-specific cytolytic activity by PECs was markedly reduced in CD4 depleted mice (19.4±14%, P≤0.05) compared to control mice (Fig. 3a). The frequency of PECs, producing IFNγ specifically in response to 6132A, was also significantly lower in CD4 depleted mice (13.8±12%; P≤0.05) compared to control mice (Fig. 3d). Similar findings were observed on day 8, following three IP 6132A tumor inoculums (not shown). The findings suggest that CD4⁺ T cells were required for an effective and persistent anti-tumor CTL response in the peritoneal cavity.

Splenocytes from mice lacking CD4⁺ T cells retard the growth of 6132A tumor cells

We next asked whether CD4 depleted mice failed to activate anti-tumor CTL or whether the CTL were activated but failed to eliminate 6132A tumor cells. Elimination of a single inoculum of 10⁷ 6132A cells by CD4 depleted mice suggests that priming of anti-tumor CTL might occur in the absence of CD4⁺ T cells. We examined the ability of adoptively transferred T cells isolated from CD4 depleted and control mice to prevent 6132A tumor growth in SCID mice. Mice were challenged with 6132A cells as described in Fig. 1. Two days after the final tumor dose, PECs and splenocytes (2×10⁷ cells) were isolated and adoptively transferred by IP injection into SCID mice. In indicated experiments, CD8⁺ T cells were enriched using MACS columns prior to adoptive transfer (2×10⁶ cells). Twenty-four hours later the SCID mice were inoculated SC with 1×10⁶ 6132A or control tumor cells. PECs and splenocytes from control mice prevented the growth of 6132A tumors (Figs. 4a, b). Splenocytes, isolated from CD4 depleted mice could also protect SCID mice from 6132A tumor growth (6/11 mice were tumor free), albeit less effectively. In contrast, PECs isolated from CD4 depleted mice failed to protect SCID mice from 6132A tumor (0/11 mice tumor free). MACS purified CD8⁺ T cells also inhibited 6132A growth, with CD8⁺ PECs isolated from control mice, followed by CD8⁺ splenocytes isolated from CD4 depleted mice, being the most potent (Fig. 4c, d). Naive splenocytes did not significantly affect tumor growth (Fig. 4c, d). Adoptive transfer of T cells from any group failed to alter the growth of a control fibrosarcoma, suggesting that the response was tumor-antigen specific (Fig. 4e). Failure to eliminate the control fibrosarcoma was not due to intrinsic resistance to immune elimination, as adoptive transfer of splenocytes from mice rejecting the control fibrosarcoma protected SCID mice from challenge with the same tumor in previous studies [46]. The findings suggest that 6132A-specific CD8⁺ T cells, capable of eliminating tumor, were activated in both control mice and CD4 depleted mice. However, in contrast to control mice, tumor-specific CTL were present at greatly reduced frequency in the peritoneal cavity of CD4 depleted mice, which may explain why such mice were susceptible to peritoneal inoculation with 6132A cells. Failure of effective anti-tumor CTL response in the peritoneal cavity of CD4 depleted mice may have been the result of immune ignorance, AICD, anergy, or trafficking away from the tumor site. Accumulation of anergic T cells in the peritoneal cavity might not be detected using functional assays.

Fig. 4.

PECs and splenocytes from control mice and splenocytes from CD4 depleted mice prevent the growth of 6132A tumor cells in SCID mice. Control and CD4 depleted mice were challenged with tumor as described in Fig. 1. Two days after the final tumor dose, PECs and splenocytes were isolated. CD8⁺ T cells were purified using MACS immunomagnetic beads. Total PECs (20×10⁶), total splenocytes (20×10⁶), purified CD8⁺ PECs (2×10⁶) or purified CD8⁺ splenocytes (2×10⁶) from each group were transferred IP into SCID mice, which were then challenged SC with 6132A or control (6139B) fibrosarcoma. a Size of 6132A tumor in SCID mice treated with total PECs or splenocytes was measured. Tumors failed to grow in mice treated with PECs (closed diamonds) or splenocytes (closed triangles) isolated from control mice. Tumors from mice treated with PECs (open diamonds) or splenocytes (open triangles) isolated from CD4 depleted mice were significantly smaller than tumors from untreated mice (closed circles). b Tumor-free survival of SCID mice following adoptive treatment with PECs or splenocytes. PECs and splenocytes isolated from control mice, as well as splenocytes isolated from CD4 depleted mice, but not PECs isolated from CD4 depleted mice, prevented the growth of SC tumors in SCID mice. “*”Denotes P value <0.05 compared to no treatment. c CD8⁺ PECs and splenocytes from control mice, and CD8⁺ splenocytes from CD4 depleted mice retarded 6132A tumor growth in SCID mice. Naive CD8⁺ splenocytes (open circles) did not significantly delay tumor growth compared to untreated mice. CD8⁺ T cells were purified using MACS. d Survival of SCID mice following adoptive transfer of CD8⁺ PECs isolated from control mice or splenocytes isolated from CD4 depleted mice prevented 6132A tumor growth in SCID mice. “*”Denotes P value <0.05 compared to no treatment or naive splenocytes. e Adoptive transfer of PECs or splenocytes did not retard the growth of control (6139B) fibrosarcoma, and all mice (3/3 in each group) succumbed to tumor

Treatment with an agonist anti-CD40 antibody restores CTL activity to the peritoneum of mice lacking CD4⁺ T cells, permitting elimination of 6132A tumor

Signaling through the CD40 receptor plays a central role in Th1 responses and anti-tumor immunity. To determine whether anti-tumor response can be restored to the peritoneal cavity, an agonistic anti-CD40 monoclonal antibody was administered IP to mice depleted of CD4⁺ cells either on days 3–5 (early) or on days 8–10 (late; Fig. 1). Mice treated either early (90%) or late (95%) were protected from repeat inoculation with 6132A tumor cells (Table 1). PECs, isolated from CD4 depleted mice treated with anti-CD40 mAb starting on day 8, specifically lysed 6132A tumor cells (82% vs. 19% of control, P<0.05; Fig. 3a) and secreted IFNγ in response to 6132A cells (71% vs. 14% of control P<0.05; Fig. 3d). A similar but less pronounced trend was observed when the antibody was administered starting on day 3 (Fig. 3a, d). The findings suggest that CD4 depleted mice, treated with agonist anti-CD40 mAb, retain tumor-specific cytolytic activity and IFNγ production in the peritoneal cavity, possibly explaining their resistance to IP inoculation with 6132A tumor cells.

IL-2 reduces 6132A tumor growth but fails to restore tumor-specific CTL to the peritoneal cavity

IL-2 can restore effector function to anergic CTL and tumor infiltrating lymphocytes (TILs) [27, 39, 47–50]. We therefore administered an intermediate dose of exogenous rhIL-2 (20,000 IU), twice daily for 5 days, starting on day 3, to CD4 depleted mice that were challenged IP with repeat inoculums of 6132A cells (Fig. 1). Administration of rhIL-2 did not prevent the depletion of CD4⁺ T cells, nor did it alter the total number of CD8⁺ T cells in the peritoneum (data not shown). We found that 70% of CD4 depleted mice treated with rhIL-2 eliminated tumor (Table 1). However, NK cell depletion abrogated the beneficial effect of IL-2 on tumor elimination (Table 1). Furthermore, in the peritoneum of CD4 depleted mice treated with rhIL-2, cytotoxicity (19% vs. 17%; Fig. 3b), and frequency of IFNγ secreting cells (14% vs. 12%; Fig. 3e) were not significantly increased compared to untreated CD4 depleted mice. These findings suggest that rhIL-2 promoted the elimination of 6132A tumor in CD4 depleted mice predominantly by stimulating NK reactivity, and not by restoring tumor-specific CTL reactivity to the peritoneum.

Fas ligation does not mediate the loss of 6132A-specific CTL from the peritoneum of CD4 depleted mice

Fas-mediated AICD serves as a negative regulatory mechanism to prevent excessive proliferation and autoimmunity [25, 51–53]. We examined whether the absence of CD4⁺ T cells led to increased susceptibility of tumor-specific CTL to Fas mediated AICD in the peritoneal cavity. Mice carrying an inactivating mutation in the Fas gene (MRL) and C3H mice were depleted of CD4⁺ T cells and challenged with 6132A tumor cells. 6132A-specific cytotoxicity by PECs (Fig. 3c) and the frequency of PECs secreting IFNγ (Fig. 3f) in response to 6132A were markedly reduced in MRL mice lacking CD4⁺ T cells. The pattern of tumor reactivity by PECs from MRL mice with lack of 6132A specificity in CD4 depleted mice, was almost identical to C3H mice. The findings suggest that Fas-mediated AICD is not the predominant mechanism for loss of 6132A-specific CTL in CD4 depleted mice.

Discussion

Peritoneal tumor spread is commonly found in patients with gastrointestinal carcinomas as well as visceral sarcomas, and has poor prognosis [41, 42]. In the current study, we examined the requirement for CD4⁺ T cells in order to generate a CTL response to an immunogenic fibrosarcoma inoculated into the peritoneal cavity. We found that CD4 depleted mice were able to eliminate single IP tumor inoculums. However, mice depleted of CD4⁺ cells developed extensive peritoneal growth of 6132A tumors after single large (5×10⁷) or repeat tumor inoculums (five injections of 1×10⁷ cells). The findings suggest significantly greater susceptibility of CD4 depleted mice to persistent or repeat IP tumor challenge. Previous studies demonstrated that mice lacking CD4⁺ T cells are able to eliminate challenges with viral strains that produce an acute infection, a scenario analogous to a single 6132A inoculum. However, similarly to repeat inoculation with 6132A tumor, mice lacking CD4⁺ T cells fail to clear viral infections that have protracted course [54–56].

Since elimination of 6132A is dependent on CD8⁺ T cells [33–35], we examined CTL function in the peritoneal cavity. The peritoneal cavity served as an approximate reflection of the tumor environment, and permitted studying the local milieu of a tumor that is being eliminated. Using the peritoneal cavity to approximate the immunological milieu of a tumor has been previously reported [28, 40]. By avoiding the need to culture T cells ex vivo prior to analysis, we avoided a potential bias resulting from expansion of a subset of T cells that does not represent the original population. The model allowed the evaluation of the endogenous CTL response, which is likely to be diverse, in contrast to TCR transgenic models where a response by a uniform T cell population to a single epitope is examined.

These studies demonstrated that CD4⁺ T cells were required for optimal anti-tumor CTL response against 6132A tumor in the peritoneal cavity. In the presence of CD4⁺ T cells, high frequency of tumor-specific IFNγ-producing cytolytic PECs was found. Analysis of PECs from CD4 depleted mice revealed a dramatic decrease in tumor-specific cytolytic activity, and frequency of cells producing IFN-γ in response to 6132A tumor. Furthermore, PECs, isolated from CD4 depleted mice and adoptively transferred into SCID mice, had greatly diminished capacity to prevent 6132A tumor growth, when compared to PECs from control mice. The finding that CD4⁺ T cells were required for effective anti-tumor CTL response in the peritoneal cavity is consistent with previous studies using model tumor antigens [27]. Hanson et al. [57] demonstrated that immunization with a tumor antigen presented in the context of MHC-II enhanced the tumor-specific CTL response and reduced the number of CTL precursors that need to be adoptively transferred in order to eliminate the tumor. Similarly Marzo et al. [9, 58] demonstrated that adoptively transferred hemagglutinin (HA)-specific CD4⁺ T cells reduced the number of adoptively transferred CTL precursors needed to eliminate a transplantable tumor expressing HA.

In the current study, failure of persistent tumor-specific CTL response in the peritoneal cavity and 6132A tumor growth occurred despite activation of the anti-tumor immunity in CD4 depleted mice. Activation of anti-tumor immunity is suggested by the finding that CD4 depleted mice eliminated single challenges with 6132A tumor cells. And, in contrast to the peritoneal cells, CD8⁺ splenic cells, from CD4 depleted mice, retarded or prevented 6132A tumor growth in SCID mice, at least as effectively as CD8⁺ spleen cells from control mice. Reconstituted SCID mice that eliminated 6132A cells, were protected from re-challenge with 6132A, but not with control tumor, indicating that protection was T cell mediated and immune memory was established (not shown). The findings suggest that anti-tumor CTL were activated in CD4 depleted mice following challenge with repeat inoculums of 6132A tumor cells. CTL activation in the absence of CD4⁺ T cells may have been due to the high immunogenicity of the 6132A tumor cells. However, despite the high immunogenicity of 6132A tumor cells and the presence of activated CTLs in the spleen of CD4 depleted mice, such mice developed tumors in the peritoneal cavity. Taken together, the findings suggest that even when anti-tumor CTL are activated, CD4⁺ T cells are required for persistent tumor-specific effector response at the tumor site.

To examine whether CTL activity could be restored to the tumor site in the absence of CD4⁺ T cells, we tested the effect of CD40 ligation using agonist anti-CD40 mAb. Recent studies have suggested that CD40 ligation can enhance CTL activity and reduce tumor growth [59–62]. A member of the TNF-receptor superfamily, CD40 plays a central role in immune activation. CD4⁺ T cells expressing CD40 ligand may engage CD40, leading to DC maturation, which is characterized by upregulation of co-stimulatory molecules, secretion of cytokines such as TNFα, IL-6 and IL-12p70, and chemokines, such as CCL19 and CXCL9, 10 and 11 [10–12, 63]. We administered agonist anti-CD40 mAb to CD4 depleted mice either early (days 3–5) or late (days 8–10), and found that nearly all CD4 depleted mice eliminated 6132A tumor cells (Table 1). Analysis of PECs revealed significantly increased tumor-specific cytolytic activity and frequency of IFNγ producing T cells in the peritoneal cavity (Fig. 3a, d). Surprisingly, late administration of anti-CD40, starting on day 8, was even more effective than early treatment. At the same time point, on day 8, 6132A-specific CTL reactivity was detectable in control mice, but could not be detected in the peritoneal cavity of CD4 depleted mice (not shown). The findings suggest that the treatment with agonist anti-CD40 mAb restored tumor-specific CTL activity to the peritoneum and led the elimination of 6132A tumors. Recent studies have suggested that CD40 signaling can enhance the effector phase of the CTL response. Bronte et al. [59] reported that in vivo ligation of CD40 enhanced the efficacy of DNA vaccination, but only when administered during the effector phase of the CTL response. Stumbles et al. [23] reported that CTL were constitutively primed in tumor bearing mice, however, treatment with agonist anti-CD40 led to dissemination of tumor-specific CTL to the periphery and promoted the elimination of tumor. It is therefore tempting to speculate that in the absence of CD4⁺ T cells, primed 6132A-specific CTLs did not persist in the peritoneal cavity, but were restored to the tumor site upon CD40 ligation.

In the current study, we did not directly differentiate between preferential trafficking by 6132A-specific CTL to the tumor site, from T cell anergy or apoptosis at the tumor site. Anergy may be reversed in the presence of IL-2. However, exogenous IL-2 administration did not restore 6132A-specific CTL reactivity to the peritoneal cavity of CD4 depleted mice (Fig. 3b, e). AICD is mediated by Fas/Fas-L interaction. But, we did not find a higher frequency of tumor-specific CTL in the peritoneal cavity of CD4 depleted MRL mice challenged with 6132A cells IP (Fig. 3c, f). The findings suggest that IL-2 reversible anergy, and Fas mediated AICD, were not responsible for the lack of anti-tumor CTL activity in the peritoneal cavity of CD4 depleted mice. Our findings do not rule out the possibility that IL-2 unresponsive anergy, or apoptosis brought about by a Fas independent mechanism, resulted in failure of 6132A-specific T cell reactivity in the peritoneum of CD4 depleted mice.

In summary, our findings support the hypothesis that CD4⁺ T cells are required for efficient anti-tumor responses. In the absence of CD4⁺ T cells, tumor-specific CTL response in the peritoneal cavity was abortive, resulting in tumor growth. However, CTLs, capable of eliminating 6132A cells upon adoptive transfer into SCID mice, were activated and were retained in the spleens of mice depleted of CD4⁺ T cells. Treatment with agonist anti-CD40 restored tumor-specific cytolytic activity and IFNγ production to the peritoneum of mice lacking CD4⁺ T cells, and promoted tumor elimination. In contrast, exogenous IL-2 protected mice primarily by enhancing NK activity. The findings led us to speculate that CTL, activated in the absence of CD4⁺ T cells, cannot sustain an effective anti-tumor response in the peritoneal cavity, but can be redirected to tumor by agonist anti-CD40 antibody treatment. Absent or ineffective anti-tumor CD4⁺ T cell responses in tumor bearers have been previously reported [64–66]. Similarly, tumor-bearing mice, generated by implanting 6132A tumor fragments SC, were unresponsive to the immunodominant MHC-II antigen expressed by the 6132A cells (our unpublished results). The findings suggest that failure of anti-tumor CD4⁺ T cell response may be in part responsible for failure of anti-tumor CTL responses. Our study suggests that stimulating anti-tumor CTL may be insufficient to affect tumor destruction, and supports the hypothesis that stimulating CD4⁺ T cells or manipulating CD40 signaling could redirect anti-tumor CTL and enhance the efficacy of cancer immunotherapy.

Abbreviations

PECs

Peritoneal exudate cells

AICD

Activation induced cell death

DLN

Draining lymph nodes

CTL

Cytotoxic T lymphocytes

IFNγ

Intereferon gamma

IL-2

Interleukin-2