SEARCH

SEARCH BY CITATION

Keywords:

  • combination immunotherapy;
  • hepatocellular carcinoma;
  • 4-1BB;
  • liver cancer;
  • therapy;
  • MDSC

Abstract

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Immunotherapy is a promising strategy against hepatocellular carcinoma (HCC). We assessed the therapeutic effects of stimulating CD137, a member of the TNF receptor family, with agonistic monoclonal antibodies (mAb). Agonistic anti-CD137 mAb treatment was tested on two in situ models of HCC in immunocompetent mice. We also studied the mediators involved at different time points. In an orthotopic HCC the treatment consistently leads to complete tumor regression in 40–60% of animals. The protection is long lasting in the animals responding to the treatment, which can reject a second tumor challenge more than 3 months after treatment and eradication of the first malignancy. The main mediators of the effect are T lymphocytes and NK cells, demonstrated through depletion experiments. In addition, adoptive transfer of splenocytes prepared from anti-CD137 mAb-treated and -cured mice to naive mice allowed them to, in turn, reject the tumor. The efficacy of anti-CD137 mAb treatment is associated with early, sustained recruitment of iNOS-positive macrophages within tumor nodules. Moreover, in the absence of treatment, tumor development is accompanied by infiltration by myeloid derived suppressor cells (MDSC) and regulatory T lymphocytes. In mice responding to the anti-CD137 mAb treatment, this infiltration is very limited, and a combination treatment with a depletion of MDSC leads to the recovery of 80% of the mice. These results demonstrate that agonistic anti-CD137 mAb is a promising therapeutic strategy for anti-tumor immunity stimulation against HCC.

Abbreviations
DEN

diethyl nitrosamine

HCC

hepatocellular carcinoma

IFNγ

gamma interferon

mAb

monoclonal antibody

MDSC

myeloid derived suppressor cell

NPC

non-parenchymal cells

TNF

tumor necrosis factor

Treg

regulatory T lymphocytes

Hepatocellular carcinoma (HCC) is the fifth most common cancer worldwide and, due to limited treatment options, the third greatest cause of cancer death.[1] Despite numerous attempts at amelioration, curative treatments continue to exhibit a very high recurrence rate.[2]

The liver possesses a specific immune system, which plays an intricate role in its functions and pathobiology, including the progressive evolution of chronic liver disease to fibrosis, cirrhosis and finally to HCC.[3] Work in recent years has demonstrated that local and systemic immunity can be harnessed to fight hepatic neoplastic development, and a variety of evidence points toward immunotherapy as a promising strategy against HCC. The presence of lymphocytic infiltrates has been shown to correlate with improved overall survival of HCC patients after surgical resection.[4] T lymphocytes specific for some tumor antigens have been detected in HCC patients.[5] In some patients, conventional treatment has been shown to enhance immune responses, suggesting that immunotherapy, as an adjuvant to, or following primary treatments,[5, 6] could be beneficial for the patient, especially as a solution to the high recurrence risk.

Antigen specific immunotherapy could be a valid option for the treatment of HCC and a number of strategies are in development at a clinical level.[7, 8] Complementary to these approaches requiring a certain level of personalization (not all HCC express all tumor antigens), there is a place for nonspecific immunotherapy strategies: to block the suppressive mechanisms that arise during tumor development and impede the anti-tumor immune response, or to directly enhance this natural anti-tumor immune response. It is noteworthy that, to date, some of the most promising clinical results for HCC immunotherapy have been obtained by adoptive transfer of autologous T lymphocytes nonspecifically stimulated after surgical tumor resection.[9]

CD137 is a member of the tumor necrosis factor (TNF) receptor superfamily. It is expressed by many cell types including activated T cells, NK and NKT cells, dendritic cells, neutrophils or monocytes.[10] The binding of its ligand CD137L to CD137 leads to the prolonged survival and sustained activation of CD8 T cells.[11] Agonistic anti-CD137 monoclonal antibody fixation to CD137 enhances T-cell proliferation and cytokine production by T-helper lymphocytes, and protects CD8 T cells from activation-induced cell death.[12, 13] Agonistic anti-CD137 antibody induces regression of established tumors in various animal models,[14-16] and prevents tumor recurrence in a model of melanoma.[17] In low tumor burden conditions this nonspecific immunotherapy strategy could be well-suited to the treatment of HCC in order to stimulate in situ the natural anti-tumor immune response. To test this hypothesis, we investigated the anti-tumor effect of an agonistic anti-CD137 antibody treatment on murine models of HCC for which the tumor develops within the liver of immunocompetent animals. We show the therapeutic efficacy of this approach. The mechanistic analyses demonstrate that the treatment mobilizes T cells, NK cells and allows the early and sustained recruitment of macrophages in the tumor nodules and that the immune response elicited by the treatment keeps potentially harmful regulatory cells at bay.

Material and Methods

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Mice and reagents

C57Bl/6 mice were purchased from Janvier Laboratory (France), and were kept in the Nantes IFR26 animal facility. Animal housing and surgical procedures were conducted according to the guidelines of the French Agriculture Ministry and were approved by the regional ethical committee. Hepa1.6 cells were cultured in DMEM (Life Technologies), 10% FBS, glutamine and antibiotics at 37°C and 5% CO2. Cells were harvested and resuspended in FBS-free DMEM for animal inoculation. Anti-CD137 mAb was produced from the 3H3 hybridoma (kindly provided by Pr R.S. Mittler).

Orthotopic model: anti-CD137 treatment, tumor challenge and adoptive transfer

Eight-weeks-old C57Bl/6J male mice received 2.5 × 106 Hepa1.6 cells in 100 µL through the portal vein, as described.[18] Four and eight days after tumor inoculation, mice were injected with 100 µg of rat anti-CD137 mAb, or rat IgG (Sigma–Aldrich). For tumor challenge, mice received a second dose of 2.5 × 106 Hepa1.6 cells by intrasplenic injection. For adoptive transfer, anti-CD137 cured mice and naive mice were sacrificed 10 days after rechallenge with Hepa1.6 cells. Total splenocytes were collected, single-cell suspensions were prepared and injected intravenously into naive mice, inoculated with Hepa1.6 cells on the same day.

DEN model

Fifteen-day-old male F1[B6xC3H] mice received one intra peritoneal (i.p.) injection of 50 μg g−1 DEN (Sigma). Immunizations with the agonistic anti-CD137 antibody (100 µg/injection) were performed at 4, 5 and 6 months post DEN-treatment.

Depletion experiments

CD8, CD4 T cells and NK cells were depleted in vivo by i.p. injection of 500 µg of 2.43, GK1.5, or PK136 mAb (BioXCell), 2 days before and 5 days after tumor inoculation. For MDSC depletion, mice were injected i.p. with 250 µg of RB6-8C5 mAb (BioXCell) twice a week starting on day 4 after tumor inoculation, for 3 weeks.

Flow cytometry

For liver non-parenchymal cell (NPC) preparation, the liver was digested with collagenase IV followed by Percoll density-gradient centrifugation and RBC lysis. Fluorochrome-conjugated anti-mouse antibodies: CD3e (145-2C11), CD4 (RM4-5), CD8 (53-6.7), CD11b (M1/70), CD11c (HL3), CD45 (RA3-6B2), CD25 (PC61), NK1.1 (PK136), Gr1 (RB6-6C5) and IFNγ (XMG1.2) (BD Pharmingen), FoxP3 (FJK-16s, eBiosciences), F4/80 (CI:A3-1, AbDSerotec). FACS analysis was conducted using a BD Pharmingen LSR-II flow cytometer and FlowJo software (Tree Star).

Proliferation test

Splenocytes or NPC were incubated with 1 µM Cell Trace™ Far Red DDAO-SE (Molecular Probes, Eugene, OR) followed by incubation with Hepa1.6 cell lysate for 72 hr in RPMI, 10% FBS and IL2 (50 U mL−1). Cells were stimulated with 40 ng mL−1 PMA, 200 ng mL−1 ionomycin for 5 hr, collected and stained for FACS analysis.

Quantitative real-time PCR

qRT-PCR was performed on RNA prepared from splenocytes or NPC with the Qiagen RNeasy Mini Kit (Qiagen). PCR reactions were done using SYBR Green Supermix (Primer sequences in Supporting Information Table 1).

Immunohistochemistry/immunofluorescence

Liver samples were formalin-fixed, and prepared as previously described,[19] or frozen in Tissue-Tek (Fisher Scientific) and cut into 10-µm sections. Immunohistochemistry was performed with anti-CD4 (RM4-5) and anti-CD8 (53-6.7) biotynilated antibodies (1:250) followed by streptavidin-peroxidase and diaminobenzidine revelation. Presence of macrophages was assessed with primary anti-mouse Ab: F4/80 (CI:A3-1; AbDSerotec) and iNOS (Abcam). Slides were analyzed using standard fluorescence microscopy and AxioVision imaging software (Carl Zeiss, Oberkochen, Germany).

Statistical analysis

The significance of differences in ratio and FACS analyses between experimental groups was determined using the non-parametric Mann–Whitney test. A log-rank test was used to analyze overall survival. Two-sided p values <0.05 were considered statistically significant between two groups.

Results

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Agonistic anti-CD137 mAb exhibits potent anti-tumor activity against HCC in vivo

The therapeutic potential of an agonistic anti-CD137 mAb was first evaluated in an orthotopic murin model of HCC. Hepa1.6 cells were injected into the portal vein of C57Bl/6 mice. On day 4 and 8 after tumor inoculation, mice received i.p. injections of anti-CD137 mAb. Animals were sacrificed on day 18 and tumor growth was assessed (Fig. 1a). All animals in the untreated group showed significant liver tumor mass development. In the anti-CD137 mAb treated group, the mean liver/body weight ratio was significantly lower (10.51% vs. 30.45% in the control group) and four out of the seven mice had normal liver morphology and weight. Hepa 1.6 cells do not express CD137 on their membrane and anti-CD137 mAb had no direct cytotoxic effect on Hepa1.6 cells in vitro, suggesting that the anti-tumor effect was not due to direct targeting of the tumor by the mAb (Supporting Information Fig. 1).

image

Figure 1. Anti-tumor activity of agonistic anti-CD137 mAb on in situ HCC models. In the orthotopic model, C57Bl/6 male mice were injected with 2.5 × 106 Hepa 1.6 through the portal vein and received 100 µg of either the anti-CD137 mAb (n = 7) or control rat IgG (n = 6) at days 4 and 8 after tumor inoculation. At sacrifice (day 18), liver morphology was checked and the tumor burden was determined by the ratio of liver weight over body weight (a). In the DEN model, 15-day-old F1[B6×C3H] males were injected with diethylnitosamine and received anti-CD137 mAb treatment (n = 17) or control rat IgG (n = 16) (100 µg) at month 4, 5 and 6 after DEN injection. At sacrifice (month 8), mouse and liver were weighed and nodules were counted and measured to determine a total tumor surface (b). Liver histology (H&E staining) was performed at the time of sacrifice in both HCC models. Some lymphocytic infiltrates were visible only in anti-CD137 treated mice (black arrows) (magnification ×100). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

The anti-CD137 mAb treatment was then assessed on an autochthonous, chemically induced HCC model. DEN-treated mice were injected with 3 monthly doses of 100 µg of anti-CD137 mAb, starting 4 months after DEN induction, when the liver is still macroscopically normal.[20] At the time of sacrifice (8 months) the anti-CD137 mAb-treated mice displayed a statistically significant lower total tumor surface (272.03 mm2 vs. 468.42 mm2) (Fig. 1b).

Histological analysis showed some inflammatory cell infiltrates in the liver parenchyma of anti-CD137 mAb treated animals of both models (Fig. 1).

Mice cured with anti-CD137 mAb develop long-lasting anti-tumor immunity

In an overall survival experiment on the orthotopic model, all the control mice died within 35 days of tumor inoculation whereas 40% of the anti-CD137 mAb-treated animals survived. When they were sacrificed ninety days after Hepa1.6 cell inoculation, they had no evidence of tumor, their liver was normal, indicating the generation of an effective anti-tumor immune response (Fig. 2a).

image

Figure 2. Anti-CD137 treatment induces long-term survival and long-lasting immunity. In the orthotopic model, C57Bl/6 mice received either control rat IgG or anti-CD137 antibody on day 4 and 8 after tumor inoculation. They were then monitored for overall survival (a). Hepa1.6 cell-inoculated mice treated and cured by anti-CD137 mAb therapy and naive mice were challenged with Hepa 1.6 tumor cells (day 0 corresponds to day 90 after the first inoculation of the cured mice). Mice were monitored for survival (b). To assess the potential of memory immune response generated by the anti-CD137 mAb treatment to eradicate the tumor in naive mice, splenocytes from anti-CD137 mAb treated and cured mice or from naive mice were injected i.v. in naive mice on the day of Hepa 1.6 tumor inoculation. The overall survival was evaluated over time (c). To identify cell populations that are involved in the therapeutic effect of the anti-CD137 mAb therapy, mice were injected with different depleting antibodies 2 days before and 5 days after the tumor inoculation (500 µg injection−1). Mice were then monitored for the overall survival (seven mice per group) (d). Statistical significance of survival observed for mice injected with the anti-CD137 mAb only compared to the other groups was measured by log-rank test, * : p < 0.05, ** : p < 0.01, *** : p < 0.005.

Download figure to PowerPoint

To test if this immune response would permit the rejection of a second tumor challenge, mice that had been successfully treated with anti-CD137 mAb were challenged with Hepa1.6 cells, 90 days after the first inoculation. Without the need for a second anti-CD137 mAb treatment, all the mice rejected this second tumor and were still alive and healthy 70 days later, whereas all naive mice had died (Fig. 2b). In another group, mice that had been cured of Hepa1.6 tumors after anti-CD137 mAb treatment were subcutaneously challenged with 5 × 105 B16F10 melanoma cells. All these mice showed significant tumor growth demonstrating the tumor-specificity of this memory immune response (data not shown).

We next investigated whether the memory response elicited by the anti-CD137 mAb treatment could be transferred into naive mice. Mice injected with Hepa1.6 cells received, on the same day, splenocytes from either anti-CD137 mAb-treated and -cured mice or from naive mice, challenged with Hepa1.6 cells 10 days before. All the mice that received the splenocytes from anti-CD137 mAb-cured mice rejected the tumors whereas all control mice died (Fig. 2c).

Altogether, these results demonstrate that treatment with agonistic anti-CD137 mAb cures 40 to 60% of the mice with HCC. In these mice the treatment induces a strong specific memory immune response that can successfully protect them against a second HCC challenge more than 3 months after treatment, and that can be extended to a naive mouse by adoptive transfer.

These first observations raised two important questions:

  • What are the mediators of the therapeutic effect of the anti-CD137 treatment?
  • Why is this treatment effective in only half of the animals?

Anti-CD137 mAb therapy requires primarily NK/NKT cells and CD8 T cells

In tumor-bearing animals, liver CD8+, CD4+ Tcells, NK and NKT cells, all express CD137 on their surface (Supporting Information Fig. 2). Depleting antibodies targeting CD8+ T cells, CD4+ T cells or NK/NKT cells were administered, 2 days before and 5 days after tumor inoculation. Anti-CD137 mAb was injected on day 4 and 8. Depletion was confirmed by FACS analysis on PBMC at day 11 (Supporting Information Fig. 3). In this experiment, 60% of anti-CD137 mAb-treated mice survived, whereas both CD8- and NK/NKT-depleted mice all died, as with the control untreated mice (Fig. 2d). When the CD4 T cells were depleted, the anti-CD137 mAb treatment led to an extended survival time, but the overall survival rate did not increase. Thus, CD8 T cells and NK/NKT cells, and also CD4+ T cells, are all involved in the therapeutic effect of the anti-CD137 mAb in this model.

Treatment with anti-CD137 mAb causes early T-lymphocytes recruitment to the liver

To identify the immune cells involved in the early stages of the anti-tumor immune response elicited by anti-CD137 mAb within the liver, mice were injected with Hepa1.6 cells, treated with the anti-CD137 mAb and sacrificed at two different time points (day 4 and 9 post-inoculation). Histology showed significant increase of infiltrates within the liver parenchyma of anti-CD137 mAb-treated mice at day 9. Immunohistochemistry and quantification of CD8+ and CD4+ T cells in these infiltrates demonstrated a significant increase in CD8+ T cells, but not in CD4+ T cells, in the liver of treated animals (Fig. 3a). FACS analysis performed on liver NPC showed an increased proportion of CD8+ T cells in anti-CD137 treated vs. control mice (36% vs. 24%), a lower percentage of CD4 T cells (23% vs. 32%) and of NK cells (8% vs. 10%), (Fig. 3b). Functionally, these cells showed an activated phenotype, and the anti-CD137 mAb treatment induced a statistically significant increase of liver CD69+CD8+ T cells (Supporting Information Fig. 4). All these cells produced more IFNγ compared to those isolated from untreated, tumor-bearing mice (Fig. 3c).

image

Figure 3. Characterization of lymphocytic infiltration induced by anti-CD137 treatment. (a) C57Bl/6 mice were sacrificed on day 4 and day 9 after tumor inoculation. Anti-CD8 and -CD4 IHC was performed on frozen liver sections. Quantification on at least 10 fields allowed the determination of the number of positive cells/µm² (n = 4/gp). (b) Percentage of CD8+ T cells, CD4+ T cells, and NK cells among the NPC were established by FACS analysis at day 9 after tumor inoculation (n = 5 to 7/gp). (c) On day 9 after tumor inoculation, liver NPC were prepared and stimulated with 100 µg mL−1 of Hepa1.6 tumor lysate for 3 days. IFNγ intracellular staining in CD8 T cells, CD4 T cells and NK cells was performed (n = 5 to 7/gp). (d) Proliferation assay was performed on NPC and splenocytes and percentage of dividing CD8+ and CD4+ T cells was determined (n = 10/gp). (e) Percentage of CD8+ T cells, CD4+ T cells, and NK cells among the NPC were established by FACS analysis at day 14 after tumor inoculation, when it was possible to discriminate responder from non-responder mice (n = 3–4 animals/gp, mean ± SEM). [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

Proliferation of CD8+ and CD4+ T cells in NPC and splenocytes prepared from mice sacrificed at day 9 were assessed after in vitro stimulation with Hepa1.6 cell lysate (Fig. 3d). There was an increased proliferation of CD8+ T cells in the splenocytes from all the anti-CD137 mAb-treated mice. Interestingly, only half of these mice showed an increased proliferation of liver-infiltrating CD8+ and CD4+ T cells. However, at this early time point it was not possible to discriminate between mice responding and mice not responding to the treatment in terms of tumor growth evaluation.

HCC bearing animals were treated with the anti-CD137 mAb as above and sacrificed 14 days after tumor inoculation, when tumor burden analysis allows the differentiation of mice responding from those not responding to the treatment. FACS analysis of NPC demonstrated that responder mice had significantly more CD8+ and CD4+ T cells in their liver than non treated and nonresponders. The frequency of NK cells was still lower in treated mice than in control, but this difference was less pronounced in the responder mice (Fig. 3e). During tumor development, the absolute number of liver NK and NKT cells tended to decrease and the anti-CD137 treatment did not modify this nor did it induce a specific recruitment of these cells within the tumor nodules (Supporting Information Fig. 5).

In summary, anti-CD137 mAb treatment induces an early infiltration of the liver tumor, mainly with CD8+ T cells. In some mice these CD8+T cells, together with CD4+ T cells, proliferate, and this most probably represents the basis of the anti-tumor response.

Role of macrophages in early stage of anti-tumor immune response

Other cell populations known to play a role in the anti-tumor immunity were also studied. There was a significant increase of macrophages among NPC from anti-CD137 mAb treated mice compared to untreated mice on day 9 (16.3 vs. 13.7%), but not on day 14 (Figs. 4a and 4b). To determine if this transient increase was related to a favorable outcome, HCC bearing mice were treated with anti-CD137 mAb, and a liver biopsy was performed at day 9. The animals were then left alive and monitored for survival to identify responders and nonresponders (Fig. 4c). There was a statistically significant increase in the frequency of F4/80+ cells in the NPC prepared from liver biopsies of responder animals compared to the nonresponders (21.45% vs. 14.13%, Fig. 4d). RT qPCR performed on the biopsies also revealed a significantly higher expression of the proinflammatory cytokine IL6 in the responders (Supporting Information Fig. 6).

image

Figure 4. Presence of macrophages at an early stage in the anti-tumor response is correlated with a good prognosis. Percentage of macrophages was determined by FACS analysis performed on NPC isolated at day 9 (n = 5 to 7/gp) (a), and day 14 (n = 3 to 4/gp) (b) after Hepa1.6 injection. (c) In another experiment, a liver biopsy was performed on day 9 after tumor inoculation and mice were monitored for overall survival. (d) NPC were extracted from the biopsies and the percentage of F4/80+ cells was established (n = 5 to 11 animals/gp, mean ± SEM).

Download figure to PowerPoint

Immunohistochemistry revealed a striking difference in the distribution of the F4/80+ cells within the liver parenchyma (Fig. 5). At day 9, when tumor nodules become visible, F4/80+ cells were found in the normal parenchyma as well as in some tumor nodules of anti-CD137 mAb treated mice. On day 14, these cells were present only in tumor nodules of responder mice. Moreover, these cells were also positive for iNOS staining, a marker of M1 proinflammatory macrophages.

image

Figure 5. Difference in macrophage localization and phenotype in anti-CD137 mAb treated mice. Liver sections of control and treated mice sacrificed at different time points (days 4, 9, and 14 after Hepa 1.6 injection, n = 4/gp) were performed and stained with an anti-F4/80 and an anti-iNOS antibody. Isolated tumor nodules are marked with dotted lines when possible. At day 14 in control and non responder animals tumor invasion is too important to identify isolated nodules. Arrows show the colocalization of the two markers in some cells within the tumor nodules in the liver of treated mice only at earlier stage and in cured-mice at day 14. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Download figure to PowerPoint

RT-qPCR was performed to evaluate the relative expression of the chemokine CCL2, and of iNOS and Arginase-1, markers of M1/M2 polarization, in NPC from control and anti-CD137 treated mice. CCL2 and iNOS expression were higher in NPC of responder mice compared to control whereas arginase-1 expression was lower (Supporting Information Fig. 7). These variations need to be confirmed on more animals. They suggest a better recruitment of macrophages and a M1 proinflammatory polarization of these macrophages in the tumor nodules of responder mice.

Increased populations of MDSC and Treg cells in nonresponder mice

Phenotypic analysis of liver NPC revealed that tumor development was accompanied by an increased infiltration of regulatory cells in untreated animals. Myeloid derived suppressor cells (MDSC), characterized by a high expression of CD11b and Gr1 in mice, were 10 times more frequent at day 14 compared to day 9 after tumor inoculation (Figs. 6a and 6b). These MDSC were mainly of monocytic origin (Ly6C+, Ly6G−) and did not express CD137 (Supporting Information Fig. 8). Infiltrating regulatory T cells also accumulated over the same time (2.9–5.2%) (Figs. 6c and 6d). The anti-CD137 mAb treatment did not modify the frequency of these cells at day 9. However, at day 14, responder animals had no accumulation of Treg cells (2.8%) and a very limited increase of MDSC (7.7%) in their livers, whereas the non-responders showed massive infiltration of MDSC and Treg cells, similarly to untreated animals.

image

Figure 6. Tumor infiltration by myeloid derived suppressor cells and regulatory T cells. Percentage of MDSC (a,b) and Treg cells (c,d) were determined by FACS analysis on NPC isolated from control and anti-CD137 mAb treated mice at day 9 (a,c, n = 5 to 7/gp) and day 14 (b,d, n = 3 to 4/gp) after tumor inoculation (mean ± SEM). (e) Groups of mice (8–10/gp) were inoculated with Hepa1.6 cells and treated with control IgG, or anti-CD137 mAb (100 µg on days 4 and 8), anti Gr1 Ab (250 µg twice a week for 3 weeks), or a combination of both, and monitored for overall survival.

Download figure to PowerPoint

This result prompted us to evaluate the therapeutic effect of a combined treatment with the anti-CD137 mAb and an anti-Gr1 Ab known to deplete MDSC[21] (Fig. 6e). By itself, treatment with anti-Gr1 Ab led to a discrete anti-tumor effect, with 2 mice out of 10 cured. As in our previous experiments, 40% of the mice treated with the anti-CD137 mAb alone survived. Finally, the combination of the anti-CD137 mAb treatment with the MDSC depletion resulted in 80% of the mice cured.

Discussion

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

In the present work we demonstrate the therapeutic efficacy of agonistic anti-CD137 mAb in two models of in situ development of HCC, with a 65% reduction of the tumor burden in the orthotopic model. On the more stringent model of chemically induced HCC, a 42% reduction of the tumor burden was obtained after only three injections of 100 µg of the anti-CD137 mAb during 8 months of liver carcinogenesis.

In the orthotopic HCC model, the treatment induces a strong memory response, which protects the mice against a second presentation of tumor cells 3 months after the first, and after the anti-CD137 mAb injections. Moreover, adoptive transfer of splenocytes prepared from mice treated and cured using anti-CD137 mAb inhibits the tumor growth in recipient naive mice. These are very significant results, which reinforce the validity of this approach to treating tumor recurrence in the liver.

This effect is primarily mediated by NK and T lymphocytes. These two cell types are the most common effectors involved in the anti-tumor effect of the agonistic anti-CD137 mAb for various malignancies.[22] In our model, the therapeutic effect of the anti-CD137 mAb treatment is limited to delayed tumor growth in CD4+ T cell-depleted mice, consistent with a role of these cells in enhancing the expansion and cytolytic function of tumor reactive CD8+ T cells[23] and in the maintenance of a long-term memory response.[14] Functionally, CD8+ and CD4+ T cells, and NK cells from the liver of treated mice have an increased capacity to secrete IFNγ. This could be responsible for the decreased percentage and number of NK cells in the liver and in the spleen. Indeed, it has recently been shown that CD137-induced IFNγ production downregulates NK cell development in bone marrow, leading to a decreased number of these cells in the periphery.[24]

Anti-CD137 mAb treatment consistently led to 40–60% of the mice completely cured after two injections, and as early as 1 day after the second mAb injection, a specific proliferation of liver CD8+ and CD4+ T cells can be detected in only half the mice, though it is not yet possible to identify responder from non responder mice. On day 14, a time when responders already show a measurably reduced tumor burden, the responders have more liver-infiltrating CD8+ and CD4+ T cells than at day 9. At the same point the anti-tumor immune response seems to have “aborted” in the nonresponders.

We observe an early variation in liver macrophages after anti-CD137 mAb treatment. Macrophages can, depending on their polarization, either favor or limit the host immune response. The M1 phenotype is induced by IFNγ and is associated with anti-tumor activity and iNOS expression. The M2 phenotype is associated with tissue remodeling, anti-inflammatory cytokine secretion and arginase-1 expression.[25] Presence of tumor infiltrating macrophages in HCC after resection has been correlated with both good and bad prognosis.[26, 27] In absence of treatment, we observed that macrophages are present in the periphery of the tumor nodules but not inside the nodules. The anti-CD137 mAb treatment allows the macrophages to penetrate within tumors early in their development and the sustained infiltration of tumor nodules with macrophages is correlated with a favorable outcome for the animals. Macrophage behavior is largely coordinated by cytokine/chemokine signaling and consistent with this, CCL2 (MCP-1), a chemokine known to have a major role in the recruitment and activation of macrophages, as well as IL6 proinflammatory cytokine, are found in increased amounts in the liver of responder mice at an early time point. This concords with results obtained in human HCC, where the expression of IL-6 and of CCL2 have been associated with a good prognosis.[28] Finally, we demonstrate that the tumor infiltrating macrophages express iNOS, a marker of M1 proinflammatory/anti tumoral polarization.

Regulatory immune cells might also be involved in the early arrest of anti-tumor immune response in the non responder mice. Increased populations of MDSC and Tregs have been found in patients with HCC.[29, 30] We observed that the liver populations of MDSC and Treg cells increase along with tumor development. This significant infiltration of regulatory cells does not happen in animals responding to the anti-CD137 mAb treatment, suggesting that in these animals the tumor environment is skewed toward a proinflammatory, efficacious anti-tumor immune response, maybe as a consequence of the action of the infiltrating M1 macrophages.

To summarize these results, we propose a model in which HCC development is accompanied both by a natural anti-tumor immune response (T lymphocytes and NK cells) and by suppressive strategies to overcome this immune response (MDSC and Treg cells). In some mice, the anti-CD137 mAb treatment tips the balance towards an efficacious anti-tumor response, and the early events identified for this to happen are a significant infiltrate of CD8+ T cells, a massive IFNγ secretion potential by various populations of NPC, and a sustained recruitment within the nodules and an activation of pro-inflammatory macrophages, which in turn increase the inflammatory environment surrounding the tumor nodules. In other mice, however, the immune response induced by the treatment is subdued, leading to an increased infiltration of regulatory immune cells, MDSC, Treg cells, which in turn inhibit the anti-tumor immune response and lead to further tumor growth. DEN-induced liver carcinogenesis is accompanied with local and systemic inflammatory reactions, with an accumulation of macrophages and lymphocytes and its clinical relevance is well established.[31, 32] Thus the mechanisms underlying the therapeutic effect might be comparable to that shown in the orthotopic model.

Anti-CD137 mAb has been used in combination with various other strategies in the treatment of other tumors.[33-35] Recently, the combination of anti-CD137 mAb with other immunostimulatory antibodies (anti-PD-L1 and anti-OX40) has been shown to lead to an extension of the survival in a transgenic model of HCC, however, none of the immunostimulatory Abs alone generated a therapeutic effect.[33] The difference in efficacy of the treatment with the anti-CD137 mAb between this study and ours could be due to the difference in the HCC models, which are of different “aggressiveness.” Indeed, the DEN treatment will lead to the initiation of some hepatocytes, and in a comparable manner, only a limited number of tumor Hepa1.6 cells are injected in the liver in the orthotopic model, whereas in the transgenic model, all the hepatocytes express the c-myc oncogene.

Increased frequency of MDSC has been demonstrated in patients with HCC, as well as in various murin models of the disease.[30, 36] We show here that the depletion of these cells with an anti-Gr1 Ab produces a modest therapeutic effect, however, the combination of the anti-CD137 mAb and anti-Gr1 Ab, leads to 80% of recovery. Although there is little neutrophil infiltration in the present model, we cannot exclude a role of these cells since they are also depleted by the anti-Gr1 Ab. Further more specific deletion experiments will be necessary to conclude on this matter.[37]

Finally, there have been some reports of liver toxicity associated with the anti-CD137 mAb treatment. A phase II clinical trial (NCT00612664) of the BMS-663513, the humanized anti-CD137 mAb, as a second line treatment of melanoma patients, was terminated due to unusual high incidence of hepatitis.[38] Liver toxicity was also observed in some mouse models[39, 40] and not in others,[34] nor in ours. These differences could be due to the antibody doses, and/or to the number and frequency of injections, which vary from one study to another, and require further in-depth analysis. A new clinical phase 1 study is currently evaluating the safety, and tolerability of low doses of BMS-663513 (NCT01471210).

In conclusion, this study demonstrates that stimulation of CD137 is a promising strategy for HCC immunotherapy. Our results provide justification for further investigation into the use of agonistic anti-CD137 mAb associated with first line conventional HCC treatments, and they also suggest that concomitant inhibition of regulatory immune cells may further increase anti-tumor immunity. Such a combination should be considered as a potential therapeutic option for treating HCC patients.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

The authors thank Pr RS Mittler for the 3H3 Hybridoma, Mr. Bernard Martinet and the academic platform “Protein production and purification” of the University of Nantes/INSERM UMR1064 for the production and purification of the anti-CD137 monoclonal antibody, and Lydie Guigand for technical help.

References

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Material and Methods
  4. Results
  5. Discussion
  6. Acknowledgements
  7. References
  8. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
ijc28943-sup-0001-suppinfo01.pdf359KSupporting Information

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.