Inhibition of STAT3 activity re-activates anti-tumor immunity but fails to restore the immunogenicity of tumor cells in a B-cell lymphoma model

Abstract

A large number of patients with advanced lymphoma become refractory or relapse after initial treatment due to the persistence of minimal residual disease. Ideal immunotherapy strategy for eradicating the minimal residual disease of lymphoma and preventing the tendency to relapse need to be developed. Here, we use a mice model mimicked the disease entities of aggressive B-cell lymphoma dynamically to analyze the host anti-lymphoma immunity during the progression of lymphoma. We have shown that STAT3 activity was gradually enhanced in host immune effector cells with the progression of lymphoma. Inhibition of the STAT3 activity with a small molecule inhibitor was able to effectively enhance the function of both host innate and adaptive immunity, and thereby delayed the progression of lymphoma. Despite the therapeutic benefits were achieved by using of the STAT3 inhibitor, disrupting of STAT3 pathway did not prevent the eventual development of lymphoma due to the presence of point mutation of β2M, which controls immune recognition by T cells. Our findings highlight the complexity of the mechanism of immune evasion; therefore a detailed analysis of genes involved in the immune recognition process should be essential before an elegant immunotherapy strategy could be conducted.

Introduction

In recent years, substantial advances have been achieved in the diagnosis and treatment of lymphoma, allowing some subtypes of lymphoma to go from uniformly lethal to potentially curable.^1,2 However, many patients with advanced lymphoma eventually become refractory to further treatment or relapse after initial treatment due to the persistence of minimal residual disease. These patients are considered incurable using the available treatment options, and there is an urgent need for the development of novel treatment modalities to improve clinical outcomes.^3,4

Active immunotherapy is a promising approach for lymphoma treatment and is now being investigated as a next generation strategy. Passive infusion of monoclonal antibodies against lymphoma antigens in combination with chemotherapy has shown remarkable clinical efficacy and become a standard front-line treatment.^1,3-5 However, much effort has been expended to search for other promising immunotherapies that are able to induce active and long-lasting immune responses against lymphoma. Many elegant immunotherapy strategies such as DNA vaccines and the adoptive transfer of anti-tumor T cells have been developed to activate immune responses in tumor-bearing hosts and are now being tested in clinical trials.^6-9 Despite some clinical efficacy of these strategies, the trial results have been generally disappointing, indicating that current immune strategies need to be further optimized for enhanced efficacy.

The activation of anti-lymphoma immunity requires both the re-activation of the host immune system and enhanced antigen-presenting function of the lymphoma cells themselves. Strategies fulfilling both requirements may lead to the effective eradication of lymphoma cells and the establishment of long-lasting immunity to inhibit the high tendency to relapse.^10,11 Signal transducer and activator of transcription 3 (STAT3), a negative regulator of both host anti-tumor immunity and the antigen-presenting function of tumor cells, is a potentially promising target for immune therapeutics.¹² In many patients with cancer, STAT3 is constitutively activated in the tumor and in various host immune cells and plays a critical role in escape from anti-tumor immunity.^12,13 Consistent with these findings, targeting STAT3 with small molecule inhibitors can elicit potent antitumor immunity and reverse immune tolerance by activating immune cells obtained from immune-compromised patients with cancer.^14-16 Furthermore, a recent study had shown that pharmacologic disruption of STAT3 activity in mantle cell lymphoma-bearing mice augmented the immunogenicity of lymphoma cells and led to the re-activation of anti-lymphoma immunity in vivo, indicating the therapeutic potential of STAT3 inhibitors as novel immune therapeutics against lymphoma.¹⁷

In this study, we have established a mouse model mimicking aggressive B-cell lymphoma. The direct effects of lymphoma on the host immune system were examined. In addition, a small molecule STAT3 inhibitor, WP1066, was tested for activity against lymphoma by reactivating host immunity in the B-cell lymphoma model. Our findings demonstrate that lymphoma can eventually escape WP1066-enhanced immunity due to the presence of a point mutation of β2M, which controls immune recognition by T cells. Our data provide a conceptual framework for the development of a novel strategy of anti-lymphoma immune therapy.

Results

Characterization of the B-cell lymphoma mouse model

To study the direct effects of lymphoma on the host immune system, we established a B-cell lymphoma mouse model following our previous description.¹⁸ This model was initially derived from a spontaneous lymphoma in a TA2 mouse and remained stably transplantable in TA2 mice by subcutaneously injecting B-cell lymphoma cells. The growth of the B lymphoma at primary inoculation sites was very rapid after injection of the lymphoma cells (Fig. 1A). No evidence of systemic dissemination to lymph nodes could be detected within 2 wk following inoculation of lymphoma cells. However, all mice eventually underwent systemic dissemination of lymphoma to distant lymph nodes 2 wk after inoculation of the lymphoma cells, which mimicked the clinical features characteristic of lymphoma in advanced clinical stages (Fig. 1B). The lymphoma tissues were sectioned for hematoxylin and eosin (H&E) staining for histological diagnosis. The malignant lymph nodes demonstrated a diffuse proliferation of immature lymphocytes with prominent nucleoli that had totally effaced the architecture typical of the non-malignant lymph node (Fig. 1C and D). The lymphoma cells expressed pan B-cell markers such as CD19 and CD20 and abnormally co-expressed CD5 antigens as detected by flow cytometry (Fig. 1E). The proliferative fraction as detected by Ki67 staining was higher than 70% (Fig. 1E). The analysis of tumor cells in the lymph node showed recurrent karyotype abnormality (Fig. 1F). The neoplastic cells of lymphoma tissues were positive for CD10 and BCL-6, negative for BCL-2 determined by immunohistochemical analysis (Fig. S1). Collectively, these results demonstrate that the present mouse model mimics aggressive B-cell lymphoma especially diffuse large B-cell lymphoma (DLBCL) and can be employed in subsequent experiments.

Figure 1. Characterization of a transplantable murine B lymphoma model. (A) Growth of lymphoma after inoculation. (B) The lymphoma TA2 mouse model shows the signs of advanced clinical stages of lymphoma, characterized by extensive growth of local and disseminated tumors. Red arrow: local tumor. Green arrow: disseminated tumor. (C) Representative histological image of a lymph node by H&E staining (100×). High-power magnification of lymph node tissue showing normal lymphocytes (200×). (D) Representative histological image of lymphoma by H&E staining (100×). High-power magnification of lymphoma tissue showing large lymphocytes with prominent nucleoli (200×). (E) Flow cytometric analysis of cells isolated from lymphoma tissues. Two-color analyses of lymphoma cells stained with CD3/CD19, CD5/CD20, and CD4/Ki67 are shown. (F) Representative image of karyotype abnormality is shown. Tumor lymph node cells displayed are hyperdiploid. Red arrow: chromosomal structure abnormality.

Host immunosurveillance was subverted during lymphoma progression

To study the changes in host immunosurveillance in the lymphoma mouse model, we constantly monitored mice for functional changes in innate and adaptive immune effector cells during tumor progression. We initially chose to determine the functional status of macrophages because they are one of the major effector cell types in the innate immune system and are capable of both inhibiting and promoting tumor progression depending on their functional status. The release of nitric oxide, a surrogate marker that reflects the capability of macrophages to kill infectious organisms or tumor cells, was significantly decreased in peritoneal macrophages from lymphoma-bearing mice in comparison to control mice, implicating that anti-tumor functions of macrophages in host mice were impaired with the progression of lymphoma (Fig. 2A). Furthermore, levels of phosphorylated STAT3 protein in macrophages were profoundly increased in lymphoma-bearing mice but not in control mice, indicating the macrophages in lymphoma-bearing mice could favor rather than inhibit lymphoma progression as increased STAT3 activity suppresses tumor immunity (Fig. 2B). During the progression of lymphoma in TA2 mice, the numbers of another major effector cells type in the innate immune system, NK cells, were significantly decreased, again indicating a compromised function of the innate immune system due to lymphoma (Fig. 2C). Next, we determined whether the functions of the adaptive immune system were also impaired by progression of lymphoma. We initially investigated the functional status of bone marrow-derived dendritic cells (DCs) in mice with or without lymphoma because these cells are the most potent antigen-presenting cells and are crucial for the induction and maintenance of anti-tumor immune responses. Bone marrow-derived DCs isolated from lymphoma-bearing mice produced low levels of IL-12 in response to ex vivo stimulation with lipopolysaccharide, indicating that the production of cytokines required for antigen-specific T cell stimulation was impaired by the progression of lymphoma (Fig. 2D). Consistently, bone marrow-derived DCs from tumor-bearing mice were also considerably less potent at inducing autologous CD4+ T cell proliferation in vitro (Fig. 2E). Similar to the finding in macrophages, the levels of phosphorylated STAT3 protein in DCs were profoundly increased in lymphoma-bearing mice but not in control mice, indicating that the DCs in lymphoma-bearing mice should functionally favor lymphoma progression (Fig. 2F). Because T cells are the principal mediators of anti-tumor immunity, we examined the functional status of T cells in mice with or without lymphoma. The T cells in lymphoma-bearing mice generated a weaker cytotoxic T cell response as determined by IFN-γ ELISPOT assays than did the T cells from the control mice (Fig. 2G). The T cells from peripheral blood displayed higher STAT3 activity in tumor-bearing mice vs. control mice (Fig. 2H). To study whether extrinsic cellular suppression of anti-tumor immunity was influenced by the progression of lymphoma, spleen cells from tumor-bearing or tumor-free mice were harvested, and numbers of regulatory T cells were determined. As expected, the progression of lymphoma promoted a significant increase in regulatory T cells, thus favoring immune evasion (Fig. 2I). Collectively, these data indicate that progression of lymphoma eventually subverts host immunosurveillance at multiple levels.

Figure 2. Host immunosurveillance was subverted during lymphoma progression. (A) Reduced production of nitric oxide was detected in peritoneal macrophages from lymphoma-bearing mice. Data are presented as means ± SD of quadruplicates. (B) Increased phospho-STAT3 expression was found in macrophages upon tumor progression. Representative results of 3 independent experiments with 4 mice per group are shown. (C) Splenic NK cells from tumor-bearing or control mice were analyzed by FACS using a CD49b-specific antibody. Data are represented as mean ± SD of quadruplicates. (D) ELISA assays of IL-12 expression in CD11c⁺ DCs purified from tumor-bearing or tumor-free mice after lipopolysaccharide stimulation. Data are represented as mean ± SD of quadruplicates. (E) Proliferation of T cells stimulated by DCs from tumor-bearing or tumor-free mice. Data are shown as means ± SD; n = 4. (F) STAT3 was constitutively activated in DCs as determined by western blotting. Representative results of 3 independent experiments with 4 mice per group are shown. (G) T cells from control TA2 mice were able to mount stronger responses against an endogenous lymphoma tumor antigen than T cells from lymphoma-bearing mice as assessed by IFN-γ ELISPOT. Data shown are the mean numbers of lymphoma-specific IFN-γ-producing spot forming cells from 8 separate mice per group analyzed individually. (H) T cells showed increased phospho-STAT3 activity along with tumor progression. (I) Population of Treg cells from tumor-bearing mice was increased.*P < 0.05; **P < 0.01; ***P < 0.001.

Optimizing the dosing schedule of WP1066 for targeted disruption of the STAT3 signaling pathway in vivo

To study the effects of inhibiting STAT3 on anti-tumor immunity in lymphoma-bearing mice, we sought to optimize the dosing schedule of WP1066, a potent STAT3 inhibitor, for targeted disruption of the STAT3 signaling pathway in vivo. The plasma WP1066 concentrations were kinetically monitored after intravenous administration of WP1066 at doses of 5, 10 or 20 mg/kg every other day for up to 14 d in the lymphoma-bearing mice (Fig. 3A; Fig. S2A). While WP1066 intravenously injected at a dose of 5 mg/kg was not sufficient to inhibit the phosphorylation of STAT3 in splenocytes from lymphoma-bearing mice (Fig. S2B), this small molecule induced persistent inhibition of the phosphorylation of STAT3 at a dose of 10 mg/kg (Fig. 3B). To determine the impact of WP1066 on STAT3 activity, apoptosis and cell cycle progression of tumor cells, lymphoma cells, and B16 cells were exposed to varying concentrations of WP1066 and subjected to further analysis. In both lymphoma cells and B16 cells, WP1066 at a concentration of 1 μM was enough to inhibit the phosphorylation of STAT3 (Fig. 3C). While B16 cells were sensitive to WP1066-induced apoptosis, lymphoma cells were resistant to killing by WP1066 even at the highest concentration of 10 μM (Fig. 3D). Furthermore, treatment of lymphoma cells with 1 μM of WP1066 did not induce cell cycle arrest (Fig. 3E). These data indicate that WP1066 at doses of 10 mg/kg in the lymphoma-bearing mice were sufficient to disrupt STAT3 signaling pathways in both tumor and immune effector cells, leading to some apoptosis. Thus, this dosing schedule of WP1066 was used for subsequent experiments.

Figure 3. Optimizing the dosing schedule of WP1066. (A) Systemic administration of WP1066 i.v. at dose of 10 mg/kg every other day for 2 wk achieved stable plasma concentrations exceeding 1 μM. Plasma was analyzed for WP1066 content using tandem liquid chromatography/mass spectrometry. (B) Western blotting analysis showed expression of phosphorylated (p) STAT3 and total STAT3 proteins in splenic cells from tumor-bearing mice treated with WP1066 or not treated with inhibitor. (C) B16 and lymphoma cells were incubated with 1 μM of WP1066 for 24 h and 48 h. Western blotting was performed to analyze the expression of p- STAT3 and total STAT3 proteins. (D) Sensitivity of tumor cells to WP1066-induced apoptosis in vitro was determined by Annexin V staining. B16 cells, sensitive to WP1066-induced apoptosis, served as a positive control. (E) Cell cycle analysis was performed by propidium iodide staining at 48 h after WP1066 treatment.

Targeted disruption of STAT3 activity re-stimulated anti-tumor immunity and delayed the progression of lymphoma in the TA2 mouse model

To investigate the impact of targeted disruption of STAT3 on the progression of lymphoma, intravenous WP1066 was given to TA2 mice every other day for up to 14 d, starting 1 d after inoculation of lymphoma cells. The lymphoma-bearing TA2 mice were then monitored for anti-tumor immunity and progression of lymphoma. Treatment with WP1066 effectively inhibited the activation of STAT3 in macrophages and bone marrow-derived DC in lymphoma-bearing mice (Fig. 4A and B). Consistent with the inhibition of STAT3 pathway, the expression of co-stimulatory molecules and the release of nitric oxide were significantly enhanced in macrophages from WP1066-treated tumor-bearing mice when compared with lymphoma-bearing mice without treatment with the STAT3 inhibitor, indicating that the capabilities of macrophages to kill infectious organisms or tumor cells were improved by WP1066 (Fig. 4C and D). In addition, the expression of co-stimulatory molecules and production of IL-12 by DC were also significantly enhanced by treatment with WP1066 (Fig. 4E and F). To directly evaluate the effects of WP1066 on the anti-lymphoma functions of T cells, CD8⁺ T cells were purified from lymphoma-bearing mice and their efficiency of killing was examined in a co-culture system. CD8⁺ T cells from WP1066-treated tumor-bearing mice displayed significantly enhanced killing efficacies when compared with those from PBS-treated lymphoma-bearing mice (Fig. 4G). Next, we determined the degree of infiltration of T cells within the lymphoma tissues, as this infiltration is considered to be crucial for the induction of an antitumor response.^19,20 Immunofluorescent staining of lymphoma tissues from WP1066-treated tumor-bearing mice showed a substantially higher infiltration of T lymphocytes compared with those from PBS-treated tumor-bearing mice (Fig. 4H).

Figure 4. Targeted disruption of STAT3 activity with WP1066 re-stimulated anti-tumor immunity. TA2 mice were inoculated with 1 × 10⁶ lymphoma cells in the inguinal groove muscle and the lymphoma-bearing TA2 mice were treated with WP1066 (injected i.v. at doses of 10 mg/kg every other day for up 2 wk) or PBS. (A) Macrophages from tumor-bearing mice treated with WP1066 or PBS were harvested at 0, 6, or 14 d post-inoculation. The P-STAT3 was detected by western blotting using an anti-p-STAT3 antibody. The results shown are representative of more than 3 independent experiments. (B) BM-derived DCs were harvested at 0, 6, and 14 d post-inoculation. The phosphorylation of STAT3 was detected by western blotting. The results shown are representative of 3 independent experiments. (C) The percentages of CD11b⁺CD86⁺ macrophages from tumor-bearing mice were increased when mice were treated with WP1066. Data are shown as means ± SD; n = 4. (D) Increased production of nitric oxide was found in peritoneal macrophages from lymphoma-bearing mice treated with WP1066. Data are shown as means ± SD; n = 4. (E) The percentages of CD11c⁺CD86⁺ DCs were upregulated when mice were treated with WP1066 as determined by flow cytometry. Data are shown as means ± SD; n = 4. (F) IL-12 production by CD11c⁺ DCs from PBS- or WP1066-treated tumor-bearing mice was determined by ELISA. Data are shown as means ± SD; n = 4. (G) CD8⁺ T lymphocytes from PBS- or WP1066-treated tumor-bearing mice were incubated with tumor lysate and DCs. The resulting CTLs were added to tumor cells and cultured for 4 h. The culture supernatant was collected and subjected to an LDH-release assay for specific lysis. Data are represented as mean ± SD of triplicates. (H) Representative immunofluorescence images (400×) of infiltrating CD4 and CD8 lymphocyte populations (green staining, arrows) are shown against the background of tumor cells (red staining) from lymphoma-bearing mice treated with PBS or WP1066. Data in columns are representative of 4 experiments and are represented as means ± SD *P < 0.05.

Because innate and adaptive immunity can be effectively activated by disrupting STAT-3 signaling, we examined the effects of WP1066 on the development and progression of lymphoma. Treatment with WP1066 delayed the onset of lymphoma and inhibited the growth of the tumor (Fig. 5A and B). However, all tumor-bearing mice eventually developed lymphoma within 15 d of inoculation of tumor cells despite the treatment with WP1066 (Fig. 5A). The systemic dissemination of the lymphoma to distant lymph nodes was significantly delayed (Fig. 5C). To further address whether the inhibitory effects of WP1066 on the development and progression of lymphoma were due to the re-activation of anti-tumor immunity and were independent of direct effects of WP1066 on the tumor cells themselves, we examined the effect of WP1066 on the progression of lymphoma in NOD/SCID immune- deficient mice. While the overall survival of the tumor-bearing mice was significantly improved by treatment with WP1066 in immunocompetent lymphoma-bearing TA2 mice, WP1066 did not demonstrate any therapeutic benefit in terms of overall survival in NOD/SCID lymphoma-bearing mice (Fig. 5D). There was not a significant difference in the kinetics of onset and growth of lymphoma between PBS- or WP1066-treated NOD/SCID lymphoma-bearing mice (Fig. 5E and F), indicating that the anti-tumor effects of WP1066 required the participation of the host immune response.

Figure 5. Targeted disruption of STAT3 activity delayed the progression of lymphoma. TA2 or NOD/SCID mice were inoculated in the inguinal groove muscle with 5 × 10⁵ lymphoma cells and treated with PBS or WP1066 (10 mg/kg i.v. every other day for up 2 wk). Each group included 10 mice. (A) Cumulative probability of tumor onset in TA2 mice was determined by the appearance of neoplasms in the inguinal groove muscle. In the PBS-treated group, lymphoma formation at the inoculation sites took an average of 10.0 ± 1.69 d, whereas WP1066 treatment significantly prolonged the formation of lymphomas at primary sites to 13.5 ± 1.20 d (P = 0.003). All mice eventually exhibited signs of tumors. (B) Tumor sizes in TA2 mice were monitored every 2–3 d and the results were converted to tumor volumes and plotted. Data are analyzed from 8 tumor-bearing mice for each point. (C) Cumulative probability of tumor dissemination in TA2 mice was determined by the appearance of neoplasms in the ipsilateral or contralateral axillary region or the contralateral inguinal region. In the PBS-treated group, lymphoma dissemination took an average of 15.4 ± 1.10 d, whereas WP1066 treatment significantly prolonged the dissemination time of lymphomas to 17.6 ± 1.10 d (P = 0.0002). (D) TA2 or NOD/SCID mice bearing lymphoma received systemic administration of WP1066 i.v. at dose of 10 mg/kg or PBS every other day for 2 wk following tumor inoculation. Survival rate of mice was monitored and n = 8 mice/group. The P < 0.01 for the TA2 mice treated with WP1066 compared with the other groups. (E) Tumor growth curve of tumor-bearing NOD/SCID mice treated with PBS or WP1066. The volume of the re-challenged tumors was calculated based on caliper measurements of tumor size at 3-d intervals. Values are the mean ± SD of 8 tumor-bearing mice per group. (F) Cumulative probability of tumor onset in NOD/SCID mice treated with PBS or WP1066. Data from 8 tumor-bearing mice are analyzed for each point.

The lymphoma cells contained mutations in genes controlling immune recognition by T cells that could not be overcome by WP1066

The failure of disruption of the STAT3 pathway to prevent the eventual development of lymphoma suggests that other dominant mechanisms may exist to enable the lymphoma cells to evade host anti-tumor immunity. Therefore, we examined the lymphoma cells for the expression of various cell-surface molecules involved in immune recognition and antigen presentation. The lymphoma cells expressed MHC class II and co-stimulatory molecules (CD40 and CD86) (Fig. 6A). However, the expression level of MHC class I was nearly undetectable on the surface of the lymphoma cells (Fig. 6A), representing a potential mechanism by which the TA2 lymphoma cells circumvent immune surveillance. We then examined the lymphoma cells for mutations in genes controlling immune recognition by T cells, including CD74, CD86, CIITA, PDL1, and β2M (Table 1). Sequencing analysis revealed a homozygous T-to-C mutation at nucleotide 65 in exon1 of the β2M gene transcript in TA2 lymphoma cells, leading to the substitution of valine to an alanine at position 5 (Fig. 6B). β2M is an essential component of MHC class I molecules and is necessary for cell-surface expression of the MHC class I molecule and the stability of the peptide-binding groove.²¹ We then examined the functional consequences of this β2M structural alteration. Functional prediction of the missense mutation using the AGVGD (http://agvgd.iarc.fr/) method based on protein sequence homology demonstrated that the gene lesion was likely to be deleterious (class C65 of substitution, most likely to interfere with function). Immunohistochemistry analysis of tumor biopsies further confirmed that the missense mutation led to loss of expression of the β2M protein (Fig. 6C). Next, we tested whether expression of the MHC class I antigen could be restored with WP1066. Although B16 cells also expressed low levels MHC class I, the β2M gene in these cells is wild-type. Interestingly, expression of the MHC class I antigen in B16 cells could be enhanced by IFN-α but not by WP1066. However, neither WP1066 nor IFN-α could restore the expression of the MHC class I antigen on lymphoma cells, indicating that the loss of MHC class I caused by gene lesions could not be overcome by pharmacologic agents (Fig. 6D).

Figure 6. Lymphoma cells with mutations in genes that control immune recognition by T cells could not be overcome by WP1066. (A) Lymphoma cells were stained with Abs against various molecules involved in signal 1 and signal 2 and analyzed using multi-parametric flow cytometry. The isotype control is shown as a green line and the respective target antigen is shown as a purple line. (B) Sequence traces of the lymphoma cells and the A20 tumor cell line. (C) Immunohistochemistry analysis of β2M in tumor biopsy samples (400×). (D) B16 cells and lymphoma cells were stained for surface expression of MHC class I after exposure to IFN-α, WP1066, or the combination of WP1066 and IFN-α. Isotype control is shown as a green line and MHC class I expression is shown as a purple line.

Table 1. Appearance of genes controlling immune recognition by T cells in lymphoma cells.

Gene	Aberrationa	Proteinb
CD74	WT	Positive
CD86	WT	Positive
CIITA	WT	Positive
PDL1	WT	Positive
β2M	M	Negative

^a Compared to the corresponding sequence of TA2 spleen lymphocytes; ^bas assessed by IHC or immunofluorescence. Abbreviations: M, mutation; WT, wild-type.

Discussion

The successful application of allogenic hematopoietic stem cell transplantation and monoclonal antibodies in treating lymphoma highlight the potential role of immunotherapy to cure refractory lymphoma and prevent the high tendency of these patients to relapse. Although passive infusion of monoclonal antibodies has achieved great clinical success to date, the novel immunotherapies that attempt to induce active and long-lasting immune responses have generally been disappointing in clinical trials, indicating that current immune strategies need to be further optimized.

To improve the clinical efficacy of active immune therapies, key mechanistic details should be investigated to better understand the strengths and limitations of the therapeutics. In the present study, we used a transplantable lymphoma TA2 mice model that mimicked aggressive B-cell lymphoma. Unlike genetically engineered lymphoma models and humanized lymphoma mouse models, the TA2 lymphoma model mimics the clinical course of established, aggressive human B-cell lymphoma and provides a valuable tool for dynamic analysis of host anti-lymphoma immunity during the progression of lymphoma. While it is well known that B-cell lymphomas express a variety of soluble or membrane-anchored factors to influence the host immune responses,^22,23 it has never been systematically and kinetically determined whether the progression of lymphoma directly disables host anti-tumor immunity in vivo. Thus, the findings presented here provide direct evidence that B-cell lymphomas actively suppress the host anti-tumor immunity at multiple levels and provide the rationale for using immunotherapies to curtail and possibly reverse the progression of lymphoma.

Current immune therapy includes several approaches such as idiotype vaccines, DNA vaccines, engineered antigen-specific T cells and small molecule inhibitors, among which STAT3 inhibitors are thought to hold promise as one of the most effective immune strategies against cancer. Constitutively activated STAT3 had been reported in many types of lymphoma cells as well other immune cells that serve as critical mediators to induce immune tolerance.^24-28 In a previous study, ablating STAT3 in hematopoietic cells activated multicomponent therapeutic anti-tumor immunity.¹³ As STAT3 inhibitors were able to enhance both anti-tumor immunity and the antigen-presenting function of the lymphoma cells in a recent study,¹⁷ these molecules could potentially be ideal immune therapeutics for eradicating the minimal residual disease of lymphoma and preventing the tendency to relapse. In this study, we show that STAT3 activity was gradually enhanced in host immune effector cells with the progression of lymphoma. Inhibition of the STAT3 activity with a small molecule inhibitor was able to effectively enhance the function of host innate and adaptive immunity and thereby delay the progression of lymphoma. While therapeutic benefits were achieved by using the STAT3 inhibitor, disrupting the STAT3 pathway did not prevent the eventual development of lymphoma in the present study. These findings imply that the STAT3 inhibitor has therapeutic limitations. In contrast to other studies, the use of the STAT3 inhibitor was not able to enhance the expression of the MHC class I antigen on the surface of TA2 lymphoma cells and failed to augment the immunogenicity of tumor cells. Obviously, the inability to restore the immunogenicity of TA2 lymphoma cells with the STAT3 inhibitor greatly limited the therapeutic efficacy and led to the eventual failure of the therapy.

The recognition and killing of tumor cells by cytotoxic T lymphocytes (CTL) requires the participation of MHC class I.²⁹ Loss of HLA class I expression is common in cancer cells and appears to be one of the major strategies used by cancer cells to evade host immunosurveillance.^30-32 Several mechanisms such as abnormal signal transduction, epigenetic modifications, and gene structural alterations are responsible for the loss of the HLA class I antigen.³³ In the case of TA2 lymphoma-bearing mice, the loss of MHC class I molecules was caused by a missense mutation in the β2M gene. Therefore, it is not surprising that the expression of the HLA class I antigen could not be restored by pharmacologic agents in TA2 lymphoma cells. In the case of B16 cells, however, expression of the HLA class I antigen could be enhanced with IFN-α because no gene structural alteration was detected in the machinery required to assemble the HLA class I antigen complex. The present findings seem highly clinically relevant and informative. Recently, investigators have shown that gene mutations necessary for the recognition and killing of lymphoma cells by CTL or NK cells were very common in diffuse large B-cell lymphomas (DLBCL).^34,35 These findings are consistent with our data and strongly support our conclusions.

Our findings have important clinical implications. Active immunotherapy is a promising approach and is now being investigated as a next generation strategy for lymphoma treatment. Whichever active immunotherapy approach is used, the recognition of lymphoma cells by immune surveillance mechanisms will be a prerequisite for yielding real therapeutic benefits. Our study highlights the complexity of the mechanisms of immune escape. While the ability of an immunotherapy to re-activate host immunity is an important factor in determining anti-tumor efficacy, gene structural alterations involved in immune recognition before an immune therapy are also critical determinants of therapeutic efficacy. We propose that a detailed analysis of genes involved in the immune recognition process is essential before any immune therapy can be conducted.

Materials and Methods

Animals and cell lines

Male and female TA2 mice were obtained from the Animal Experimental Center of Tianjin Cancer Institute. NOD/SCID mice were purchased from the Shanghai Experimental Animal Center of the Chinese Academy of Sciences. Animals were handled in the Animal Facility at Huazhong University of Science and Technology (HUST). All studies involving mice were approved by the Animal Care and Use Committee of HUST. The murine melanoma B16 cell line and murine lymphoma A20 cell line were purchased from the China Center for Type Culture Collection.

STAT3 inhibitor

WP1066 was purchase from Merck and dissolved in dimethyl- sulfoxide (DMSO, Sigma-Aldrich) to make a stock solution that was serially diluted to the desired concentration with RPMI 1640 medium.

Animal experiments

All animal experiments described in this study was performed in accordance with and approved by the Institutional Review Board for Animal Welfare at the TongJi Hospital and Medical College at the Huazhong University of Science and Technology. TA2 mice were inoculated with 5 × 10⁵ lymphoma cells in the left inguinal groove muscle. The functional changes in innate and adaptive immune effector cells during tumor progression were observed. To optimizing the dosing schedule of WP1066, WP1066 was injected i.v. at doses of 5, 10, and 20 mg/kg every other day for 2 wk. Plasma was harvested after blood collection, and the WP1066 concentration was analyzed using tandem liquid chromatography/mass spectrometry. To determine the effect of WP1066 on antitumor immunity, WP1066 was given to TA2 mice every other day for up to 14 d starting 1 d after inoculation of lymphoma cells. The immune function and the STAT3 activity of immune cells were examined. The onset of the primary tumor was defined as a tumor diameter over 0.2 cm³, and lymphoma dissemination was defined as a remote lymph node diameter over 0.2 cm³. In all animal experiments, the growth and dissemination of lymphoma was monitored 3 times per week until the mice were sacrificed.

Preparation of cells

Bone marrow cells were harvested from femurs and tibias and single-cell suspensions were cultured in control or lymphoma-conditioned medium in the presence of 20 ng/mL recombinant murine GM-CSF (PeproTech) as described elsewhere.³⁶ Eighteen hours before harvesting the cells, dendritic cells (DC) were matured with 1 mg/mL LPS (Sigma) and pulsed with tumor lysate. Peritoneal macrophages were prepared as previously described.³⁷ Briefly, mice were injected intraperitoneally (i.p.) with RPMI-1640 medium without fetal bovine serum. Peritoneal macrophages were obtained by peritoneal lavage.

Flow cytometry and cell sorting

For phenotyping and sorting, spleens, lymph nodes, and tumors were processed into single-cell suspensions and these cells were labeled with fluorochrome-conjugated antibodies (Abs). Isotype-matched mAbs were used in the control samples. Flow cytometry was performed using a FACSCalibur flow cytometer (BD Biosciences), and data were analyzed using CellQuest (BD Biosciences) and FlowJo (Tree Star) software. CD8⁺ cells were sorted using a FACS Vantage cell sorter (BD Biosciences). Sorted cells were reanalyzed by flow cytometry and found to be 95% pure.

Western blotting and immunofluorescence

Preparation of protein samples and western blots were performed as previously described.³⁸ Phospho-STAT3 (Tyr705) Antibody (Cell Signaling Technology) was used to detect the activation of STAT3. For immunofluorescence, tumor tissues were surgically excised and fixed in 4% paraformaldehyde, embedded in paraffin and sectioned. Biotinylated anti-mouse CD4 (clone GK1.5) and CD8 antibodies (clone 53–6.7) purchased from Biolegend were used as primary antibodies. After incubation with primary antibodies, the sections were then incubated with avidin-FITC for 30 min and observed with a confocal laser-scanning microscope (German Leica, true confocal scanner spectrophotometry).

ELISPOT assay and cytokine release assay

Splenocytes were harvested from mice challenged subcutaneously with 5 × 10⁵ lymphoma cells. One million splenocytes were seeded into each well of a 96-well filtration plate in the presence or absence of tumor lysate and incubated at 37 °C for 24 h. We performed an ELISPOT assay to detect specific IFN-γ–positive spots according to the manufacturer’s protocol (Cell Sciences) and scanned and quantified the plates using an Immunospot Analyzer from Cellular Technology Ltd. Culture supernatants were assayed for mIL-12 using a specific ELISA kit (Ebioscience).

Mutation analysis

RNA was extracted from lymphoma cells and normal lymphocytes from TA2 mice and then cDNA encoding genes controlling immune recognition by T cells, including CD74, CD86, CIITA, PDL1, and β2M, was obtained by RT-PCR. Mutation analysis was performed by sequencing PCR products using an ABI 3130XL analyzer.

Histology and immunohistochemistry

Lymphoma tissues from TA2 tumor-bearing mice were surgically excised, fixed overnight in 4% formalin, embedded in paraffin, and sectioned for H&E staining and immunohistochemistry. An anti-mouse β2M antibody was purchased from Lifespan Biosciences, Inc.

Statistics

Results are expressed as means ± SD and were analyzed with an ANOVA test for repeated measures and a Kaplan–Meier test. Differences were considered to be statistically significant when P < 0.05.

10.4161/cbt.29453