ORIGINAL ARTICLE
See related commentary on pg 1794
Intratumoral CD4 þ T Lymphodepletion Sensitizes Poorly Immunogenic Melanomas to Immunotherapy with an OX40 Agonist Susumu Fujiwara1, Hiroshi Nagai1, Noriko Shimoura1, Shuntaro Oniki1, Takayuki Yoshimoto2 and Chikako Nishigori1 Previous studies have shown that the antitumor effects of OX40 agonists depend on the immunogenicity of the tumor and that poorly immunogenic tumors such as B16F10 melanomas do not respond to OX40 agonist treatment. In this study, we have shown that intratumoral CD4 þ T lymphodepletion sensitized poorly immunogenic B16F10 melanomas to immunotherapy with an OX40 agonist. CD4 þ T lymphodepletion dramatically altered the tumor immune microenvironment, making it more susceptible to the antitumor effects of an OX40 agonist by enhancing the accumulation of CD8 þ T cells and natural killer (NK) cells in tumor tissue. However, unexpectedly, the number of CD11b þ Gr-1 þ myeloid-derived suppressor cells (MDSCs) within tumor tissues also significantly increased as a result of CD4 þ T lymphodepletion. As a countermeasure against CD8 þ T-cell accumulation, CCR2positive CD11b þ Gr-1int (monocytic) MDSCs predominantly increased. Treatment with an OX40 agonist under CD4 þ T lymphodepletion neither reduced MDSCs nor increased CD8 þ T cells and NK cells, but further enhanced the expression of cytotoxic molecules from tumor-infiltrating effector cells. Our results suggest that combined immunotherapy using both an OX40 agonist and CD4 þ T lymphodepletion could be a promising therapeutic strategy for poorly immunogenic tumors and might be more effective if further combined with a therapeutic strategy targeting MDSCs. Journal of Investigative Dermatology (2014) 134, 1884–1892; doi:10.1038/jid.2014.42; published online 27 February 2014
INTRODUCTION OX40 (CD134) is an B50-kDa glycoprotein with three complete and one truncated cysteine-rich domains that are characteristic of the TNF receptor superfamily. OX40 is predominantly expressed on activated CD4 þ T cells and CD8 þ T cells; however, other cell types, including natural killer (NK) cells, NK T cells, neutrophils and CD4 þ Foxp3 þ regulatory T cells (Tregs), also express OX40 (Sugamura et al., 2004; Croft, 2010; Ishii et al., 2010; Weinberg et al., 2011). Several studies have shown that OX40 agonists have potent antitumor activities in preclinical models (Weinberg et al., 2000; Piconese et al., 2008; Jensen et al., 2010). However, these antitumor effects were found to be dependent on the immunogenicity of the tumor, because poorly immunogenic tumors did not respond to OX40 agonist treatment (Kjaergaard 1
Division of Dermatology, Department of Internal Related, Kobe University Graduate School of Medicine, Kobe, Japan and 2Department of Immunoregulation, Institute of Medical Science, Tokyo Medical University, Tokyo, Japan Correspondence: Hiroshi Nagai, Division of Dermatology, Department of Internal Related, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan. E-mail: [email protected] Abbreviations: MDSC, myeloid-derived suppressor cell; NK, natural killer Received 7 March 2013; revised 8 December 2013; accepted 23 December 2013; accepted article preview online 27 January 2014; published online 27 February 2014
1884 Journal of Investigative Dermatology (2014), Volume 134
et al., 2000). A recent study highlighted the significance of combining OX40 agonist therapy with other therapeutic approaches for poorly immunogenic tumors (HirschhornCymerman et al., 2009). B16F10 melanoma cells are poorly immunogenic tumor cells that originally spontaneously developed in C57BL/6 mice and are thought to reflect the poor immunogenicity of metastatic tumors in humans (Fidler, 1973). B16F10 melanoma, along with other poorly immunogenic tumors, has been shown to be unresponsive to immunotherapy with an OX40 agonist (Hirschhorn-Cymerman et al., 2009). Accumulating evidence has shown that Tregs have a crucial role in refractoriness to immunotherapy, and high levels of Tregs have been detected in B16F10 tumor tissues (Kryczek et al., 2007). Furthermore, other CD4 þ regulatory T cells, such as TR1 cells and Th3 cells, have also been shown to be involved in tumor immune escape (Seo et al., 2001; Lo´pez et al., 2006). As an OX40 agonist can activate not only antitumor effector cells but also Tregs (Ruby et al., 2009; Xiao et al., 2012), we hypothesized that depletion of the CD4 þ regulatory T cells in tumor tissue might sensitize B16F10 melanomas to immunotherapy with an OX40 agonist. To test this possibility, we intratumorally administered an anti-CD4 mAb on day 7 after tumor inoculation. This approach efficiently depleted Tregs, as well as helper T cells and other CD4 þ cells in B16F10 melanomas and sensitized these tumor tissues to immunotherapy with an OX40 agonist. Interestingly, & 2014 The Society for Investigative Dermatology
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
RESULTS Intratumoral CD4 þ T lymphodepletion sensitizes poorly immunogenic melanomas to immunotherapy with an OX40 agonist
Initially, we examined the antitumor effects of an OX40 agonist on poorly immunogenic B16F10 melanomas. As shown in Figure 1a, treatment with OX40 agonist (starting on day 0 after tumor inoculation) did not exert antitumor effects. Next, we tested whether the OX40 agonist could exert antitumor effects on B16F10 melanomas when combined with cytokines that are involved in antitumor immunity. Treatment with the OX40 agonist exerted significant antitumor effects on IL-12 cDNA-transfected B16F10 melanomas, but not on IL-27 cDNA-transfected B16F10 melanomas (Figure 1b and c). In contrast, when combined with a single, intratumoral administration of an anti-CD4 mAb (for CD4 þ T lymphodepletion), OX40 agonist treatment had significant antitumor effects when started on day 7 (Figure 1d) or on day 0 (data not shown). Flow cytometric analysis showed that, in control mice, which were intratumorally administered rat IgG, 23.8% of the tumor-infiltrating mononuclear cells were CD4 þ after 7 days of administration, whereas in anti-CD4 mAb-treated mice, the proportion of CD4 þ cells decreased to 0.67% at 7 days after administration of the anti-CD4 mAb. The proportion of Treg cells (CD4 þ Foxp3 þ cells) among the tumor-infiltrating mononuclear cells was 0.11% after antiCD4 mAb administration (vs. 7.4% in the control) (Supplementary Figure S1 online). Other CD4 þ cell populations (CD4 þ F4/80 þ cells, CD4 þ CD11b þ cells, and CD4 þ CD11c þ cells) were barely (o0.3% of tumor-infiltrating mononuclear cells) detected in the control tumor tissues on day 14 and further decreased (to o0.03%) following antiCD4 mAb administration (data not shown). These results indicated that the CD4-positive cells, including Tregs, in tumor tissues were almost completely depleted on day 7 after intratumoral injection of an anti-CD4 mAb. Intratumoral CD4 þ T lymphodepletion dramatically alters the immune microenvironment in the tumor, making it more susceptible to antitumor effects by enhancing the accumulation of CD8 þ T cells and NK cells
To clarify why intratumoral CD4 þ T lymphodepletion sensitizes B16F10 melanomas to immunotherapy with an OX40 agonist, we examined the cellular and humoral alterations in CD4 þ T-lymphodepleted tumor tissues. Flow cytometric analysis showed that, in the control group, CD8 þ T-cell infiltration into B16F10 tumors was scarce and NK cell infiltration was considerable. However, after CD4 þ
a
b
600
600 No treatment OX40 agonist
500
Tumor size (mm2)
administration of just a single dose of anti-CD4 mAb into B16F10 tissues dramatically altered the immune conditions and augmented the infiltration of effector cells into the tumors. However, unexpectedly, CD4 þ T lymphodepletion also augmented the recruitment of myeloid-derived suppressor cells (MDSCs) into the tumor tissues. Finally, we analyzed the mechanism by which combined treatment with an OX40 agonist and CD4 þ T lymphodepletion exerts antitumor effects.
400
400
300
300
200
200
100
100
0
0
10
20
30
0
40
c
d
600
600
500
No treatment OX40 agonist
500
No treatment OX40 agonist
*
0
400
300
300
200
200
100
100
20
30
40
No treatment Anti–CD4 mAb Anti–CD4 mAb + OX40 agonist
500
400
10
*
0 0
10
20
30
40
0
*
0
10
20
30
40
Day
Figure 1. Antitumor effects of an OX40 agonist on B16F10 melanomas. Onehundred thousand tumor cells (a and d, B16F10 cells; b, IL-12 cDNA-transfected B16F10 cells; c, IL-27 cDNA-transfected B16F10 cells) were subcutaneously inoculated into the right flank of C57BL/6 mice (n ¼ 5 per group) on day 0. (a–c) Mice were intraperitoneally (i.p.) administered 1 mg of OX86 on day 0 and every 5 days thereafter. (d) Anti-CD4 mAb (500 mg) was intratumorally injected only once on day 7. OX86 was i.p. administered on day 7 and every 5 days thereafter. The tumor size was measured as the product of the longest diameter and the perpendicular diameter. The bars represent mean þ SEM (*Po0.05). The data shown are from one of three individual experiments.
T lymphodepletion, both CD8 þ T cells and NK cells markedly infiltrated into tumor tissues (Figure 2). The cytokine profile of the tumor was also dramatically altered by CD4 þ T lymphodepletion. Cytokines with antitumor activities, such as IFN-g, TNF-a, IL-2, and IL-12, were significantly upregulated by CD4 þ T lymphodepletion (Figure 3a). However, interestingly, the expression of IL-23, which has been reported to inhibit the accumulation of antitumor CD8 þ T cells (Langowski et al., 2006), significantly reduced. After CD4 þ T lymphodepletion, the expression levels of the immunosuppressive cytokine TGF-b did not change significantly compared with the levels in control tumor tissues (data not shown). However, the expression of IL-10 showed a tendency to increase after CD4 þ T lymphodepletion, although the difference from the levels in control tumor tissues did not reach statistical significance (Figure 3a). Both CXCL9 and CXCL10, that have antiangiogenic properties and the chemoattractant effect of CXCR3 þ lymphocytes, were significantly upregulated by CD4 þ T lymphodepletion. The number of CXCR3 þ CD8 þ T cells was significantly increased by CD4 þ T lymphodepletion (Figure 3b). In vivo depletion assays showed that the antitumor effects induced by CD4 þ T lymphodepletion were mainly mediated by CD8 þ T cells (Supplementary Figure S2 online). www.jidonline.org 1885
S Fujiwara et al.
Cell numbers per 1×105 cells in tumors
Therapy Using OX40 Agonist and CD4 Depletion
600
600
500
500
*
* 400
400
300
300
200
200
100
100
0 Control Anti–CD4 mAb CD8+ T cells
0
Control
Anti–CD4 mAb NK cells
Figure 2. Intratumoral CD4 þ T lymphodepletion enhances the accumulation of CD8 þ T cells and natural killer (NK) cells within the tumor tissue. B16F10 cells (105) were inoculated into the right flank of C57BL/6 mice (n ¼ 5 per group) on day 0, followed by intratumoral administration of 500 mg of an antiCD4 mAb or control rat IgG on day 7. On day 14, the tumors were resected and the numbers of CD8 þ T cells (CD8 þ CD3 þ ) and NK cells (NK1.1 þ CD3 ) within tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (*Po0.05). Data shown are from one of three individual experiments.
CD11b þ Gr-1 þ MDSCs within B16F10 tumor tissues also increase following CD4 þ T lymphodepletion
To further clarify the immune alteration induced as a result of CD4 þ T lymphodepletion, we examined CD11b þ Gr-1 þ MDSCs, which inhibit antitumor immunity (Peranzoni et al., 2010; Gabrilovich et al., 2012). We originally expected that the number of MDSCs within the tumor tissue would decrease, because marked antitumor effects were achieved mainly via CD8 þ T cells after CD4 þ T lymphodepletion (Figure 1d and Supplementary Figure S2 online). Unexpectedly, CD11b þ Gr-1 þ MDSCs (particularly CD11b þ Gr-1int MDSCs but not CD11b þ Gr-1high MDSCs) significantly increased in CD4 þ T-lymphodepleted tumor tissues (Figure 4a). Recent studies revealed that CCL2-CCR2 interaction has an important role in the recruitment of MDSCs to tumor sites (Zhu et al., 2011; Lesokhin et al., 2012). Indeed, after CD4 þ T lymphodepletion, CCL2 markedly increased in the tumor tissue (Figure 4b). In addition, the proportion of CCR2 þ MDSCs that infiltrated as a result of CD4 þ T lymphodepletion was higher than those for control tumors. The CD11b þ Gr-1int MDSCs that increased following CD4 þ T lymphodepletion were mostly (490%) positive for CCR2 (Figure 4c). These results suggest that CCL2-CCR2 interaction may be involved in the increased MDSC recruitment into tumor tissue following CD4 þ T lymphodepletion. The increase in the number of MDSCs is caused by the infiltration of CD8 þ T cells and not NK cells
We hypothesized that the recruitment of MDSCs into tumor tissue induced by CD4 þ T lymphodepletion may result from the accumulation of effector cells. To verify the hypothesis, we treated mice bearing B16F10 tumors with anti-CD8 mAb and anti-asialoGM1 sera to deplete both CD8 þ T cells and NK 1886 Journal of Investigative Dermatology (2014), Volume 134
cells, and then intratumorally administered an anti-CD4 mAb. As shown in Figure 5a, MDSCs remarkably decreased following depletion of both CD8 þ T cells and NK cells. CCL2 was also significantly decreased by depletion of these effector cells. Both arginase-1 and inducible nitric oxide synthase increased following CD4 þ T lymphodepletion but decreased following depletion of both CD8 þ T cells and NK cells (Figure 5b). Next, we attempted to clarify which effector cells have a crucial role in the recruitment of MDSCs. Depletion of CD8 þ T cells did not affect the recruitment of CD11b þ Gr-1high MDSCs but reduced recruitment of CD11b þ Gr-1int MDSCs (Figure 5c). Interestingly, depletion of NK cells did not affect the recruitment of either MDSC population. These results suggest that the recruitment of MDSCs by CD4 þ T lymphodepletion may result from the accumulation of CD8 þ T cells within the tumor tissue. The antitumor effects of the OX40 agonist on CD4 þ T-lymphodepleted tumors are not related to the reduction in the number of MDSCs and are mediated by enhancing the cytotoxic activities of the effector cells
Recent studies revealed that an OX40 agonist could enhance the infiltration of therapeutic effector cells, including CD8 þ T cells, and decrease MDSCs (Gough et al., 2008; Pardee et al., 2010). Therefore, we determined whether the OX40 agonist affected the MDSC populations, CD8 þ T cells, and NK cells in this model. As shown in Figure 6a, the OX40 agonist did not decrease the MDSC populations. The OX40 agonist also did not affect the proportion of CD8 þ T cells that infiltrated into the tumor tissues. The NK cell count showed a tendency to decrease following OX40 agonist treatment; however, this decrease was not statistically significant. Nevertheless, treatment with the OX40 agonist markedly enhanced the expression of IL-2 and IFN-g in the tumor tissues (Figure 6b). In addition, treatment markedly enhanced the expression of granzyme B and perforin. These results indicate that the OX40 agonist neither decreased the recruitment of MDSCs nor increased that of effector cells, but augmented the cytotoxic activities of the effector cells that accumulated in the tumors as a result of CD4 þ T lymphodepletion, thereby further improving the antitumor efficacy of this model. DISCUSSION Recent accumulating evidence has shown that tumor immune escape is related to increases in immunosuppressive populations such as Tregs and MDSCs within tumor tissues (Lindau et al., 2013). In this study, we showed that CD4 þ T lymphodepletion dramatically altered the tumor cytokine milieu and significantly augmented CD8 þ T-cell and NK cell infiltration in established, poorly immunogenic B16F10 melanoma tissues, which resulted in inhibition of tumor growth. These dramatic changes could be partly attributed to the concomitant depletion of Tregs within tumor tissues following CD4 þ T lymphodepletion (Supplementary Figure S1 online). However, CD4 þ T lymphodepletion also depletes the T helper cells that have a crucial role in the priming, expansion, and survival of CD8 þ cytotoxic T lymphocytes.
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
a
TNF-α
IFN-γ 300
**
250
12
IL-2
IL-12
600
**
**
10
**
180 160
500 140
200
8
400
150
6
300
120 100 80
100
4
200
50
2
100
60
Relative mRNA value
40 20 0
Control Anti–CD4 mAb
0
Control
IL-23 1.2
**
0
Anti–CD4 mAb IL-10
Anti–CD4 mAb
CXCL9
Control
Anti–CD4 mAb
CXC10
**
70
90
7
80
60
1
0 Control
**
6
70 50
0.8
5
60
40
50
4
30
40
3
0.6 0.4
30
20
2
20 0.2
10
0
0
0 Control
Anti–CD4 mAb
Control
b Cell numbers per 1×105 cells in tumors
1
10
20
0
Anti-CD4 mAb
Control
**
Anti-CD4 mAb
Control
Anti-CD4 mAb
20
18
18
16
16
14
14
12
12
10
10
8
8
6
6
4
4
2
2 0
0 Control Anti–CD4 mAb CD8+ CXCR3+ cells
Control
Anti–CD4 mAb
NK1.1+ CXCR3+ cells
Figure 3. Intratumoral CD4 þ T lymphodepletion dramatically alters the tumor immune microenvironment. Tumor tissues (day 14) were resected as mentioned in the Figure 2 legend. (a) Total RNA was then extracted from the tissues. The mRNA expression levels of IFN-g, TNF-a, IL-2, IL-12p35, IL-23p19, IL-10, CXCL9, and CXCL10 were determined by real-time RT-PCR and the expression levels relative to those of GAPDH mRNA (mean±SEM) are shown (**Po0.01). Data shown are from one of three individual experiments. (b) The numbers of CXCR3-positive CD8 þ T cells and natural killer (NK) cells within the tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (**Po0.01). Data shown are from one of two individual experiments.
Therefore, the initial infiltration of CD8 þ T cells into tumor tissues following CD4 þ T lymphodepletion might not require T helper cells. In our model, T helper cells might contribute to the cytotoxic T lymphocyte-mediated antitumor effects in the late period, because the number of CD4 þ T cells in both tumors and tumor-draining lymph nodes gradually increased from B2 weeks after anti-CD4 mAb injection, concomitant with decrease in the effect of the anti-CD4 mAb (data not shown). Unexpectedly, the number of MDSCs in tumor tissues
also increased significantly following CD4 þ T lymphodepletion. MDSCs are a heterogeneous population of cells that consists of polymorphonuclear, mononuclear, and immature myeloid cells. In mice, MDSCs express the myeloid antigens CD11b and Gr-1 and are classified as CD11b þ Gr-1high cells (granulocytic MDSCs) or CD11b þ Gr-1int cells (monocytic MDSCs). Granulocytic MDSCs suppress antigen-specific CD8 þ T cells, predominantly through the production of reactive oxygen species. In contrast, monocytic MDSCs www.jidonline.org 1887
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
Control
Anti-CD4 mAb
Gr-1high
Gr-1
Gr-1int
CD11b
*
Cell numbers per 1×105 cells in tumors
500
500
450
450
400
400
350
350
300
300
250
250
200
200
150
150
100
100
50
50 0
0 Control
Control
Anti–CD4 mAb
CD11b+Gr-1intMDSC 100
**
30
***
CCR2-positive cells %
CCL2 relative mRNA value
90 25 20 15 10 5 0 Anti–CD4 mAb
100 90
80
80
70
70
60
60
50
50
40
40
30
30
20
20
10
10
0 Control
Anti–CD4 mAb
CD11b+Gr-1highMDSC
0 Control
Anti–CD4 mAb
CD11b+Gr-1highMDSC
Control
Anti–CD4 mAb
CD11b+Gr-1intMDSC
Figure 4. The number of CCR2-positive myeloid-derived suppressor cells (MDSCs) increased in B16F10 tumor tissues after CD4 þ T lymphodepletion. Tumor tissues (day 14) were resected as mentioned in the Figure 2 legend. (a) The numbers of CD11b þ Gr-1 þ MDSCs within tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (*Po0.05). Data shown are from one of four individual experiments. (b) CCL2 mRNA expression was determined by real-time RT-PCR and has been shown relative to GAPDH mRNA expression (mean±SEM values) (**Po0.01). Data shown are from one of two individual experiments. (c) The proportion of CCR2-positive MDSCs to total MDSCs within tumor tissues was analyzed by flow cytometry. The bars represent mean þ SEM (***Po0.001). Data shown are from one of two individual experiments.
suppress CD8 þ T cells, predominantly via inducible nitric oxide synthase and arginase-1 enzymes and through the production of reactive nitrogen species. Monocytic MDSCs have been shown to be more suppressive than granulocytic MDSCs in antigen-stimulated CD8 þ T cells (Dolcetti et al., 2010; Youn and Gabrilovich, 2010; Gabrilovich et al., 2012; Lu and Gabrilovich, 2012). In our model, a larger number of 1888 Journal of Investigative Dermatology (2014), Volume 134
monocytic MDSCs infiltrated into tumors than granulocytic MDSCs. The infiltration of monocytic MDSCs was almost completely abrogated by in vivo depletion of CD8 þ T cells but not by depletion of NK cells (Figure 5c). Therefore, monocytic MDSCs appear to increase in tumor tissues as a specific response to the infiltration of CD8 þ T cells. This increase in the number of MDSC and the increase in IL-10
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
b 800 700 600
CCL2 relative mRNA value
CD11b+Gr–1high MDSC CD11b+Gr–1int MDSC
*
500 400 300 200 100 +
+
CD8 depletion
–
+
NK depletion
–
+
c
25 20 15 10 5
7 6 5 4 3 2 1 0
800 700
12
CD11b+Gr–1high MDSC CD11b+Gr–1int MDSC
iNOS relative mRNA value
Cell number per 1×105 cells in tumors
30
0 8
0 Anti-CD4 mAb
Arginase-1 relative mRNA value
Cell number per 1×105 cells in tumors
a
*
600 500 400 300 200 100 0
10 8 6 4 2
Anti-CD4 mAb
+
+
+
0 Anti-CD4 mAb
–
+
+
CD8 depletion
–
+
–
CD8 depletion
–
–
+
NK depletion
–
–
+
NK depletion
–
–
+
Figure 5. The increase in the number of myeloid-derived suppressor cells (MDSCs) depends on the infiltration of CD8 þ T cells. C57BL/6 mice (n ¼ 5 per group) were subcutaneously inoculated with 105 B16F10 cells on day 0. Some mice were intraperitoneally administered with anti-CD8 mAb and/or anti-asialo GM1 on days 6 and 11. On day 7, anti-CD4 mAb or control rat IgG were intratumorally administered (a and c). On day 14, the numbers of CD11b þ Gr-1 þ MDSCs within tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (*Po0.05). (b) CCL2, arginase-1, and inducible nitric oxide synthase (iNOS) mRNA expression was determined by real-time RT-PCR, and their expression is shown relative to that of GAPDH mRNA (mean þ SEM values). Data shown are from one of two individual experiments.
expression (Figure 3a) are homeostatic responses for counteracting the antitumor immune reactions dominated by CD8 þ T cells. However, these MDSCs could not fully suppress the antitumor immune reactions induced by intratumoral CD4 þ T lymphodepletion because the reactions were so strong. Recent studies strongly suggest that CCL2-CCR2 interaction has an important role in the recruitment of MDSCs to tumor sites (Zhu et al., 2011; Gehad et al., 2012; Lesokhin et al., 2012). In contrast, Fridlender et al. (2011) showed that CCL2 blockade inhibited tumor growth but did not affect the recruitment of MDSCs. In fact, CCL2 blockade altered the macrophage phenotype, thereby activating CD8 þ T cells (Fridlender et al., 2011). Recently, Zhao et al. (2012) showed that TNF/TNF-R2 signaling has a crucial role in MDSC accumulation within tumor tissues and that the major source of TNF for MDSC accumulation could be the MDSCs themselves. Furthermore, Sevko et al. (2012) showed that lowdose cyclophosphamide therapy augmented the accumulation of MDSCs in melanoma lesions concomitantly with increased production of chronic inflammatory mediators, including TNF,
in the ret transgenic murine melanoma model. Indeed, as shown in Figure 3a, TNF expression significantly increased in tumor tissues following CD4 þ T lymphodepletion. In addition, the expression of TNF was significantly reduced by depletion of CD8 þ T cells but not by depletion of NK cells (data not shown). Although the initial trigger of TNF induction in MDSCs has not yet been determined, infiltration of CD8 þ T cells into tumor tissues might be one of the key triggers. In contrast to our results, Kjaergaard et al. (2000) showed that the antitumor effects of an OX40 agonist on a GL261 mouse glioma were abrogated in mice depleted of either CD4 þ or CD8 þ T cells. The reason for this discrepancy is not known; however, we suppose that it might be caused by the difference in the immunogenicity of tumor cells. The GL261 tumor was originally induced by intracranial injection of 3methylcholanthrene into C57BL/6 mice and is considered to be moderately immunogenic (Szatma´ri et al., 2006). In highly or moderately immunogenic tumors, CD8 þ T cells might easily accumulate and exert antitumor effects in cooperation with CD4 þ T cells. In fact, Giovarelli et al. (2000) showed www.jidonline.org 1889
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
Cell numbers per 1×105 tumor cells
a
800
*
700 600 500
CD11b+Gr–1high MDSC CD11b+Gr–1int MDSC CD8+T cells NK cells
400 300 200 100 0 Control
b
Anti-CD4 mAb + OX40 agonist
IL-2
IFN-γ
**
600
Relative mRNA value
Anti-CD4 mAb
*
500
500
400
400
300
300
200
200
100
100
0
*
600
0 Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist
Control
Control
Granzyme B
***
140
*
120
Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist Perforin
60
*
** 50
100
40
80 30 60 20
40
10
20
0
0 Control
Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist
Control
Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist
Figure 6. The OX40 agonist exerts antitumor effects on CD4 þ T-lymphodepleted tumors by further enhancing the cytotoxic activities of effector cells. C57BL/6 mice (n ¼ 5 per group) were subcutaneously inoculated with B16F10 cells on day 0 and intratumorally administered anti-CD4 mAb or control IgG on day 7. Some mice were administered OX86 on days 7 and 12. On day 14, tumors were resected. (a) The numbers of myeloid-derived suppressor cells (MDSCs), CD8 þ T cells, and natural killer (NK) cells within tumor tissues were analyzed. Data shown are from one of two individual experiments. (b) IL-2, IFN-g, granzyme B, and perforin mRNA levels were determined and are shown relative to GAPDH mRNA levels. (*Po0.05; **Po0.05; ***Po0.001). Data shown are from one of three individual experiments. The bars represent mean þ SEM.
that large numbers of both CD4 þ T cells and CD8 þ T cells infiltrated an immunogenic variant of a TSA (a mouse mammary adenocarcinoma) tumor, whereas both cells sparsely infiltrated a parental poorly immunogenic TSA tumor. Therefore, we hypothesized that, in highly or moderately immunogenic tumors, treatment with an OX40 agonist can activate both CD4 þ T cells and CD8 þ T cells that accumulate within the tumor, thereby exerting their antitumor effects. In 1890 Journal of Investigative Dermatology (2014), Volume 134
fact, the OX40 agonist has been shown to directly costimulate both CD4 þ T cells and CD8 þ T cells in vitro (Taraban et al., 2002). Activated CD4 þ T helper cells help control CD8 þ T-cell-dependent tumor eradication in vivo. Therefore, antitumor effects on highly or moderately immunogenic tumors would be abolished by depletion of either CD4 þ or CD8 þ T cells. To verify this hypothesis, we examined the antitumor effects of an OX40 agonist using a Cloudman S91 (clone M-3) murine melanoma (of DBA/2 origin), which is generally considered to be more immunogenic than B16F10 melanomas. The Cloudman S91 melanoma had a slight tendency to respond to OX40 agonist treatment alone in the late stage; however, the difference from the control was not statistically significant. Intratumoral CD4 þ T lymphodepletion caused more marked shrinkage of the S91 melanoma than of the B16F10 melanoma. CD4 þ T lymphodepletion plus OX40 agonist treatment led the greatest reduction in tumor size; however, the difference from the size of CD4 þ T-lymphodepleted tumors was not statistically significant, and lymphodepletion plus OX40 agonist treatment led to rejection of the tumors in some mice on day 27 (when mice in the other groups showed no tumor rejection) (Supplementary Figure S3 online). Therefore, our hypothesis could not be verified; however, it might result from the still poor immunogenicity of Cloudman S91 melanoma. However, in another respect, these results seem significant because CD4 þ T lymphodepletion plus OX40 agonist treatment was found to be effective for melanomas with different immunogenicities and from different mouse strains (Figure 1d and Supplementary Figure S3 online). Although previous studies have shown that the therapeutic efficacy of OX40 agonists depends on tumor immunogenicity, this efficacy might be closely correlated with the degree of CD8 þ T-cell infiltration within tumor tissues. In fact, in B16F10 tumors, the infiltration of CD8 þ T cells has been found to be sparse (Nagai et al., 2000a). Because the antitumor effects of IL-12 gene-transfected B16F10 cells were mediated by both CD8 þ T cells and NK cells (Nagai et al., 2000b), whereas those of IL-27 gene-transfected B16F10 cells were mediated by NK cells but not CD8 þ T cells (Oniki et al., 2006), the OX40 agonist might have induced antitumor effects on IL-12 gene-transfected B16F10 cells but not on IL-27 gene-transfected B16F10 cells (Figure 1b and c). As CD4 þ T lymphodepletion permits the accumulation of CD8 þ T cells in B16F10 tumors (Figure 2), the OX40 agonist could exert antitumor effects even if tumor cells are poorly immunogenic, by further activating the tumor-infiltrating CD8 þ T cells. In fact, it has been shown that OX40 can directly costimulate anti-CD3-activated CD8 þ T cells and antigen-specific CD8 þ T cells in vitro (Taraban et al., 2002). Recently, several studies have focused on the effect of OX40 on the immunosuppressive population, that is, Tregs and MDSCs. OX40 engagement has been shown to inhibit Tregs in tumor tissues and exert antitumor activities (Piconese et al., 2008; Hirschhorn-Cymerman et al., 2009). In contrast, it was found that OX40 stimulation promotes Tregs in an appropriate cytokine milieu (Ruby et al., 2009; Xiao et al., 2012). The latter finding means that OX40 stimulation can inhibit antitumor immune responses, depending on the tumor
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
cytokine milieu, via activation of Tregs. However, in our model, OX40-stimulated Treg-mediated inhibition of antitumor immune responses could be almost completely avoidable in the initiation of OX40 agonist treatment, because intratumoral administration with anti-CD4 mAb efficiently depleted Tregs within tumors, along with helper T cells (Supplementary Figure S1 online). On the other hand, concerning MDSCs, OX40 agonist therapy has been found to significantly enhance CD8 þ T-cell infiltration and decrease MDSC population in the MCA205 H12 tumor site. The reduction in the number of MDSCs was caused by the indirect effects of the OX40 agonist because OX40 expression in MDSCs was not detected at the tumor site (Gough et al., 2008). In our study, systemic treatment with the OX40 agonist did not reduce the number of MDSCs that were recruited into the tumor site following CD4 þ T lymphodepletion (Figure 6a). In contrast, OX40 agonist treatment further enhanced the expression of arginase-1 and inducible nitric oxide synthase in CD4 þ T-lymphodepleted tumors, which may be indicative of a more functionally active condition of MDSCs (Supplementary Figure S4 online). We speculate that these functional changes might occur in response to the enhancement of cytotoxic activities of effector cells by the OX40 agonist. The activation of MDSCs as a protective response might increase with increase in the antitumor activities. Although OX40 agonists seem to be promising candidates for immunotherapy for human tumors (Weinberg et al., 2011), the effects of OX40 agonist monotherapy for patients with malignant tumors will be limited because most human tumors are poorly immunogenic. On the other hand, several studies have shown that autologous adoptive tumor-infiltrating lymphocyte transfer after lymphodepletion can cause the regression of refractory metastatic melanomas (Rosenberg and Dudley, 2004; Dudley et al., 2005). Thus, our results support the hypothesis that a combination of OX40 agonist therapy and CD4 þ T lymphodepletion might exert efficient antitumor effects even on poorly immunogenic human melanomas and suggest that lymphodepletion could facilitate the accumulation of tumor-infiltrating lymphocytes into tumors, thereby inducing antitumor effects; however, the number of MDSCs might simultaneously increase within tumors and be activated as a countermeasure against accumulation and cytotoxicity mediated by tumor-infiltrating lymphocytes. In conclusion, this combination therapy could be a promising therapeutic strategy for patients with poorly immunogenic melanomas and might be more effective if further combined with therapeutic strategies that can selectively deplete or inhibit MDSCs (Nagaraj et al., 2010; Meyer et al., 2011; Kodumudi et al., 2012; Lesokhin et al., 2012; Srivastava et al., 2012). MATERIALS AND METHODS Tumor cell lines and animal studies B16F10 melanoma cells (provided by the RIKEN BRC through the National Bio-resource Project of MEXT, Japan), IL-12 cDNA-transfected B16F10 cells, and IL-27 cDNA-transfected B16F10 cells (Oniki et al., 2006) were maintained in Eagle’s modified essential medium supplemented with 5% fetal bovine serum. Cloudman S91 (clone
M-3) melanoma cells were obtained from American Type Culture Collection (ATCC, Rockville, MD). C57BL/6N mice and DBA/2 mice (6- to 8-week-old females) were purchased from Charles River Laboratories (Tokyo, Japan). All animal experiments were conducted according to the Guidelines for Animal Experimentation at the Kobe University Graduate School of Medicine.
In vivo antibody treatments We administered 500 mg of rat GK1.5 mAb (ATCC) into the tumors on day 7 after tumor inoculation to deplete CD4 þ T cells. CD8 þ T cells in vivo were depleted using the rat mAb 2.43 (ATCC). NK cells were depleted using a polyclonal anti-asialo GM1 antibody (Wako Fine Chemicals, Osaka, Japan). OX40 agonist immunotherapy was performed with a rat anti-OX40 agonistic mAb (OX86; ATCC). One milligram of OX86 was administered intraperitoneally on day 0 or day 7 and every 5 days thereafter. Rat IgG (Wako Fine Chemicals) was used as the control antibody.
Flow cytometry Resected tumors were digested with Liberase (0.2 Wunsch units/ml; Roche Diagnostics K.K., Basel, Switzerland) and DNase I (0.02 mg/ ml; Roche) at 37 1C for 60 min. These cells were stained with different combinations of the following antibodies: CD3-FITC, CD8-APC, NK1.1-PE, CXCR3-PE-Cy7, CD11b-PE-Cy7, CCR2-APC (BioLegend, San Diego, CA), CD4-FITC Gr-1-PE, Foxp3-PE (Miltenyi Biotec, Bergisch Gladbach, Germany). Cytometric acquisition was performed with a MoFlo XDP system (Beckman Coulter, Brea, CA) or BD FACSVerse (BD Biosciences) and the Summit 5.2 (Beckman Coulter) or Flowjo (Tree Star, Ashland, OR). Supplementary Figure S5 online shows gating procedures of tumor samples.
Real-time quantitative reverse transcriptase–PCR (qRT-PCR) analysis Total RNA was extracted by using the QuickGene RNA tissue kit (Fujifilm Corporation, Tokyo, Japan). cDNA synthesis was performed using the PrimeScript RT reagent kit (Takara Bio, Shiga, Japan). Realtime qRT-PCR was performed using an Applied Biosystems 7500 realtime PCR system (Applied Biosystems, Foster City, CA) and SYBR Premix Ex Taq II (Takara Bio). mRNA expression values were normalized according to the internal control glyceraldehyde 3phosphate dehydrogenase (GAPDH), and the relative expression values were plotted.
Statistical analysis The statistical significance of differences between the means of two groups was determined using Student’s t-test. The statistical significance of three or more groups was determined using Dunnett’s test or Tukey’s HSD post hoc test. The differences were considered statistically significant at Po0.05. Data are shown as mean±SEM values.
CONFLICT OF INTEREST The authors state no conflict of interest.
ACKNOWLEDGMENTS This study was supported in part by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan, Japanese Society for Pigment Cell Research, and Japanese Dermatological Association basic medical research grant (Shiseido donation).
www.jidonline.org 1891
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid
REFERENCES Croft M (2010) Control of immunity by the TNFR-related molecule OX40 (CD134). Annu Rev Immunol 28:57–78 Dolcetti L, Peranzoni E, Ugel S et al. (2010) Hierarchy of immunosuppressive strength among myeloid-derived suppressor cell subsets is determined by GM-CSF. Eur J Immunol 40:22–35 Dudley ME, Wunderlich JR, Yang JC et al. (2005) Adoptive cell transfer therapy following non-myeloablative but lymphodepleting chemotherapy for the treatment of patients with refractory metastatic melanoma. J Clin Oncol 23:2346–57 Fidler IJ (1973) Selection of successive tumour lines for metastasis. Nat New Biol 242:148–9 Fridlender ZG, Kapoor V, Buchlis G et al. (2011) Monocyte chemoattractant protein-1 blockade inhibits lung cancer tumor growth by altering macrophage phenotype and activating CD8 þ cells. Am J Respir Cell Mol Biol 44:230–7 Gabrilovich DI, Ostrand-Rosenberg S, Bronte V (2012) Coordinated regulation of myeloid cells by tumours. Nat Rev Immunol 12:253–68 Gehad AE, Lichtman MK, Schmults CD et al. (2012) Nitric oxide-producing myeloid-derived suppressor cells inhibit vascular E-selectin expression in human squamous cell carcinomas. J Invest Dermatol 132:2642–51 Giovarelli M, Cappello P, Forni G et al. (2000) Tumor rejection and immune memory elicited by locally released LEC chemokine are associated with an impressive recruitment of APCs, lymphocytes, and granulocytes. J Immunol 164:3200–6 Gough MJ, Ruby CE, Redmond WL et al. (2008) OX40 agonist therapy enhances CD8 infiltration and decreases immune suppression in the tumor. Cancer Res 68:5206–15 Hirschhorn-Cymerman D, Rizzuto GA, Merghoub T et al. (2009) OX40 engagement and chemotherapy combination provides potent antitumor immunity with concomitant regulatory T cell apoptosis. J Exp Med 206:1103–16 Ishii N, Takahashi T, Soroosh P et al. (2010) OX40-OX40 ligand interaction in T-cell-mediated immunity and immunopathology. Adv Immunol 105:63–98 Jensen SM, Maston LD, Gough MJ et al. (2010) Signaling through OX40 enhances antitumor immunity. Semin Oncol 37:524–32 Kjaergaard J, Tanaka J, Kim JA et al. (2000) Therapeutic efficacy of OX-40 receptor antibody depends on tumor immunogenicity and anatomic site of tumor growth. Cancer Res 60:5514–21 Kodumudi KN, Weber A, Sarnaik AA et al. (2012) Blockade of myeloid-derived suppressor cells after induction of lymphopenia improves adoptive T cell therapy in a murine model of melanoma. J Immunol 189:5147–54 Kryczek I, Wei S, Zou L et al. (2007) Cutting edge: Th17 and regulatory T cell dynamics and the regulation by IL-2 in the tumor microenvironment. J Immunol 178:6730–3 Langowski JL, Zhang X, Wu L et al. (2006) IL-23 promotes tumour incidence and growth. Nature 442:461–5 Lesokhin AM, Hohl TM, Kitano S et al. (2012) Monocytic CCR2( þ ) myeloidderived suppressor cells promote immune escape by limiting activated CD8 T-cell infiltration into the tumor microenvironment. Cancer Res 72:876–86 Lindau D, Gielen P, Kroesen M et al. (2013) The immunosuppressive tumour network: myeloid-derived suppressor cells, regulatory T cells and natural killer T cells. Immunology 138:105–15 Lu T, Gabrilovich DI (2012) Molecular pathways: tumor-infiltrating myeloid cells and reactive oxygen species in regulation of tumor microenvironment. Clin Cancer Res 18:4877–82 Lo´pez M, Aguilera R, Pe´rez C et al. (2006) The role of regulatory T lymphocytes in the induced immune response mediated by biological vaccines. Immunobiology 211:127–36
1892 Journal of Investigative Dermatology (2014), Volume 134
Meyer C, Sevko A, Ramacher M et al. (2011) Chronic inflammation promotes myeloid-derived suppressor cell activation blocking antitumor immunity in transgenic mouse melanoma model. Proc Natl Acad Sci USA 108:17111–6 Nagai H, Hara I, Horikawa T et al. (2000a) Antitumor effects on mouse melanoma elicited by local secretion of interleukin-12 and their enhancement by treatment with interleukin-18. Cancer Invest 18:206–13 Nagai H, Hara I, Horikawa T et al. (2000b) Elimination of CD4( þ ) T cells enhances anti-tumor effect of locally secreted interleukin-12 on B16 mouse melanoma and induces vitiligo-like coat color alteration. J Invest Dermatol 115:1059–64 Nagaraj S, Youn JI, Weber H et al. (2010) Anti-inflammatory triterpenoid blocks immune suppressive function of MDSCs and improves immune response in cancer. Clin Cancer Res 16:1812–23 Oniki S, Nagai H, Horikawa T et al. (2006) Interleukin-23 and interleukin-27 exert quite different antitumor and vaccine effects on poorly immunogenic melanoma. Cancer Res 66:6395–404 Pardee AD, McCurry D, Alber S et al. (2010) A therapeutic OX40 agonist dynamically alters dendritic, endothelial, and T cell subsets within the established tumor microenvironment. Cancer Res 70:9041–52 Peranzoni E, Zilio S, Marigo I et al. (2010) Myeloid-derived suppressor cell heterogeneity and subset definition. Curr Opin Immunol 22:238–44 Piconese S, Valzasina B, Colombo MP (2008) OX40 triggering blocks suppression by regulatory T cells and facilitates tumor rejection. J Exp Med 205:825–39 Rosenberg SA, Dudley ME (2004) Cancer regression in patients with metastatic melanoma after the transfer of autologous antitumor lymphocytes. Proc Natl Acad Sci USA 101(Suppl 2):14639–45 Ruby CE, Yates MA, Hirschhorn-Cymerman D et al. (2009) Cutting Edge: OX40 agonists can drive regulatory T cell expansion if the cytokine milieu is right. J Immunol 183:4853–7 Seo N, Hayakawa S, Takigawa M et al. (2001) Interleukin-10 expressed at early tumour sites induces subsequent generation of CD4( þ ) T-regulatory cells and systemic collapse of antitumour immunity. Immunology 103:449–57 Sevko A, Sade-Feldman M, Kanterman J et al. (2012) Cyclophosphamide promotes chronic inflammation-dependent immunosuppression and prevents antitumor response in melanoma. J Invest Dermatol 133:1610–9 Srivastava MK, Zhu L, Harris-White M et al. (2012) Myeloid suppressor cell depletion augments antitumor activity in lung cancer. PLoS ONE 7:e40677 Sugamura K, Ishii N, Weinberg AD (2004) Therapeutic targeting of the effector T-cell co-stimulatory molecule OX40. Nat Rev Immunol 4:420–31 Szatma´ri T, Lumniczky K, De´saknai S et al. (2006) Detailed characterization of the mouse glioma 261 tumor model for experimental glioblastoma therapy. Cancer Sci 97:546–53 Taraban VY, Rowley TF, O’Brien L et al. (2002) Expression and costimulatory effects of the TNF receptor superfamily members CD134 (OX40) and CD137 (4-1BB), and their role in the generation of anti-tumor immune responses. Eur J Immunol 32:3617–27 Weinberg AD, Morris NP, Kovacsovics-Bankowski M et al. (2011) Science gone translational: the OX40 agonist story. Immunol Rev 244: 218–31 Weinberg AD, Rivera MM, Prell R et al. (2000) Engagement of the OX-40 receptor in vivo enhances antitumor immunity. J Immunol 164:2160–9 Xiao X, Gong W, Demirci G et al. (2012) New insights on OX40 in the control of T cell immunity and immune tolerance in vivo. J Immunol 188: 892–901 Youn JI, Gabrilovich DI (2010) The biology of myeloid-derived suppressor cells: the blessing and the curse of morphological and functional heterogeneity. Eur J Immunol 40:2969–75 Zhao X, Rong L, Li X et al. (2012) TNF signaling drives myeloid-derived suppressor cell accumulation. J Clin Invest 122:4094–104 Zhu X, Fujita M, Snyder LA et al. (2011) Systemic delivery of neutralizing antibody targeting CCL2 for glioma therapy. J Neurooncol 104:83–92
See related commentary on pg 1794
Intratumoral CD4 þ T Lymphodepletion Sensitizes Poorly Immunogenic Melanomas to Immunotherapy with an OX40 Agonist Susumu Fujiwara1, Hiroshi Nagai1, Noriko Shimoura1, Shuntaro Oniki1, Takayuki Yoshimoto2 and Chikako Nishigori1 Previous studies have shown that the antitumor effects of OX40 agonists depend on the immunogenicity of the tumor and that poorly immunogenic tumors such as B16F10 melanomas do not respond to OX40 agonist treatment. In this study, we have shown that intratumoral CD4 þ T lymphodepletion sensitized poorly immunogenic B16F10 melanomas to immunotherapy with an OX40 agonist. CD4 þ T lymphodepletion dramatically altered the tumor immune microenvironment, making it more susceptible to the antitumor effects of an OX40 agonist by enhancing the accumulation of CD8 þ T cells and natural killer (NK) cells in tumor tissue. However, unexpectedly, the number of CD11b þ Gr-1 þ myeloid-derived suppressor cells (MDSCs) within tumor tissues also significantly increased as a result of CD4 þ T lymphodepletion. As a countermeasure against CD8 þ T-cell accumulation, CCR2positive CD11b þ Gr-1int (monocytic) MDSCs predominantly increased. Treatment with an OX40 agonist under CD4 þ T lymphodepletion neither reduced MDSCs nor increased CD8 þ T cells and NK cells, but further enhanced the expression of cytotoxic molecules from tumor-infiltrating effector cells. Our results suggest that combined immunotherapy using both an OX40 agonist and CD4 þ T lymphodepletion could be a promising therapeutic strategy for poorly immunogenic tumors and might be more effective if further combined with a therapeutic strategy targeting MDSCs. Journal of Investigative Dermatology (2014) 134, 1884–1892; doi:10.1038/jid.2014.42; published online 27 February 2014
INTRODUCTION OX40 (CD134) is an B50-kDa glycoprotein with three complete and one truncated cysteine-rich domains that are characteristic of the TNF receptor superfamily. OX40 is predominantly expressed on activated CD4 þ T cells and CD8 þ T cells; however, other cell types, including natural killer (NK) cells, NK T cells, neutrophils and CD4 þ Foxp3 þ regulatory T cells (Tregs), also express OX40 (Sugamura et al., 2004; Croft, 2010; Ishii et al., 2010; Weinberg et al., 2011). Several studies have shown that OX40 agonists have potent antitumor activities in preclinical models (Weinberg et al., 2000; Piconese et al., 2008; Jensen et al., 2010). However, these antitumor effects were found to be dependent on the immunogenicity of the tumor, because poorly immunogenic tumors did not respond to OX40 agonist treatment (Kjaergaard 1
Division of Dermatology, Department of Internal Related, Kobe University Graduate School of Medicine, Kobe, Japan and 2Department of Immunoregulation, Institute of Medical Science, Tokyo Medical University, Tokyo, Japan Correspondence: Hiroshi Nagai, Division of Dermatology, Department of Internal Related, Kobe University Graduate School of Medicine, 7-5-1 Kusunoki-cho, Chuo-ku, Kobe 650-0017, Japan. E-mail: [email protected] Abbreviations: MDSC, myeloid-derived suppressor cell; NK, natural killer Received 7 March 2013; revised 8 December 2013; accepted 23 December 2013; accepted article preview online 27 January 2014; published online 27 February 2014
1884 Journal of Investigative Dermatology (2014), Volume 134
et al., 2000). A recent study highlighted the significance of combining OX40 agonist therapy with other therapeutic approaches for poorly immunogenic tumors (HirschhornCymerman et al., 2009). B16F10 melanoma cells are poorly immunogenic tumor cells that originally spontaneously developed in C57BL/6 mice and are thought to reflect the poor immunogenicity of metastatic tumors in humans (Fidler, 1973). B16F10 melanoma, along with other poorly immunogenic tumors, has been shown to be unresponsive to immunotherapy with an OX40 agonist (Hirschhorn-Cymerman et al., 2009). Accumulating evidence has shown that Tregs have a crucial role in refractoriness to immunotherapy, and high levels of Tregs have been detected in B16F10 tumor tissues (Kryczek et al., 2007). Furthermore, other CD4 þ regulatory T cells, such as TR1 cells and Th3 cells, have also been shown to be involved in tumor immune escape (Seo et al., 2001; Lo´pez et al., 2006). As an OX40 agonist can activate not only antitumor effector cells but also Tregs (Ruby et al., 2009; Xiao et al., 2012), we hypothesized that depletion of the CD4 þ regulatory T cells in tumor tissue might sensitize B16F10 melanomas to immunotherapy with an OX40 agonist. To test this possibility, we intratumorally administered an anti-CD4 mAb on day 7 after tumor inoculation. This approach efficiently depleted Tregs, as well as helper T cells and other CD4 þ cells in B16F10 melanomas and sensitized these tumor tissues to immunotherapy with an OX40 agonist. Interestingly, & 2014 The Society for Investigative Dermatology
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
RESULTS Intratumoral CD4 þ T lymphodepletion sensitizes poorly immunogenic melanomas to immunotherapy with an OX40 agonist
Initially, we examined the antitumor effects of an OX40 agonist on poorly immunogenic B16F10 melanomas. As shown in Figure 1a, treatment with OX40 agonist (starting on day 0 after tumor inoculation) did not exert antitumor effects. Next, we tested whether the OX40 agonist could exert antitumor effects on B16F10 melanomas when combined with cytokines that are involved in antitumor immunity. Treatment with the OX40 agonist exerted significant antitumor effects on IL-12 cDNA-transfected B16F10 melanomas, but not on IL-27 cDNA-transfected B16F10 melanomas (Figure 1b and c). In contrast, when combined with a single, intratumoral administration of an anti-CD4 mAb (for CD4 þ T lymphodepletion), OX40 agonist treatment had significant antitumor effects when started on day 7 (Figure 1d) or on day 0 (data not shown). Flow cytometric analysis showed that, in control mice, which were intratumorally administered rat IgG, 23.8% of the tumor-infiltrating mononuclear cells were CD4 þ after 7 days of administration, whereas in anti-CD4 mAb-treated mice, the proportion of CD4 þ cells decreased to 0.67% at 7 days after administration of the anti-CD4 mAb. The proportion of Treg cells (CD4 þ Foxp3 þ cells) among the tumor-infiltrating mononuclear cells was 0.11% after antiCD4 mAb administration (vs. 7.4% in the control) (Supplementary Figure S1 online). Other CD4 þ cell populations (CD4 þ F4/80 þ cells, CD4 þ CD11b þ cells, and CD4 þ CD11c þ cells) were barely (o0.3% of tumor-infiltrating mononuclear cells) detected in the control tumor tissues on day 14 and further decreased (to o0.03%) following antiCD4 mAb administration (data not shown). These results indicated that the CD4-positive cells, including Tregs, in tumor tissues were almost completely depleted on day 7 after intratumoral injection of an anti-CD4 mAb. Intratumoral CD4 þ T lymphodepletion dramatically alters the immune microenvironment in the tumor, making it more susceptible to antitumor effects by enhancing the accumulation of CD8 þ T cells and NK cells
To clarify why intratumoral CD4 þ T lymphodepletion sensitizes B16F10 melanomas to immunotherapy with an OX40 agonist, we examined the cellular and humoral alterations in CD4 þ T-lymphodepleted tumor tissues. Flow cytometric analysis showed that, in the control group, CD8 þ T-cell infiltration into B16F10 tumors was scarce and NK cell infiltration was considerable. However, after CD4 þ
a
b
600
600 No treatment OX40 agonist
500
Tumor size (mm2)
administration of just a single dose of anti-CD4 mAb into B16F10 tissues dramatically altered the immune conditions and augmented the infiltration of effector cells into the tumors. However, unexpectedly, CD4 þ T lymphodepletion also augmented the recruitment of myeloid-derived suppressor cells (MDSCs) into the tumor tissues. Finally, we analyzed the mechanism by which combined treatment with an OX40 agonist and CD4 þ T lymphodepletion exerts antitumor effects.
400
400
300
300
200
200
100
100
0
0
10
20
30
0
40
c
d
600
600
500
No treatment OX40 agonist
500
No treatment OX40 agonist
*
0
400
300
300
200
200
100
100
20
30
40
No treatment Anti–CD4 mAb Anti–CD4 mAb + OX40 agonist
500
400
10
*
0 0
10
20
30
40
0
*
0
10
20
30
40
Day
Figure 1. Antitumor effects of an OX40 agonist on B16F10 melanomas. Onehundred thousand tumor cells (a and d, B16F10 cells; b, IL-12 cDNA-transfected B16F10 cells; c, IL-27 cDNA-transfected B16F10 cells) were subcutaneously inoculated into the right flank of C57BL/6 mice (n ¼ 5 per group) on day 0. (a–c) Mice were intraperitoneally (i.p.) administered 1 mg of OX86 on day 0 and every 5 days thereafter. (d) Anti-CD4 mAb (500 mg) was intratumorally injected only once on day 7. OX86 was i.p. administered on day 7 and every 5 days thereafter. The tumor size was measured as the product of the longest diameter and the perpendicular diameter. The bars represent mean þ SEM (*Po0.05). The data shown are from one of three individual experiments.
T lymphodepletion, both CD8 þ T cells and NK cells markedly infiltrated into tumor tissues (Figure 2). The cytokine profile of the tumor was also dramatically altered by CD4 þ T lymphodepletion. Cytokines with antitumor activities, such as IFN-g, TNF-a, IL-2, and IL-12, were significantly upregulated by CD4 þ T lymphodepletion (Figure 3a). However, interestingly, the expression of IL-23, which has been reported to inhibit the accumulation of antitumor CD8 þ T cells (Langowski et al., 2006), significantly reduced. After CD4 þ T lymphodepletion, the expression levels of the immunosuppressive cytokine TGF-b did not change significantly compared with the levels in control tumor tissues (data not shown). However, the expression of IL-10 showed a tendency to increase after CD4 þ T lymphodepletion, although the difference from the levels in control tumor tissues did not reach statistical significance (Figure 3a). Both CXCL9 and CXCL10, that have antiangiogenic properties and the chemoattractant effect of CXCR3 þ lymphocytes, were significantly upregulated by CD4 þ T lymphodepletion. The number of CXCR3 þ CD8 þ T cells was significantly increased by CD4 þ T lymphodepletion (Figure 3b). In vivo depletion assays showed that the antitumor effects induced by CD4 þ T lymphodepletion were mainly mediated by CD8 þ T cells (Supplementary Figure S2 online). www.jidonline.org 1885
S Fujiwara et al.
Cell numbers per 1×105 cells in tumors
Therapy Using OX40 Agonist and CD4 Depletion
600
600
500
500
*
* 400
400
300
300
200
200
100
100
0 Control Anti–CD4 mAb CD8+ T cells
0
Control
Anti–CD4 mAb NK cells
Figure 2. Intratumoral CD4 þ T lymphodepletion enhances the accumulation of CD8 þ T cells and natural killer (NK) cells within the tumor tissue. B16F10 cells (105) were inoculated into the right flank of C57BL/6 mice (n ¼ 5 per group) on day 0, followed by intratumoral administration of 500 mg of an antiCD4 mAb or control rat IgG on day 7. On day 14, the tumors were resected and the numbers of CD8 þ T cells (CD8 þ CD3 þ ) and NK cells (NK1.1 þ CD3 ) within tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (*Po0.05). Data shown are from one of three individual experiments.
CD11b þ Gr-1 þ MDSCs within B16F10 tumor tissues also increase following CD4 þ T lymphodepletion
To further clarify the immune alteration induced as a result of CD4 þ T lymphodepletion, we examined CD11b þ Gr-1 þ MDSCs, which inhibit antitumor immunity (Peranzoni et al., 2010; Gabrilovich et al., 2012). We originally expected that the number of MDSCs within the tumor tissue would decrease, because marked antitumor effects were achieved mainly via CD8 þ T cells after CD4 þ T lymphodepletion (Figure 1d and Supplementary Figure S2 online). Unexpectedly, CD11b þ Gr-1 þ MDSCs (particularly CD11b þ Gr-1int MDSCs but not CD11b þ Gr-1high MDSCs) significantly increased in CD4 þ T-lymphodepleted tumor tissues (Figure 4a). Recent studies revealed that CCL2-CCR2 interaction has an important role in the recruitment of MDSCs to tumor sites (Zhu et al., 2011; Lesokhin et al., 2012). Indeed, after CD4 þ T lymphodepletion, CCL2 markedly increased in the tumor tissue (Figure 4b). In addition, the proportion of CCR2 þ MDSCs that infiltrated as a result of CD4 þ T lymphodepletion was higher than those for control tumors. The CD11b þ Gr-1int MDSCs that increased following CD4 þ T lymphodepletion were mostly (490%) positive for CCR2 (Figure 4c). These results suggest that CCL2-CCR2 interaction may be involved in the increased MDSC recruitment into tumor tissue following CD4 þ T lymphodepletion. The increase in the number of MDSCs is caused by the infiltration of CD8 þ T cells and not NK cells
We hypothesized that the recruitment of MDSCs into tumor tissue induced by CD4 þ T lymphodepletion may result from the accumulation of effector cells. To verify the hypothesis, we treated mice bearing B16F10 tumors with anti-CD8 mAb and anti-asialoGM1 sera to deplete both CD8 þ T cells and NK 1886 Journal of Investigative Dermatology (2014), Volume 134
cells, and then intratumorally administered an anti-CD4 mAb. As shown in Figure 5a, MDSCs remarkably decreased following depletion of both CD8 þ T cells and NK cells. CCL2 was also significantly decreased by depletion of these effector cells. Both arginase-1 and inducible nitric oxide synthase increased following CD4 þ T lymphodepletion but decreased following depletion of both CD8 þ T cells and NK cells (Figure 5b). Next, we attempted to clarify which effector cells have a crucial role in the recruitment of MDSCs. Depletion of CD8 þ T cells did not affect the recruitment of CD11b þ Gr-1high MDSCs but reduced recruitment of CD11b þ Gr-1int MDSCs (Figure 5c). Interestingly, depletion of NK cells did not affect the recruitment of either MDSC population. These results suggest that the recruitment of MDSCs by CD4 þ T lymphodepletion may result from the accumulation of CD8 þ T cells within the tumor tissue. The antitumor effects of the OX40 agonist on CD4 þ T-lymphodepleted tumors are not related to the reduction in the number of MDSCs and are mediated by enhancing the cytotoxic activities of the effector cells
Recent studies revealed that an OX40 agonist could enhance the infiltration of therapeutic effector cells, including CD8 þ T cells, and decrease MDSCs (Gough et al., 2008; Pardee et al., 2010). Therefore, we determined whether the OX40 agonist affected the MDSC populations, CD8 þ T cells, and NK cells in this model. As shown in Figure 6a, the OX40 agonist did not decrease the MDSC populations. The OX40 agonist also did not affect the proportion of CD8 þ T cells that infiltrated into the tumor tissues. The NK cell count showed a tendency to decrease following OX40 agonist treatment; however, this decrease was not statistically significant. Nevertheless, treatment with the OX40 agonist markedly enhanced the expression of IL-2 and IFN-g in the tumor tissues (Figure 6b). In addition, treatment markedly enhanced the expression of granzyme B and perforin. These results indicate that the OX40 agonist neither decreased the recruitment of MDSCs nor increased that of effector cells, but augmented the cytotoxic activities of the effector cells that accumulated in the tumors as a result of CD4 þ T lymphodepletion, thereby further improving the antitumor efficacy of this model. DISCUSSION Recent accumulating evidence has shown that tumor immune escape is related to increases in immunosuppressive populations such as Tregs and MDSCs within tumor tissues (Lindau et al., 2013). In this study, we showed that CD4 þ T lymphodepletion dramatically altered the tumor cytokine milieu and significantly augmented CD8 þ T-cell and NK cell infiltration in established, poorly immunogenic B16F10 melanoma tissues, which resulted in inhibition of tumor growth. These dramatic changes could be partly attributed to the concomitant depletion of Tregs within tumor tissues following CD4 þ T lymphodepletion (Supplementary Figure S1 online). However, CD4 þ T lymphodepletion also depletes the T helper cells that have a crucial role in the priming, expansion, and survival of CD8 þ cytotoxic T lymphocytes.
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
a
TNF-α
IFN-γ 300
**
250
12
IL-2
IL-12
600
**
**
10
**
180 160
500 140
200
8
400
150
6
300
120 100 80
100
4
200
50
2
100
60
Relative mRNA value
40 20 0
Control Anti–CD4 mAb
0
Control
IL-23 1.2
**
0
Anti–CD4 mAb IL-10
Anti–CD4 mAb
CXCL9
Control
Anti–CD4 mAb
CXC10
**
70
90
7
80
60
1
0 Control
**
6
70 50
0.8
5
60
40
50
4
30
40
3
0.6 0.4
30
20
2
20 0.2
10
0
0
0 Control
Anti–CD4 mAb
Control
b Cell numbers per 1×105 cells in tumors
1
10
20
0
Anti-CD4 mAb
Control
**
Anti-CD4 mAb
Control
Anti-CD4 mAb
20
18
18
16
16
14
14
12
12
10
10
8
8
6
6
4
4
2
2 0
0 Control Anti–CD4 mAb CD8+ CXCR3+ cells
Control
Anti–CD4 mAb
NK1.1+ CXCR3+ cells
Figure 3. Intratumoral CD4 þ T lymphodepletion dramatically alters the tumor immune microenvironment. Tumor tissues (day 14) were resected as mentioned in the Figure 2 legend. (a) Total RNA was then extracted from the tissues. The mRNA expression levels of IFN-g, TNF-a, IL-2, IL-12p35, IL-23p19, IL-10, CXCL9, and CXCL10 were determined by real-time RT-PCR and the expression levels relative to those of GAPDH mRNA (mean±SEM) are shown (**Po0.01). Data shown are from one of three individual experiments. (b) The numbers of CXCR3-positive CD8 þ T cells and natural killer (NK) cells within the tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (**Po0.01). Data shown are from one of two individual experiments.
Therefore, the initial infiltration of CD8 þ T cells into tumor tissues following CD4 þ T lymphodepletion might not require T helper cells. In our model, T helper cells might contribute to the cytotoxic T lymphocyte-mediated antitumor effects in the late period, because the number of CD4 þ T cells in both tumors and tumor-draining lymph nodes gradually increased from B2 weeks after anti-CD4 mAb injection, concomitant with decrease in the effect of the anti-CD4 mAb (data not shown). Unexpectedly, the number of MDSCs in tumor tissues
also increased significantly following CD4 þ T lymphodepletion. MDSCs are a heterogeneous population of cells that consists of polymorphonuclear, mononuclear, and immature myeloid cells. In mice, MDSCs express the myeloid antigens CD11b and Gr-1 and are classified as CD11b þ Gr-1high cells (granulocytic MDSCs) or CD11b þ Gr-1int cells (monocytic MDSCs). Granulocytic MDSCs suppress antigen-specific CD8 þ T cells, predominantly through the production of reactive oxygen species. In contrast, monocytic MDSCs www.jidonline.org 1887
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
Control
Anti-CD4 mAb
Gr-1high
Gr-1
Gr-1int
CD11b
*
Cell numbers per 1×105 cells in tumors
500
500
450
450
400
400
350
350
300
300
250
250
200
200
150
150
100
100
50
50 0
0 Control
Control
Anti–CD4 mAb
CD11b+Gr-1intMDSC 100
**
30
***
CCR2-positive cells %
CCL2 relative mRNA value
90 25 20 15 10 5 0 Anti–CD4 mAb
100 90
80
80
70
70
60
60
50
50
40
40
30
30
20
20
10
10
0 Control
Anti–CD4 mAb
CD11b+Gr-1highMDSC
0 Control
Anti–CD4 mAb
CD11b+Gr-1highMDSC
Control
Anti–CD4 mAb
CD11b+Gr-1intMDSC
Figure 4. The number of CCR2-positive myeloid-derived suppressor cells (MDSCs) increased in B16F10 tumor tissues after CD4 þ T lymphodepletion. Tumor tissues (day 14) were resected as mentioned in the Figure 2 legend. (a) The numbers of CD11b þ Gr-1 þ MDSCs within tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (*Po0.05). Data shown are from one of four individual experiments. (b) CCL2 mRNA expression was determined by real-time RT-PCR and has been shown relative to GAPDH mRNA expression (mean±SEM values) (**Po0.01). Data shown are from one of two individual experiments. (c) The proportion of CCR2-positive MDSCs to total MDSCs within tumor tissues was analyzed by flow cytometry. The bars represent mean þ SEM (***Po0.001). Data shown are from one of two individual experiments.
suppress CD8 þ T cells, predominantly via inducible nitric oxide synthase and arginase-1 enzymes and through the production of reactive nitrogen species. Monocytic MDSCs have been shown to be more suppressive than granulocytic MDSCs in antigen-stimulated CD8 þ T cells (Dolcetti et al., 2010; Youn and Gabrilovich, 2010; Gabrilovich et al., 2012; Lu and Gabrilovich, 2012). In our model, a larger number of 1888 Journal of Investigative Dermatology (2014), Volume 134
monocytic MDSCs infiltrated into tumors than granulocytic MDSCs. The infiltration of monocytic MDSCs was almost completely abrogated by in vivo depletion of CD8 þ T cells but not by depletion of NK cells (Figure 5c). Therefore, monocytic MDSCs appear to increase in tumor tissues as a specific response to the infiltration of CD8 þ T cells. This increase in the number of MDSC and the increase in IL-10
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
b 800 700 600
CCL2 relative mRNA value
CD11b+Gr–1high MDSC CD11b+Gr–1int MDSC
*
500 400 300 200 100 +
+
CD8 depletion
–
+
NK depletion
–
+
c
25 20 15 10 5
7 6 5 4 3 2 1 0
800 700
12
CD11b+Gr–1high MDSC CD11b+Gr–1int MDSC
iNOS relative mRNA value
Cell number per 1×105 cells in tumors
30
0 8
0 Anti-CD4 mAb
Arginase-1 relative mRNA value
Cell number per 1×105 cells in tumors
a
*
600 500 400 300 200 100 0
10 8 6 4 2
Anti-CD4 mAb
+
+
+
0 Anti-CD4 mAb
–
+
+
CD8 depletion
–
+
–
CD8 depletion
–
–
+
NK depletion
–
–
+
NK depletion
–
–
+
Figure 5. The increase in the number of myeloid-derived suppressor cells (MDSCs) depends on the infiltration of CD8 þ T cells. C57BL/6 mice (n ¼ 5 per group) were subcutaneously inoculated with 105 B16F10 cells on day 0. Some mice were intraperitoneally administered with anti-CD8 mAb and/or anti-asialo GM1 on days 6 and 11. On day 7, anti-CD4 mAb or control rat IgG were intratumorally administered (a and c). On day 14, the numbers of CD11b þ Gr-1 þ MDSCs within tumor tissues were analyzed by flow cytometry. The bars represent mean þ SEM (*Po0.05). (b) CCL2, arginase-1, and inducible nitric oxide synthase (iNOS) mRNA expression was determined by real-time RT-PCR, and their expression is shown relative to that of GAPDH mRNA (mean þ SEM values). Data shown are from one of two individual experiments.
expression (Figure 3a) are homeostatic responses for counteracting the antitumor immune reactions dominated by CD8 þ T cells. However, these MDSCs could not fully suppress the antitumor immune reactions induced by intratumoral CD4 þ T lymphodepletion because the reactions were so strong. Recent studies strongly suggest that CCL2-CCR2 interaction has an important role in the recruitment of MDSCs to tumor sites (Zhu et al., 2011; Gehad et al., 2012; Lesokhin et al., 2012). In contrast, Fridlender et al. (2011) showed that CCL2 blockade inhibited tumor growth but did not affect the recruitment of MDSCs. In fact, CCL2 blockade altered the macrophage phenotype, thereby activating CD8 þ T cells (Fridlender et al., 2011). Recently, Zhao et al. (2012) showed that TNF/TNF-R2 signaling has a crucial role in MDSC accumulation within tumor tissues and that the major source of TNF for MDSC accumulation could be the MDSCs themselves. Furthermore, Sevko et al. (2012) showed that lowdose cyclophosphamide therapy augmented the accumulation of MDSCs in melanoma lesions concomitantly with increased production of chronic inflammatory mediators, including TNF,
in the ret transgenic murine melanoma model. Indeed, as shown in Figure 3a, TNF expression significantly increased in tumor tissues following CD4 þ T lymphodepletion. In addition, the expression of TNF was significantly reduced by depletion of CD8 þ T cells but not by depletion of NK cells (data not shown). Although the initial trigger of TNF induction in MDSCs has not yet been determined, infiltration of CD8 þ T cells into tumor tissues might be one of the key triggers. In contrast to our results, Kjaergaard et al. (2000) showed that the antitumor effects of an OX40 agonist on a GL261 mouse glioma were abrogated in mice depleted of either CD4 þ or CD8 þ T cells. The reason for this discrepancy is not known; however, we suppose that it might be caused by the difference in the immunogenicity of tumor cells. The GL261 tumor was originally induced by intracranial injection of 3methylcholanthrene into C57BL/6 mice and is considered to be moderately immunogenic (Szatma´ri et al., 2006). In highly or moderately immunogenic tumors, CD8 þ T cells might easily accumulate and exert antitumor effects in cooperation with CD4 þ T cells. In fact, Giovarelli et al. (2000) showed www.jidonline.org 1889
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
Cell numbers per 1×105 tumor cells
a
800
*
700 600 500
CD11b+Gr–1high MDSC CD11b+Gr–1int MDSC CD8+T cells NK cells
400 300 200 100 0 Control
b
Anti-CD4 mAb + OX40 agonist
IL-2
IFN-γ
**
600
Relative mRNA value
Anti-CD4 mAb
*
500
500
400
400
300
300
200
200
100
100
0
*
600
0 Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist
Control
Control
Granzyme B
***
140
*
120
Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist Perforin
60
*
** 50
100
40
80 30 60 20
40
10
20
0
0 Control
Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist
Control
Anti-CD4 Anti-CD4 mAb mAb + OX40 agonist
Figure 6. The OX40 agonist exerts antitumor effects on CD4 þ T-lymphodepleted tumors by further enhancing the cytotoxic activities of effector cells. C57BL/6 mice (n ¼ 5 per group) were subcutaneously inoculated with B16F10 cells on day 0 and intratumorally administered anti-CD4 mAb or control IgG on day 7. Some mice were administered OX86 on days 7 and 12. On day 14, tumors were resected. (a) The numbers of myeloid-derived suppressor cells (MDSCs), CD8 þ T cells, and natural killer (NK) cells within tumor tissues were analyzed. Data shown are from one of two individual experiments. (b) IL-2, IFN-g, granzyme B, and perforin mRNA levels were determined and are shown relative to GAPDH mRNA levels. (*Po0.05; **Po0.05; ***Po0.001). Data shown are from one of three individual experiments. The bars represent mean þ SEM.
that large numbers of both CD4 þ T cells and CD8 þ T cells infiltrated an immunogenic variant of a TSA (a mouse mammary adenocarcinoma) tumor, whereas both cells sparsely infiltrated a parental poorly immunogenic TSA tumor. Therefore, we hypothesized that, in highly or moderately immunogenic tumors, treatment with an OX40 agonist can activate both CD4 þ T cells and CD8 þ T cells that accumulate within the tumor, thereby exerting their antitumor effects. In 1890 Journal of Investigative Dermatology (2014), Volume 134
fact, the OX40 agonist has been shown to directly costimulate both CD4 þ T cells and CD8 þ T cells in vitro (Taraban et al., 2002). Activated CD4 þ T helper cells help control CD8 þ T-cell-dependent tumor eradication in vivo. Therefore, antitumor effects on highly or moderately immunogenic tumors would be abolished by depletion of either CD4 þ or CD8 þ T cells. To verify this hypothesis, we examined the antitumor effects of an OX40 agonist using a Cloudman S91 (clone M-3) murine melanoma (of DBA/2 origin), which is generally considered to be more immunogenic than B16F10 melanomas. The Cloudman S91 melanoma had a slight tendency to respond to OX40 agonist treatment alone in the late stage; however, the difference from the control was not statistically significant. Intratumoral CD4 þ T lymphodepletion caused more marked shrinkage of the S91 melanoma than of the B16F10 melanoma. CD4 þ T lymphodepletion plus OX40 agonist treatment led the greatest reduction in tumor size; however, the difference from the size of CD4 þ T-lymphodepleted tumors was not statistically significant, and lymphodepletion plus OX40 agonist treatment led to rejection of the tumors in some mice on day 27 (when mice in the other groups showed no tumor rejection) (Supplementary Figure S3 online). Therefore, our hypothesis could not be verified; however, it might result from the still poor immunogenicity of Cloudman S91 melanoma. However, in another respect, these results seem significant because CD4 þ T lymphodepletion plus OX40 agonist treatment was found to be effective for melanomas with different immunogenicities and from different mouse strains (Figure 1d and Supplementary Figure S3 online). Although previous studies have shown that the therapeutic efficacy of OX40 agonists depends on tumor immunogenicity, this efficacy might be closely correlated with the degree of CD8 þ T-cell infiltration within tumor tissues. In fact, in B16F10 tumors, the infiltration of CD8 þ T cells has been found to be sparse (Nagai et al., 2000a). Because the antitumor effects of IL-12 gene-transfected B16F10 cells were mediated by both CD8 þ T cells and NK cells (Nagai et al., 2000b), whereas those of IL-27 gene-transfected B16F10 cells were mediated by NK cells but not CD8 þ T cells (Oniki et al., 2006), the OX40 agonist might have induced antitumor effects on IL-12 gene-transfected B16F10 cells but not on IL-27 gene-transfected B16F10 cells (Figure 1b and c). As CD4 þ T lymphodepletion permits the accumulation of CD8 þ T cells in B16F10 tumors (Figure 2), the OX40 agonist could exert antitumor effects even if tumor cells are poorly immunogenic, by further activating the tumor-infiltrating CD8 þ T cells. In fact, it has been shown that OX40 can directly costimulate anti-CD3-activated CD8 þ T cells and antigen-specific CD8 þ T cells in vitro (Taraban et al., 2002). Recently, several studies have focused on the effect of OX40 on the immunosuppressive population, that is, Tregs and MDSCs. OX40 engagement has been shown to inhibit Tregs in tumor tissues and exert antitumor activities (Piconese et al., 2008; Hirschhorn-Cymerman et al., 2009). In contrast, it was found that OX40 stimulation promotes Tregs in an appropriate cytokine milieu (Ruby et al., 2009; Xiao et al., 2012). The latter finding means that OX40 stimulation can inhibit antitumor immune responses, depending on the tumor
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
cytokine milieu, via activation of Tregs. However, in our model, OX40-stimulated Treg-mediated inhibition of antitumor immune responses could be almost completely avoidable in the initiation of OX40 agonist treatment, because intratumoral administration with anti-CD4 mAb efficiently depleted Tregs within tumors, along with helper T cells (Supplementary Figure S1 online). On the other hand, concerning MDSCs, OX40 agonist therapy has been found to significantly enhance CD8 þ T-cell infiltration and decrease MDSC population in the MCA205 H12 tumor site. The reduction in the number of MDSCs was caused by the indirect effects of the OX40 agonist because OX40 expression in MDSCs was not detected at the tumor site (Gough et al., 2008). In our study, systemic treatment with the OX40 agonist did not reduce the number of MDSCs that were recruited into the tumor site following CD4 þ T lymphodepletion (Figure 6a). In contrast, OX40 agonist treatment further enhanced the expression of arginase-1 and inducible nitric oxide synthase in CD4 þ T-lymphodepleted tumors, which may be indicative of a more functionally active condition of MDSCs (Supplementary Figure S4 online). We speculate that these functional changes might occur in response to the enhancement of cytotoxic activities of effector cells by the OX40 agonist. The activation of MDSCs as a protective response might increase with increase in the antitumor activities. Although OX40 agonists seem to be promising candidates for immunotherapy for human tumors (Weinberg et al., 2011), the effects of OX40 agonist monotherapy for patients with malignant tumors will be limited because most human tumors are poorly immunogenic. On the other hand, several studies have shown that autologous adoptive tumor-infiltrating lymphocyte transfer after lymphodepletion can cause the regression of refractory metastatic melanomas (Rosenberg and Dudley, 2004; Dudley et al., 2005). Thus, our results support the hypothesis that a combination of OX40 agonist therapy and CD4 þ T lymphodepletion might exert efficient antitumor effects even on poorly immunogenic human melanomas and suggest that lymphodepletion could facilitate the accumulation of tumor-infiltrating lymphocytes into tumors, thereby inducing antitumor effects; however, the number of MDSCs might simultaneously increase within tumors and be activated as a countermeasure against accumulation and cytotoxicity mediated by tumor-infiltrating lymphocytes. In conclusion, this combination therapy could be a promising therapeutic strategy for patients with poorly immunogenic melanomas and might be more effective if further combined with therapeutic strategies that can selectively deplete or inhibit MDSCs (Nagaraj et al., 2010; Meyer et al., 2011; Kodumudi et al., 2012; Lesokhin et al., 2012; Srivastava et al., 2012). MATERIALS AND METHODS Tumor cell lines and animal studies B16F10 melanoma cells (provided by the RIKEN BRC through the National Bio-resource Project of MEXT, Japan), IL-12 cDNA-transfected B16F10 cells, and IL-27 cDNA-transfected B16F10 cells (Oniki et al., 2006) were maintained in Eagle’s modified essential medium supplemented with 5% fetal bovine serum. Cloudman S91 (clone
M-3) melanoma cells were obtained from American Type Culture Collection (ATCC, Rockville, MD). C57BL/6N mice and DBA/2 mice (6- to 8-week-old females) were purchased from Charles River Laboratories (Tokyo, Japan). All animal experiments were conducted according to the Guidelines for Animal Experimentation at the Kobe University Graduate School of Medicine.
In vivo antibody treatments We administered 500 mg of rat GK1.5 mAb (ATCC) into the tumors on day 7 after tumor inoculation to deplete CD4 þ T cells. CD8 þ T cells in vivo were depleted using the rat mAb 2.43 (ATCC). NK cells were depleted using a polyclonal anti-asialo GM1 antibody (Wako Fine Chemicals, Osaka, Japan). OX40 agonist immunotherapy was performed with a rat anti-OX40 agonistic mAb (OX86; ATCC). One milligram of OX86 was administered intraperitoneally on day 0 or day 7 and every 5 days thereafter. Rat IgG (Wako Fine Chemicals) was used as the control antibody.
Flow cytometry Resected tumors were digested with Liberase (0.2 Wunsch units/ml; Roche Diagnostics K.K., Basel, Switzerland) and DNase I (0.02 mg/ ml; Roche) at 37 1C for 60 min. These cells were stained with different combinations of the following antibodies: CD3-FITC, CD8-APC, NK1.1-PE, CXCR3-PE-Cy7, CD11b-PE-Cy7, CCR2-APC (BioLegend, San Diego, CA), CD4-FITC Gr-1-PE, Foxp3-PE (Miltenyi Biotec, Bergisch Gladbach, Germany). Cytometric acquisition was performed with a MoFlo XDP system (Beckman Coulter, Brea, CA) or BD FACSVerse (BD Biosciences) and the Summit 5.2 (Beckman Coulter) or Flowjo (Tree Star, Ashland, OR). Supplementary Figure S5 online shows gating procedures of tumor samples.
Real-time quantitative reverse transcriptase–PCR (qRT-PCR) analysis Total RNA was extracted by using the QuickGene RNA tissue kit (Fujifilm Corporation, Tokyo, Japan). cDNA synthesis was performed using the PrimeScript RT reagent kit (Takara Bio, Shiga, Japan). Realtime qRT-PCR was performed using an Applied Biosystems 7500 realtime PCR system (Applied Biosystems, Foster City, CA) and SYBR Premix Ex Taq II (Takara Bio). mRNA expression values were normalized according to the internal control glyceraldehyde 3phosphate dehydrogenase (GAPDH), and the relative expression values were plotted.
Statistical analysis The statistical significance of differences between the means of two groups was determined using Student’s t-test. The statistical significance of three or more groups was determined using Dunnett’s test or Tukey’s HSD post hoc test. The differences were considered statistically significant at Po0.05. Data are shown as mean±SEM values.
CONFLICT OF INTEREST The authors state no conflict of interest.
ACKNOWLEDGMENTS This study was supported in part by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology of Japan, Japanese Society for Pigment Cell Research, and Japanese Dermatological Association basic medical research grant (Shiseido donation).
www.jidonline.org 1891
S Fujiwara et al. Therapy Using OX40 Agonist and CD4 Depletion
SUPPLEMENTARY MATERIAL Supplementary material is linked to the online version of the paper at http:// www.nature.com/jid
REFERENCES Croft M (2010) Control of immunity by the TNFR-related molecule OX40 (CD134). Annu Rev Immunol 28:57–78 Dolcetti L, Peranzoni E, Ugel S et al. (2010) Hierarchy of immunosuppressive strength among myeloid-derived suppressor cell subsets is determined by GM-CSF. Eur J Immunol 40:22–35 Dudley ME, Wunderlich JR, Yang JC et al. (2005) Adoptive cell transfer therapy following non-myeloablative but lymphodepleting chemotherapy for the treatment of patients with refractory metastatic melanoma. J Clin Oncol 23:2346–57 Fidler IJ (1973) Selection of successive tumour lines for metastasis. Nat New Biol 242:148–9 Fridlender ZG, Kapoor V, Buchlis G et al. (2011) Monocyte chemoattractant protein-1 blockade inhibits lung cancer tumor growth by altering macrophage phenotype and activating CD8 þ cells. Am J Respir Cell Mol Biol 44:230–7 Gabrilovich DI, Ostrand-Rosenberg S, Bronte V (2012) Coordinated regulation of myeloid cells by tumours. Nat Rev Immunol 12:253–68 Gehad AE, Lichtman MK, Schmults CD et al. (2012) Nitric oxide-producing myeloid-derived suppressor cells inhibit vascular E-selectin expression in human squamous cell carcinomas. J Invest Dermatol 132:2642–51 Giovarelli M, Cappello P, Forni G et al. (2000) Tumor rejection and immune memory elicited by locally released LEC chemokine are associated with an impressive recruitment of APCs, lymphocytes, and granulocytes. J Immunol 164:3200–6 Gough MJ, Ruby CE, Redmond WL et al. (2008) OX40 agonist therapy enhances CD8 infiltration and decreases immune suppression in the tumor. Cancer Res 68:5206–15 Hirschhorn-Cymerman D, Rizzuto GA, Merghoub T et al. (2009) OX40 engagement and chemotherapy combination provides potent antitumor immunity with concomitant regulatory T cell apoptosis. J Exp Med 206:1103–16 Ishii N, Takahashi T, Soroosh P et al. (2010) OX40-OX40 ligand interaction in T-cell-mediated immunity and immunopathology. Adv Immunol 105:63–98 Jensen SM, Maston LD, Gough MJ et al. (2010) Signaling through OX40 enhances antitumor immunity. Semin Oncol 37:524–32 Kjaergaard J, Tanaka J, Kim JA et al. (2000) Therapeutic efficacy of OX-40 receptor antibody depends on tumor immunogenicity and anatomic site of tumor growth. Cancer Res 60:5514–21 Kodumudi KN, Weber A, Sarnaik AA et al. (2012) Blockade of myeloid-derived suppressor cells after induction of lymphopenia improves adoptive T cell therapy in a murine model of melanoma. J Immunol 189:5147–54 Kryczek I, Wei S, Zou L et al. (2007) Cutting edge: Th17 and regulatory T cell dynamics and the regulation by IL-2 in the tumor microenvironment. J Immunol 178:6730–3 Langowski JL, Zhang X, Wu L et al. (2006) IL-23 promotes tumour incidence and growth. Nature 442:461–5 Lesokhin AM, Hohl TM, Kitano S et al. (2012) Monocytic CCR2( þ ) myeloidderived suppressor cells promote immune escape by limiting activated CD8 T-cell infiltration into the tumor microenvironment. Cancer Res 72:876–86 Lindau D, Gielen P, Kroesen M et al. (2013) The immunosuppressive tumour network: myeloid-derived suppressor cells, regulatory T cells and natural killer T cells. Immunology 138:105–15 Lu T, Gabrilovich DI (2012) Molecular pathways: tumor-infiltrating myeloid cells and reactive oxygen species in regulation of tumor microenvironment. Clin Cancer Res 18:4877–82 Lo´pez M, Aguilera R, Pe´rez C et al. (2006) The role of regulatory T lymphocytes in the induced immune response mediated by biological vaccines. Immunobiology 211:127–36
1892 Journal of Investigative Dermatology (2014), Volume 134
Meyer C, Sevko A, Ramacher M et al. (2011) Chronic inflammation promotes myeloid-derived suppressor cell activation blocking antitumor immunity in transgenic mouse melanoma model. Proc Natl Acad Sci USA 108:17111–6 Nagai H, Hara I, Horikawa T et al. (2000a) Antitumor effects on mouse melanoma elicited by local secretion of interleukin-12 and their enhancement by treatment with interleukin-18. Cancer Invest 18:206–13 Nagai H, Hara I, Horikawa T et al. (2000b) Elimination of CD4( þ ) T cells enhances anti-tumor effect of locally secreted interleukin-12 on B16 mouse melanoma and induces vitiligo-like coat color alteration. J Invest Dermatol 115:1059–64 Nagaraj S, Youn JI, Weber H et al. (2010) Anti-inflammatory triterpenoid blocks immune suppressive function of MDSCs and improves immune response in cancer. Clin Cancer Res 16:1812–23 Oniki S, Nagai H, Horikawa T et al. (2006) Interleukin-23 and interleukin-27 exert quite different antitumor and vaccine effects on poorly immunogenic melanoma. Cancer Res 66:6395–404 Pardee AD, McCurry D, Alber S et al. (2010) A therapeutic OX40 agonist dynamically alters dendritic, endothelial, and T cell subsets within the established tumor microenvironment. Cancer Res 70:9041–52 Peranzoni E, Zilio S, Marigo I et al. (2010) Myeloid-derived suppressor cell heterogeneity and subset definition. Curr Opin Immunol 22:238–44 Piconese S, Valzasina B, Colombo MP (2008) OX40 triggering blocks suppression by regulatory T cells and facilitates tumor rejection. J Exp Med 205:825–39 Rosenberg SA, Dudley ME (2004) Cancer regression in patients with metastatic melanoma after the transfer of autologous antitumor lymphocytes. Proc Natl Acad Sci USA 101(Suppl 2):14639–45 Ruby CE, Yates MA, Hirschhorn-Cymerman D et al. (2009) Cutting Edge: OX40 agonists can drive regulatory T cell expansion if the cytokine milieu is right. J Immunol 183:4853–7 Seo N, Hayakawa S, Takigawa M et al. (2001) Interleukin-10 expressed at early tumour sites induces subsequent generation of CD4( þ ) T-regulatory cells and systemic collapse of antitumour immunity. Immunology 103:449–57 Sevko A, Sade-Feldman M, Kanterman J et al. (2012) Cyclophosphamide promotes chronic inflammation-dependent immunosuppression and prevents antitumor response in melanoma. J Invest Dermatol 133:1610–9 Srivastava MK, Zhu L, Harris-White M et al. (2012) Myeloid suppressor cell depletion augments antitumor activity in lung cancer. PLoS ONE 7:e40677 Sugamura K, Ishii N, Weinberg AD (2004) Therapeutic targeting of the effector T-cell co-stimulatory molecule OX40. Nat Rev Immunol 4:420–31 Szatma´ri T, Lumniczky K, De´saknai S et al. (2006) Detailed characterization of the mouse glioma 261 tumor model for experimental glioblastoma therapy. Cancer Sci 97:546–53 Taraban VY, Rowley TF, O’Brien L et al. (2002) Expression and costimulatory effects of the TNF receptor superfamily members CD134 (OX40) and CD137 (4-1BB), and their role in the generation of anti-tumor immune responses. Eur J Immunol 32:3617–27 Weinberg AD, Morris NP, Kovacsovics-Bankowski M et al. (2011) Science gone translational: the OX40 agonist story. Immunol Rev 244: 218–31 Weinberg AD, Rivera MM, Prell R et al. (2000) Engagement of the OX-40 receptor in vivo enhances antitumor immunity. J Immunol 164:2160–9 Xiao X, Gong W, Demirci G et al. (2012) New insights on OX40 in the control of T cell immunity and immune tolerance in vivo. J Immunol 188: 892–901 Youn JI, Gabrilovich DI (2010) The biology of myeloid-derived suppressor cells: the blessing and the curse of morphological and functional heterogeneity. Eur J Immunol 40:2969–75 Zhao X, Rong L, Li X et al. (2012) TNF signaling drives myeloid-derived suppressor cell accumulation. J Clin Invest 122:4094–104 Zhu X, Fujita M, Snyder LA et al. (2011) Systemic delivery of neutralizing antibody targeting CCL2 for glioma therapy. J Neurooncol 104:83–92
