In vivo anti-tumor effect is associated with increased T cells infiltrations
To evaluate possible antitumor effects of IDX on the growth of nasopharyngeal NPC43+ve tumors in nude mice, test animals were treated every-other-day with IDX and cis at a dosage of 10 and 20mg/kg and 5mg/kg weight/day, respectively. The growth curves demonstrated that high combination-treatment (IDX20+cis5) significantly decreased tumor growth, compared to the control group (Figure 7A, B). Although athymic nude mice are grossly deficient in peripheral T cells, the number of lymphocytes bearing T-cell markers (CD4, CD8) increases steadily with age (Kennedy, Pierce & Lake, 1992). Analogously, in our model, while CD14+ (monocytes) and CD19+ (B cells) lymphocytes represented as the main two subsets, CD4+ and CD8+ T cells were also detectable in the circulation across all experimental groups. (Supplementary Figure 4A-C; before and after drug treatment for CD4+ T cells was 19.47±5.877 and 18.01±5.411, respectively; CD8+ T cells was 4.971±1.390 and 9.675±5.213, respectively). In relation to tumor-infiltrating lymphocytes (TILs), we observed significantly higher number of infiltrated CD8+ T cells within CD19- population in the tumor of IDX20+Cis-treated mice than that detected in the cis group (Figure 7C; 2.860±1.548 and 8.482±1.758, respectively). Importantly, the percentages of CD8+ T cells, albeit at low levels among all live TILs, was significantly higher in IDX20+cis-treated mice than that observed in the DMSO group (Figure 7D; 1.668±0.5130 and 5.400±1.426, respectively), which was confirmed by immunostaining (Supplementary Figure 4D). However, there were no significant differences in the percentage of infiltrating CD4+ T cells among all groups.
Finally, we evaluated the clinical relevance of our signature findings by investigating the expression of ENOX2 as a potential link to IDX’s pronounced anti-cancer effect in NPC. From a publicly available sequencing cohort GSE150430, we observed that malignant cells were the main cell type expressed ENOX2 at single-cell resolution (Figure 7E-F), this feature was further confirmed at protein level by histological staining (Figure 7G). Another three publicly microarray GEO datasets (GSE12452, GSE53819 and GSE34573) illustrated that ENOX2 gene was present at all stages of cancer progression (I–IV) and expressed at a significantly higher levels in tumors as compared to normal tissues, which was associated with decreased DNA methylation as evident in the GSE52068 dataset (Figure 7G). As such, the antitumor-responsive ENOX2 activity might be closely associated with NPC cancer, and thus IDX treatment could potentially elicit a tumor-killing and immune-mediated destruction in NPC tumor cells, while adding cis in combination could further enhance this effect.
Discussion and Conclusions
Previously, most studies with IDX focused on its anticancer properties in association with cell cycle arrest and via a range of apoptotic pathways in tumor cells, with little attention given to the drugs immuno-oncology potential. Therefore, for the first time, we have discovered that low-dose IDX acts as an immune stimulus that can induce immunogenic tumor cell death by facilitating T-cell priming and trafficking into tumors. Three experimental challenges were overcome to allow this finding: determining the appropriate concentration of IDX enabling an effective immune response in addition to its anti-tumor effect, elucidating the functions of IDX on tumor-lymphocyte interactions and finally, seeking the activation pathways involved in fighting NPC disease. The immune-stimulating properties of IDX might therefore lay a new foundation toward giving this unique molecule a second chance at contributing to the future of cancer treatment.
We first determined the cytotoxicity of IDX against a range of NPC cancer cell lines. Although IDX was found to be less cytotoxic than cis, the relative restricted range of IC50 values reflected its homogenous response across the various cancer cell lines. Additionally, we confirmed previous findings showing that IDX further sensitizes these cells to apoptosis in combination with a chemotherapeutic drug. Our interest in determining how IDX impacts immune function stems in part from results obtained during a past human clinical study. In 2004, phenoxodiol gained fast track approval from the FDA and was tested in phase II/III trials in humans (Mor, Fu & Alvero, 2006) against various types of primary and metastatic cancers. However, lymphocytopenia was associated with phenoxodiol use (Herst, Davis, Neeson, Berridge & Ritchie, 2009), suggesting that dosing was not fully optimised in these studies. Herein, we confirmed previous findings (Georgaki et al., 2009) and demonstrated that whilst IDX at high concentrations (4µM) could severely inhibit T cell proliferation in vitro , if given at appropriate dosages (1-2µM), IDX can act as an immune-enhancer in addition to its anticancer effects by allowing proliferating T cells to unleash their cytotoxic ability. This finding further supports the application of this agent in future therapeutic interventions aiming to activate tumor-reactive lymphocytes.
As T cells are one of the main effectors of the antitumor immune response and are often suppressed during the disease, our finding that IDX positively modulates tumor-lymphocyte interactions is of significance. In our 2D coculture assay, we observed that conditioning of tumors with IDX, and to a larger extent upon combination treatment of IDX and cis, drives the expansion, activation and migration of T-cells. The effect is thought to be partially type I IFN/CXCL10-dependent in this study. To date, the STING field has been focused on the concept that it activates a predominant IRF3 response which has the potential to help recruit dendritic cells and CD8+ T cells to the TME, thereby promoting anti-tumor activities. While activation of STING in the myeloid compartment clearly promotes antitumor effects, the role of STING in tumor cells, and the role it may play in the antitumor response is starting to gain attention. Herein, we observed that low concentrations of IDX had no impact on endogenous STING expression, whilst higher IDX concentrations in combination with cisplatin down-regulated STING protein in NPC cell lines. Unexpectedly, a strong reduction in IFNα which is responsible for the upregulation of T-helper 1 cell chemokines, such as CXCL10, was only observed following cis treatment in NPC cells. Accordingly, the expressions of CXCL10, an important chemokine for the homing of antigen-presenting cells and trafficking of CD8+ T-cells, were attenuated. This raised an important question as to why IDX is capable of sustaining IFNα and drive up-regulation of CXCL10 expressions in NPC cells given that the STING pathway was partially impaired. As such, we drew our attention toward a recent discovery pointing at caspase cascade functions during apoptosis to prevent dying cells from producing IFN (White et al., 2014). Likewise, we observed that IDX modestly induced the release of mitochondrial DNA (mtDNA, data not shown), while cis was more likely to trigger cleaved apoptotic caspase-7, and the latter fed back to inhibit IFNα production by the dying cell. This finding highlights that, at least when it comes to the mode of action, IDX and cis exert their functions through different apoptotic downstream signaling pathways. Whether the elevation in IFNα caused by loss of the caspase cascade leading to the induction of chemoattractant is the result of upstream mitochondrial damage or an alternative signaling mechanism, and whether perturbation in STING activity might contribute to the immunological impact an apoptotic cell has on the NPC cells by inducing mtDNA remain to be established.
We further developed heterotypic cocultures of NPC tumor spheroids with T cells to study the infiltration, phenotype and function of infiltrates. Critically, this study showed that activated/memory DP CD4+CD8+ T cells, acquiring lower expression of CD62L receptor and PD-1 marker, were able to infiltrate spheroids and kill tumor cells, and that CXCL10 was involved in these processes. Expression of CD45RO was the classical marker of primed memory CD8 T cells, which could be further divided into CCR7+CD62L+ central memory and CCR7-CD62L- effector memory cells based on the assumption that the presence of CD62L and CCR7 would facilitate entry of central memory cells to secondary lymphoid tissue from blood, whereas effector memory cells home to peripheral tissues. As such, our results pointed at another potential mechanism through which IDX could render the accumulation of exhausted PD1+TILs into tumor cells. Given the memory phenotype, cytotoxic profile, and migratory capacity that we observed for DP T lymphocytes, we could speculate that multiple immune activation circumstances exerted by IDX could favor the presence of these cells in the periphery of, and potentially inside, the TME. Further studies are deemed vital to better understand the mechanisms controlling the co-expression of CD4 and CD8 lymphocytes and whether such co-expression confers any functional advantage. We further confirmed our results in an in vivotherapeutic animal tumor model, confirming that this combination strategy was capable of inducing the accumulation of T cells within tumor sites and abrogate tumor growth.
Collectively, the multifaceted behavior of IDX provides the potential to synergize with, or complement, a wide range of cancer therapies, and overcome resistance mechanisms. It is therefore possible that IDX could condition or sensitize NPC patients to chemotherapy by converting “cold” or “altered” tumors into “hot” tumors, thereby increasing the chance of a durable response to targeted therapy.