Role of hypoxia and hypoxia-inducible factor-1 in tumour immune escape

Previous studies revealed that, upon exposure to hypoxia, tumour cells acquire resistance to the cytolytic activity of IL-2-activated lymphocytes. The MHC class I chain-related (MIC) molecules – comprised of MICA and MICB – are ligands for the activating NKG2D receptor on Natural Killer (NK) and CD8+ T cells. MIC-NKG2D interactions lead to the activation of NK and CD8+ T cells and the subsequent lysis of the tumour cells. The study also showed that the mechanism of the hypoxia-mediated immune escape involves the shedding of MIC, specifically MICA, from the tumour cell surface. The objective of the present study was to determine whether the shedding of MICA requires the expression of hypoxia inducible factor-1 (HIF-1), a transcription factor that regulates cellular adaptations to hypoxia. Exposure to hypoxia (0.5% O2 vs. 20% O2) led to the shedding of MIC from the surface of MDA-MB-231 human breast cancer cells and DU-145 human prostate cancer cells as determined by flow cytometry. Knockdown of HIF-1α mRNA using siRNA technology resulted in inhibition of HIF-1α accumulation under hypoxic conditions as determined by Western blot analysis. Parallel study revealed that knockdown of HIF-1α also blocked the shedding of MICA from the surface of MDA-MB-231 cells exposed to hypoxia. These results indicate that HIF-1 is required for the hypoxia-mediated shedding of MICA and, consequently, that HIF-1 may play an important role in tumour immune escape. Ongoing studies aim to determine the HIF-1 target genes involved in the shedding of MICA under hypoxia.

Role of Nitric Oxide-Signalling in Hypoxia-Induced Immune Escape in Cancer

A key step in malignant progression is the acquired ability of tumour cells to escape immune-mediated lysis. A potential mechanism by which tumour cells avoid immune destruction involves the shedding of MHC Class I Chain-Related Protein A (MICA), a Natural Killer (NK) cell-activating ligand, from the tumour cell membrane. Hypoxia has been shown to cause increased MICA shedding; however, this hypoxia-induced effect can be attenuated by pharmacological activation of the cyclic guanosine monophosphate (cGMP)-dependent nitric oxide (NO)-signalling pathway in cancer cells. The primary objective of the present study was to determine whether treatment of tumour-bearing nude mice with the NO-mimetic glyceryl trinitrate (GTN) attenuates in vivo tumour growth and if so, whether this effect is dependent on the presence of an intact NK cell compartment. Results indicated that continuous transdermal administration of GTN (1.8 µg/h) can significantly attenuate the growth of transplanted human DU-145 prostate tumours but that this effect of GTN is lost in mice whose NK-cells have been depleted. Tumours and serum from the mice in this study were analysed to determine whether GTN treatment had any effect on the expression levels of proteins integral to the proposed MICA shedding mechanism; however, the results of these studies were inconclusive. As phosphodiesterase (PDE) inhibition represents a potential method to enhance NO-signalling, experiments were performed to determine whether treatment with the PDE5/6 inhibitor zaprinast could also attenuate hypoxia-induced MICA shedding and decrease in vivo growth of DU-145 tumours. Results demonstrated that treatment with zaprinast (10 mg/kg) significantly attenuates MICA shedding in DU-145 cancer cells and significantly decreases in vivo tumour growth. Taken together, the results of these experiments indicate that GTN attenuates tumour growth by sensitising tumour cells to innate immunity, likely by increasing membrane-associated tumour cell MICA levels through the reactivation of NO-signalling, and that zaprinast decreases tumour growth likely through a similar mechanism. These findings are important because they indicate that agents capable of reactivating NO-signalling, such as NO-mimetics and PDE inhibitors, can potentially be used as immunosensitisers in the treatment and/or prevention of cancer.

The PD-1/PD-L1 Axis: An Immune-Mediated Mechanism of Chemoresistance in Cancer

The ability of tumour cells to avoid immune destruction (immune escape) and their acquired resistance to anti-cancer drugs constitute important barriers to the successful management of cancer. The interaction between specific molecules on the surface of tumour cells with their corresponding receptors on immune effector cells can result in inhibition of these effector cells, consequently allowing tumour cells to evade the host’s anti-tumour immune response. The interaction of the Programmed Death Ligand 1 (PD-L1) on the surface of tumour cells with the Programmed Death-1 (PD-1) receptor on cytotoxic T lymphocytes leads to inactivation of these immune effectors, and is a specific example of an immune escape mechanism tumour cells use to avoid immune destruction. Clinically, antibodies capable of blocking the PD-1/PD-L1 interaction have demonstrated significant therapeutic benefit, and are currently being used to help bolster patients’ immune response against malignant cells in a variety of cancer types. Here we show that the PD-1/PD-L1 interaction also leads to tumour cell resistance to conventional chemotherapeutic agents. Incubation of PD-L1-expressing human and mouse tumour cells with PD-1-expressing Jurkat T cells or purified recombinant PD-1 resulted in tumour cell resistance to doxorubicin and docetaxel. Interference with the PD-1/PD-L1 interaction using blocking anti-PD-1 or anti-PD-L1 antibody or shRNA-mediated gene silencing resulted in attenuation of PD-1/PD-L1-mediated drug resistance. Moreover, inhibition of the PD-1/PD-L1 signalling axis using anti-PD-1 antibody enhanced the effect of doxorubicin chemotherapy to inhibit 4T1 tumour cell metastasis in an in vivo mouse model of mammary carcinoma. These findings indicate that blockade of the PD-1/PD-L1 axis may be a useful approach to immunosensitize and chemosensitize tumours in cancer patients and provide a rationale for the use of anti-PD-1/PD-L1 antibodies as adjuvants to chemotherapy.

FES KINASE SIGNALING PROMOTES MAST CELL RECRUITMENT TO TUMOURS

FES protein-tyrosine kinase (PTK) activation downstream of the KIT receptor in mast cells (MC) promotes cell polarization and migration towards the KIT ligand Stem cell factor (SCF). A variety of tumours secrete SCF to promote MC recruitment and release of mediators that enhance tumour vascularization and growth. This study investigates whether FES promotes MC migration via regulation of microtubules (MTs), and if FES is required for MC recruitment to the tumour microenvironment. MT binding assays showed that FES has at least two MT binding sites, which likely contribute to the partial co-localization of FES with MTs in polarized bone marrow-derived mast cells (BMMCs). Live cell imaging revealed a significant defect in chemotaxis of FES-deficient BMMCs towards SCF embedded within an agarose drop, which correlated with less MT organization compared to control cells. To extend these results to a tumour model, mouse mammary carcinoma AC2M2 cells were engrafted under the skin and into the mammary fat pads of immune compromised control (nu/nu) or FES-deficient (nu/nu:fes-/-) mice. A drastic reduction in tumour-associated MCs was observed in FES-deficient mice compared to control in both mammary and skin tissue sections. This correlated with a trend towards reduced tumour volumes in FES-deficient mice. These results implicate FES signaling downstream of KIT, in promoting MT reorganization during cell polarization and for chemotaxis of MCs towards tumour-derived SCF. Thus, FES is a potential therapeutic target to limit recruitment of stromal mast cells or macrophages to solid tumours that enhance tumour progression.

ROLE OF STAT1 AND CXCL10 IN THE TUMOUR IMMUNE MICROENVIRONMENT OF HIGH-GRADE SEROUS OVARIAN CANCER

High-grade serous ovarian cancer (HGSC) is the most prevalent epithelial ovarian cancer characterized by late detection, metastasis and resistance to chemotherapy. Previous studies on the tumour immune microenvironment in HGSC identified STAT1 and CXCL10 as the most differentially expressed genes between treatment naïve chemotherapy resistant and sensitive tumours. Interferon-induced STAT1 is a transcription factor, which induces many genes including tumour suppressor genes and those involved in recruitment of immune cells to the tumour immune microenvironment (TME), including CXCL10. CXCL10 is a chemokine that recruits tumour infiltrating lymphocytes (TILs) and exhibits angiostatic function. The current study was performed to determine the effects of differential STAT1 and CXCL10 expression on HGSC disease progression and TME. STAT1 expression and intratumoural CD8+ T cells were evaluated as prognostic and predictive biomarkers via immunohistochemistry on 734 HGSC tumours accrued from the Terry Fox Research Institute-Canadian Ovarian Experimental Unified Resource. The combined effect of STAT1 expression and CD8+ TIL density was confirmed as prognostic and predictive companion biomarkers in the second independent biomarker validation study. Significant positive correlation between STAT1 expression and intratumoral CD8+ TIL density was observed. The effects of enforced CXCL10 expression on HGSC tumour growth, vasculature and immune tumour microenvironment were studied in the ID8 mouse ovarian cancer cell engraftment in immunocompetent C57BL/6 mice. Significant decrease in tumour progression in mice injected with ID8 CXCL10 overexpressing cells compared to mice injected with ID8 vector control cells was observed. Multiplexed cytokine analysis of ascites showed differential expression of IL-6, VEGF and CXCL9 between the two groups. Endothelial cell marker staining showed differences in tumour vasculature between the two groups. Immune transcriptomic profiling identified distinct expression profiles in genes associated with cytokines, chemokines, interferons, T cell function and apoptosis between the two groups. These findings provide evidence that STAT1 is an independent biomarker and in combination with CD8+ TIL density could be applied as novel immune-based biomarkers in HGSC. These results provide the basis for future studies aimed at understanding mechanisms underlying differential tumour STAT1 and CXCL10 expression and its role in pre-existing tumour immunologic diversity, thus potentially contributing to biomarker guided immune modulatory therapies.

The role of vitamin D signalling in the interaction between mesenchymal mammary tumour cells and macrophages

The transition of epithelial-like tumour cells to those exhibiting mesenchymal characteristics (Epithelial-to-mesenchymal Transition; EMT) is an integral process in breast cancer metastasis. EMT can be promoted by Transforming growth factor-beta (TGF-β) which can be found at high levels in the tumour stroma. Tumour-associated macrophages (TAMs) can also induce EMT in breast cancer cells, which is one way that they promote breast cancer metastasis. Vitamin D signalling has been implicated in EMT suppression and plays a role in modulating macrophage differentiation and stimulating their anti-inflammatory functions. This project had two major aims. First, we aimed to create and verify a unique fluorescent reporter gene construct designed to evaluate the dynamics of EMT in real-time and at the single-cell level. While some components of this reporter system were successfully validated, work to complete the final reporter construct is ongoing. The second and main aspect of this project focused on exploring the ability of 1,25-dihydroxyvitamin D3 (1,25D3) to modulate the interaction between mesenchymal mammary tumour cells and TAMs. Unexpectedly, in short-term treatment (48 hours) studies of 4T1 murine mammary tumour cells, we observed that 1,25D3 and TGF-β signalling work together to increase expression of the mesenchymal markers, Snai1, Fn1, and Col1a1. 1,25D3 and TGF-β also synergistically activate transcription of the gene encoding the 1,25D3-catabolizing enzyme, Cyp24a1. The ability of 1,25D3 and TGF-β to enhance expression of these genes was diminished in a long-term treatment (14 days) of 4T1 cells, and this effect was accompanied by a decrease in cell proliferation. 1,25D3 may also cooperate with cytokines produced by normal macrophages and macrophages considered to be TAM-like. Conditioned media experiments revealed that in the presence of factors from normal macrophages, 1,25D3 enhanced expression of Fn1, and in the presence of factors from TAM-like macrophages, 1,25D3 enhanced expression of Fn1 and Cyp24a1. Rather than mitigating the interaction as hypothesized, 1,25D3 may exacerbate the tumour-promoting effects of the EMT-TAM relationship. Also, signalling pathways involved in the EMT-TAM relationship may synergize with 1,25D3 to upregulate Cyp24a1 expression. These findings are important for understanding the potential of vitamin D compounds to be used in the treatment of breast cancer.

Impaired GABA Plasticity at ovBNST Synapses Predicts Compulsive Drinking in Schedule-Induced Polydipsia

Schedule-Induced Polydipsia (SIP) is an animal model of adjunctive drinking induced when a hungry rat receives food on a fixed interval of time. This model has been implemented as a model of compulsive behaviour and may represent a powerful tool to understand the neural mechanisms of compulsion. The bed nucleus of the stria terminalis (BNST) is thought to translate challenges to energy homeostasis into consummatory behaviours, and is therefore likely to contribute to drinking behaviours displayed by food restricted rats in the SIP paradigm. Furthermore, the BNST seems implicated in various compulsive behaviors, including compulsive water drinking in rats. Therefore, the goal of this project was to determine whether compulsive drinking in the SIP paradigm was associated with alterations in transmission at oval BNST (ovBNST) synapses. Rats undergoing the SIP procedure had restricted food access (1-hours/day) for a total of 29 days. After 7 days of food restriction and for the next 21 consecutive days, the rats had daily 2-hour access to operant conditioning chambers where they were presented with a 45-mg food pellet every minute. Water consumed during these 2-hour sessions was measured and the rats that drank 15 ml or more water for a minimum of 3 consecutive days were considered High Drinkers (HD; n=17) or otherwise, Low Drinkers (LD; n=13). Brain slices whole-cell patch clamp recordings conducted 18-hours after the last SIP training showed that chronic food restriction changed low frequency stimulation (LFS) - induced long-term potentiation of ovBNST inhibitory synaptic transmission (iLTP) into LFS - induced long-term depression (iLTD) in a majority of neurons, regardless of drinking behaviours. However, ad libitum access to food between the last day of SIP training and brain slice recordings (18-hour refeed) rescued LFS-induced iLTP in LD but not in HD, suggesting that impaired bi-directional plasticity of ovBNST synapses may contribute to compulsive drinking in the SIP paradigm.