CDK7 and CDK9 inhibition interferes with transcription, translation, and stemness, and induces cytotoxicity in GBM irrespective of temozolomide sensitivity

BACKGROUND

Glioblastoma (GBM) is refractory to current treatment modalities while side effects of treatments result in neurotoxicity and cognitive impairment. Here we test the hypothesis that inhibiting CDK7 or CDK9 would effectively combat GBM with reduced neurotoxicity.

METHODS

We examined the effect of a CDK7 inhibitor, THZ1, and multiple CDK9 inhibitors (SNS032, AZD4573, NVP2, and JSH150) on GBM cell lines, patient-derived temozolomide (TMZ)-resistant and responsive primary tumor cells and glioma stem cells (GSCs). Biochemical changes were assessed by western blotting, immunofluorescence, multispectral imaging, and RT-PCR. In vivo, efficacy was assessed in orthotopic and subcutaneous xenograft models.

RESULTS

CDK7 and CDK9 inhibitors suppressed the viability of TMZ-responsive and resistant GBM cells and GSCs at low nanomolar concentrations, with limited cytotoxic effects in vivo. The inhibitors abrogated RNA Pol II and p70S6K phosphorylation and nascent protein synthesis. Furthermore, the self-renewal of GSCs was significantly reduced with a corresponding reduction in Sox2 and Sox9 levels. Analysis of TCGA data showed increased expression of CDK7, CDK9, SOX2, SOX9, and RPS6KB1 in GBM; supporting this, multispectral imaging of a TMA revealed increased levels of CDK9, Sox2, Sox9, phospho-S6, and phospho-p70S6K in GBM compared to normal brains. RNA-Seq results suggested that inhibitors suppressed tumor-promoting genes while inducing tumor-suppressive genes. Furthermore, the studies conducted on subcutaneous and orthotopic GBM tumor xenograft models showed that administration of CDK9 inhibitors markedly suppressed tumor growth in vivo.

CONCLUSIONS

Our results suggest that CDK7 and CDK9 targeted therapies may be effective against TMZ-sensitive and resistant GBM.

Abstract 1011: Tank binding kinase (TBK1) inhibitor BX795 impairs anti-tumor efficacy of T cells by abrogating granzyme and perforin mobilization

TANK binding kinase 1 (TBK1) is a non-canonical IκB kinase that regulates the innate immune response and is highly expressed in lung, breast, pancreatic and colon cancers. It phosphorylates and activates downstream targets such as IRF3 and c-Rel and mediates NF-κB activation, facilitating the expression of pro-inflammatory genes and interferons. In addition to its role in regulating innate immunity, TBK1 also promotes oncogenesis downstream of K-Ras by phosphorylating Akt and promoting cell survival, proliferation, autophagy and mitophagy. Therefore, TBK1 inhibitors have been proposed to be potentially efficacious against inflammatory diseases and cancer. Unfortunately, TBK1 inhibitors have not been found to be highly efficacious anti-cancer agents and we made efforts to elucidate the underlying reasons.

Earlier findings from our lab revealed a novel role for TBK1 in regulating mitosis. Specifically, we found that active phospho-TBK1 levels were increased during mitosis and TBK1 localized to the centrosomes, mitotic spindles and midbody; selective inhibition or silencing of TBK1 triggered defects in spindle assembly and prevented mitotic progression (Pillai et al., Nature Communications, 2015). TBK1 was found to facilitate the binding of the kinesis Kif2b to CEP170 and dynein to NuMA; depletion or inhibition of TBK1 significantly altered microtubule stability and dynamics, probably due to its regulation of dyenein-NuMA interaction. Since dynein function and appropriate microtubule dynamics is necessary for proper trafficking of vesicles for secretion, hypothesized that inhibition of TBK1 might adversely affect the tumor microenvironment by interfering with the mobilizing of cytotoxic molecules like granzyme and perforin by T cells. Hence we tested if inhibition of TBK1 would impair the anti-tumor activity of T-cells.

Pilot immunofluorescence experiments conducted on JURKAT T-cells treated with the TBK1 inhibitor BX795 showed higher percent of cells retaining granzyme-b and perforin as compared to control cells. Further, we carried out in vitro cytotoxicity assay using normal human peripheral blood mononuclear cells (PBMC's) treated with BX795 (2.5uM for 24 hrs), activated with T cell mitogen concanavalin A (1.25ug/ml) and co-cultured with KRAS mutant A549 lung cancer cells for 72 hrs. Our results showed that BX795 treatment of PBMC's significantly inhibited the ability of the T cells to induce apoptosis in A549 cells, as measured by Annexin-V / PI +ve signal.

These findings suggest that inhibiting TBK1 may be causing dysfunction of immune mediated T cell functions and interference with the anti-tumor immunity. Further experiments are in progress to study the regulatory role of TBK1 on different T cell populations and the associated cytokine signaling.

Citation Format: Fatema Khambati, Jaya Padmanabhan, SriKumar Chellappan. Tank binding kinase (TBK1) inhibitor BX795 impairs anti-tumor efficacy of T cells by abrogating granzyme and perforin mobilization [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1011.

Abstract 2395: A bispecific T- cell engager antibody (BiTE) against TAG-72/ CD3 for targeted cancer immunotherapy

Background: Immunotherapies involving cytotoxic T lymphocytes (CTL’s) have been central to cancer immunotherapy. Bi-specific T cell engager antibody (BiTE) therapy has emerged as an effective immunotherapy by redirecting cytotoxic T cells against cell surface protein on tumor cells. Our BiTE antibody targets human tumor associated glycoprotein 72 (TAG-72). TAG-72 is a pan-carcinoma marker expressed on the surfaces of many cancer tissues including breast, prostate, ovary, endometrium, stomach, esophagus, and pancreas. It is not expressed in normal tissues making it an ideal target therapy for cancer. Our BiTE antibody can bind CD3 to activate the T-cell’s while simultaneously binding TAG-72 on the surface of the cancer cell leading to perforin induced cancer cell death. The in vitro and in vivo effects of TAG-72/ CD3 BiTE antibody on different cancer cells are presented here.

Methods: We studied in vitro and in vivo effects of TAG-72/ CD3 BiTE antibody on different cancer cells; breast (DOX resistant MCF-7, SKBR3, MDA-MB-175 VII), ovarian (OVCAR-3), endometrium (KLE) and colon (Ls174T). In vitro cytotoxicity assay was performed by flow cytometry and immunofluorescence staining using live/ dead cell labelling fluorescent dyes. The cells were co-cultured with naïve human CD8 T cells with/ without BiTE antibody (1 µg, 5 µg and 10 µg) for different time intervals and assayed. In vivo study was carried out in NSG mice implanted with MCF-7 breast cancer cells + naïve human CD8 T cells and treated with a total of 4 doses of BiTE antibody (200µg/ dose).

Results: Different tumor cells showed varying (%) of TAG-72 expression determined by flow cytometry, viz; DOX (55%), SKBR3 (16%), MDA 175 VII (8%), OVCAR-3 (27%), KLE (27%), Ls174T (30%). The BiTE antibody also showed dose dependent binding to these cells. Flow cytometry analysis showed that addition of the BiTE antibody to naïve CD8 T-cells and incubation for 4 days significantly increases tumor cell death as indicated by the Annexin V/ PI+ve signal from DOX (52.4% vs. 41.4% CD8 alone), SKBR-3 (47.9% vs. 30% CD8 alone), MDA 175 VII (23% vs 11% CD8 alone), OVCAR-3 (31.1% vs. 23.8% CD8 alone), KLE (27% vs 13% CD8 alone) and Ls174T (54% vs 29% CD8 alone) respectively (p<0.05). Further, a significant level of IFN-γ secreted by activated T-cells was only detected in the bispecific T-cell engager Ab + CD8 conditions. In vivo data showed good tumor regression at day 45 with tumors measuring (8 mm2) in group treated with BiTE antibody as compared to untreated group (31 mm2) (p<0.01).

Conclusion: Our in vitro and in vivo data indicates that the TAG-72/ CD3 BiTE antibody can effectively redirect cytotoxic T cells against different cancer cells expressing TAG-72 glycoprotein and may be effective in the treatment of breast, ovarian, endometrium and colon cancers.

Citation Format: Fatema Khambati, Hatem Soliman. A bispecific T- cell engager antibody (BiTE) against TAG-72/ CD3 for targeted cancer immunotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 2395.

Abstract 4701: Verteporfin inhibits surface PD-L1 expression in triple-negative breast cancer (TNBC) cells

Background: Programmed death ligand 1 (PDL1) is commonly expressed on the surface of many tumor cells, including breast cancer. The activity of tumor-infiltrating lymphocytes (TIL) is inhibited by PDL1. High PD-L1 basal cells (particularly basal B) overexpress genes involved in invasion, motility, and chemoresistance (Soliman et al., 2014). Targeting and blocking PD-L1 may enhance eradication of aggressive breast cancer cells by the immune system. Verteporfin (VP), a photosensitizer used to treat macular degeneration, was found to inhibit PD-L1 expression in a high-throughput screen. Studies demonstrated VP inhibits YAP activation by disrupting YAP-TEAD interactions and preventing YAP induced oncogenic growth (Zhang et al., 2016). Here we demonstrate the inhibitory effect of VP on PD-L1 expression in TNBC cell lines.

Methods: MDA-MB-231 cells co-cultured with human PBMCs were treated with VP (Sigma) in presence of concanavalin A (ConA) and analyzed using flow cytometry to study levels of CD8+IFNg+ cells. TNBC cell lines (MDA-MB-231, BT-20, HCC-1143, Hs-578T) were treated with VP at doses ranging from 1µM-10µM for 24 hrs. The cells were processed for flow cytometry, Western blot and RT-PCR to check PD-L1 in mechanistic studies looking at manipulation of YAP pathway genes.

Results: When MDA-MB-231 cells were co-cultured with normal human PBMCs in the presence of ConA, the CD8+IFNg+ stained cells were reduced compared to PBMC + ConA alone. Interestingly, in the group treated with VP, rescue of CD8+IFNg+ cells was observed. Moreover, MDA-MB-231, BT-20, HCC-1143 and Hs-578T TNBC cells treated with VP showed a significant dose-dependent inhibition of PD-L1 expression by flow cytometry. Western blot analysis also showed complete clearance of PD-L1 protein band with the lowest dose (1µM) used. However, RT-PCR analysis did not show a significant fold change in mRNA levels of PD-L1 in MDA-MB-231 treated cells. Surprisingly, mechanistic studies performed by silencing YAP1, E2F1, and TBK1 in MDA-MB-231, BT-20 and HS-578 T showed a decline in PD-L1 in E2F1 silenced cells, highlighting a plausible role of E2F1-PDL1 signaling axis. However, no change in PD-L1 expression was seen in cells silenced with YAP1 and TBK1. Further, chromatin-immunoprecipitation assay demonstrated E2F1 binding to PD-L1 promoter.

Conclusion: Our data so far demonstrate that verteporfin treatment leads to inhibition of PD-L1 in TNBC cell lines and improvement in CD8+IFNg+ cells, indicating that VP might have potential for treatment approaches. Our study warrants further attention towards understanding the mechanism of action of VP in inhibiting PD-L1 and the role of E2F1 in the process.

Citation Format: Fatema Khambati, Neha Jaiswal, Srikumar Chellappan, Hatem Soliman. Verteporfin inhibits surface PD-L1 expression in triple-negative breast cancer (TNBC) cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 4701.

A phase-1/2 study of adenovirus-p53 transduced dendritic cell vaccine in combination with indoximod in metastatic solid tumors and invasive breast cancer

Background

Indoleamine 2, 3-dioxygenase is an enzyme that causes immunosuppression in tumors. Indoximod inhibits the indoleamine 2, 3-dioxygenase pathway and enhances immunologic responses to dendritic cell (DC) vaccines preclinically. Adenovirus p53 (Ad.p53) is used to generate DC vaccines against p53. A phase-1/2 trial of indoximod with Ad.p53-DC vaccine was conducted.

Materials and Methods

The phase-1 study combined 7 indoximod dose levels with < 6 Ad.p53-DC vaccinations every 2 weeks. Primary endpoints were maximum-tolerated dose in phase 1 and objective response in phase 2. Flow cytometry measured immune responses.

Results

Thirty-nine patients were treated. In combination with Ad.p53-DC vaccine, the maximum-tolerated dose of indoximod was 1600 mg twice daily. Attributable toxicities were grade 1–2. Best response was stable disease in 4 patients. Immunologic responses were detected in 7 out of 23 evaluable patients. Median progression-free survival was 13.3 weeks (95% confidence interval, 12.97–21.85) and median overall survival was 20.71 weeks (95% confidence interval, 25.75–46.15). Nine out of 22 patients (40%) benefitted from chemotherapy after vaccination. Median overall survival in chemotherapy responders was 69.4 weeks (30.1–122.1).

Conclusions

Indoximod 1600 mg twice daily with Ad.p53-DC was well tolerated. There may have been a chemosensitization effect. Future trials should explore combining this treatment with chemotherapy.

Abstract 3144: Detection of HER2-HER3 heterodimers in patient circulating tumor cells using proximity ligation assay

BACKGROUND: The HER2 protein is a key oncogene in approximately 20% of breast cancers, and preferentially forms heterodimers with other members of the HER family (such as HER3) to activate its signaling cascade. Dual HER2 signaling blockade by using both trastuzumab and pertuzumab (a HER2 dimerization inhibiting monoclonal antibody) yields superior clinical benefit in HER2+ patients compared to trastuzumab monotherapy. A phase 1/2 clinical trial is ongoing to assess the safety and efficacy of the combination of gemcitabine with trastuzumab and pertuzumab in metastatic HER2+ pretreated patients. (NCT02139358) One of the correlative experiments is looking at the feasibility of detecting HER2-HER3 heterodimers using a proximity ligation assay on circulating tumor cells isolated from patients on treatment as a possible pharmcodynamic marker.

METHODS: Whole blood was collected in two 10cc Cellsave tubes (Veridex, Cat#: 7900005) at baseline, week 4, 7, 10. Samples were drawn before treatment was administered, to avoid contamination of the samples with cytotoxic agents. The blood was subjected to red cell lysis with 1XRBC lysis buffer (Santa Cruz). The lysed cells were fixed with 4% paraformaldehyde followed by negative selection utilizing CD45+ magnetic beads (EasySep CD45 depletion kit, cat#18259). The CD45 depleted cell suspension was re-fixed with 4% PFA on coated glass slides following cytospin. These cells were then stained for PLA to test for HER2-HER3 using Sigma Aldrich DuoLink PLA kit + DAPI fluorescence mounting medium (DUO92013) with anti-HER2 (PA5, Thermo Scientific, diluted 1:50) and anti-HER3 (MA5-12675, Thermo Scientific diluted 1:50) as primary antibodies plus cytokeratin (CK) and CD45 antibodies. A set of HER2+ and HER2- PLA controls using SKBR3 were included in the analysis. Control samples using SKBR3 cells were used to optimize the method.

RESULTS: Using the spiked SKBR3 HER2+ blood samples it was possible to isolate CK+ CD45- DAPI+ CTCs demonstrating HER2-HER3 heterodimers following optimization of the PLA method. Using this method in the first four patients we successfully collected CTCs in 7/11 samples processed. In those 7 CTC+ samples, 4 contained CTCs with positive HER2-HER3 PLA signals.

CONCLUSIONS: We have demonstrated a feasible methodology for the detection of HER2-HER3 heterodimers on the surface of circulating tumor cells from HER2+ metastatic breast cancer patients. This can provide investigators with a non-invasive method to monitor the pharmacodynamics of pertuzumab therapy on tumor cells. The collection of these samples is ongoing and correlation with benefit from therapy will be analyzed in the phase 2 portion of the study.

Citation Format: Hatem Soliman, Fatema Khambati. Detection of HER2-HER3 heterodimers in patient circulating tumor cells using proximity ligation assay. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 3144.

A phase I study of indoximod in patients with advanced malignancies

PURPOSE

Indoximod is an oral inhibitor of the indoleamine 2,3-dioxygenase pathway, which causes tumor-mediated immunosuppression. Primary endpoints were maximum tolerated dose (MTD) and toxicity for indoximod in patients with advanced solid tumors. Secondary endpoints included response rates, pharmacokinetics, and immune correlates.

EXPERIMENTAL DESIGN

Our 3+3 phase I trial comprised 10 dose levels (200, 300, 400, 600, and 800 mg once/day; 600, 800, 1200, 1600, and 2000 mg twice/day). Inclusion criteria were measurable metastatic solid malignancy, age ≥18 years, and adequate organ/marrow function. Exclusion criteria were chemotherapy ≤ 3 weeks prior, untreated brain metastases, autoimmune disease, or malabsorption.

RESULTS

In 48 patients, MTD was not reached at 2000 mg twice/day. At 200 mg once/day, 3 patients previously treated with checkpoint inhibitors developed hypophysitis. Five patients showed stable disease >6 months. Indoximod plasma AUC and Cmax plateaued above 1200mg. Cmax (~12 μM at 2000 mg twice/day) occurred at 2.9 hours, and half-life was 10.5 hours. C reactive protein (CRP) levels increased across multiple dose levels.

CONCLUSIONS

Indoximod was safe at doses up to 2000 mg orally twice/day. Best response was stable disease >6 months in 5 patients. Induction of hypophysitis, increased tumor antigen autoantibodies and CRP levels were observed.