Blocking PD1/PDL1 Interactions Together with MLN4924 Therapy is a Potential Strategy for Glioma Treatment

Abstract

Objective: MLN4924, a pharmacological inhibitor of cullin neddylation, resulted in glioma cell apoptosis, deregulation of the S-phase of DNA synthesis and thus, offers great potential for the treatment of brain tumours. However, targeting the neddylation pathway with an MLN4924 treatment stabilized the hypoxia-inducible factor 1A (HIF1A), which is one of the main transcriptional enhancers of the immune checkpoint molecule PDL1 (programmid death ligand-1) in cancer cells. The influence of immune checkpoint molecules on glioma progression has recently been discovered; PDL1 overexpression in gliomas corresponds to a significant shortening of patient survival and a decrease of the anti-tumour immune response. We hypothesize that i) PDL1 is up-regulated in gliomas after treatment with MLN4924 and induces T-cell energy; ii) co-utilization of the PD1/PDL1 blockage with MLN4924 therapy may reduce T-cell energy and may engage MLN4924-induced tumour disruption with the immune response.

Methods: PDL1 expression and its immunosuppressive role in gliomas, glioma microenvironments, and after treatments with MLN4924 were assessed by utilizing methods of immunohistochemistry, molecular biology, and biochemistry.

Results: We confirmed PDL1 overexpression in clinical brain tumour samples, PDGx and established glioma cell lines, extracellular media from glioma cells, and CSF (cerebrospinal fluid) samples from tumour-bearing mice. Our primary T-cell based assays verified that the up-regulation of PDL1 in tumour cells protects gliomas from T-cell treatment and reduces T-cell activation. We found that a pharmacological inhibitor of cullin neddylation, MLN4924, exhibited strong cytotoxicity towards PDGx and established glioma cell lines, in vitro, with an IC50's range from 0.2 to 3 uM. However, we observed a significant increase of HIF1A and PDL1 in mRNA and protein levels in all glioma cell lines after treatment with MLN4924. The MLN4924-dependent induction of PDL1 in gliomas resulted in T-cell energy, which was blocked by a blockage of the PD1/PDL1 interaction.

Conclusion: We conclude that i) PDL1 up-regulation in gliomas and the glioma microenvironment is an important chemotherapeutic target; ii) MLN4924 therapy, combined with a blockage of the PD1/PDL1 pathway, should be considered as a potential strategy for glioma treatment.

Keywords: Brain tumour; Cancer therapy; Gliomas; HIF; Immune checkpoint; MLN4924; PDL1.

Figure 1:

PDL1 expression in clinical brain tumor samples, glioma cell lines and CSF samples from the immunocompetent glioma mice model. A) Graph (left) represents PDL1 mRNA expression in brain tumors and melanomas (cBioPortal (http://www.cbioportal.org)). Note that the average PDL1 mRNA level in GBM is similar to the average PDL1 mRNA level in melanomas. The PDL1 values in GBM and low-grade gliomas which exceed the average PDL1 mRNA level in melanomas are highlighted in the red box. Kaplan-Meier plot (right) illustrates survival rates of glioma patients with high and low PDL1 expression levels; 9% versus 24% of 2 year survival for high and low PDL1 expression; the difference is significant, P=0.02 (data has been obtained from the Human Pathology Atlas). B) Immunohistochemical detection of PDL1 in the tissue microarray of normal and brain tumor samples. The images were taken at 40x magnification. C) Western blot illustrates PDL1 and Actin protein levels in control and brain tumor clinical samples. Graph represents PDL1 to Actin ratios for corresponding protein samples; 0.87 ± 0.32 (n=7) versus 0.19 ± 0.1 (n=7) for tumor and normal samples, respectively, the difference is significant, P=0.003. Two samples (marked by light grey color) of patients with hemorrhage have been excluded from the average. D) Western blots illustrate PDL1 protein levels in established and PDGx glioma cell lines (left) and in the extracellular media collected from these cell lines (right). The protein content of the collected media was concentrated six folds before an analysis (see method). The graph (right) provides PDL1 concentrations in CSF samples from control and tumor-bearing mice (see method). The immunocompetent glioma mice model with GL261 cells was utilized for this experiment. E) Images illustrate interactions of tumor neuro spheres from XD456 (a) and from two parental U251-IDH1-R132H cell lines with different PDL1 expression levels (b) with primary T-cells. Note that tumor neuro spheres with high PDL1 levels keep their integrity after 48 hours of interaction with primary T-cells (illustrated in the insert). The graph represents the average percent of neuro spheres after 48 hours of interaction with T-cells, 89 ± 7% (n=4) and 14 ± 5% (n=4) for cell lines with high and low PDL1 levels, respectively. The difference is significant, P=0.0002. Primary T-cells were loaded with calcein, AM (green) before experiments for visualization.

Figure 2:

HIF1A accumulation evokes PDL1 up-regulation in glioma cell lines. A) HIF1A accumulations induced by CoCl2 treatment of U251 and U87 cell lines evoke PDL1 up-regulation in protein and mRNA levels. Western blot illustrates HIF1A and PDL1 levels in nuclear and cytoplasmic fractions, respectively, in control and after treatment with CoCl2 (85uM). Lamin A/C and alpha Tubulin were utilized to verify nuclear and cytoplasmic fractions, respectively. Graphs illustrate normalized PDL1/GAPDH mRNA ratios after cell treatment with CoCl2 (85uM) at different time points. In each experiment, data has been normalized to the corresponding ratios in untreated cells, results are presented as mean ± S.D. B) Transcription initiation complexes encoded by pAC154-dual-dCas9VP160 plasmids, guided by sgRNAs to the HIF1A binding domains in the first intron of the PDL1 gene, evoke PDL1 overexpression. Western blot illustrates PDL1 and Actin protein levels in U251 cells after transfection with plasmids encoding control (scrambled sequence) sgRNA or HIF1A sgRNAs. Note that transcription initiation complexes guided by sgRNAs to the HIF1A binding domains (A), (B), and (A) with (B) simultaneously increased PDL1 expression by 1.5, 3.9 and 1.8 folds, respectively, compared to the control (PDL1 expression in the presence of transcription initiation complexes guided by scrambled sgRNA). In each experiment, PDL1 expression was normalized to the Actin expression.

Figure 3:

MLN4924 treatment induces up-regulation of HIF1A and PDL1 in glioma cell lines. A) The inhibitory dose- response curves for MLN4924 in established, PDGx and PDGX-stem glioma cell lines. The IC50s are 0.3 ± 0.2 uM (n=4), 2.7 ± 1 uM (n=6), 3 ± 2 uM (n=3), 3 ± 1 uM (n=4), 2.9 ± 0.5 uM (n=4), 0.8 ± 0.2 uM (n =4), 0.2 ± 0.1uM (n=4) for LN221, U251, U87, XD456, JX10, XD456-stem, X14P-stem cell lines, respectively, after treatment with MLN4924 for 5 days. B) Western blots illustrate HIF1A and PDL1 protein levels in nuclear and cytoplasmic fractions in the control and after treatment with MLN4924 (1 uM, 5 days). LaminA/C and alpha Tubulin were utilized to verify nuclear and cytoplasmic fractions, respectively. C) The graph illustrates normalized PDL1/18S mRNA ratios after treatment with MLN4924 (1uM, 5 days) for different cell lines. Note the significant enhancement of the PDL1/18 mRNA ratio for all cell lines after MLN4924 treatment: 8 ± 3, 25 ± 5, 5 ± 1, 8 ± 3, 4.5 ± fold increase compared to the corresponding control values for U251, Ln229, U87, XD456, JX6 cell lines, respectively, P<0.05, n=3.

Figure 4:

Glioma cells treated with MLN4924 decrease T-cell proliferation via utilization of PD1/PDL1 signaling pathway. A) The graph illustrates an inhibition of T-cell proliferation after T-cell encounter with U251 cells (1); with U251 cells in the presence of an inhibitor of PD1/PDL1 interaction (2); with U251 cells treated with MLN4924 (1uM, for 4 days) (3); with U251 cells treated with MLN4924 (1uM, for 4 days) and in the presence of an inhibitor of PD1/PDL1 interaction (4). After MLN4924 treatment, glioma cells were washed and placed in the media with/and without an inhibitor of PD1/PDL1 interaction (4 uM). Note, that glioma cells treated with MLN4924 induce a stronger decrease of T-cell proliferation compared to untreated cells (55 ± 8% (n=4) versus 30 ± 8% (n=4), respectively, the difference is significant with P=0.0005). The reduction of T-cell proliferation induced by glioma cells treated with MLN4924 is inhibited in the presence of an inhibitor of PD1/PDL1 interaction (P=0.0009, n=4). B) The graph illustrates an inhibition of T-cell proliferation after T-cell encounter with XD456 cells (1); with UXD456 cells in the presence of an inhibitor of PD1/PDL1 interaction (2); with XD456 cells treated with MLN4924 (1uM, for 4 days) (3); with XD456 cells treated with MLN4924 (1uM, for 4 days) and in the presence of an inhibitor of PD1/PDL1 interaction (4). After MLN4924 treatment, glioma cells were washed and placed in media with/and without an inhibitor of PD1/PDL1 interaction. Note, that glioma cells treated with MLN4924 induce a stronger decrease of T-cell proliferation compared to untreated cells (74 ± 8% (n=4) versus 41 ± 7% (n=4), respectively, the difference is significant with P=0.0003). The reduction of T-cell proliferation induced by glioma cells treated with MLN4924 is inhibited in the presence of an inhibitor of PD1/PDL1 interaction (P=0.007, n=4).