Psy Gangnam Style Reversed

0 views

Skip to first unread message

Shanta Plansinis

unread,

Aug 3, 2024, 4:28:03 PM8/3/24

to mierustelal

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Cancer immunotherapy has emerged as a promising cancer treatment. However, the presence of immune-refractory tumor cells limits its clinical success by blocking amplification of anti-tumor immunity. Previously, we found that immune selection by immunotherapy drives the evolution of tumors toward multi-modal resistant and stem-like phenotypes via transcription induction of AKT co-activator TCL1A by NANOG. Here, we report a crucial role of HSP90A at the crossroads between NANOG-TCL1A axis and multi-aggressive properties of immune-edited tumor cells by identifying HSP90AA1 as a NANOG transcriptional target. Furthermore, we demonstrate that HSP90A potentiates AKT activation through TCL1A-stabilization, thereby contributing to the multi-aggressive properties in NANOGhigh tumor cells. Importantly, HSP90 inhibition sensitized immune-refractory tumor to adoptive T cell transfer as well as PD-1 blockade, and re-invigorated the immune cycle of tumor-reactive T cells. Our findings implicate that the HSP90A-TCL1A-AKT pathway ignited by NANOG is a central molecular axis and a potential target for immune-refractory tumor.

Here, we report that HSP90A is a clinically actionable target for NANOG-mediated multi-aggressive properties of immune-edited tumor cells. Mechanistically, transcriptional induction of HSP90AA1 by NANOG leads to stabilization of TCL1A, which contributes to subsequent activation of the AKT-signaling pathway. Furthermore, we demonstrate that HSP90A inhibition with AUY-922 renders tumor susceptible to T cell-based immunotherapy including ACT and anti-PD-1 therapy, and leads to increase of the infiltration of tumor-reactive T cells via amplification of anti-tumor immunity. Thus, we provide proof of principle in a preclinical model that the inhibition of HSP90A signaling is an appealing therapeutic strategy to incorporate with various cancer therapeutic modalities, particularly an immune-based modality, and overcome NANOGhigh immune-refractory tumors.

We next attempted to elucidate the underlying mechanism responsible for HSP90AA1 up-regulation in immune-edited tumor. In this regards, we previously demonstrated that NANOG is a key TF driving multi-modal resistance and stem-like phenotype of the immune-refractory tumor25. Therefore, we hypothesized that NANOG might be responsible for transcriptional activation of HSP90AA1 gene. Indeed, silencing of NANOG in P3 cells resulted in decrease of HSP90A protein, which was accompanied by decreased HSP90AA1 mRNA expression (Fig. 2a, b). Conversely, introduction of NANOG into P0 cells raised HSP90A protein level and HSP90AA1 mRNA expression level (Fig. 2c, d). Notably, NANOG WT profoundly increased levels of HSP90AA1 mRNA and HSP90A protein, while NANOG MUT which was previously characterized for its weak transcriptional activity16, had no significant impact on both of HSP90AA1 mRNA and HSP90A protein levels, indicating that NANOG regulates HSP90AA1 expression through its transcriptional function (Fig. 2e, f). To further elucidate the underlying mechanism by which NANOG regulates HSP90AA1 transcription, we identified the HSP90AA1 promoter region containing a putative NANOG-binding site, suggesting the possibility that NANOG is a direct transcriptional activator of HSP90AA1 (Fig. 2g). Luciferase assays showed a significant increase in HSP90AA1 promoter activity upon co-transfection with NANOG WT but not upon co-transfection with NANOG MUT (Fig. 2h). Moreover, mutation of the NANOG-binding site in the HSP90AA1 promoter region eliminated the promoter activation by NANOG WT (Fig. 2h). Chromatin immunoprecipitation (ChIP) assays confirmed the direct binding of NANOG to the regulatory region of HSP90AA1 gene (Fig. 2i), and also validated in the P0 and P3 cells, where we noted more NANOG occupancy in P3 cells, relative to P0 cells (Fig. 2j). Altogether, these findings demonstrate that NANOG up-regulates HSP90AA1 transcription by directly binding to its promoter region.

Previously, we demonstrated that NANOG promotes tumorigenicity and immune resistance of tumor cells through AKT-dependent up-regulation of Cyclin A and MCL-1 (ref. 16). Despite NANOG overexpression, knockdown of HSP90AA1 robustly dampened levels of pAKT, Cyclin A, and MCL-1 (Fig. 4a). Consistently, silencing of HSP90AA1 in P3 cells but not in P0 cells (Supplementary Fig. 12a) and NANOG up-regulated human cancer cells (Supplementary Fig. 12b) markedly reduced the levels of them. Importantly, overexpression of constitutively active AKT (CA-AKT) decreased the sensitivity of siHSP90AA1-transfected CaSki-NANOG cells to CTLs and cisplatin (Supplementary Fig. 13), suggesting that HSP90A plays a crucial role in NANOG-mediated phenotypes by reinforcing the link between NANOG and AKT-signaling pathway.

We next aimed to elucidate the role of HSP90A in NANOG-induced activation of AKT signaling. In this regard, we previously demonstrated that NANOG hyper-activates the AKT signaling through transcriptional up-regulation of TCL1A, a co-activator of AKT kinase27. Given the role of HSP90A as a chaperone in protein stabilization, we questioned whether HSP90A affects protein levels of AKT or TCL1A. The down-regulation of pAKT level upon HSP90AA1 knockdown, however, was not accompanied with loss of AKT protein (Fig. 4a and Supplementary Fig. 12). In contrast, knockdown of HSP90AA1 in CaSki-NANOG cells significantly decreased the levels of TCL1A protein (Fig. 4a). To assess whether HSP90A affects TCL1A stability, we measured the half-life of TCL1A protein upon HSP90AA1 knockdown by performing a cycloheximide-chase assay. Notably, the half-life of endogenous TCL1A protein in siHSP90AA1-transfected CaSki-NANOG cells decreased, compared with those of the siGFP-transfected cells (Fig. 4b). In addition, treatment of MG132 blocked TCL1A protein down-regulation upon HSP90AA1 knockdown (Fig. 4c), indicating the proteasome-mediated degradation of TCL1A protein. Since poly-ubiquitination is required for proteasome-dependent protein degradation, we examined whether HSP90A affects ubiquitination of TCL1A protein. As shown in Fig. 4d, ubiquitination of TCL1A was significantly increased by HSP90AA1 knockdown. Notably, we noted that decreased pAKT level and increased susceptibility to CTLs and cisplatin after HSP90AA1 knockdown in CaSki-NANOG cells were reversed upon restoration of TCL1A expression (Fig. 4e and Supplementary Fig. 14). These results suggest that HSP90A-mediated TCL1A stabilization is important for NANOG-induced AKT activation as well as multi-modal resistance.

We then assessed the effect of HSP90AA1 overexpression on CaSki P0 cells. Overexpression of HSP90AA1 in P0 cells increased TCL1A stability and pAKT level (Supplementary Fig. 15a and b). Importantly, HSP90AA1-transfected P0 cells were more resistant to apoptosis induced by CTLs, compared to empty vector-transfected P0 cells (Supplementary Fig. 15c), indicating that HSP90AA1 expression by itself is sufficient to confer resistance to CTLs. Taken together, our data indicate that HSP90A leads to activation of AKT signaling through TCL1A stabilization, and thus contributes to multi-modal resistance of tumor cells.

Given our observations in vitro, we reasoned that in vivo administration of AUY-922 should reverse resistance to T cell-base immunotherapy. To test this possibility, we treated MART1+ MDA-MB231 P3-bearing NOD-SCID mice with MART-1-specific CTLs along with AUY-922 (Fig. 6a). While immunotherapy alone had no effect on tumor growth, dual therapy with E7-specific CTLs and AUY-922 retarded tumor growth (Fig. 6b, c) and prolonged survival of the mice (Fig. 6d). Consistent with our in vitro results, we observed reduced levels of TCL-1, pAKT, MCL-1, and Cyclin A in tumor tissue from AUY922-treated mice compared with that from un-treated mice (Fig. 6e). In addition, measurements by Ki67 staining demonstrated that AUY-922-treated tumors contained fewer proliferating cells than PBS-treated tumors, as this was unaffected by the adoptive transfer of CTL (Fig. 6f). Although we observed a slight, but not statistically significant decrease in the frequency of antigen-specific CTLs in the tumors of AUY-922-treated mice compared with those in PBS-treated mice, the overall cytotoxic effect of these CTLs was greater after treatment of AUY-922 relative to that than PBS control, as indicated by the percentage of apoptotic tumor cells (Fig. 6g, h). Taken together, we conclude that inhibition of HSP90A can incapacitate the immune resistance of immune-edited tumor cells and represents an attractive strategy for the control of human cancer, as a synergistically, as part of a T cell-mediated immunotherapy.

It has been reported that multi-gene signature is associated with clinical efficacy of PD-1 blockade30. To gain insights into NANOG as a TF responsible for response to anti-PD-1 therapy, we performed TF analysis using up-regulated or down-regulated DEGs in NR relative to R to anti-PD-1 therapy. From this analysis, we noted that many of the up-regulated DEGs in NR were directly linked to NANOG (Fig. 7c). On the other hand, genes regulated by IRF8, a TF in the interferon gamma (IFN-γ)-signaling pathway, were down-regulated in NR to anti-PD-1 therapy (Supplementary Fig. 17a and b). This finding is consistent with recent studies demonstrating that an IFN-γ signature was found to be differentially down-expressed in the pretreatment tumor biopsies from non-responding patients31,32,33.

To validate preliminary NANOG-responsive genes from this analysis, we further performed a complementary analysis followed by filtering to include only genes that have been previously identified as responding to NANOG34,35,36,37,38. Subsequently, we acquired the refined NANOG-responsive six genes (FGFBP1, SPESP1, HMGA2, PERP, FGF1, and DKK1). To quantify the expression of NANOG-responsive genes, we calculated the average expression of each gene. Notably, the expression level of the NANOG-responsive genes was significantly higher in NR compared with R to anti-PD-1 therapy (Fig. 7d). Furthermore, we also found a strong correlation between NANOG-responsive genes and HSP90AA1 expression in NR (Fig. 7e, f).