Expression levels analysis and high NAPSB inferred a better prognosis for HCC
NAPSB transcription levels in different human tumors were showed in Fig. 1a. Compared with adjacent normal tissues, expression of NAPSB in BLCA (bladder urothelial carcinoma), COAD (colon adenocarcinoma), LIHC (liver hepatocellular carcinoma), LUAD (lung adenocarcinoma), LUSC (lung squamous cell carcinoma) and READ (rectal adenocarcinoma) was significantly decreased but in BRCA (breast invasive carcinoma), ESCA (esophageal carcinoma), KICH (kidney chromosome) and KIRP (kidney renal papillary cell carcinoma) was significantly increased. For TCGA cohort, we analyzed paired samples by paired Student’s t-test to verify the above results in HCC (Fig. 1b). To fully demonstrate this expression difference, we validated it with multiple datasets, including ICGC, GSE55092, GSE54236, and GSE121248, finding that NAPSB was indeed significantly decreased in HCC tissues (Fig. 1c). Moreover, NAPSB expression was examined in 13 paired HCC and adjacent normal tissues of Zhongnan cohort from our hospital by RT-PCR and we obtained consistent results. (Fig. 1d).
The correlation between NAPSB and clinicopathologic characteristics for TCGA and ICGC cohorts were presented in Supplementary Table 5 and 6. In addition, we investigated its survival value. K-M survival analysis showed that high expression of NAPSB was linked to better overall survival than its low expression (Fig. 1e) and more significantly associated with better disease-free survival (Fig. 1f) and progression-free survival (Fig. 1g). Its overall survival value was also verified in the ICGC cohort (Fig. 1h). Given that there are multiple factors affecting survival, univariate Cox regression analysis was performed using age, tumor stages, gender, and NAPSB expression as inputs; results showed that NAPSB expression were significantly associated with better DFI and PFI outcomes (Supplementary Fig. 1a) and multivariate Cox regression analysis further validated it (Supplementary Fig. 1b). Therefore, NAPSB expression was beneficial to overall survival and could serve as an independent predictor of disease-free survival and progression-free survival in HCC.
Enrichment analyses inferred NAPSB was related to immune activation
To investigate the potential biological function of NAPSB, we firstly performed a correlation analysis between NAPSB and other genes using RNA-seq data of HCC patients from the TCGA cohort. The results showed that there were 930 genes significantly associated with NAPSB (p-value < 0.01, |Spearman`s correlation| ≥ 0.45; Supplementary Table 7). The correlation of NAPSB and the top 50 co-expressed genes was showed in Fig. 2a, which contained some immune-related molecules like CD48, CD37, IL6, HLA-DQA1, etc. Meanwhile, we analyzed the differentially expressed genes (DEGs) between the NAPSB high and low expression groups and found that there were 993 upregulated in the NAPSB-high group compared with the NAPSB-low group (adjusted p-value < 0.05 and | log2 (fold change) | value ≥ 1.3; Supplementary Table 8). The top 10 upregulated genes contained immune-related molecules, such as CD48, CD37, CCR5 (Fig. 2b). Both co-expressed genes and upregulated DEGs contained well-known molecules associated with immunity, suggesting that NAPSB may also be involved in immunity.
Thereafter, the intersection of co-expressed genes and upregulated DEGs including 476 common genes were obtained as the most closely related genes to NAPSB (Fig. 2c; Supplementary Table 9) and enrichment analyses were conducted in an attempt to reflect biological functions of NAPSB indirectly. The GO analysis showed that the common genes were enriched in processes such as T cell activation, leukocyte cell-cell adhesion, regulation of T cell activation, and regulation of immune effector process, etc (Fig. 2d; Supplementary Table 10). The KEGG pathway analysis demonstrated that the common genes were associated with chemokine signaling pathway, Th17 cell differentiation and T cell receptor signaling pathway (Fig. 2e; Supplementary Table 11). Most biological functions and signaling pathways were immune-related, strongly implying that NAPSB may mediate the TME in HCC.
Even further, we conducted GSEA and GSVA between NAPSB high- and low- groups and also identified many significant pathways related to immunity (Fig. 2f, g; Supplementary Table 12, 13) in the enrichment of MSigDB Collection (gene sets of “c5.cp.v7.4.symbols.gmt” and “h.all.v7.4.symbols”). These findings paralleled the above results.
NAPSB shaped an immuno-hot and inflamed TME in HCC
Given that the enrichment analyses indicated NAPSB was associated with immune activation in HCC, we subsequently comprehensively explored its immunological role using TCGA and ICGC cohorts. NAPSB was found to upregulated the expression of critical immunomodulators (including MHC, immunostimulator, chemokine, and receptor) (Fig. 3a), which may upregulate the activities of the cancer–immunity cycle subsequently. Additionally, ESTIMATE algorithm was applied to calculate the immune score, stromal score, estimated score and tumor purity. We found these scores were significantly increased NAPSB-high group (Fig. 3b). On the contrary, tumor purity was negatively correlated with the expression of NAPSB (Fig. 3c). Subsequently, TME immune cell infiltration was assessed by ssGSEA and a significant difference between NAPSB-high and NAPSB-low groups was observed. Almost all immune cells were significantly enriched in NAPSB-high group (Fig. 3d). Furthermore, we calculated the infiltration level of TIICs using other six independent algorithms (Fig. 3e). Consistent with the result of ssGSEA analysis, the infiltration levels of CD8+ T cells, CD4+ T cells, NK cells, B cells, DCs and macrophages were almost positively correlated with NAPSB in different algorithms. In particular, CD8+ T cells were positively correlated with NAPSB with statistical significance in all algorithms. In line with these, NAPSB was positively correlated with the marker genes of these six major types of immune cells (Fig. 3f). These results suggested NAPSB was associated with an inflamed TME. Even further, we observed the NAPSB expression positively correlated with the T cell inflamed score (TIS) established using IFN-γ-related mRNA profiles [40] and all of genes within this signature (Fig. 3g, h), further confirming its roles in shaping a hot inflamed TME. These findings were all verified in ICGC cohort and obtained consistent results (Supplementary Fig. 2).
Finally, we evaluated the correlation between NAPSB and seven steps of cancer-immunity cycle, which conceptualized the anti-cancer immune response [38]. Overall, In the NAPSB-high group, the activities associated with the majority of the steps in the cycle were notably upregulated (Fig. 3i), including the release of cancer cell antigens (Step 1), priming and activation (Step 3), trafficking of immune cells to tumors (Step 4), and infiltration of immune cells into tumors (Step 5). In summary, these data consistently indicated that high expression of NAPSB was to transform a non-inflamed TME into an immuno-hot and inflamed microenvironment, consequently triggering anti-cancer immune response.
NAPSB highly expressed in hot tumors and may enhance immunotherapy response
As mentioned earlier, solid tumors can be divided into hot tumors and cold tumors, and hot tumors respond to cancer immunotherapy [14, 51]. Unsupervised clustering was conducted to classified HCC samples into hot tumors and cold tumors based on the hot tumor signature genes (Supplementary Table 14; Fig. 4a-d) [41]. Then, the expression of NAPSB was compared between hot and cold tumors and we found that it was overexpressed significantly in hot tumors (Fig. 4e), suggesting that NAPSB could play a role in distinct hot/cold tumor states and be associated with therapeutic response to immunotherapy. The same methods were used to validate above results in the ICGC cohort (Supplementary Fig. 3a-e).
In addition, NAPSB expression was found to be positively correlated with BTLA (B and T lymphocyte associated), CTLA-4 (cytotoxic T-lymphocyte-associated protein 4), IDO1 (indoleamine 2,3-dioxygenase 1), LAG-3 (lymphocyte activating 3), PD-1(programmed cell death 1), PD-L1 (programmed cell death 1 ligand 1), TIGIT (T cell immunoreceptor with Ig and ITIM domains), and TIM-3 (T-cell immunoglobulin and mucin domain-containing protein 3) expression (Fig. 4f), which were well-known predictors of response to immunotherapy. Also, the enrichment scores of therapeutic signatures, predicting clinical response, were compared in NAPSB subgroups. As exhibited in Fig. 4g, i, NAPSB was negatively correlated with the enrichment scores of PPARG network, β-catenin signaling pathway, VEGFA and IDH1, which were all immunosuppressive gene signatures [15-17, 52]. However, in the NAPSB-high group, immunotherapy-positive pathways such as IFN-γ-signature, APM-signal, EGFR-ligands, hypoxia and KDM6B (lysine demethylase 6b) were activated (Fig. 4h) [18-21, 53], indicating immune-activated state and beneficial to immunotherapy response. These observations were also validated using ICGC samples (Supplementary Fig. 3f-h).
The last but important, the role of the NAPSB in predicting the immune checkpoint blockade (ICB) response was explored in tow immunotherapy-related melanoma cohorts. In GSE91061, we found the ICB response rates were obviously higher in the NAPSB-high group than in the NAPSB-low group (Fig. 4j) and the expression of NAPSB was significantly high in response group (Fig. 4k). Similar results were observed in the GSE78220 cohort (Supplementary Fig. 3i). These evidences reconfirmed that NAPSB may be a valuable predictor of immunotherapy response across cancers.
NAPSB was associated with increased sensitivity to chemotherapy
As for whether NAPSB plays a role in chemotherapy sensitivity, the correlation between NAPSB expression and IC50 of multiple drugs in cell lines was analyzed using data from GDSC and CTRP. Intriguingly, NAPSB expression was negatively associated with IC50 of most agents in GDSC and in CTRP (Fig. 5a and Supplementary Fig. 4; Supplementary Table 15), supporting that NAPSB can enhance the therapeutic response to chemotherapy. Two heat maps showed that the IC50 of some commonly used drugs is lower in the NAPSB-high group in GDSC and CTRP databases, respectively (Fig. 5b, c). Results above speculated that high expression of NAPSB is beneficial to the sensitive response of chemotherapy.
Thereafter, by analyzing GSE104580, a HCC cohort of transarterial chemoembolization (TACE), we found the expression of NAPSB was significantly higher in TACE response group (Fig. 5d) and the response rates were obviously higher in the NAPSB-high group than in the NAPSB-low group (Fig. 5e). This data further illustrated that high expression of NAPSB may be beneficial to chemotherapy response.
Association of NAPSB with cell death of tumor cells
Given that cell death had been reported in recent years to play a significant role in tumor therapy [54], we investigated the association between NAPSB and various forms of cell death, including pyroptosis, necroptosis, apoptosis, autophagy and ferroptosis, using ssGSEA. As showed in Fig. 6a-c, and E, NAPSB expression was markedly correlated with pyroptosis, apoptosis and necroptosis, but negatively correlated with ferroptosis. Autophagy had no correlation with NAPSB expression (Fig. 6d). We also verified these discoveries with ICGC cohort. In line with these findings, the correlations between NAPSB and several cell death forms were consistent with that in the ICGC cohort (Fig. 6f). In addition, the enrichment scores of pyroptosis, apoptosis and necroptosis in NAPSB-high groups were markedly higher than NAPSB-low groups (Fig. 6g). However, similar to results of TCGA, ferroptosis scores were lower in NAPSB-high group than in NAPSB-low group and autophagy scores had no significant differences in NAPSB subgroups. Among the above results, the correlation between NAPSB and pyroptosis was the most significant. Results above inferred that NAPSB may have a beneficial effect on immunotherapy and chemotherapy responses by promoting pyroptosis, necroptosis and apoptosis (PANoptosis) in tumor therapy.