As for tumorigenesis, genetic alteration endows neoantigens to be expressed and allows recognition under immune surveillance. The important intermediate step requires normal function of TAP1 to get antigens well presented for CD8+ CTL [29]. Within the elaborate immune regulation, TAP1 acts as an important factor vulnerable to be hijacked by tumor cells as strategies to evade immune response for survival and progression. The altered TAP1 expression in tumor tissues, as well as the pivotal function in immune responses, probably indicate its potential role in immune therapy. In this study, we performed a systemic bioinformatic analysis on TAP1, excavated its potential in predicting the clinical prognosis and effect of immunotherapy.
Based on data mining, the plotted transcriptomic data of TAP1 was evident enough that almost all pan-cancer types revealed an elevated RNA level of TAP1 in tumor tissues (except for ACC, KICH and UCS). Our Western blot in GBM samples kept consistent with the RNA-seq results in protein level. However, previous studies had demonstrated a down-regulation of TAP1 in both mRNA and protein level, which is controversial to our results [10–15]. Thus, it is reasonable to speculate the alteration in genomics that counteracts the increased amount of TAP1. In the genetic alteration analysis, a maximum frequency of 8% in TAP1 gene alteration occurred in the tested pan-cancer cohort, and the mutation types were non-specific, accounting little for the cancer development. In addition, an alteration in transcriptomics may result in changes in protein. Sometimes, disability of a protein is reflected in the distribution that determines a different function. In present study, immunofluorescence images of melanoma cells and normal epithelial cells revealed a strict distribution of TAP1 on ER. Thus, the elevated expression of TAP1 was neither interpreted by specific types of genetic alteration or alteration in protein distribution. The reason why TAP1 was aberrantly expressed in tumor tissue remains to be explored.
Generally, tumor tissue is composed of parenchyma and mesenchyme, more than just malignant cells but also resident stromal cells and infiltrated immune cells. The derived TAP1 expression level in tumor tissue is actually a summation of individual cells. Thus, a single-cell expression analysis was performed across samples from pan-cancer cohorts. In these tumor tissues, TAP1 expression was highly concentrated in various immune cells, especially the adaptive immune members like CD4+ and CD8+ T lymphocytes, followed by innate immune cells like monocyte/macrophage and DC. There was also scattered TAP1 expression among all the candidate tumor cell lineages, though, with less amount. The results, to some degree, may account for the contradictory opinion between the Big Data and individual studies. As the experiment conducted before were based on the cellular level, while our results were derived from a tissue-based analysis, without a separation of tumor cells from adjacent mesenchyme. Thus, the expression value of all non-malignant cells was counted and might cause an excess. Anyway, these detailed information of TAP1 expression atlas helps enrich the briefing of TAP1, which is set as the basic for our further investigation.
In another side, we also focused on clinical significance of TAP1, and hopefully it would enlighten a practical application. TAP1 expression data was allocated to bivariate and continuous variable and analyzed for correlation with cancer prognosis, using Kaplan-Meier and univariate Cox regression methods, respectively. Here, clinical prognostic outcomes were presented as four indexes: OS, DFI, DSS and PFI, each of which is characterized by a specific endpoint and available to reflect prognosis in different conditions. Risky and protective indicators were primarily confirmed, suggesting a distinct effect of TAP1 in each cancer. In the forest plot of univariate Cox regression using OS data, the prognostic role of TAP1 in association with survival probability was varied in all the 32 cancer types. Results obtained in terms of OS showed TAP1 was a risk factor for 11 cancer types and protective for 9 types. For BLCA, BRCA, KICH, KIRC, LIHC, SKCM, STAD and THCA, they showed positive correlation with TAP1 expression and admitted it a protective role. While results of most cancer types (15 in 32) like HNSC, LGG and UVM demonstrated TAP1 as a net risk factor. Specifically, Kaplan-Meier survival curve in LGG suggested a protective effect of high-TAP1 expression in terms of OS, while the clinical outcomes were opposite in BLCA, HNSC and SKCM. Thus, TAP1 could be a promising and powerful prognostic biomarker for various cancers.
With such a significant result, we wondered what functionating processes TAP1 may involve in. By taking the advantage of GSEA, we set emphasis on the enrichment of TAP1 in hallmarks gene sets, and our results showed a prominent enrichment in immune-related pathways. Here, based on TAP1 expression, the enriched pathways exhibited a consistent correlation across pan-cancer cohorts. TNF-α signaling pathway, IFN-γ response, IFN-α response, inflammatory response, IL6-JAK-STAT3 signaling, IL2-STAT5 signaling and allograft rejection with positive NES and little FDR were of great significance. According to previous study, IFN as well as TNF molecules promoted in-vivo MHC-I expression by inducing the transcription activity [30, 31]. Besides, the inductive role of IFN-γ, IFN-α/β was more evident in TAP1, and IFN-γ is capable to facilitate TAP-dependent peptide transport [32, 33]. Although MHC-I molecule and TAP as components ubiquitously expressed in all nucleated cells of distinct levels, they are mainly expressed in the site of inflammation in a short time after recognition and warning by immune system [34]. Our GSEA results just conformed to the proposed opinion in published papers. A strong correlation was observed between TAP1 expression and pathways of interest. Allograft rejection, an immune rejection response against grafts from the same species, is a typical inflammatory response of different severity [35]. The most common form is acute rejection that is mainly triggered by T-cell mediated immune responses [35]. For interleukin-mediated signaling pathway, IL6 and IL2 as well-known inflammatory factors also involve regulation of tumor immunity by facilitating the growth and function of lymphocytes [36, 37]. In all, the surprising results point at immune-related mechanism, which encourages us to further explore the potential of TAP1 to predict the responses to immunotherapy.
Actually, development and progression of tumor rely on adjacent environment, TME, which comprises of a complexity of non-malignant cell types (immune cells, fibroblasts, endothelia) and extracellular components (cytokines, hormones) [38]. Although the composition of TME for each cancer is diverse, some common features applied to all types. For instance, vascular network in most tumor is relative leaky and disorganized, allowing infiltration of multiple immune cells for tumor immunity [39]. Considering the results that TAP1 was enriched in immune-related pathways, an immune cell infiltration analysis was conducted to ascertain the association between TAP1 expression and infiltrated immune cells in TME. Scanning our results, an elaborate infiltration pattern was portrayed. It can be indicated that TAP1 expression was positively correlated with multiple immune cells, especially the CD8+ T cells, DC and macrophages. Interestingly, the results are in conformity with single-cell analysis, thus verifying mutually. The cell types highlighted in both analyses were CD8+ T cells and monocyte/macrophages, the killer cells in immune system. Within so many cell types from TIMER 2.0, M2 macrophage showed an opposite association. Probably, the distinct manifestation may result from its stimulative property in anti-inflammation, T helper 2 cell activation (assisting humoral immunity) and immunoregulation, which are contradictory to our proposal about TAP1-associated cell-mediated immunity [38]. Additionally, the IL2-STAT5 signaling pathway, inflammatory response and complement activity highlighted in GSEA are realized by macrophages and CD8+ CTL. Combining the currently available results, it is concluded that TAP1 expression is highly correlated with immune regulation, and it is corresponding to distinct immune signature for each pan-cancer type. Even though there are abundant immune cells infiltrated for tumor immunity, the relationship between TME and immune cells is quite complicated. T cells mediated tumor immunity is either pro-tumorous or anti-tumorous depending on the cells and regulators they encountered in the process of immune responses [39]. In our study, 47 ICPs were recruited to be tested for their correlation with TAP1 mRNA expression across pan-cancer. Tumor cells adopt strategies to activate the suppressive ICP pathways, thus silencing the effector lymphocytes and evading immune surveillance [40]. Through the heatmap, TAP1 expression was positively correlated with most ICPs in majority of pan-cancer types, especially the BRCA, KIRC, PRAD, TGCT, THCA and UVM. TMB and MSI are reported biomarker to predict TME condition and anti-tumor efficacy of ICI therapy [41]. A Spearman’ method was also conducted to test the correlation between TMB, MSI and TAP1 expression. The analyzed results highlighted specific cancer with significance association. For instance, COAD, KIRC and LUAD were correlated with TAP1 expression both in TMB and MSI analysis. Hence, our results may support the availability of TAP1 to predict the responses of immunotherapy that targets immune regulatory process.
Based on distinct TAP1 expression level, precise therapy targeting tumor immunity shows a promising future for cancer patients. The anti-tumor immunity is regulated by a complex of factors in TME, including the ICP, TMB, MSI we discussed already, and then responds with different immune outcomes [40]. PD-1 and PD-L1 as well as CTLA-4 are the well-known immunosuppressive ICPs that determine the suppression of immune responses, usually recruited by tumor cells for immune evasion [42, 43]. Up to now, monoclonal antibodies with high selectivity against PD-1 and CTLA-4 are approved and widely used in the clinical market. However, expected responses are only observed in a portion of patients. As the novel ICI therapy becomes popular, whether it will trigger a favorable response for certain individual remains a problem. In the case that TAP1 expression is highly correlated with immunotherapeutic biomarkers, it is reasonable to expect the feasibility of immunotherapy for patients who bore significant correlation with TAP1 expression. Cohort information of clinical outcomes and transcriptomic profiles from the patients receiving immune therapy was collected and analyzed. The obtained results may guide the therapeutic scheme for the patients waiting for therapy decision. In our study, we cited previous studies where cohorts of patients with primary or metastatic urothelial cancer, breast cancer and melanoma were treated with single or combined monoclonal antibody against PD-L1, PD-1 and CTLA-4, and all the clinical outcomes suggested a protective role of TAP1[24–27]. In our study, BLCA, BRCA and SKCM all exhibited a better prognosis in high TAP1 expression group, in line with the cohort search. However, TAP1 is not a favorable factor for immune therapy responses in all the cancer types. Just as what we concluded in prognostic and ICPs correlation analysis, the correlation of TAP1 expression was diverse among all the cancer types, which may determine diverse predictive role of TAP1. In LGG, high-TAP1 expression displayed a risk effect in tested cohort. Besides, cancer in different stages may bear varied TAP1 expression level and clinical outcomes. In stage 1 and 2 breast cancer, TAP1 expression is reduced, while the trend is reversed in stage 3 and 4, but TAP1 was considered a protective factor in our study [14]. Thus, we proposed TAP1 as a promising and powerful biomarker to predict the effect of immunotherapy for cancer patients. Besides the immunotherapy, previous study had reported a success in increasing tumor-specific immune responses by restoration of TAP1 expression via a TAP1 expressing adenovirus [10]. Surprisingly, novel treatment like this inspires us to foresee the clinical prognosis and make the best treatment option on the basis of specific cancer types as well as individual transcriptomic pattern of biomarkers like TAP1.
Although the present study provides rigorous evidence to demonstrate the predictive role of TAP1 in clinical prognosis and potential responses of immunotherapy across pan-cancer, it still bears limitations. TAP1 is conventionally tumor associated gene, however, whose correlation with prognosis showed diversity in pan-cancer analysis. Although we have proposed the possible explanation, a series of elaborate experiments are still required. What’s more, we have just proposed an essential role of TAP1 as predictor without verifying the clinical use in practical, allowing inaccuracy to occur. Furthermore, our investigation focused on population, whereas the individual difference was neglected. However, clinical therapy protocol is specific to individual, which also determines limitations. In turn, the left issues will indicate research directions for future study, and hopefully bring advantages to those who needs novel treatment for survive.
In conclusion, a systemic pan-cancer analysis with novel design and character is conducted. Our results revealed an aberrant expression of TAP1 in most pan-cancer types, and this expression is significantly correlated with clinical prognosis, immune cell infiltration, expression of ICPs, TME biomarkers and efficacy of immunotherapy. Hence, we propose the TAP1 as a novel biomarker to predict the prognosis and immunotherapeutic responses in different cancer types, opening a new chapter in the exploration of TAP1.