APPL1 Is a Prognostic Biomarker and Correlated with Treg Cell Infiltration via Oxygen-Consuming Metabolism in Renal Clear Cell Carcinoma

Kidney renal clear cell carcinoma (KIRC) is one of the most hazardous tumors in the urinary system. The regulation of oxygen consumption in renal clear cell carcinoma is a consequence of adaptive reprogramming of oxidative metabolism in tumor cells. APPL1 is a signaling adaptor involved in cell survival, oxidative stress, inflammation, and energy metabolism. However, the correlation of APPL1 with regulatory T cell (Treg) infiltration and prognostic value in KIRC remain unclear. In this study, we comprehensively predicted the potential function and prognostic value of APPL1 in KIRC. For KIRC patients, relatively low expression of APPL1 was associated with high degree of metastasis, pathological stage, and shorter overall time or poor prognosis. Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses suggested that low expression of APPL1 may be adapted to the malignant progression of tumors via affecting oxygen-consuming metabolism. In addition, the expression level of APPL1 was negatively correlated with Treg cell infiltration and chemotherapy sensitivity, which indicated that APPL1 may regulate the tumor immune infiltration and chemotherapy resistance by decrease oxygen-consuming metabolic process in KIRC. Therefore, APPL1 may become one of the important prognostic factors, and it may serve as a candidate prognostic biomarker in KIRC.


Introduction
Renal cell carcinoma (RCC) is a lethal cancer of the urinary system, which is second only to prostate and bladder cancer in mortality worldwide [1][2][3].Kidney renal clear cell carcinoma (KIRC) is a predominant heterogeneous histological subtype (~70%) of kidney cancer and is one of the most malignant diseases [4,5].Tumor metastases occur in approximately 30% in diagnosed KIRC patients, and it has a 40% of recurrence rate after surgery [6,7].Currently, patients with KIRC could obtain a favorable prognosis after chemotherapy, radiotherapy, and especially immunotherapy and chemotherapy, but chemotherapy resistance is still a very challenging issue [8][9][10].The poor prognosis of KIRC may be related to the reprogramming of energy metabolism in tumor cells, such as the downregulation of the oxygenconsuming tricarboxylic acid cycle (TAC) and damage caused by oxidative stress [11].The immune system can inhibit tumor cell growth, while some T cells from tumor microenvironment can also promote tumor malignant progression via promoting immunosuppressive state [12].Numerous studies have shown that CD4 + T cells, CD8 + T cells, etc. played dual roles in KIRC: functioning as antitumor or leading to a poor prognosis [12][13][14].Therefore, there is an urgent need to clarify the immunophenotype of tumorimmune interactions and related therapeutic targets for improving the prognosis of KIRC therapy.
Adaptor protein containing a pH domain, a PTB domain, and a leucine zipper motif 1 (APPL1) is a mediator protein that associates with regulation of cellular energy metabolism, survival, oxidative stress, and inflammation regulation [15][16][17].In addition, the APPL1/AMPK or APPL1/Nrf2 pathways promote the uptake and metabolism of glucose and fatty acids for anti-inflammatory and antioxidant responses in normal cells [18,19].Thus, expression level of APPL1 may be correlated with change of tumor microenvironment in KIRC.However, the mechanism and poor prognosis of APPL1 expression on Treg cell infiltration in KIRC remain unclear.
In this study, we integrated multiple bioinformatics methods based on database online to explore whether APPL1 affects tumor Treg infiltration and try to find its correlation with poor prognosis in KIRC patients.We found that the expression of APPL1 was significantly downregulated in KIRC tumor tissues compared with adjacent normal tissues.Besides, the expression level of APPL1 was correlated with survival rate and Treg infiltration and immune checkpoints in KIRC.Our further analysis revealed that APPL1 and related genes may affect the oxygen-consuming metabolic process, which contribute to chemotherapy resistance of KIRC patients.These results highlight that APPL1 may serve as a biomarker in predicting the prognosis of KIRC.
The data on KIRC tumor protein expression levels were obtained from Clinical Proteomic Tumor Analysis Consortium (CPTAC).The difference of APPL1 mRNA and protein level for tumor tissue and adjacent normal tissue in KIRC was analyzed via UALCAN database online (http:// Ualcan.path.uab.edu/analysis).We also investigate the various clinicopathological parameters for KIRC (gender, tumor stage, lymph node metastasis status, etc.) in UALCAN.

Cell Migration and Proliferation
Assays.The culture conditions of 786O, 769P, and Caki-1 cells were DMEM-F12 (90%, Gibcol) + FBS (10%, Gibcol) + penicillin and streptomycin (1%, Biosharp).Scratch wound and Transwell assay were used in cell migration experiments.The scratch wound assay was photographed at 0 h and 24 h, respectively; Transwell was fixed with 4% paraformaldehyde and stained with crystal violet at 24 h.The proliferation assay of cells in each group was conducted by adding CCK8 (DOJINDO) to determine the OD450 absorption value from 1 to 5 days.All experiments were independently repeated three times.

Kaplan-Meier
Plotter and ROC Analysis.RNA-sequencing expression (level 3) profiles and related clinical information for KIRC were from TCGA dataset (https://portal.gdc.com).Log-rank test was used to compare differences in overall survival (OS), progression-free interval (PFI), and disease specific survival (DSS) between these groups, and it compared the predictive accuracy of APPL1 mRNA by the time ROC (v 0.4).For the Kaplan-Meier curves, p values and hazard ratio (HR) with 95% confidence interval (CI) were generated by log-rank tests and univariate Cox proportional hazards regression.

Prognostic Model (Nomogram).
Univariate and multivariate cox regression analyses were performed to identify the proper terms to construct the nomogram.The forest plot was used to display the p value, HR, and 95% CI of each variable via R package (forest plot).The nomogram was constructed based on the results of multivariate cox proportional hazards analysis to predict the 1-, 3-, 5-year overall recurrence.The nomogram provided a graphical representation of the factors, which can be used to calculate the risk of recurrence for an individual patient by R package (rms).

Correlation Analysis.
Immune correlation analysis of APPL1 expression and the two-gene correlation map in KIRC were conducted by the R software gg stats plot package.We used Spearman's correlation analysis to describe the correlation between quantitative variables without a normal distribution.All the analysis methods and R package were implemented by R version ( * p < 0:05, * * p < 0:01, and * * * p < 0:001).TISIDB is a web portal for tumor and immune system interaction, which integrates multiple heterogeneous data types.Immune checkpoint analysis of APPL1 expression in KIRC was conducted through TISIDB database online (http://cis.hku.hk/TISIDB/index.php).We used the R software GSVA package to analyze, choosing parameter as method = "ssgsea."The correlation between genes and pathway scores was analyzed via Spearman's correlation.All the analysis methods and R packages were implemented by R version 4.0.3 ( * p < 0:05, * * p < 0:01, and * * * p < 0:001).

Enrichment Analysis of Genes Correlated with APPL1.
Metascape (https://metascape.org) is a tool that integrated the enrichment functions of GO and KEGG [20].The genes positively correlated with APPL1 expression were analyzed, and we obtain the related biological processes and pathways.The p < 0:01 and an enrichment factor > 1:5 indicate significance of the results.GSEA enrichment analysis was conducted by software R (version 3.6.3)and ggplot2 package (version 3.3.3),and the threshold for significant enrichment was false discovery rate ðFDRÞ < 0:25 and p:adjust < 0:05.

Immunohistochemistry (IHC).
Twelve patient samples involved in this study were obtained and used for validation experiment, under the approval by the Ethics Committee of Sichuan Provincial People's Hospital, and informed consent was obtained from the participants.Each tissue sample was fixed in 4% neutral formaldehyde solution for 48 hours, then embedded in paraffin, section, dewaxing, and antigen retrieval, and the rest of steps were completed according to the immunohistochemical protocol [23,24].The images were collected by light microscope (200x, Nikon).Antibodies used in the experiment included rabbit monoclonal antibody APPL1 (ab180140, Abcam) and Donkey anti rabbit HRP (bs-0295D, BIOSS).In addition, another IHC images of clear cell carcinoma were obtained from the Human Protein Atlas online database (https://www.proteinatlas.org/learn/dictionary/cell).To further investigate the variations in APPL1 expression during the progression of KIRC, we analyzed different pathological stages of KIRC patients from the UALCAN database.As expected, the results showed that the expression of APPL1 mRNA was significantly downregulated, with the progression of pathological stage, N classification, and histological grade in KIRC ( * * p < 0:01 and * * * p < 0:001) (Figures 1(e)-1(g)).These results confirmed the correlation between APPL1 and tumor progression.
We further verified the expression of APPL1 in KIRC based on histological and cellular experiments.We found that APPL1 expressed in the cytoplasm was significantly downregulated in tumor tissues relative to normal adjacent tissues ( * * * p < 0:001) (Figures 2(a) and 2(b)).In addition, the expression levels of APPL1 in KIRC tumor cell lines Caki-1 and 769P were significantly lower than that of HK-2 cells (a normal proximal tubular cell line) ( * p < 0:05 and * * p < 0:01) (Figures 2(c) and 2(d)), and this result was consistent with the gene expression distribution from CCLE database (Figure 2(e)).The Caki-1 cells belong to the metastatic cell line form KIRC. Based on the difference of APPL1 protein expression between Caki-1 and 786O cells, we speculate that the metastasis of KIRC tumor cells may be related to the expression level of APPL1.

APPL1 Expression Inhibits Caki-1 Cell Migration and
Proliferation.APPL1 is expressed at low levels in Caki-1 cells, and we evaluated the effect on metastatic renal cancer cells by restoration to overexpression of APPL1.The results of scratch wound assay showed that the wound closure of the group overexpressing APPL1 (oeAPPL1, 12.81%) was significantly lower than control (oeCTL, 33.18%) for Caki-1 cells (Figure 2(f), * * p < 0:01).Meanwhile, cell migration assay showed that the cell migration ability of the group oeAPPL1 was significantly lower than oeCTL ( * * * p < 0:001) Based on these two results, we confirmed that low expression of APPL1 was significantly associated with the migratory ability of Caki-1 cells.In addition, the cell  Oxidative Medicine and Cellular Longevity proliferation activity of the group oeAPPL1 was also significantly lower than oeCTL from day 3 to 5 (Figure 2(h), * p < 0:05 and * * p < 0:01).Altogether, the expression level of APPL1 is associated with the malignant phenotype of KIRC cells and may affect the survival and prognosis of KIRC patients.

Decreased APPL1 Expression Correlates with Poor
Prognosis in KIRC Patients.Since downregulation of APPL1 was correlated with progression of KIRC, we investigated the effects of APPL1 expression level on the prognosis.Through the Kaplan-Meier plotter (K-M) database, we found that the KIRC patients with low expression of APPL1 had a significantly poor prognosis, which included decreased overall survival (OS), progression-free interval (PFI), and disease specific survival (DSS) (Figures 3(a)-3(c)).To evaluate the accuracy of APPL1 as a prognostic factor, we drew receiver operating characteristic (ROC) curves and calculated area under the curve (AUC) values for APPL1 from TCGA database.As shown in Figures 3(d)-3(f), APPL1, as a potential prognosis factor for KIRC, is of reliability in 1, 3, and 5 years, respectively (OS, AUC > 0:6; PFI, AUC > 0:6; and DSS, AUC > 0:6).Subsequently, the gene APPL1, age, gender, and different classification of KIRC patients were included in univariate and multivariate Cox analyses.The results indicated that APPL1, age, and M stage were significantly independent prognostic factors for OS (Figures 3(g) and 3(h)).Moreover, a nomogram and calibration analysis for KIRC patients also supported the above findings (Figures 3(i) and 3(j)).Taken together, these results suggested that APPL1 may be a prognostic factor for KIRC patients.

Correlation of APPL1 Expression with Treg Cell
Infiltration in KIRC.KIRC is a heterogeneous tumor with high immune infiltration, which affects tumor microenvironment [25,26].Among those immune cells, tumorinfiltrating T lymphocytes are closely correlated with prognostic survival [12].Therefore, we explored the correlation of between APPL1 expression level and typical T cell infiltration.The results indicated that APPL1 expression level had significantly negative correlation with regulatory T cells KIRC has a higher T cell infiltration score among 19 cancer types detected [25].To further investigate the mechanism of APPL1 as an important factor in KIRC, T cell populations and APPL1 had been included for correlation analysis via TIMER2.0database.As expected, the results showed that APPL1 was significantly negatively correlated with exhausted T cells ( * p < 0:05 and * * * p < 0:001) (Table 1).In addition, Th2 and effector T cells were found to have significantly positive correlation with APPL1 expression level ( * p < 0:05 and * * * p < 0:001) (Table 1).In summary, the expression level of APPL1 was negatively correlated with Treg cell infiltration.

Functional Enrichment of Genes Positively Correlated with APPL1 in KIRC.
To further explore the mechanism of APPL1 for Treg cell infiltration in KIRC, we found 246 genes shared between the 6389 genes that positively correlated with APPL1 (APPL1-PCGs) and 1300 differentially expressed genes (downregulated/differentially expressed genes (DEGs)) in KIRC from UALCAN and TCGA databases, respectively (Figure 6(a)).These genes had been included in GO and KEGG enrichment analyses.The results showed the metabolic pathways of several different substances, which may be involved in energy metabolic processes (Figure 6(b)).We further chose the 4 pathways related to metabolism for energy (R-HSA-70895, GO: 0032787, R-HSA-556833, and hsa05230) to conduct GO and KEGG enrichment analyses, respectively.
The results of enrichment focus on some oxygen-consuming metabolic processes, such as carboxylic acid catabolic process, tricarboxylic acid cycle enzyme complex, mitochondrial matrix, fatty acid metabolic process, central carbon metabolism in cancer, cellular respiration, and energy derivation by oxidation of organic compounds (Figures 6(c)-6(f)).
In addition, the differentially expressed genes from TCGA and GEO (GSE53757) databases were included in GSEA enrichment analysis, respectively.The results showed that two major processes involved in branched-chain amino acid and fatty acid metabolism were consistent with results of GO/KEGG enrichment, and both showed a significant downregulation trend (FDR < 0:25 and p:adjust < 0:05) (Figures 7(a .Taken together, the low expression of APPL1 and its related genes may be positively correlated with the process of oxygenconsumption metabolism in KIRC.

Correlation between the Oxygen-Consumption Metabolism and Treg/Tcm Infiltration in KIRC.
To confirm the association between oxygen-consumption metabolism-related genes and Treg/Tcm cell infiltration, we selected the pathway "central carbon metabolism in cancer" as a representative verification and analysis.The central carbon metabolism is the main source of energy required by organisms and plays a major role in oxidative energy supply.Therefore, we screened 9 genes by Cox univariate regression analysis (Supplementary Figure 1A), LASSO regression analysis (Supplementary Figure 1B, C), and analysis of risk factor signature (Supplementary Figure 1D) based on 21 genes from central carbon metabolism in cancer.These genes include THRB, ATP1B1, TEK, NNT, MTOR, FGFR2, KL, FDX1, and ACADM, which showed a significant downregulation trend in KIRC (Figure 8(a)).In addition, there were significant differences in survival and prognosis (Figures 8(b) and 8(c), p < 0:05 and AUC > 0:8).
We further assessed the correlation of these 9 genes with immune infiltration used by Treg and Tcm cells, respectively.The results showed that only FDX1 had no significant correlation with Tcm (R = 0:002, p = 0:969), and others had significantly negative and positive correlations with Tregs and Tcm (p < 0:05), respectively (Figure 8(d)).In addition, the immunohistochemical protein levels revealed that 8 proteins were significantly downregulated in renal cancer tissues relative to adjacent tissues from the Human Protein Atlas (THPA) database (Supplementary Figure 2) (the protein KL was not included in the THPA database).Taken together, this part suggested that low expression of 9 genes, which related to oxidative metabolism processes, was associated with immune infiltration and antitumor immunity.

Discussion
Renal cancer is second only to prostate cancer and bladder cancer among the malignant tumors of the urinary system,  [4,27].Immune infiltration and immune evasion are the main causes for poor prognosis in KIRC patients [12].Therefore, it is important to explore the mechanisms that affect immune infiltration or immunosuppression and find reliable tumor prognostic markers [28].APPL1 is an adaptor of adiponectin signaling, and it is also closely correlated with the development and differentiation of lymphocytes, such as T cells and macrophages [29,30].However, the effects of APPL1 expression on tumor Treg cell infiltration and correlation with poor prognosis have not been reported in KIRC.
In this study, we performed bioinformatics analysis based on the TCGA, UALCAN, TIMER, and TISIDB data-bases and found that APPL1 was downregulated in KIRC tumor tissues relative to adjacent normal tissues (Figures 1  and 2(a) and 2(b)).Meanwhile, the results of cellular level in vitro also reflected that the expression of APPL1 was associated with the malignant phenotype of tumor cells (Figures 2(f)-2(h)).This finding could also be observed in other 11 types of tumors, which suggesting that downregulation of APPL1 expression may be associated with the development of some tumors.However, this finding was different from a few studies, which found that APPL1 was highly expressed in cholangiocarcinoma (CHOL), liver hepatocellular carcinoma (LIHC), stomach adenocarcinoma (STAD), and breast cancer.Ding    that the expression of APPL1 was upregulated in breast cancer (MCF-7) and liver cancer cells (HepG2), and they verified that APPL1 could promote the proliferation and migration of tumor cells via leptin-mediated phosphorylation of STAT3, ERK1/2, and AKT, which could explain the relevant mechanisms [31][32][33].The clinical prognosis analysis for KIRC showed that the survival rate was significantly lower in an APPL1-downregulated condition (Figure 3).
Thus, these results indicated that APPL1 may be a prognostic factor for KIRC.Although nephrectomy has always been a common treatment for kidney cancer, 30% of the patients have localization after surgery and then develop into metastases and finally contribute to a high mortality rate [34][35][36].Previous studies have shown that APPL1 can inhibit cell migration via inhibiting activation of AKT pathway in tumor [37,        38].Therefore, downregulation of APPL1 expression may contribute to tumor cell metastasis in KIRC.Moreover, the high aggressiveness of KIRC tumors was mainly associated to the hypoxia induction [39,40], high infiltration of T cells [41,42], Th2 cells, and macrophages from tumor microenvironment [25,[43][44][45].We found that APPL1 expression was significantly and negatively correlated with Treg cells in KIRC (Figure 4).These correlations suggest that immune response of antitumor may decreased with downregulation of APPL1, and it was consistent with suboptimal chemotherapy outcome or poor prognosis in KIRC [46].Meanwhile, immunosuppression is closely related to the immune checkpoints of Treg cells in KIRC microenvironment [43,47], which is a predominant immunosuppressive effect of tumor [25,47].However, the results of correlation analysis indicated that downregulation of APPL1 may lead to upregulation of immune checkpoints (CTLA4, LAG3, PDCD1, TIGIT, and TGFB1) in KIRC (Figure 5 and Table 1).The infiltrating T cells could have some typical phenotype that may be closely related to the tumor metabolic microenvironment in KIRC [48][49][50], but this study has not yet further explored the molecular mechanism of the effect of APPL1 expression on T cell phenotype transformation in tumor microenvironment.
Our study identified the correlation of APPL1 expression with Treg cell infiltration and poor prognosis in KIRC for the first time.We found the expression of APPL1 associated with oxygen-consuming metabolic process in KIRC (Figure 6(b), the red arrow).The proliferation of renal cancer cells undergoes oxidative metabolic reprogramming during adaptation to the hypoxic tumor microenvironment, such as the downregulation processes of hypoxia-induced fatty acid catabolism [51][52][53].We confirmed that expression level of APPL1 may affect the Treg cell infiltration (or exhausted T cells) in KIRC via downregulated the processes of oxygen-consuming metabolism in tumor microenvironment.(Figures 6(c)-6(f), 7(c) and 7(d), and 8(d)).However, we still need to further verify this conclusion via restoration experiments.In addition, gene MTOR belong to one of the 9 genes related to oxidative metabolism of central carbon, and the expression level of APPL1 may decrease chemosensitivity of KIRC targeting PI3K/AKT/ MTOR (Figure 9).Therefore, low expression of APPL1 may be a reason for chemoresistance in KIRC.
In conclusion, we preliminarily explored the relationship between APPL1 and KIRC and the effect of APPL1 expression.The APPL1 was significantly downregulated with tumor progression in KIRC tissues and might lead to increased infiltration level of Treg cells.Meanwhile, downregulation of APPL1 expression may affect the sensitivity of chemotherapeutics, and it may have contributed to the poor prognosis of KIRC patients.However, we need to further verify or explore the correlation of APPL1 expression  level on different subtypes of KIRC and deeper impact mechanism of immune infiltration based on animal level.In summary, APPL1 may serve as a potential biomarker for the diagnosis or prognosis of KIRC and a potential immunotherapy target.

Figure 1 :Figure 2 Figure 2 :Figure 3 :
Figure 1: Expression of APPL1 in KIRC.(a) APPL1 expression levels of human different tumor types in the TIMER2.0database.(b) APPL1 mRNA expression between tumors and normal tissues of KIRC patients from TCGA database.(c) APPL1 mRNA expression between 72 pairs of tumors and adjacent normal tissues of KIRC patients from TCGA database.(d) APPL1 protein expression between primary tumors and normal tissues of KIRC patients from the CPTAC database.(e-g) The downregulation of APPL1 mRNA was significantly correlated with KIRC patients' pathological stage (e), N classification (f), and histological grade (g).

Figure 3 :
Figure 3: Effect of APPL1 expression on the survival and prognostic value of KIRC patients.(a-c) OS, PFI, and DSS survival curves of KIRC (n = 539 and n = 72).(d-f) Receiver operating characteristic (ROC) analysis curve of APPL1 for OS, PFI, and DSS survival curves in KIRC.(g) Univariate Cox analysis.(h) Multivariate Cox analysis.(i) Nomogram for OS in KIRC patients.(j) The calibration curves for each year.

Figure 8 :
Figure 8: Analysis of survival prognosis and Treg/Tcm cell infiltration from the expression of the 9 genes in KIRC.(a) The expression differences of the 9 genes in KIRC from TCGA database ( * * * p < 0:001).(b) The K-M survival curves of 9 genes, respectively.(c) The ROC curves of the 9 genes, respectively.(d) The correlation analysis of Treg and Tcm cells for the 9 genes, respectively.

Table 1 :
Correlation analysis between APPL1 and gene markers of different types of T cells form TIMER.
et al., Zhai et al., and Liu et al. found