Anti-Tumor Role of CAMK2B in Remodeling the Stromal Microenvironment and Inhibiting Proliferation in Papillary Renal Cell Carcinoma

Abstract

The tumor microenvironment (TME) is variable across tumor types and has diverse effects on malignant progression, based on the type and number of infiltrating stromal cells. In particular, TME effector genes and their competitive endogenous RNA (ceRNA) networks play a critical role in regulating malignant tumor progression. However, the core effector molecules involved in TME modulation of kidney renal papillary cell carcinoma (KIRP) are poorly understood. To address this question, a cohort containing 233 KIRP patients was derived from The Cancer Genome Atlas (TCGA) database, and the data were processed using the ESTIMATE algorithm. We further evaluated the relationship between immune scores (ISs) and stromal scores (SSs) and disease progression and found that high SSs were associated with a poor prognosis in KIRP. Differentially expressed genes (DEGs) were therefore screened based on SS scores, resulting in 2509 DEGs, including 1668 mRNAs, 783 long noncoding (lnc)RNAs, and 58 micro (mi)RNAs. DEGs were then filtered using the random variance and subjected to hierarchical clustering using EPCLUST. Weighted gene co-expression network analysis (WGCNA) was used to assess the prognostic capacity of these DEGs and identify target ceRNA networks, and lncRNA GUSBP11/miR-432-5p/CAMK2B in the turquoise module was selected as a promising ceRNA network. From this analysis CAMK2B was selected as the core gene predicted to be involved in stromal TMA regulation. We therefore explored the expression and function of CAMK2B in vitro and in vivo and provide evidence that this protein promotes stromal TME remodulation and inhibits proliferation in KIRP. Lastly, we show that vascular endothelial growth factor (VEGF), transforming growth factor (TGF)β, and close homolog of L1 (CHL1) act as downstream effectors of CAMK2B in KIRP. Thus, in this study, we show that the TME determines prognosis of KIRP patients via the core effector molecule CAMK2B, which mediates both microenvironmental remodeling and tumor progression. Based on these findings, we propose that remodeling of the stromal microenvironment could represent an improved therapeutic approach relative to immunotherapy for KIRP.

Keywords: CAMK2B; kidney renal papillary cell carcinoma (KIRP); lncRNA GUSBP11; miR-432-5p; tumor microenvironment.

Figure 1

Overview of the data preparation, processing, analysis, and validation pipeline for this study and association between prognosis and both stromal scores (SCs) and immune scores (ISs). (A) Flowchart describing the bioinformatics analysis pipeline and experimental validation strategy in this study. (B, a) Survival analysis for patients in the IS-high and IS-low groups and (b) for patients in the SS-high and SS-low groups, using the R package Survival Analysis. High SSs are associated with a poor prognosis, whereas no survival difference was observed for those with high vs. low ISs.

Figure 2

Differentially expressed genes (DEGs) in patients with high vs. low SSs were evaluated using weighted gene co-expression network analysis (WGCNA), and the turquoise module was selected based on positive pathological roles in KIRP. (A) Unsupervised hierarchical clustering analyses for DEGs, including differentially expressed messenger (m)RNAs, long noncoding (lnc)RNAs, and micro (mi)RNAs, in SS-high and SS-low KIRP tissue samples. The cluster analysis heat map shows the correlation between expression maps and group conditions. The rows represent differentially expressed miRNAs, lncRNAs, and mRNAs, and the columns represent the samples. (B) Sample clustering detection revealed no outlier samples. (C) Soft-threshold power (β) for co-expression of lncRNAs/mRNAs was determined by analyzing the network topology with a soft-threshold power ranging from 1 to 20. (D) Different modules were identified by the Dynamic Tree Cutting method, and each module was assigned a color as an identifier. Six modules were generated after merging based on the correlation of modules with WGCNA. (E) Heatmap plot of the adjacencies in the hub gene network; red represents positive correlation with high adjacency, and blue represents negative correlation with low adjacency. Squares of red color along the diagonal represent the meta-module. (F) Matrix of module–trait relationships and P-values for selected traits. Each column corresponds to a module eigengene, and each row corresponds to a histopathological trait. Each cell contains a corresponding correlation and P-value. The table is color-coded by correlation according to the color legend. (G, a) Gene ontology (GO) enrichment analyses of turquoise mRNAs; significant top 20 GO terms are shown. (G, b) Kyoto Encyclopedia of Genes and genomes (KEGG)-pathway enrichment analyses of turquoise mRNAs; significant top 20 signaling pathways are shown.

Figure 3

Regulation and co-expression of competing endogenous (ce)RNA networks in the turquoise module by WGCNA and selection of CAMK2B as a key gene based on results of correlation strength and survival analyses. (A) Co-expression of ceRNA networks in the turquoise module, including lncRNA GUSBP11/miR-432-5p/CAMK2B, was analyzed by WGCNA. Diamonds represent miRNAs, circles represent mRNAs, and cones represent lncRNAs. Upregulated genes are shown in red, and downregulated genes are shown in green. (B) Hazard ratios and corresponding 95% confidence intervals were estimated using the Cox proportional hazard regression model, and the prognostic effects of these (a) lncRNAs, (b) miRNAs, and (c) mRNAsare presented in a forest plot. (C, a) The prognostic impact of CAMK2B in 33 types of human tumors was analyzed using GEPIA2, revealing that high expression of CAMK2B is associated with poor prognosis in KIRP, liver hepatocellular carcinoma (LIHC), and ovarian cancer (OV). (b) Kaplan–Meier plotter was used to analyze survival, and high expression of CAMK2B was found to be associated with longer overall survival (OS) and recurrence-free survival (RFS).

Figure 4

Expression and clinical role of CAMK2B and its associated ceRNA network in KIRP samples. (A) Expression of the selected ceRNA network, including CAMK2B, mir-432-5p, and lncRNA GUSBP11, was measured using quantitative reverse transcription (qRT)-PCR in KIRP tissue and normal tissue adjacent to the carcinoma. Ns, no significance. (B, a) In cancer patients, high expression of lncRNA GUSBP11 is associated with no metastasis, (b) early-stage disease, and (c) small tumor size. (C, a) High expression of mir-432-5p was associated with tumor metastasis, (b) late-stage disease, and (c) large tumor size. (D, a) High expression of CAMK2B is associated with no metastasis, (b) early-stage disease, and (c) small tumor size. (E) Protein levels of CAMK2B were measured by immunohistochemistry (IHC) using tissue microarrays. (F a, b) Protein levels of CAMK2B are lower in tumor tissue than in paired tumor-adjacent tissue. (c) Expression of CAMK2B is higher in tumors <3 cm in size than in tumors >3 cm in size. (d) High expression of CAMK2B is positively correlated lower tumor grade. (e) High expression of CAMK2B is negatively correlated with tumor metastasis.

Figure 5

Upregulation of CAMK2B is associated with decreases in proliferation and fibroblast infiltration, as well as inhibition of angiogenesis. (A) CAMK2B was overexpressed or subjected to by short-hairpin (sh) RNA-mediated silencing in SK-RC-39 cells using lentiviral transfection. (B) CAMK2B inhibits cell proliferation (a), and this is reversed by CAMK2B silencing, as determined using the CCK8 assay (b). (C) Upregulation of CAMK2B inhibits cell proliferation ability in a sphere formation assay (a), and this is reversed by CAMK2B silencing (b). (D) CAMK2B inhibits cell migration ability (a), whereas migration is enhanced after CAMK2B silencing (b), as measured using the Boyden Chamber migration assay. (E) Diminished subcutaneous tumor growth is observed in mice injected with SK-RC-39-CAMK2B cells relative to those injected with SK-RC-39 cells. (F) Immunohistochemical staining confirms significantly decreased expression of CD34, α-SMA, and PCNA in subcutaneous tumor tissues derived from SK-RC-39-CAMK2B cells vs. tissue from control cells. (G) Negative correlation is observed between expression of CAMK2B and both α-SMA and CD34 in human KIRP tissue from TCGA database using GEPIA2. (H) Significant negative correlations are observed between expression of CAMK2B and both endothelial cells (R=-0.455) and cancer-associated fibroblasts (R=-0.384) using Xcell and EPIC, whereas the correlation with immune cells is weak.

Figure 6

CHL1, VEGF, and TGFβ act as downstream effectors of CAMK2B. (A) Genes co-expressed genes with CAMK2B in KIRP tissue were screened in TCGA database using LinkedOmics and Metascape and visualized in a volcano plot. (B) The protein levels of CAMK2B, CHL1, and VEGF in tissue from KIRP patients were evaluated by immunoblotting. (C) The correlation between the expression levels of CAMK2B and CHL1 (a), CAMK2B and CHL1 (b), and CAMK2B and TGFβ1 (c) were evaluated in TCGA database using GEPIA2. (D) Expressed of several predicted effector molecules, including TGFβ1, VEGF, vimentin, PCNA, CHL1, and E-cadherin were measured in KIRP cells overexpressing or silenced for CAMK2B.