Introduction
Hepatocellular carcinoma (HCC) as a kind of malignant tumor that leads to very high morbidity and mortality around the world, especially in China.1 First-line conventional therapies, such as chemotherapy, radiotherapy and surgery, have demonstrated limited results, and the prognosis of patients remains quite poor. Moreover, recurrence is the main cause of the disease. In the early stages of HCC, surgery is still the main treatment option, while in the later stages, surgery is typically combined with platinum-based adjuvant therapy and cytotoxic chemotherapy and/or radiotherapy.
Immune biology plays a crucial part in oncogenesis and development, and immunotherapy is deemed to be a promising direction of cancer treatment,2 whereby researchers are trying to modulate the body’s own immune system to fight and prevent various type of cancers.3 In recent years, immunotherapies including monoclonal antibodies and adoptive cell transfers have been increasingly integrated into the clinic for the treatment of various types of cancer, such as melanoma and lung cancer.4 In the past decade, the discovery of antibodies against immune checkpoints (i.e. PD-1 and PD-L1) have remodeled the treatment of non-small cell lung cancer (NSCLC).4 Immunotherapy, as represented by PD-1 and PD-L1 inhibitors, has showed promising anti-tumor effects in various types of cancer, including non-small cell lung cancer and melanoma.5 Anti-CTLA4 also showed a clinical curative effect in HCC, similar to blockade of PD-1 or PD-L1 having shown partial response in advanced liver cancer.6 Additionally, more and more studies have demonstrated that the tumor-infiltrating lymphocytes (TILs) play an important role in response to chemotherapy and improving prognosis of various types of cancer,7 such as tumor-associated macrophages (TAMs)8 and tumor-infiltrating neutrophils (TINs), which also contribute to prognosis.9,10 However, the bull’s-eye for HCC treatment via targeting tumor microenvironment (TME) remains unelucidated. It has been demonstrated that the hepatic microenvironment epigenetically shapes lineage commitment in mosaic mouse models of liver tumorigenesis.11 TME of HCC harbors a significant level of T cells, as indicated by small conditional RNA-sequencing, however, and the TILsare incapable of killing tumor cells,12 implying that TME is very complicated. So, there is an urgent demand to clarify the immunophenotype of tumor-immune interactions and identify new immune-related therapeutic targets for liver cancer.
Lysophosphatidic acid (LPA) is a kind of lipid that is involved in tumor proliferation and one of their receptors (LPAR6) is the latest determinate G protein-coupled receptor (GPCR) of LPA family,13,14 and it has been revealed to be associated with several types of cancer, including colorectal, prostate, pancreatic cancer and HCC.15–18 However, the function of LPAR6 remains highly controversial, as demonstrated by the previous studies. LPAR6 acts as a tumor suppressor and inhibits tumor migration in colorectal cancer, whereas in the other tumors mentioned, the LPAR6 protein might act as a facilitator.16–18 All these findings indicate that the proteins encoded by LPAR6 may play an essential role in cancer, but the association between LPAR6 and tumor progression and the underlying mechanism is still not well understood.
Our previous study showed that LPAR6 was highly expressed in several T cell subgroups, including naïve T cells, CD38+ T cells and plasmablasts, while showing low expression in T regulatory cells (Tregs) and exhausted T cells in chemical-induced cancer mouse model (unpublished data), based on single-cell RNA sequencing. These findings suggest that LPAR6 may have multifaceted functional roles in modulating or recruiting TILs; in this way, they are able to remodel the tumor microenvironment. However, the underlying functions and mechanisms of LPAR6 in tumor progression and tumor immunology is still not well understood.
In this work, we extensively studied the expression level of LPAR6 and the relationship with prognosis of cancer patients according databases as Oncomine, PrognoScan, and Kaplan-Meier Plotter. Moreover, we investigated the correlation of LPAR6 with tumor-infiltrating immune cells in the different tumor microenvironments.
All the findings in this report throw the light on the key role of LPAR6 in HCC and provide insight into a potential relationship and a fundamental mechanism between LPAR6 and tumor-immune interactions.
Results
mRNA expression levels of LPAR6 in different types of human cancers
To determine differences of LPAR6 expression in tumor and normal tissues, the LPAR6 mRNA levels in different tumors and normal tissues of multiple cancer types were analyzed using the Oncomine database. This analysis revealed that the LPAR6 expression was higher in kidney cancer, leukemia, liver cancer and lymphoma compared to the normal tissues, and lower expression of LPAR6 was observed in bladder cancer, breast cancer, cervical cancer and esophageal cancer compared to the normal tissues (cancer vs. normal) (Fig. 1A). In addition, LPAR6 expression was higher in brain and central nervous system (CNS), leukemia, ovarian cancer and sarcoma in tumor tissues (cancer vs. cancer), whereas in brain and CNS, kidney, ovarian cancer and sarcoma, LPAR6 is low expression (Fig. 1A). The detailed results of LPAR6 expression in different cancer types are summarized in Supplementary Table 1.
To further evaluate LPAR6 expression in human cancers, we examined LPAR6 expression using the RNA-sequencing data of multiple malignancies in TCGA. The differential expression between the tumor and adjacent normal tissues for LPAR6 across all TCGA tumors is shown in Figure 1B. LPAR6 expression was significantly lower in the tumor tissue of bladder urothelial carcinoma (BLCA), breast invasive carcinoma (BRCA), colon adenocarcinoma (COAD), head and neck cancer (HNSC), kidney chromophobe (KICH), lung adenocarcinoma (LUAD), prostate adenocarcinoma (PRAD), rectum adenocarcinoma (READ) and uterine corpus endometrial carcinoma (UCEC) compared with adjacent normal tissues and was significantly higher in esophageal carcinoma (ESCA), kidney renal clear cell carcinoma (KIRC), kidney renal papillary cell carcinoma (KIRP), thyroid carcinoma (THCA) compared with adjacent normal tissues and slightly lower in liver hepatocellular carcinoma (LIHC) tumor tissue (Fig. 1 B).
GEPIA2 generates dot plots profiling gene/isoform expression across multiple cancer types and paired normal adjacent samples. The differential expression level for LPAR6 between tumor and matched TCGA normal data across all TCGA is shown in Figure 1C. The expression level of LPAR6 was significantly higher in cholangiocarcinoma (CHOL), ESCA and KIRC and lower in KICH, BRCA, KICH and UCEC compared with the adjacent normal tissues, and slightly higher in LIHC tumor tissue.
Prognostic potential of LPAR6 in cancers
We investigated whether LPAR6 level was associated with various types of cancer. The relationships between LPAR6 expression and prognosis of different cancers are shown in Supplementary Table 2. Notably, LPAR6 expression significantly affects OS in several types of cancer, including lung, bladder, breast, colorectal, eye and ovarian (Fig. 2A–M). Notably, LPAR6 expression significantly impacts OS in two types of cancers, namely lung and breast (Fig. 2A, B, D). Two cohorts (GSE3141 and GSE4573) of lung cancer showed that high LPAR6 expression was associated with better prognosis (OS HR=0.53, 95% CI=0.36 to 0.80, Cox p=0.00206181; OS HR=0.53, 95% CI=0.31 to 0.91, Cox p=0.0219869) (Fig. 2A, B). Therefore, it is conceivable that higher LPAR6 expression is an independent risk factor and leads to a better prognosis in lung cancer patients, and HR below 0 indicates LPAR6 expression is a protective factor. Also, high LPAR6 expression significantly impacts DSS in bladder cancer and RFS and DFS in the breast cancer (Fig. 2C, E, F). Moreover, three cohorts (GSE19615, GSE9195 and GSE11121) of breast cancer showed that the higher LPAR6 expression was associated with a better prognosis of DMFS (Fig. 2G–I). Higher expression level of LPAR6 was associated with better prognosis in some other types of cancer (Fig. 2J–M).
To further explore the prognostic potential of LPAR6 in various types of cancer, the Kaplan-Meier Plotter database was employed to determine the LPAR6 prognostic value based on Affymetrix microarrays and RNA-sequencing data. Similarly, better prognosis in breast cancer (OS and DMSF) and lung cancer (OS and PPS) was shown to correlate with higher LPAR6 expression (Fig. 2N–P, T–V). Better prognosis (OS, PFS and RFS) in liver cancer was shown to correlate with higher LPAR6 expression level (Fig. 2Q–S). These data supported the prognostic value of LPAR6 in some specific types of cancers and that increased and decreased LPAR6 expression have different prognostic values depending on the specific cancer type.
The RNA-sequencing data in the TCGA were also used to explore the prognostic potential of LPAR6 in different cancers via GEPIA2. We analyzed the association between LPAR6 expression and prognostic values in 33 types of cancer. The expression level of LPAR6 significantly impacts prognosis in three types of cancers, including adrenocortical carcinoma (ACC), lower grade glioma (LGG) (Supplementary Fig. 1). High LPAR6 expression levels were associated with better prognosis of OS in skin cutaneous melanoma (SKCM) but appeared to have less influence on DFS. These results confirmed the prognostic value of LPAR6 in some specific types of cancers and that decreased and increased LPAR6 expression have different prognostic values depending on the type of cancers.
High LPAR6 expression impacts the prognosis of liver cancer in different clinical characteristics
To better study the relevance and underlying mechanisms of LPAR6 expression in cancer, we investigated the correlation between the LPAR6 expression level and clinical characteristics of liver cancer, especially the different clinical stages.
High expression of LPAR6 was correlated with better OS in the early stage of the progress of cancer (Table 1). Higher expression of LPAR6 was correlated with better OS in both female and male patients, whereas different association patterns were shown among different races (Table 1). Besides the Asian race, LPAR6 was associated with better OS in people of the White race.
Table 1Correlation of LPAR6 mRNA expression and clinical prognosis in HCC with different clinicopathological factors by the Kaplan-Meier Plotter
| Clinicopathological characteristics | OS, n=364
| Progression-free survival, n=370
|
|---|
| n | HR | p | n | HR | p |
|---|
| Sex |
| Female | 118 | 0.45 (0.24–0.86) | 0.013 | 120 | 0.75 (0.41–1.35) | 0.33 |
| Male | 246 | 0.57 (0.36–0.91) | 0.017 | 246 | 0.66 (0.46–0.95) | 0.024 |
| Race |
| White | 181 | 0.5 (0.3–0.84) | 0.0073 | 183 | 0.7 (0.47–1.06) | 0.088 |
| Black or African American | 17 | – | – | 17 | – | – |
| Asian | 155 | 0.68 (0.37–1.25) | 0.21 | 155 | 0.62 (0.38–1.01) | 0.054 |
| Stage |
| 1 | 170 | 0.47 (0.25–0.9) | 0.02 | 170 | 1.55 (0.94–2.55) | 0.084 |
| 1+2 | 253 | 0.6 (0.37–0.97) | 0.037 | 254 | 0.54 (0.36–0.83) | 0.0039 |
| 2 | 83 | 0.52 (0.22–1.24) | 0.13 | 84 | 0.59 (0.32–1.06) | 0.075 |
| 2+3 | 166 | 1.45 (0.85–2.45) | 0.17 | 167 | 0.76 (0.51–1.14) | 0.19 |
| 3 | 83 | 1.57 (0.85–2.89) | 0.14 | 83 | 1.28 (0.72–2.28) | 0.4 |
| 3+4 | 87 | 1.57 (0.87–2.84) | 0.13 | 88 | 1.25 (0.71–2.19) | 0.44 |
| 4 | 4 | – | – | 5 | – | – |
| AJCC_T |
| 1 | 180 | 0.5 (0.27–0.92) | 0.024 | 180 | 1.48 (0.91–2.4) | 0.11 |
| 2 | 90 | 0.56 (0.25–1.26) | 0.16 | 92 | 0.6 (0.35–1.05) | 0.07 |
| 3 | 78 | 1.65 (0.88–3.1) | 0.11 | 78 | 1.36 (0.75–2.49) | 0.31 |
| 4 | 13 | – | | 13 | – | |
| Vascular invasion |
| None | 203 | 0.46 (0.27–0.78) | 0.0034 | 204 | 0.64 (0.4–1.02) | 0.06 |
| Micro | 90 | 2.27 (0.95–5.38) | 0.057 | 91 | 0.7 (0.36–1.38) | 0.3 |
| Macro | 16 | – | | 16 | | |
| Risk factors |
| Alcohol consumption | | | | | | |
| Yes | 115 | 0.56 (0.29–1.06) | 0.07 | 115 | 0.59 (0.35–0.98) | 0.041 |
| None | 202 | 0.58 (0.35–0.94) | 0.026 | 204 | 0.77 (0.51–1.15) | 0.2 |
| Hepatitis virus | | | | | | |
| Yes | 150 | 0.56 (0.29–1.07) | 0.076 | 152 | 0.53 (0.33–0.87) | 0.01 |
| None | 167 | 0.57 (0.34–0.98) | 0.03 | 167 | 0.76 (0.49–1.17) | 0.21 |
Higher LPAR6 expression level was associated with better OS in stage 1 and 1+2 and better PFS in stage 1+2 of HCC patients respectively but was not correlated with OS and PFS of other stages. This phenomenon was also verified by the American Joint Committee on Cancer classification, we found that higher expression of LPAR6 was associated with better OS only in HCC early stage (OS HR=0.5, p=0.024) (Table 1). The vascular invasion could also function as an indicator of cancer staging during progression.30 We discovered that higher LPAR6 expression was correlated with better OS in the non-vascular invasion liver cancer patients. All these consequences suggest that LPAR6 expression level can affect the prognosis in early HCC staged patients but not associated with PFS and OS of late stage HCC patients.
Low promoter methylation levels of LPAR6 impacts the clinicopathological parameters of liver cancer patients
Low promoter methylation levels of LPAR6 were associated with the earlier stage of the progress of cancer in LIHC (Fig. 3), which implies that the earlier stage, the lower promoter methylation levels of LPAR6. This could be an explanation to the higher LPAR6 expression level being associated with better OS in earlier stage HCC (Fig. 3).
Interaction network of LPAR6
We found that LPAR6 co-expressed with 19 proteins, and shared protein domains with ADRB2 and physical interactions with dystrophin by constructing a LPAR6 interaction network (Fig. 4A). The top 50 negatively [green spot; false discovery rate (FDR) <0.05] and positively correlated genes (red spot; FDR<0.05), with the expression of LAPR6 displayed as a volcano plot by LinkedOmics online tools (Fig. 4B). These results indicate that LPAR6 serves a critical role in cancer development. A strong positive association between the expression levels of LPAR6 was revealed by Pearson’s correlation coefficient analysis. Biological process and molecular function analyses showed that LPAR6-associated differentially expressed genes were involved in a number of biological processes and molecular functions, including ‘interleukin production’, ‘cytokine production’, ‘respiratory burst’, ‘inflammatory cell apoptotic process’, ‘inflammatory response’, and some other immune biology process in LIHC (Fig. 4C). All the above imply that LPAR6 serves a key role in immune system activation, cellular responses to stimulation, metabolism and a number of other processes.
LPAR6 expression is correlated with immune infiltration level in HCC
Tumor-infiltrating lymphocytes are an independent predictor of survival in cancers.31,32 Therefore, we investigated whether LPAR6 expression was correlated with immune infiltration levels in different types of cancer. We assessed the correlations of LPAR6 expression with immune infiltration levels in 39 cancer types from TIMER. The results showed that LPAR6 expression had significant negative correlations with tumor purity in 26 types (BLCA, BRCA, BRCA-Basal, BRCA-Her2, BRCA-Luminal, CESC, CHOL, DLBC, GBM, HNSC, HNSC-HPVneg, KICH, KIRC, KIRP, LGG, LIHC, LUAD, LUSC, OV, PCPG, SARC, SKCM, SKCM-Metastasis, STAD, TGCT, UCEC, and UVM) of cancer, which indicates LPAR6 is somehow related to recruiting lymphocytes to the tumor, and significant correlations with B cell infiltration levels in 13 types of cancers (BRCA-Basal, BRCA-Luminal, CHOL, COAD, GBM, HNSC-HPVpos, KIRC, LGG, LIHC, LUAD, THCA, THYM, UCEC) (Supplementary Fig. 2). In addition, LPAR6 expression had significant correlations with infiltrating levels of CD8+ T cells in 24 types of cancer (ACC, BRCA, BRCA-Basal, BRCA-Her2, BRCA-Luminal, CESC, COAD, HNSC, HNSC-HPVpos, HNSC-HPVneg, KIRC, KIRP, LGG, LIHC, LUAD, MESO, OV, READ, SKCM, SKCM-Primary, SKCM-Metastasis, STAD, THCA, UCEC), CD4+ T cells in 26 types of cancer (ACC, BLCA, BRCA, BRCA-Basal, BRCA-Her2, BRCA-Luminal, CESC, CHOL, COAD, DLBC, KICH, KIRC, KIRP, LGG, LIHC, LUAD, MESO, OV, PAAD, PCPG, STAD, TGCT, UCEC, USC), macrophages in 20 types of cancer (ACC, BRCA, BRCA-Basal, BRCA-Her2, BRCA-Luminal, CHOL, COAD, HNSC, HNSC-HPVneg, KIRP, LGG, LIHC, LUAD, MESO, OV, PCPG, SKCM, SKCM-Metastasis, STAD, UCEC), neutrophils in 33 types of cancer (ACC, BLCA, BRCA, BRCA-Basal, BRCA-Her2, BRCA-Luminal, CESC, COAD, ESCA, HNSC, HNSC-HPVpos, HNSC-HPVneg, KIRC, KIRP, LGG, LIHC, LUAD, LUSC, OV, PAAD, READ, SKCM, SKCM-Primary, SKCM-Metastasis, STAD, TGCT, THCA, THYM, UCEC, UCS, UVM), and dendritic cells in 21 types of cancer (ACC, BRCA, BRCA-Basal, BRCA-Her2, BRCA-Luminal, CESC, CHOL, COAD, ESCA, KIRC, KIRP, LGG, LIHC, LUAD, OV, PCPG, SKCM, STAD, THCA, THYM, UCEC). (Supplementary Fig. 2).
In view of the correlation between the expression level of LPAR6 and the level of immune infiltration in various types of cancer, we can determine that LPAR6 is associated with prognosis and immune infiltration in specific cancer types. Tumor purity reflects the degree of immune infiltration of clinical tumor samples, and purity is negatively correlated with the degree of immune infiltration.22,28 Therefore, we selected cancer types in which the expression level of LPAR6 was significantly negatively correlated with tumor purity in TIMER and significantly correlated with prognosis. What attracted our attention is that the expression level of LPAR6 is associated with a better OS prognosis and higher immune infiltration in HCC.
The expression level of LPAR6 in LIHC was significantly negatively correlated with tumor purity (Fig. 5). The expression level of LPAR6 was significantly positively correlated with the infiltration level of T cells (CD8 + T, CD4 + T) B cells, neutrophils, macrophages and dendritic cells in LIHC (Fig. 5). These findings strongly indicate that LPAR6 plays a specific role in immune infiltration in different types of HCC.
Correlation analysis between LPAR6 expression and immune marker sets
In order to study the relationship between LPAR6 and various immune infiltrating cells, we studied the correlation between LPAR6 and the immune marker sets of various immune cells of LIHC in the TIMER and GEPIA databases. We analyzed the correlation of the expression level of LPAR6 of different immune cells (including T cells (CD8 + T and T general), B cells, monocytes, neutrophils, tumor association macrophages, natural killer cells, M1 and M2 macrophages and dendritic cells between immune marker genes’ expression in LIHC (Table 2 and Fig. 6). We also analyzed the different functional T cells, such as Th1, Th2, Th17, follicular helper T cells, and Tregs, as well as exhausted T cells.33
Table 2Correlation analysis between LPAR6 and related marker genes of immune cells
| Description | Gene markers | LIHC
| THYM
|
|---|
None
| Purity
| None
| Purity
|
|---|
| Cor | p | Cor | p | Cor | p | Cor | p |
|---|
| CD8+ T cell | CD8A | 0.505 | *** | 0.341 | *** | −0.108 | 0.239 | −0.177 | 0.0588 |
| CD8B | 0.451 | *** | 0.292 | *** | −0.219 | 0.0162 | −0.291 | * |
| T cell, general | CD3D | 0.474 | *** | 0.312 | *** | −0.163 | 0.0745 | −0.24 | * |
| CD3E | 0.6 | *** | 0.421 | *** | −0.119 | 0.196 | −0.192 | 0.0398 |
| CD2 | 0.597 | *** | 0.436 | *** | −0.146 | 0.112 | −0.225 | 0.0158 |
| Naive T cell | CCR7 | 0.59 | *** | 0.389 | *** | 0.163 | 0.0748 | 0.148 | 0.114 |
| LEF1 | 0.202 | *** | 0.095 | 0.0775 | −0.067 | 0.466 | −0.137 | 0.145 |
| TCF7 | 0.147 | * | −0.161 | * | −0.132 | 0.15 | −0.211 | 0.0237 |
| SELL | 0.484 | *** | 0.315 | *** | 0.18 | 0.0497 | 0.137 | 0.142 |
| Effector T cell | CX3CR1 | 0.406 | *** | 0.318 | *** | −0.026 | 0.782 | 0.01 | 0.918 |
| FGFBP2 | 0.138 | * | 0.092 | 0.0896 | −0.13 | 0.157 | −0.194 | 0.0377 |
| FCGR3A | 0.43 | *** | 0.267 | *** | 0.083 | 0.369 | 0.112 | 0.235 |
| Effector memory T cell | PDCD1 | 0.416 | *** | 0.257 | *** | −0.055 | 0.55 | −0.134 | 0.154 |
| DUSP4 | 0.439 | *** | 0.261 | *** | 0.284 | * | 0.278 | * |
| GZMK | 0.556 | *** | 0.374 | *** | 0.426 | * | 0.439 | *** |
| GZMA | 0.503 | *** | 0.34 | *** | 0.383 | *** | 0.399 | *** |
| IFNG | 0.296 | *** | 0.158 | * | 0.157 | 0.0865 | 0.16 | 0.0872 |
| Resident memory T cell | CD69 | 0.587 | *** | 0.413 | *** | 0.307 | ** | 0.274 | * |
| ITGAE | 0.188 | ** | 0.144 | * | −0.034 | 0.708 | −0.108 | 0.251 |
| CXCR6 | 0.564 | *** | 0.394 | *** | 0.229 | 0.0122 | 0.246 | * |
| MYADM | 0.335 | *** | 0.25 | *** | −0.102 | 0.268 | −0.079 | * |
| B cell | CD19 | 0.408 | *** | 0.268 | *** | 0.322 | ** | 0.301 | * |
| CD79A | 0.526 | *** | 0.335 | *** | −0.023 | 0.804 | −0.093 | 0.322 |
| Monocyte | CD86 | 0.594 | *** | 0.414 | *** | 0.118 | 0.198 | 0.118 | 0.208 |
| CD115 (CSF1R) | 0.587 | *** | 0.391 | *** | 0.186 | 0.0422 | 0.215 | 0.0209 |
| TAM | CCL2 | 0.559 | *** | 0.36 | *** | 0.098 | 0.286 | 0.099 | 0.291 |
| CD68 | 0.405 | *** | 0.213 | *** | 0.183 | 0.0454 | 0.172 | 0.0667 |
| IL10 | 0.527 | *** | 0.342 | *** | 0.296 | * | 0.273 | * |
| M1 Macrophage | INOS (NOS2) | 0.117 | 0.025 | 0.07 | 0.198 | −0.256 | * | −0.231 | 0.013 |
| IRF5 | 0.142 | * | 0.124 | 0.022 | 0.455 | *** | 0.467 | *** |
| COX2 (PTGS2) | 0.605 | *** | 0.434 | *** | 0.115 | 0.209 | 0.14 | 0.137 |
| M2 Macrophage | CD163 | 0.514 | *** | 0.319 | *** | 0.339 | *** | 0.328 | *** |
| VSIG4 | 0.55 | *** | 0.376 | *** | 0.207 | 0.0238 | 0.227 | 0.0148 |
| MS4A4A | 0.543 | *** | 0.34 | *** | 0.269 | * | 0.251 | * |
| Neutrophil | CD66b (CEACAM8) | 0.116 | 0.025 | 0.084 | 0.12 | 0.053 | 0.566 | 0.029 | 0.756 |
| CD11b (ITGAM) | 0.362 | *** | 0.2 | ** | 0.055 | 0.553 | 0.05 | 0.595 |
| CCR7 | 0.59 | *** | 0.389 | *** | 0.163 | 0.0748 | 0.148 | 0.114 |
| Natural killer cell | KIR2DL1 | 0.096 | 0.064 | 0.056 | 0.3 | 0.104 | 0.259 | 0.135 | 0.149 |
| KIR2DL3 | 0.139 | * | 0.145 | 0.405 | −0.026 | 0.781 | −0.026 | 0.785 |
| KIR2DL4 | 0.217 | *** | 0.147 | * | 0.184 | 0.0437 | 0.231 | 0.0132 |
| KIR3DL1 | 0.118 | 0.023 | 0.044 | 0.414 | 0.112 | 0.222 | 0.149 | 0.111 |
| KIR3DL2 | 0.221 | *** | 0.132 | 0.014 | 0.112 | 0.224 | 0.12 | 0.203 |
| KIR3DL3 | 0.089 | 0.088 | 0.076 | 0.158 | 0.022 | 0.811 | 0.057 | 0.548 |
| KIR2DS4 | 0.154 | * | 0.178 | ** | 0.043 | 0.642 | 0.079 | 0.399 |
| Dendritic cell | HLA-DPB1 | 0.558 | *** | 0.37 | *** | 0.21 | 0.0216 | 0.203 | 0.0297 |
| HLA-DQB1 | 0.473 | *** | 0.29 | *** | 0.154 | 0.0924 | 0.115 | 0.22 |
| HLA-DRA | 0.538 | *** | 0.345 | *** | 0.303 | ** | 0.305 | ** |
| HLA-DPA1 | 0.577 | *** | 0.399 | *** | 0.277 | * | 0.29 | * |
| BDCA-1 (CD1C) | 0.568 | *** | 0.421 | *** | −0.149 | 0.105 | −0.219 | 0.0185 |
| BDCA-4 (NRP1) | 0.316 | *** | 0.24 | *** | 0.123 | 0.181 | 0.187 | 0.0456 |
| CD11c (ITGAX) | 0.593 | *** | 0.438 | *** | 0.305 | ** | 0.315 | ** |
| Th1 | TBX21 (T-bet) | 0.529 | *** | 0368 | *** | 0.242 | * | 0.224 | 0.0159 |
| STAT4 | 0.479 | *** | 0.39 | *** | 0.078 | 0.395 | 0.09 | 0.336 |
| STAT1 | 0.276 | *** | 0.152 | * | 0.259 | * | 0.292 | * |
| IFNG (IFN-g) | 0.296 | *** | 0.158 | * | 0.157 | 0.0865 | 0.16 | 0.0872 |
| TNF-a (TNF) | 0.513 | *** | 0.343 | *** | 0.271 | * | 0.263 | * |
| Th2 | GATA3 | 0.601 | *** | 0.45 | *** | −0.145 | 0.114 | −0.225 | 0.0156 |
| STAT6 | 0.035 | 0.502 | 0.012 | 0.82 | 0.076 | 0.411 | 0.139 | 0.137 |
| STAT5A | 0.357 | *** | 0.203 | ** | −0.31 | ** | −0.291 | * |
| IL13 | 0.125 | 0.016 | 0.11 | 0.041 | 0.255 | * | 0.284 | * |
| Tfh | BCL6 | 0.019 | 0.712 | 0.012 | 0.831 | 0.045 | 0.622 | 0.035 | 0.711 |
| IL21 | 0.079 | 0.127 | 0.005 | 0.922 | −0.084 | 0.362 | −0.096 | 0.31 |
| Th17 | STAT3 | 0.185 | ** | 0.023 | 0.665 | 0.097 | 0.294 | 0.155 | 0.0977 |
| IL17A | 0.103 | 0.048 | 0.112 | 0.037 | 0.223 | 0.0146 | 0.226 | 0.0152 |
| Treg | FOXP3 | 0.299 | *** | 0.193 | ** | 0.215 | 0.0187 | 0.241 | * |
| CCR8 | 0.463 | *** | 0.348 | *** | 0.08 | 0.387 | 0.051 | 0.59 |
| STAT5B | −0.084 | 0.105 | −0.013 | 0.816 | 0.022 | 0.808 | 0.034 | 0.714 |
| TGFB1 (TGFb) | 0.491 | *** | 0.332 | *** | 0.033 | 0.722 | 0.022 | 0.816 |
| T cell exhaustion | PDCD1 (PD-1) | 0.416 | *** | 0.257 | *** | −0.055 | 0.55 | −0.134 | 0.154 |
| CTLA4 | 0.45 | *** | 0.296 | *** | 0.248 | * | 0.245 | * |
| LAG3 | 0.331 | *** | 0.25 | *** | 0.138 | 0.131 | 0.138 | 0.141 |
| HAVCR2 (TIM-3) | 0.575 | *** | 0.386 | *** | 0.315 | ** | 0.306 | ** |
| GZMB | 0.32 | *** | 0.156 | * | 0.133 | 0.147 | 0.115 | 0.219 |
After correlation adjustment for purity, the results showed that the expression level of LPAR6 was significantly related to most immune marker sets of various immune cells and T cell subtypes, especially the effector T cells in LIHC, which were negatively related to THYM, and THYM was related to poor prognosis. We employed THYM as a negative control here. (Table 2 and Fig. 6).
We found that the expression levels of the marker genes in general T cells, CD8+ T cells, naive T cells, effector T cells, natural killer cells, M1 macrophages and dendritic cells have strong correlations with LPAR6 expression in LIHC (THYM as a negative control which with poor prognosis) (Table 2). Specifically, we showed NOS2, IRF5 and PTGS2 of M1 phenotype were significantly correlated with LPAR6 expression in LIHC (p<0.0001). It is reported that M1 could prevent tumor development 8,10. Further studies need to be done on whether LPAR6 is a crucial factor mediating the de-polarization of macrophages and remodeling the tumor microenvironment. In addition, for Tregs, LPAR6 did not demonstrate a correlation with the Treg markers, such as STAT5B in LIHC (Table 2). We further analyzed the correlation between LPAR6 expression and the above markers of monocytes and various types of T cells in normal and tumor tissue in the GEPIA database, and we found that the correlation results between LPAR6 and markers of monocytes and TAMs were similar to those in TIMER (Supplementary Table 3, Figs. 7, 8).
Different correlation patterns between tumor and normal tissue in LIHC patients
We found that the expression levels of most marker sets of immune cells, including resident memory T cells, effector Treg, Th1-like, have strong correlations with LPAR6 expression both at a similar level in tumor and normal tissue in the LIHC. The more interesting thing is that in the naive T cell, effector T cell, effector memory T cell, central memory T cell and T cell exhaustion populations, the correlation coefficients were higher in normal tissue, whereas an inverse phenomenon had been detected in the resting Treg population (Fig. 8 and Supplementary Table 3). High LPAR6 expression relates to a high infiltration level of dendritic cells in the tumor tissue of LIHC patients, dendritic cells markers such as HLA-DQB1, CD1C and NRP1 show significant correlations with LPAR6 expression both in the tumor tissue in LIHC (Supplementary Table 3). HLA-DPB1, HLA-DRA, HLA-DPA1 and CD11c also showed significant correlations with LPAR6 expression in both tumor and normal tissue in LIHC (Supplementary Table 3 and Fig. 8). These results further revealed that there is a strong relationship between LPAR6 and dendritic cell infiltration. This finding suggests that there are different correlation patterns between tumor and normal tissues in LIHC patients. This exciting finding indicates that LPAR6 may regulate dendritic cell infiltration in the tumor microenvironment of the LIHC patient and LPAR6 may be a novel target for HCC therapy.
Discussion
LPA receptors are GPCRs that bind the LPA and activate multiple cellular responses, such as cell proliferation, apoptosis, cytoskeletal rearrangements and motility.34–36 To date, five LPA receptors (LPAR1-5) have been well characterized and extensively studied.37 LPAR6 is a newly identified receptor, known as ARWH1, HYPT8, LAH3 and P2RY5, and the originally-referred-to purinergic receptor P2Y5 that is involved in inherited hair loss.14,38 Although LPAR6 has not been extensively studied, it was reported that LPAR6 negatively regulates tumor cell migration in colorectal cancer,15 and LPAR6 expression was down-regulated in P53-mutated cases. It was also reported that the LPA axis plays an important role in HCC by stimulating the recruitment and trans-differentiation of peritumoral fibroblasts into carcinoma-associated fibroblasts.39,40 This gives us a clue that LPAR6 is involved in the tumor microenvironment. Nowadays, immunotherapy is applied as a novel treatment for patients with advanced cancer. Immunotherapy has shown good results in the treatment of NSCLC, but in HCC treatment, tumor immunotherapy is not effective.
In this study, we demonstrated that variations in LPAR6 expression levels are associated with the prognosis of different types of cancer. Higher expression level of LPAR6 is associated with a better prognosis of three types of cancers, including liver, breast and lung cancer. Here, we used the independent data set in Oncomine and 33 types of TCGA data in GEPIA2 to determine the mRNA expression level of LPAR6 in different types of cancer and the prognosis of the system. Differential expression patterns of LPAR6 between cancer and normal tissues is observed in many types of cancer.
When we looked into the Oncomine database, we found that the discrepancies in levels of LPAR6 expression in different cancer types among different databases might reflect the data collection approaches and underlying mechanisms pertinent to different biological properties. Nevertheless, in these databases we also found consistent prognostic correlation patterns between LPAR6 mRNA expression level in breast, bladder, colorectal, cervical, lung, esophageal and prostate cancers. The analysis of the TCGA database revealed that higher mRNA expression level of LPAR6 is associated with better prognosis in LGG, ACC, SKCM (Supplementary Fig. 2). Furthermore, analysis of data from two databases (Kaplan-Meier Plotter and PrognoScan) showed a higher mRNA level of LPAR6 expression was correlated with better prognosis in lung, breast, colorectal, bladder, ovarian and eye cancers (Fig. 2).
When we looked into two datasets in the PrognoScan database, we found that the mRNA expression level of LPAR6 could act as independent risk factors for prognosis in liver cancer and LUAD. That is, higher level of LPAR6 expression was shown to be associated with better prognosis of liver cancer in the early stages (stage 1 and stage 1+2) with the lowest HR for a better OS when LPAR6 was highly expressed. Considering these findings collectively, we believe that LPAR6 is a prognostic biomarker in HCC.
Another significant findings of this study is that the expression level of LPAR6 is correlated with a variety of immune infiltration levels in cancer (especially LIHC). Our results indicate that there is a strong positive correlation between the infiltration levels of T cells (CD8 + T and CD4 + T), neutrophils, macrophages, and dendritic cells and the mRNA expression level of LPAR6 in LIHC (Fig. 5). The correlation between LPAR6 expression and immune cell marker genes suggests a role for LPAR6 in regulating tumor immunology in these types of cancers. The possible explanation for this striking effect is that LPAR6 could orchestrate the functions of multiple immune marker genes. This supports the idea that LPAR6 tumor levels are important contributors to human disease and indicators of the prognosis of specific cancer types.
First, gene markers of M1 macrophages, such as PTGS2 and IRF5, show significant correlations with LPAR6 expression in LIHC (Table 2). Since macrophages are functionally plastic cells, M1 macrophages produce type 1 cytokines to prevent tumors from developing, whereas M2 macrophages induce type 2 cytokines to facilitate tumor growth. Especially in tumor tissue of LIHC, both NOS2 and IRF5 show significant correlations with LPAR6 expression and PTGS2 shows a significant correlation with LPAR6 expression in the tumor tissue of LIHC (Supplementary Table 3). These results reveal the potential regulating role of LPAR6 in depolarization of macrophages against tumor tissue that activated macrophages can be repolarized towards the opposite functional phenotypes by microenvironmental modifications and then inhibit tumor growth.
Second, our results indicated that LPAR6 has the potential to activate CD8+ T cells, naive T cells, effector T cells and natural killer cells and to inactivate Tregs and decrease T cell exhaustion. CD8A, a crucial surface protein on T cells, is highly correlated with LPAR6 expression in LIHC, which are types of cancers with better prognosis. Moreover, CD8A negatively correlated in THYM, which has poor prognosis (Table 2). This pattern also occurs with the general T cell markers, such as CD3D, CD3E and CD2, and most markers of naive T cells, effector T cells, effector memory T cells and natural killer cells. Consider, LEF1, which has been proven as a predictor of better treatment response in acute myelocytic leukemia (AML), due to high expression level being associated with favorable RFS in patients and predicted a significantly better overall survival for AML patients.41 Furthermore, the LPAR6 expression does not positively correlate with the Treg markers, such as STAT5B in LIHC (Table 2).
Third, different correlation patterns can be found between LPAR6 expression and the regulation of several markers of Th cells (Th1, Th2, follicular helper T cells, and Th17) in these different cancers. Interferon-gamma is a Th1 cytokine with both pro- and anti-cancer properties42 and is highly correlated with LPAR6 expression in LIHC, whereas it did not demonstrate significant correlations in THYM (Table 2). Interleukin (IL)-13 is an important immunoregulatory cytokine which is mainly produced by activated Th2 cells, and is widely involved in tumorigenesis and development, fibrosis and inflammation.[43,44] We found that IL-13 is highly correlated with LPAR6 expression in THYM, but it did not demonstrate significant correlations in THYM (adjusted by purity) and a similar situation was observed for IL-21. So, these could be explanations as to why LPAR6 indicates a poor prognosis in THYM and a better prognosis in LIHC.
All these correlations listed above could be indications of a potential mechanism whereby LPAR6 regulates T cell functions in LIHC. Together, these findings suggest that the LPAR6 plays an important role in recruitment and regulation of effective T cells infiltrating in LIHC, leading to a better prognosis.
Supporting information
Supplementary Fig. 1
Correlation of LPAR6 expression with prognostic values in diverse types of cancer.
OS and DFS curves comparing the high and low expression of LPAR6 in adrenocortical carcinoma (ACC) (A–B), BLCA (C–D), BRCA (E–F), CESC (G–H), CHOL (I–J), COAD (K–L), DLBC (M–N), ESCA (O–P), GBM (Q–R), HNSC (S–T), KICH (U–V), KIRC (W–X), KIRP (Y–Z), LAML (AA–AB), LGG (AC–AD), LIHC (AE–AF), LUAD (AG–AH), LUSC (AI–AJ), MESO (AK–AL), OV (AM–AN), PAAD (AO–AP), PCPG (AQ–AR), PRAD (AS–AT), READ (AU–AV), SARC (AW–AX), SKCM (AY–AZ), STAD (BA–BB), TGCT (BC–BD), THCA (BE–BF), THYM (BG–BH), UCEC (BI–BJ), UCS (BK–BL), and UVM (BM–BN). OS, overall survival; DFS, disease-free survival; BLCA, bladder urothelial carcinoma; BRCA, breast invasive carcinoma; CESC, Cervical squamous cell carcinoma and endocervical adenocarcinoma; CHOL, cholangiocarcinoma; COAD, colon adenocarcinoma; DLBC, Lymphoid Neoplasm Diffuse Large B-cell Lymphoma; ESCA, esophageal carcinoma; GBM, Glioblastoma multiforme; HNSC, head and neck cancer; KICH, kidney chromophobe; KIRC, kidney renal clear cell carcinoma; KIRP, kidney renal papillary cell carcinoma; LGG, lower grade glioma; LIHC, liver hepatocellular carcinoma; LUAD, lung adenocarcinoma; LUSC, Lung squamous cell carcinoma; OV, Ovarian serous cystadenocarcinoma; PCPG, Pheochromocytoma and Paraganglioma; PRAD, prostate adenocarcinoma; READ, rectum adenocarcinoma; SARC, Sarcoma; SKCM, skin cutaneous melanoma; STAD, Stomach adenocarcinoma; TGCT, Testicular Germ Cell Tumors; THCA, thyroid carcinoma; THYM, Thymoma; UCEC, uterine corpus endometrial carcinoma; UVM, Uveal Melanoma.
(PDF)
Supplementary Fig. 2
Correlation of the expression of LPAR6 with immune infiltration level in cancers.
(PDF)
Supplementary Table 1
LPAR6 expression in the Oncomine database.
(DOCX)
Supplementary Table 2
Relation between LPAR6 expression and patient progonsis of different cancer in Prognoscan database.
(XLSX)
Supplementary Table 3
Correlation analysis between LPAR6 and related genes and markers of immune cells in tumor and normal tissues in LIHC.
(DOCX)