CCL17/CCL22 Predict the Survival and Response to Immune Checkpoint Blockade Therapy of Patients with HNSCC

doi:10.21203/rs.3.rs-936585/v1

Download PDF

Primary research

CCL17/CCL22 Predict the Survival and Response to Immune Checkpoint Blockade Therapy of Patients with HNSCC

https://doi.org/10.21203/rs.3.rs-936585/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Background: Head and neck squamous cell carcinoma (HNSCC) was considered as an immunosuppressive disease. Cross-talk between cancer cells and immune cells is modulated in part by chemokines CC ligands (CCL). However, the prognostic value and function of CCL in HNSCC have not been clarified.

Methods: To identify the key members of CCL family, CCL17 and CCL22 were selected through co-expression analysis. We further evaluated the correlation between the expression of CCL17/CCL22 and cancer immune infiltration through TIMER and scRNA-seq data. TIDE database was used to evaluate whether these molecules can be used as targets for immunotherapy.

Results: The results indicated high expression of CCL17 and CCL22 were associated with a better prognosis. CCL17/CCL22 might regulate the T cells activation and were positively correlated with the CD4+ T cells infiltration levels. CCL17/CCL22 can also act as a predictive biomarker for the response to immune checkpoint blockade in HNSCC patients.

Conclusions: High expression of CCL17/CCL22 may predict a better response of immune checkpoint blockade therapy in HNSCC. These findings may elucidate the roles played by CCL17/CCL22 in HNSCC progression and identify a new strategy in promoting collaborative activities in the context of immunotherapy.

Cancer Biology

HNSCC

CCL

immune microenvironment

immune checkpoint blockade

single-cell RNA sequence

Head and neck squamous cell carcinoma (HNSCC) is the most common malignant lesion in the head and neck region, with and incidence of 600,000 cases worldwide, with 40–50% mortality[1]. Treatment of HNSCC has evolved significantly over the past half century. With the great improvement of whole genome sequencing, a more detailed understanding of the molecular pathogenesis of HNSCC was made possible. Despite of the great progress in comprehensive treatment, the five-year survival rate for HNSCC patients still remains around 50%[2].

HNSCC was considered as an immunosuppressive disease and the progression and recurrence was associated with immune dysfunction[3, 4]. The expression of immune checkpoint ligands (e.g. PD-L1/2) by cancer cells can lead to an inactivation of T cells and suppress the immune response. Therefore, the immune checkpoint inhibitors like anti-PD-1 and anti-PD-L1 antibodies were approved by FDA in 2016 for platinum-resistant HNSCC patients[5, 6]. However, only ~ 20% patients can benefit from immune checkpoint blockade (ICB) therapy[7]. It is necessary to find a biomarker that can predict the response of ICB.

Chemokines are a large family of chemoattractant cytokines, with low molecular weight, which are classified into 4 highly conserved groups: CXC, CC, CX3C and C based on the number and location of the cysteine residues at the N-terminus of the protein. The chemokines are involved in varieties of biological processes in cancer including angiogenesis, tumor growth metastasis[8]. CC chemokine ligands (CCL) share characteristic is the N-terminal CC domain and play an important role in recruiting immune cells. The properties of CCL can be divided to two sides, including anti-cancer and pro-cancer. CCL2 can bind to CCR2 and recruit type 1 cytotoxic γδT lymphocytes to tumor beds, which can destroy cancer cells[9]. In turn, some CCL chemokines can recruit tumor promoting cells, such as regulatory T cells, and promote the proliferation, migration or invasion of cancer cells[10–12].

In HNSCC, CCL family members also play a different role in promoting or suppressing cancer by regulating the immune microenvironment. Many researchers suggest that most of CCL family members will promote the malignant phenotype of HNSCC. For example, CCL2 can drive radioresistance-associated metastasis in nasopharyngeal carcinoma[13], CCL4 can increase VEGF-C expression and promoted lymphangiogenesis in oral cancer[14]. At the same time, some CCL members were reported to be a tumor-suppressor gene in HNSCC. CCL28 can induce RARβ expression and inhibit oral squamous cell carcinoma bone invasion[15]. As a key regulator of immune cells in HNSCC, limited studies about the relationship between CCL family and immune therapy in HNSCC were reported.

In this study, aiming to find effective biomarkers for prognosis and prediction of immune therapy response, we collected datasets and RNA-seq data to determine the expression levels of CCL family members in HNSCC patients by using available data in the public database. We then conducted a step-by-step analysis of the clinical correlation and functions of CCL family members in HNSCC, and verified their correlation with immune cells expression and immunotherapy response through single-cell datasets and online databases.

Data Acquisition and Processing

Gene expression data, including the count and fragments per kilobase of transcript per million mapped reads (FPKM) type, and related clinical data of HNSCC patients were downloaded from UCSC Xena online database(http://xena.ucsc.edu/). The FPKM data were transformed into transcripts per million reads for further analyses. The scRNA-seq data of GSE103322[16] was downloaded from GEO database (http://www.ncbi.nlm.nih.gov/). This study met the publication guidelines described by TCGA. The data used in this study were obtained from TCGA; hence, ethics approval and informed consent were not required.

Oncomine database analysis

Oncomine (www.oncomine.org) is a powerful translation bioinformatics service that provides genome-wide expression analysis[17]. Data were extracted to evaluate the expression of CCL family members in HNSCC. In our study, the cut-off p-value and

fold-change were as follows: p-value: 0.01, fold-change: 2, gene rank:10%. Only the analysis of gene expression between tumor tissues and normal precancerous tissues is in line with the conditions for further analysis.

cBioPortal analysis

cBioPortal (www.cbioportal.org) is a comprehensive web tools, which can visualize and analyze multidimensional cancer genomics data[18]. Based on TCGA database, genetic alterations of CCL chemokines and correlated neighboring genes of CCL17 or CCL22 in HNSCC were obtained from cBioPortal.

Functional analysis of correlated neighboring genes

Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses of correlated neighboring genes were conducted using enrichGO and enrichKEGG functions of clusterProfiler R package[19]. Adjust P values (false-discovery rate [FDR]) less than 0.05 were considered to be statistically significant.

GeneMANIA

GeneMANIA (http://www.genemania.org) was used to construct protein-protein interaction networks (PPI), pathways, co-expression, co-localization, and protein domain similarity of submitted genes. [20].

TRRUST

TRRUST (https://www.grnpedia.org/trrust/) is a powerful web tool for predicting the trascriptional regulatory networks in human and mouse[21]. The TRRUST database contains 8444 transcription factor (TF) target regulatory relationships among 800 human TFs, which can provide information regarding how these interactions are regulated.

STRING database analysis

STRING database (https://string-db.org) was used to predict protein interaction network. A confidence value of greater than 0.7 was set for interacted proteins selected.

Tumor Purity analysis

ESTIMATE is a method that uses gene expression signatures to infer the fraction of stromal and immune cells in tumour samples[22]. The tumor purity of TCGA samples was calculated by ESTIMATE package in R.

TIMER database analysis

TIMER (http://timer.cistrome.org/) is a reliable and intuitive tool for evaluating immune infiltration levels from TCGA datasets[23]. The immune infiltration level scores of TCGA samples were downloaded from TIMER for further analyzing.

Single-cell sequence data analysis

The expression matrix of scRNA-seq of GSE103322 has already normalized by the researcher of GSE103322. The matrix was combined in R and converted to a Seurat object using Seurat R package[24]. To reduce dimensionality, principal component analysis was used to summarize the resulting variably expressed genes, and then the t-SNE dimensionality reduction (RunTSNE function) was used to further summarize the principal components. The clusters of each cell type were annotated by the expression of following genesets: CD4+ T cells (CD4, IL7R), Tregs (CD4, FOXP3, IL2RA), CD8+ Tcells (CD8A, CD8B), CD8+ dysfunction T cells (CD8A, CD8B, PDCD1, CTLA4, LAG3), Myleoid cells ( TPSB2, CD1A, CD14, CD163, C1QA, TREM1), Fibroblasts (COL1A2, DCN), NK cells (NCAM1, KLRD1, KLRC1, XCL1), Endothelial cells, (PECAM1, VWF) and B cells (CD79A, CD19). The expression of CCL17 or CCL22 were calculated by sum the expression in each cell types.

Differentially Expressed Genes (DEGs) Analysis

The cells were divided into high-expression and low-expression groups according to the level of CCR7 (CD4+ T cells) and CCR8 (Treg cells) expression (median as cut-off value). Then expression data were compared between high and low FOXD2-AS1 expression groups to identify differentially expressed genes (DEGs) using the limma package in R. log2FoldChange >1 were set as threshold values for DEGs[25].

Tumor immune dysfunction and exclusion (TIDE)

TIDE (http://tide.dfci.harvard.edu) is a computational framework developed to evaluate the potential of tumor immune escape from the gene expression profiles of cancer samples[26, 27]. We used TIDE to analyze the relationship between the expression of CCL17 or CCL22 and immune-related phenotypes and to analyze the prognosis of CCL17 and CCL22 in different ICB-treated cancer types.

Statistical analysis and data visualization

Statistical analysis was performed using R (4.0.2). Gene expression of CCL family members was compared by using Wilcoxon rank-sum tests. The association between CCL family members and survival were analyzed by Cox regression models. The relationships between clinical parameters and expression of CCL family members were analyzed using the Wilcoxon rank-sum test. Spearman correlation tests were performed to analyze the correlation between expression of each CCL members. The further correlation analysis between the expression of CCL17 or CCL22 and tumor purity, the expression of PD1 or PD-L1 or immune infiltration levels were also performed by Spearman correlation tests. The visualization of the results was used by ggplot2, pheatmap, clusterProfiler and Seurat package.

Overview of the expression and mutation of different CCL family members.

To investigate the potential prognostic value of CCL family members, the mRNA levels of CCL family members (not including CCL6, CCL9, CCL10, CCL12) in tumor and normal tissues in different cancers were analyzed according to the Oncomine database (Figure.1A). Based on the data from Oncomine, the mRNA expression levels of CCL2, CCL3, CCL4, CCL5, CCL8, CCL11, CCL13, CCL18, CCL20 and CCL24 were significantly elevated in HNSCC tumor tissues, while the expression levels of CCL14, CCL15, CCL16, CCL19 and CCL21 were reduced (Table S1). Next we analyzed the data from TCGA HNSCC project. The genetic alternation status was shown in Figure.1B. Except CCL24 and CCL28, the mutation rate of other CCL family members was less than 10%. The expression level of CCL family members between tumor and normal tissues in TCGA HNSCC samples were compared and the results showed that CCL3, CCL4, CCL5, CCL7, CCL8, CCL11, CCL13, CCL18, CCL20, CCL22, CCL24, CCL26 were upregulated in tumor tissues, while CCL2, CCL14, CCL15, CCL16, CCL17, CCL19, CCL23, CCL28 were significantly reduced in tumor tissue (Figure.1C). The result from TCGA was basically consistent with the one from Oncomine. The expression status of CCL family members in pan-cancer from TCGA datasets was shown in Fig.S1 and Fig.S2. The expression and mutation overview of CCL family members indicated that CCL family members had different functions in HNSCC.

Prognostic role of CCL family members in TCGA HNSCC samples.

To evaluate the prognostic value of differentially expressed CCL family members in HNSCC, univariate Cox regression model was used to analyzed the correlation between the expression of CCL family members and overall survival in TCGA HNSCC samples. The results were showed in Figure.2. In the forest plot, the items with p value less than 0.1 were selected and bold-shown. Among CCL family members, CCL1, CCL5, CCL11, CCL14, CCL17, CCL20, CCL22, CCL25, CCL28 were associated with the prognosis. Further Kaplan-Meier curves were performed to verify the association between a high/ low expression of CCL family members and overall survival. The K-M curves of nine CCL family members mentioned above were shown in Figure.3A, while other members were shown in Fig.S3.

A strong co-expression relationship between CCL17 and CCL22 among CCL family members.

To further identify the key genes in nine CCL family members, we performed the correlation test to explore the co-expression relationship between each CCL family member. The correlation coefficient was clustered and visualized in the heat map (Figure.3B). In the heat map we found CCL17 was adjacent to CCL22, while CCL20 was adjacent to CCL25 (Blue and red boxes in heat map). The scatter plot of the expression of CCL17 and CCL22 as well as CCL20 and CCL25 was shown in Figure.3C and D. Based on the correlation test, CCL17 had a strong correlation with CCL22 (Spearman R = 0.620, p value < 0.001), while CCL20 had a weak correlation with CCL25 (Spearman R = 0.088, p value = 0.05). Considering the correlation with prognosis and co-expression relationship, we identified CCL17 and CCL22 as the key genes in HNSCC. We further analyzed the correlation between CCL17 and CCL22 expression levels with clinical parameters. The lower expression level of CCL22 was associated with the more advanced clinical stage of HNSCC, while there was no correlation between the expression of CCL17 and clinical stage (Figure.3E). A more detailed analysis of clinical parameters (including T stage, N stage, pathological gradie, and lympho Vascular invasion) was shown in Fig.S4. We also explored the diagnostic capability of CCL17 and CCL22. However, both of two genes had a poor capability in diagnosis (Figure.3F). The results above screened out CCL17 and CCL22 for a further analysis.

Functional Enrichment Analysis of CCL17 and CCL22 correlated genes in TCGA HNSCC samples

Neighboring genes correlated with CCL17 and CCL22 were first selected according to the cBioPortal database. The correlated neighboring genes of CCL17 and CCL22 were shown in Table S2 and Table S3. GO and KEGG functional analysis was performed by clusterProfiler in R and the results were shown in Figure.4A and C. The detailed items of functional analysis were displayed in Table S4 and Table S5. We also used GeneMANIA to construct molecular regulatory networks between correlated genes (Figure.4B and D). The functional analysis results indicated that the correlated genes were mostly associated with the immune cell activation but not cancer-associated pathways. To further identify the upstream and downstream molecules of CCL17 and CCL22, TRRUST database was used to explore the upstream factors while STRING database was used to identify the downstream molecules. Several transcriptional factors targets CCL family members were predicted by TRRUST (Table 1). Among them, NFKB1 and RELA were predicted to target CCL22, while STAT6 was predicted to target CCL17. The downstream receptors including CCR1, CCR2, CCR3, CCR4, CCR5, CCR7, CCR8 and CXCR3 were predicted by STRING database as the interaction score was set as 0.70. CCR2, CCR4, CCR7, CCR8 and CXCR3 were the targets for both CCL17 and CCL22. The regulation network of CCL17 and CCL22 was shown in Figure.4E. The scatter plot between CCL17 or CCL22 and CCR2, CCR4, CCR7, CCR8 and CXCR3 were shown in Fig.S5. The results of functional analysis suggested that CCL17 and CCL22 might mainly function in immune cells rather than cancer cells.

Table.1

Key regulated transcriptional factor of CCL chemokines in HNSC (TRRUST)

Key TF	Description	Regulate gene	P value	Q value
RELA	v-rel reticuloendotheliosis viral oncogene homolog A (avian)	CCL4,CCL19,CCL20,CCL22,CCL2, CCL3,CCL11,CCL5,CCL13	6.25E-11	3.32E-10
NFKB1	nuclear factor of kappa light polypeptide gene enhancer in B-cells 1	CCL5,CCL3,CCL20,CCL11,CCL13, CCL4,CCL22,CCL19,CCL2	6.63E-11	3.32E-10
STAT6	signal transducer and activator of transcription 6, interleukin-4 induced	CCL11,CCL17,CCL26	1.26E-05	4.18E-05
IRF3	interferon regulatory factor 3	CCL2,CCL5	0.000161	0.000403
REL	v-rel reticuloendotheliosis viral oncogene homolog (avian)	CCL2,CCL5	0.000353	0.000705
SPI1	spleen focus forming virus (SFFV) proviral integration oncogene spi1	CCL2,CCL5	0.0028	0.00466
STAT1	signal transducer and activator of transcription 1, 91kDa	CCL2,CCL3	0.00507	0.00725
STAT3	signal transducer and activator of transcription 3 (acute-phase response factor)	CCL11,CCL2	0.0139	0.017
JUN	jun proto-oncogene	CCL5,CCL2	0.0153	0.017
SP1	Sp1 transcription factor	CCL2,CCL5	0.12	0.12

CCL17 and CCL22 expression correlated with immune infiltration levels in HNSCC

As mentioned above, CCL17 and CCL2 might affect immune cells in HNSCC patients. Therefore, we used TIMER database and ESITMATE algorithm to focus on the relationship between CCL17 or CCL22 and immune infiltration cells in HNSCC samples. First, tumor purity of samples in TCGA HNSCC project was calculated by using ESTIMATE algorithm in R. The ESTIMATE scores and purity scores were displayed in Table S6. Correlation test between CCL17 or CCL22 expression and tumor purity scores were performed using Spearman correlation test. As shown in Figure.5A and B, the expression of CCL17 and CCL22 had a negative correlation with the tumor purity in TCGA HNSCC samples (Spearman R=-0.34, p value < 0.001 for CCL17 and Spearman R=-0.52, p value < 0.001 for CCL22). The results validated the functions of CCL17 and CCL22 as the chemokines that can recruit the immune cells, which caused the reduction of the tumor purity in tumor. Then the immune infiltration levels of each sample were calculated by TIMER database (Table S7). Spearman correlation test was performed between the expression of CCL17 or CCL22 and each immune cells infiltration score. The correlation R was displayed in Figure.5C and D. At the same time, we performed the Spearman correlation test between the expression of CCL17 or CCL22 and the expression of PD1 (PDCD1) and PD-L1 (CD274). As shown in the Figure.5E and F, CCL17 expression was positively correlated with PD1 (PDCD1) expression, but had not correlation with PD-L1 (CD274). The expression of CCL22 was both positively correlated with PD1 (PDCD1) and PD-L1 (CD274) in TCGA HNSCC samples (Spearman R=-0.38, p value < 0.001 for PD1 and Spearman R=-0.19, p value < 0.001 for PD-L1) (Figure.5G and H).

High expression of CCL22 recruited more CD4 + T cells according to scRNA-seq data

We downloaded the scRNA-seq data of GSE103322 from GEO database. Cancer cells were first excluded and rest of cells were re-clustered into 9 groups (Figure.6A, B and C). The expression and distribution of CCL17 and CCL22 were explored in each cell type The t-SNE plots indicated that CCL17 and CCL22 were mainly expressed in the endothelial cells (Figure.6D and E). Then we calculated the sum of CCL17 and CCL22 expression of each sample. The samples were divided into CCL17/CCL22 high expression group and low expression group according to the median expression. Then we counted the cell numbers of each type of T cells including CD4 + T cells, Tregs, CD8 + T cells and CD8 + dysfunction T cells. Finally, only CD4 + T cells had a different expression between high and low CCL22 expression group (p < 0.05) (Figure.6F and Table S8). The results were consistent with the analysis of TCGA samples which confirmed that CCL22 can increase the CD4 + T cells infiltration in HNSCC patients.

High expression of CCR7 activated the mTORC1 signaling in CD4 + T cells

The expression and distribution of receptors of CCL17 and CCL22 (including CCR2, CCR4, CCR7, CCR8 and CXCR3) were shown in Fig. 7A, which suggested that CCR7 and CCR8 were the main receptors in HNSCC. These two receptors were mainly located in T cells. Besides, since CCR7 and CCR8 were the major receptors in HNSCC, we further investigate the potential signaling pathway of CCR7 and CCR8 in CD4 + T cells and Treg cells, respectively. As shown in Fig. 7B and C, CCR7 was mainly expressed in CD4 + T cells and CD8 + T cells, while CCR8 was expressed in Treg cells. In CD4 + T cells, we divided cells into high or low CCR7 expression group, while in Treg cells, we divided cells into high or low CCR7 expression group. Differentially expressed genes (DEGs) were analyzed in two cell types and were visualized using volcano plot (Fig. 7D and E). The detailed information of DEGs was shown in Table S9. GO analysis was performed on the DEGs (Fig. 7F and G). To identify the different pathways, GSEA analysis showed in CD4 + T cells, high CCR7 expression cells were mainly enriched in MYC-targets and MTORC1 signaling (Fig. 7H). In Treg cells, high CCR8 expression cells were mainly enriched in IL6-JAK-STAT3 signaling and KRAS signaling DN (Fig. 7I). The detailed items of GSEA were shown in Table S10

Prediction of clinical response to immune checkpoint blockade (ICB) through CCL17/CCL22 expression

Immune checkpoint blockade therapy has exhibited a great improvement in the treatments of a variety of cancers. Since we have found that CCL17 and CCL22 had a recruitment effect on T cells, we next explored whether CCL17 and CCL22 could be used to predict the response of ICB therapy. TIDE database was used to investigate the role of CCL17 and CCL22 played in cancer ICB therapy. The results suggested that in different type of cancer, CCL17 and CCL22 had a different effect on the T cell dysfunction phenotype (Figure.8A left panel). The high expression of CCL17 and CCL22 was associated with a good prognosis in most cancer ICB therapies (Figure.8A second to left panel. The red color means a poor ICB response and prognosis, while blue means a good ICB response and prognosis). In clinical studies of blocking specific immune checkpoints, the relationship between gene expression of CCL17/CCL22 and treatment outcomes is shown in Table S11. A Risk scores calculated by TIDE algorithms based on a variety of datasets were shown in Figure.8B and Table S12. We then extracted the part of HNSCC datasets and found that based on TCGA datasets, high CCL17 and CCL22 expression had a lower risk scores which means a better response to ICB therapy. However, the role of CCL17 and CCL22 in other datasets were conflicted with TCGA datasets (Figure.8C). Finally, we used the function of “Biomarker Evaluation” in TIDE to see the prediction power of CCL17 and CCL22. This function can calculated the AUC value of biomarkers based on the data from a study of immunotherapy (PD1) on a mouse oral squamous cell carcinoma model[28]. Compared with several biomarkers currently used, we found a combination of CCL17, CCL22, PD-L1 (CD274) and PD1 (PDCD1) had a better prediction power than other biomarkers in pre-surgical groups (Figure.8D. The AUC of combination was 0.69). Therefore, according to the analysis results, CCL17 and CCL22 might be selected as potential biomarkers for prediction of HNSCC ICB therapy.

CC chemokine ligands (CCL) are chemotactic cytokines with an N-terminal CC domain and play an important role in immune system cells, such as CD4 + and CD8 + T cells, myeloid cells and NK cells, as well in neoplasia. Previous studies have shown that many CCL family members were abnormally expressed in HNSCC samples[13, 29–32]. Thus, combining the results of Oncomine and TCGA datasets, we found that most CCL family members’ expression, excluding CCL1, CCL21 and CCL27, were increased or decreased in HNSCC patients.

Next we used the clinical data to explore the prognostic value of abnormally expressed CCL family members. By calculating the overall-survival rate of each CCL family members by Cox regression model, we further selected 9 key CCL molecules that might have effects on the prognosis of HNSCC patients. According to the result, CCL1, CCL5, CCL11, CCL14, CCL17, CCL22, CCL 25 and CCL28 were protective factors of the prognosis in HNSCC patients. Only CCL20 was a risk factor that might be related with poor survival. CCL20 was confirmed by many researches and identified as a tumor promoting factor in many cancer types[33, 34], which was consistent with the result in our study.

The functions of CCL14 and CCL28 varied from tumor to tumor. In hepatocellular carcinoma and colorectal cancer, CCL14 was considered to be a tumor suppressor[35, 36] while in breast cancer it can promote angiogenesis and metastasis[37]. CCL28 was reported to promote breast cancer growth and metastasis via MAPK-mediated cellular anti-apoptosis and pro-metastasis mechanism[38]. Another study, however, indicated that CCL28 can inhibit the bone invasion in oral squamous cell carcinoma[15].

Some studies suggested that CCL1 can recruit Treg cells and promote stemness property of cancer cells in breast cancer[39, 40], while CCL5 can promote tumor progression in both solid tumors and hematological malignancies[41]. CCL11 and CCL25 were also reported to be oncogenes in several types of cancer[42–45]. CCL17 and CCL22 were both the ligands of CCR4, which was expressed on the surface of Treg cells. They also participated in the differentiation of Th2 cells. Most studies indicated CCL17 and CCL22 could promote cancer progression by recruiting Treg cells and suppress the immune response[46–49], which means CCL17 and CCL22 were both oncogenes.

In this way, the mechanism of the CCL family members in cancers is quite complicated. Moreover, there are still few studies in HNSCC. In order to further determine the molecules that play an important role in HNSCC, further analysis was performed to screened the above CCL members.

Co-expression analysis was performed to find the CCL members with similar functions. Finally, we selected CCL17 and CCL22, which had a relatively high correlation with Spearman R = 0.62. Further analysis of the correlation between CCL17/CCL22 and clinical parameters was performed and the results indicated a low CCL22 expression was associated with the advanced stage in HNSCC. We both investigated the diagnostic power of these two genes and found that both of them cannot be used in the diagnosis, which suggested that CCL17/CCL22 might have no association with the occurrence of HNSCC.

As we mentioned that these two molecules were both the ligands of CCR4 and can regulate the differentiation and recruitment of T cells. We analyzed the functions of neighboring genes with CCL17/CCL22. The results suggested these neighboring genes were enriched in the activation of immune cells. Except CCR4, CCR2, CCR7, CCR8 and CXCR3 were also the receptors for both of CCL17 and CCL22. Correlation test between CCL17/CCL22 and their receptors also proved their positive correlation in HNSCC. We wonder whether CCL17/CCL22 were associated with the tumor purity or immune cells infiltration levels, TIMER database and ESTIMATE were used to calculated the immune cells expression and tumor purity. As we expected, the expression of CCL17 and CCL22 was positively correlated with the expression of CD4 + T cells.

To validate the findings from TCGA datasets, we analysis a scRNA-seq data of HNSCC[16]. We first checked the expression and distribution of receptors in these data and found that CCR7 and CCR8 were the major receptors in HNSCC samples. And these two receptors were mainly located in T cells, including CD4 + T cells, Treg cells and CD8 + T cells. This results suggested that CCL17/CCL22 might mainly regulate T cells by binding to CCR7 or CCR8 in HNSCC.

Some studies suggested the activation of CCR7 signaling can promote HNSCC progression and metastasis via MAPK, JAK-STAT3 or PI3K pathway[50–52]. However, these studies mainly explored the phenotype changes caused by overexpression of CCR7 in tumor cells. A study on the function of CCR7 in CD8 + T cells of patients with HNSCC confirmed CCR7 signaling can protect effector T cells from apoptosis through the phosphatidylinositol 3-kinase/Akt pathway[53]. In our study, we found a high expression of CCR7 in CD4 + T cells can activate the mTORC1 signaling pathway, which can regulate the differentiation of CD4 + T cells into T helper 1 or T helper 17 cells[54]. These two kinds of Th cells can activate the immune response and kill tumors. Whats’more, high expression of CCR7 in CD4 + T cells can also activate MYC target signaling, which is reported to promote the activation and proliferation of CD4 + T cells[55]. The results give us a new sight that the interaction between CCL17/CCL22 and CCR7 might be a factor that promotes immune response in HNSCC.

Till now, there is no study on the function of CCR8 in HNSCC. However, many studies have verified a subgroup of Treg cells, which had a highly expressed CCR8, in breast, colon and lung cancer[56–58]. Blockade of CCR8 in these cancer can suppress the function of Treg cells and inhibit the tumor growth. As we found in our study, CCR8 was mainly expressed in Treg cells. And a high expression of CCR8 was related with the IL6-JAK-STAT3 signaling and the down-regulation of KRAS signaling. The functions of these two pathways in Tregs cells were not clear at now. Perhaps CCL17 and CCL22 can acted as a blockade molecules and inhibit the activation of CCR8 downstream signaling pathways and reverse the immunosuppression caused by Treg cells.

Except the analysis on receptors, we also explored the number of T cells in different CCL17/CCL22 expression group. The results was consistent with the finding in TCGA data that the number of CD4 + T cells were increased when the CCL17 and CCL22 was highly expressed. The results indicated that CCL17 and CCL22 can recruit CD4 + T cells in HNSCC.

Finally, we used TIDE to evaluate the response of CCL17/CCL22 in the immune checkpoint blockade (ICB) treatment. The results showed that high expression of CCL17 and CCL22 was associated with a good prognosis in most cancer ICB trials, which also proved that CCL17 and CCL22 acted as a factor that can promote immune response. However, the prediction of ICB response in HNSCC was not consistent among different datasets. In TCGA datasets, CCL17 and CCL22 were predicted to be a positive marker to ICB response, which was consistent to our previous conclusion. While in other datasets from GEO or PRECOG, the role of CCL17 and CCL22 was not clear. We believe that the differences between different datasets may be caused by other regulatory factors in the immune microenvironment. Further analysis and clarification of these key factors in the tumor microenvironment may provide a broader understanding of the research on immunotherapy in HNSCC. At last we confirmed the predictive power of CCL17 and CCL22 as a biomarker of ICB response. CCL17 or CCL22 alone have poor diagnostic efficiency as a biomarker, but the combination of CCL17, CCL22, PD1, and PD-L1 greatly improves the predictive ability, and is equal to the predictive ability of some existing predictive markers.

This study has several limitations that need to be addressed. First, although this study combines several database or datasets, most of these datasets are collected retrospectively. In addition, although current studies have obtained useful findings, more laboratory studies and large-scale clinical trials are needed in the future to prove our findings.

Based on our study, we believe that high expression of CCL17 and CCL22 is associated with the good survival in HNSCC. Besides, CCL17 and CCL22 may promote the immune response in patients with HNSCC by regulating the functions and recruitment of T cells. At the same time, CCL17 and CCL22 might play an important role in ICB treatment and be a potential biomarker that can predict the response of ICB treatment.

Ethics approval and consent to participate

Informed consent for participating in the clinical study was signed by all involved patients. The ethics approval was granted by the Ethical Committee of the Shanghai Ninth People’s Hospital affiliated to Shanghai Jiaotong University School of Medicine.

Consent for publication

All authors agree to publish.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request

Competing interests:

The authors have no conflicts of interest to declare

Funding

This work was supported by National Natural Science Foundation of China (81903056). The funders had no role in study design, data collection, analysis, and interpretation; preparation of the manuscript or decision to submit the article for publication.

Authors’ contributions

ZQL and YZ designed the study. WKZ, XZ collected and analyzed the data and wrote the main manuscript. XHL and YL performed the statistical analysis. ZQL and YZ did the manuscript review. All authors reviewed and approved the final version of the manuscript.

Acknowledgments

The authors thank the study participants for their involvement in each of the individual studies.

Ferlay J, Soerjomataram I, Dikshit R, Eser S, Mathers C, Rebelo M, Parkin DM, Forman D, Bray F: Cancer incidence and mortality worldwide: sources, methods and major patterns in GLOBOCAN 2012. Int J Cancer2015, 136(5):E359-386.
Shield KD, Ferlay J, Jemal A, Sankaranarayanan R, Chaturvedi AK, Bray F, Soerjomataram I: The global incidence of lip, oral cavity, and pharyngeal cancers by subsite in 2012. CA: a cancer journal for clinicians2017, 67(1):51-64.
Varilla V, Atienza J, Dasanu CA: Immune alterations and immunotherapy prospects in head and neck cancer. Expert Opin Biol Ther2013, 13(9):1241-1256.
Economopoulou P, Perisanidis C, Giotakis EI, Psyrri A: The emerging role of immunotherapy in head and neck squamous cell carcinoma (HNSCC): anti-tumor immunity and clinical applications. Ann Transl Med2016, 4(9):173.
Seiwert TY, Burtness B, Mehra R, Weiss J, Berger R, Eder JP, Heath K, McClanahan T, Lunceford J, Gause Cet al: Safety and clinical activity of pembrolizumab for treatment of recurrent or metastatic squamous cell carcinoma of the head and neck (KEYNOTE-012): an open-label, multicentre, phase 1b trial. Lancet Oncol2016, 17(7):956-965.
Ferris RL, Blumenschein G, Jr., Fayette J, Guigay J, Colevas AD, Licitra L, Harrington K, Kasper S, Vokes EE, Even Cet al: Nivolumab for Recurrent Squamous-Cell Carcinoma of the Head and Neck. N Engl J Med2016, 375(19):1856-1867.
Zandberg DP, Algazi AP, Jimeno A, Good JS, Fayette J, Bouganim N, Ready NE, Clement PM, Even C, Jang RWet al: Durvalumab for recurrent or metastatic head and neck squamous cell carcinoma: Results from a single-arm, phase II study in patients with ≥25% tumour cell PD-L1 expression who have progressed on platinum-based chemotherapy. Eur J Cancer2019, 107:142-152.
Nagarsheth N, Wicha MS, Zou W: Chemokines in the cancer microenvironment and their relevance in cancer immunotherapy. Nat Rev Immunol2017, 17(9):559-572.
Lança T, Costa MF, Gonçalves-Sousa N, Rei M, Grosso AR, Penido C, Silva-Santos B: Protective role of the inflammatory CCR2/CCL2 chemokine pathway through recruitment of type 1 cytotoxic γδ T lymphocytes to tumor beds. J Immunol2013, 190(12):6673-6680.
Dagouassat M, Suffee N, Hlawaty H, Haddad O, Charni F, Laguillier C, Vassy R, Martin L, Schischmanoff PO, Gattegno Let al: Monocyte chemoattractant protein-1 (MCP-1)/CCL2 secreted by hepatic myofibroblasts promotes migration and invasion of human hepatoma cells. Int J Cancer2010, 126(5):1095-1108.
Mou T, Xie F, Zhong P, Hua H, Lai L, Yang Q, Wang J: MiR-345-5p functions as a tumor suppressor in pancreatic cancer by directly targeting CCL8. Biomed Pharmacother2019, 111:891-900.
Low-Marchelli JM, Ardi VC, Vizcarra EA, van Rooijen N, Quigley JP, Yang J: Twist1 induces CCL2 and recruits macrophages to promote angiogenesis. Cancer Res2013, 73(2):662-671.
Guo SS, Liu R, Wen YF, Liu LT, Yuan L, Li YX, Li Y, Hao WW, Peng JY, Chen DNet al: Endogenous production of C-C motif chemokine ligand 2 by nasopharyngeal carcinoma cells drives radioresistance-associated metastasis. Cancer Lett2020, 468:27-40.
Lien MY, Tsai HC, Chang AC, Tsai MH, Hua CH, Wang SW, Tang CH: Chemokine CCL4 Induces Vascular Endothelial Growth Factor C Expression and Lymphangiogenesis by miR-195-3p in Oral Squamous Cell Carcinoma. Front Immunol2018, 9:412.
Park J, Zhang X, Lee SK, Song NY, Son SH, Kim KR, Shim JH, Park KK, Chung WY: CCL28-induced RARβ expression inhibits oral squamous cell carcinoma bone invasion. J Clin Invest2019, 129(12):5381-5399.
Puram SV, Tirosh I, Parikh AS, Patel AP, Yizhak K, Gillespie S, Rodman C, Luo CL, Mroz EA, Emerick KSet al: Single-Cell Transcriptomic Analysis of Primary and Metastatic Tumor Ecosystems in Head and Neck Cancer. Cell2017, 171(7):1611-1624.e1624.
Rhodes DR, Yu J, Shanker K, Deshpande N, Varambally R, Ghosh D, Barrette T, Pandey A, Chinnaiyan AM: ONCOMINE: a cancer microarray database and integrated data-mining platform. Neoplasia (New York, NY)2004, 6(1):1-6.
Gao J, Aksoy BA, Dogrusoz U, Dresdner G, Gross B, Sumer SO, Sun Y, Jacobsen A, Sinha R, Larsson Eet al: Integrative analysis of complex cancer genomics and clinical profiles using the cBioPortal. Science signaling2013, 6(269):pl1.
Yu G, Wang LG, Han Y, He QY: clusterProfiler: an R package for comparing biological themes among gene clusters. Omics : a journal of integrative biology2012, 16(5):284-287.
Warde-Farley D, Donaldson SL, Comes O, Zuberi K, Badrawi R, Chao P, Franz M, Grouios C, Kazi F, Lopes CTet al: The GeneMANIA prediction server: biological network integration for gene prioritization and predicting gene function. Nucleic acids research2010, 38(Web Server issue):W214-220.
Han H, Cho JW, Lee S, Yun A, Kim H, Bae D, Yang S, Kim CY, Lee M, Kim Eet al: TRRUST v2: an expanded reference database of human and mouse transcriptional regulatory interactions. Nucleic acids research2018, 46(D1):D380-d386.
Yoshihara K, Shahmoradgoli M, Martínez E, Vegesna R, Kim H, Torres-Garcia W, Treviño V, Shen H, Laird PW, Levine DAet al: Inferring tumour purity and stromal and immune cell admixture from expression data. Nature communications2013, 4:2612.
Li T, Fan J, Wang B, Traugh N, Chen Q, Liu JS, Li B, Liu XS: TIMER: A Web Server for Comprehensive Analysis of Tumor-Infiltrating Immune Cells. Cancer Res2017, 77(21):e108-e110.
Stuart T, Butler A, Hoffman P, Hafemeister C, Papalexi E, Mauck WM, 3rd, Hao Y, Stoeckius M, Smibert P, Satija R: Comprehensive Integration of Single-Cell Data. Cell2019, 177(7):1888-1902.e1821.
Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, Smyth GK: limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic acids research2015, 43(7):e47.
Jiang P, Gu S, Pan D, Fu J, Sahu A, Hu X, Li Z, Traugh N, Bu X, Li Bet al: Signatures of T cell dysfunction and exclusion predict cancer immunotherapy response. Nature medicine2018, 24(10):1550-1558.
Fu J, Li K, Zhang W, Wan C, Zhang J, Jiang P: Large-scale public data reuse to model immunotherapy response and resistance. 2020, 12(1):21.
Friedman J, Moore EC, Zolkind P, Robbins Y, Clavijo PE, Sun L, Greene S, Morisada MV, Mydlarz WK, Schmitt Net al: Neoadjuvant PD-1 Immune Checkpoint Blockade Reverses Functional Immunodominance among Tumor Antigen-Specific T Cells. Clin Cancer Res2020, 26(3):679-689.
Fujita S, Ikeda T: The CCL2-CCR2 Axis in Lymph Node Metastasis From Oral Squamous Cell Carcinoma: An Immunohistochemical Study. J Oral Maxillofac Surg2017, 75(4):742-749.
Tsujikawa T, Yaguchi T, Ohmura G, Ohta S, Kobayashi A, Kawamura N, Fujita T, Nakano H, Shimada T, Takahashi Tet al: Autocrine and paracrine loops between cancer cells and macrophages promote lymph node metastasis via CCR4/CCL22 in head and neck squamous cell carcinoma. Int J Cancer2013, 132(12):2755-2766.
Wang J, Seethala RR, Zhang Q, Gooding W, van Waes C, Hasegawa H, Ferris RL: Autocrine and paracrine chemokine receptor 7 activation in head and neck cancer: implications for therapy. J Natl Cancer Inst2008, 100(7):502-512.
Qin Y, Wang J, Zhu G, Li G, Tan H, Chen C, Pi L, She L, Chen X, Wei Met al: CCL18 promotes the metastasis of squamous cell carcinoma of the head and neck through MTDH-NF-κB signalling pathway. J Cell Mol Med2019, 23(4):2689-2701.
Kadomoto S, Izumi K, Mizokami A: The CCL20-CCR6 Axis in Cancer Progression. Int J Mol Sci2020, 21(15).
Liu W, Wang W, Wang X, Xu C, Zhang N, Di W: Cisplatin-stimulated macrophages promote ovarian cancer migration via the CCL20-CCR6 axis. Cancer Lett2020, 472:59-69.
Yan G, Li S, Yue M, Li C, Kang Z: Lysine demethylase 5B suppresses CC chemokine ligand 14 to promote progression of colorectal cancer through the Wnt/β-catenin pathway. Life Sci2020:118726.
Zhu M, Xu W, Wei C, Huang J, Xu J, Zhang Y, Zhao Y, Chen J, Dong S, Liu Bet al: CCL14 serves as a novel prognostic factor and tumor suppressor of HCC by modulating cell cycle and promoting apoptosis. Cell Death Dis2019, 10(11):796.
Li Q, Shi L, Gui B, Yu W, Wang J, Zhang D, Han X, Yao Z, Shang Y: Binding of the JmjC demethylase JARID1B to LSD1/NuRD suppresses angiogenesis and metastasis in breast cancer cells by repressing chemokine CCL14. Cancer Res2011, 71(21):6899-6908.
Yang XL, Liu KY, Lin FJ, Shi HM, Ou ZL: CCL28 promotes breast cancer growth and metastasis through MAPK-mediated cellular anti-apoptosis and pro-metastasis. Oncol Rep2017, 38(3):1393-1401.
Kuehnemuth B, Piseddu I, Wiedemann GM, Lauseker M, Kuhn C, Hofmann S, Schmoeckel E, Endres S, Mayr D, Jeschke Uet al: CCL1 is a major regulatory T cell attracting factor in human breast cancer. BMC Cancer2018, 18(1):1278.
Xu Y, Dong X, Qi P, Ye Y, Shen W, Leng L, Wang L, Li X, Luo X, Chen Yet al: Sox2 Communicates with Tregs Through CCL1 to Promote the Stemness Property of Breast Cancer Cells. Stem Cells2017, 35(12):2351-2365.
Aldinucci D, Colombatti A: The inflammatory chemokine CCL5 and cancer progression. Mediators Inflamm2014, 2014:292376.
Xu B, Deng C, Wu X, Ji T, Zhao L, Han Y, Yang W, Qi Y, Wang Z, Yang Zet al: CCR9 and CCL25: Areview of their roles in tumor promotion. J Cell Physiol2020, 235(12):9121-9132.
Zhang Z, Sun T, Chen Y, Gong S, Sun X, Zou F, Peng R: CCL25/CCR9 Signal Promotes Migration and Invasion in Hepatocellular and Breast Cancer Cell Lines. DNA Cell Biol2016, 35(7):348-357.
Nolen BM, Lokshin AE: Targeting CCL11 in the treatment of ovarian cancer. Expert Opin Ther Targets2010, 14(2):157-167.
Levina V, Nolen BM, Marrangoni AM, Cheng P, Marks JR, Szczepanski MJ, Szajnik ME, Gorelik E, Lokshin AE: Role of eotaxin-1 signaling in ovarian cancer. Clin Cancer Res2009, 15(8):2647-2656.
Mizukami Y, Kono K, Kawaguchi Y, Akaike H, Kamimura K, Sugai H, Fujii H: CCL17 and CCL22 chemokines within tumor microenvironment are related to accumulation of Foxp3+ regulatory T cells in gastric cancer. Int J Cancer2008, 122(10):2286-2293.
Röhrle N, Knott MML, Anz D: CCL22 Signaling in the Tumor Environment. Adv Exp Med Biol2020, 1231:79-96.
Yang P, Li QJ, Feng Y, Zhang Y, Markowitz GJ, Ning S, Deng Y, Zhao J, Jiang S, Yuan Yet al: TGF-β-miR-34a-CCL22 signaling-induced Treg cell recruitment promotes venous metastases of HBV-positive hepatocellular carcinoma. Cancer Cell2012, 22(3):291-303.
Gobert M, Treilleux I, Bendriss-Vermare N, Bachelot T, Goddard-Leon S, Arfi V, Biota C, Doffin AC, Durand I, Olive Det al: Regulatory T cells recruited through CCL22/CCR4 are selectively activated in lymphoid infiltrates surrounding primary breast tumors and lead to an adverse clinical outcome. Cancer Res2009, 69(5):2000-2009.
Liu FY, Safdar J, Li ZN, Fang QG, Zhang X, Xu ZF, Sun CF: CCR7 regulates cell migration and invasion through MAPKs in metastatic squamous cell carcinoma of head and neck. International journal of oncology2014, 45(6):2502-2510.
Wang J, Zhang X, Thomas SM, Grandis JR, Wells A, Chen ZG, Ferris RL: Chemokine receptor 7 activates phosphoinositide-3 kinase-mediated invasive and prosurvival pathways in head and neck cancer cells independent of EGFR. Oncogene2005, 24(38):5897-5904.
Chen Y, Shao Z, Jiang E, Zhou X, Wang L, Wang H, Luo X, Chen Q, Liu K, Shang Z: CCL21/CCR7 interaction promotes EMT and enhances the stemness of OSCC via a JAK2/STAT3 signaling pathway. J Cell Physiol2020, 235(9):5995-6009.
Kim JW, Ferris RL, Whiteside TL: Chemokine C receptor 7 expression and protection of circulating CD8+ T lymphocytes from apoptosis. Clin Cancer Res2005, 11(21):7901-7910.
Delgoffe GM, Pollizzi KN, Waickman AT, Heikamp E, Meyers DJ, Horton MR, Xiao B, Worley PF, Powell JD: The kinase mTOR regulates the differentiation of helper T cells through the selective activation of signaling by mTORC1 and mTORC2. Nat Immunol2011, 12(4):295-303.
Zeng Z, Wang K, Li Y, Xia N, Nie S, Lv B, Zhang M, Tu X, Li Q, Tang Tet al: Down-regulation of microRNA-451a facilitates the activation and proliferation of CD4(+) T cells by targeting Myc in patients with dilated cardiomyopathy. J Biol Chem2017, 292(14):6004-6013.
Wang L, Simons DL, Lu X, Tu TY, Solomon S, Wang R, Rosario A, Avalos C, Schmolze D, Yim Jet al: Connecting blood and intratumoral T(reg) cell activity in predicting future relapse in breast cancer. Nat Immunol2019, 20(9):1220-1230.
Villarreal DO, L'Huillier A, Armington S, Mottershead C, Filippova EV, Coder BD, Petit RG, Princiotta MF: Targeting CCR8 Induces Protective Antitumor Immunity and Enhances Vaccine-Induced Responses in Colon Cancer. Cancer Res2018, 78(18):5340-5348.
Plitas G, Konopacki C, Wu K, Bos PD, Morrow M, Putintseva EV, Chudakov DM, Rudensky AY: Regulatory T Cells Exhibit Distinct Features in Human Breast Cancer. Immunity2016, 45(5):1122-1134.

FigS1.tif
Figure.S1 Expression of CCL gene family (CCL1- CCL20) in pan-cancer.
FigS2.tif
Figure.S2 Expression of CCL gene family (CCL21- CCL28) in pan-cancer.
FigS3.tif
Figure.S3 Kaplan-Meier curves of the rest of CCL family.
FigS5.tif
Figure.S5 The correlation analysis between CCL17/CCL22 and CCR2/CCR4/CCR7/CCR8/CXCR3.
SupplementaryTable.1.docx
SupplementaryTable.2.docx
SupplementaryTable.3.docx
SupplementaryTable.4.docx
SupplementaryTable.5.docx
SupplementaryTable.6.docx
SupplementaryTable.7.docx
SupplementaryTable.9.docx
SupplementaryTable.10.docx
SupplementaryTable.11.docx
SupplementaryTable.12.docx

Download PDF

Editor assigned by journal
26 Sep, 2021
Submission checks completed at journal
26 Sep, 2021
Editor invited by journal
26 Sep, 2021
First submitted to journal
24 Sep, 2021

You are reading this latest preprint version

CCL17/CCL22 Predict the Survival and Response to Immune Checkpoint Blockade Therapy of Patients with HNSCC

Status:

Version 1

Abstract

Figures

Introduction

Material And Methods

Data Acquisition and Processing

Oncomine database analysis

cBioPortal analysis

Functional analysis of correlated neighboring genes

GeneMANIA

TRRUST

STRING database analysis

Tumor Purity analysis

TIMER database analysis

Single-cell sequence data analysis

Differentially Expressed Genes (DEGs) Analysis

Tumor immune dysfunction and exclusion (TIDE)

Statistical analysis and data visualization

Results

Overview of the expression and mutation of different CCL family members.

Prognostic role of CCL family members in TCGA HNSCC samples.

A strong co-expression relationship between CCL17 and CCL22 among CCL family members.

Functional Enrichment Analysis of CCL17 and CCL22 correlated genes in TCGA HNSCC samples

CCL17 and CCL22 expression correlated with immune infiltration levels in HNSCC

High expression of CCL22 recruited more CD4 + T cells according to scRNA-seq data

Prediction of clinical response to immune checkpoint blockade (ICB) through CCL17/CCL22 expression

Discussion

Conclusions

Declarations

Ethics approval and consent to participate

Consent for publication

Availability of data and materials

Competing interests:

Funding

Authors’ contributions

Acknowledgments

References

Supplementary Files

Status:

Version 1