Development and validation of an immune-related prognostic signature in cervical cancer

doi:10.21203/rs.3.rs-227250/v1

Download PDF

Research Article

Development and validation of an immune-related prognostic signature in cervical cancer

https://doi.org/10.21203/rs.3.rs-227250/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background

Cervical cancer is the fourth most frequently gynecological malignancy across the world. Immunotherapies have proved to improve prognosis of cervical cancer. However, few studies on immune-related prognostic signature had been reported in cervical cancer.

Methods

Raw data and clinical information of cervical cancer samples were download from TCGA and UCSC Xena website. Immunophenoscore of immune infiltration cells in cervical cancer samples was calculated through ssGSEA method using GSVA package. WGCNA, Cox regression analysis, LASSO analysis and GSEA analysis were performed to classify cervical cancer prognosis and explore the biological signaling pathway.

Results

There were 8 immune infiltration cells associated with prognosis of cervical cancer. Through WGCNA, 153 genes from 402 immune-related genes were significantly correlated with prognosis of cervical cancer. A 15-gene signature demonstrated powerful predictive ability in prognosis of cervical cancer. GSEA analysis showed multiple signaling pathways containing PD-L1 expression and PD-1 checkpoint pathway differences between high risk and low risk groups. Furthermore, the 15-gene signature was associated with multiple immune cells and immune infiltration in tumor microenvironment.

Conclusion

The 15-gene signature is an effective potential prognostic classifier in the immunotherapies and surveillance of cervical cancer.

Cancer Biology

Obstetrics & Gynecology

cervical cancer

weighted gene co-expression network analysis (WGCNA)

least absolute shrinkage and selection operator (LASSO)

prognostic signature

gene set enrichment analysis (GSEA)

According to estimates from GLOBOCAN 2018, cervical cancer was the fourth most common cancer among women, with approximately 570,000 new cases and 311,000 deaths[1]. Cure rate for earliest stages are more than 90%, while locally advanced lesions treated with multimodality therapy can only achieve 65% cure rate in stage II lesions and 55% in stage III lesions[2]. Therapies to improve survivorship are in desperate need for cervical cancer especially locally advanced disease.

With rapid development of precision medicines, novel therapeutic strategies, especially immunotherapies, have been proposed to significantly improve clinical outcomes of cervical cancer[3, 4]. Harnessing an antitumor immune response has been a fundamental strategy in cancer immunotherapy. A paradigm shift has appeared in cancer immunotherapy: from traditional immune enhancement with low objective responses and frequent adverse events to more effective and less toxic reactions immune normalization[5, 6].

Cancer immunotherapies, such as administration of the cytokine IL-2, adoptive cell transfer (ACT) and the checkpoint modulators CTLA-4 and PD-1 have proved effective in clinical practice[6]. Blockade of the checkpoint modulators CTLA-4 and PD-1 starts the field of immune normalization in immunotherapy. Upregulated PD-1 in tumor microenvironment inhibits an effector T cell antitumor immune response, and therapies blocking this pathway have proven effective against multiple tumor types[5]. Anti-PD therapy perform antitumor immunity mainly through the following three principles:1) targeting a tumor-induced immune escape mechanism; 2) selectively modulating immunity in the tumor microenvironment; 3) resetting immunity in the tumor microenvironment[5]. It has been proved that patients with high mutation burden and burden of potential neoepitopes benefit more from immunological checkpoint blockade[7–11]. Immunotherapy plays a dispensable role in management of cervical cancer.

The objective response rate was 26.3% with a disease control rate of 68.4% for immunotherapy in cervical cancer[4]. Only small proportion patients benefited from immunotherapy, this proportion will hopefully increase with better patient selection and combination therapy. So it is urgent to find potential biomarkers for prediction of response to checkpoint immunotherapy and the rationale for the use of checkpoint immunotherapy.

In the present study, we qualify immune cells infiltration in cervical cancer and analyze the correlation between immune cells and cancer prognosis. Hub genes regulating prognosis through immune infiltration in cervical cancer were identified by weighted gene co-expression network analysis (WGCNA) and least absolute shrinkage and selection operator (LASSO). It was suggested by infiltrated immune cells and pathway enrichment analysis that our immune-related signature was closely related to tumor prognosis and could predict response of immunotherapy. Finally, a robust immune-related prognostic signature based on transcriptomics in cervical cancer was constructed and validated.

Data Source and Processing

The gene expression profiles and clinical information of cervical cancer were downloaded from the Cancer Genome Atlas (TCGA) Genomic Data Commons Data Portal (https://portal.gdc.cancer.gov/). Patients with pathologically confirmed cervical cancer and complete information about transcriptomics overall survivals (OS) was included in this study. Finally, a total of 304 primary cervical cancer samples, 2 metastatic cervical cancer samples and 3 normal cervix samples from TCGA was analyzed in our study. Details on clinical information of the included samples were summarized in Table 1. A workflow of this study was indicated in Fig. 1.

Table 1

Clinical characteristics of 309 samples from TCGA
	Alive (N = 235)	Dead (N = 74)	Overall (N = 309)
Age (years)
Mean (SD)	47.4 (13.2)	50.8 (15.4)	48.2 (13.8)
Median [Min, Max]	46.0 [20.0, 88.0]	48.0 [21.0, 79.0]	46.0 [20.0, 88.0]
Pathology
Adenocarcinoma	24 (10.2%)	5 (6.8%)	29 (9.4%)
Adenosquamous carcinoma	4 (1.7%)	1 (1.4%)	5 (1.6%)
Basaloid squamous cell carcinoma	1 (0.4%)	0 (0%)	1 (0.3%)
Endometrioid adenocarcinoma	3 (1.3%)	0 (0%)	3 (1.0%)
Mucinous adenocarcinoma	12 (5.1%)	5 (6.8%)	17 (5.5%)
Papillary squamous cell carcinoma	1 (0.4%)	0 (0%)	1 (0.3%)
Squamous cell carcinoma	190 (80.9%)	63 (85.1%)	253 (81.9%)
Stage
I	125 (53.2%)	38 (51.4%)	163 (52.8%)
II	60 (25.5%)	11 (14.9%)	71 (23.0%)
III	35 (14.9%)	11 (14.9%)	46 (14.9%)
IV	8 (3.4%)	14 (18.9%)	22 (7.1%)
Missing	7 (3.0%)	0 (0%)	7 (2.3%)
Sample
Metastatic	1 (0.4%)	1 (1.4%)	2 (0.6%)
Primary Tumor	232 (98.7%)	72 (97.3%)	304 (98.4%)
Solid Tissue Normal	2 (0.9%)	1 (1.4%)	3 (1.0%)

Infiltration of immune cells

Tumor-infiltrating immune cells can be quantified from RNA sequencing data of human tumors using bioinformatics approaches. Single-sample gene set enrichment analysis (ssGSEA) calculated and qualified the infiltration of immune cells through RNA-seq data. The infiltration of 28 immune cells was obtained. Immunophenoscore (normalized enrichment score, NES) of each TCGA cervical cancer sample were calculated through ssGSEA. The ssGSEA ranks the genes by their absolute expression in a sample and computes Enrichment score by integrating the differences between the empirical cumulative distribution functions of the gene ranks[12].

Survival analysis

Univariate cox regression analysis was performed to identify the association between cervical cancer survival and immunophenoscore of infiltrated immune cells. Forest plot was drawn to demonstrate the influence of infiltrated immune cells on survival. The best separation statistic was performed by use of “survminer” package which divides gene expression into high groups and low groups based on best separation. Next, Kaplan-Meier curve was made to further analyze relationship between survival and infiltration immune cells.

Construction of weighted gene co-expression network analysis (WGCNA)

Weighted correlation network analysis (WGCNA) is an algorithm for finding genetic interactions in a weighted manner. It is used to build a gene co-expression networks to mine network modules closely associated with clinical traits through systematic biological method[13]. In this study, immunophenoscore of the infiltrated immune cells was regarded as target clinical traits. Genes expressing NA were removed. The top 25% genes with most median absolute deviation (MAD) used as a robust measure of variability were selected for WGCNA analysis[14, 15].

Using the WGCNA function adjacency, an adjacency matrix is constructed by computing the Pearson correlation between all pairs of genes in the selected sample. Genes were divided into different gene modules based on the dissimilarity measure. A hierarchical clustering tree was constructed with different branches of the tree representing different gene modules[14–16].

The WGCNA network was built. There were 8 immune cells associated with cervical cancer survival. The magenta module was the module that most related to infiltration of immune cells. A total of 402 genes contained in magenta module were analyzed for survival using the standardized expression data FPKM. Finally, we found that 153 genes were associated with survival (p < 0.05) in magenta module.

Screening hub genes by LASSO and survival analysis

The 153 genes with the highest correlation associated with survival was picked out from WGCNA analysis. Then hub genes were further screened from the 153 genes by the use of LASSO. Survival analysis of hub genes was performed using “survival” and “survminer” packages to verify whether differently expressed genes affected tumor prognosis. Data of cervical cancer from UCSC Xena was downloaded and applied to compare the expression of hub genes in normal cervix and cervical cancer.

Construction of prognostic scoring model

The TCGA cervical cancer patients were randomly divided into two groups by stage: the training group 70% and the test group 30% by the use of caret R package. Finally, a total of 199 patients were used as training group and 74 patients were regarded as test group. The selected key genes for support vector machine analysis were used to fit a LASSO Cox-Proportional Hazards (Cox-PH) model for selecting an optimal panel of predictive genes with penalized package of R[17]. Optimal lambda value was computed through a 10 cross-validations. Next, Cox-PH coefficients and infiltration of immune cells levels of these selected genes were used to calculate prognostic score as follows:

Risk Score = Σ (coef_gene * immunophenoscore_gene),

where Coef_gene and immunophenoscore_gene suggest Cox-PH coefficient and immunophenoscore level of a gene, respectively.

Validation of prognostic scoring model

We calculated risk score of every patient in the training and test group by the model. We separated the training group and test group into a high-risk group and a low-risk group with the median risk score as cutoff respectively. Kaplan-Meier curve was applied to obtain and compare the overall survival time of two risk groups. ROC (operating characteristic curve) was performed to evaluate the predictive accuracy of the model.

Comparison of 8 immune cell subtypes between high risk and low risk groups

To explore the differences of immune cell subtypes between high risk and low risk groups, the 8 immune cell subtypes associated with overall survival in cervical cancer were assessed in test and train cohorts. Mann-Whitney U test was used to compare differences in immune cell subtypes in the high risk and low risk groups.

Gene set enrichment analysis (GSEA)

All 304 primary cervical cancer patients were evaluated by prognostic scoring model and then divided into high risk group and low risk group based on the standard cutoff. The global gene expression was analyzed and displayed with volcanic maps using the limma package in R. GSEA was conducted respectively to search “All gene sets” enriched in the samples with high risk group and low risk group. The differentially expressed genes were enriched in PD-L1 pathway functional sets.

Quantify immune cell infiltration and survival analysis

We used ssGSEA to quantify mRNA data for immune cell infiltration. Finally, 28 infiltrating immune cells were included. Immunophenoscores of 304 primary cervical cancer samples, 2 metastatic cervical cancer samples and 3 normal cervix samples for 28 immune cells were calculated and demonstrated in Fig. 2A. Univariate Cox regression analysis was performed to identify association between immune cells and overall survival of cervical cancer, the results are shown in Fig. 2B. It indicated that 8 immune cells were correlated with OS significantly. Moreover, these 8 immune cells (activated B cell, activated CD8 T cell, eosinophil, monocyte, activated CD4 T cell, effector memory CD8 T cell, immature B cell, plasmacytoid dendritic cell) were all protective factors for OS. Kaplan-Meier curve validated that high expression of these 8 immune cells were related to longer overall survival time suggested worse prognosis (Fig. 3).

The weighted gene co-expression network analysis (WGCNA) construction and key module identification

The top 25% of the gene expression of variance was screened by the quartile of gene expression level, and 609 genes were screened out to construct co-expressed gene networks and the sample dendrogram and trait heatmap are constructed (Fig. 4A). In this study, the power of β = 8 was selected to ensure a scale-free network (Fig. 4B, C) (scale free R² = 0.9, slope = -1.65).

After merging the modules with the high similarity of feature genes in the gene cluster dendrogram through a cut-line (0.25) (Fig. 5A), 9 modules were identified by the dynamic tree cut method. The clustering dendrograms of genes were shown in Fig. 5B. A heat map illustrating the correlation between immunophenoscores of infiltrated immune cells and key genes in the module was created (Fig. 5C).

It was obviously that magenta module was the most correlated with immune infiltration in the heat map. So magenta module was selected as the clinical significant module for further analysis. We found that the correlation coefficient between magenta module and 11 immune infiltration cells was ≥ 0.7 which suggested strong correlation. This indicates that the genes in this module are most relevant to tumor overall survival. All 402 genes in magenta module were analyzed for survival of cervical cancer. Through univariate cox regression survival analysis, 153 genes found to be significantly correlated with prognosis of cervical cancer were selected for further analysis.

Screening the key genes and survival analysis

To develop a prognostic scoring model based on WGCNA, the 153 key genes were used to fit the LASSO Cox-PH model. Using parameter lambda (0.038) obtained upon performing 1000 cross-validations, a combination of 15 genes was obtained. These 15 genes were as follows: LAG3, CD74, CCL22, CH25H, OLR1, MIAT, BATF, IKZF3, TRARG, ACSL6, C11orf21, GTSF1, APOL1, CD1C, LINC00158 (Table 2, Fig. 6). Risk score of each sample in the training set was calculated. Then the samples in the training cohort were divided into high- and low-risk groups according to median risk score. As demonstrated in Fig. 6, the high-risk group patients in the training cohort has shorter overall survival time than the low-risk group (p < 0.001, HR: 5.1, 95%CI: 3.4–7.5), with an AUC of 0.803 in 1 year, 0.809 in 3 year, 0.800 in 5 year. Next, the 15- gene signature was then verified by the validation cohort from TCGA. Consistently, the high-risk group patients in the validation cohort had worse prognosis than the low-risk group (p < 0.001, HR: 1.3, 95%CI: 1.1–1.4) and the AUC was 0.845 in 1 year, 0.703 in 3 year and 0.761 in 5 year (Fig. 7). All these suggested that the 15- gene signature associated with immune infiltration is able to predict prognosis in cervical cancer patients.

Table 2
Fifteen genes from LASSO analysis
Gene	HR	95% CI	p.value	Lasso Coefficient
TRARG1	660	(0.00027-1.6e + 09)	0.39	4.7797999
ACSL6	3.4	(1.4–8.6)	0.008	0.653170468
C11orf21	1.3	(1.1–1.5)	0.00022	0.128185283
GTSF1	1.1	(1.0-1.2)	0.002	0.081672124
CD74	1.0	(1.0–1.0)	0.03	-8.18E-05
APOL1	1.0	(1.0–1.0)	0.016	0.001359139
OLR1	1.0	(1.0–1.0)	0.02	0.008057682
MIAT	1.0	(1.0–1.0)	5.00E-04	0.01207222
BATF	0.96	(0.93-1.0)	0.025	-0.006690799
LAG3	0.93	(0.86–0.99)	0.029	-0.012618
CCL22	0.90	(0.79-1.0)	0.073	-0.003652351
CH25H	0.88	(0.77-1.0)	0.053	-0.017915424
CD1C	0.78	(0.63–0.97)	0.027	-0.100839181
IKZF3	0.76	(0.63–0.93)	0.0061	-0.086985406
LINC00158	0.00062	(6.7e-07-0.58)	0.034	-0.708766884

Immune cell subtypes between high risk and low risk groups

The expression levels of the 15 genes in test and validation cohorts are shown in Fig. 8A. Different immune scores had differential overall survival in cervical cancer samples. In both test cohorts and train cohorts, these 8 immune cell subtypes (activated B cell, activated CD8 T cell, eosinophil, monocyte, activated CD4 T cell, effector memory CD8 T cell, immature B cell, plasmacytoid dendritic cell) expressed differentially in high risk and low risk groups. (Fig. 8B-C).

GSEA analysis

All 304 primary cervical cancer samples were divided into high-risk group and low-risk group based on the 15-gene signature. To identify potential function of the 15 key genes, GSEA was conducted respectively to search “All gene sets” enriched in all 304 samples. GSEA showed 33 significant KEGG pathway associated with risk score, including PD-L1 expression and PD-1 checkpoint pathway in cancer (Fig. 9, Table 3). These results validated that the key genes in clinical significant module were mainly involved in the regulation of immune system.

Table 3
Significant KEGG pathways from GSEA analysis
ID	Description	P.value	p.adjust	q.values
hsa04060	Cytokine-cytokine receptor interaction	3.29E-05	0.000668	0.000554
hsa04062	Chemokine signaling pathway	3.40E-05	0.000668	0.000554
hsa05152	Tuberculosis	3.42E-05	0.000668	0.000554
hsa04514	Cell adhesion molecules	3.45E-05	0.000668	0.000554
hsa05322	Systemic lupus erythematosus	3.49E-05	0.000668	0.000554
hsa04659	Th17 cell differentiation	3.52E-05	0.000668	0.000554
hsa04061	Viral protein interaction with cytokine and cytokine receptor	3.53E-05	0.000668	0.000554
hsa04660	T cell receptor signaling pathway	3.53E-05	0.000668	0.000554
hsa04640	Hematopoietic cell lineage	3.53E-05	0.000668	0.000554
hsa04658	Th1 and Th2 cell differentiation	3.54E-05	0.000668	0.000554
hsa04662	B cell receptor signaling pathway	3.56E-05	0.000668	0.000554
hsa05416	Viral myocarditis	3.63E-05	0.000668	0.000554
hsa05320	Autoimmune thyroid disease	3.64E-05	0.000668	0.000554
hsa04672	Intestinal immune network for IgA production	3.64E-05	0.000668	0.000554
hsa05340	Primary immunodeficiency	3.67E-05	0.000668	0.000554
hsa05330	Allograft rejection	3.67E-05	0.000668	0.000554
hsa05310	Asthma	3.70E-05	0.000668	0.000554
hsa04630	JAK-STAT signaling pathway	6.88E-05	0.001174	0.000974
hsa05145	Toxoplasmosis	0.000526	0.008506	0.007058
hsa04612	Antigen processing and presentation	0.000647	0.009924	0.008234
hsa05140	Leishmaniasis	0.000717	0.010486	0.008701
hsa05321	Inflammatory bowel disease	0.001157	0.01606	0.013326
hsa05235	PD-L1 expression and PD-1 checkpoint pathway in cancer	0.001203	0.01606	0.013326
hsa05169	Epstein-Barr virus infection	0.001524	0.01949	0.016172
hsa05170	Human immunodeficiency virus 1 infection	0.001867	0.022921	0.019019
hsa04950	Maturity onset diabetes of the young	0.002213	0.026135	0.021685
hsa03320	PPAR signaling pathway	0.002867	0.032598	0.027048
hsa04390	Hippo signaling pathway	0.003763	0.040506	0.03361
hsa03010	Ribosome	0.003912	0.040506	0.03361
hsa04080	Neuroactive ligand-receptor interaction	0.004057	0.040506	0.03361
hsa04020	Calcium signaling pathway	0.00416	0.040506	0.03361
hsa05166	Human T-cell leukemia virus 1 infection	0.004222	0.040506	0.03361
hsa04940	Type I diabetes mellitus	0.00475	0.044185	0.036663

Cervical cancer is the fourth most frequently diagnosed cancer and the fourth leading cause of cancer death in women across the world[18]. Cervical cancer is a major challenge for global health[1]. The incidence and mortality vary widely with geographic location. Approximately 90% of cervical cancers occur in developing countries that lack organized screening and HPV vaccination[4]. In lower human development index settings, cervical cancer ranks second in incidence and mortality[18]. There are more than half a million new cervical cancer cases diagnosed and over 300 000 deaths worldwide each year[4]. Cervical cancer represents a major public health problem even in developed countries such as United States. There were 13800 new cases and 4290 deaths estimated in 2020 in United States[19]. Moreover, cervical cancer was the second leading cause of death in women aged 20–39 years in United States[19]. Increasing incidence of cervical adenocarcinoma and attenuation of earlier declines for cervical squamous cell carcinoma emphasize the importance of the need for improved therapeutic options to reduce the burden of cervical cancer of the U.S[20].

Cervical cancer in early stage remains a curable disease that could be treated by surgical resection and concurrent chemoradiation[2]. Women with bulky or locally advanced cervical cancer (IIB-IVA) when lesions are not amenable to high cure rates with surgery or radiation (RT) account for the majority of cervical cancer death[2]. The most important risk factors affecting prognosis of cervical cancer are stage, status of the lymph nodes, tumor volume, depth of tumor invasion into the cervical stroma, and LVSI[21]. The 5-year overall survival decreased significantly with rising stages, with FIGO stage IA, IB, IIB and IIIB were 90%, 83%, 70%, 42% respectively[22]. A substantial percentage of advanced cervical cancer patients will undergo recurrence and poor prognosis[2, 23]. Recurrence rate of cervical cancer fluctuates from 10–74%[24]. Tumor stage was one of the most pivotal factors related to recurrence. The recurrence rate by stage was as follows: 10% for stage IB, 17% for stage IIA, 23% for stage IIB, 42% for stage III and 74% for stage IVA[25]. Prognosis of metastatic and recurrent cervical cancer was extremely poor with a median survival time of 12 months[26].

High-risk subtypes of the human papilloma virus (HPV) are the cause of cervical cancer[27]. Viral oncoproteins E6 and E7 leads to dysregulation of p53 and HIF-1 alpha, thus affecting cell cycle proteins and VEGF expression[27–30]. HPV is highly immunogenic and elicits immune responses in humans, thus immune might play important roles in carcinogenesis of cervical cancer. Incidence of cervical cancer was substantial declined in countries with high HPV vaccine coverage[31–33]. Immunotherapy fights against tumor cells through activating endogenous immune response which seems to switch on the new frontier of the anticancer treatment[34]. Immunotherapy has different approaches, such as active immunotherapy (vaccine), passive immunotherapy (adoptive cellular transfer, antibodies and cytokines) and immunomodulation (cyclooxygenase 2 inhibitor)[34]. Discovery of immune checkpoint such as CTLA-4 and PD-1 plays an indispensable role in the development of cancer immunotherapy. It was surprising that immune checkpoint inhibitors anti-CTLA-4 and anti-PD-1 displayed enormous success in solid tumors[35]. Similar to other solid tumors, novel immunotherapeutic approaches, such as immune checkpoint inhibitors have shown promising results in cervical cancer[4]. Based on high response observed in KEYNOTE-158, the US Food and Drug Administration approved pembrolizumab inhibiting PD-1 for use in advanced cervical cancer with progressive disease either during or after chemotherapy[36].

Immunotherapy has shown encouraging results in cervical cancer treatment. Thus, implementing immunotherapeutic approaches earlier in advanced cervical cancer would seem to be most appropriate. However, the objective response rate with anti-PD-1/PD-L1 monotherapy is hovering at 20%. Moreover, immune-related toxicities and severe adverse effects can occur during PD-1/PD-L1 blockade therapy[37]. PD-1/PD-L1 blockade therapy does not demonstrate efficacy in almost 80% cervical cancer patients, suggesting the potential mechanism of PD-1/PD-L1 in immunotherapy remain to be further clarified. So new immune checkpoint inhibitors or comprehensive understanding of specific mechanism underlying PD-1/PD-L1 regulation in carcinogenesis is in urgent need.

Immune infiltration of cervical cancer determines the immune activation of tumor microenvironment and is related with clinical outcome of patients. In this study, immunophenoscore of 28 infiltrated immune cell in TCGA cervical cancer was calculated. Univariate Cox regression analysis showed that activated B cell, activated CD8 T cell, eosinophil, monocyte, activated CD4 T cell, effector memory CD8 T cell, immature B cell, plasmacytoid dendritic cell were strongly associated with overall survival of cervical cancer. WGCNA revealed that magenta module was the most relevant module to immune infiltration. In total, we identified 10 kinds of infiltrated immune cells with strong correlation with magenta module. These were as follows: activated B cell (R = 0.88), activated CD8 T cell (R = 0.82), activated dendritic cell (R = 0.79), central memory CD4 T cell (R = 0.78), effector memory CD8 T cell (R = 0.85), immature B cell (R = 0.93), macrophage (R = 0.72), MDSC (R = 0.85), regulatory T cell (R = 0.71), T follicular helper cell (R = 0.72), type 1 T helper cell (R = 0.83).

By the use of WGCNA, we found magenta module showed the highest correlation with the immune infiltration of cervical cancer. The magenta module contained 609 immune-related genes. There were 153 genes picked up from the magenta module for further analysis after taking intersection of prognosis and immune infiltration. Through LASSO analysis, 15 immune-related genes (LAG3, CD74, CCL22, CH25H, OLR1, MIAT, BATF, IKZF3, TRARG, ACSL6, C11orf21, GTSF1, APOL1, CD1C, LINC00158) were included in prognostic classifier. Except CTLA-4 and PD-1/PD-L1 targeted cancer immunotherapy, LAG3 (lymphocyte activation gene-3, CD223) is the third inhibitory receptor to be targeted in the clinic[38]. CD74 (invariant chain) plays a dispensable role in process of immune systems that it participates in antigen presentation, B-cell differentiation and inflammatory signaling. CD74 has the potential to be a therapeutic target in cancer and autoimmune disease[39]. The chemokine CCL22 promoted regulatory T cell communication with dendritic cells to control immunity and was associated with poor prognosis[40, 41]. CH25H produces 25-hydroxycholesterol, which inhibited tumor-derived extracellular vesicles uptake and correlated with prognosis in melanoma patients[42]. OLR1 (oxidized LDL receptor 1) was a possible link between obesity, dyslipidemia and cancer. OLR1 played carcinogenic role by activating NF-kB pathway to promote proliferation, migration and inhibit apoptosis and de novo lipogenesis[43]. MIAT lncRNA was overexpressed in a number of malignancies and caused poor prognosis[44–46]. BATF was an important transcription factor regulating differentiation of early effector CD8 + T cells[47] and was a prognostic indicator for patients with colon cancer[48]. IKZF3 promoted growth of multiple tumors to cause poor prognosis[49].

Immune infiltration played important roles in the survival of cervical cancer. This study identified 15 immune infiltration associated genes in cervical cancer. Immune scores depended on the expression of these 15 genes and were associated with the survival of cervical cancer. In this study, high immune scores meant good prognosis. A 15-gene prognostic signature associated with immune system in cervical cancer was built. GSEA analysis showed that the 15-gene prognostic signature was obviously associated with PD-L1 expression and PD-1 checkpoint pathway in cancer. The prognostic signature could provide basis for potential immunotherapy in the future. However, the study has several limitations. First, no vitro or vivo molecular experiment was performed to verify our analysis, Second, our study was a retrospective study. So prospective study is in need to validate the findings of our study.

In conclusion, we successfully constructed a 15-gene prognostic signature with powerful predictive function. Differences in the overall survival of high and low risk groups are implicated in immune infiltration, tumor microenvironment, PD-L1 expression and PD-1 checkpoint pathway. These findings revealed the underlying mechanism of immunotherapy and provided basis for cervical cancer pathogenesis and clinical treatment.

ETHICS APPROVAL AND CONSENT TO PARTICIPATE

Not Applicable.

CONSENT TO PUBLISH

Not Applicable.

CONFLICT OF INTEREST

No conflict of interest exits in the submission of this manuscript, and manuscript is approved by all authors for publication.

DATA AVAILABILITY STATEMENT

The raw data supporting the conclusions of this article sourced the Cancer Genome Atlas (TCGA) Genomic Data Commons Data Portal (https://portal.gdc.cancer.gov/).

AUTHOR CONTRIBUTIONS

S-RJ is responsible for experimental design. J-CW and B-HL are responsible for instrument operation. W-XY and K-MH are responsible for data analysis. X-JD and J-CJ are for providing overall ideas. All authors contributed to the article and approved the submitted version.

FUNDING

This work was supported by the National Natural Science Foundation of China (grant NO. 82002736).

ACKNOWLEDGEMENTS

We thank Dr. Jianming Zeng (University of Macau), and all the members of his bioinformatics team, biotrainee, for generously sharing their experience and codes，especially Xiaojie Sun.

Arbyn M, Weiderpass E, Bruni L, de Sanjosé S, Saraiya M, Ferlay J, Bray F: Estimates of incidence and mortality of cervical cancer in 2018: a worldwide analysis. The Lancet Global health 2020, 8(2):e191-e203.
Monk BJ, Tewari KS, Koh WJ: Multimodality therapy for locally advanced cervical carcinoma: state of the art and future directions. Journal of clinical oncology: official journal of the American Society of Clinical Oncology 2007, 25(20):2952–2965.
Chung HC, Ros W, Delord JP, Perets R, Italiano A, Shapira-Frommer R, Manzuk L, Piha-Paul SA, Xu L, Zeigenfuss S et al: Efficacy and Safety of Pembrolizumab in Previously Treated Advanced Cervical Cancer: Results From the Phase II KEYNOTE-158 Study. Journal of clinical oncology: official journal of the American Society of Clinical Oncology 2019, 37(17):1470–1478.
Cohen PA, Jhingran A, Oaknin A, Denny L: Cervical cancer. Lancet (London, England) 2019, 393(10167):169–182.
Sanmamed MF, Chen L: A Paradigm Shift in Cancer Immunotherapy: From Enhancement to Normalization. Cell 2018, 175(2):313–326.
Tran E, Robbins PF, Rosenberg SA: 'Final common pathway' of human cancer immunotherapy: targeting random somatic mutations. Nature immunology 2017, 18(3):255–262.
Asaoka Y, Ijichi H, Koike K: PD-1 Blockade in Tumors with Mismatch-Repair Deficiency. The New England journal of medicine 2015, 373(20):1979.
Hugo W, Zaretsky JM, Sun L, Song C, Moreno BH, Hu-Lieskovan S, Berent-Maoz B, Pang J, Chmielowski B, Cherry G et al: Genomic and Transcriptomic Features of Response to Anti-PD-1 Therapy in Metastatic Melanoma. Cell 2016, 165(1):35–44.
Powles T, Eder JP, Fine GD, Braiteh FS, Loriot Y, Cruz C, Bellmunt J, Burris HA, Petrylak DP, Teng SL et al: MPDL3280A (anti-PD-L1) treatment leads to clinical activity in metastatic bladder cancer. Nature 2014, 515(7528):558–562.
Rizvi NA, Hellmann MD, Snyder A, Kvistborg P, Makarov V, Havel JJ, Lee W, Yuan J, Wong P, Ho TS et al: Cancer immunology. Mutational landscape determines sensitivity to PD-1 blockade in non-small cell lung cancer. Science (New York, NY) 2015, 348(6230):124–128.
van Rooij N, van Buuren MM, Philips D, Velds A, Toebes M, Heemskerk B, van Dijk LJ, Behjati S, Hilkmann H, El Atmioui D et al: Tumor exome analysis reveals neoantigen-specific T-cell reactivity in an ipilimumab-responsive melanoma. Journal of clinical oncology: official journal of the American Society of Clinical Oncology 2013, 31(32):e439-442.
Finotello F, Trajanoski Z: Quantifying tumor-infiltrating immune cells from transcriptomics data. Cancer immunology, immunotherapy: CII 2018, 67(7):1031–1040.
Langfelder P, Horvath S: WGCNA: an R package for weighted correlation network analysis. BMC bioinformatics 2008, 9:559.
Dong J, Horvath S: Understanding network concepts in modules. BMC systems biology 2007, 1:24.
Horvath S, Dong J: Geometric interpretation of gene coexpression network analysis. PLoS computational biology 2008, 4(8):e1000117.
Langfelder P, Zhang B, Horvath S: Defining clusters from a hierarchical cluster tree: the Dynamic Tree Cut package for R. Bioinformatics (Oxford, England) 2008, 24(5):719–720.
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. International journal of cancer 2015, 136(5):E359-386.
Bray F, Ferlay J, Soerjomataram I, Siegel RL, Torre LA, Jemal A: Global cancer statistics 2018: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA: a cancer journal for clinicians 2018, 68(6):394–424.
Siegel RL, Miller KD, Jemal A: Cancer statistics, 2020. CA: a cancer journal for clinicians 2020, 70(1):7–30.
Islami F, Fedewa SA, Jemal A: Trends in cervical cancer incidence rates by age, race/ethnicity, histological subtype, and stage at diagnosis in the United States. Preventive medicine 2019, 123:316–323.
Peters WA, 3rd, Liu PY, Barrett RJ, 2nd, Stock RJ, Monk BJ, Berek JS, Souhami L, Grigsby P, Gordon W, Jr., Alberts DS: Concurrent chemotherapy and pelvic radiation therapy compared with pelvic radiation therapy alone as adjuvant therapy after radical surgery in high-risk early-stage cancer of the cervix. Journal of clinical oncology: official journal of the American Society of Clinical Oncology 2000, 18(8):1606–1613.
Sturdza A, Pötter R, Fokdal LU, Haie-Meder C, Tan LT, Mazeron R, Petric P, Šegedin B, Jurgenliemk-Schulz IM, Nomden C et al: Image guided brachytherapy in locally advanced cervical cancer: Improved pelvic control and survival in RetroEMBRACE, a multicenter cohort study. Radiotherapy and oncology: journal of the European Society for Therapeutic Radiology and Oncology 2016, 120(3):428–433.
Liontos M, Kyriazoglou A, Dimitriadis I, Dimopoulos MA, Bamias A: Systemic therapy in cervical cancer: 30 years in review. Critical reviews in oncology/hematology 2019, 137:9–17.
Peiretti M, Zapardiel I, Zanagnolo V, Landoni F, Morrow CP, Maggioni A: Management of recurrent cervical cancer: a review of the literature. Surgical oncology 2012, 21(2):e59-66.
Perez CA, Grigsby PW, Nene SM, Camel HM, Galakatos A, Kao MS, Lockett MA: Effect of tumor size on the prognosis of carcinoma of the uterine cervix treated with irradiation alone. Cancer 1992, 69(11):2796–2806.
Elit LM, Hirte H: Management of advanced or recurrent cervical cancer: chemotherapy and beyond. Expert review of anticancer therapy 2014, 14(3):319–332.
Wardak S: Human Papillomavirus (HPV) and cervical cancer. Medycyna doswiadczalna i mikrobiologia 2016, 68(1):73–84.
Burd EM: Human papillomavirus and cervical cancer. Clinical microbiology reviews 2003, 16(1):1–17.
Monk BJ, Willmott LJ, Sumner DA: Anti-angiogenesis agents in metastatic or recurrent cervical cancer. Gynecologic oncology 2010, 116(2):181–186.
Nakamura M, Bodily JM, Beglin M, Kyo S, Inoue M, Laimins LA: Hypoxia-specific stabilization of HIF-1alpha by human papillomaviruses. Virology 2009, 387(2):442–448.
Brotherton JM, Fridman M, May CL, Chappell G, Saville AM, Gertig DM: Early effect of the HPV vaccination programme on cervical abnormalities in Victoria, Australia: an ecological study. Lancet (London, England) 2011, 377(9783):2085–2092.
Drolet M, Bénard É, Pérez N, Brisson M: Population-level impact and herd effects following the introduction of human papillomavirus vaccination programmes: updated systematic review and meta-analysis. Lancet (London, England) 2019, 394(10197):497–509.
Gargano JW, Zhou F, Stokley S, Markowitz LE: Human papillomavirus vaccination in commercially-insured vaccine-eligible males and females, United States, 2007–2014. Vaccine 2018, 36(23):3381–3386.
Ventriglia J, Paciolla I, Pisano C, Cecere SC, Di Napoli M, Tambaro R, Califano D, Losito S, Scognamiglio G, Setola SV et al: Immunotherapy in ovarian, endometrial and cervical cancer: State of the art and future perspectives. Cancer treatment reviews 2017, 59:109–116.
Balar AV, Weber JS: PD-1 and PD-L1 antibodies in cancer: current status and future directions. Cancer immunology, immunotherapy: CII 2017, 66(5):551–564.
Marabelle A, Fakih M, Lopez J, Shah M, Shapira-Frommer R, Nakagawa K, Chung HC, Kindler HL, Lopez-Martin JA, Miller WH, Jr. et al: Association of tumour mutational burden with outcomes in patients with advanced solid tumours treated with pembrolizumab: prospective biomarker analysis of the multicohort, open-label, phase 2 KEYNOTE-158 study. The Lancet Oncology 2020, 21(10):1353–1365.
Topalian SL, Hodi FS, Brahmer JR, Gettinger SN, Smith DC, McDermott DF, Powderly JD, Carvajal RD, Sosman JA, Atkins MB et al: Safety, activity, and immune correlates of anti-PD-1 antibody in cancer. The New England journal of medicine 2012, 366(26):2443–2454.
Andrews LP, Marciscano AE, Drake CG, Vignali DA: LAG3 (CD223) as a cancer immunotherapy target. Immunological reviews 2017, 276(1):80–96.
Borghese F, Clanchy FI: CD74: an emerging opportunity as a therapeutic target in cancer and autoimmune disease. Expert opinion on therapeutic targets 2011, 15(3):237–251.
Rapp M, Wintergerst MWM, Kunz WG, Vetter VK, Knott MML, Lisowski D, Haubner S, Moder S, Thaler R, Eiber S et al: CCL22 controls immunity by promoting regulatory T cell communication with dendritic cells in lymph nodes. The Journal of experimental medicine 2019, 216(5):1170–1181.
Wang Q, Schmoeckel E, Kost BP, Kuhn C, Vattai A, Vilsmaier T, Mahner S, Mayr D, Jeschke U, Heidegger HH: Higher CCL22 + Cell Infiltration is Associated with Poor Prognosis in Cervical Cancer Patients. Cancers 2019, 11(12).
Ortiz A, Gui J, Zahedi F, Yu P, Cho C, Bhattacharya S, Carbone CJ, Yu Q, Katlinski KV, Katlinskaya YV et al: An Interferon-Driven Oxysterol-Based Defense against Tumor-Derived Extracellular Vesicles. Cancer cell 2019, 35(1):33–45.e36.
Khaidakov M, Mitra S, Kang BY, Wang X, Kadlubar S, Novelli G, Raj V, Winters M, Carter WC, Mehta JL: Oxidized LDL receptor 1 (OLR1) as a possible link between obesity, dyslipidemia and cancer. PloS one 2011, 6(5):e20277.
Huang X, Gao Y, Qin J, Lu S: lncRNA MIAT promotes proliferation and invasion of HCC cells via sponging miR-214. American journal of physiology Gastrointestinal and liver physiology 2018, 314(5):G559-g565.
Sha M, Lin M, Wang J, Ye J, Xu J, Xu N, Huang J: Long non-coding RNA MIAT promotes gastric cancer growth and metastasis through regulation of miR-141/DDX5 pathway. Journal of experimental & clinical cancer research: CR 2018, 37(1):58.
Liu W, Wang Z, Wang C, Ai Z: Long non-coding RNA MIAT promotes papillary thyroid cancer progression through upregulating LASP1. Cancer cell international 2019, 19:194.
Kurachi M, Barnitz RA, Yosef N, Odorizzi PM, DiIorio MA, Lemieux ME, Yates K, Godec J, Klatt MG, Regev A et al: The transcription factor BATF operates as an essential differentiation checkpoint in early effector CD8 + T cells. Nature immunology 2014, 15(4):373–383.
Dai L, Cui X, Zhang X, Cheng L, Liu Y, Yang Y, Fan P, Wang Q, Lin Y, Zhang J et al: SARI inhibits angiogenesis and tumour growth of human colon cancer through directly targeting ceruloplasmin. Nature communications 2016, 7:11996.
Liu XP, He L, Zhang QP, Zeng XT, Liu SQ: Baicalein Inhibits Proliferation of Myeloma U266 Cells by Downregulating IKZF1 and IKZF3. Medical science monitor: international medical journal of experimental and clinical research 2018, 24:2809–2817.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Development and validation of an immune-related prognostic signature in cervical cancer

Status:

Version 1

Abstract

Background

Methods

Results

Conclusion

Figures

Introduction

Materials And Methods

Data Source and Processing

Infiltration of immune cells

Survival analysis

Construction of weighted gene co-expression network analysis (WGCNA)

Screening hub genes by LASSO and survival analysis

Construction of prognostic scoring model

Validation of prognostic scoring model

Comparison of 8 immune cell subtypes between high risk and low risk groups

Gene set enrichment analysis (GSEA)

Results

Quantify immune cell infiltration and survival analysis

The weighted gene co-expression network analysis (WGCNA) construction and key module identification

Screening the key genes and survival analysis

Immune cell subtypes between high risk and low risk groups

GSEA analysis

Discussion

Conclusion

Declarations

References

Additional Declarations

Status:

Version 1