This article is mentioned in:

Introduction

Laryngeal squamous cell carcinoma (LSCC) is one of the most common types of head and neck squamous cell carcinoma, the incidence of which continues to rise. Despite the continuous improvement in surgical, radiotherapeutic and chemotherapeutic techniques, an efficient prolongation of the survival rate of patients with LSCC has yet to be achieved (1,2). The 5-year survival rate of patients with LSCC is ~30% (3). Due to the specificity of the anatomic location of the throat, patients with LSCC not only have to endure painful surgery, but they also experience postoperative impairment and loss of throat function. Subsequent physical and psychological afflictions have a serious impact on the outcome of postoperative rehabilitation and the quality of life of these patients (4). In recent years, cancer immunotherapy has gradually become a subject undergoing intense study in cancer research and has achieved satisfactory therapeutic efficacy in the clinical treatment of cancers (5). However, compared with other types of cancer treatment, progress in the development of LSCC-related immunotherapy has been fairly slow. One of the main reasons behind this slow progress is the lack of effective molecular targets for immunotherapy. Therefore, it is necessary to determine the immunoregulatory mechanisms underlying LSCC and to identify factors that are critical for the success of immunotherapy.

Regulatory T cells (Tregs) comprise a subpopulation of T cells with unique immunomodulatory functions. Tregs are able to suppress immune responses and therefore serve important roles in the maintenance of immune homeostasis, tumor immunity and transplant tolerance (6). Forkhead box P3 (FOXP3) is a characteristic Treg surface marker, and a key factor for Treg development and maintenance of function (7). Franco et al (8) silenced the expression of FOXP3, to reduce the infiltration of FOXP3+ Tregs, and identified that melanoma growth was significantly inhibited, and that the immunosuppressive microenvironment was restored. Treg-secreted cytokines, including interleukin (IL)-10, serve an important role in the function of Tregs. These cytokines may initiate inhibitory signals and suppress the activity of functional T lymphocytes, which inhibits the tumor immune response (9). In addition, it has previously been demonstrated that certain costimulatory molecules exhibit an important influence on the role of Tregs in cancer immune evasion. For example, the costimulatory molecule cluster of differentiation (CD)274 (also known as programmed death-ligand 1) is abnormally expressed in various tumor tissues. CD274 is an important molecule that mediates cancer immune evasion (10). The expression status of CD274 is closely associated with the prognosis of various tumors, including ovarian cancer, bladder cancer and pancreatic cancer (11–13). These previous findings suggest that Tregs positive for the tumor immunity-associated molecules IL-10, CD274 and FOXP3 serve an important role in tumor development and progression. However, correlations among the expression of these three molecules, and the association between these three molecules and the prognosis of LSCC remain to be elucidated.

Numerous studies have demonstrated that the coding and non-coding gene regulatory associations are extensive in tumor immunity, and that they serve essential roles in tumor development and progression. Liang et al (14) indicated that microRNA (miR)-22 was able to impair antitumor immunity through targeting p38. In addition, various regulatory associations exist between coding and non-coding genes in complex gene regulatory networks. For instance, miRNAs bind to sites in the 3′-termini of their target genes to regulate gene expression, and long non-coding RNAs (lncRNAs) or circular RNAs (circRNAs) act as competitive endogenous RNAs (ceRNAs) to regulate target gene expression (15). Therefore, it is useful to identify the non-coding genes that regulate IL-10, FOXP3 and CD274 in order to understand the molecular mechanism underlying tumor immunity in LSCC.

In the present study, immunohistochemistry (IHC) was employed to examine the expression of IL-10, CD274 and FOXP3 in tumor samples derived from 133 patients with LSCC. Subsequently, associations between the expression of IL-10, CD274 and FOXP3 and the clinical characteristics of LSCC were analyzed. In addition, Spearman's rank correlation method was used to analyze the correlations among IL-10, CD274 and FOXP3 protein expression. The Kaplan-Meier (KM) method and the Cox regression model were employed in combination with the clinical data to analyze the effects of the expression of these three proteins on the survival rate of patients with LSCC. The present study aimed to increase understanding regarding the immune evasion mechanisms of LSCC and to aid the development of LSCC immunotherapy.

Materials and methods

Collection of LSCC samples

The present study was approved by the Clinical Research Ethics Committee of Beijing Tongren Hospital, Capital Medical University (Beijing, China), and written informed consent was obtained from the patients. This investigation was conducted according to the Declaration of Helsinki. A total of 133 LSCC samples were collected for use in the present study; these samples consisted of paraffin-embedded tissue blocks obtained from patients with primary LSCC that underwent surgical treatment between January 2012 and July 2013 at Beijing Tongren Hospital affiliated with Capital Medical University. The patients received a definite pathological diagnosis, and complete clinical information was available. In addition, none of the patients received preoperative radiotherapy or chemotherapy. The inclusion criteria were established based on tumor size and location, invasion status of the primary tumors and the status of lymph node metastasis. The LSCC samples were subjected to histopathological classification and tumor stage evaluation, which were conducted according to the Koppel standards and the tumor-node-metastasis (TNM) staging system, respectively. All pathological diagnoses were verified by senior pathologists through a review of the tissue sections (Table I). The normal laryngeal mucosa tissue and vocal cord polyp tissue were derived from the patients with recurrent cysts and vocal cord polyps who had undergone surgery between January 2012 and July 2013 at Beijing Tongren Hospital affiliated with Capital Medical University (Beijing, China).

Table I. General information and clinical characteristics of patients with LCSS.

Table I.

General information and clinical characteristics of patients with LCSS.

Variable	Cases	%
Gender
Male	128	96.2
Female	5	3.8
Age
<60 years	48	36.1
≥60 years	85	63.9
Tumor location
Supraglottic	30	22.6
Glottic	101	75.9
Subglottic	2	1.5
TNM
T1/T2	91	68.4
T3/T4	42	31.6
LNM
No	98	73.7
Yes	35	26.3
Clinical stage
I	48	36.1
II	38	28.6
III	26	19.5
IV	21	15.8
Course of disease
≤6 months	53	39.8
6–12 months	28	21.1
≥12 months	52	39.1
Pathological stage (differentiation)
Well	52	39.1
Moderate	53	39.8
Poor	28	21.1

[i] LNM, lymph node metastasis; LSCC, laryngeal squamous cell carcinoma; TNM, tumor-node-metastasis classification.

Hematoxylin and eosin (H&E) staining

Phosphate-buffered saline (PBS) was used to wash the sections of the normal and tumor tissues for 5 min. The nucleus was stained with hematoxylin for 5 min and was differentiated with hydrochloric acid and ethanol for 30 sec. Tap water and distilled water was used to wash the sections for 1 min. Eosin staining was performed for 30 sec. The slices were washed with a 70% ethanol for 30 sec, a 90% ethanol rinse for 30 sec, a 95% ethanol rinse for 30 sec, two 100% ethanol rinses for 2 min each time and xylene clearing twice for 5 min each time. Finally, an inverted light microscope was used to observe the stained sections.

IHC staining

Paraffin-embedded tissue sections (4 µm) were deparaffinized and rehydrated according to conventional methods. The tissue sections were washed in PBS and were subjected to antigen retrieval. EDTA buffer (pH 9.0) was used to retrieve antigen as a repair solution. The antigen was placed in an autoclave for 2 min. Once cool, the tissue sections were rinsed with distilled water and then stored in Tris Buffered Saline (pH 7.4). Following antigen retrieval, each section was incubated with 50 µl 3% H2O2 at room temperature for 10 min to block endogenous peroxidase activity. Anti-CD274 antibody (NBP1-76769; 1:200) and anti-FOXP3 antibody (NB100-39002; 1:200) were purchased from Novus Biologicals, LLC (Littleton, CO, USA), and the anti-IL-10 antibody (BA1201-1; 1:200) was purchased from Wuhan Boster Biological Technology, Ltd. (Wuhan, China). Following washing with PBS, the primary antibodies (1:200 dilution) were added dropwise to the sections, followed by incubation at 4°C overnight. Subsequently, the sections were washed again with PBS. According to the protocol of an IHC two-step detection kit (PV-6001; OriGene Technologies, Beijing, China), the section was incubated with 100 µl antibodies at room temperature for 30 min. Subsequently, the sections were exposed to the chromogenic substrate 3,3′-diaminobenzidine (OriGene Technologies) and observed under a light microscope for 3–10 min. The appearance of brownish-yellow particles indicated positive staining. Following washing with distilled water, the sections were counterstained with hematoxylin at room temperature for 10 min, differentiated in 0.1% hydrochloric acid and rinsed with tap water. The sections were then dehydrated in a graded series of alcohol, cleared in xylene and mounted in neutral gum. Negative controls were established by replacing the primary antibodies with PBS. Gastric cancer tissues from the tissue bank of Beijing Tongren Hospital were used as positive controls. Uniformly stained sections that exhibited clear tissue structure were subjected to semi-quantitative analysis. The staining results were evaluated by two pathologists using the double-blind method.

Assessment of the IHC results

The IHC results were scored based on the semi-quantitative immunoreactivity scoring system developed by Remmele and Stegnerde (16). For each tissue section, 5 areas within the tumor were randomly selected under ×200 magnification. The percentage of cells in the selected areas that exhibited positive cytoplasmic/plasma membrane staining was calculated and assigned a score between 0 and 4, as follows: 0, No cells exhibited positive staining; 1, 1–10% of the cells exhibited positive staining; 2, 11–50% of the cells exhibited positive staining; 3, 51–75% of the cells exhibited positive staining; and 4, >75% of the cells exhibited positive staining. In addition, the staining intensity was scored as follows: 0, No staining; 1, weak intensity (light yellow); 2, moderate intensity (brownish yellow); and 3, strong intensity (dark brown). The two scores were multiplied together to generate IHC scores, and the samples were graded according to their IHC scores as follows: Negative (−), 0; weakly positive (+), 1, positive (++), 2–3; and strongly positive (+++), >3. This semi-quantitative scoring system was used to determine staining in tumor-infiltrating lymphocytes. Initially, the distribution characteristics of the positive cells in whole tissue sections were examined under low magnification (×40), which allowed cancer foci to be distinguished from tumor stroma. To quantify the tumor-infiltrating lymphocytes, the 5 areas richest in tumor-infiltrating lymphocytes were selected from each tissue section under low magnification (×40), and the tumor-infiltrating lymphocytes were counted under a high-power field (×400). The number of tumor-infiltrating lymphocytes was expressed as the average number of cells in the five fields of view. The distribution characteristics of the CD274-, FOXP3- and IL-10-positive and -negative cells in whole tissue sections were examined under low magnification. The appearance of brownish-yellow precipitate in the cytoplasm and/or the cell membrane indicated positive staining. The staining results were evaluated using a semi-quantitative scoring system as described above. The infiltrating lymphocytes were classified into the following four categories based on the IHC staining scores for CD274, FOXP3 and IL-10: (−), <1; (+), 1–5; (++), 6–19; and (+++), >20. In addition, (−) and (+) indicated low expression, whereas (++) and (+++) indicated high expression. These scores were determined independently by two pathologists and were averaged and subjected to statistical analysis.

KM survival rate analysis and Cox regression model

The KM model was used to assess the effects of FOXP3, IL-10 and CD274 expression on the survival rate of patients with LSCC. Due to the advantages associated with the KM model, including the accurate assessment of the survival rates and the median survival time, and the pairwise comparison of survival time distributions between numerous groups, this model has been widely adopted. In addition, the Cox regression model was employed to analyze and compare the effects of various factors on the survival time of patients with LSCC (17). The Cox regression model is commonly used to analyze the prognosis of tumors and other chronic diseases. The equation used in the Cox regression analysis is as follows: h(t/X)=h0(t) exp (β1 X1 + β2 X2 + … + βp Xp), where h(t/X) is the baseline hazard function, namely, the hazard function at time t when all independent variables assume a value of zero. The influencing factors (variables) are denoted by X1, X2, …, Xp, whereas the regression coefficients are represented by β1, β2, …, βp.

Identification of miRNA, lncRNA and circRNA that regulate FOXP3, IL-10 and CD274

To identify the potential mechanism underlying the correlations among the three genes, the miRTarBase database (http://mirtarbase.mbc.nctu.edu.tw/) was used to predict the miRNAs of FOXP3, IL-10 and CD274. Based on the miRNAs, StarBase version 2.0 database (http://starbase.sysu.edu.cn/) was used to identify the lncRNAs and circRNAs that regulate the miRNAs. These interactions were confirmed by experiments including western blot analysis and PCR. Based on these interactions, Cytoscape software (http://www.cytoscape.org/) was used to establish the target gene-miRNA-lncRNA and the target gene-miRNA-circRNA networks.

Statistical analysis

The data were statistically analyzed using SPSS software version 21.0 (IBM Corp., Armonk, NY, USA). Comparisons between the two groups were performed using the χ2 test. A correlation analysis was conducted using Spearman's rank correlation test. The KM method and log-rank test were used to analyze survival rates. The association between the expression levels of CD274, Foxp3 and IL-10 and various clinicopathological characteristics was subjected to a rank transformation and entered into Spearman rank correlation test tests. P<0.05 was considered to indicate a statistically significant difference.

Results

CD274, FOXP3 and IL-10 expression in normal laryngeal mucosa, vocal cord polyp and LSCC

The hematoxylin and eosin staining of LSCC tissues is presented in Fig. 1A. The results of IHC demonstrated that the positive expression rate of CD274 was 58.6% (Fig. 1B), the positive expression rate of IL-10 was 73.7% (Fig. 1C) and the positive expression rate of FOXP3 was 68.4% (Fig. 1D) in the 133 LSCC tissues. However, their expression in normal laryngeal mucosa and vocal cord polyp tissues was negative or at a low level (Table II).

Figure 1. CD274, FOXP3 and IL-10 expression in LSCC. (A) Hematoxylin and eosin staining of LSCC tissue. Immunohistochemical staining of LSCC tissue: (B) CD274; (C) FOXP3; and (D) IL-10 expression. Magnification, ×40. CD, cluster of differentiation; FOX, forkhead box; IL, interleukin; LSCC, laryngeal squamous cell carcinoma.

Table II. Expression of CD274, FOXP3 and IL-10 in three tissue types.

Table II.

Expression of CD274, FOXP3 and IL-10 in three tissue types.

		CD274				FOXP3				IL-10
Tissue type	Cases	+	−	χ2	P-value	+	−	χ2	P-value	+	−	χ2	P-value
NLM	15	0	0	/	/	3	12	31.10	<0.001	2	13	38.07	<0.001
VCP	15	0	0			1	14			2	13
LCSS	133	78	55			91	42			98	35

[i] CD, cluster of differentiation; FOX, forkhead box; IL, interleukin; LCSS, laryngeal squamous cell carcinoma; NLM, normal laryngeal mucosa; VCP, vocal cord polyp.

Analysis of the correlations between the clinicopathological features and CD274, FOXP3 and IL-10 expression

In patients with LSCC, CD274 expression demonstrated no correlation with tumor location, TNM stage, clinical stage or pathologic grade (P>0.05). However, CD274 expression was significantly correlated with lymph node metastasis, disease course prior to hospital admission, the age of the patients and prognosis (P<0.05). FOXP3 expression appeared to exhibit no correlation with the age of the patients, tumor location, T stage, clinical stage or pathological grade (P>0.05). However, FOXP3 expression exhibited clear correlations with disease course prior to hospital admission, lymph node metastasis and prognosis (P<0.05). IL-10 expression demonstrated correlations with all factors, with the exception of patient age and tumor location (P<0.05; Table III).

Table III. CD274, FOXP3 and IL-10 expression in LCSS.

Table III.

CD274, FOXP3 and IL-10 expression in LCSS.

	CD274			FOXP3			IL-10
Variable	−	+	P-value	−	+	P-value	−	+	P-value
Age
≤60 years	14	34	0.032	16	32	0.744	14	34	0.575
>60 years	41	44		26	59		21	64
Tumor location
Supraglottic	11	19	0.497	9	21	1.000	6	24	0.453
Glottic	44	57		33	68		29	72
Subglottic	0	2		0	2		0	2
TNM
T1/T2	41	50	0.202	33	58	0.087	31	60	0.003
T3/T4	14	28		9	33		4	38
LNM
No	48	50	0.003	37	61	0.010	32	66	0.005
Yes	7	28		5	30		3	32
Clinical stage
I	24	25	0.570	17	32	0.579	20	29	0.008
II	14	23		13	24		10	27
III	10	16		8	18		4	22
IV	7	14		4	17		1	20
Course of disease
≤6 months	13	40	0.004	7	46	<0.001	8	45	0.024
6–12 months	13	15		9	19		7	21
≥12 months	29	23		26	26		20	32
Pathological stage (differentiation)
Well	26	26	0.169	20	32	0.076	18	34	0.026
Moderate	21	32		18	35		15	38
Poor	8	20		4	24		2	26
Prognosis
Recurrence (−)	45	25	<0.001	34	36	<0.001	30	40	<0.001
Recurrence (+)	10	53		8	55		5	58

[i] CD, cluster of differentiation; FOX, forkhead box; IL, interleukin; LNM, lymph node metastasis; LSCC, laryngeal squamous cell carcinoma; TNM, tumor-node-metastasis classification.

Correlation of CD274, FOXP3 and IL-10 expression in LSCC

The results of a Spearman analysis demonstrated that there was a positive expression correlation between CD274 and FOXP3 (rs=0.767, P<0.001), between CD274 and IL-10 (rs=0.640, P<0.001), and between IL-10 and FOXP3 (rs=0.637, P<0.001; Fig. 2).

Figure 2. Correlation of CD274, FOXP3 and IL-10 expression in LSCC. Spearman analysis revealed that there was a positive expression correlation between CD274 and FOXP3 (rs=0.767, P<0.001), between CD274 and IL-10 (rs=0.640, P<0.001), and between IL-10 and FOXP3 (rs=0.637, P<0.001). CD, cluster of differentiation; FOX, forkhead box; IL, interleukin; LSCC, laryngeal squamous cell carcinoma.

Correlation between CD274, FOXP3 and IL-10 expression and the prognosis of patients with LSCC

The KM analysis revealed that the CD274-, FOXP3- and IL-10-negative patients with LSCC experienced a significantly prolonged mean survival time compared with the CD274-, FOXP3- and IL-10-positive patients with LSCC (P<0.05; Fig. 3). A multivariate Cox regression analysis of the effects of the expression of CD274, FOXP3 and IL-10 on the prognosis of LSCC yielded hazard ratios >1 and P<0.05, which indicated that the expression of these three proteins was a significant risk factor for LSCC (Table IV).

Figure 3. Correlation between the expression of CD274, FOXP3 and IL-10 and the prognosis of patients with LSCC. The Kaplan-Meier analysis revealed that (A) CD274-, (B) FOXP3- and (C) IL-10-negative patients with LSCC experienced a significantly prolonged mean survival time compared with the CD274-, FOXP3- and IL-10-positive patients with LSCC (P<0.05). CD, cluster of differentiation; FOX, forkhead box; IL, interleukin; LSCC, laryngeal squamous cell carcinoma.

Table IV. Analysis of prognostic factors in LSCC.

Table IV.

Analysis of prognostic factors in LSCC.

Factor	B	SE	Wald	P-value	HR	95% CI
CD274	0.109	0.045	5.986	0.014	1.115	1.022	1.217
FOXP3	0.132	0.042	9.777	0.002	1.141	1.050	1.239
IL-10	0.076	0.037	4.239	0.040	1.079	1.004	1.160

[i] B, regression coefficients; CD, cluster of differentiation; CI, confidence interval; FOX, forkhead box; HR, hazard ratio; IL, interleukin; LSCC, laryngeal squamous cell carcinoma; SE, standard error; Wald, Wald test χ² value.

CD274, FOXP3 and IL-10-associated ceRNA networks

Using the miRTarBase and StarBase v2.0 databases, it was established that lncRNA- and circRNA-associated ceRNA networks regulated CD274, FOXP3 and IL-10 (Figs. 4 and 5). The two networks revealed that the three genes can be regulated by specific miRNAs including hsa-miR-106a-5p, which regulated IL-10; hsa-miR-31-5p and hsa-miR-210-3p, which regulated FOXP3; and hsa-miR-124-3p, hsa-miR-103a-3p, hsa-miR-214-3p, hsa-miR-16-5p, hsa-miR-326, hsa-miR-708-5p and hsa-miR-29c-3p, which regulated CD274 (Fig. 4). The number of neighboring nodes directly connected to the node indicates the importance of the node in the network. The degree values of the network were then analyzed using Cytoscape software. The results demonstrated that miRNA nodes, including hsa-miR-16-5p, hsa-miR-106a-5p and hsa-miR-103a-3p in the network, exhibited larger degree values (Figs. 4B and 5B). Based on the network, 35 IL-10-associated lncRNAs [including long intergenic non-protein coding RNA 116, X-inactive specific transcript (XIST) and zinc ribbon domain containing 1-AS1], 12 FOXP3-associated lncRNAs (including membrane-associated guanylate kinase, WW, and PDZ domain-containing protein 2-AS3, small nucleolar RNA host gene 16 and DNA polymerase δ-interacting protein 2) and 134 CD274-associated lncRNAs (including long intergenic non-protein coding RNA 662, H19 and HOX transcript antisense RNA) were identified. These lncRNAs may regulate the three genes through their ceRNA effects. The results demonstrated that lncRNA nodes in the network, including XIST, RP11-467L20.9 and AC084219.4, exhibited larger degree values. (Fig. 4C). In addition to lncRNAs, 408 circRNAs that regulate the three genes in LSCC were identified (Fig. 5A). Among them, it was identified that ISY1 splicing factor homolog hsa_circ_001859, tubulin γ complex associated protein 3 hsa-circ_001373 and tubulin α 1b hsa_circ_002179 in the network exhibited larger degree values (Fig. 5C).

Figure 4. lncRNA-associated ceRNA network of CD274, FOXP3 and IL-10 in LSCC. (A) Based on the association among lncRNA, miRNA and mRNA, databases were used to establish the ceRNA network. The number of neighboring nodes directly connected to the node indicates the importance of the node in the network. The degree values of the network were then analyzed using Cytoscape software. (B) Top 10 miRNAs, according to degree values. (C) Top 10 lncRNAs, according to degree values. CD, cluster of differentiation; ceRNA, competitive endogenous RNA; FOX, forkhead box; IL, interleukin; LSCC, laryngeal squamous cell carcinoma; lncRNA, long non-coding RNA; miRNA, microRNA.

Figure 5. circRNA-associated ceRNA network of CD274, FOXP3 and IL-10 in LSCC. (A) Based on the association among circRNA, miRNA and mRNA, databases were used to establish the ceRNA network. The number of neighboring nodes directly connected to the node indicates the importance of the node in the network. The degree values of the network were then analyzed using Cytoscape software. (B) Top 10 miRNAs, according to degree values. (C) Top 10 circRNAs, according to degree values. CD, cluster of differentiation; circRNA, circular RNA; FOX, forkhead box; IL, interleukin; LSCC, laryngeal squamous cell carcinoma; miRNA, microRNA.

Discussion

As an important mechanism of tumorigenesis, immune evasion has been a subject undergoing intense study in cancer research. Tumor development and progression depend on the tumor microenvironment. The infiltration and function of T cells in the tumor microenvironment determine the outcome of tumor immunosurveillance. The recently identified Tregs are a subpopulation of T cells that possess immunoregulatory functions. Tregs possess two major functions: Induction of immune incompetence and immunosuppression (18,19). FOXP3 is a characteristic surface marker of Tregs. It has previously been identified that the proportion of FOXP3+ Tregs is increased in the peripheral blood of patients with cancer, including lung, breast, pancreatic and colon cancer (20–23). These findings indicated that FOXP3+ Tregs are closely associated with tumor development. The present study identified that FOXP3 expression was not only increased in LSCC but was also negatively correlated with the survival rate of patients with LSCC. Such a finding indicated that FOXP3+ Tregs may promote the malignant progression of LSCC. The correlation analysis of clinical factors revealed that the expression levels of FOXP3 were significantly elevated in patients with a long disease course and in those who developed lymph node metastasis. These results demonstrated that FOXP3 was closely associated with the degree of malignancy of LSCC.

The promoter effect of Tregs on cancer immune evasion depends on the expression of inhibitory molecules, including IL-10, which are secreted by Tregs themselves (24). Previous studies have demonstrated that Tregs initiate inhibitory signals through the secretion of IL-10 and other cytokines, which inhibits the activity of functional T lymphocytes and suppresses tumor immune responses (9,25). IL-10 not only induces immune tolerance through the inhibition of the dendritic cell-mediated transformation of T cells into cytotoxic T lymphocytes, but also induces immunosuppression through the inhibition of the antigen-presenting function of antigen-presenting cells (26–28). In certain patients with cancer, IL-10 expression may serve as an independent negative prognostic factor (29,30). The present study identified that IL-10 was highly expressed in LSCC. In addition, the expression levels of IL-10 exhibited a significant negative correlation with the survival rate of patients with LSCC. These findings indicated that IL-10 may be closely associated with the malignant progression of LSCC. Analysis of the clinicopathological factors demonstrated that the expression levels of IL-10 were significantly positively correlated with the TNM stage of cancer, lymph node metastasis, course of the disease and the pathological grade of cancer in patients with LSCC. The results further demonstrated that IL-10 may be involved in the malignant progression of LSCC. In addition, a significant correlation between IL-10 and FOXP3 expression was identified, which indicated that FOXP3+ Tregs may affect the malignant progression of LSCC through the secretion of IL-10.

It has previously been demonstrated that the expression of negative costimulatory molecules in the tumor microenvironment is involved in tumor escape from the immune system (10). Negative costimulatory molecules are important factors that constitute the tumor microenvironment. CD274 is one of the primary members of the B7 family of negative costimulatory molecules, which is closely associated with the progression of various tumors, including melanoma and glioma (11,31). Previous studies have reported that CD274 exerts its effects on tumors through the regulation of tumor immunity. For example, CD274 suppresses type 1 T helper cell-based immune responses and induces T cell apoptosis in vitro via the inhibitory receptor programmed cell death protein 1, which aids the escape of tumor cells from the immune system (32,33). The present study demonstrated that the expression levels of CD274 in LSCC were positively correlated with the TNM stage of cancer, lymph node metastasis and the pathological stage of cancer. In addition, the survival rate was significantly reduced in patients with high CD274 expression compared with those with low CD274 expression. These findings indicated that CD274 may be involved in the malignant progression of LSCC. In addition, CD274 and FOXP3 co-expression was detected in LSCC tissues, thus indicating that CD274 not only inhibits the proliferation of activated T cells but also exerts an immunosuppressive effect through an increase in the number of FOXP3+ Tregs, which promotes the progression of LSCC. Therefore, CD274 may serve as a potential target for LSCC immunotherapy, whereas CD274 expression may serve as an important prognostic indicator of LSCC. Furthermore, in addition to the finding that CD274 was co-expressed along with FOXP3, CD274 expression was also positively correlated with that of IL-10. Based on the aforementioned results, the following hypothesis was proposed: A high level of CD274 expression increases the number of FOXP3+ Tregs, and, as a result, the expression of the immunosuppressive molecule IL-10 is increased, ultimately promoting LSCC immune evasion and enhancing the degree of malignancy of LSCC. This hypothesis is likely to represent an important mechanism underlying LSCC immune evasion; however, this hypothesis needs to be verified in the future through large-scale scientific studies.

To understand the potential molecular mechanism underlying the correlation among these three genes in LSCC, miRTarBase and StarBase databases were used to identify the miRNAs and lncRNAs that regulated the three genes. The results demonstrated that certain miRNAs in the network exhibited larger degree values, including hsa-miR-16-5p and hsa-miR-103a-3p. Previous studies have demonstrated that miRNAs may serve important roles in cancer progression. Hudcova et al (34) identified that reduced expression of miR-29c-3p in tumor tissue was associated with worse relapse-free survival rate in head and neck cancer. Ye et al (35) indicated that miR-106a-5p was associated with the expression of circulating exosomes and immune escape in human nasopharyngeal carcinoma. In addition to miR-29c-3p and miR-106a-5p, the association between other miRNAs and head and neck cancer is not clear; however, their roles in other cancers have been reported. Li et al (36) reported that miR-16-5p is able to control the malignant development of osteoarthritis in chondrocycytes. Wang et al (37) demonstrated that miR-326 can regulate cell proliferation and migration in lung cancer. In addition, miR-106a-5p can inhibit the proliferation and migration of astrocytoma cells and promote apoptosis (38). Based on these findings, it may be hypothesized that these miRNAs exhibit the potential to serve important roles in LSCC.

Salmena et al (15) introduced the concept of ceRNAs in 2011, which states that all types of RNA molecules (e.g., lncRNA and circRNA) are able to mutually regulate one another by competitively binding to miRNAs as long as they share common miRNA-binding sites. In the present study, the effects of ceRNAs in the 1ncRNA-miRNA-mRNA regulatory network in LSCC were evaluated. The results demonstrated that certain lncRNAs in the network exhibited larger degree values, including XIST, H19 and metastasis associated lung adenocarcinoma transcript 1 (MALAT1). Numerous studies have identified that these lncRNAs are associated with the process of cancer. Huang et al (39) demonstrated that a reduction in XIST expression in breast tissue can upregulate Akt phosphorylation and tumor cell viability. Li et al (40) reported that the lncRNA H19 can promote cell invasion in nasopharyngeal carcinoma. In addition, it has been identified that MALAT1 can serve roles in numerous types of cancer, including breast and pancreatic cancer (41,42). Chen et al (43) revealed that MALAT1 is an important target of cisplatin and paclitaxel, and may have the potential as a novel molecular target for the treatment of patients with LSCC. Based on these findings, it may be hypothesized that these lncRNAs exercise their ceRNA effects to regulate LSCC through upregulating IL-10, FOXP3 or CD274. The lncRNAs predicted in the present study may serve as a basis for new studies regarding the mechanism of tumor immunity in LSCC. In addition, the circRNA-related ceRNA network identified 408 circRNAs that regulated the three genes. It has been identified that some of these genes bind with miRNAs and act as natural miRNA sponges to inhibit the activities of associated miRNAs. For instance, Xie et al (44) demonstrated that circ_001569 can promote the proliferation and invasion of colorectal cancer via its ceRNA effect. The present study identified circ_001569 in the ceRNA network, thus suggesting that circ_001569 may serve roles in LSCC. As research on circRNA remains in the early stages, the effects of more circRNAs on cancer are yet to be elucidated. However, the present study indicated that these circRNAs may bind with the miRNAs that regulate CD274, FOXP3 and IL-10. Based on these findings, it may be hypothesized that these circRNAs exhibit the potential to be ceRNAs of the three genes that regulate the progression of LSCC.

In conclusion, the present study reported that the expression levels of CD274, FOXP3 and IL-10 were higher in LSCC tissue compared with their expression levels in normal tissues. Notably, there was a strong positive correlation among the expression of these three proteins. In addition, it was identified that the expression of the three proteins was closely correlated with the clinical characteristics of LSCC, and the expression levels of FOXP3, IL-10 and CD274 had a negative association with the survival rate of patients with LSCC. The results of the Cox regression model suggested that the three proteins were prognostic risk factors for LSCC. In addition, the ceRNA network of the three genes was established. Data generated from the present study provided novel information, which may increase understanding regarding the immune evasion mechanisms of LSCC, and aid the development of LSCC immunotherapy.

Acknowledgements

The present study was supported by the Inner Mongolia Natural Science Foundation (grant nos. 2014MS0856 and 2017MS0838), the Natural Science Foundation of China (grant no. 81473499) and the General Research Projects of the Affiliated Hospital of Inner Mongolia Medical University (grant no. NYFY YB 027).

References