Epigenetically‐Upregulated CD98 Shed Light on the Precancerous Diagnosis and Prognosis Prediction of Esophageal Cancer

ABSTRACT

Esophageal squamous cell carcinoma (ESCC) has a significantly low survival rate, primarily due to the lack of diagnostic markers for early diagnosis and effective therapies. Recently, CD98 has been proposed as a specific marker of mature esophageal basal cells, which may be associated with esophageal carcinogenesis. Therefore, we aimed to investigate the clinical significance and biological function of CD98 in ESCC progression, as well as the value of CD98 as a potential new marker for the early diagnosis of ESCC. Through MassARRAY system spectroscopy, DIA proteomics analysis, immunohistochemical and functional experiments, we found hypomethylation‐linked upregulation of CD98 expression, which was associated with poor prognosis, promoted cell proliferation by regulating the cell cycle in ESCC. Furthermore, we not only demonstrated that the range of CD98 expression was consistent with that of dysplastic cells in 73.81% of low‐grade intraepithelial neoplasia (LGIN) and 83.08% of high‐grade intraepithelial neoplasia (HGIN) cases, but also confirmed the expression level of CD98 was positively correlated with the classical diagnosis marker P53. Compared to using P53 alone, the combination of the immunohistochemical markers CD98 and P53 (either one was positive) provided more accurate diagnostic data for LGIN (92.86% vs. 88.10%, p < 0.01) and HGIN (93.85% vs. 73.85%, p < 0.01). In summary, we propose that CD98 is involved in a crucial step in ESCC carcinogenesis, and when combined with P53, may serve as a diagnostic marker for ESCC precancerous lesions.

This study confirmed that hypomethylation‐induced upregulation of CD98 promoted ESCC cell proliferation by regulating the cell cycle and was correlated with poorer overall survival of ESCC patients. Combined panels of immunohistochemistry (CD98 and P53) are recommended for accurate diagnosis of esophageal precancerous lesions.

Abbreviations

DIA

data‐independent acquisition

ESCC

Esophageal squamous cell carcinoma

HGIN

high‐grade intraepithelial neoplasia

IHC

immunohistochemical

LGIN

low‐grade intraepithelial neoplasia

TNM

tumor‐node‐metastasis

1. Introduction

Esophageal squamous cell carcinoma (ESCC) is one of the deadliest human malignancies worldwide. Due to its nonspecific symptoms and the lack of early diagnostic techniques, most patients are diagnosed at advanced stages, with a 5‐year survival rate of less than 20% [1, 2]. Studies have revealed that ESCC undergoes a multifactorial, multi‐stage, and bidirectional transformation process, which can manifest as varying degrees of inflammation, dysplasia, low‐grade intraepithelial neoplasia (LGIN), and high‐grade intraepithelial neoplasia (HGIN), and eventually develop into invasive carcinoma [3, 4, 5]. Patients diagnosed in the early stages of ESCC have a significantly better prognosis, with a 5‐year survival rate of over 90%, after prompt treatment [6, 7]. Unfortunately, there remains a lack of suitable diagnostic biomarkers for early stage ESCC. Although recent work has begun to identify candidate genes for distinguishing ESCC precancerous lesions and has suggested P53 as the most commonly used immunomarker, it lacks high sensitivity and specificity [8, 9, 10]. Thus, exploring specific molecular markers for the early detection of ESCC is crucial for the timely diagnosis and treatment of precancerous lesions, ultimately aiming to reduce the incidence and mortality rates associated with ESCC.

CD98 is a transmembrane protein on the cell surface, encoded by the solute carrier family 3 member 2 (SLC3A2) gene [11]. CD98 not only functions as a crucial chaperone for amino acid transporters, but also forms complexes with integrin β1 to activate integrin‐dependent signaling, thereby contributing to cell adhesion, proliferation and survival [12, 13]. CD98 has been extensively documented as a biomarker associated with adverse outcomes in numerous cancers, including non‐small cell lung cancer [14], astrocytic neoplasms [15], triple negative breast cancer [16], and colorectal cancer [17], emphasizing the crucial role of CD98 in tumorigenesis. The derivation and early inflammatory relevance of CD98, a marker originating from esophageal basal cells, have been corroborated [18]. In addition, CD98 has garnered significant attention as a crucial gene involved in ferroptosis [19, 20]. Although numerous studies have established ferroptosis as a vital mechanism influencing the occurrence and progression of ESCC [21, 22], the clinicopathological significance of CD98 expression in ESCC progression remains unclear.

Aberrant DNA methylation is critical in human cancer progression. In ESCC, CpG island promoter hypermethylation and global DNA hypomethylation are key epigenomic features, while recent studies identify specific hypomethylation events during its initiation and progression [23, 24, 25, 26, 27]. In this study, we investigated the expression and function of CD98 in precancerous lesions and ESCC, and explored the possibility of CD98 as a prognostic indicator and potential therapeutic target for patients with ESCC, especially in the early stage.

2. Material and Methods

2.1. Patients and Tissue Samples

A total of 318 esophageal cancer specimens diagnosed with ESCC, 232 paired normal tissue samples (more than 2 cm from the tumor), and 157 endoscopic submucosal dissection specimens (comprising 42 LGIN, 65 HGIN, and 50 invasive squamous cell carcinoma) were collected from patients diagnosed with histological pathology at the Department of Pathology, Nanjing Drum Tower Hospital. All the cases were diagnosed by two senior pathologists, without any age, gender, or disease stage restrictions. As shown in Tables S1 and S2, detailed clinicopathological variables, such as depth of infiltration, presence of distant metastasis, and tumor location, were extracted from the patients' medical records. Lymph node status, tumor‐node‐metastasis (TNM) stage, and differentiation grade were classified based on the Cancer Staging Manual (8th edition, issued in 2016 by the American Joint Committee on Cancer/Union for International Cancer Control). Our study was approved by the Ethics Committee of Nanjing Drum Tower Hospital.

2.2. Histology and Immunohistochemical Staining

Tissue microarray (TMA) and immunohistochemical (IHC) analyses were conducted as previous studies [28]. All pathological specimens were fixed in 10% formalin phosphate buffered saline overnight, gradient dehydrated, embedded in paraffin, and then sectioned into slices of approximately 4 μm thickness for hematoxylin and eosin or IHC staining. IHC staining was performed using an automatic Ventana Bench Mark Ultra system (Roche Diagnostics, Basel, Switzerland), with negative and positive controls. The primary antibody against CD98 was purchased from Proteintech (Chicago, USA) at a dilution of 1:1000, and the P53 and Ki67 antibodies were purchased from ZSGB‐BIO (Beijing, China) at a dilution of 1:50.

All IHC results were assessed independently by two senior pathologists, and CD98 expression was semi‐quantitatively scored as described previously to obtain the IHC score (IS) [28].

2.3. Quantitative Analysis of CD98 DNA Methylation

CD98 methylation was detected in 178 samples, including 118 ESCC samples and 60 normal tissue samples. A DNA extraction kit (Qiagen Inc., Shanghai, China) was used to isolate DNA from the tissues. A NanoDrop spectrophotometer (NanoDrop Technologies Inc.) and gel electrophoresis were utilized to guarantee the purity and quality of the DNA. Genomic DNA was maintained at −20°C until used as a template for PCR. PCR detection of bisulfite‐transformed genomic DNA samples was performed with human CD98 primers 5ʹ‐AGG AAG AGA GTG GGT AAG GGG TAG TTT AGA ATA GG‐3ʹ (forward) and 5ʹ‐CAG TAA TAC GAC TCA CTA TAG GGA GAA GGC TCC AAA CAA CTA CAA CAC AAA AAA CA‐3ʹ (reverse). Genomic DNA was treated with bisulfite according to the manufacturer's protocol (ZYMO, CA, USA) using an EZ DNA methylation kit. CpG Island Prediction was used to determine the sequence of the CpG islands (https://www.ebi.ac.uk/Tools/seqstats/emboss_cpgplot/, Version: 2.3.0 + 24213590). The analyzed region and CpG sites of the CD98 promoter are shown in Figure 2. We used EpiDesigner software to design the primer set for methylation analysis of the CD98 promoter region (http://epidesigner.com). Matrix‐assisted laser desorption/ionization‐time‐of‐flight mass spectrometry (MALDI‐TOF MS) was performed as described previously [3, 29]. The assessment of methylation levels in each sample was based on the average methylation value of all CpG units within the CD98 promoter region.

FIGURE 2.

CD98 specific CpG unit methylation is negatively correlated with its expression and the prognosis of ESCC. (A) Comparison of CpG units methylation of CD98 promoter region in ESCC and normal samples. Each row represents one sample. Each column shows clusters of CpG units, which are individual CpG sites or combinations of CpG sites. The color gradient between red and blue indicates the methylation of each CD98 CpG unit in each sample, ranging from 0 to 1. Gray represents insufficient or missing data. (B) Methylation levels of 14 informative CpG sites in the CD98 promoter in normal and ESCC samples. *p < 0.05 (Mann–Whitney U test). (C) Analysis of scatterplots and simple linear regression graphically displaying the correlation between the methylation level of each CpG_23.24.25.26.27 and CD98 expression in ESCC samples by Spearman correlation coefficient analysis. (D) Kaplan–Meier survival curves of ESCC patients stratified by CD98 CpG_23.24.25.26.27 methylation.

2.4. Cell Culture and Cell Transfection

Two ESCC cell lines (EC9706 and Eca109) were obtained from the Institute of Biochemistry and Cell Biology of the Chinese Academy of Sciences. These cell lines were cultured in 1640 Medium (Gibco, CA, USA), which was supplemented with 10% fetal bovine serum, 100 U/mL penicillin, and 100 mg/mL streptomycin, and maintained in a moist air at 37°C with 5% CO₂. They were authenticated by Short Tandem Repeat DNA profiling. Mycoplasma contamination was not found in these cell lines.

Oligonucleotide small interfering RNA (siRNA) duplexes for CD98 were obtained from GenePharma (Shanghai, China). The specific siRNA sequences employed for CD98 were as follows: CD98 siRNA‐1, 5′‐CUC AAC UUC UCC GAC UCU A‐3′ and CD98 siRNA‐2, 5′‐CAG AUC CUG AGC CUA CUC GAA‐3′. The negative control siRNA sequences was: 5′‐UUC UCC GAA CGU GUC ACG UTT‐3′. The siRNA sequences were subjected to BLAST analysis to minimize any potential off‐target effects. The cells were transfected with siRNA targeting CD98 or RNAi negative control duplexes for 48 h using Lipofectamine 2000 (Invitrogen, CA, USA), according to the manufacturer's instructions.

2.5. Data‐Independent Acquisition Analysis and Proteomics Analysis

Data‐independent acquisition (DIA) and bioinformatics analyses were carried out by Genedenovo Biotechnology Company (Guangzhou, China). Six ESCC cell samples (three in each group) were pre‐processed using the iST Sample Preparation Kit (PreOmics, Germany), and total proteins were obtained. Protein identification and analysis were performed using the UniProt protein database, and FDR cutoffs at the precursor and protein levels were applied at 1%. Raw data of DIA were processed and analyzed by Spectronaut X (Biognosys AG, Switzerland). A criterion of |fold change| > 1.2 and p‐value < 0.05 was used to define the significantly different proteins between the two groups. Finally, we subjected the significantly different proteins to the Gene Ontology (GO) database (http://www.geneontology.org/) and Kyoto Encyclopedia of Genes and Genomes (KEGG) database analysis to recognize the main biological functions and enriched pathways that significantly different proteins exercise.

2.6. Additional Methods

Additional methods are described in Data S1.

2.7. Statistical Analysis

The R programming language, SPSS22.0 (SPSS, Chicago, USA) and GraphPad Prism 5.0 (GraphPad Software Inc.) were used to visualize the data. The Mann–Whitney U test was used to compare methylation between groups. Correlations were calculated by Pearson or Spearman correlations. Fisher's exact test was used if necessary. Survival curves were plotted using the Kaplan–Meier method and compared with a log‐rank test. For data that follow a normal distribution, we used mean ± SE to represent the data. Differences were statistically significant at p < 0.05 (two sided).

3. Results

3.1. Hypomethylation‐Related Overexpression of CD98 Indicates a Poor Prognosis for ESCC

Initially, we conducted a comparative analysis of CD98 expression levels between 33 cancer and matched normal samples based on TCGA data. As depicted in Figure 1A, pan‐cancer analysis revealed that CD98 mRNA was overexpressed in 19 types of cancer, particularly in esophageal carcinoma (ESCA). Subsequently, we focused on the role of CD98 in ESCC, which accounts for nearly 90% of all ESCA cases in China. IHC analysis revealed a significant upregulation of CD98 protein, localized on the cell membrane, in ESCC tumors. In contrast, its expression was confined to the basal cell layer in normal esophageal tissues (Figure 1B,C). Consistent with this, CD98 expression with a four‐grade distribution (0–1, 2–3, 4–8, and 9–12) was significantly different between the adjacent normal tissues and ESCC tissues (Figure 1D,E). As shown in Figure 1D,E, the CD98 expression score was notably elevated in the tumor group. The high expression rate of CD98 in ESCC tissues was 77.0% (245/318), whereas it was only 0.9% in adjacent normal tissues (Table 1). Interestingly, we found that CD98 was highly expressed in the differentiation status of ESCC tissues (p = 0.001), but there was no significant correlation between CD98 expression and other clinicopathological variables, such as gender, age, and TNM stage (Table 2). Kaplan–Meier survival curves revealed that the median disease‐specific survival was 42 months for patients with high CD98 expression and 78 months for patients with low CD98 expression in the cohort, suggesting that high CD98 expression was correlated with a low survival rate of patients with ESCC (p = 0.022, log‐rank test, Figure 1F).

FIGURE 1.

CD98 is highly expressed in ESCC and correlates with poor prognosis. (A) CD98 mRNA levels in different cancer types from TCGA data. (B) IHC analysis of the CD98 protein in adjacent normal and ESCC tissues. Right image magnification ×40; left image magnification ×200. (C) Quantification of positive staining for the CD98 protein in (B). (D, E) Distribution ratios of adjacent normal tissues and ESCC tissues in four‐level score (0–1, 2–3, 4–8 and 9–12) of CD98 expression. (F) Kaplan–Meier survival curves of ESCC patients stratified by CD98 expression. ACC, adrenocortical carcinoma; BLCA, bladder urothelial carcinoma; BRCA, breast invasive carcinoma; CESC, cervical squamous cell carcinoma and endocervical adeno carcinoma; CHOL, cholangio carcinoma; COAD, colon adenocarcinoma; DLBC, lymphoid neoplasm diffuse large B‐cell lymphoma; ESCA, esophageal carcinoma; GBM, glioblastoma multiforme; HNSC, head and neck squamous cell carcinoma; KICH, kidney chromophobe; KIRC, kidney renal clear cell carcinoma; KIRP, kidney renal papillary cell carcinoma; LAML, acute myeloid leukemia; LGG, brain lower grade glioma; LIHC, liver hepatocellular carcinoma; LUAD, lung adenocarcinoma; LUSC, lung squamous cell carcinoma; MESO, mesothelioma; OV, ovarian serous cystadenocarcinoma; PAAD, pancreatic adenocarcinoma; PCPG, pheochromocytoma and paraganglioma; PRAD, prostate adenocarcinoma; READ, rectum adenocarcinoma; SARC, sarcoma; SKCM, skin cutaneous melanoma; STAD, stomach adenocarcinoma; TGCT, testicular germ cell tumors; THCA, thyroid carcinoma; THYM, thymoma; UCEC, uterine corpus endometrial carcinoma; UCS, uterine carcinosarcoma; UVM, uveal melanoma.

TABLE 1.

Expression of CD98 in ESCC and normal tissues.

Group	CD98		p
	Low expression (< 6)	High expression (≥ 6)
Normal	230 (99.1%)	2 (0.9%)
ESCC	73 (23.0%)	245 (77.0%)	< 0.001***

^*** p < 0.001, as determined by Pearson's χ ² test.

TABLE 2.

Correlation between CD98 expression and clinicopathological features of ESCC.

Variables	CD98		χ2	p
	Low expression(< 6)	High expression (≥ 6)
Gender
Male	98	90	0.802	0.371
Female	61	45
Age
≥ 65	100	88	0.166	0.683
< 65	59	47
Tumor location
Upper + Middle	101	98	2.746	0.097
Lower	58	37
Differentiation status
Well + Moderate	135	130	10.654	0.001**
Poor	24	5
Infiltration depth
T1 + T2	57	50	0.045	0.833
T3 + T4	102	85
Lymph node metastasis
No	77	67	0.042	0.837
Yes	82	68
Metastasis
No	151	129	0.055	0.814
Yes	8	6
TNM stage
I and II	86	71	0.066	0.798
III and IV	73	64

Note: Metastasis: metastasis confirmed at the time of diagnosis of primary ESCC.

^** p < 0.01, indicates a significant association among the variables.

In addition, to investigate whether abnormal expression of CD98 is regulated by DNA methylation, we further analyzed the methylation status of the CD98 promoter region. Although no differentially methylated regions were detected (Figure S1A), CpGs residing within CD98 were found to be strongly hypomethylated in tumor samples compared to adjacent normal tissues (Figure 2A,B), particularly at CpG_23.24.25.26.27 (Table 3). Notably, methylation of CpG_23.24.25.26.27 was significantly and negatively correlated with CD98 protein expression (Figure 2C). In the TCGA Illumina 450 k infinite methylation bead chip, we further observed that the methylation status of three CpG sites (cg03193698, cg13414478, and cg17300723) was negatively correlated with CD98 expression (Figure S1B–D). Moreover, the results from the Kaplan–Meier analysis demonstrated a significantly shorter overall survival for esophageal cancer patients with CpG_23.24.25.26.27 hypomethylation compared to those with relatively hypermethylation (log‐rank p = 0.034) (Figure 2D and Figure S1E–O). Subsequently, we evaluated the association between the quantitative methylation patterns of CpG_23.24.25.26.27 of CD98 and the clinicopathological parameters of 66 patients with ESCC. The hypomethylation rate of CpG_23.24.25.26.27 was positively correlated with patient age, as shown in Table 4 (r = 0.210, p = 0.043).

TABLE 3.

Comparison of methylation levels of CpG units in the CD98 promoter region of ESCC and para‐cancer normal esophageal mucosal epithelial tissues.

CpGs	Normal	Tumor	Fold (%)	Diff (−)	p
	X ± SE	X ± SE
CpG_1	0.0869 ± 0.0302	0.0610 ± 0.0184	0.7025	−0.0259	0.4533
CpG_2	0.3931 ± 0.0543	0.3785 ± 0.0364	0.9627	−0.0147	8.20E‐01
CpG_3.4	0.0219 ± 0.0112	0.0467 ± 0.0169	2.1275	0.0247	0.3321
CpG_6	0.1191 ± 0.0309	0.1156 ± 0.0248	0.9706	−0.0035	9.30E‐01
CpG_9.10	0.4881 ± 0.0399	0.5172 ± 0.0274	1.0596	0.0291	5.50E‐01
CpG_12.13.14.15.16	0.5123 ± 0.0485	0.6084 ± 0.0338	1.1878	0.0962	1.10E‐01
CpG_17	0.0515 ± 0.0499	0.1353 ± 0.0436	2.6278	0.0838	0.2545
CpG_18	0.1945 ± 0.0648	0.1472 ± 0.0345	0.7566	−0.0473	0.4845
CpG_21	0.1312 ± 0.0430	0.0936 ± 0.0291	0.7135	−0.0376	0.4703
CpG_23.24.25.26.27	0.7697 ± 0.0409	0.6382 ± 0.0365	0.8292	−0.1315	3.50E‐02*
CpG_29.30	0.1512 ± 0.0571	0.1506 ± 0.0346	0.9958	−0.0006	0.9921
CpG_33	0.4614 ± 0.0591	0.3788 ± 0.0444	0.821	−0.0826	2.90E‐01
CpG_34	0.0789 ± 0.0224	0.0953 ± 0.0187	1.2082	0.0164	6.10E‐01
CpG_37	0.1312 ± 0.043	0.0936 ± 0.0291	0.7134	−0.0376	0.4703

Note: Fold (%): the ratio of the methylation levels of the two groups; Diff (−): the difference of methylation levels between the two groups.

^* p‐value < 0.05.

TABLE 4.

Correlation between CD98 CpG_23.24.25.26.27 methylation levels and clinicopathological features of ESCC.

Clinical characteristics	n	X ± SE	p
Gender
Male	47	0.6217 ± 0.0453	0.501
Female	19	0.6789 ± 0.0603
Age
≥ 65	33	0.720 ± 0.220	0.014*
< 65	33	0.780 ± 0.300
Tumor location
Upper	6	0.8150 ± 0.1221	0.202
Middle	43	0.6284 ± 0.0458
Lower	17	0.6006 ± 0.0683
Differentiation status
Well	20	0.5960 ± 0.0821	0.945
Moderate	32	0.6656 ± 0.0463
Poor	14	0.6357 ± 0.0731
Lymph node metastasis
Yes	31	0.6013 ± 0.0530	0.266
No	35	0.6709 ± 0.0505
TNM stage
I and II	35	0.6723 ± 0.0507	0.232
III and IV	31	0.5997 ± 0.0526

Note: N stand for the number of analyzed patients.

^* p‐value < 0.05.

Taken together, the aforementioned results indicated that upregulation of CD98 expression, induced by promoter hypomethylation, plays a crucial role in the development of ESCC.

3.2. Overexpression of CD98 Is Relevant to the Proliferation of ESCC

It is well established that ESCC progression is associated with an increased degree of inflammation, which was corroborated by our results (Figure S2A,B). Furthermore, we found a significant positive correlation between any two of the inflammation degree, expression of CD98, and proliferative marker Ki67 (Figure S2C–E). Our results implied that CD98 is involved in inflammation and cell proliferation during ESCC progression.

To investigate the underlying mechanisms, we initially constructed CD98 knockdown (KD) cells by siRNA transfection (Figure 3A), and then employed a DIA‐MS based quantitative proteomics strategy to obtain the protein expression profiles in the CD98 KD group compared with the control group. A total of 82,616 peptides and 8053 proteins were finally screened in three biological replicates, and subsequently, 185 up‐regulated proteins and 121 down‐regulated proteins in the CD98 KD group were identified. The heat map showed the TOP40 genes with high and low expression differences (Figure 3B). To gain insight into the possible biological functions of CD98, we subjected 306 significantly different proteins in CD98 KD cells to bioinformatic enrichment analysis using the GO and KEGG databases. The results showed that the significantly different proteins were mainly enriched in megakaryocytes, myeloid cell differentiation‐related processes, and DNA replication (Figure 3C). In addition to the enrichment of known CD98 participating pathways, such as immune and inflammation pathways, our data revealed that CD98 could also play a role in multiple previously less well‐appreciated cellular functions, such as fatty acid and amino acid metabolism, oxidative phosphorylation, and estrogen signaling pathways (Figure 3D). Proteome analysis suggested multiple new roles for CD98 in cellular physiology, especially in inflammation and proliferation.

FIGURE 3.

Differential expression analysis of CD98 KD and control ESCC cell lines. (A) Western blotting showing the expression of CD98 in EC9706 and Eca109 cells transfected with RNAi negative control (control) or si‐CD98 (siRNA). (B) Heat map of TOP40 differentially expressed genes. (C) Circle diagram of the GO term analysis. The first circle indicates the top 20 GO terms; the second circle indicates the number of this GO term and the Q value in the background of differential proteins; and the third circle represents the bar chart of the proportion of upregulated and downregulated proteins. (D) Classification statistical bubble diagram of the KEGG enrichment (top 20). Different colors indicate different functional categories.

To further explore the correlation between CD98 expression and tumorigenesis, we determined whether CD98 influences the clonogenic potential of ESCC cells via colony formation assays. We first assessed the expression levels of CD98 in different esophageal cancer cell lines and selected EC9706 and Eca109 cell lines for subsequent functional experiments (Figure S3A). After knockdown of CD98 by specific siRNA, the results revealed a significant reduction in both colony size and colony number in EC9706 and Eca109 cells (Figure 4A,B). Subsequently, we assessed cell proliferation by CCK‐8 assays and observed a significant suppression in the viability of EC9706 and Eca109 cells after CD98 silencing (Figure 4C,D). Furthermore, cell cycle analysis indicated that knockdown of CD98 expression led to a G1 phase block in ESCC cells (Figure 4E,F). Consistent with this, overexpression of CD98 promoted cell proliferation (Figure S3B–D). Collectively, these results suggested that CD98 is indispensable for ESCC cell proliferation.

FIGURE 4.

Knockdown of CD98 inhibits ESCC cell proliferation via G1 arrest. (A, B) Knockdown of CD98 inhibits cell growth as determined by colony formation assays. (C, D) CCK‐8 assays showed that silencing of endogenous CD98 inhibited cell proliferation. (E, F) Knockdown of CD98 induced cell cycle arrest at the G1 phase. *p < 0.05, **p < 0.01.

3.3. Correlation of CD98 Distribution and Expression Pattern With ESCC Progression

Next, a total of 255 cases, encompassing normal, LGIN, HGIN, and invasive carcinoma, were employed to evaluate the IHC staining patterns of CD98 in esophageal precancerous lesions (Figure 5A). In normal specimens, CD98 was exclusively expressed on the cell membrane of the basal layer cells of the esophagus, exhibiting a linear growth pattern. However, the membrane protein CD98 exhibited bulky growth and was predominantly expressed in the lower 1/2 of the epithelial layer in 73.81% of LGIN samples, exceeded 1/2 the height of the epithelium in 83.08% of HGIN samples, exhibiting a notable concurrence with the confined extent of dysplasia cells. Furthermore, CD98 portrayed a diffusely invasive growth pattern in invasive carcinoma samples. Based on this, we summarized the expression pattern of CD98 in the esophageal epithelium (Figure 5A and Table 5), suggesting that the expression pattern of CD98 varies with the development of ESCC.

FIGURE 5.

Correlation of CD98 distribution and expression patterns with ESCC progression. (A) Representative HE staining (left panel) and CD98 IHC staining (middle panel) in normal, LGIN, HGIN, and invasive carcinoma tissues. Scale bar, 50 μm. Cartoons demonstrating CD98 expression (CD98‐positive cells appear dark brown cell membranes) in the development of esophageal invasive carcinoma (right panel). Relationship between pathological changes and (B) CD98 and (C) P53 score. (D) SPSS correlation analysis between CD98 score and P53 score.

TABLE 5.

Expression pattern of CD98 in esophageal epithelium.

Cancer progression	Histology		Immune phenotype
	Cytology	dysplastic cell range	CD98+ cell distribution
Normal	Normal epithelium or Mild hyperplasia	—	Linear outgrowth, basal cell layer
LGIN	Mild–moderate dysplasia	Epithelial layer < 1/2	Bulky outgrowth, involving epithelial layer < 1/2
HGIN	Severe dysplasia	Epithelial layer > 1/2	Bulky outgrowth, involving epithelial layer > 1/2
Infiltrating carcinoma	Severe epithelium	Breakthrough of basement	Infiltrating diffuse growth, into lamina propria

The distribution of CD98 expression in four‐grade (0–1, 2–3, 4–8, and 9–12) exhibited a significant dissimilarity in the multi‐stage development of ESCC (Figure 5B). P53, a classic IHC marker for esophageal squamous intraepithelial neoplasia, exclusively displays nuclear expression. As shown in Figure 5C, the expression of P53 was intimately associated with different stages of ESCC, paralleling the distribution of CD98. Correlation analysis unequivocally revealed a significant positive correlation between CD98 and P53 expression (p < 0.01, Figure 5D).

3.4. CD98 Combined With P53 Enhance the Detection Rate of Precancerous Lesions in ESCC

The unique distribution and expression pattern of CD98 prompted us to hypothesize that it may serve as a diagnostic criterion for ESCC progression. Remarkably, we identified four prevalent distribution patterns of CD98 and P53 proteins in basal and proliferating cells of LGIN (Figure 6A) and HGIN (Figure 6B). As depicted in Figure  6A and Table 6, co‐expression of CD98 and P53 proteins (CD98⁺/P53⁺) was observed in basal cells and proliferative cells, accounting for 69.05% (29 out of 42) in pattern 1; esophageal epithelial tissue containing CD98⁺/P53⁺ basal cells and CD98⁻/P53⁺ proliferative cells accounted for 19.05% (8 out of 42) in pattern 2; CD98⁺/P53⁻ basal cells and CD98⁺/P53⁻ proliferative cells accounted for 4.76% (2 out of 42) in pattern 3; and approximately 7.14% (3 out of 42) of the cases displayed CD98⁺/P53⁻ basal cells and CD98⁻/P53⁻ proliferative cell staining patterns (pattern 4). IHC staining was also conducted for CD98 and P53 in 65 patients with HGIN (Figure 6B and Table 6) and found that 63.08% (41 out of 65) of samples showed co‐expression of CD98 and P53 proteins in basal cells and proliferative cells (pattern 1); esophageal epithelial tissue containing CD98⁺/P53⁺ basal cells and CD98⁻/P53⁺ proliferative cells accounted for 10.77% (7 out of 65, pattern 2); CD98⁺/P53⁻ basal cells and CD98⁺/P53⁻ proliferative cells accounted for 20.00% (13 out of 65, pattern 3); the remaining 4 cases exhibited CD98⁺/P53⁻ basal cells and CD98⁻/P53⁻ proliferative cell staining patterns (pattern 4).

FIGURE 6.

Expression patterns of CD98 and P53 in esophageal intraepithelial neoplasia. (A) Representative images showing four different expression patterns in LGIN tissues: CD98⁺/P53⁺ (pattern 1), CD98⁻/P53⁺ (pattern 2), CD98⁺/P53⁻ (pattern 3), and CD98⁻/P53⁻ (pattern 4). (B) Representative images showing four different expression patterns in HGIN tissues: CD98⁺/P53⁺ (pattern 1), CD98⁻/P53⁺ (pattern 2), CD98⁺/P53⁻ (pattern 3), and CD98⁻/P53⁻ (pattern 4). Scale bar, 50 μm. TOP panel: HE staining; middle panel: CD98 IHC staining; bottom panel: P53 IHC staining.

TABLE 6.

Four expression patterns of CD98 and P53 in esophageal intraepithelial neoplasia samples.

Group	Pattern	N (%)	Distribution	CD98	P53
LGIN	CD98+/P53+	29 (69.05%)	Basal cells	+	+
			Simple hyperplastic cells	+	+
	CD98−/P53+	8 (19.05%)	Basal cells	+	+
			Simple hyperplastic cells	−	+
	CD98+/ P53−	2 (4.76%)	Basal cells	+	−
			Simple hyperplastic cells	+	−
	CD98−/ P53−	3 (7.14%)	Basal cells	+	−
			Simple hyperplastic cells	−	−
HGIN	CD98+/P53+	41 (63.08%)	Basal cells	+	+
			Simple hyperplastic cells	+	+
	CD98−/P53+	7 (10.77%)	Basal cells	+	+
			Simple hyperplastic cells	−	+
	CD98+/ P53−	13 (20.00%)	Basal cells	+	−
			Simple hyperplastic cells	+	−
	CD98−/ P53−	4 (6.15%)	Basal cells	+	−
			Simple hyperplastic cells	−	−

Note: N stand for the number of analyzed patients.

Therefore, compared to the application of P53 alone, the combination of the IHC markers CD98 and P53 (either one was positive) demonstrated a significantly improved accuracy in the diagnosis of LGIN (92.86% vs. 88.10%, p = 0.001) and HGIN (93.85% vs. 73.85%, p = 0.004). Finally, Figure 7 visualized the diagnostic patterns of CD98 and P53 proteins in basal and proliferating cells in ESCC precancerous lesions, including CD98⁺/P53⁺ (Figure 7A), CD98⁺/P53⁻ (Figure 7B), and CD98⁻/P53⁺ (Figure 7C).

FIGURE 7.

Schematic graph demonstrating different expression patterns of CD98 and P53 in the development of invasive esophageal carcinoma. CD98‐positive cells appeared as dark brown cell membranes and P53‐positive cells showed dark brown nuclei. In the occurrence and progression of esophageal squamous cell carcinoma, CD98 and P53 showed distinct combinatorial expression patterns, including CD98⁺/P53⁺ (A), CD98⁺/P53⁻ (B), and CD98⁻/P53⁺ (C).

4. Discussion

CD98, which plays a pivotal role in numerous biological processes through its unique domain, has been reported to be upregulated in various human solid and hematologic cancers, and is associated with tumor growth [30, 31]. One study utilized an anti‐CD98 antibody against heavy chains and showed that the growth of human tumor cells was inhibited in vitro [32]. Additionally, it has been reported that suppressing CD98 expression significantly reduces the survival of ovarian cancer cells treated with cisplatin and promotes cell migration [33]. Overall, these studies suggest that CD98 may be a potential target for therapeutic interventions in cancer. We found for the first time that CD98 was highly expressed in ESCC and associated with a poor prognosis. Intensive molecular biological studies have shown that epigenetic dysregulation plays a crucial role in the occurrence and development of almost all human cancers. Among them, DNA methylation has been extensively studied as an important epigenetic modification, and a key switch controlling gene expression. Indeed, Zou et al. suggested a negative correlation between CD98 expression and methylation level by pan‐cancer analysis [34], which was confirmed by our results showing elevated CD98 expression in ESCC tissues due to promoter hypomethylation, specifically at CpG_23.24.25.26.27 sites associated with poor prognosis. Thus, epigenetically upregulated CD98 expression is essential for its transcription. Additionally, previous reports have shown that MicroRNA‐7 inhibits CD98 expression during intestinal epithelial cell differentiation [35], while MicroRNA‐7 is lowly expressed in ESCC tissues [36, 37], which also raises another possibility of high CD98 expression in ESCC.

CD98 is widely expressed on the cell membrane and mediates the transmembrane transport of almost all essential amino acids, promoting cell cycle progression, and maintaining normal biological cellular functions [38]. In a study on osteosarcoma, Zhu et al. demonstrated that the downregulation of CD98 resulted in cell cycle arrest in the G2/M phase and reduced cell proliferation [39]. Recently, Wang et al. reported that interferon released by CD8⁺ T cells promotes tumor cell ferroptosis by downregulating CD98 expression [40]. In most cases of esophageal cancer, abnormal activation of the PI3K/AKT pathway can induce changes in esophageal cancer cells, leading to proliferation and metastasis [41, 42]. In this study, the expression of CD98 was parallel to the progression of tumor, and the down‐regulation of CD98 expression resulted in G1 phase arrest and inhibited proliferation of ESCC cells, which may be associated with ferroptosis or PI3K/AKT pathway and requires further exploration. More importantly, we used DIA proteomics to screen out the differentially expressed proteins in ESCC cells after CD98 inhibition. We observed aberrant alterations in ATF2, a key regulator required for the activity of cell cycle proteins such as p53, p21, and cyclin D1. Extensive literature has highlighted the pivotal role of ATF2 in modulating oxidative stress, regulating the cell cycle, and promoting tumor cell proliferation [43, 44, 45]. Our findings suggest a potential correlation between CD98, inflammation, and P53 expression in the progression of esophageal cancer, with the precise role of ATF2 in this context warranting further investigation. Although its specific regulatory mechanism has not been fully elucidated, it is sufficient to demonstrate that CD98 may be necessary for tumor cell growth in ESCC patients with aggressive characteristics, suggesting that the regulation of CD98 could provide a therapeutic approach for ESCC.

Precancerous lesions of ESCC mainly refer to esophageal squamous cell dysplasia, and an increase in dysplasia grade is closely related to an increased risk. The WHO now refers to it as intraepithelial neoplasia, which can be classified into LGIN (atypical cells limited to less than 1/2 of the epithelium) and HGIN (heterotypic cells involving subepithelial 1/2 and above), based on the degree of cell dysplasia and depth of epithelial involvement [4, 46]. Precancerous cells are presumed to originate in the basal layer, and Ye's study suggested that CD98 is specifically expressed in the basal cell layer of normal esophageal epithelium and is associated with chronic inflammation in the early stages of ESCC [18]. Consistent with this, our study provided further evidence that CD98 expression is associated with chronic esophageal inflammation and the precancerous status of ESCC. Unfortunately, the underlying mechanisms of CD98 in the immune response were not described in depth in this study, which will be elaborated in detail in our subsequent studies. As previously reported, we found that CD98 protein expression can be used as a histological marker of the basal layer of normal esophageal squamous epithelium, and was uniformly expressed on the cell membrane of the basal layer. More importantly, we demonstrated that CD98 was expressed in most dysplastic epithelial cells in the precancerous state (73.81% in LGIN and 83.08% in HGIN), indicating the degree of dysplasia. IHC‐positive staining of P53 results from the accumulation of mutant P53, which is frequently observed in both LGIN and HGIN, but it also implies that the P53 protein does not possess diagnostic sensitivity and specificity [47, 48, 49]. In this study, we found that the combination of CD98 and P53 expression effectively increased the detection rate of precancerous lesions dependent solely on P53 expression (92.86% vs. 88.10% in LGIN; 93.85% vs. 73.85% in HGIN, p < 0.001). Thus, we proposed an algorithm that reflects changes in CD98 and P53 expression patterns in esophageal precancerous lesions and the diagnostic recommendations. This provides a powerful basis for the accurate diagnosis of early ESCC.

In addition, this study has several limitations, including the sample size and single‐center design. Large‐scale cohort studies are required to further elucidate the role of CD98 in esophageal cancer. As we did not employ multiple experimental methods for cross‐validation, there may be inherent limitations associated with the specific methods employed. This study elucidates the diagnostic value of the distinct expression patterns of CD98 and P53 in the early detection of esophageal cancer. However, the precise clinical implications of these distinct expression profiles in esophageal cancer remain to be further explored. Ultimately, while our data provide preliminary insights into the role of CD98 in promoting ESCC progression, in vivo studies have yet to be conducted; the feasibility of CD98 as a therapeutic target and its impact on clinical outcomes and survival benefits require more comprehensive evaluation.

In summary, this study confirmed that the upregulation of CD98 expression resulting from promoter hypomethylation was an independent factor for ESCC progression and prognosis, and fully demonstrated the diagnostic value of CD98 protein as a biomarker for esophageal precancerous lesions. Finally, we proposed auxiliary diagnostic suggestions based on the combined expression of CD98 and P53 (either one was positive). Targeting CD98 represents a promising strategy for early prevention and treatment of ESCC.

Author Contributions

Xue Zhang: conceptualization, methodology, visualization, writing – original draft. Mengfei Sun: data curation, formal analysis, methodology. Na Li: formal analysis, funding acquisition, methodology. Jie Yu: methodology. Wenjie Wang: methodology, resources. Ju Yang: methodology, resources. Hongyan Wu: methodology. Linyue Zhao: methodology. Huakun Zhang: methodology. Lan Yang: methodology. Feng Li: supervision. Qi Sun: investigation. Yunzhao Chen: conceptualization, funding acquisition, investigation, writing – review and editing. Xiaobin Cui: conceptualization, data curation, funding acquisition, investigation, supervision, writing – review and editing.

Ethics Statement

The study was ethically approved by the Ethics Committee of Nanjing Drum Tower Hospital (Approval ID: 2023–648‐02).

Consent

All informed consent was obtained from the subjects and/or guardians.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Data S1. Additional methods.

Figure S1. Relationship between CD98 promoter methylation and its protein expression and prognosis in patients with esophageal cancer.

Figure S2. The expression of CD98 is correlated with inflammation and Ki67.

Figure S3. Overexpression of CD98 promotes ESCC cell proliferation.

Table S1. Baseline clinicopathological characteristics of 318 patients with ESCC.

Table S2. Baseline table of clinicopathological indicators of 157 esophageal patients undergoing endoscopic submucosal dissection (ESD).

Data Availability Statement

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