Article Text
PDF
Abstract
Aims To clarify claudin18.2 expression and its clinicopathological features in various cancers, especially in lung adenocarcinoma.
Methods Immunohistochemistry staining and fluorescence in situ hybridisation (FISH) were performed to detect claudin18.2 expression and CLDN18 gene rearrangement in adenocarcinoma from different organs.
Results The results showed that claudin18.2 expression was found in 68% (27 of 40) of lung mucinous adenocarcinoma, 52% (16 of 31) of cholangiocarcinoma, 2% (10 of 423) of colorectal adenocarcinoma tissue microarray, 27% (6 of 22) of colorectal mucinous adenocarcinoma and 30% (3 of 10) of cervical adenocarcinoma, but not in all 39 cases of invasive breast adenocarcinoma by immunohistochemistry staining. There was significantly positive correlation between ratio of claudin18.2-positive carcinoma cells and staining intensity in lung mucinous adenocarcinoma and cholangiocarcinoma. Claudin18.2 expression was much more in female patients than male patients with lung mucinous adenocarcinoma. In addition, cholangiocarcinoma with claudin18.2 expression was more aggressive and had perineural invasion. Intraductal papillary neoplasm of the bile duct and epithelial dysplasia of the adjacent bile in cholangiocarcinoma also showed claudin18.2 expression. All three cases of cervical adenocarcinoma with claudin18.2 expression were moderately differentiated adenocarcinoma including one human papillomavirus (HPV)-associated carcinoma, two non-HPV-associated and gastric-type carcinoma. CLDN18 gene rearrangement was not found in all 22 cases with high claudin18.2 expression by FISH.
Conclusions Our results suggest claudin18.2 might be a potential biomarker for targeted therapy on lung mucinous adenocarcinoma, cholangiocarcinoma, colorectal mucinous adenocarcinoma and gastric-type cervical adenocarcinoma.
- IMMUNOHISTOCHEMISTRY
- Lung Neoplasms
- Colorectal Neoplasms
Data availability statement
No data are available.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Claudin18.2 protein is expressed in gastric cancer; zolbetuximab targeting claudin18.2 achieved better clinical effects in advanced gastric cancer. Claudin18.2 expression in adenocarcinoma from other sites remains unknown.
WHAT THIS STUDY ADDS
Claudin18.2 expression was found in lung mucinous adenocarcinoma, cholangiocarcinoma, colorectal mucinous adenocarcinoma and gastric-type cervical adenocarcinoma, but not in invasive breast adenocarcinoma.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
Claudin18.2 might be a potential biomarker for targeted therapy on lung mucinous adenocarcinoma, cholangiocarcinoma, colorectal mucinous adenocarcinoma and gastric-type cervical adenocarcinoma.
Introduction
Claudin18.2 belongs to the claudins protein family and is one of the important proteins in the tight junction between the epithelium and the endothelium.1 Claudin18.2 is specifically expressed in differentiated gastric mucosal epithelial cells, but not in gastric stem cells.2 When carcinogenesis occurs, claudin18.2 protein is expressed in primary and metastatic gastric cancer.3 A few studies show that claudin18.2 protein is abnormally expressed in pancreatic neoplasms,4 biliary neoplasms,5 ovarian cancer6 and lung cancer.7 In advanced gastric cancer, zolbetuximab (IMAB362) targeting claudin18.2 achieved better clinical effects compared with the control group; the combination treatment of epirubicin, oxaliplatin and capecitabine+zobetal in phase II trials achieved the best effect of prolonging the overall survival to 13.2 months, and the safety events were tolerable.8
The gene encoding claudin18.2 protein is CLDN18. Gastric cancers with CLDN18::ARHGAP fusion gene occurring in young adults are considered to be a subgroup with unique clinicopathological features including gender parity, advanced disease and diffuse histology.9 Studies have shown that CLDN18::ARHGAP fusion gene is more common in young patients with gastric cancer and is related to tumour aggressiveness, and its fusion status is an independent predictor of distant organ metastasis.10 The prognosis of patients with gastric cancer carrying CLDN18::ARHGAP26/6 fusion is worse, and chemotherapy does not affect their prognosis, but patients who do not carry the CLDN18::ARHGAP26/6 fusion gene can benefit from chemotherapy.11
Claudin18.2 expression in adenocarcinoma from other sites remains unknown. In this study, we investigated claudin18.2 expression in adenocarcinoma from various parts including lung adenocarcinoma, cholangiocarcinoma, colorectal adenocarcinoma, cervical adenocarcinoma and invasive breast adenocarcinoma, and explored whether its expression was associated with different histological types of adenocarcinoma. Furthermore, the relationship between claudin18.2 expression and clinicopathological features was analysed to clarify the impact of claudin18.2 on cancer aggressiveness and prognosis. In addition, CLDN18 break-apart gene was detected in various carcinomas with high claudin18.2 expression by fluorescence in situ hybridisation (FISH).
Materials and methods
Specimen
Specimens including 40 cases of invasive mucinous adenocarcinoma from March 2013 to March 2023, 57 other types of lung adenocarcinoma, 31 cholangiocarcinomas, 22 colorectal mucinous adenocarcinomas and colorectal cancer (CRC) tissue microarray including 423 CRC and 285 non-cancer colorectal tissues, 10 cervical adenocarcinomas and 39 invasive breast adenocarcinomas were collected between January 2019 and December 2021 from the Department of Pathology of the First Affiliated Hospital, Sun Yat-sen University. All patients had not received radiotherapy and/or chemotherapy before tumour resection.
Immunohistochemistry staining and quantification
Immunohistochemistry (IHC) staining was performed as we previously described.12 The mouse-derived claudin18.2 antibody was donated by Guangzhou LBP Medical Technology Co (clone number LBP-CLDN18.2). The working concentration of claudin18.2 antibody was 1:1000. The positive signals of claudin18.2 were located in the membrane of tumour cells. The known positive gastric signet ring cell carcinoma tissue was used as a positive control.
Any membranous staining of tumour cells for claudin18.2 was evaluated semiquantitatively accompanied with the staining intensity and positive ratio. Signal intensity was scored as 0 (negative), 1+ (weak), 2+ (moderate) and 3+ (strong). The positive ratio of tumour cells was divided into 0 (0~25%), 1+ (26~50%), 2+ (51~75%) and 3+ (76~100%). ≥2+ in ≥50% of tumour cells staining was defined as high expression; the others are defined as low expression or negative. For TTF-1 (8G7G3/1, Guangzhou LBP Medical Technology Co, China), CK7, napsin A, CDX2 and CK20, which were purchased from DAKO Company, positive expression was defined as positive signal in more than 25% of tumour cells; otherwise, it was classified as negative. The results of IHC were determined by two pathologists independently.
FISH detection and next-generation sequencing
FISH was performed as we previously described.12 The paraffin specimens of tumour tissues with claudin18.2 staining intensity of ≥2+ in ≥50% tumour cells by IHC were determined by FISH using the Claudin18.2 gene fragmentation probe reagent (Guangzhou LBP Medical Technology Co, China). Gene rearrangement was reported as positive if ≥10% of the tumour nuclei showed split signals defined as separation of signals.
The tissue DNA was extracted from all tissue samples using the QIAamp DNA FFPE tissue kit (Qiagen, Valencia, California, USA) according to the manufacturer’s instructions. Capture-based next-generation sequencing (NGS) with a panel of 168 cancer-related genes (Burning Rock Biotech, Guangzhou, China) was conducted in paraffin specimens of lung cancer. Alterations of eight well-established driver genes, EGFR, ERBB2, ALK, ROS1, RET, KRAS, BRAF and MET, were all included in sequencing panels. In brief, DNA was sheared using Covaris M220 (Covaris, Woburn, Massachusetts, USA), which was then followed by end-repair, phosphorylation and adaptor ligation. Fragments that were 200–400 bp in size were selected by beads (Agencourt AMPure XP Kit, Beckman Coulter, Brea, California, USA), followed by hybridisation with capture probes or baits, hybrid selection with magnetic beads and PCR amplification. A high-sensitivity DNA assay was performed using a bioanalyser to evaluate the DNA quality and size. Indexed samples were then sequenced on a NextSeq 500 (Illumina, San Diego, California, USA) with pair-end reads. The sequencing data were in an FASTQ format and mapped to the human genome (hg19) using Burrows-Wheeler Aligner V.0.7.10. Local alignment optimisation, mark duplication and variant calling were performed using the Genome Analysis ToolKit V.3.2, Picard (http://picard.sourceforge.net/) and VarScan. Gene translocations were called with FACTERA, and the copy number variation was analysed with an in-house algorithm based on sequencing depth.
Statistical analysis
Associations between clinicopathological parameters and claudin18.2 expression were analysed using Fisher’s exact test and Χ2 test, the correlation between claudin18.2 staining intensity and positive ratio was determined using the Spearman’s correlation test. P<0.05 was considered as statistically significant.
Results
Expression of claudin18.2 in lung adenocarcinoma
97 cases of lung adenocarcinoma including 40 cases of invasive mucinous adenocarcinoma (IMA), 14 cases of lepidic adenocarcinoma, 9 cases of solid adenocarcinoma, 12 cases of papillary adenocarcinoma, 7 cases of micropapillary adenocarcinoma and 15 cases of acinar adenocarcinoma according to 2021 WHO Classification of Thoracic Tumours criteria were collected. Claudin18.2 positive signal was located in the carcinoma cell membrane, and the cell membrane near the basal part of the tumour cells was deeply stained by IHC staining. The result showed that 68% (27 of 40) of IMA of the lung was positive for claudin18.2. Among them, 19 cases were scored 3+ and 7 cases were scored 2+ according to claudin18.2 positive signalling intensity. Most positive samples were characterised to highly express claudin18.2 including 25 cases (93%, 25 of 27) with staining intensity ≥2+ in ≥50% of carcinoma cells (figure 1). The ratio of claudin18.2-positive carcinoma cells and staining intensity was closely correlated (r=0.8549, p<0.0001, figure 2A). In order to investigate the relationship between claudin18.2 expression and clinicopathological features of lung IMA, no significant differences were noted in terms of age, tumour size, tumour location, smoking, clinical stage, and the expression of CK7 and CK20 in lung IMA. Claudin18.2 expression was observed more frequently in female patients (p=0.0002); 8% TTF-1 (2 of 25), 8% napsin A (2 of 25), 36% CK20 (9 of 25) and no CDX2 (0 of 25) expression were positive in 25 cases of IMA with strong claudin18.2 expression, respectively. In our study, there was only one case-mixed lung mucinous and non-mucinous adenocarcinoma; both mucinous and non-mucinous areas were negative for the claudin18.2 expression. In addition, NGS was performed in 15 IMAs including 9 cases with strong claudin18.2 positivity and 6 cases with weak or negative claudin18.2 expression. The result showed four cases were KRAS gene mutation (44%, 4 of 9) in IMA with strong claudin18.2 positivity, compared with only one case (17%, 1 of 6) of KRAS mutation in IMA with weak or negative claudin18.2 expression (figure 3). Among the lung lepidic adenocarcinoma, eight cases (57%, 8 of 14) present uniform and weak claudin18.2 expression (score 1+), and the ratio of positive carcinoma cells was more than 50% in four cases. However, claudin18.2 expression was not found in all acinar, solid, papillary and micropapillary adenocarcinomas as well as in adjacent non-tumour lung tissue (figure 1 and tables 1 and 2).
Claudin18.2 expression in lung adenocarcinoma. Invasive mucinous adenocarcinoma, H&E 200× (A); claudin18.2 staining with score 3, IHC 200× (B); lepidic adenocarcinoma, H&E 200× (C); claudin18.2 staining with score 1, IHC 200× (D); solid adenocarcinoma, H&E 200× (E); no claudin18.2 expression, IHC 200× (F); papillary adenocarcinoma, H&E 200× (G); no claudin18.2 staining, IHC 200× (H); micropapillary adenocarcinoma, H&E 200× (I); no claudin18.2 staining, IHC 200× (J); acinar adenocarcinoma, H&E 200× (K); no claudin18.2 staining, IHC 200× (L). IHC, immunohistochemistry.
Claudin18.2 staining intensity associated with positive ratio of carcinoma cells in lung mucinous adenocarcinoma (A) and cholangiocarcinoma (B), respectively.
Gene alterations in lung invasive mucinous adenocarcinoma by next-generation sequencing: (A) lung invasive mucinous adenocarcinoma with strong claudin18.2 positivity; (B) lung invasive mucinous adenocarcinoma with weak or negative claudin18.2 expression.
Claudin18.2 expression in various types of adenocarcinoma
Correlation between different parameters and claudin18.2 expression in pulmonary mucinous adenocarcinoma
Claudin18.2 expression in adenocarcinoma of other parts
16 (52%, 16 of 31) samples of cholangiocarcinoma were found positive for claudin18.2 expression with 2 cases scoring 3+, 10 cases scoring 2+ and 4 cases scoring 1+ by IHC staining. The ratio of claudin18.2-positive carcinoma cells was less than 50% in seven cases of cholangiocarcinoma which were all moderately differentiated adenocarcinoma. Meanwhile, the positive signal distribution was variable within one tumour. The ratio of claudin18.2-positive carcinoma cells was more than 50% in other nine cases including two well-differentiated carcinomas and seven moderately differentiated carcinomas. Moreover, there was a positive correlation between the ratio of claudin18.2-positive carcinoma cells and the staining intensity (r=0.8597, p<0.0001, figure 2B). Some intraductal papillary neoplasm of the bile duct and epithelial dysplasia of the adjacent bile in cholangiocarcinoma also showed claudin18.2 expression (figure 4C,D). Most of the cholangiocarcinoma with claudin18.2 expression had perineural invasion (81%, 13 of 16) (figure 4E,F), However, there was no significant difference between claudin18.2 expression and tumour lymph node metastasis (p=0.7224) and prognosis including survival rate (p=0.7224) and tumour recurrence (p=0.7166), respectively.
Claudin18.2 expression in cholangiocarcinoma. Cholangiocarcinoma, H&E 200× (A); claudin18.2 staining with score 3, IHC 200× (B); intraductal papillary neoplasm of the bile duct, H&E 200× (C); claudin18.2 expression in papillary tumour cells of the bile duct, IHC 200× (D); cholangiocarcinoma perineural invasion, H&E 200× (E); claudin18.2 expression in cholangiocarcinoma, IHC 200× (F). IHC, immunohistochemistry.
IHC staining was performed in colorectal cancer tissue microarray including 285 non-tumour colorectal mucosal tissues and 423 cases of colorectal adenocarcinoma. The result showed that only 10 cases (2%, 10 of 423) of colorectal adenocarcinoma including 2 mucinous adenocarcinoma, 2 moderately differentiated adenocarcinoma and 6 poorly differentiated adenocarcinoma were positive for claudin18.2 with 1 case scoring 1+ and 9 cases scoring 2+ by IHC staining. The ratio of claudin18.2-positive carcinoma cells was more than 50% in two cases with score 2+ and poorly differentiated adenocarcinoma. All non-tumour colorectal mucosal tissues in tissue microarray were negative for claudin18.2. Claudin18.2 expression was found in small foci of carcinoma tissue on tissue microarray which could not represent the overall tumour tissue. Then, another 22 samples of colorectal mucinous adenocarcinoma were collected: 6 cases (27%, 6 of 22) were positive for claudin18.2 expression with 1 case scoring 1+ and 5 cases scoring 2+. The ratio of claudin18.2-positive carcinoma cells reached more than 50% in three cases with score 2+ (figure 5A–D). Claudin18.2 expression was not found in non-tumour colorectal mucosal tissue.
Claudin18.2 expression in other types of adenocarcinoma. Claudin18.2 expression in poorly differentiated colorectal adenocarcinoma of tissue microarray at low power and high power, respectively, IHC (A: 100×, B: 200×); colorectal mucinous adenocarcinoma, H&E 200× (C); claudin18.2 staining, IHC 200× (D); HPV-associated endocervical adenocarcinoma, H&E 200× (E); claudin18.2 staining with score 2, IHC 200× (F); non-HPV-associated gastric-type endocervical adenocarcinoma, H&E 200× (G); claudin18.2 staining with score 1, IHC 200× (H); invasive breast adenocarcinoma, H&E 200× (I); no claudin18.2 expression, IHC 200× (J). HPV, human papillomavirus; IHC, immunohistochemistry.
We detected claudin18.2 expression in 10 cases of endocervical adenocarcinoma including 2 non-human papillomavirus (HPV)-associated adenocarcinoma and 8 HPV-associated adenocarcinomas by IHC staining. Claudin18.2 expression was found in three cases (30%) including two cases with score 3+ and one case with score 2+. All three cases with claudin18.2 expression were moderately differentiated adenocarcinoma including one HPV-associated carcinoma, and two non-HPV-associated and gastric-type carcinoma (figure 5E–H).
Claudin18.2 expression was not found in all 39 cases of invasive breast adenocarcinoma including 10 cases of mucinous adenocarcinomas and 29 cases of invasive ductal carcinomas, not otherwise specified (NOS) (figure 5I,J and table 1).
Detection of CLDN18 gene translocation by FISH
To explore whether there is CLDN18 gene rearrangement in a variety of adenocarcinomas, FISH was performed in 27 cases of adenocarcinoma with high claudin18.2 expression including 16 cases of mucinous adenocarcinoma of the lung, 8 cases of cholangiocarcinoma and 3 cases of colorectal mucinous adenocarcinoma. The result showed that CLDN18 gene rearrangement was not found in all 22 cases except that 5 cases were failed because of weak or no signals in cancer tissues by FISH (figure 6).
Claudin18.2 expression by immunohistochemistry staining and CLDN18 rearrangement by FISH: lung mucinous adenocarcinoma (A), cholangiocarcinoma (C), colorectal mucinous adenocarcinoma (E) with high claudin18.2 expression by immunohistochemistry staining, 200×; no CLDN18 rearrangement was found in lung mucinous adenocarcinoma (B), cholangiocarcinoma (D) and colorectal mucinous adenocarcinoma (F) by FISH (400×). FISH, fluorescence in situ hybridisation.
Discussion
Claudins constitute a family of tight junction proteins that maintain and control the exchange of molecules between cells.13 Claudin18.2 is one of the isoforms, which is highly tissue selective.3 14–17 In normal tissue, it is only hidden in the gastric mucosa, and monoclonal antibodies cannot be obtained. However, in tumour cells, claudin18.2 epitope exposure makes it a specific target and endows the specificity of solid tumour targeted therapy.3 A current phase IIa (MONO) trial of tumour-targeted drugs evaluated the safety and efficacy of IMAB362 as monotherapy in patients with advanced gastric, gastro-oesophageal junction or oesophageal adenocarcinoma with ≥50% of tumour cells moderately or strongly expressing claudin18.2.18 IMAB362 showed good safety for patients with ≥70% of tumour cells moderately or strongly expressing claudin18.2. The novel antibody IMAB362, whether used as monotherapy or in combination with chemotherapy, has shown significantly superior efficacy and safety.18 An ongoing study of claudin18.2 CAR-T cell therapy provides another possibility for solid tumour treatment.19
The latest global cancer data in 2020 show that lung cancer ranks second in cancer incidence and first in mortality,20 and non-small cell lung cancer (NSCLC) accounts for 80% of all lung cancer cases.21 Adenocarcinoma is the most common pathological type of NSCLC, and its prevalence has been shown to increase.22 Invasive lung adenocarcinomas are mainly divided into invasive non-mucinous adenocarcinoma including five subtypes, which are lepidic, papillary, acinar, solid and micropapillary types, and IMA, colloid adenocarcinoma, fetal adenocarcinoma and enteric-type adenocarcinoma. Other reports stated that IMA accounts for 5–10% of all lung adenocarcinomas,23 and there is no effective targeted therapy in clinical practice. Consistent with previous report, the cancer cells express CK7, CK20 and CDX2, while TTF-1 and napsin A were mostly negative.22 Our results showed that expression of napsin A, TTF-1, CK20 and CDX2 in IMA with strongly positive expression of claudin18.2 was commonly lower or negative, and IMA cases with strong claudin18.2 expression had higher KRAS gene mutation compared with those cases with weak or negative claudin18.2 expression. Our result unexpectedly showed that claudin18.2 was highly expressed in 68% of lung mucinous adenocarcinoma cases, especially in female patients. 57% of lung lepidic adenocarcinomas showed uniform and weak claudin18.2 expression. The result suggests that claudin18.2 might be a potential biomarker for lung mucinous adenocarcinoma targeted therapy.
Cholangiocarcinoma is a malignant tumour of the bile duct system, originating from bile duct epithelial cells, and accounts for about 10–20% of all primary liver cancers.24 Due to typical advanced clinical manifestations and adverse reactions to current treatments, the clinical prognosis is very poor.25 Our result showed that 52% of cholangiocarcinoma was positive for claudin18.2 expression. Meanwhile, we first found that intraductal papillary neoplasm of the bile duct and epithelial dysplasia of the adjacent bile in cholangiocarcinoma also showed claudin18.2 expression. 13 cases (81%, 13 of 16) of cholangiocarcinoma with claudin18.2 expression had perineural invasion. There was no relationship between cholangiocarcinoma with claudin18.2 expression and lymph node metastasis, survival rate and tumour recurrence, respectively. This issue needs further study with more samples.
Iwaya et al have reported that CLDN18 was positive in 27% (15 of 56) of colitis-associated colorectal adenocarcinomas (CACs) and in 5% (3 of 56) of sporadic CRCs (p=0.004), and CLDN18-positive CACs were more likely to have lymph node metastasis than CLDN18-negative CACs (67% vs 36%; p=0.017).26 Our data showed that only 10% of CRCs including two mucinous adenocarcinomas, two moderately differentiated adenocarcinomas and six poorly differentiated adenocarcinomas were positive for claudin18.2 on tissue microarray. However, claudin18.2 expression was found in 27% of colorectal mucinous adenocarcinoma. The result suggests that claudin18.2 might be a potential biomarker for colorectal mucinous adenocarcinoma treatment.
In cervical adenocarcinoma, we found that claudin18.2 expression was found in one HPV-associated carcinoma and two non-HPV-associated and gastric-type carcinoma. The result was consistent with Kiyokawa et al’s study which showed that claudin18 was more frequently expressed in gastric-type (21 of 22) than non-gastric-type (8 of 53) cervical adenocarcinoma. In addition, some clear cell carcinomas were positive for claudin18, but none showed intense staining.27 In our study, other types of cervical carcinoma were not included. It needs further study with much more samples of different types of cervical carcinoma. Our results suggest that claudin18.2 is a promising surrogate biomarker for gastric-type carcinoma and HPV-associated cervical adenocarcinoma.
The clinical treatment of invasive breast cancer is mainly based on the expression of oestrogen receptor, progesterone receptor and human epidermal growth factor receptor-2 (HER2), and the corresponding endocrine and targeted therapy are carried out to obtain good clinical results. However, there are still a small number of breast cancer cases that are triple negative and not sensitive to hormones and targeted therapy. For this reason, it is urgent to identify another effective therapeutic target. Studies have shown that in gastric cancer, claudin18.2 has a higher positive rate than HER2.28 To this end, we collected 39 cases of invasive breast adenocarcinoma including 10 cases of mucinous adenocarcinoma and 29 cases of invasive ductal carcinoma, NOS and found no claudin18.2 expression in all cases by IHC. The result suggests that claudin18.2 expression is not a potential biomarker for breast cancer therapy.
The CLDN18::ARHGAP fusion is one of the molecular characteristics of diffuse gastric cancer and is also an independent prognostic risk factor for gastric cancer. In addition, it is also related to multiple clinical characteristics, including age, sex, lymph node metastasis and tumour stage.29CLDN18::ARHGAP26 mediates epithelial disintegration, possibly leading to stomach H(+) leakage, and the fusion might contribute to invasiveness once a cell is transformed.30 However, our data showed that CLDN18 break-apart gene was not found in all 22 cases with high claudin18.2 expression except that 5 cases were failed because of weak or no signals in cancer tissues which might be related to the old specimen for more than 3 years. This issue needs further study in large series of samples.
Data availability statement
No data are available.
Ethics statements
Patient consent for publication
Ethics approval
Ethics approval and consent to participate were obtained from the Ethics Committee of the First Affiliated Hospital, Sun Yat-sen University.
References
Footnotes
Handling editor Deepa Patil.
PY and YD contributed equally.
Contributors PY and YD performed the experiments and analysed data. FZ, TZ and JL collected clinical data and analysed results. PY and HS drafted the manuscript. AH as guarantor accepts full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish.
Funding This work was supported by Natural Science Foundation of Guangdong Province (nos. 2021A1515012379; 2022A1515012550).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.