Cancer‐associated fibroblast‐derived exosomal microRNA‐24‐3p enhances colon cancer cell resistance to MTX by down‐regulating CDX2/HEPH axis

Abstract MicroRNA‐24‐3p (miR‐24‐3p) has been implicated as a key promoter of chemotherapy resistance in numerous cancers. Meanwhile, cancer‐associated fibroblasts (CAFs) can secret exosomes to transfer miRNAs, which mediate tumour development. However, little is known regarding the molecular mechanism of CAF‐derived exosomal miR‐24‐3p in colon cancer (CC). Hence, this study intended to characterize the functional relevance of CAF‐derived exosomal miR‐24‐3p in CC cell resistance to methotrexate (MTX). We identified differentially expressed HEPH, CDX2 and miR‐24‐3p in CC through bioinformatics analyses, and validated their expression in CC tissues and cells. The relationship among HEPH, CDX2 and miR‐24‐3p was verified using ChIP and dual‐luciferase reporter gene assays. Exosomes were isolated from miR‐24‐3p inhibitor–treated CAFs (CAFs‐exo/miR‐24‐3p inhibitor), which were used in combination with gain‐of‐function and loss‐of‐function experiments and MTX treatment. CCK‐8, flow cytometry and colony formation assays were conducted to determine cell viability, apoptosis and colony formation, respectively. Based on the findings, CC tissues and cells presented with high expression of miR‐24‐3p and low expression of HEPH and CDX2. CDX2 was a target gene of miR‐24‐3p and could up‐regulate HEPH. Under MTX treatment, overexpressed CDX2 or HEPH and down‐regulated miR‐24‐3p reduced cell viability and colony formation and elevated cell apoptosis. Furthermore, miR‐24‐3p was transferred into CC cells via CAF‐derived exosomes. CAF‐derived exosomal miR‐24‐3p inhibitor diminished cell viability and colony formation and increased cell apoptosis in vitro and inhibited tumour growth in vivo under MTX treatment. Altogether, CAF‐derived exosomal miR‐24‐3p accelerated resistance of CC cells to MTX by down‐regulating CDX2/HEPH axis.


| INTRODUC TI ON
Colon cancer (CC) represents the 2nd most common reason for cancer-related death, which arises from a local adenoma detected in the intestinal epithelium to an aggressive malignant tumour and might potentially metastasize to the liver. 1 Currently, the therapeutic regimens for CC include chemotherapy. 2 Methotrexate (MTX), a well-known antitumour agent, is a folic acid analogue with high cytotoxicity, 3 which is widely used as a standard chemotherapy for a wide array of cancers such as CC 4 and colorectal cancer (CRC). 5 However, resistance to chemotherapy that develops in tumour cells limits their therapeutic efficacy in patients. 6 Moreover, chemotherapy is not the only available treatment, and target therapy has been demonstrated to better the survival and quality of life of cancer patients life. 7,8 Cancer-associated fibroblasts (CAFs), amongst the most abundant cell type occurringin tumour matrix, play a pivotal role in context of the development of cancers, including breast cancer, pancreatic cancer and colon cancer. 9 It is reported that CAFs can secret exosomes to regulate tumour progression. 10 Exosomes belong to extracellular vesicles and can function as non-invasive biomarkers for cancer treatment 11 influencing drug resistance. 12 Currently, the transfer of microRNAs (miRNAs) from fibroblasts via microvesicles to tumour cells is a promising approach in the cessation of chemoresistance of tumour cells. 13 Furthermore, CAFderived exosome carrying miRNAs contributes to chemoresistance of various cancers, including pancreatic cancer and head and neck cancer. 14,15 miRNAs participate in numerous cellular processes, whose dysregulation has vital functionality in the context of the development and progression of cancers. 16 At present, mounting evidence has revealed that miRNAs promote chemotherapy resistance of CC cells. [17][18][19] Furthermore, microRNA-24-3p (miR-24-3p) has been documented to regulate chemoresistance in several cancers such as in breast cancer 20 and ovarian cancer. 21 Moreover, a study previously confirmed that CAF-derived exosome carrying lncRNA H19 plays a role in the enhancement of CRC cell chemoresistance. 22 Furthermore, microarray data of our study predicted that miR-24-3p were capable of targeting caudal-related homeobox 2 (CDX2) in CC. Moreover, another study showed that CDX2 directly regulates the transcription of hephaestin (HEPH). 23 With the aforementioned preliminary predictions and previous studies considered as the basis, we hypothesized that exosomal miR-24-3p derived from CAFs played a role in CC resistance to MTX via CDX2/HEPH axis and conducted a study, including the co-culturing of exosomes isolated from CAFs with CC cells to test this hypothesis.

| Ethical statement
The study was conducted with the approval of the Ethics Committee of Shanghai Tenth People's Hospital, Tongji University School of Medicine.
All of the participants or their caregivers signed informed consent prior to the enrolment in this study. All animal studies were undertaken in accordance with the guidelines issued in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.
The expression of CDX2 and HEPH in CC of Cancer Genome Atlas (TCGA) database analysis was obtained from the UALCAN database available at (http://ualcan.path.uab.edu/analy sis.html). miRNA microarray data in fibroblast microvesicles were downloaded from EVmiRNA database (http://bioin fo.life.hust.edu.cn/EVmiRNA). The mirDIP database acquired from (http://ophid.utoro nto.ca/mirDI P/index.jsp#r) and TargetScan database acquired from (http://www.targe tscan.org/ vert_71/) were used to predict miRNAs regulating CDX2. The expression of miR-24-3p in CC of TCGA database was searched in the star-Base database (starb ase.sysu.edu.cn). Prior to surgery, the complete medical history was obtained from all 28 patients and each patient received a physical examination.

| Cell culture
Normal human colonic epithelial cells NCM-460 and CC cell lines including SW620, HT29, LoVo and SW1116 were purchased from

| Cell transfection
The cells were subjected to seeding into 6-well plates containing

| Western blot analysis
Total protein content was extracted using lysis buffer that contained was subjected to determination with the application of Gel densitometry (Bio-Rad, Hercules, CA, USA).  Helsinki, Finland). Three replicate wells were set for each sample.

| Clone formation assay
Transfected cells were subjected to seeding into 6-well plates at 500 cells/well for the purpose of overnight incubation. The following day, cells were treated with 50 nmol/L MTX for duration of 24 hours.
Then, the medium was replaced with MTX-free medium. After 14 d, the cells were subjected to fixation by means of 4% paraformaldehyde, staining by means of 0.1% crystal violet and counting visually.
Three replicate wells were set for each sample.

| Flow cytometry
Total 60,000 cells were cultured in 1 mL medium for duration of was adopted with the aim to examine cell apoptosis.

| Isolation of exosomes
In order to isolate exosomes from the culture medium, CAFs or NFs were cultured for duration of 48 hours in DMEM/F12 medium which was supplemented with 10% FBS (exosome removal).
The supernatant was collected, after which centrifugation was carried out at 300 × g for 10 minutes, at 2000 × g for 15 min-

| Identification of exosomes
The morphology of the exosomes was observed by means of a transmission electron microscopy (TEM), FEI Tecnai 12, Philips, Eindhoven, the Netherlands. Briefly, exosomes were fixed on a copper grid with 4% paraformaldehyde. The copper grid was dried at room temperature for 10 minutes. The sample was subjected to staining with 2% uranyl acetate, drying for duration of 10 minutes, followed by observation at 100 KV. The size distribution of exosomes was analysed by measuring the Brownian motion rate using a Nanosight LM20 system equipped with fast video capture and particle tracking software (Nanosight, Amesbury, UK).
Exosomal markers CD63, CD81 and TSG101 were determined by Western blot analysis.

| Tumour xenografts in nude mice
Eighty BALB/c nude mice (at the age of 5 weeks, and weighing 12-18 g) were attained from Hunan SJA Laboratory Animal Co., Ltd

| Statistical analysis
Statistical analysis for all data in the current study was implemented

| CDX2 and HEPH are down-regulated in CC tissues and cells
Firstly, the correlation of CC with CDX2 and HEPH was explored.
After analysing CC-related differentially expressed gene in the TCGA database, it was revealed that expression of HEPH and CDX2 was markedly down-regulated in CC samples compared with normal samples ( Figure 1A,B). Furthermore, RT-qPCR and Western blot analyses were adopted to examine CDX2 and HEPH expression in clinical samples of CC tissues and adjacent tissues.
The results showed that CDX2 and HEPH expression in CC tissues was noticeably lower than that in adjacent tissues ( Figure 1C Figure S1. The aforementioned results indicated that CDX2 and HEPH were down-regulated in CC and they were positively correlated in CC cells. CDX2 and HEPH were able to bind to each other, and CDX2 could increase HEPH expression.

| Overexpression of CDX2 elevates HEPH expression to reduce CC cell resistance to MTX
To investigate the effect of CDX2 and HEPH on CC cell resistance . Measurement data were expressed as mean ± standard deviation. Data of clinical samples were compared using paired t test, and data between two groups were analysed by unpaired t test. Comparisons among multiple groups were conducted by one-way analysis of variance (ANOVA), followed by Bonferroni's post hoc test. Each cell experiment was repeated three times and HT29 cells exhibited decreased cell viability and colony formation and increased cell apoptosis. After MTX treatment, SW620 and HT29 cell viability and colony formation were reduced secondary to overexpressed CDX2, which was reversed by HEPH silencing ( Figure 2B-D). CDX2 was silenced, and HEPH was overexpressed in SW620 and HT29 cells ( Figure 2E). In MTX-treated SW620 and HT29 F I G U R E 2 CDX2 overexpression up-regulates HEPH to repress CC cell resistance to MTX. A, Heat map of differentially expressed genes in MTX-resistant CC expression profile GSE11440. The abscissa refers to the sample number, the ordinate refers to the gene name, the left tree indicates the gene expression cluster, and the upper tree indicates the sample cluster. The upper right histogram was colour gradation, and each square in the FIGURE represents the expression of a gene in a sample. B, Cell viability of SW620 and HT29 cells transfected with CDX2 and si-HEPH after 5 d of MTX treatment detected by CCK-8 assay. C, Colony formation of MTX-treated SW620 and HT29 cells after transfection with CDX2 and si-HEPH assessed by colony formation assay. D, Apoptosis of MTX-treated SW620 and HT29 cells after transfection with CDX2 and si-HEPH examined by flow cytometry assay. E, CDX2 and HEPH expression in MTX-treated SW620 and HT29 cells after transfection with si-CDX2 and HEPH measured by RT-qPCR normalized to GAPDH. F, Cell viability of SW620 and HT29 cells transfected with CDX2 and si-HEPH after 5 d of MTX treatment detected by CCK-8 assay. G, Colony formation of MTX-treated SW620 and HT29 cells after transfection with si-CDX2 and HEPH assessed by colony formation assay. H, Apoptosis of MTX-treated SW620 and HT29 cells after transfection with si-CDX2 and HEPH analysed by flow cytometry. * P < .05 vs. SW620 and HT29 cells without any treatment or MTX-treated SW620 and HT29 cells. # P < .05 vs. MTX-treated SW620 and HT29 cells transfected with pCMV6-XL5. @ P < .05 vs. MTXtreated SW620 and HT29 cells transfected with CDX2 or si-CDX2. Measurement data were expressed as mean ± standard deviation.
Comparisons among multiple groups were conducted by one-way analysis of variance (ANOVA). Data at different time-points were compared by repeated-measures ANOVA, followed by Bonferroni's post hoc test. Each cell experiment was repeated three times cells, cell viability and colony formation were strikingly elevated, and the apoptotic rate was remarkably diminished by si-CDX2 treatment, which was abrogated by HEPH treatment (Figure 2F-H). These findings suggested that after cells exposed to MTX, overexpression of CDX2 up-regulated HEPH to reduce the resistance of CC cells to MTX.

| CDX2 is the target gene of miR-24-3p
The mirDIP and TargetScan were applied to predict the upstream regulatory miRNAs of CDX2. Meanwhile, miRNAs with the expression value of more than 5000 in fibroblast microbubbles were obtained in the EVmiRNA database. The intersection between the predicted results and EVmiRNA results showed that only one miRNA, miR-24-3p, was found in the intersection. The binding site between miR-24-3p and CDX2 was predicted on the TargetScan database ( Figure 3A). miR-24-3p expression of CC in TCGA was searched, and the result revealed a high expression of miR-24-3p in CC ( Figure 3B). miR-24-3p expression in CC tissues and adjacent tissues was detected by RT-qPCR. It was suggested that miR-24-3p expression was prominently higher in CC tissues than in adjacent tissues ( Figure 3C). Moreover, it was demonstrated that miR-24-3p expression was negatively correlated with the mRNA expression of CDX2 in clinical samples of CC ( Figure 3D).

miR-24-3p expression was determined in NCM-460 cells and CC cells
(SW620, HT29, LoVo and SW1116). The results displayed that miR-24-3p highly expressed in CC cells compared with NCM-460 cells ( Figure 3E), which was concurred with the trend in the clinical samples.
Subsequently, dual-luciferase reporter gene assay was adopted to further verify whether CDX2 was a target gene of miR-24-3p. The results documented that the luciferase activity of CDX2-WT in HEK-293T cells transfected with miR-24-3p mimic was severely reduced.
There was no significant change in luciferase activity of the CDX2-MUT ( Figure 3F). In addition, we successfully transfected miR-24-3p inhibitor and si-CDX2 into SW620 and HT29 cells ( Figure 3G), followed by measurement of CDX2 expression in SW620 and HT29 cells by Western blot analysis. The results presented that the protein expression of CDX2 and HEPH was significantly elevated in miR-24-3p inhibitor-transfected cells, which was rescued by si-CDX2 ( Figure 3H). The above were compared using paired t test, and data between cells were compared by unpaired t test. Data between two groups were compared by paired t test, and comparisons among multiple groups were conducted by one-way analysis of variance (ANOVA), followed by Bonferroni's post hoc test results collectively demonstrated that miR-24-3p was overexpressed in CC tissues and cells, and that CDX2 was a target gene of miR-24-3p.

| miR-24-3p promotes CC cell resistance to MTX through down-regulation of CDX2
CDX2 was proven to be a downstream target gene of miR-24-3p. Next, we explored whether miR-24-3p regulated the resistance of CC cells to MTX via CDX2. SW620 and HT29 cells were transfected with miR-24-3p mimic and CDX2, followed by MTX treatment. CCK-8 assay, colony formation assay and flow cytometry were conducted, the results of which showed that in MTX-treated cells, viability and colony formation of SW620 and HT29 cells were markedly elevated, and apoptotic rate was reduced following transfection with miR-24-3p mimic, which was eliminated following the overexpression of CDX2 ( Figure 4A-D).
The results of tumour xenografts in nude mice revealed that size of subcutaneous tumours was noticeably enhanced in MTX-treated F I G U R E 4 miR-24-3p enhances MTX resistance in CC via CDX2 down-regulation. SW620 and HT29 cells were transfected with mimic NC, miR-24-3p mimic or miR-24-3p mimic + CDX2, followed by MTX treatment. A, The expression of miR-24-3p in SW620 and HT29 cells determined by RT-qPCR normalized to U6. B, Cell viability of transfected SW620 and HT29 cells after 5 d of MTX treatment detected by CCK-8 assay. C, Colony formation assessed by colony formation assay. D, Flow cytometry analysis of transfected SW620 and HT29 cell apoptosis. Mice were transfected with mimic NC, miR-24-3p mimic or miR-24-3p mimic + CDX2, followed by MTX treatment. E, Microscopic observation of tumour size and tumour volume (n = 10). F, TUNEL staining of apoptosis in tumour tissues (n = 10; ×200). * P < .05 vs cells or mice treated with mimic NC + MTX. # P < .05 vs cells or mice treated with miR-24-3p mimic + MTX. Measurement data were expressed as mean ± standard deviation. Comparisons among multiple groups were conducted by one-way analysis of variance (ANOVA). Data at different time-points were compared by repeated-measures ANOVA, followed by Bonferroni's post hoc test mice treated with miR-24-3p mimic, which was abrogated after CDX2 overexpression ( Figure 4E). In addition, MTX-treated mice presented with reduced TUNEL-positive apoptotic cells after treatment with miR-24-3p mimic, which was neutralized by up-regulating CDX2 ( Figure 4F). In summary, miR-24-3p enhanced the resistance to MTX by down-regulating CDX2 in CC.

| miR-24-3p is highly expressed in exosomes derived from CAFs
Cancer-associated fibroblasts and Normal fibroblasts were obtained from CC tissues and normal colonic mucosa. CAFs were positive for α-SMA, FAP, FSP-1 and vimentin, while these proteins were poorly expressed in NFs ( Figure 5A-C), which indicated that the isolated cells were CAFs and NFs. HT29 and SW620 cells were treated with CAFs-conditioned medium (CAFs-CM) or NFs-CM. As revealed in Figure 5D-F, compared with cells treated with NFs-CM, cells treated with CAFs-CM displayed stronger resistance to MTX. We hypothesized that CAFs might regulate CC cells by secreting exosomes. In order to test this hypothesis, exosomes were isolated from CAFs-CM (CAFs-exos) and NFs-CM (NFs-exos), respectively, by differential ultracentrifugation. Under the TEM, exosomes were in a discoid vesicle structure ( Figure 5G). Nanosight analysis showed a mean particle size of 50-100 nm, which was a typical feature of exosomes ( Figure 5H). Furthermore, results of Western blot analysis revealed that the expression of the exosomal marker proteins (CD63, CD81 and TSG101) was positive ( Figure 5I). miR-24-3p expression in CAFs, NFs, CAFs-exo and NFs-exo was examined by RT-qPCR, which demonstrated that miR-24-3p was highly expressed in CAFs and CAFsexo ( Figure 5J). The aforementioned results showed the presence of high miR-24-3p expression in CAF-derived exosomes.

| miR-24-3p can be transferred from CAFs to CC cells through exosomes
CAFs-exo or NFs-exo were labelled with the green fluorescent dye PKH67 and added to CC culture medium to track whether these ex- However, the expression of miR-24-3p was significantly boosted in those cells after incubation with CAFs-exo ( Figure 6D). HT29 cells or SW620 cells were further co-cultured with control CAFs or CAFs transfected with Cy3-labelled miR-24-3p using the Transwell system. Interestingly, red fluorescence was observed in HT29 cells or SW620 cells co-cultured with CAFs transfected with Cy3-labelled miR-24-3p ( Figure 6E). The results of Western blot analysis displayed that CDX2 and HEPH expression was sharply increased in F I G U R E 6 miR-24-3p can be transferred directly from CAFs to CC cells. A, The internalization of CAFs-exo or NFs-exo in HT29 and SW620 cells observed under a LSCM (×200). B, Expression of miR-24-3p in SW620 and HT29 cells co-cultured with CAFs-exo or NFs-exo measured by RT-qPCR normalized to U6. C, The expression of miR-24-3p in CC cells treated with CAFs-exo after treatment of actinomycin D determined by RT-qPCR normalized to U6. D, The expression of miR-24-3p in CC cells after treatment of miR-24-3p inhibitor and exosomes examined by RT-qPCR normalized to U6. E, Images of HT29 cells or SW620 cells co-cultured with CAFs transfected with Cy3-labelled miR-24-3p observed under LSCM (×400). F, The protein expression of CDX2 and HEPH in CC cells after treatment of miR-24-3p inhibitor and exosomes analysed by Western blot analysis normalized to GAPDH. * P < .05 vs untreated HT29 cells or SW620 cells or HT29 or SW620 cells transfected with inhibitor-NC. # P < .05 vs HT29 or SW620 cells transfected with miR-24-3p inhibitor + NFs-exo. Measurement data were expressed as mean ± standard deviation. Data between two groups were compared using unpaired t test. Comparisons among multiple groups were conducted by one-way analysis of variance (ANOVA), followed by Bonferroni's post hoc test. The cell experiment was repeated three times CC cells transfected with miR-24-3p inhibitor and strikingly reduced after treatment with CAFs-exo, but there was no significant difference in CDX2 and HEPH expression in cells treated with NFs-exos ( Figure 6F). Based on these findings, miR-24-3p could be directly transferred from CAFs to CC cells through exosomes, resulting in a significant decrease in CDX2 and HEPH expression in CC cells.

| CAF-derived exosome carrying miR-24-3p enhances CC cell resistance to MTX
The effect of CAF-derived exosomal miR-24-3p on the treatment of CC by MTX in vitro and in vivo was explored. CAFs-exo/miR-24-3p inhibitor significantly reduced the viability and colony formation of HT29 and SW620 cells and enhanced apoptosis compared with HT29 and SW620 cells treated with CAFs-exo, which was reversed following treatment with miR-24-3p mimic ( Figure 7A-D). The results of Western blot analysis showed a significant increase in protein expression of CDX2 and HEPH in HT29 and SW620 cells following treatment of CAFs-exo/miR-24-3p inhibitor, which was abolished when miR-24-3p mimic was re-transfected into HT29 and SW620 cells ( Figure 7E).
In in vivo experiments, the results displayed that the largest subcutaneous tumour was formed in mice by HT29 cells treated with 0.9% saline, while MTX-treated HT29 cells formed the smallest subcutaneous tumour in nude mice. Under MTX treatment, HT29 cells treated with CAFs-exo/miR-24-3p inhibitor formed smaller tumours than cells treated with CAFs-exo but formed larger tumours than those treated with NFs-exo ( Figure 7F). In addition, apoptosis was reduced in tumour tissues of mice treated with CAFs-exo compared with those treated with NFs-exos. CAFs-exo/miR-24-3p inhibitor induced apoptosis in tumour tissues of MTX-treated mice ( Figure 7G).
Collectively, CAF-derived exosomal miR-24-3p could enhance CC cell viability, inhibit apoptosis and promote tumour growth in vivo under MTX treatment.

| D ISCUSS I ON
While chemotherapy has been a successful therapeutic option for CC, the current challenge faced in these patients and main causes of death include drug resistance and metastasis. 25  In addition, our result revealed a high expression of miR-24-3p in CC clinical samples and CC cells, which was also confirmed in breast cancer based on a previous study. 38 Moreover, another study also validated the up-regulation of miR-24-3p in CRC. 39 In addition, our findings elucidated that miR-24-3p overexpression promoted cell proliferation and reduced cell apoptosis, thus inducing MTX resistance, which was supported by various studies.
An example of such is the ability of miR-24-3p to accelerate cell migration and proliferation in lung cancer. 40 Moreover, miR- 24-3p has been proven to contribute to facilitation of cell proliferation, migration and invasion in bladder cancer. 41 A study conducted F I G U R E 8 Schematic representation of function of CAF-derived exosomal miR-24-3p in CC. Exosomal miR-24-3p derived from CAFs was released after endocytosis by CC cells. miR-24-3p could target CDX2, and CDX2 could positively regulate HEPH. miR-24-3p contributed to promotion of CC cell resistance to MTX, while CDX2 and HEPH inhibited CC cell resistance to MTX

MTX resistant
Colon cancer cells

Exosomal miR-24-3p
HEPH mcroenviroment by Liu et al revealed that miR-24-3p was one of the miRs, which could result in cell chemoresistance to cisplatin in ovarian cancer. 21 Meanwhile, a prior study also reported that down-regulated miR-24-3p resulted in the reduction of cell proliferation, colony formation and chemoresistance in head and neck squamous cell carcinoma. 42 Another crucial finding from our study was that miR-24-3p can be transferred from CAFs to CC cells through exosomes and that CAF-derived exosome carrying miR-24-3p leads to the enhancement of CC cell resistance to MTX. CAFs are mesenchymal cells in tumours, which are generated via the activation of resident mesenchymal cell populations and recruitment of bone marrow-derived mesenchymal stem cells and fibrocytes, with potential to promote CC progression. 43 It is documented that secretion of exosomal miR-NAs can communicate with CAFs with cancer cells. 44 For instance, miR-92a-3p is confirmed to have a high expression in CAFs and CAFderived exosomal miR-92a-3p stimulates resistance of CRC cells to chemotherapy. 45 Similarly, another study portrayed that miR-21 can be transferred into tumour cells via CAF-derived exosomes, promoting CRC development. 46 These findings indirectly supported our results suggesting the presence of a high expression of miR-24-3p in CAFs and the transfer of miR-24-3p by exosomes derived from CAFs promoted MTX resistance in CC in vitro and in vivo.
In conclusion, the aforementioned findings elucidated that CAFderived exosomal miR-24-3p could function as an oncogene and was capable of enhancing the resistance of CC cells to MTX by activating the CDX2/HEPH axis (Figure 8), which provided a potential target to increase MTX sensitivity in clinical CC treatment. However, further studies are required to determine whether the therapeutic target is applicable to human beings.

CO N FLI C T O F I NTE R E S T
The authors declare that they have no competing interests.

ACK N OWLED G EM ENTS
We acknowledge and appreciate our colleagues for their valuable suggestions and technical assistance for this study. Formal analysis (equal); Supervision (equal).

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.