Multi-omics analysis reveals the effects of Jinfeng pills on the outcomes of IVF-ET in patients with poor ovarian response: a randomized controlled trial

Abstract

Background

Jinfeng pills (JFP), a well-known traditional Chinese medicine (TCM) consisted of 9 herbs, benefit Qi, nourish the blood, effectively regulate the menstrual cycle, promote ovulation and improve endometrial thickness and blood flow. Previous studies have shown that JFP has certain effect on premature ovarian insufficiency (POI). In this prospective cohort study, the effect and mechanisms of JFP on the outcomes of in vitro fertilization and embryo transfer (IVF-ET) in poor ovarian response(POR) was investigated.

Methods

Patients diagnosed with POR based on the Bologna criteria were prospectively randomized into two groups: the JFP group and the POR group. Young patients with normal ovarian reserve were included in the Control (Con) group. The live birth rate (LBR), levels of estradiol (E₂), progesterone, total testosterone (TT), and insulin-like growth factor-1 (IGF-1) in the follicular fluid (FF) and outcomes of IVF-ET were compared. Network pharmacology, combined with transcriptomics and proteomics, was used to identify core targets and signaling pathways. Molecular docking was performed to evaluate the interactions between the core components of JFP and the identified targets.

Results

Compared with the Con group, the POR group presented significantly lower levels of E₂ [1302.00 (985.50, 1464.50) ng/mL vs. 2561.00 (2190.50, 3084.00) ng/mL] and TT [2.41 (1.45, 3.15) ng/mL vs. 5.12 (4.00, 6.79) ng/mL] in FF (P < 0.017). These levels were greater in the JFP group [E₂: 2450.00 (1603.00, 2842.00) ng/mL; TT: 2.85 (2.08, 4.55) ng/mL] than in the POR group (P < 0.017). IGF-1 levels in FF [79.60 (64.00, 88.50) ng/mL vs. 87.90 (78.80, 112.00) ng/mL], the implantation rate, and the LBR per treatment were significantly lower in the POR group than in the Con group (P < 0.017), while there was no significant difference compared with those in the JFP group (P > 0.05). The birth weight and height of the infants did not differ significantly among the three groups (P > 0.05). Transcriptomic and proteomic analyses revealed seven core genes (ALPL, HPSE, KRT1, LDLR, METTL7B, TGFBI, and TMOD1) and one key pathway (glycosaminoglycan degradation). Molecular docking confirmed strong binding activity between the core components of JFP and these genes.

Conclusion

JFP enhances the ovarian levels of steroid hormones in poor ovarian responders undergoing IVF-ET. The therapeutic effects of JFP may involve the regulation of ALPL, HPSE, KRT1, LDLR, METTL7B, TGFBI, TMOD1, and the glycosaminoglycan degradation pathway.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s13048-026-02196-6.

Introduction

Poor ovarian response (POR) affects 9% ~ 24% of women undergoing in vitro fertilization and embryo transfer (IVF-ET) and is characterized by reduced oocyte and embryo yields and a low live birth rate (LBR) [1]. The etiology of POR is multifactorial, and involves advanced maternal age, diminished ovarian reserve (DOR), genetic predisposition, immune dysregulation, oxidative stress, and gonadotropin receptor insensitivity [1, 2]. Although interventions aimed at improving the quantity and quality of oocytes, such as pretreatment with growth hormone (GH), coenzyme Q10 (CoQ10) or dehydroepiandrosterone (DHEA), and luteinizing hormone (LH) adjuvants during controlled ovarian stimulation (COS), their efficacy remains controversial [3].

Traditional Chinese medicine (TCM) offers a unique therapeutic approach for POR, as its pathogenesis is attributed to Kidney Qi deficiency and blood stasis [4]. JFP, a Qing Dynasty herbal formula standardized by Tongyitang Pharmaceuticals since 1886, exemplifies this approach. Each 0.18 g water-honeyed pill contains nine botanicals formulated under sovereign-minister-assistant-messenger principles [5]. Epimedium brevicornu Maxim. (Yinyanghuo) [6] and Cervus elaphus Linnaeus (Lurong) [7] act as sovereign herbs to increase Yang and benefit Qi. Panax ginseng C.A. Mey. (Renshen), Equus asinus collagen (Ejiao), Ligustrum lucidum W.T. Aiton (Nvzhenzi), and Reynoutria multiflora (Thunb.) Moldenke (Heshouwu) serve as minister herbs to nourish Yin and tranquilize the mind. Curculigo orchioides Gaertn. (Xianmao), Neolitsea cassia (L.) Kosterm. (Rougui), and Leonurus japonicus Houtt. (Yimucao) functions as an assisting herb to dispel cold and resolve blood stasis. Honey serves as a messenger drug and plays a harmonious role, such as an adhesion agent and a sweetening agent [5] .

Since 1990, clinical studies have demonstrated the efficacy of JFP in treating infertility, DOR, and thin endometrium. JFP improves ovarian and uterine blood perfusion, endometrial thickness and receptivity [8]. In patients with polycystic ovary syndrome (PCOS), Xu et al. reported that JFP increased the ovarian blood flow, basal estradiol (E₂) level and pregnancy rate, but reduced serum levels of testosterone and follicle-stimulating hormone (FSH) [9]. A prospective study reported that JFP reduced the gonadotropin requirement and increased embryo yield in patients with DOR [10]. However, the specific mechanisms underlying the effects of JFP on POR remain unclear.

Animal studies have demonstrated that the components of JFP enhance ovarian function through diverse mechanisms. Epimedium brevicornum Maxim. improves ovarian endocrine function by increasing the expression of CYP17 and CYP19, thereby promoting the secretion of E₂ and progesterone in rats [11]. Curculigo orchioides Gaertn. exhibits estrogen-like biological activity [12], whereas Cervus elaphus Linnaeus exerts estrogen-like effects, manifested as improved bone trabecular structure in ovariectomized mice [13]. Panax ginseng C.A. Mey. increases follicular development by increasing follicle diameter, maturation, and hormone secretion in mice [14]. Although these components individually contribute to ovarian function, their combined effects on JFP differ from their individual effects. Notably, only one study reported that JFP improved ovarian reproductive function by inhibiting IL-17 A/IL-6 and MEK/ERK signaling in cyclophosphamide-induced premature ovarian insufficiency (POI) rats [5]. To further elucidate the clinical efficacy of JFP in poor ovarian responders undergoing IVF-ET, we conducted a prospective, randomized controlled trial (RCT) and evaluated the mechanisms of JFP via multi-omics analysis.

Materials and methods

Drug preparation and quality control

JFP (Lot. 20210301) is the exclusive product (NMPA Approval Number: ZGB2022-34) of Tongyitang Pharmaceutical Co., Ltd. (Zhangzhou, Fujian, China). JFP is a black-coated, water-honeyed pill formulated with nine herbal components, including Panax ginseng C.A.Mey. (Renshen), Neolitsea cassia (L.) Kosterm. (Rougui), Epimedium brevicornu Maxim. (Yinyanghuo), Ligustrum lucidum W.T. Aiton (Nvzhenzi), and Reynoutria multiflora (Thunb.) Moldenke (Heshouwu). The preparation involves powdering selected herb, and decocting others (e.g., Epimedium brevicornu Maxim. (Yinyanghuo) and Curculigo orchioides Gaertn. (Xianmao), and combine extracts with honey to produce 1000 pills (1.8 g/10 pills, which is equivalent to 2.0 g of crude herbs). Moreover, the effective components in JFP were determined via liquid chromatography-mass spectrometry (LC-MS), and five components (icariin, ginsenoside Rb1, 2,3,5,4’-tetrahydroxystilbene 2-O-beta-D-glucoside, specnuezhenide, and curculigoside) were compared with the LuMet-TCM standard library.

Study design and subjects

From January 2021 to March 2023, poor ovarian responders diagnosed according to the Bologna criteria [15] were recruited from the Reproductive Medicine Centre of Sichuan Provincial Women’s and Children’s Hospital. Using computer-generated random numbers, poor ovarian responders aged 23 ~ 45 years were randomized 1:1 to either the JFP group (pretreated with JFP; Tongyitang Pharmaceutical Co., Ltd.) or the POR group (no pretreatment). During the same period, young women aged 23–34 years with normal ovarian reserve [antral follicle count (AFC) > 5, and anti-Müllerian hormone (AMH) ≥ 1.2 ng/mL] and regular menstrual cycles were enrolled in the control (Con) group [16]. The exclusion criteria included congenital uterine malformation, intrauterine adhesions, endometriosis, systemic lupus erythematosus, sicca syndrome, uncontrolled endocrinopathy, or the need for intracytoplasmic sperm injection (ICSI) due to male infertility. Additionally, poor ovarian responders were excluded if they had a known allergy to JFP or had received pretreatment with GH, CoQ10 or DHEA.

Controlled ovarian stimulation

All patients underwent the gonadotropin-releasing hormone (GnRH) antagonist protocol. Starting on day 2 of the menstrual cycle, 150 ~ 225 IU/day recombinant follicle-stimulating hormone (rFSH; Gonal-F, Merck Serono KGaA, Darmstadt, Germany) was administered. Patients in the JFP group received pretreatment with 10 pills of oral JFP twice daily beginning on day 2 of the menstrual cycle preceding COS. When the dominant follicle reached 14 mm in diameter or the serum level of LH exceeded 10 mIU/mL, 0.25 mg of GnRH antagonist (Ganirelix, Ocalon, USA) was subcutaneously injected daily until the trigger day. Triggering was performed with 250 µg of recombinant human chorionic gonadotropin (rHCG; Merck-Serono, Germany) when one or two follicles reached 18 mm in diameter. Oocyte retrieval was performed under transvaginal ultrasound guidance 36.5 h later. Cancellation of oocyte retrieval occurs in cases of follicular growth failure or premature ovulation.

In vitro fertilization and fresh embryo transfer

After 4 ~ 6 h of conventional fertilization at 37℃ in a 6% CO₂ and 5% O₂ culture environment, mature oocytes exhibiting the first polar body were classified into the metaphase II (MII) stage. Fertilization was confirmed 17 ~ 18 h later by the presence of two pronuclei. On day 3, cleaved embryos were morphologically evaluated according to the Istanbul criteria, with high-quality embryos defined as having eight equally sized, mononucleated blastomeres and < 10% fragmentation [17].

One or two top-grade embryos were transferred under transabdominal ultrasound guidance. Luteal phase support consisting of 90 mg/day vaginal progesterone (Crinone 8% gel, Merck, Germany) and 30 mg/day oral dydrogesterone (Duphaston; Abbott Healthcare Products B.V., Netherlands) was initiated on the day following oocyte retrieval.

Frozen-thawed embryo transfer and lutein phase support

Ovarian hyperstimulation syndrome (OHSS) was diagnosed and graded [18]. Fresh embryo transfer was cancelled if patients met any of the following criteria: a diagnosis of moderate or severe OHSS, a serum level of progesterone > 1.5 ng/mL on the trigger day, or endometrial abnormalities. At least one month later, one or two top-grade embryos were thawed and transferred to a frozen embryo transfer (FET) cycle via hormone replacement therapy. Luteal phase support was the same as that used for fresh embryo transfer. Only the first FET cycle was included in this study.

Outcomes of embryo transfer

Clinical pregnancy and live birth were defined as reported previously [19]. The implantation rate (IR), clinical pregnancy rate (CPR), multiple pregnancy rate and early miscarriage rate were calculated as reported previously [19]. The LBR was expressed as the number of live births per treatment cycle initiated.

Determination of the levels of IGF-1 and steroid hormones in follicle fluid

Upon oocyte retrieval, follicle flushing was avoided, and follicular fluid (FF) was collected and immediately centrifuged at 700 × g for 5 min at room temperature. The supernatant was stored at − 80 °C. The concentration of insulin-like growth factor-1 (IGF-1) was measured via an IMMULITE 2000 chemiluminescence assay system (Siemens Healthcare Diagnostics Inc., MA). E₂, progesterone, and total testosterone (TT) levels in FF were quantified via a Roche Cobas e602 electrochemiluminescence analyzer (Roche Diagnostics).

Network pharmacological analysis

Active components and targets of JFP

The chemical components of JFP were retrieved from the Traditional Chinese Medicine Systems Pharmacology (TCMSP) database (https://old.tcmsp-e.com/index.php) and the HERB database (http://herb.ac.cn/). Components meeting the criteria of oral bioavailability (OB) ≥ 30% and drug-likeness (DL) ≥ 0.18 were selected for further analysis. The targets of these compounds were identified, and the target gene names were standardized via the UniProt database (https://www.uniprot.org/).

POR-associated target genes

Target genes related to POR were identified by querying the GeneCards database (https://www.genecards.org/) via the keywords “poor ovarian response”, “decreased ovarian reserve”, and “premature ovarian insufficiency”. To screen for highly associated targets, the median of the relevance scores for all the targets was defined as the threshold. Only targets with scores above this threshold were retained for subsequent analysis.

Protein-protein interaction (PPI) network

The overlapping targets were imported into the STRING database (https://cn.string-db.org/) to generate PPI networks and TSV files. The TSV files were analyzed via Cytoscape 3.10.0 software to visualize the protein interaction network. The core targets were identified via the CytoHubba plugin.

GO and KEGG pathway analysis

The potential anti-POR targets of JFP were analyzed via the DAVID database (https://david.ncifcrf.gov/) for Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment. Significant biological process (BP), cellular component (CC), molecular function (MF), and pathway (P < 0.05) terms were identified to elucidate the molecular mechanisms of JFP.

Transcriptomic analysis

Total RNA was extracted via TRIzol reagent (Invitrogen, Carlsbad, California, USA) following the manufacturer’s instructions. RNA quality and quantity were assessed via an Agilent 2100 Bioanalyzer (Agilent, CA, USA). For mRNA enrichment, total RNA was denatured at an appropriate temperature to disrupt secondary structures, followed by mRNA isolation via oligo (dT)-attached magnetic beads. The enriched mRNA was converted into single-stranded circular DNA through denaturation and cyclization reactions. Linear DNA molecules were subsequently digested to obtain the final sequencing library. After quality control, RNA sequencing was performed, and gene expression levels were quantified via RSEM (v1.3.1). The Benjamini-Hochberg (BH) method was applied to adjust the P values to control the false discovery rate (FDR). Differentially expressed genes were identified with the criteria of a Log2 |FC| > 1 and an adjusted P value < 0.05.

Proteomic analysis

The protein samples were lysed via lysis buffer and ultrasonication, followed by centrifugation at 12,000 × g for 10 min at 4 °C to collect the supernatant. Dithiothreitol (DTT) and iodoacetamide (IAA) were added to achieve a final concentration of 11 mM. Equal amounts of protein from each sample were enzymatically digested. Peptides were separated via the 60-SPD (samples per day) method on an EvoSep One liquid chromatography system (EvoSep) and analyzed via timsTOF Pro 2 mass spectrometry with a capillary ion source. The precursors and fragments were detected via a time-of-flight (TOF) detector. Data-independent acquisition (DIA) data were processed via the DIA-NN search engine (v1.8). The FDRs of the proteins and peptides were adjusted to less than 1%. The differentially expressed proteins were subjected to GO secondary classification, subcellular localization analysis, and KEGG pathway enrichment.

Verification of molecular docking

The core compounds of JFP and their key targets for treating POR were selected for molecular docking. The three-dimensional (3D) structures of the core components of JFP were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/), whereas the 3D structures of the core target proteins were obtained from the Protein Data Bank (PDB) database (https://www.rcsb.org/). PyMOL software was used to prepare the ligand and protein structures by adding hydrogen atoms, removing water molecules, and performing other preprocessing steps. AutoDock Tools 1.5.6 was employed for protein and small molecule preparation. Binding activity was assessed on the basis of binding energy, and PyMOL software was used to visualize the results with the highest binding affinities.

Statistical analysis

The data were analyzed via SPSS 26.0 software. Continuous variables are expressed as median and interquartile range (IQR) and were analyzed via Kruskal-Wallis H test across three groups. Categorical variables are presented as numbers and percentages and were compared via the chi-square test across three groups. The P value < 0.05 was considered statistically significant. For pairwise comparison, Bonferroni correction was applied for post hoc analysis, and P value < 0.017 was considered statistically significant.

The primary outcome was LBR per treatment cycle. AT our center, LBR per treatment cycle among poor ovarian responders is approximately 25%, which was raised to 47% after the pretreatment with JFP in our preliminary experiment. Therefore, the sample size was calculated by assuming an increase in the LBR per treatment cycle from 25% to 47% following JFP pretreatment for poor ovarian responders. With an α of 0.05 and a β error of 0.2 (power = 80%), 58 patients per group were needed. Accounting for a dropout rate of 2 participants per group, the final sample size was set at 60 participants per group.

Results

Basal characteristics of the patients

The participant recruitment flowchart is shown in Fig. 1. In the initial cohort, 26 patients were excluded, and 199 were enrolled. During follow-up, 3, 1, and 2 patients were lost to follow-up in the JFP, POR, and Con groups, respectively (for career or personal reason). Consequently, 60, 64, and 69 patients in the JFP, POR, and Con groups, respectively, completed the study.

Fig. 1.

CONSORT flow diagram. Progression from recruitment to completion

Compared with those in the Con group, patients in the JFP and POR groups were significantly older and had higher serum basal level of FSH, but lower AFC, AMH, and TT levels (P < 0.017). No significant differences were detected among the three groups in terms of infertility duration, infertility type, gravidity, parity, body mass index (BMI), or basal levels of LH, E₂, progesterone, or prolactin (P > 0.05) (Table 1).

Table 1.

Basal characteristics of the patients

	JFP group (n = 60)	POR group (n = 64)	Con group (n = 69)	P value
Age (y) a, b	34 (31.50, 39.00)	35.50 (32.00, 39.25)	30.00 (27.75, 33.00)	< 0.001
Infertility duration (y)	2.00 (1.00, 5.00)	2.50 (1.00, 8.25)	3.00 (2.00, 4.00)	0.411
Type of infertility (%)				0.339
Primary	31.67% (19/60)	34.38% (22/64)	43.48% (30/69)
Secondary	68.33% (41/60)	65.62% (42/64)	56.52% (39/69)
History of gravidity (%)				0.602
0 gravidity	31.67% (19/60)	34.38% (22/64)	43.48% (30/69)
1 gravidity	33.33% (20/60)	29.69% (19/64)	30.43% (21/69)
> 1 gravidity	35.00% (21/60)	35.93% (23/64)	26.09% (18/69)
History of parity (%)				0.176
0 parity	66.67% (40/60)	60.94% (39/64)	79.71% (55/69)
1 parity	25.00% (15/60)	29.69% (19/64)	17.39% (12/69)
2 parities	8.33% (5/60)	9.37% (6/64)	2.90% (2/69)
BMI (kg/m2)	22.03 (20.36, 24.37)	23.17 (20.98, 25.68)	22.05 (20.28, 23.96)	0.123
AFC a, b	6.00 (4.00, 8.00)	6.50 (4.00, 8.50)	16.00 (11.75, 19.25)	< 0.001
AMH (ng/mL) a, b	0.73 (0.49, 1.03)	0.85 (0.33, 1.08)	3.48 (2.23, 4.58)	< 0.001
bFSH (mIU/mL) a, b	7.25 (6.36, 10.05)	7.22 (5.32, 9.40)	6.06 (4.91, 7.14)	< 0.001
bLH (mIU/mL)	4.14 (2.70, 6.54)	4.62 (3.12, 6.14)	4.37 (3.50, 5.74)	0.649
bE2 (pg/mL)	39.97 (27.94, 53.98)	35.95 (23.16, 48.71)	32.77 (27.62, 40.93)	0.081
Basal progesterone (ng/mL)	0.41 (0.38, 0.52)	0.35 (0.28, 0.54)	0.40 (0.30, 0.57)	0.040
TT (ng/mL) a, b	0.16 (0.10, 0.26)	0.13 (0.07, 0.22)	0.24 (0.19, 0.37)	< 0.001
prolactin (uIU/mL)	277.10 (183.40, 362.60)	311.10 (268.55, 448.55)	301.60 (229.95, 423.90)	0.157

JFP Jinfeng pills, POR poor ovarian response, Con control, BMI body mass index, AFC antral follicle count, AMH anti-Müllerian hormone, bFSH basal follicle-stimulating hormone, bLH basal luteinizing hormone, bE₂ basal estradiol, TT total testosterone

^a P < 0.017 JFP group vs. Control group

^b P < 0.017 POR group vs. Control group

Component identification of JFP

A total of 975 compounds were identified from JFP, among which the most abundant components were carbohydrates and glycosides, terpenes, flavonoids and phenylpropanoids. Five components (icariin, ginsenoside Rb1, 2,3,5,4’-tetrahydroxystilbene 2-O-beta-D-glucoside, specnuezhenide, and curculigoside) were compared with the LuMet-TCM standard library. The extracted ion chromatograms (EICs) of these components and their secondary mass spectra compared with those of the LuMet-TCM standard library are shown in Fig. 2. The fragment information of these five components in JFP corresponds well with the standard library.

Fig. 2.

Composition information and quality control of JFP. A Quantitative distribution diagram of JFP component classification. B Icariin. C Ginsenoside Rb1. D 2,3,5,4’-Tetrahydroxystilbene 2-O-beta-D-glucoside. E Specnuezhenide. F Curculigosid

JFP improved the levels of hormones in FF

Oocyte retrieval was cancelled in 5 patients each from the JFP and POR groups. Additionally, 6, 10, and 2 patients in the JFP, POR, and Con groups, respectively, had no transferable embryos (Fig. 1). One patient in the Con group developed moderate OHSS.

Compared with those in the Con group, the JFP and POR groups received a significantly greater dose of rFSH and a greater proportion of cycles without transferable embryos; however, these groups presented shorter rFSH durations, lower serum levels of E₂ on the trigger day and fewer oocytes retrieved (P < 0.017). The levels of E₂ and TT in FF were highest in the Con group, followed by those in the JFP and POR groups (P < 0.017). The MII rate and level of IGF-1 in FF were significantly lower in the POR group than in the Con group (P < 0.017). No significant differences were detected among the three groups in terms of the serum levels of progesterone and LH, endometrial thickness on the trigger day, rates of normal fertilization, cleavage embryo, and high-quality cleavage embryo, or incidence of moderate and severe OHSS (P > 0.05) (Table 2).

Table 2.

Controlled ovarian stimulation, in vitro fertilization and hormones in follicle fluid

	JFP group (n = 60)	POR group (n = 64)	Con group (n = 69)	P value
rFSH dose (IU) a, b	2400.00 (2025.00, 2700.00)	2362.50 (1968.75, 2700.00)	1675.00 (1371.88, 2100.00)	< 0.001
rFSH duration a, b	9.00 (8.00, 10.00)	9.00 (8.00, 10.00)	11.00 (10.75, 12.00)	< 0.001
E2 on the trigger day (pg/mL) a, b	973.80 (553.85, 1282.07)	1215.00 (877.40, 1551.50)	2515.50 (1623.25, 3989.75)	< 0.001
LH on the trigger day (mIU/L)	2.83 (2.18, 4.13)	3.54 (1.74, 4.60)	3.25 (1.03, 4.42)	0.192
progesterone on the trigger day (ng/mL)	0.68 (0.51, 1.07)	0.78 (0.64, 1.07)	0.87 (0.59, 1.19)	0.053
Endometrial thickness (mm)	10.00 (9.00, 12.00)	10.00 (8.75, 12.00)	9.00 (8.00, 11.25)	0.966
Oocytes retrieved a, b	3.00 (2.00, 4.00)	4.50 (3.00, 5.00)	13.00 (9.00, 18.25)	< 0.001
MII rate (%) b	85.31% (209/245)	77.69% (202/260)	85.71% (852/994)	0.006
Normal fertilization rate (%)	57.14% (140/245)	54.62% (142/260)	57.95% (576/994)	0.626
Cleavage embryo rate (%)	89.66% (156/174)	87.57% (148/169)	88.82% (564/635)	0.827
High-quality embryo rate (%)	33.33% (58/174)	27.81% (47/169)	35.12% (223/635)	0.202
Proportion of cycle without transferable embryo (%) a, b	18.33% (11/60)	23.44% (15/64)	2.90% (2/69)	0.002
Moderate and severe OHSS rate (%)	0	0	2.90% (2/69)	0.163
In FF
E2 (ng/mL) a, b, c	2450.00 (1603.00, 2842.00)	1302.00 (985.50, 1464.50)	2561.00 (2190.50, 3084.00)	< 0.001
progesterone (µg/mL)	50.30 (42.50, 62.52)	40.90 (36.10, 52.93)	48.00 (36.79, 63.87)	0.113
TT (ng/mL) a, b, c	2.85 (2.08, 4.55)	2.41 (1.45, 3.15)	5.12 (4.00, 6.79)	< 0.001
IGF-1 (ng/mL) b	81.20 (69.63, 104.25)	79.60 (64.00, 88.50)	87.90 (78.80, 112.00)	0.028

JFP Jinfeng pills, POR poor ovarian response, Con control, rFSH recombinant follicle-stimulating hormone, E2 estradiol, LH luteinizing hormone, MII mature oocyte, FF follicle fluid, TT total testosterone, IGF-1 insulin-like growth factor 1

^a P < 0.017 JFP group vs. Con group

^b P < 0.017 POR group vs. Con group

^c P < 0.017 JFP group vs. POR group

Effects of JFP on outcomes of embryo transfer

In the JFP, POR, and Con groups, 39, 43, and 56 patients, respectively, underwent fresh embryo transfer, respectively., whereas fresh embryo transfer was cancelled in 10, 6, and 11 patients, respectively, and FET was proceeded one month later (Fig. 1). The IR and CPR in embryo transfer cycles, as well as the LBR per treatment cycle were significantly lower in the POR group than in the Con group (P < 0.017), whereas there was no significant difference compared with those in the JFP group (P > 0.05). No significant differences were observed among the three groups in terms of the number of embryos transferred, the proportion of cycles involving blastocyst transfer, the rates of early miscarriage and multiple pregnancies, the gestational age at delivery, or the birth weight and height of the infants (P > 0.05) (Table 3).

Table 3.

Outcomes of embryo transfer

	JFP group	POR group	Con group	P value
Patients randomized (n)	63	65	71
Treatment cycles started (n)	60	64	69
Embryo transfer cycles (n)	49	49	67
Number of embryos transferred	2.00 (1.00, 2.00)	2.00 (2.00, 2.00)	2.00 (1.00, 2.00)	0.140
Proportion of cycle transferred with blastocyst (%)	14.29% (7/49)	8.16% (4/49)	20.90% (14/67)	0.165
IR (%) a	29.29% (29/99)	21.59% (19/88)	40.35% (46/114)	0.015
CPR (%) a	46.94% (23/49)	34.69% (17/49)	59.70% (40/67)	0.028
Early miscarriage (%)	13.04% (3/23)	11.76% (2/17)	10.00% (4/40)	0.932
Multiple pregnancy rate (%)	26.09% (6/23)	11.76% (2/17)	15.00% (6/40)	0.420
LBR per treatment cycle started (%) a	33.33% (20/60)	23.44% (15/64)	50.72% (35/69)	0.004
Gestational age at delivery (week)	38.00 (37.00, 39.00)	38.00 (37.00, 39.00)	38.00 (37.00, 40.00)	0.490
Birth weight (g)	2800.00 (2447.50, 3007.50)	3200.00 (2580.00, 3525.00)	3100.00 (2550.00, 3342.50)	0.220
Birth height (cm)	49.00 (48.00, 50.00)	49.50 (48.00, 50.00)	49.00 (48.00, 50.00)	0.531

JFP Jinfeng pills, POR poor ovarian response, Con control, IR implantation rate, CPR clinical pregnancy rate, LBR live birth rate

^a P < 0.017 POR group vs. Con group

One case of ectopic pregnancy had occurred in the Con group. A total of 81 infants (42 boys and 39 girls) were born without birth defects. The mean gestational age at delivery was 38 weeks (range: 33 ~ 41 weeks), with a mean birth weight of 2947 g (range: 1600 ~ 4300 g) and a mean birth height of 49 cm (range: 41 ~ 52 cm).

Prediction of key targets and potential pathways of JFP in POR treatment via network pharmacology

Herbal compounds and targets of JFP

Using the TCMSP database, 88 compounds were identified, including 23 from Epimedium brevicornu Maxim., 7 from Curculigo orchioides Gaertn., 8 from Leonurus japonicus Houtt., 15 from Neolitsea cassia (L.) Kosterm., 13 from Ligustrum lucidum W.T. Aiton, and 22 from Panax ginseng C.A.Mey. The HERB database yielded 96 compounds, with 3 from Equus asinus, 69 from Reynoutria multiflora (Thunb.) Moldenke, and 24 from Cervus elaphus Linnaeus. After removing duplicates, 94 herbal compounds and 1043 targets were retained. Table 4 lists the components with high OB and DL values for the selected JFP ingredients. Among these compounds, stigmasterol has been reported to have an intestinal absorption rate in humans of approximately 95%, confirming its potential for efficient utilization [20]. In addition, following oral administration, icariin is metabolized primarily into icaritin, which is rapidly absorbed. Pharmacokinetic analysis revealed that icaritin has a relatively high oral bioavailability [21–22]. Therefore, these pharmacokinetic data provide preliminary in vivo relevance for the predictions of network pharmacology, suggesting that JFP may possess promising in vivo activity.

Table 4.

Main compounds of JFP

MOL ID	Compound names	OB (%)	DL	Source
MOL004382	Yinyanghuo A	56.96	0.77	Epimedii Folium
MOL004373	Anhydroicaritin	45.41	0.44	Epimedii Folium
MOL004425	Icariin	41.58	0.61	Epimedii Folium
MOL000449	Stigmasterol	43.83	0.76	Curculiginis Rhizoma
MOL004146	Curculigosaponin C	39.31	0.19	Curculiginis Rhizoma
MOL003578	Cycloartenol	38.69	0.78	Curculiginis Rhizoma
MOL001421	Preleoheterin	85.97	0.33	Leonuri Herba
MOL001418	Galeopsin	61.02	0.38	Leonuri Herba
MOL000354	Isorhamnetin	49.6	0.31	Leonuri Herba
MOL005190	Eriodictyol	71.79	0.24	Ligustri Lucidi Fructus
MOL004576	Taxifolin	57.84	0.27	Ligustri Lucidi Fructus
MOL005146	Lucidumoside D	48.87	0.71	Ligustri Lucidi Fructus
MOL005314	Celabenzine	101.88	0.49	Ginseng Radix Et Rhizoma
MOL005321	Frutinone A	65.9	0.34	Ginseng Radix Et Rhizoma
MOL003648	Inermin	65.83	0.54	Ginseng Radix Et Rhizoma

JFP Jinfeng pills, MOL ID molecule identification‌, OB oral bioavailability, DL drug-likeness

Screening of POR-associated target genes

By combining the results from the Genecards database and removing duplicates, 921 disease targets were identified. A Venn diagram (Fig. 3A) revealed 132 overlapping targets between JFP and POR, defined as potential therapeutic targets for JFP in POR treatment (Supplemental Table 1).

Fig. 3.

Prediction of key targets and pathways of JFP in POR treatment via network pharmacology. A Venn diagram of overlapping target genes between JFP and POR. B PPI network of potential targets of JFP in POR treatment; node colour intensity reflects the degree value. C The top 10 enriched GO terms for biological process (BP), cellular component (CC), and molecular function (MF) related to overlapping targets. D The top 30 enriched KEGG pathways for overlapping targets

PPI network of targets

The 132 cotarget genes were analyzed via the STRING database with H. sapiens as the organism and a high-confidence interaction score threshold (0.700). The resulting TSV file was imported into Cytoscape 3.10.0 to construct a PPI network (Fig. 3B). Key targets, including TP53, AKT1, STAT3, TNF, IL6, PIK3CA, ESR1, IFNG, BRCA1, and MAPK3, were identified on the basis of degree, closeness, and betweenness centrality (Supplemental Table 2).

GO and KEGG analyses of cotarget genes

The 132 targets were subjected to GO analysis, with P < 0.05 used as the screening standard. Biological processes included positive regulation of gene expression, positive regulation of transcription by RNA polymerase II, positive regulation of cell population proliferation, and so forth. Cellular components included mainly the extracellular space, protein-containing complex, and extracellular region. The main molecular functions included identical protein binding, cytokine activity, and enzyme binding (Fig. 3C).

A stricter threshold (P < 0.01) was used to identify pathways via KEGG analysis, and the top 30 signaling pathways were mapped (Fig. 3D). The key pathways included pathways in cancer, PI3K-Akt signaling, proteoglycans in cancer, lipids and atherosclerosis, and the AGE-RAGE signaling pathway in diabetic complications.

Transcriptomic analysis of JFP in POR treatment

We first assessed the dispersion of gene expression across the Con, POR, and JFP groups, and revealed no significant differences in overall gene expression levels (Fig. 4A). Pearson correlation coefficients were calculated to evaluate the similarity of gene expression between samples, with higher coefficients indicating greater similarity (Fig. 4B). By statistically analyzing the difference information of each group and generating a histogram (Fig. 4C), the distributions of DGEs between different groups can be intuitively compared. Compared with the Con group, the POR group presented 550 upregulated and 318 downregulated genes (Fig. 4D). In contrast, compared with the POR group, the JFP group presented 226 upregulated and 991 downregulated genes (Fig. 4E).

Fig. 4.

Transcriptomic analysis of JFP in POR treatment. A Boxplot of gene expression levels across samples. B Correlation heatmap of sample similarity. C Differential gene expression statistics among the groups. D Volcano plot of differentially expressed genes (DEGs) between the POR and Con groups. E Volcano plot of DEGs between the JFP and POR groups. F Histogram of the results of the GO enrichment analysis of the DEGs. G KEGG pathway enrichment analysis of the DEGs

GO and KEGG analyses were performed to identify the functional enrichment and key pathways of all the DEGs. Biological processes included positive regulation of heart rate, skeletal system development, and mitotic spindle assembly, and so forth. Cellular components included mainly the extracellular region, collagen-containing extracellular matrix, and extracellular matrix. The main molecular functions included voltage-gated sodium channel activity, cytokine activity, ATP binding, and so forth (Fig. 4F). The key pathways included cytokine-cytokine receptor interaction, rheumatoid arthritis, vascular smooth muscle contraction, Cushing syndrome, TGF-beta signaling, and cortisol synthesis and secretion (Fig. 4G).

Proteomic analysis of JFP in POR treatment

This study revealed 38,745 identified peptides, 38,025 unique peptides, 6,036 identified proteins, and 4,576 comparable proteins (Fig. 5A). The analysis revealed that the majority of the peptides were 7–20 amino acids in length (Fig. 5B), which is consistent with the principles of enzymatic digestion and mass spectrometry. Additionally, most proteins corresponded to two or more peptide segments (Fig. 5C), enhancing the accuracy and reliability of the quantitative results. Differential protein analysis revealed distinct expression patterns among the groups (Fig. 5D). Compared with the Con group, the POR group presented 240 upregulated and 872 downregulated proteins (Fig. 5E). In contrast, compared with the POR group, the JFP group presented 832 upregulated and 353 downregulated proteins (Fig. 5F).

Fig. 5.

Proteomic analysis of JFP in POR treatment. A Summary of identified peptides, proteins, and quantifiable proteins. B Distribution of peptide lengths. C Distribution of peptide counts per protein. D Statistics of differentially expressed proteins among the groups. E Volcano plot of differentially expressed proteins (DEPs) between the POR and Con groups. F Volcano plot of DEPs between the JFP and POR groups. G Histogram of the results of the GO enrichment analysis of DEPs. H The top 20 enriched pathways from KEGG analysis. I COG/KOG functional classification of DEPs between the POR and Con groups. J COG/KOG functional classification of DEPs between the JFP and POR groups

GO and KEGG analyses revealed key biological processes, including platelet aggregation, cytoplasmic translation, and hydrogen peroxide catabolic processes. Cellular components were enriched in extracellular exosomes, cytosol, focal adhesion, and cytoplasm. The molecular functions included protein binding, RNA binding, and actin filament binding (Fig. 5G). The significant pathways included metabolic pathways, coronavirus disease-COVID-19, regulation of the actin cytoskeleton, neutrophil extracellular trap formation, and focal adhesion (Fig. 5H).

The COG/KOG database system was employed to classify homologous protein families and predict their functions. COG annotations provided insights into protein functional states, whereas KOG annotations facilitated the prediction of novel functions across species. The differentially expressed proteins were categorized into four functional groups: cellular processes and signaling, information storage and processing, metabolism, and poorly characterized. The POR/Con (Fig. 5I) and JFP/POR (Fig. 5J) groups presented the greatest number of proteins involved in cellular processes and signaling.

Integrated analysis of JFP in POR treatment

In this study, 17,438 transcripts and 6,036 proteins were quantified through transcriptomic and proteomic analyses, respectively. Comparative analysis revealed 5,883 genes quantified at both the transcriptome and proteome levels (Fig. 6A).

Fig. 6.

Integrated transcriptomic and proteomic analyses of JFP in POR treatment. A Venn diagram of overlapping genes between the transcriptome and proteome. B Distribution of differentially expressed proteins and transcripts between the POR and Con groups. C Distribution of differentially expressed proteins and transcripts between the JFP and POR groups. D Venn diagram of overlapping differentially expressed proteins and transcripts between the POR and Con groups. E Venn diagram of overlapping differentially expressed proteins and transcripts between the JFP and POR groups. F Venn diagram of overlapping genes from the transcriptome, proteome, and network pharmacology analyses. G Histogram of the results of the GO enrichment analysis of the overlapping genes. H KEGG enrichment analysis of the overlapping genes

To reduce the false positive rate in the transcriptomic differential analysis, we applied BH correction to the P values. Owing to the smaller dataset of differentially expressed proteins, P value validation was not performed. The distributions of the differentially expressed transcripts and proteins in the POR/Con (Fig. 6B) and JFP/POR (Fig. 6C) groups are shown. Venn diagrams illustrate the overlap between differentially expressed proteins and transcripts for the POR/Con (Fig. 6D) and JFP/POR (Fig. 6E) groups.

Integrated transcriptomic and proteomic analyses revealed 7 core genes regulated by JFP in POR: alkaline phosphatase (ALPL), heparanase (HPSE), keratin 1 (KRT1), low-density lipoprotein receptor (LDLR), methyltransferase-like protein 7B (METTL7B), tropomodulin-1 (TMOD1), and transforming growth factor beta-induced (TGFBI). Among these genes, ALPL was the only gene common to all three datasets, including the network pharmacology results (Fig. 6F).

GO and KEGG analyses of these 7 genes highlighted key biological processes, including peptide cross-linking, endochondral ossification, and negative regulation of cell adhesion. Cellular components were enriched in the cortical cytoskeleton, collagen-containing extracellular matrix, membrane, extracellular space, extracellular region and cytoskeleton. The enriched molecular functions included heparanase activity, phosphoethanolamine phosphatase activity, pyrophosphatase activity, hydrolase activity, acting on glycosyl bonds, structural constituents of the skin epidermis and cell adhesion molecule binding (Fig. 6G). The significant pathways included proteoglycans in cancer and glycosaminoglycan degradation (Fig. 6H).

Molecular docking verification of core genes and main compounds in JFP treatment of POR

Molecular docking was performed between the 7 core genes regulated by JFP in POR and the top 5 compounds identified in the drug-component-target-disease network of JFP (Supplemental Table 3). The docking outcomes are summarized in Table 5, with the top 5 molecular docking interactions illustrated in Fig. 7. A binding energy threshold of ≤ -5.0 kcal/mol indicates stable ligand-receptor interactions and strong binding activity. All the ligand-receptor pairs presented binding energies below − 5.0 kcal/mol, with ALPL showing the lowest binding energies for beta-sitosterol and estrone. These results demonstrate that the core components of JFP exhibit strong binding affinity with the core targets involved in POR treatment.

Table 5.

Binding energy of the molecular docking of the compound to the target

Targets	Binding energy(Kcal/mol)
	beta-sitosterol	estrone	kaempferol	quercetin	luteolin
ALPL	-9.8	-9.8	-8.5	-8.4	-8.3
HPSE	-7.7	-8.1	-7.4	-8	-8
KRT1	-7.9	-7.8	-7.2	-7	-7.1
LDLR	-5.8	-5.8	-6.4	-6.5	-6.6
METTL7B	-8.3	-7.8	-8.2	-8.3	-8.9
TMOD1	-7.3	-7.4	-7.2	-7.2	-7.1
TGFBI	-7.6	-8.1	-8.4	-8.4	-8.5

ALPL alkaline phosphatase, tissue-nonspecific isozyme, HPSE heparanase, KRT1 keratin, type II cytoskeletal 1, LDLR low-density lipoprotein receptor, METTL7B methyltransferase-like protein 7B, TMOD1 tropomodulin-1, TGFBI transforming growth factor-beta-induced protein ig-h3

Fig. 7.

Molecular docking verification. A Heatmap of binding energies from molecular docking. B ALPL-beta-sitosterol interaction. C ALPL-estrone interaction. D ALPL-kaempferol interaction. E METTL7B-luteolin interaction. F TGFBI-luteolin interaction

Discussion

This study revealed that JFP increased the levels of steroid hormones in the FF of poor ovarian responders undergoing IVF-ET. Integrated transcriptomic and proteomic analyses revealed 7 core genes (ALPL, HPSE, KRT1, LDLR, METTL7B, TGFBI, and TMOD1) as key targets of the therapeutic effects of JFP, with glycosaminoglycan degradation emerging as the primary regulatory pathway.

Granulosa cells (GCs) play a critical role in supporting oocyte development through steroid hormone synthesis, metabolism, and material transport (e.g., cAMP, IP3, and Ca²) [23–25]. In POR, GC dysfunction leads to reduced steroidogenesis and impaired folliculogenesis [26]. Testosterone, which is synthesized in theca cells, serves as a substrate for estrogen production in GCs [27]. During COS, the levels of E₂ and TT in FF typically increase with follicular development [27–28]; however, these levels are diminished in POR because fewer oocytes and dysfunctional GCs are present [26]. In our study, JFP significantly increased the levels of E₂ and TT in FF, suggesting improved endocrine function in GCs. IGF-1, a GC-secreted polypeptide, promotes folliculogenesis by increasing GC proliferation, LH receptor expression, aromatase activity, and steroidogenesis [29]. Although FF IGF-1 levels were lower in POR patients, which is consistent with prior studies [26, 30], JFP treatment resulted in a nonsignificant increase, potentially due to the small sample size.

Hu et al. [5] reported that JFP promotes folliculogenesis and mitigates cyclophosphamide-induced ovarian damage in rats with POI. In our study, JFP did not significantly improve the rates of MII, normal fertilization, or the quality of embryos in poor ovarian responders. Although JFP has been shown to increase endometrial receptivity by increasing uterine volume, estrogen receptor expression, and blood flow [9], its effects on IR, CPR and LBR in our study were not significantly increased despite trends toward improvement. This may be attributed to the following factors: (1) the small sample size; (2) the possibility of insufficient dosage and/or duration of JFP treatment; (3) patient heterogeneity with regard to age, ovarian reserve and basal levels of hormone; (4) limited improvement in GC function. Furthermore, the possibility that JFP’s effects are limited to specific subgroups (for example, based on age or ovarian reserve) was not analyzed due to small sample size. Although the sample size was calculated for the LBR, the difference in the LBR among the groups was not statistically significant. Thus, the present study may be underpowered to detect a modest between-group difference in LBR, and the possibility of a Type II error cannot be excluded. Therefore, the nonsignificant LBR findings should be interpreted cautiously. In the future, multicenter RCTs with large sample sizes and specific hypotheses or biomarkers are warranted to validate the potential benefit of JFP on the LBR. Notably, JFP did not adversely affect neonatal outcomes, supporting its safety in IVF-ET.

Notably, the baseline indicators (age, AMH, AFC, and so forth) were comparable between the JFP and POR groups, but significantly different between the two groups and the Con group, indicating that the ovarian reserve was poorer in the JFP and POR groups than in the Con group. After JFP intervention, the MII rate and level of IGF-1 in FF increased and did not significantly differ from those in the Con group, which suggested that JFP may improve ovarian function close to normal.

Integrated transcriptomic and proteomic analyses revealed seven core genes regulated by JFP in POR. Although the direct links between ALPL, HPSE, LDLR, TGFBI, TMOD1, and POR remain unexplored, these genes are involved in reproductive processes. ALPL is associated with ovulation, embryonic development, and embryo implantation [31–34]. HPSE, which is expressed in GCs and theca cells, may facilitate follicular rupture via extracellular matrix (ECM) remodeling [35–36]. LDLR promotes follicular development and steroidogenesis, with knockout models exhibiting reduced antral follicle and estrogen levels [37–39]. TGFBI is involved in embryogenesis through regulation of Wnt signaling [40]. TMOD1 is essential for cardiac development, with knockout embryos exhibiting developmental abnormalities [41–42]. KRT1 and METTL7B are linked to uterine and skeletal disorders, respectively, although their roles in POR remain unclear [43–44]. KEGG pathway analysis of these 7 core genes highlighted glycosaminoglycan degradation as a key mechanism. This pathway is associated with reproductive performance in Langshan chickens [45] and is enriched in mature equine cumulus-oocyte complexes [46], suggesting its potential role in oocyte metabolism and quality improvement in POR.

This study demonstrated that JFP enhances the levels of steroid hormones in FF and improves the function of GCs in poor ovarian responders undergoing IVF-ET. While JFP did not significantly improve pregnancy outcomes, it did not adversely affect neonatal outcomes, indicating its safety. Integrated multi-omics analyses revealed 7 core genes (ALPL, HPSE, KRT1, LDLR, METTL7B, TGFBI, and TMOD1) and the glycosaminoglycan degradation pathway as key mediators of the effects of JFP. These findings suggest that JFP has the potential to modulate ovarian function in POR. However, whether the therapeutic effects of JFP on POR through the regulation of the expression or activity of the aforementioned seven core genes and key signaling pathways was not confirmed in vitro or in vivo experiments in our study, and this is a major limitation. In subsequent studies, we will employ cultural models of GCs and animal experiments, to validate the association of the predicted targets and signaling pathways with the therapeutic effects of JFP on POR.

Abbreviations

AMH

anti-Müllerian hormone

ALPL

alkaline phosphatase

BH

Benjamini-Hochberg

BP

Biological processes

CC

Cellular components

Con

Control

CoQ10

Coenzyme Q10

COS

Controlled ovarian stimulation

CPR

Clinical pregnancy rate

DHEA

Dehydroepiandrosterone

DIA

Data-independent acquisition

DL

Drug-likeness

DOR

Diminished ovarian reserve

DTT

Dithiothreitol

E2

Estradiol

FET

Frozen embryo transfer

FF

Follicular fluid

FSH

Follicle-stimulating hormone

GH

Growth hormone

GnRH

Gonadotropin-releasing hormone

GO

Gene Ontology

HPSE

Heparanase

IAA

Iodoacetamide

ICSI

Intracytoplasmic sperm injection

IGF-1

Insulin-like growth factor-1

IR

Implantation rate

IVF-ET

in vitro fertilization and embryo transfer

JFP

Jinfeng Pills

KEGG

Kyoto Encyclopedia of Genes and Genomes

KRT1

Keratin 1

LBR

Live birth rate

LDLR

Low-density lipoprotein receptor

LH

Luteinizing hormone

MF

Molecular functions

MII

Metaphase II

OB

Oral bioavailability

OHSS

Ovarian hyperstimulation syndrome

PCOS

Polycystic ovary syndrome

PDB

Protein Data Bank

POI

Premature ovarian insufficiency

POR

Poor ovarian responders

RCT

Prospective randomized controlled trial

SD

Standard deviation

TCM

Traditional Chinese medicine

TCMSP

Traditional Chinese Medicine Systems Pharmacology

TGFBI

Transforming growth factor beta-induced

TMOD1

Tropomodulin-1

TOF

Time-of-flight

TT

Total testosterone

Authors' contributions

YG and FYL conceptualized and designed the study. XMK, THH and FRC performed the experiments and wrote the manuscript. QWL and XYY performed the network pharmacology analysis. YNZ, DSX, YQ and WL helped with sample collection. XMK and QWL conducted the data analysis. All the authors have read and approved the final manuscript.

Funding

This study has been jointly supported by grants from the Scientific and Technological Innovation Project of the Sichuan Maternal and Child Health Care Association (22FXZD01), the Scientific Research Project of the Sichuan Medical Association (S22015), the Scientific Research Project of the Chengdu Municipal Health Commission (2022155), the National Center for Women and Children’s Health, the China CDC “Maternal and Infants Nutrition and Health Research Programs” (2023FYH012), and the Natural Science Foundation of Sichuan Province (2024NSFSC0695).

Data availability

The data will be made available upon request.

Declarations

Ethics approval and consent to participate

This RCT was approved by the Medical Ethics Committee of Sichuan Provincial Women’s and Children’s Hospital and registered with the Chinese Clinical Trial Registry Centre (Registration No. ChiCTR2100051925). Written informed consent was obtained from all participants prior to enrollment. The study adhered to the ethical standards of relevant national and institutional committees on human experimentation and complied with the principles of the Helsinki Declaration (1975, revised in 2013).

Consent for publication

This manuscript is original, has not been published or submitted elsewhere, and is not under consideration by any other journal. All authors approved the final version and consent to its submission. All participants provided written informed consent for the publication of this manuscript.

Competing interests

The authors declare no competing interests.