Perinatal exposure to a low dose of bisphenol-A advances pubertal spermatogenesis of F1 adolescent male rats


 Background

To investigate the impact of perinatal exposure to a low dose of bisphenol A (BPA) on spermatogenesis in male rats and the underlying mechanism.
Methods

Female rats were injected subcutaneously with 2 µg BPA/kg/day from gestation day 10 through lactation day 7. The spermatogenesis and expression of key regulatory genes in the testes as well as the central modulators of the hypothalamic-pituitary-gonadal axis were determined in male offspring on postnatal day 18, 21, and 24 (PND18, 21, and 24).
Results

1) Perinatal BPA exposure led to an increase in the weight of body and testis in PND21-24 male offspring. The seminiferous tubular diameter and the number of round spermatids were significantly increased in PND21 BPA-rats, while the volumes of the Sertoli cells, spermatogonia and spermatocytes were not significantly altered. 2) Compared to the control rats, the expression levels of key meiotic regulators such as cyclinA1, c-jun and c-fos in the seminiferous tubules were significantly elevated in PND21 BPA-rats. 3) The plasma levels of FSH and LH (PND21 and PND24) as well as the frequency of pulsatile LH secretion (PND21) were significantly increased in BPA-rats, although the plasma levels of testosterone and estrogen showed no significant difference between the two groups. 4) In comparison with control rats, the levels of GnRH mRNA in the preoptic area (POA) and kiss1 mRNA in arcuate nucleus (ARC) were significantly increased in the BPA-rats, whereas the level of ERα mRNA in ARC was decreased, although the number of GnRH-positive cells and ARC kisspeptin-positive cells were unchanged. Interestingly, neither the number of kisspeptin-positive cells nor the level of kiss1 mRNA in the anteroventral periventricular nucleus (AVPV) showed a difference between the two groups.
Conclusion

Perinatal exposed to a low dose of BPA leads to an increased meiosis of spermatocytes and promotes the spermatogenesis in male offspring, most likely through activation of the hypothalamic-pituitary-gonadal axis.


Abstract Background
To investigate the impact of perinatal exposure to a low dose of bisphenol A (BPA) on spermatogenesis in male rats and the underlying mechanism.

Methods
Female rats were injected subcutaneously with 2 µg BPA/kg/day from gestation day 10 through lactation day 7. The spermatogenesis and expression of key regulatory genes in the testes as well as the central modulators of the hypothalamic-pituitary-gonadal axis were determined in male offspring on postnatal day 18, 21, and 24 (PND18, 21, and 24). Results 1) Perinatal BPA exposure led to an increase in the weight of body and testis in PND21-24 male offspring.
The seminiferous tubular diameter and the number of round spermatids were signi cantly increased in PND21 BPA-rats, while the volumes of the Sertoli cells, spermatogonia and spermatocytes were not signi cantly altered. 2) Compared to the control rats, the expression levels of key meiotic regulators such as cyclinA1, c-jun and c-fos in the seminiferous tubules were signi cantly elevated in PND21 BPA-rats. 3) The plasma levels of FSH and LH (PND21 and PND24) as well as the frequency of pulsatile LH secretion (PND21) were signi cantly increased in BPA-rats, although the plasma levels of testosterone and estrogen showed no signi cant difference between the two groups. 4) In comparison with control rats, the levels of GnRH mRNA in the preoptic area (POA) and kiss1 mRNA in arcuate nucleus (ARC) were signi cantly increased in the BPA-rats, whereas the level of ERα mRNA in ARC was decreased, although the number of GnRH-positive cells and ARC kisspeptin-positive cells were unchanged. Interestingly, neither the number of kisspeptin-positive cells nor the level of kiss1 mRNA in the anteroventral periventricular nucleus (AVPV) showed a difference between the two groups. Conclusion Perinatal exposed to a low dose of BPA leads to an increased meiosis of spermatocytes and promotes the spermatogenesis in male offspring, most likely through activation of the hypothalamic-pituitarygonadal axis.

Background
Environmental estrogens are synthetic chemicals that could interfere with the production, metabolism and physiological effects of endogenous estrogens, thereby inducing endocrine disorders including Page 3/14 neurological, developmental and reproductive diseases in humans as well as animals [1]. Bisphenol-A (BPA), as an environmental estrogenic chemical, has attracted much attention because of its high risk of human exposure via BPA-containing products such as food containers, dental sealants and industrial pollution [2,3]. It is noteworthy that BPA is capable of passing through the placental and blood-brain barriers, and has been detected in placental tissues, amniotic uids and milk of lactating mothers [4]. Particularly, the fetuses and neonates are more sensitive to the toxicity of BPA by their fast developing and immature nature [5]. Numerous studies have demonstrated a close link between BPA exposure and infertility, embryonic defects, low sperm count and increased morphological abnormalities in spermatozoa [6,7,8]. Moreover, maternal exposure to a high dose of BPA has been shown to negatively impact the sperm count and motility parameters of F1 adult mice [7], suggesting a hereditary hazard. It is unclear whether these effects are caused by a direct action of BPA on testis or due to the alterations in reproductive endocrine system induced by BPA. In this context, the effects and mechanism of low-dose or environmental level of BPA exposure is of a particular concern.
Spermatogenesis is a complex and tightly controlled biological process including mitosis and meiosis stages [9]. Diploid spermatogonia differentiate into primary spermatocytes, which divide into round haploid spermatids without additional DNA replication. Subsequently, round spermatids differentiate into mature sperm through a series of morphological changes. Cyclins and cyclin-dependent kinases (Cdk), especially cyclin A1, play an important role(s) in both mitosis and meiosis stages of spermatogenesis [10,11,12]. Upregulation of Cyclin A1 and its activation of Cdk2 represents a critical step of meiosis during normal spermatogenesis in mammalian [13,14,15]. While some studies did not observe a signi cant effect of the low dose BPA on spermatogenesis [7], some data suggested that perinatal exposure to either low dose (L-BPA) or high dose (H-BPA) of BPA could cause sperm malformation in the male offspring testis [8]. It was also reported that neonatal exposure of male rats to BPA could alter the testis structure and advance the rst wave of spermatogenesis at puberty [16].
The rst wave of spermatogenesis occurs as a consequence of activated hypothalamic-pituitary-gonadal (HPG) axis [17]. The activation of gonadotropin-releasing hormone (GnRH) neurons in POA of the hypothalamus leads to an increased frequency of pulsatile GnRH secretion [18]. GnRH acts on the anterior pituitary to stimulate the secretion of luteinizing hormone (LH) and follicle stimulating hormone (FSH), which regulates sexual maturation and gametogenesis [18]. Approximately 90% of GnRH neurons express G protein-coupled receptor 54 (GPR54) and are intensely activated by kisspeptin-expressing neurons located in AVPV and ARC of the hypothalamus [19]. Kisspeptin-expressing neurons in ARC are considered to be the generator of pulsatile GnRH secretion [20]. Redmond et al report that kiss1 expression in the middle ARC is increased gradually during juvenile development, which is involved in the acceleration of pulsatile LH release [21]. Neonatal exposure to BPA has been reported to increase the GnRH pulse frequency, as indicated by an increased number of peaks per hour in female rats [22]. Our previous studies have demonstrated that perinatal exposure to a low dose of BPA caused a sustained increase of AVPV kisspeptin neurons in male rats, which led to the generation of a normally estradiolinduced LH-surge in adult male rats [23]. However, it remains unclear if a low dose of BPA exposure during perinatal period can affect kisspeptin neurons in ARC of F1 adolescent male rats.
In this study, female breeders were exposed to a relatively low dose of BPA from gestation through early lactation period of time. In their male offspring (BPA-rats), pubertal changes around the rst wave of spermatogenesis was examined, and histological and functional changes in HPG axis, especially those occurred in the ARC kisspeptin neurons, were determined and analyzed.

Experimental animals
The usage of animals was approved by the Animal Care and Ethical Committee of Nanjing Medical University. Every effort was made to use the minimal number of animals and to limit animal suffering.
Adult Sprague-Dawley rats (Oriental Bio Service, Inc, Nanjing, China), weighing 250-300 g before experiments, were used in the study. Animal cages were maintained on a 12:12 light-dark cycle starting at 0700 h and kept at a temperature of 22-23°C. The animals were permitted free access to food and tap water.

Preparation of BPA-rat model
The female rats with two consecutive regular 4-day estrous cycles were used to mate with adult male rats. Adult male and female rats were placed in the same cages, and the morning of nding a vaginal plug was designated as day 0 of pregnancy. BPA (>99% purity; Sigma-Aldrich Inc., St. Louis, MO) was dissolved in dimethyl sulfoxide, and diluted with olive oil for injection. Pregnant females were injected subcutaneously with BPA solution at a dose of 2 μg/kg per day from day 10 of pregnancy to day 7 of lactation. The dose was chosen based on a recent report that environmental exposure of BPA is approximately at a level equivalent to 2.4 μg/kg per day [24]. The control female rats were treated with vehicle at the same volume. The offspring were weaned on postnatal day 22, and transferred to plastic hanging cages (2-4 rats/cage). Testicular and brain tissues as well as blood samples were collected on the postnatal day (PND) 18, 21 and 24, respectively.

Histological examination of testis
After xation in Bouin's Fluid, testis tissues were dehydrated through graded series of alcohol, cleared in xylene and embedded in para n wax. Sections (5-μm-thick) were stained with hematoxylin and eosin with routine processing. Average diameter of seminiferous tubule (ST) was determined by measuring 200 ST with ocular micrometer using a conventional light microscope (Olympus PD70) at 10× objective. The percentage of ST containing round spermatids was determined in 100 cross-sectioned tubules at 20× objective. To assess the rst wave of spermatogenesis, standard point-counting of cell nuclei was used to determine the nuclear volume per testis of Sertoli cells and germ cells as previously reported [16]. The nuclear volume of the spermatocytes and spermatogonia per unit of Sertoli cell nuclear volume was calculated as an index of spermatogenic e ciency. Using a systematic clock-face sampling pattern from a random starting point, 16 elds in 4 sections were counted under oil immersion using a Leitz×363 plan apochromatic objective tted to a Leitz Laborlux microscope and a 121-point eyepiece graticule. Points falling over the nuclei of Sertoli cells and germ cells were scored and expressed as a percentage of the 121 points. The relative nuclear volumes per testis were converted to absolute nuclear volumes by reference to testis weight.

Measurement of hormones
Blood samples were taken from rats anesthetized with 10% chloral hydrate (400 mg/kg i.p.) during the light phase 0900-1000 h by jugular venipuncture. To determine the pulsatile LH secretion, blood samples were collected every 10 min from 0900 h for 3 h through a silicon cannula (o.d. 1.0 mm, i.d. 0.5 mm; Shin-Etsu Polymer, Tokyo, Japan) that was inserted into the right atrium through the jugular vein. An equal volume of heparinized saline was given through the atrial cannula after each blood collection. Plasma (300 μl per rat) was separated by centrifugation at 4°C and stored at -80°C until use. Plasma levels of LH, FSH, testosterone (T) and E2 were measured using a radioimmunoassay (RIA) kit provided by the National Hormone and Peptide Program (Baltimore, MD USA). The intra-and inter-assay coe cients of variation were 5.5% and 8.9% for LH, 4.3% and 10.3% for FSH, 6.2% and 7.4% for T, 6% and 5.8% for E2, respectively. Pulsatile LH secretion (min) was calculated by measuring the mean inter-pulse interval in increments of LH secretion.

Immunohistochemistry of kisspeptin and GnRH
The rats were anesthetized and perfused intraventricularly with ice-cold phosphate-buffered saline followed by 4% paraformaldehyde between 1600-1800 h. Brain tissues were removed and xed overnight, and were incubated in 15% and 30% sucrose gradually until settled. Sections (40-μm-thick) were cut in the coronal plane using a cryostat.
For kisspeptin immunohistochemistry, brain sections were incubated in 1% normal fetal goat serum Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) Total RNA was isolated from testis and hypothalamus using the Trizol reagents (Invitrogen Life Technologies, France). RNA (1 μg) was reverse transcribed using the High-Capacity cDNA Reverse Transcription Kit (TaKaRa Biotechnology CO., Ltd, Shanghai, China). Published primer sequences of target genes (Table 1) were used for PCR [25]. Quantitative PCR was performed using a Light Cycler Fast Start DNA Master SYBR Green I kit and an ABI Prism 7300 Sequence Detection System (Applied Biosystems, Foster City, California, USA). Relative mRNA levels were quanti ed using the 2 -ΔΔCt method via normalization to GAPDH expression. The mRNA level of each experimental group was expressed as a percentage of the value in corresponding control group.

Data analysis/statistics
Data were retrieved and processed with the software Micro Cal Origin 6.1. Data were expressed as the means ± standard error (SE). The difference between two groups was evaluated using the Student's t test.
When analyzing data from more than 2 groups, differences among groups was evaluated using one-way ANOVA followed by Bonferroni's post-hoc tests. Statistical analysis was performed using the State7 software (STATA Corporation, USA). Differences with P<0.05 were considered statistically signi cant.

BPA effects on body and testis weights
Data analysis showed that there was no difference in the body and testis weights between the control and BPA groups prior to PND18. However, from PND21 to PND24, body and testis weights of the BPA-rats became signi cantly higher than those of the age-matched controls (P<0.05, Fig. 1). For this reason, further studies were focused on changes of testicular parameters and the upstream regulators in PND18, 21 and 24, respectively.

BPA-induced changes of testis morphology and spermatogenesis
Since the round spermatids were derived from meiosis of spermatocyte cells, their number indicates meiotic activities. As shown in Fig. 2, round spermatids started to be observed in the testes of PND24 control and BPA groups. Interestingly, both the number of round spermatids in seminiferous tubule (ST) and the ratio of ST with round spermatids were signi cantly increased in PND24 BPA-rats compared to controls (P<0.01, Table 2), whereas the Sertoli cell nuclear volume, spermatogonia and spermatocyte nuclear volume per unit Sertoli cell nuclear volume showed no signi cant difference between the two groups. In addition, the diameter of seminiferous tubules was signi cantly increased in PND24 BPA-rats, but not in PND18 or 21 BPA-rats, compared to the correspondent controls (P<0.05, Table 2). These results indicated that perinatal exposure to the low dose of BPA could promote pubertal development of the seminiferous tubule and the meiosis of germ cells.

Expression of meiosis-related factors in control and BPA groups
To explore the mechanism by which BPA promotes meiosis of testicular spermatogenic cells, the expression of the key regulators controlling testicular meiosis such as cyclins and AP-1 were examined in PND 18, 21 and 24 among the control and BPA groups. The levels of Cyclin A1 were found to be signi cantly elevated in BPA-rats on PND21 and PND24, but not on PND18, compared to age-matched control rats (P<0.05; P<0.01, Fig. 3A). However, the Cyclin B1 mRNA level was nearly the same in the two groups (Fig. 3B). The expressions of AP-1 factors, c-jun and c-fos, were also determined since they are upstream regulators of the Cyclin A1 gene [26]. While the levels of c-jun and c-fos mRNA in PND21 BPArats were increased approximately 2-fold compared to age-matched control rats (P<0.01, Fig. 3C and D), there was no signi cant difference in c-Jun and c-fos mRNA levels on either PND18 or PND24 control and BPA-rats.
BPA treatment led to changes in the levels of reproductive hormones In comparison with control rats, the plasma levels of LH and FSH were signi cantly elevated in PND21 and 24 BPA groups (P<0.05, Fig. 4A and B), but not in the PND18 BPA group. In addition, the frequency of pulsatile secretion of LH in PND21 BPA-rats (IPI=36.7±4.02 min) increased approximately 20% in comparison with the control rats (IPI=47.2±4.6 min, P<0.05, Fig. 4C), suggesting an activated production of pituitary peptide hormone following perinatal BPA exposure. Thus, the time window of changes in LH and TSH levels was in a general agreement with that of changes in spermatogenesis. On the other hand, while the levels of T and E2 tend to be lower in the BPA groups at all the time points, the difference failed to reach a statistical signi cance ( Fig. 4D and E), highlighting the effect of BPA on endocrine system rather than directly on the testes.

Changes in hypothalamic GnRH and kisspeptin neurons following BPA treatment
Since the pituitary secretion of LH and FSH is controlled by hypothalamic peptide hormones, we further examined the effects of BPA on the GnRH and kisspeptin neurons. The results showed that in comparison with the control rats, the average level of POA GnRH mRNA was signi cantly increased in PND21 BPA-rats (P<0.05, Fig. 5A). While the number of brown bipolar GnRH-immunoreactive (GnRH + ) cells showed a trend of increase, the change did not reach a signi cant level (Fig. 5B). Similarly, the level of ARC kiss1 mRNA in PND21 BPA-rats was obviously higher than that in controls (P<0.01, Fig. 5C), but there had no signi cant difference in the number of ARC kisspeptin + cells (Fig. 5D). A very limited amount of kisspeptin + cells, lying close to the third ventricle, was observed in AVPV of PND21 rats (Fig. 5F).
Interestingly, neither the level of kiss1 mRNA nor the number of kisspeptin + cells in AVPV showed a signi cant difference between the control and BPA groups ( Fig. 5E and F). In addition, the level of ARC ERα mRNA was signi cantly decreased in PND21 BPA-rats (P<0.05, Fig. 5G), while the level of ARC ERβ mRNA showed no difference between the two groups. As discussed later, these results indicated that the perinatal BPA exposure increases kiss1 gene expression likely through reducing the E2-mediated negative feedback regulation in ARC kisspeptin-expressing neurons, which led to the activation of the hypothalamic-pituitary-gonadal axis.

Discussion
Perinatal BPA exposure advancesspermatogenesis through affectingmeiosis BPA, as an environmental endocrine disruptor, can mimic or antagonize the effects of endogenous estrogens and interfere with normal reproductive endocrine functions. It was reported that different doses of environmental estrogens may exert divergent, and sometimes seemingly opposite effects on reproductive functions. Atanassova et al. found that neonatal exposure of male rats to a high dose (10 µg/per rat) of diethylstilbestrol caused a retardation of pubertal spermatogenesis, as evidenced by decreases in testis weight, lumen formation, and spermatocyte nuclear volume per unit Sertoli cell, while low doses (0.1-0.01 µg/per rat) of diethylstilbestrol signi cantly increased spermatocyte nuclear volume per unit Sertoli cell [16]. In addition to the dose, the effect of BPA on male reproductive development also varied depending on the endogenous levels of estrogens, time windows, and route of administration. It was reported that exposure to a low dose of BPA (2.4 µg/kg/d) during early adolescence resulted in decreased testicular weight as well as the activities of testicular steroidogenic enzymes [24]. However, Atanassova et al observed that neonatal exposure to a similar dose of BPA led to an increase in testicular weight and promotion of spermatogenesis in male offspring [16]. We found that perinatal exposure to a low dose (2 μg/kg/day) of BPA had little or no effect on the Sertoli cell nuclear volume, spermatogonia or spermatocyte nuclear volume per unit Sertoli cell nuclear volume, but resulted in a signi cant increase in the number of round spermatids in PND24 BPA-rats, which is reminiscent of the report by Atanassova et al. It is noteworthy that this effect appeared to be mostly manifested on the spermatocyte meiosis, and in a speci c time window of PND24.
In order to explore the possible mechanism by which BPA promotes meiosis of spermatogenic cell, we examined the gene expression of the relevant regulatory factors in meiosis. Cyclins play an important role in the process of spermatogenic cell division. Cyclin A1, a member of the mammalian A type cyclin family involved in both G1/S and G2/M transitions, is abundantly expressed in spermatocytes and is essential for spermatogenesis in the mice [15]. It was reported that the expression of cyclin A1 increased signi cantly during the initiation of spermatogenesis [27]. On the other side, cyclin A1 de ciency in spermatocytes was associated with the arrest of spermatogenesis at late meiotic prophase, leading to spermatocyte apoptosis and sterility of male mice [27]. In this study, we found that perinatal exposed to a low dose of BPA caused a signi cant increase in the expression of cyclin A1 mRNA in the testis of the prepubertal male offspring, while the mRNA levels of cyclin B did not signi cantly change. These results highlight the signi cance of cyclin A1 for mediating the BPA effects in the testis. It was reported that AP1 family members of c-jun and c-fos serve as upstream positive regulators of the cyclins [28]. Cyclin A1 was elevated in c-jun-overexpressing Rat1a cells [26]. In addition, overexpression of c-fos in serumstarved vascular smooth muscle cells resulted in an induction of cyclin A gene promoter activity in a CREdependent manner [29]. Our study demonstrates that the expression of c-jun and c-fos genes were increased in the testis of PND21 BPA-rats, suggesting that BPA may activate the transcription of cyclin A1 gene most likely through the up-regulation of c-jun and c-fos genes in the testis.
Perinatal BPA exposure promotes the secretion of gonadotropins Spermatogenesis is tightly controlled by local production of androgen in response to the pituitary hormone FSH and LH. Both FSH and androgen regulate the late stages of Sertoli cell proliferation. Johnston et al reported that the number of Sertoli cell was signi cantly reduced in either FSH or androgen receptor knockout mice during pre-puberty [30]. It was also reported that FSH increased the number of spermatocyte nuclear volume per unit Sertoli cell nuclear volume [16]. Our results showed that the plasma FSH levels in PND21 and 24 BPA-rats were signi cantly increased in comparison with controls, and indeed, the number of spermatids were increased. However, no signi cant change was detected in Sertoli cell nuclear volume. The possible reason for this discrepancy could be that the best time window to observe Sertoli cell proliferation might be missed. Sertoli cell proliferation may have completed even before PND18, and the later increase of FSH level induced by BPA around PND21 failed to affect Sertoli cell proliferation. Testosterone is also required for spermatocytes to complete meiosis to form spermatids [31]. No difference was found in testosterone levels between the BPA-rats and controls in spite of elevation of the LH levels at PND24. BPA was found to inhibit the expression of LH receptor in human decidual cells in vitro [32], and this effect may offset the effect of increased LH in spermatogenic cells.
The rst wave of spermatogenesis was induced by the activation of HPG axis. The onset of HPG axis activation is characterized by an increased frequency of pulsatile GnRH/LH secretion [18]. As shown in this study, the frequency of pulsatile LH secretion in PND21 BPA-rats increased approximately 20% compared to the control rats. Meanwhile, the level of GnRH mRNA in POA was also signi cantly elevated in PND21 BPA-rats. Taken together, our results suggested that low dose of BPA exposure may promote spermatogenesis by the activation of GnRH neurons and increased gonadotropin secretion in male offspring. The BPA effect in testis and spermatogenesis is mediated by increased gonadotropins, although the direct effect of BPA on testis could not be excluded.
Perinatal BPA exposure promotes the activation of ARC kisspeptin-expressing cells Kisspeptin-expressing neurons in ARC are considered to be the generator of pulsatile GnRH/LH secretion, and activation of kisspeptin-expressing neurons represents a key event for the puberty onset of the male rats [20]. ARC kisspeptin-expressing neurons in male rats began to be observed at 3 days after birth, and the expression of ARC kisspeptin-expressing neurons increased sharply at the onset of puberty [33].
Although there was no sexual difference in the number of ARC kisspeptin-expressing neurons between male and female rats, the density of ARC kisspeptin-expressing nerve bers in adult female rats was much higher than that in male rats [34]. It has been reported that ARC kisspeptin mRNA levels in adolescent rats are 1.5 times higher than that in infancy [25]. One interesting observation in present study is that the level of ARC-kiss1 mRNA in PND21 BPA-rats increased approximately 1.5-fold compared to that in control rats, although the number of kisspeptin-expressing neurons was not altered. Thus, perinatal BPA exposure is able to promote pubertal activation of ARC kisspeptin-expressing neurons in male offspring. Kisspeptin expression is known to be negatively regulated by the circulating E2 through an ERα-mediated action in ARC [35]. We observed that while the E2 levels did not signi cantly changed, ERα expression in ARC of PND21 BPA-rats was reduced. These ndings suggested that the perinatal BPA exposure increased the levels of ARC-kisspeptin mRNA probably through antagonizing the E2 activity in ARC. It has been well-recognized that phytoestrogens could exhibit either estrogenic or anti-estrogenic effects depending on the endogenous estrogen levels and ER expression in different tissues [36].
Similarly, both estrogenic and anti-estrogenic activities of BPA were reported. The detailed molecular mechanisms by which BPA increases kisspeptin expression remain to be investigated.

Conclusions
The developmental processes of steroid-sensitive network must be carefully orchestrated during the late embryonic and early postnatal periods. Exogenous endocrine-disrupting chemicals during the critical periods of hypothalamic development may exert their effects upon hypothalamic brain regions, resulting in improper hypothalamic programming and reproductive endocrine disruption. Importantly, these effects may be transmitted to future testicle development by epigenetic and other mechanisms. This study makes use of rats, and the experimental protocol for the use of animal was approved by the Animal Care and Ethical Committee of Nanjing Medical University.

Consent for publication
Not applicable.

Availability of supporting data
All data generated or analysed during the current study are available from the corresponding author on reasonable request.