|Home | About | Journals | Submit | Contact Us | Français|
Progesterone can block the oestradiol-induced GnRH/LH surge and inhibit LH pulse frequency. Recent studies reported that progesterone prevented premature LH surges during ovarian hyperstimulation in women. As the most potent stimulator of GnRH/LH release, kisspeptin is believed to mediate the positive and negative feedback effects of oestradiol in the hypothalamic anteroventral periventricular (AVPV) and arcuate (ARC) nuclei, while the region-specific role of progesterone receptors in these nuclei remains unknown. This study examined the hypothesis that progesterone inhibits LH surge and pulsatile secretion via its receptor in the ARC and/or AVPV nuclei. Adult female rats received a single injection of pregnant mare serum gonadotropin followed by progesterone or vehicle. Progesterone administration resulted in a significant prolongation of the oestrous cycle and blockade of LH surge. However, microinjection of the progesterone receptor antagonist, RU486, into the AVPV reversed the prolonged cycle length and rescued the progesterone blockade LH surge, while RU486 into the ARC shortened LH pulse interval in the progesterone treated rats. These results demonstrated that progesterone’s inhibitory effect on the GnRH/LH surge and pulsatile secretion is mediated by its receptor in the kisspeptin enriched hypothalamic AVPV and ARC respectively, which are essential for progesterone regulation of oestrous cyclicity in rats.
The positive and negative feedback effects of oestradiol (E2) and progesterone are essential in regulating the cyclical activity of the hypothalamic-pituitary-ovarian axis. Acting as a major inhibitory brake in the luteal phase of the ovarian/menstrual cycle, progesterone inhibits gonadotrophin-releasing hormone (GnRH) and luteinising hormone (LH) secretion1. Moreover, when administered before or concurrent with E2, progesterone inhibits E2 positive feedback and abolishes the preovulatory GnRH and gonadotrophin surge. This blockade of the LH surge, observed in many species, including the rat2, ewe3, monkey4 and women5–7, is critical for synchronizing the wave of follicular development in the ovary and maintaining the length of luteal phase8, 9. Disturbances in progesterone inhibitory feedback have been implicated in infertility associated with enhanced GnRH/LH secretion10. Clinically, premature spontaneous LH surges are a major cause of cycle cancellation in women11 and recently, progesterone has been shown to successfully prevent premature LH surges in women undergoing ovarian stimulation, thereby validating its use in in vitro fertilization (IVF) regimes6, 7, 12. Despite its clinical and physiologic importance, the neural mechanisms underlying the inhibitory actions of progesterone on surge, and indeed pulsatile, release of GnRH/LH remain poorly understood.
Although progesterone regulates GnRH secretion via its hypothalamic receptors, the lack of progesterone receptors (PR) on GnRH neurones2 suggests its action may involve interneurons expressing sex steroid receptors. The most probable upstream afferents are kisspeptin expressing neurones, which are highly enriched with both oestradiol receptor (ER) and PR13, and project directly to the GnRH neurones14, 15. Additionally, knockout of PR in kisspeptin neurones in mice, results in loss of LH surges, irregular oestrous cycles and infertility16, 17.
In rodent, there are two major populations of kisspeptin neurones: those in hypothalamic anteroventral periventricular nucleus (AVPV) which are thought to underlie the LH surge by mediating positive feedback effect of E2, and the arcuate kisspeptin population which coexpress neurokinin B and dynorphin (KNDy) and mediate steroid negative feedback actions on LH secretion18–20. In humans, KNDy neurones of the infundibular nucleus21 (homologue to the ARC in other species) relay both steroid negative and positive feedback actions22. The KNDy neurones are strongly implicated in GnRH pulse generation. Neurokinin B and dynorphin are thought to stimulate and inhibit KNDy neurone activity respectively, driving episodic release of kisspeptin, which in turn drives pulses of GnRH release23. Interestingly, ARC kisspeptin neurones may also participate in the regulation of the LH surge. Ablation of KNDy neurones robustly increased the magnitude of the E2-induced LH surge in ovariectomised rats24, which may be mediated by their projections to the AVPV kisspeptin neurones25 or directly to the GnRH neuronal cell bodies in the medial preoptic area (mPOA)26 and/or terminals in the median eminence25.
Since PR are expressed in virtually all KNDy neurones, they are thought to be critical in mediating progesterone feedback effects on GnRH and LH secretion. Studies in sheep have shown that dynorphin from KNDy neurones may meditate the action of progesterone on GnRH/LH pulsatile frequency27. However, the role of hypothalamic PR in control of the LH surge is controversial. Previous studies have indicated that progesterone might act via its receptors in the mPOA to block the LH surge in rats28. In contrasts, the expression of PR on kisspeptin neurones is required for the LH surge and normal oestrous cyclicity in mice16, 17. Given the afore-mentioned, the role of AVPV and ARC PR in progesterone negative feedback effect on pulsatile LH secretion as well as the LH surge in the rodent remains unclear.
In the present study, we used bilateral microinjections of the potent PR antagonist, RU486, into the AVPV or ARC nuclei to investigate the inhibitory role of progesterone on the LH surge induced by pregnant mare serum gonadotropin (PMSG), which induces follicle development mimicking clinical ovarian stimulation and LH surge induction. Additionally, intra-nuclear administration of RU486 was used to examine the effects of progesterone on pulsatile release of LH and oestrous cyclicity in female Sprague-Dawley rats.
The location of the intra-AVPV and intra-ARC cannulae were confirmed by microscopic histological inspection of cresyl‐violet stained brain sections. Only animals with appropriate bilateral cannulae placement in the AVPV or ARC were included in the analysis (Fig. 1). Of the 16 rats that underwent hypothalamic cannulation, 13 and 12 were confirmed with correct bilateral cannulae in the AVPV or ARC, respectively. The remaining rats were excluded from the analysis due to inaccurate probe placement.
To test whether exposure to progesterone influences ovarian cyclicity, progesterone was administered on metestrus to normal cycling rats, as well as to pregnant mare serum gonadotropin (PMSG) primed rats to mimic an ovarian stimulation background as described in the method below. In the normal cycling rats, vaginal cytology reveals that the progesterone treatment prolongs cycle length (Progesterone vs Vehicle: 6.71±0.67 vs 4.68±0.35 d; n=7 and 5, respectively; P<0.05) (Fig. 2B), with a significant increase in time spent in metestrus but decrease in the diestrus phase (P<0.05) (Fig. 2C). Representative examples of typical oestrous cyclicity in vehicle and progesterone treated animals and shown in Fig. 2A. Similarly, in PMSG-primed animals, progesterone administration lengthened the oestrous cycle (PMSG+Progesterone vs PMSG+Vehicle: 5.57±0.81 vs 4.50±0.71 d; n=7 and 5, respectively; P<0.05) (Fig. 2B), especially the metestrus phase (Fig. 2C). Of note, all vehicle injected PMSG-primed animals were proestrus on day 2 (day 0, PMSG treatment), whereas proestrus was delayed in the progesterone treated rats, and representative examples are illustrated in Fig. 2A.
As expected from the proestrus smear, all vehicle injected PMSG-primed animals displayed a robust LH surge on the afternoon/evening of day 2 (day 0, PMSG treatment) (Fig. 3A,E and F). In contrast, none of the PMSG-primed females administered progesterone displayed an LH surge (Fig. 3B,E and F). Area under curve (AUC) analysis of the LH profile (12:00–20:00h) on day 2 is shown in Fig. 3F.
Circulating progesterone levels were significantly elevated at 18:00h on experimental day 0 post injection (124.62±43.62 vs 22.79±5.64 ng/ml vehicle control; P<0.05) (Fig. 3C). Circulating levels of E2 measured on the afternoon of day 1, before the expected LH surge, as a measure of follicle development, did not differ between PMSG-primed female treated with vehicle or progesterone (224.1±66.1 vs 254.2±76.3pg/ml, respectively; P>0.05) (Fig. 3D).
To determine whether progesterone signalling within the AVPV or ARC may play a role in its inhibitory feedback effects on the LH surge and oestrous cyclicity, the progesterone receptor antagonist, RU486, was micro-infused into these brain areas in progesterone treated PMSG-primed animals described above. Administration of RU486 into the AVPV rescued the PMSG stimulated LH surge on day 2 (day 0, PMSG treatment) in 5 out of 7 animals (Fig. 4C), while no LH surge was evident in controls (n=6) (Fig. 4A and E). In contrast, intra-ARC administration of RU486 failed to rescue the PMSG stimulus LH surge on day 2 (P>0.05; n=6 per group) (Fig. 4B,D and E). A summary of these data with AUC analysis of LH levels is provided in Fig. 4E.
To verify the effects of progesterone on oestrous cyclicity, daily vaginal cytology was examined. Compared with vehicle infusion, bilateral intra-AVPV administration of RU486 significantly decreased cycle length in the presence of progesterone (P<0.05; n=6–7 per group) (Fig. 5A,B and C). However, intra-ARC administration of RU486 showed only a tendency to reduce cycle length compared with vehicle (P>0.05; n=6 per group) (Fig. 5D,E and F).
To investigate the site of the inhibitory action of progesterone on LH pulse frequency, RU486 was bilaterally infused into the AVPV or ARC. Intra-ARC administration of RU486 significantly reduced LH pulse interval compared with controls (P<0.05; n=6–7 per group) (Fig. 6B,D and F). In contrast, intra-AVPV administration of RU486 had no effect on LH pulse interval in the presence of progesterone (P>0.05; n=6 per group) (Fig. 6A,C and E). LH pulse amplitude was not affected in any treatment group (data not shown).
The present study provides direct evidence of a role for PR in the hypothalamic AVPV and ARC nuclei in mediating the inhibitory effects of progesterone on surge and pulsatile release of LH, and oestrous cyclicity in female rats. Intra-AVPV administration of the progesterone antagonist, RU486, attenuated the inhibitory effect of progesterone on LH surges in PMSG-primed female rats, while antagonism of PR in the ARC restored LH pulse frequency in progesterone-treated animals.
Previous studies have shown that an increase in circulatory level of progesterone between oestrous to metestrus is related to prolonged oestrous cycle length in rats29, 30. Nequin L. G. et al.30 reported that higher endogenous serum progesterone level (around 88.5 ng/ml) after ovulation resulted in longer cycle length in rats. In accordance with these data, we confirm that exogenous administration of progesterone (serum level 124.62±43.62 ng) on metestrus prolonged oestrous cycle lengths, particularly the time spent in metestrus stage. PMSG has previously been used to stimulate and synchronize follicular growth and ovulation, mimicking ovarian stimulation in women31, 32. In the present study, progesterone extended oestrous cycle length, regardless of the prevailing oestrogenic milieu stimulated by PMSG. Everett et al.29 reported cycle prolongation by progesterone was accompanied by delayed ovulation, suggesting an antagonistic role of progesterone on oestrogenic stimulation of ovulation. Steroid feedback on LH surge20 and pulse19 generation are mediated by kisspeptin signalling in the AVPV and ARC respectively. Kisspeptin neurone specific ER knockout mice arrest between the oestrous and diestrus phases and are anovulatory33. Similarly, antagonism of GnRH signalling inhibits spontaneous ovulation and oestrous cyclicity34. These data suggest that cycle arrest in metestrus and diestrus is accompanied by interruption of E2 feedback via kisspeptin on GnRH/LH secretion which is essential for ovulation33. Indeed, moderate knockdown of kisspeptin signalling in the AVPV, results in extended cycle length, and increased the time spent in metestrus and/or oestrus20. Of note, the present study shows that progesterone antagonism within the AVPV reversed the prolonged oestrous cycle length induced by progesterone. Additionally, PR knockout in kisspeptin neurones in adult female mice shows abnormal cycle with persistent diestrus and consequently anovulatory16. It therefore appears that progesterone may interrupt oestrogen-kisspeptin-GnRH signalling in the AVPV, thus affecting ovulation and reproductive cycles.
Progesterone can either facilitate or inhibit E2-induced LH surges, depending on their temporal relationship. When administered after E2 priming, progesterone augments and synchronizes the preovulatory LH surge35, whereas high levels of progesterone, as during luteal phase of oestrous/menstrual cycle, block the stimulatory effects of E2 on LH surges in rats2, sheep3, 36 and primates4, including human5. In the present study, our data are consistent with these previous findings and demonstrates the inhibitory effect of progesterone on the GnRH/LH surge under a high E2 milieu induced by PMSG in female rats. Of note, these results are in line with our observation of prolonged cycle after administration of progesterone to mimic the component of the luteal phase of the human menstrual cycle. Indeed, the ability of progesterone to blockade the preovulatory LH surge is the basis of the progesterone primed ovarian stimulation regime widely adopted for infertility patient in IVF clinic. Either endogenous progesterone in the luteal phase6, 37 or exogenous progesterone in follicular phase7 is sufficient to block the LH surge without compromising oocyte competence for women during ovarian stimulation.
Previous studies have shown that progesterone inhibits E2–induced GnRH release38 and decreases GnRH output in response to electrical stimulation of the mPOA39. Although early studies pointed to the hypothalamic preoptic area as a possible site for mediating progesterone inhibition of the LH surge28, there are no PR expressed in mPOA GnRH neurones2. However, kisspeptin neurones are enriched with both PR and ER40, 41, and project directly to the mPOA GnRH neurones14, suggesting they are the target of progesterone. It is generally accepted that E2 acts via ER in the AVPV kisspeptin neurones to activate GnRH neurones resulting in the LH surge40. However, there is yet no direct evidence for the specific role of PR in the AVPV kisspeptin neurones in regulating the LH surge. In the present study, we confirm the blockade of LH surges by progesterone, and further provide evidence for the site specific role of PR in the AVPV in mediating the inhibitory effect of progesterone on LH surges, based on the observation that progesterone antagonism within the AVPV rescues LH surges blocked by progesterone. A role of PR in the ARC seems inhibitory because intra-ARC administration of RU486 failed to rescue the inhibitory effect of progesterone on the LH surge. These data are in keeping with the role of AVPV kisspeptin in mediating the E2-induced LH surge20. Additionally, PR knockout in kisspeptin neurones result in blockade of E2-induced LH surges and lose of oestrous cyclicity in mice16, 17. It therefore appears that progesterone inhibits the pre-ovulatory LH surge by interrupting activation of the AVPV kisspeptin neurones to elicit the GnRH/LH surge. Paradoxically, the ability of E2 to induce the LH surge is dependent on the presence and activation of PR42. While E2 induces PR expression in the AVPV43, progesterone down regulates its own receptors44, 45 and may counteract the induction of PR by E2. Blockade of the LH surge by progesterone may therefore result from PR in the AVPV, thereby attenuating the ability of E2 to induce the LH surge.
The present study indicates that the ARC is a crucial site for progesterone’s inhibitory effect on LH pulse frequency in keeping with the role of this region in LH pulses generation20. Intra-ARC injection of RU486 significantly increased LH pulse frequency in progesterone treated rats. This result supports the findings of Goodman et al.46, where local antagonism of PR in the ARC interfered with the action of progesterone in reducing LH pulse frequency in sheep. As a critical component of the GnRH pulse generator, the ARC KNDy neurones are highly enriched with PR13, and have direct projections to GnRH cell bodies26 and terminals25. It has been reported that dynorphin released by KNDy neurones inhibit pulsatile GnRH secretion and may mediate progesterone negative feedback particularly in ewe27, 47, and a similar mechanism may operate in the rodent. Given the inhibitory role of dynorphin on LH secretion, it is suggested that progesterone may be act on ARC KNDy neurones and through dynorphin release inhibits pulsatile GnRH and LH secretion.
In conclusion, these data provide evidence that the inhibitory effect of progesterone on preovulatory surge and pulsatile secretion of LH is mediated by its receptor in the kisspeptin enriched hypothalamic AVPV and ARC respectively, which are essential for progesterone regulation of ovarian cyclicity.
Female Sprague-Dawley rats weighing 150–200g, obtained from Charles-Rivers (Margate, UK), were housed individually under controlled temperature (22±2°C), light conditions (12:12h light/dark, with lights on at 07:00h) and fed with standard laboratory chow and water ad libitum. All surgical procedures were carried out under ketamine (100mg/kg i.p.; Pharmacia and Upjohn, Crawley, UK) and Rompun (10mg/kg i.p.; Bayer, Leverkusen, Germany) anaesthesia. All procedures were conducted in accordance with the British Home Office Animals Scientific Procedures Animals Act 1986 (Project Licence 70/6237) and all experimental protocols were approved by the Animal Welfare and Ethical Review Body at King’s College London.
Animals were implanted with two custom-made cardiac catheters via the jugular veins to enable serial blood sampling for profiling of LH levels48. The catheters were exteriorised at the back of the head and enclosed within a 30-cm light-weight metal spring tether (Instec Laboratories, Boulder, CO, USA) secured to the slotted screw (Instec Laboratories) affixed to the surface of the skull using dental cement (Dental Filling Ltd., Swindon, UK). The distal end of the tether was attached to a fluid swivel (Instec Laboratories), which allowed the rat freedom to move around the enclosure. After surgery, all rats were housed individually and experiments commenced about 3 days later.
To assess the effect of the anti-progesterone, RU486, in the hypothalamic AVPV or ARC nuclei, animals were secured in a David Kopf stereotaxic frame and implanted with bilateral guide cannulae (33-gauge; Plastics One, Roanoke, VA, USA) targeted towards the AVPV or ARC for later microinjection of drug. The respective coordinates for the ARC and AVPV were 0.40 mm lateral, 0.00 mm posterior to bregma, 8.60 mm below the surface of the dura, and 0.50 mm lateral, 3.40 mm anterior to bregma, 10.10 mm below the surface of the dura according to the rat brain atlas of Paxinos and Watson49. A stainless steel slotted screw (Instec Laboratories) was affixed to the surface of the skull posterior to the guide cannulae and both were secured using dental cement (Dental Filling). The guide cannulae were then fitted with obturators (Plastics One) to maintain patency. After a 7-day recovery period the rats were implanted with cardiac catheters as described above. Experimentation commenced after a further 3-day recovery from surgery.
Treatment with PMSG has been validated for induction of follicle development to mimic ovarian stimulation50, 51 and induction of the LH surge32, and facilitates study of mechanisms underlying these key reproductive processes. To test the action of progesterone on the LH surge female rats receiving a single injection of PMSG (150 IU/kg, i.p., Sigma-Aldrich, Poole, UK) on the morning of metestrus, which results in an LH surge approximately 55h later31, 32, were administered progesterone (5mg/kg, i.p., in peanut oil, Sigma-Aldrich) twice daily (at 09:00 and 17:00h) for two consecutive days starting immediately after the PMSG injection. For controls, PMSG-primed rats were injected with vehicle (0.3ml, peanut oil, i.p., Sigma-Aldrich).
As an additional control group, to evaluate the effect of progesterone per se in normal cycling rats, progesterone or vehicle only was given twice daily for two days starting on metestrus as described above.
Vaginal lavage was performed daily between 07:00 and 08:00h to detect the stages of the oestrous cycle; proestrus, oestrus, metestrus and diestrus. Rats that displayed at least two consecutive 4- to 5-day oestrous cycles, with positive classification for all 4-stages were used. After at least one control oestrous cycle, PMSG-primed (n=12) and non-PMSG (n=10) treated control rats were randomly assigned on the morning of metestrus to receive either progesterone (5mg/kg) or oil vehicle (0.3ml) injections twice daily for 2 days as described above. Oestrous cyclicity was monitored for a further 8–10 days. The oestrous cycle length was calculated as the number of days between successive occurrences of oestrus. The time spent in each cycle stage was calculated as the proportion of days classified in each cycle stage.
On the morning of day 2 (day 0, PMSG treatment) at 11:00h the PMSG-primed animals were attached via one of the two cardiac catheter to a computer-controlled automated blood sampling system, which allows for the intermittent withdrawal of 25-µl blood samples without disturbing the animals48. Once connected, animals were left undisturbed for 1h before sampling commenced at 12:00h, when samples were collected hourly for 8h (till 20:00h) for LH measurement. After removal of each 25-µl blood sample, an equal volume of heparinized saline (50 U/ml heparin sodium/ml normal saline; CP Pharmaceuticals, Wrexham, UK) was automatically infused into the animal to maintain patency of the catheter and blood volume. Blood samples were frozen at −20°C for later assay to determine LH concentrations. Blood samples (25-µl) were also collected at 18:00h on day 0 and day 1 for LH measurement during experiment. Additional blood samples (200µl) were collected at 18:00h on days 0 and 1 for progesterone and E2 measurement, respectively, thus confirming hormone injection and status of follicular development.
A separate group of animals implanted with guide cannulae in the AVPV (n=16) or ARC (n=16) nuclei and cardiac catheters were monitored for normal oestrous cyclicity and then treated with the identical PMSG and progesterone regime described above. To examine the role of AVPV and ARC nuclear PR in surge release of LH, rats received bilateral intra-AVPV or intra-ARC injections of the anti-progesterone, RU486 (4 ng in 0.8μl; Sigma‐Aldrich) or vehicle (0.8μl, 60% artificial cerebrospinal fluid (aCSF), 20% DMSO and 20% ethylene glycol; Sigma‐Aldrich) 1h before each intraperitoneal injected of progesterone (i.p. at 08:00 and 16:00h for 2 days starting on metestrus). Internal cannulae (Plastics One) with extension tubing, preloaded with RU486 or vehicle, were inserted into the guide cannulae and extended 1.0 mm beyond the tips to reach the AVPV or ARC nuclei. The distal end of the extension tubing was connected to a 5‐μl syringe (SGE Analytical Science, Milton Keynes, UK). The tubing was held outside the cage to allowing remote infusion without disturbing the animal. The RU486 or vehicle was micro-infused at a rate of 0.1μl/min for 8min and the internal cannulae were kept in the guide cannulae for a further 2min to prevent backflow. This procedure was repeated for each of the 4 injections.
At 11:00h on day 2 (day 0, PMSG treatment) the animals were attached via a cardiac catheter to the blood sampling system, and left undisturbed for 1h before sampling commenced at 12:00h. Samples were collected hourly for 8h for LH measurement as described above.
Oestrous cyclicity was monitored in this experimental group by examining daily vaginal cytology.
On completion of the experiments described above, the same animals were used to examine the role of AVPV and ARC nuclear PR in pulsatile release of LH, after a washout period of about two weeks during which recovery of normal oestrous cyclicity was observed. At 08:00h on morning of metestrus, rats were given bilateral intra-AVPV or intra-ARC injection of RU486 (4 ng in 0.8μl: Sigma‐Aldrich) or vehicle (0.8μl, 60% artificial cerebrospinal fluid (aCSF), 20% DMSO and 20% ethylene glycol; Sigma‐Aldrich) as described above. One hour later (09:00h), rats received an intraperitoneal injection of progesterone (5mg/kg), and automated blood sampling commenced, as previously described, 3h later (12:00h) with collection of 25-μl blood every 5min for 2h for LH measurement. All the samples were frozen at −20°C for later assay.
After experimentation, 0.5µl India ink was injected through internal cannulae inserted into the guide cannulae for the purpose of site verification. Animal were then killed by decapitation. The brains were removed and snap frozen on dry ice, and then stored at −80°C followed by sectioning (30-µm) using a cryostat (Bright Instrument Co Ltd., Luton, UK). To evaluate the cannulae position, every fourth section throughout the AVPV or ARC region corresponding to bregma −0.60 to 0.80 mm and 2.80 to 3.60 mm49, respectively, was mounted and stained with cresyl violet. Slides were then viewed under a light microscope and images were taken using a digital camera (Zeiss, Oberkochen, Germany). Only data from animals with correct cannulae placement were analysed.
A double-antibody radioimmunoassay (RIA) supplied by the National Institute of Diabetes and Digestive and Kidney Diseases (Bethesda, MD, USA) was used to determine LH concentration in the 25-μlwhole blood sample52. The reference preparation was rLH-RP-3. The sensitivity of the assay was 0.093 ng/ml. The intra-assay coefficient of variation was 6.5% and the inter-assay coefficient of variation was 7.4%. The E2 and progesterone plasma samples (50μl) were measured by using the EIA kit (Labor Diagnostika Nord, Germany) which use was verified in rat53.
The effect of progesterone and RU486 on LH surges was determined by analysing the area under the LH profile curve (AUC) using Sigma Plot 12.0 (Systat Software, Inc., Chicago, IL). Detection of LH pulses was established through the use of the algorithm ULTRA54. Two intraassay coefficients of variation of the assay were used as the reference threshold for the pulse detection. The effect of intra-AVPV and intra-ARC RU486 microinjection on parameters of LH secretion was calculated by comparing the mean basal levels of LH, mean LH pulses interval, or mean amplitude of LH pulses, within the experimental period between the two groups. A Student’s t test was used to analyse baseline LH, E2 and progesterone levels. All other results were analysed by a one-way ANOVA. Data are presented as the mean±SEM and P<0.05 was considered statistically significant.
This work is supported by National Natural Science Foundation of China (Project Number: 81571397), Science and Technology Commission of Shanghai Municipality (Project Number: 15ZR1424900) and Doctoral Innovation Fund Projects from Shanghai Jiao Tong University School of Medicine (Project Number: BXJ201434).
W.H., X.F.L., Q.F.L. and Y.P.K. designed research. W.H., Y.L.L., H.L. and L.W. performed research. W.H., X.F.L. and H.L. analysed the data. W.H. and X.F.L. prepared figures, W.H., D.A. and K.T.O. wrote the paper. All the authors reviewed the manuscript.
The authors declare that they have no competing interests.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Yanping Kuang, Email: moc.621@pnaygnauK.
Kevin T. O’Byrne, Email: firstname.lastname@example.org.