Nootropic dipeptide, Noopept can modulate hyperalgesia by effecting on spinal microglia dependent Brain Derived Neurotropic Factor (BDNF) and pro-BDNF expression and apoptosis during persistent inammation

Background: There are largely unknown associations between changes in pain behavior responses during persistent peripheral inammation and spinal cell alteration such as apoptosis. Some evidence suggests that microglia and microglia related mediators play notable roles in induction and maintenance of central nervous system pathologies and inammatory pain. By considering those relationships and microglia related nootrophic factors, such as the Brain Derived Neurotrophic Factor (BDNF) in CNS, we attempted to assess the relationship between microglia dependent BDNF and its precursor with pain behavior through spinal cell apoptosis as well as the effect of Noopept on this relationship. Methods: Persistent peripheral inammation was induced by a single subcutaneous injection of Complete Freund’s Adjuvant (CFA) on day 0. Thermal hyperalgesia, paw edema, microglial activity, microglia dependent BDNF, pro-BDNF expression, and apoptosis were assessed in different experimental groups by conrmed behavioral and molecular methods on days 0, 7, and 21 of the study. Results: Our ndings revealed hyperalgesia and spinal cell apoptosis signicantly increased during the acute phase of CFA-induced inammation but was then followed by a decrement in the chronic phase of the study. Aligned with these variations in spinal microglial activity, microglia dependent BDNF signicantly increased during the acute phase of CFA-induced inammation. Our results also indicated that daily administration of Noopept (during 21 days of the study) not only caused a signicant decrease in hyperalgesia and microglia dependent BDNF expression but also changed the apoptosis process in relation to microglia activity alteration. Conclusions: It appears that the administration of Noopept can decrease spinal cell apoptosis and hyperalgesia during CFA-induced inammation due to its direct effects on microglial activity and microglia dependent BDNF and pro-BDNF expression.


Background
In the case of persistent pain states, mounting evidence has illustrated that the release of a variety of in ammatory mediators, such as cytokines and growth factors from neuronal and non-neuronal cells, play a crucial role in the induction of in ammatory pain [1]. Although tremendous efforts have been made to understand the cellular and molecular basis of chronic pain, its physiological aspects remains mostly unknown [2,3]. Multiple mechanisms contribute to in ammatory pain, each of which is subject to cellular and molecular alterations of nociceptive neurons in the dorsal horn of the spinal cord. These mechanism also underwrite peripheral in ammation, which is accompanied with in ammatory symptoms such as hyperalgesia and edema. Cell apoptosis, as an important consequence of central nervous system (CNS) pathological conditions, is induced by the increment of in ammatory mediators such as cytokines and growth factors [4,5]. Data from diverse animal models support the idea that apoptosis can induce neuronal sensitization and hyperalgesia [6,7]. It is noteworthy that these irreversible processes might be accompanied by different nervous system damages such as neuropathic pain [8]. In this regard, Kim et al. showed that induction of spinal nerve ligation (SNL) in rats resulted in DRG neuronal apoptosis, which precedes pain behaviors such as mechanical allodynia [9]. Microglia, as resident macrophages within the CNS, are involved in mediating in ammatory processes and can release pro-in ammatory mediators such as TNF-α [10]. Like other cells, microglia can produce not only pro/anti-in ammatory mediators and neurotrophic factors such as Brain Derived Neurotrophic Factor (BDNF), but can also play neurotoxic or neuroprotective roles in CNS depending on the released mediators [11,12]. It has become increasingly clear that microglial cells can powerfully control and modulate pain when they are activated in response to peripheral in ammation [13].
Data generated thus far indicates that BDNF is a neurotrophin with important biologic roles in neuronal survival and differentiation [14]. Also, it has a crucial role in regulating in ammatory pain thresholds and secondary hyperalgesia [15,16]. Previous studies have shown that BDNF synthesis is greatly increased in different populations of DRG neurons in response to peripheral in ammation, and it also has apoptotic effects on neural cells [17,18]. Groth and Aanonsen showed that BDNF induces an acute dose-dependent thermal hyperalgesic response and also affects central sensitization in NMDA receptor activation dependent processes [19]. On the other hand, Siuciak et al. presented that midbrain infusion of BDNF can decrease behavioral paw inch response to subcutaneous formalin injection in both early and late phases of the test [20].
Accumulating evidence supports that Pro-Brain Neurotrophic Factor (pro-BDNF), as a BDNF precursor, binds to the p75 neurotrophin receptor (p75NTR) and exerts the opposing biological functions of mature BDNF. Pro-BDNF plays an active role in regulation of axon pruning, dendritic complexity, and also neuronal apoptosis in the hippocampus. An important piece of the puzzle is missing from the previous studies of BDNF, the microglial activity-dependent pro-BDNF and BDNF relationship with spinal cell apoptosis and pain behaviors during peripheral persistent in ammation [21,22].
The term nootropics (cognition enhancer) is applied to a group of psychoactive substances which stimulate neuronal functions and enhance cognitive performances such as memory [23,24]. N-Phenylacetyl-l-prolylglycine ethyl ester (Noopept), a dipeptide analog of piracetam, is one of these synthesized cognition enhancers [25,26]. There is evidence to suggest that administration of Noopept can signi cantly suppress peripheral in ammatory responses to carrageenan [27]. Moreover, it has been shown that Noopept treatment can prevent accelerating neuronal death, which may be related to its capacity to block the neurotoxic effects of calcium ions and glutamate and speci cally its antiin ammatory effects. As such, it deserves consideration as a therapeutic target [28,29].
In light of the principal role of microglia related mediators in in ammatory pain and cell death induction, the identi cation of neuroprotective and anti-in ammatory roles of Noopept should generate novel therapeutics for the treatment of in ammatory pain. In the context of in ammatory pain, it is also interesting to consider the relationship between spinal cell apoptosis and the pain process. We, therefore, aimed to investigate the relationships between microglia dependent BDNF and pro-BDNF and spinal cell apoptosis and pain behavioral responses during different stages of CFA-induced in ammatory pain. Moreover, in a second step, we focused on the effects of Noopept administration through this pathway.

Laboratory animals and experimental procedures
All experimental procedures were performed on adult male Wistar rats (200-220 g). The animals were kept under standard temperature, humidity and lighting condition (22±2ºC, humidity 60-70%, 12h light/dark cycle) with free access to food and water. All procedures were approved by the Ethics Committee of Shahid Beheshti University of Medical Science (IR.SBMU.MSP.REC.1396.818), and were also in accordance with NIH Guidelines for the Care and Use of Laboratory Animals [5,29]. The rats were divided into ve experimental groups as follows: the (a) CFA group, (b) CFA+saline (vehicle) group, (c) Mineral oil group, (d) CFA+Noopept group, and (e) CFA+minocycline group. According to the study procedure, each group was further divided into 3 subgroups of six animals to receive injections at different time points in the study (days 0, 7, and 21). Similar to our previous studies, a persistent in ammatory pain model was induced by CFA on day zero. The mineral oil group (control group) received a single subcutaneous injection of sterile mineral oil (100 μL) without the CFA on day zero. From the rst day after in ammatory pain induction, all experimental groups received the daily drug interventions. Behavioral tests, detection of expression of spinal proteins, and a TUNEL assay were done on days 0, 7, and 21 of the study.
In ammatory pain induction CFA-induced in ammatory pain was caused by a single subcutaneous injection of (100µL) heat-killed Mycobacterium tuberculosis suspended in sterile mineral oil (10 mg/ml; CFA; Sigma, St Louis, MO, USA) into the rats' hind paw on day zero under light anesthesia with methoxyiso urane. The rst week after CFA injection was considered the in ammatory phase and the third week was the arthritic phase [30,31].

Drug administration
In this study, Minocycline (Sigma-Aldrich, Germany), as a microglia activation inhibitor, was diluted in physiological saline (0.9% NaCl) and injected intraperitoneally (i.p.) daily at a dose of 40 ml/kg from the rst day after CFA injection to the 21 st day of the study in the CFA+minocycline group [32]. Noopept (a nootropic dipeptide) was dissolved in saline solution and injected (i.p.) daily at a dose of 5 mg/kg from the rst day after CFA injection till day 21 in the CFA+Noopept group [27,28,33].

Assessment of paw edema
Paw edema assessment was done by measuring paw volume variations during the different stages of the study using a Plethysmometer (model 7141; Ugo Basile, Comerio-Varese, Italy). Brie y, the rats' hind paws were submerged to the tibiotarsal joint in the electrolyte-lled perspex cell of the plethysmometer. The amount of liquid displaced, which is related to the paw volume, was shown on a digital display.
Volume measurements were done twice for each paw, and the average was calculated. Edema was considered by measuring the differences between day zero and the other time points, days 7 and 21 of the study [34].
Thermal hyperalgesia assessment Paw withdrawal latency (PWL) was used to measure thermal hyperalgesia by applying the Hargreaves test [35]. For this purpose, the rats were placed in Plexiglass chambers (Ugo Basile, Verse, Italy) and habituated for 15 minutes to the test environment. The heat source was adjusted under the plantar surface of the injected and non-injected hind paws and an infrared light was projected focally, and then PWL was measured. Each paw was tested three times at 5 min intervals and the average value of the withdrawal latency of three consecutive tests was calculated and deducted from the other paw. In the present study, the cut-off time in the absence of a response was 20s.

Western blot
After the behavioral tests, western blot was used to examine the amounts of BDNF, pro-BDNF, Iba1, and Caspase-3 expression in the lumbar spinal cord. In this regard, the lumbar part of the spinal cord (L1-L5) was removed and homogenized in a lysis buffer. Protein samples were separated on a sodium dodecyl sulfate-polyacrylamide gel electrophoresis (12% SDS-PAGE) and transferred to an activated Polyvinylidene di uoride (PVDF) membrane (Roche Diagnostics, Indianapolis, lN, USA), then blocked with 5% Bovine serum albumin (BSA) and incubated overnight at 4°C with primary antibodies against BDNF (1:1000, ab205067), pro-BDNF (1:1000, NB100-98754), Iba1 (1:1000, ab178846), and Caspase-3 (1;3000, ab184787). The membrane then was washed three times and incubated for 1h at room temperature with a secondary antibody (1:20000, cell signaling #7074). Finally, the blots were visualized with a chemiluminescence detection system (ECL, Amersham, RPN2235). Band intensity was measured densitometrically using Image J and expressed as the ratio of the intensity of the BDNF, pro-BDNF, Iba1, and Caspase-3 band to β-actin [36].

TUNEL assay
The apoptotic cells in the lumbar section of the spinal cord were detected by TUNEL staining. Brie y, lumbar segments of the spinal cord were embedded in para n after xing in 4% paraformal dehyde for 24h. Then, ve slides from each block were prepared for assessing the apoptotic index by means of the TUNEL technique. After depara nization in xylene, the sections were rehy drated with different dilutions of alcohol and treated with H 2 O 2 and methanol at room temperature. Afterwards, they were incubated in proteinase K. The TUNEL technique was carried out using a TUNEL-based assay kit (SIGMA, Roche, Germany). The sections were incubated rst in horseradish per oxidase (POD) and then in diaminobenzidine (DAB) solution and subsequently counterstained with hematoxylin. Five watch elds were chosen randomly on each section under a microscope (Nikon-E600). Positive brown cells and total cells were counted and analyzed using GraphPad Prism 6 software [37,38].

Statistical analysis
Data were shown as mean±standard error of mean (SEM). The statistical analysis used for the behavioral, cellular, and molecular tests was a one way ANOVA. Statistical comparisons were analyzed by repeated measurements followed by Post hoc Tukey's for evaluation of variants within the groups. An unpaired Student's t test was used to compare the changes of variants on the same days between the two groups. Statistical signi cance was accepted at the p≤0.05 level. Statistical analysis were performed using the GraphPad Prism 6.0 program.

Paw volume variation during different stages of CFA-induced in ammatory pain
There were no signi cant differences in behavioral and molecular parameters between the mineral oil and CFA groups on day zero (data not shown). CFA injection into the rat's hind paw caused a signi cant increase in paw volume, which persisted until the 21st day of the study. Paw volume showed a signi cant increase on the 7th and 21st days after CFA injection compared with day zero in the CFA group (p < 0.001). Our data also indicated that paw volume signi cantly increased on day 21 compared with day 7 after CFA injection in the CFA group (p < 0.001). On the other hand, three weeks administration of Noopept in the CFA + Noopept group signi cantly reduced paw volume on the 7th (p < 0.001) and 21st (p < 0.001) days of the study compared to the same days in the CFA group. Furthermore, our results revealed that paw volume signi cantly decreased in the CFA + Noopept group in comparison with the CFA + Minocycline group on days 7 and 21 of the study (p < 0.001). Saline administration, as a vehicle, caused no signi cant differences in paw volume in comparison with the CFA group at all-time points of the study (Fig. 1).

Thermal hyperalgesia variations during different stages of CFA-induced in ammatory pain
Our results revealed that hyperalgesia noticeably increased on the 7th day after CFA injection compared with day zero (p < 0.001). However, it signi cantly decreased on day 21 in comparison with day 7 (p < 0.01) in the CFA group, although hyperalgesia on day 21 was still more than day zero (p < 0.001). Long-term Noopept administration in the CFA + Noopept group signi cantly reduced thermal hyperalgesia compared with the CFA group on days 7 and 21 (p < 0.001). Moreover, the results illustrated that thermal hyperalgesia signi cantly decreased in the CFA + Noopept group in comparison with CFA + Minocycline group on days 7 and 21 (p < 0.001). The CFA + Saline (vehicle group) group indicated no signi cant differences in paw withdrawal latency in comparison to the CFA group at all times of the study (Fig. 2).
Western blot analysis of spinal BDNF expression during different stages of CFA-induced in ammatory pain Following intraplantar injection of CFA, lumbar spinal BDNF expression demonstrated various changes during the different stages of the study. The results revealed that CFA injection caused signi cant increase in BDNF expression on day 7 in comparison with day zero (p < 0.001). There was a signi cant decrease in BDNF expression on day 21 in comparison with day 7 in the CFA group (p < 0.001). Our ndings indicated that the administration of Noopept in the CFA + Noopept group caused a signi cant decrease in spinal BDNF expression on the 7th (p < 0.001) and 21st (p < 0.001) days of study in comparison with the same day in the CFA group. In addition, in comparison with the CFA group, the decrease in BDNF expression in the the CFA + Noopept group was signi cantly more than what was observed in the CFA + Minocycline group on the 7th (p < 0.001) and 21st (p < 0.001) days of study. The CFA + Saline (vehicle) group indicated no signi cant differences in spinal BDNF expression in comparison to the CFA group at all-time points of the study (Fig. 3).

Assessment of spinal pro-BDNF expression during different stages of CFA-induced in ammatory pain
Western blotting analysis of pro-BDNF expression in the lumbar part of the spinal cord showed a signi cant increase on day 7 in comparison with day zero in the CFA group (p < 0.001). On the other hand, spinal pro-BDNF expression in the CFA group considerably decreased on day 21 compared with days zero (p < 0.05) and 7 (p < 0.001). Our ndings also indicated that the Noopept treatment in the CFA + Noopept group signi cantly decreased pro-BDNF expression on day 7 (p < 0.001 compared with the same day in the CFA group. Based on our results, administration of Noopept in the CFA injected rats was signi cantly more effective in decreasing pro-BDNF expression compared with the CFA + Minocycline and CFA groups during acute phase of the study (p < 0.001). The CFA + Saline group presented no signi cant differences in spinal pro-BDNF expression in comparison with the CFA group at all time points of the study (Fig. 4).

Spinal microglial activation variations during different stages of CFA-induced in ammatory pain
Following intra plantar injection of CFA, expression of Iba1, as a microglial marker at the lumbar spinal cord, revealed a signi cant increase on 7th day of the study compared with day zero (p < 0.001). There was signi cant difference in the Iba1 expression level between day 21 compared with days 0 (p < 0.05) and 7 (p < 0.001) in the CFA group. As shown in Fig. 5, administration of Noopept in the CFA + Noopept group considerably decreased the Iba1 expression level on days 7 (p < 0.001) and 21 (p < 0.05) in comparison with the same days in the CFA group. Furthermore, our ndings showed that the decrease in spinal Iba1/β-actin expression level in the CFA + Noopept group was signi cantly greater than what was observed in the CFA + Minocycline group on 7th day compared with the CFA group (p < 0.001). The CFA + Saline (vehicle group) indicated no signi cant differences in spinal Iba1 expression in comparison with the CFA group at all times of the study (Fig. 5).

Measurement of spinal apoptotic cells during different stages of CFA-induced in ammatory pain
After CFA injection, the apoptosis index signi cantly increased in the super cial laminas of the ipsilateral lumbar segment of the spinal cord on day 7 (p < 0.001) in the CFA group compared with day zero. In addition, There was signi cant decrease in the apoptosis index on day 21 compared with day 0 (p < 0.01). Regarding our ndings, the number of apoptotic cells considerably decreased on the 7th (p < 0.001) and 21st (p < 0.01) days in the CFA + Noopept group compared with the same days in the CFA group. As presented in Fig. 6, There was signi cant decrease in the percentage of apoptotic cells on the 7th (p < 0.05) and 21st (p < 0.05) days in the CFA + Noopept compared with the same days in the CFA + Minocycline group. There was no signi cant difference in the apoptosis index in the CFA + Saline group in comparison with the CFA group at all times of the study (Fig. 6).

Cleaved spinal caspase-3 variation during different phases of CFA-induced in ammatory pain
Our results revealed that CFA injection caused a signi cant increase in cleaved caspase-3 levels on day 7 (p < 0.001) compared with day 0 in the CFA group. There was also signi cant decrease in cleaved caspase-3 levels on 21st day of study compared with day 7 (p < 0.001) in the CFA group. Long-term Noopept treatment in the CFA + Noopept group signi cantly decreased CFA-induced caspase-3 activity on the 7th day of study (p < 0.05) compared with the same days in the CFA group. Based on our results, administration of Noopept in the CFA + Noopept group more signi cantly decreased cleaved caspase-3 levels on both the 7th (p < 0.05) and 21st (p < 0.05) days when compared with the same day in the CFA + Minocycline group. The CFA + Saline group presented no signi cant differences in spinal cleaved caspase-3 levels in comparison with the CFA group in all time points of the study (Fig. 7).

Discussion
The current study was designed to clarify the involvement of microglia dependent BDNF in the alteration of pain behaviors and spinal cell apoptosis. We also examined microglial activity and the spinal level of BDNF in Noopept-treated rats in the context of in ammatory pain and spinal cell apoptosis. Our ndings in the rst step of the study indicated that microglia dependent BDNF and pro-BDNF expressions aligned with thermal hyperalgesia, and spinal dorsal horn cell apoptosis increased in the CFA injected rats on day 7 but decreased on day 21. Thus, inhibition of microglial activity by minocycline caused a decrement of these variations during different phase of the study.
Large bodies of evidence exist to substantiate that increment of pain sensitivity is initially due to neural function variations, and microglia has a pivotal role in the pathogenesis of pain [13,39]. In line with the results from our previous studies, this study shows that CFA-induced in ammation caused persistent spinal microglial activation, which occurred in parallel with hyperalgesia increment, and can be inhibited by Minocycline as a speci c microglia inhibitor [30,40]. It is therefore conceivable that microglia has a vitally important role in the initiation and maintenance of pain, but little attention has been paid to the speci c mechanisms of spinal glia cells and their roles in the induction and development of pain behaviors [41,42].
It is currently well established that, except for certain neurons, activated microglia can release various mediators such as BDNF in response to homeostatic changes [43,44]. Ulmann et al. demonstrated that following peripheral nerve injury in in vivo and in vitro models, the stimulation of purinergic receptor P2X leads to the release of BDNF from activated microglia [4,45]. In this study, we found that microglia dependent BDNF and pro-BDNF expressions in the lumbar part of the spinal cord increased on day 7, but decreased on day 21. These results matched with the behavioral outcomes. It is intriguing to note that inhibition of microglia led to a decrement of spinal BDNF and pro-BDNF expression on day 7; however, its effects on BDNF and its precursor on day 21 was not signi cant. BDNF, a member of the neurotrophin family, is produced following intra and extra cellular processing of its precursor, pro-BDNF [15,46]. Many studies have reported that pro-BDNF is not an inactive biological molecule and has diverse cellular functions [47,48]. Numerous data convincingly support the notion that the effects of BDNF and pro-BDNF are, in fact, opposite [21]. It is widely accepted that BDNF promotes neuronal survival through activation of the TrKB receptor, while pro-BDNF induces cell apoptosis through p75 receptor activation [49]. However, there is a dearth of information about the relationship of microglia dependent BDNF and pro-BDNF on the spinal dorsal horn in pain behaviors during in ammatory pain. BDNF primarily synthesized as a larger precursor, pro-BDNF, which undergoes proteolytic processing by proteases [50,51]. It is noteworthy that the conversion of pro-BDNF to BDNF depends on neuronal activity [52]. Matsumoto et al. have shown that the capacity of a neuron for pro-BDNF processing is limited [53]. In this regard, Yoshimura et al. and Wysokiński have reported that a disturbance in BDNF and pro-BDNF balances may be involved in the pathogenesis of mental disorders [54,55]. Moreover, in vitro studies have indicated that pro-BDNF enhances apoptosis of basal forebrain neurons and sympathetic neurons [56,57].
Apoptosis is a crucial physiological process and any defect in apoptotic mechanisms is closely related to different disorders like pain [58,59]. This study results showed that the number of apoptotic cells that aligned with the spinal cleaved caspase-3 level during the rst week after CFA injection signi cantly increased in the ipsilateral dorsal horn of the spinal cord in comparison to day zero, which shows the changes in apoptosis following CFA injection were also well related to hyperalgesia during various time points of peripheral in ammation [60].
The above nding was consistent with changes in microglia dependent BDNF and pro-BDNF expression during the study. So, it seems that microglial BDNF and pro-BDNF imbalances during different phases of the study, speci cally on day 7, may lead to apoptosis more than neuronal survival. Therefore, a greater understanding of the role of the microglia signaling pathways in spinal apoptosis and their interaction with hyperalgesia variations opens up the intriguing possibility of manipulating microglial cells to eliminate or reduce pain.
Our observations in the second step of the study suggested that the long-term administration of Noopept during persistent peripheral in ammation not only decreases the activity of spinal microglia but also decreases microglia dependent BDNF and pro-BDNF expressions in parallel with thermal hyperalgesia and dorsal horn cell apoptosis. Despite the many studies done on the effects of Nootropics, such as Noopept, on cognitive disorders like Attention-De cit/Hyperactivity Disorder (ADHD), Alzheimer's and Parkinson's diseases [61,62], the mechanism of the effect of this novel dipeptide on persistent in ammatory pain and its involved central mechanisms remained largely uncharacterized. Our results demonstrated that Noopept treatment was not only effective in reducing pain behaviors but also reduced spinal cell apoptosis, which is related to pain behavior variations in the context of CFA-induced in ammation. It is of interest to note that during this study the administration of Noopept not only reduced spinal Iba1 level, as a microglial activity marker, but also reduced microglia dependent BDNF and pro-BDNF expression levels in the lumbar part of the spinal cord. Although there are considerable studies on the anti-in ammatory effect of Noopept [28,61], the current study stated the novel dipeptide antihyperalgesic and anti-in ammatory effects were mediated via inhibition of microglia activity and apoptosis. Ostrovskaya et al. showed that Noopept, through normalizing incretin system parameters, had an anti-apoptotic effect on pancreatic β cells in the context of experimental Diabetes [63], and in another study they indicated neuroprotective effects [64]. Moreover, the effects reported in this paper in response to Noopept were signi cantly more effective than those reported for minocycline administration in the CFA-in amed groups, which indicates that there are other pathways involved in Noopept antiin ammatory effects which need to be investigated. The results of the present experiment and similar reports dealing with experimental models of in ammatory pain suggest that alleviation of pain symptoms and cell apoptosis associated with long-term Noopept treatment may have been mediated more by decreasing the levels of microglia dependent BDNF than pro-BDNF expression. Data obtained from different part of this study did not rule out the involvement of other pathways, speci cally in glia cells.

Conclusion
The results of this study highlight that an alteration of spinal cell apoptosis and thermal hyperalgesia was linked with the ampli cation of microglial activity and increment of microglia dependent levels of BDNF and pro-BDNF. Furthermore, the anti-apoptotic and analgesic effects of Noopept may be mediated via decreasing microglial activity and microglia dependent levels of BDNF and pro-BDNF during CFAinduced persistent in ammatory pain. In our opinion, nootrophic peptides administration, such as Noopept, can be considered as a new and important target for in ammatory pain reduction, although this conclusion requires further in-depth research.

Declarations
Ethics approval and consent to participate All animal procedures were in compliance with the Ethics Committee of Shahid Beheshti University of Medical Science (IR.SBMU.MSP.REC.1396.818), and also NIH Guidelines for the Care and Use of Laboratory Animals.

Consent for publication
Not applicable.

Availability of data and materials
Please contact corresponds for data requests.

Competing interests
The authors declared that they have no competing interests.