Emerging evidences show that some hormone such as osteocalcin (OST) and the neurotrophins Brain-derived Neurotrophic Factor (BDNF) and Nerve Growth Factor (NGF) as well as oxytocin (Oxt) share common actions suggesting a coordinated regulation (Yano and Chao, 2000; Karsenty, 2011, 2012; Oury et al., 2011, 2013; Camerino et al., 2012; Karsenty and Oury, 2014; Brennan-Speranza and Conigrave, 2015).

The actions of osteocalcin, BDNF and NFG and Oxt are related to the regulation of reproduction, energy and bone homeostasis (Camerino, 2009a,b; Karsenty, 2011; Camerino et al., 2012). Specifically, osteocalcin exists as fully carboxylated or partially carboxylated forms and completely uncarboxylated. The carboxylated osteocalcin has an exclusive effect on bone matrix remodeling stimulating osteoblast formation (Ferron et al., 2010; Karsenty, 2011). The uncarboxylated osteocalcin acts as an hormone and promotes insulin secretion and sensitivity (Fulzele et al., 2010; Karsenty, 2011; Oury et al., 2011). In fact, mice lacking osteocalcin (Ocn)^−/− display decreased beta-cell proliferation, glucose intolerance, insulin resistance and increased fat pad (Lee et al., 2007). Uncarboxylated osteocalcin acts on the Gprc6a receptor that activates the AMPc/CREB pathway, which induces the expression of several proteins including testosterone. The Gprc6a gene is expressed in skeletal muscle, heart, lung, spleen, kidney, liver, fat, testis and pancreatic beta-cells but it seems absent in the brain. Gprc6a is not expressed in the ovary, indicating that the action of osteocalcin on reproductive maturation is gender dependent. The uncarboxylated form of osteocalcin crosses the blood brain barrier (BBB) binds to neurons of the brainstem, midbrain, and hippocampus, enhances the synthesis of monoamine neurotransmitters, inhibits GABA synthesis, prevents anxiety and depression, and favors learning and memory independently of its metabolic functions in mice (Ferron et al., 2010; Oury et al., 2013).

Some of the actions of osteocalcin cannot be easily explained on the basis of the current data. For instance, the neuronal actions of osteocalcin were observed in the absence of expression of the Gprc6a gene suggesting that some alternative pathways may play a role in mediating the osteocalcin action in neurons (Oury et al., 2013).

Furthermore, in (Ocn)^−/− mice lacking osteocalcin gene, a reduced level of testosterone production was reported in Leydig cells while the circulating levels of LH, the major regulator of testosterone production, were increased 2.5-fold (Yadav et al., 2009; Karsenty, 2011; Oury et al., 2011; Ratto et al., 2012). These data suggest that a compensatory unknown mechanism is functioning in these mice.

The neurotrophins NGF and BDNF besides their well-known classical role in neurogenesis and in synaptic plasticity (Yano and Chao, 2000), are implicated in energy, reproduction and bone metabolism in mice (Rios et al., 2001; Yamashiro et al., 2001; Yao et al., 2002; Camerino et al., 2012). NGF has been reported to play a pivotal role in reproduction inducing for instance Leydig cell maturation (Müller et al., 2006; Ratto et al., 2012). NGF is a potent stimulator of LH secretion, has a dose dependent effect on ovulation and acts via a systemic pathway at physiologically relevant doses. NGF is the ovulation inducing factor (OIF) in seminal plasma; by eliciting LH release, OIF triggers trkA up-regulation and neurite development confirming the NGF-like properties of OIF (Ratto et al., 2012). This peculiar action of NGF in regulating the LH levels may be helpful in those conditions associated with lack of regulatory LH release mechanism as in the case of (Ocn)^−/− mice.

Therefore, it may be possible that NGF and BDNF molecules and osteocalcin share common pathways in these tissues, leading to cross talk between different ligand-receptor pathways.

To investigate on the potential relationships between ligands and their receptors, we analyzed by RT-PCR, in the same plate of reaction, the mRNA levels of Ngf and associated receptors Ngfr (nerve growth factor receptor) and Ntrk1 (neurotrophic tyrosine kinase, receptor, type 1) genes, of Bdnf and associated receptor Ntrk2 (neurotrophic tyrosine kinase, receptor, type 2) genes, of Bglap (osteocalcin), osteocalcin receptor Gprc6a, oxytocin Oxt and Oxtr genes in brain, bone, fat and reproductive organs of 3 months old mice of both genders. In particular, the gene-relationship hypothesis was tested and validated using linear correlation analysis. This statistical approach leads us to compare the gene expression profiles of different tissues in male and female mice. Other methodologies based on protein content evaluation such as ELISA or Western blot do not allow simultaneous multiple comparisons between proteins derived from different tissues in the same plate of reaction.

The transcript levels of all the four factors and their relative receptors were measured in different tissues of male and female mice using a unique plate of reaction.

Using bone data as positive control for osteocalcin gene and its receptor Gprc6a, we found that osteocalcin gene was markedly expressed also in brain and less in WAT and BAT and reproductive organs, in both male and female mice (Figure 1). A large variance ratio was calculated indicating differences within groups and between groups as determined by one way analysis of variance for Bglap (F = 11.04 female; F = 29.22 male), Gprc6a (F = 5.1 female; F = 13.49 male) and Ggcx (F = 21.1 female; F = 18.49 male). The rank order of tissue expression of the osteocalcin gene in both male and female mice was bone ≥ brain > WAT = BAT ≥ reproductive organs. In agreement with previous reports (Oury et al., 2011, 2013), in our experiments the osteocalcin receptor Gprc6a was expressed in testis. An extremely low expression of the receptor gene was observed in the ovaries, uterus, and fat tissues in agreement with other reports (Karsenty, 2011, 2012; Karsenty and Oury, 2014; Figure 1).

We additionally investigated the expression levels of the gamma-glutamyl carboxylase (Ggcx) gene involved in the decarboxylation of several proteins including osteocalcin. We found a significant expression of this gene in fat tissues, brain and testis. A large variance ratio was calculated indicating differences within groups and between groups as determined by one way analysis of variance for carboxylated osteocalcin (F = 66 female; F = 24.67 male) and uncarboxylated osteocalcin (F = 62.16 female; F = 251 male). The rank order of tissue expression of Ggcx gene in male mice was: testis > BAT > brain > bone = WAT; in female was BAT ≥ WAT > brain > bone > reproductive organs (Figure 1). Furthermore, ELISA assay showed a significant level of the carboxylated osteocalcin in bone of both male and female mice, according to the role of this type of protein in bone homeostasis, while the uncarboxylated form predominates in all other tissues (Figure 2). The uncarboxylated protein crosses the BBB and acts on its receptor, according to the role of this protein as a peripheral signal released from bone and acting on other tissues (Oury et al., 2013). Despite the high expression of osteocalcin gene in brain (Figure 1), the level of the carboxylated osteocalcin is low, as demonstrated by ELISA experiments (Figure 2). According to previous reports (Karsenty, 2011; Oury et al., 2011; Karsenty and Oury, 2014), a low level of this protein was found in the uterus and ovaries (Figure 2).

We found that the mRNA levels of Ngf and Ngfr genes were markedly elevated in BAT other than in brain in either female and male mice (Figure 3). The brain data were used as positive controls for Ngf and Bdnf gene expression. A large variance ratio was calculated indicating differences within groups and between groups as determined by one way analysis of variance for Ngf (F = 41.31 female; F = 29.24 male), Ngfr (F = 7.6 female; F = 15.02 male), and Ntrk1 (F = 630.5 female; F = 23.18 male). The rank order of expression of Ngf genes was comparable in either gender and was: BAT > brain > WAT > reproductive tissues > bone. The non-specific receptor gene Ntrk1 was more expressed in brain than in BAT and WAT showing a low expression in bone in both genders (Figure 3). These findings are consistent with our hypothesis of NGF as a regulator of energy, playing a minor role in bone.

The Ngfr (p75NTR) receptor gene was significantly expressed in the reproductive tissues in female and male mice, in this context the expression of Ngfr (p75NTR) receptor gene in the testis was 300% higher than in brain (Figure 3).

Conversely, the expression levels of the Bdnf gene were significantly higher in bone than brain (Figure 4). A large variance ratio was calculated indicating differences within groups and between groups as determined by one way analysis of variance for Bdnf (F = 8.9 female; F = 20.4 male), and Ntrk2 (F = 12.74 female; F = 95.44 male). The rank order of expression of the Bdnf gene in either gender was indeed: bone > brain >> WAT and BAT. The expression profile of the receptor gene was inversely related to that of the Bdnf gene in the same tissues, a result that is expected for a very specific ligand/receptor pathway (Figure 4). The rank order of expression of Ntrk2 gene in either gender was indeed brain > BAT = WAT >> bone. These findings showed that the BDNF pathway plays a major role in bone, playing a minor role in the reproductive tissues.

Thus, it appears that the neurotrophins have distinct peripheral roles: the NGF pathway seems to regulate mostly energy balance and reproduction, while BDNF pathway is important for bone homeostasis.

We next investigated the expression levels of Oxt and its receptor in the same tissues (Figure A1). We found that the Oxt gene was markedly expressed in brain and less expressed in bone; Oxtr was also expressed in fat tissues and bone in both genders, although at a lower level. In the reproductive tissues, Oxtr was markedly expressed in ovaries. The ovaries data were considered as a positive control of Oxt and Oxtr gene expression. A large variance ratio was calculated indicating differences within groups and between groups as determined by one way analysis of variance for Oxt (F = 42.2 female; F = 14.62 male), and Oxtr (F = 48.67 female; F = 22.6 male).

The potential relationship between osteocalcin Bglap, Ngf, Bdnf, and Oxt genes and their respective gene receptors Gprc6a, Ngfr/Ntrk1, Ntrk2, Oxtr was investigated by a linear correlation analysis focusing on gene expression data collected from different tissues. The gene expression data obtained in each tissue were fitted using a linear regression equation (see Materials and Methods). An high level of correlation was calculated in brain, BAT and in testis for NGF, BDNF, and osteocalcin (Table 1). The Oxt/Oxtr gene expression data were responsible for the loss of correlation or low correlation calculated in the tissues other than brain and BAT. These findings suggest a tissues specific correlation between the genes expressing ligands and their receptors in the investigated tissues.

In the present work we compared the expression profile of the Ngf/Bdnf/Oxt-Osteocalcin genes and their receptors in different tissues of 3 months old male and female mice in the same plate of reaction.

We found a significant linear correlation between the expression levels of the hormone genes Bglap-NGF-BDNF and their receptors Gprc6a-Ngfr/Ntrk1-Ntrk2 in brain and BAT in either male and female. It is therefore expected that changes in the expression levels of osteocalcin, Ngf and Bdnf genes may mutually affect the expression levels of the others in brain and BAT. Taking into account also the effects of osteocalcin in male fertility, osteocalcin may act in a “male specific estrogen-like” manner.

The Oxt/Oxtr genes in our experiments were, however responsible for the loss of correlation in different tissues.

The relationship between neurotrophins Ngf or Bdnf genes and osteocalcin gene found in our experiments may help to explain the actions of osteocalcin in those brain areas lacking the osteocalcin receptor Gprc6a. It is likely that neurotrophins can mediate most of the actions of osteocalcin in various brain areas, transducing the peripheral signaling represented by the uncarboxylated osteocalcin that crosses the BBB into neuronal signaling (Figure 5).

In this context, we propose that NGF may have a distinct compensatory role in regulating reproduction and energy balance of BAT and increasing LH levels in the absence of osteocalcin, thereby explaining the enhanced LH levels observed in the (Ocn)^−/− mice (Yadav et al., 2009; Karsenty, 2011; Oury et al., 2011; Ratto et al., 2012). Interestingly the Ngfr (p75NTR) receptor gene is expressed in testis at levels greater than brain. This matches the expression of Gprc6a genes. In our experiments the NGF genes were indeed highly expressed in the reproductive tissues. NGF has been reported to plays a pivotal role in reproduction inducing for instance Leydig cell maturation (Müller et al., 2006; Ratto et al., 2012). NGF is a potent stimulator of LH secretion, has a dose dependent effect on ovulation and acts via a systemic pathway at physiological relevant doses. NGF is the OIF in seminal plasma; by eliciting LH release, OIF triggers TrkA up-regulation and neurite development confirming the NGF-like properties of OIF (Ratto et al., 2012).

These observations imply that these hormones may share common pathways in tissues, leading to cross talk between different ligand-receptor pathways. The existence of alternative pathways for the hormonal actions has been described for estrogen and androgen that act through genomic and non-genomic pathways (Falkenstein et al., 2000; Camerino, 2012; Karsenty and Oury, 2012; Prossnitz and Barton, 2014). The biological advantage of this mechanism is that the lack of effects of a specific gene following pathophysiological conditions or gene downregulation can be counterbalanced by the action of other genes (Karsenty and Oury, 2012).

Ngf and Bdnf gene expression appears to show an opposite trend, with NGF regulating reproduction and energy in both genders and BDNF regulating bone homeostasis and energy (Figure 5). The Bdnf and Ntrk2 genes are indeed not expressed in the reproductive organs in our experiments, in contrast to Ngf, Ngfr, Ntrk1 genes, and Oxtr gene which are expressed in these tissues (Dissen et al., 2009; Melmed et al., 2011; Ratto et al., 2012). The Bdnf and Ntrk2 genes are instead highly expressed in bone. In agreement with this conclusion, previous work demonstrated that brain-targeted BDNF conditional knockout mice (Bdnf(2lox/2lox)/93), in which the Bdnf gene has been specifically deleted in brain, show high bone mass and are a metabolic phenocopy of the leptin deficient ob/ob mice, but independent of adrenaline and serotonin pathway (Ducy et al., 2000; Yadav et al., 2009; Camerino et al., 2012).

In regard to the expression of the osteocalcin receptor Gprc6a gene, our data are in line with other reports showing elevated levels of the mRNA of this gene in testis and extremely low levels in ovaries (Oury et al., 2011; Karsenty, 2012). However, the investigation of the role of Gprc6a in regulation of bone mass accrual has generated contradicting findings (Pi et al., 2005; Wellendorph et al., 2009; Karsenty, 2012). In line with these findings, our data reported extremely low levels of the related mRNA in bone.

The Ngfr (p75NTR) receptor gene was significantly expressed in the reproductive tissues in female and male mice suggesting that the gonads receive NGF signaling released from CNS through NGF receptors p75NTR/NTRK1. BDNF signaling may also contributes to this cascade of reaction through p75NTR albeit with a low affinity (Sandhya et al., 2013).

Therefore, NGF may act as a physiologic mediator of osteocalcin both centrally through the sprouting of new synapses, and in energy regulation, and peripherally probably regulating steroid production in Leydig cells in the testosterone deficient Ost^−/− mice (Oury et al., 2013).

Should be underlined, that we measured mRNA levels and no functional proteins were studied. The mRNA levels in our work were extracted from whole organs with no evaluation of cell specific expression of the investigated genes. However, here we investigated on the potential correlation between genes by using a multiple comparison of mRNA levels of various tissues collected from female and male mice in the same plate of reaction. In addition, the high sensitivity of the RT-PCR technique in comparison to other methodology allows the determination of very small quantity of mRNA in the tissues in which some genes of interest may show extremely low level of expression while protein may be undetectable in the same sample.

In the present work, we focused our attention on young adult mice to keep sample homogeneous but in the future work it can be interesting to evaluate the gene-relationship hypothesis even during development. Future experiments will further confirm our hypothesis in animal models.

CC, elaborated the hypothesis and theory; CC and DT designed the studies; EC, MC, and RC conducted the experiments; CC and EC analyzed the results; CC and DT wrote the manuscript and all authors approved the manuscript.