Neurodevelopmental Expression Profile of Dimeric and Monomeric Group 1 mGluRs: Relevance to Schizophrenia Pathogenesis and Treatment

Group 1 metabotropic glutamate receptors (mGluR1/mGluR5) play an integral role in neurodevelopment and are implicated in psychiatric disorders, such as schizophrenia. mGluR1 and mGluR5 are expressed as homodimers, which is important for their functionality and pharmacology. We examined the protein expression of dimeric and monomeric mGluR1α and mGluR5 in the prefrontal cortex (PFC) and hippocampus throughout development (juvenile/adolescence/adulthood) and in the perinatal phencyclidine (PCP) model of schizophrenia. Under control conditions, mGluR1α dimer expression increased between juvenile and adolescence (209–328%), while monomeric levels remained consistent. Dimeric mGluR5 was steadily expressed across all time points; monomeric mGluR5 was present in juveniles, dramatically declining at adolescence and adulthood (−97–99%). The mGluR regulators, Homer 1b/c and Norbin, significantly increased with age in the PFC and hippocampus. Perinatal PCP treatment significantly increased juvenile dimeric mGluR5 levels in the PFC and hippocampus (37–50%) but decreased hippocampal mGluR1α (−50–56%). Perinatal PCP treatment also reduced mGluR1α dimer levels in the PFC at adulthood (−31%). These results suggest that Group 1 mGluRs have distinct dimeric and monomeric neurodevelopmental patterns, which may impact their pharmacological profiles at specific ages. Perinatal PCP treatment disrupted the early expression of Group 1 mGluRs which may underlie neurodevelopmental alterations observed in this model.


Results
Group 1 mGluRs (dimeric and monomeric forms) were examined at postnatal days (PN)12, 35 and 96, corresponding to juvenile, adolescent and adult time points, which are key periods for the emergence of psychiatric disorders 15,31 . mGluR1α and mGluR5 monomeric and dimeric proteins were clearly identified in the rat PFC and hippocampus at all time points (Fig. 1A,B, respectively), with the exception of the mGluR5 monomer, which showed relatively low expression at PN35 and PN96 in both regions. mGluR1α and mGluR5 presented as two dimer bands (~270-280 kDa) and a single monomer band (~150 kDa), as previously reported 9,32,33 . The two dimeric bands were quantified as the dimer; these bands have previously been reported using alternative mGluR1α and mGluR5 antibodies, and have been confirmed as specific, using respective knockout mice 33,34 . The sums of the dimer and monomer densitometry values were taken to equal total mGluR1α and mGluR5. mGluR1α total, dimeric and monomeric expression in the PFC and hippocampus. Two-way ANOVAs revealed significant main effects of age on mGluR1α protein expression in the PFC (Total: F 2,28 = 55.086; p < 0.001; Dimer: F 2,29 = 93.889; p < 0.001; Monomer: F 2,26 = 12.104; p < 0.001). Post-hoc analyses in control rats revealed total mGluR1α expression significantly increased from PN12 to PN35 (160%, p < 0.001) with no significant difference between PN35 and PN96. Dimeric mGluR1α expression similarly increased from PN12 to PN35 in control rats (209%, p < 0.001) followed by a modest increase from PN35 to PN96 (26%, p = 0.049). mGluR1α monomer levels showed a general trend toward increased expression at each developmental time point in the control group, yet a significant difference was only observed between PN12 and PN96 (117%, p = 0.044) (Fig. 2). There was a main effect of PCP treatment on dimeric mGluR1α (F 1,29 = 12.107; p = 0.002) and age x treatment interactions on total (F 2,28 = 5.066; p = 0.013) and dimeric mGluR1α protein expression in the PFC (F 2,29 = 4.244; p = 0.024); PCP treated rodents displayed a reduction in total (− 31%, p = 0.018) and dimeric (− 31%, p = 0.006) mGluR1α expression compared to saline treated rats at the PN96 time point only (Fig. 2). There was no effect of PCP treatment on mGluR1 monomer.
p < 0.001) thus was analysed using non-parametric analyses. There was a significant age effect in control rats (X 2 (2) = 11.014, p = 0.004); peak mGluR5 monomer levels occurred at PN12, followed by a dramatic reduction at PN35 (− 97%, U = 0.000, p < 0.001) with levels remaining low at PN96 (Fig. 3). While there were no main effects of PCP treatment on mGluR5 total (F 1,29 = 1.506, p = 0.230) or dimer (F 1,29 = 0.035, p = 0.854), total mGluR5 expression in the PFC was influenced by an age x treatment interaction (F 2,29 = 5.863; p = 0.007) in which perinatal Bars represent mean values. *p < 0.05, **p < 0.01 and ***p < 0.001 indicate statistical significance between saline treated age groups. § p < 0.05 and § § p < 0.01 indicate statistical significance between perinatal PCP and saline treatment at specific age group. Within the respective age groups the saline group is on the left (blue) and PCP treated group on the right (red).

Figure 3. mGluR5 protein expression reduces throughout development and is acutely increased following perinatal PCP treatment in the PFC and hippocampus.
Relative total, dimeric and monomeric protein levels mGluR5 in the PFC (a-c) and hippocampus (d-f) of phencyclidine (PCP) treated rats compared to controls at postnatal days (PN) 12, 35 and 96 (n = 5-6 per treatment/time point). Bars represent mean values. *p < 0.05, **p < 0.01 and ***p < 0.001 indicate statistical significance between saline treated age groups. § p < 0.05 and § § p < 0.01 indicate statistical significance between perinatal PCP and saline treatment at specific age group. Bars represent mean values + SEM. Within the respective age groups the saline group is on the left (blue) and PCP treated group on the right (red).

Discussion
In this study, we report the neurodevelopmental profile of Group 1 mGluRs, including dimeric and monomeric forms, in the PFC and hippocampus. Furthermore, we detailed the effects of perinatal PCP treatment, an established neurodevelopmental model of schizophrenia, on Group 1 mGluR protein expression in these brain regions. Whilst mGluR1α and mGluR5 share similar structural and functional homology 35 , we report here that they display differential monomeric and dimeric neurodevelopmental profiles and expression patterns following perinatal PCP treatment in rats. In view of the recent literature regarding mGluR-targeting compounds for the treatment of various psychiatric and neurodevelopmental disorders 36 , the present results may have implications for the pharmacological profiles of these compounds at specific ages.
During development, total mGluR1α protein and importantly its functional dimeric form, is relatively low in the PFC and hippocampus following birth, with expression peaking at adolescence (hippocampus) or adulthood (PFC). This is largely consistent with the one previous study examining dimeric and monomeric mGluR1α in cortex and hippocampus 10 . mGluR5 dimer levels were relatively abundant and stable throughout the three stages of development examined. Whilst mGluR5 monomers were expressed at the juvenile time point, they dramatically reduced at adolescence and adulthood, to the lowest limits of detection in both the hippocampus and PFC. Further investigation is required to determine if this trend is specific to certain cell types. However, we also report that this abundance of monomeric mGluR5 early in neurodevelopment is also present in the nucleus accumbens region (see Supplementary Fig. S1), which is characterised by a vastly different neuronal population and neuronal network than the PFC 37 and hippocampus 38 suggesting this finding is not brain region specific.
Previous studies have reported a more subtle reduction over time when examining total mGluR5 expression in these regions 7,10,39 , however this is the first time the expression of the mGluR5 monomer has been reported under non-reduced conditions, relative to the dimeric form during neurodevelopment. These dimeric and monomeric neurodevelopmental expression profiles may have implications for the mGluR pharmacological profiles at specific ages. Positive allosteric modulators (PAMs), particularly targeting mGluR5, are being developed for the potential treatment of schizophrenia 36 . mGluR5 PAMs bind within the 7-transmembrane membrane region of mGluR5 and whilst exerting no G-protein activation alone when bound to dimers, PAMs are capable of potentiating glutamatergic signalling when glutamate is bound. In contrast, recent works from the Pin research team have demonstrated that monomeric mGluRs are capable of G-protein activity, similar to that of a full agonist response, in the presence of PAMs and the absence of glutamate 12 . This is important since glutamate-induced neurotoxic effects and seizures have been observed following the use of Group 1 mGluR agonists [40][41][42] . The absence of the monomer at adolescent and adult time points confers a potential advantage in avoiding these agonist-like effects of mGluR5 PAMs. It is not known if the same abundance of monomeric mGluR5 is present in humans or at what neurodevelopmental time point monomeric expression is reduced, however administration of mGluR5 PAMs at a time point when monomeric mGluR5 is highly expressed may have adverse consequences. Further studies will need to explore whether the abundant monomeric expression is capable of reaching the cell surface where it is accessible to endogenous and exogenous targets. We investigated the Group I mGluR regulators Homer1b/c and Norbin, which can regulate trafficking of mGluR5 to the cell surface 26 . Interestingly, these endogenous regulators were in lower abundance at PN12 in both the PFC and hippocampus, at a time when there was high monomer, suggesting possible differences in regulation and trafficking of mGluR5 at the different developmental stages. Perinatal PCP treatment substantially increased total mGluR5 expression in the PFC and hippocampus at the juvenile time period, yet this did not extend to adolescence or adulthood time points. These findings are consistent with postnatal administration of the NMDA receptor antagonist, MK-801, which increased mGluR5 mRNA expression in the cortex and hippocampus, 4 hrs post treatment 43 . Whilst the aforementioned study did not investigate the long-term effect on mGluR5 mRNA or protein levels, Owaczarek et al., reported that perinatal PCP treatment had no long term effect on mGluR5 protein levels in the PFC or hippocampus, when measured at the single adulthood time point, consistent with our findings 17 . Together, these results suggest that perinatal NMDA receptor antagonism may acutely upregulate cortical and hippocampal mGluR5 mRNA and protein expression, but does not affect its long-term expression at adulthood. Despite inducing acute alterations in mGluR5 expression, changes at such a critical neurodevelopmental period may have long-term consequences on excitatory/inhibitory tone. Investigation of mGluR5 KO mice has shown mGluR5 activation critical for driving the normal neurodevelopmental NMDA receptor subunit, NR2B/2A switch 44 . The composition of NR2A/B subunits plays an important role in the calcium dynamics of the NMDA receptor and synaptic plasticity 45,46 . Although, we observed an increase in mGluR5 expression following perinatal PCP treatment, it is possible that alterations in mGluR5 expression and/or activity during the juvenile neurodevelopmental period may consequently alter the development of NMDA receptor subunit composition, which has indeed been extensively reported in this model 14,15,17 .
In contrast to mGluR5, the present study reported that total mGluR1α expression was reduced in the juvenile hippocampus, of perinatal PCP treated rats. mGluR1α and mGluR5 share similar sequence and second messenger coupling homology, although several studies have identified differences in their cellular location, intracellular signalling and functionality [47][48][49] . In addition, we have shown in the present study that they have differential dimeric and monomeric neurodevelopmental profiles, suggesting that the regulation of dimerization may differ. Similar to mGluR5, mGluR1α is co-localised with the NMDA receptor and can modulate its activity 50 . However, in contrast to mGluR5, which is abundantly found throughout the hippocampus and on various cell types, mGluR1α mRNA and protein expression is found primarily in the CA1 region and localises to GABAergic interneurons 49,[51][52][53] . Hippocampal interneurons have previously been reported to be prone to insult and neuronal death, following perinatal PCP treatment 54 and reductions in hippocampal mGluR1α may be a consequence of a reduction in interneuron cell number. The results from the present study largely suggest that perinatal PCP treatment causes acute changes in mGluR1α and mGluR5 at the early stages of perinatal neurodevelopment, which may contribute to altered glutamatergic signalling and neurodevelopmental abnormalities previously reported in this model 14,15,17 .
In summary, we report that Group 1 mGluRs, mGluR1α and mGluR5, exhibit differential temporal and regional neurodevelopmental expression patterns, especially in relation to dimeric and monomeric expression. These findings may have implications for the effectiveness and side effect profile of mGluR PAMs at specific developmental stages, however, pharmacological studies are required to investigate this. Furthermore, we show that perinatal PCP treatment disrupts the normal developmental expression of these proteins in the PFC and hippocampus, particularly at the juvenile period; this may contribute to ongoing developmental glutamatergic alterations that are believed to contribute to the development of schizophrenia.

Materials and Methods
Animals. Pregnant Sprague-Dawley rats were obtained at gestational day 14 from the Animal Resources Centre (WA, Australia). Animals were housed under constant temperature control (20 °C) and 12:12 hour lightdark cycle. Rats were provided food and water ad libitum. Pups were sexed on post-natal day (PN) 7 and subsequently allocated into PCP or saline treatment groups. Dam and littermates were housed together until 3 weeks of age, when they were weaned and housed 2/per cage of the same sex and treatment. Only female rats were used in subsequent protein analyses in this study. This study was carried out in accordance with the Australian Code of Practice for the Care and Use of Animals for Scientific Purposes (8th edition) and was approved by the University of Wollongong Animal Ethics Committee (AE13/03).

Perinatal PCP Treatment.
Female Sprague-Dawley rats were administered PCP (10 mg/kg, s.c.) (Sigma, Castle Hill, Australia) or saline (0.9% NaCl) on PN7, 9 and 11. Rats were subsequently euthanised by carbon dioxide asphyxiation and decapitation on PN12, 35 and 96 (n = 6/group) coinciding with juvenile, adolescent and adult time points as previously described in Du Bois et al., 15,31 . The prefrontal cortex and hippocampus were dissected according to a standard rat brain atlas 55 , snap frozen in liquid nitrogen and stored at − 80 °C. Rat Brain Tissue Preparation. Tissue samples (left and right hemispheres combined) were homogenised in buffer containing 0.1 M Tris-HCl, 2 mM EDTA, 10% glycerol, 2% SDS, 100 mM iodoacetamide, 0.5 mM PMSF, Protease Inhibitor Cocktail (P8340; Sigma, Australia) and Phosphatase Inhibitor Cocktail 2 (Sigma, Australia). Samples were then centrifuged at 10 000 g for 10 minutes at 4 °C, supernatant collected and stored at − 80 °C. Total protein concentration was determined using a DC assay kit as per the manufacturer's instructions (Bio-Rad, Australia).
Statistical Analysis. Two-way Analyses of Variance (ANOVA) were used to determine any effects of age and treatment on relative protein densities. Where significant interactions were found, independent t-tests were used to identify differences between treatment groups at specific ages. One-way ANOVAs were used to determine significance between age groups within the control group only. While all data showed a normal distribution (Kolmogorov-Smirnov p > 0.05), normality testing can be unreliable in small samples 56 . However, given the findings from the KS test coupled with the observation that biological variables tend to be normally distributed 57 , statistical analyses were completed using parametric analyses as per Dean et al. 58 and our previous studies 15 . However, where there were unequal variances between the groups (mGluR5 monomer in the PFC, mGluR5 total, dimer and monomer in the hippocampus), non-parametric analyses were employed. Kruskall-Wallis tests were used to determine the significance between age groups within the control groups. Mann-Whitney U tests were used to determine the effect of PCP treatment at specific ages. Significance was set at an alpha level of p = 0.05.