The PPARα agonist fenofibrate attenuates disruption of dopamine function in a maternal immune activation rat model of schizophrenia

Summary Aims Prenatal maternal immune activation (MIA) is associated with a risk to develop schizophrenia and affects dopamine systems in the ventral tegmental area (VTA), key region in the neurobiology of psychoses. Considering the well‐described sex differences in schizophrenia, we investigated whether sex affects MIA impact on dopamine system and on schizophrenia‐related behavioral phenotype. Furthermore, considering peroxisome proliferator‐activated receptor‐α (PPARα) expression in the CNS as well as its anti‐inflammatory and neuroprotective properties, we tested if PPARα activation by prenatal treatment with a clinically available fibrate (fenofibrate) may mitigate MIA‐related effects. Methods We induced MIA in rat dams with polyriboinosinic‐polyribocytidylic acid (Poly I:C) and assessed prepulse inhibition and dopamine neuron activity in the VTA by means of electrophysiological recordings in male and female preweaned and adult offspring. Results Poly I:C‐treated males displayed prepulse inhibition deficits, reduced number and firing rate of VTA dopamine neurons, and paired‐pulse facilitation of inhibitory and excitatory synapses. Prenatal fenofibrate administration attenuated detrimental effects induced by MIA on both the schizophrenia‐like behavioral phenotype and dopamine transmission in male offspring. Conclusion Our study confirms previous evidence that females are less susceptible to MIA and highlights PPARα as a potential target for treatments in schizophrenia.

infection and has been shown to induce schizophrenia-like phenotypes in rodents. 2 These phenotypes display behavioral abnormalities, including sensorimotor gating impairment as well as alterations in brain regions key in the neuropathology of psychoses, such as dopaminergic ventral tegmental area (VTA) or prefrontal cortex. While the pathophysiology of schizophrenia has not been conclusively determined, it is believed that dopamine transmission dysregulation importantly alters information processing in multiple domains and results in the global symptoms observed in schizophrenia. 3 Neuroimaging studies in humans indicate that dopamine signaling is altered in both the mesocortico-limbic 4,5 and nigrostriatal pathway.
Hence, dysregulation of dopamine transmission in the striatum, especially in the rostral caudate, is considered an additional hallmark of schizophrenia. 3,7 Schizophrenia is a neurodevelopmental disorder that displays robust epidemiological sex differences, with males and females exhibiting different symptoms, disease prevalence, and treatment response. Indeed, it is well established that females are less susceptible to autism and to early onset schizophrenia than males. 8 Nevertheless, most of the studies investigating the Poly I:C model have been limited to male offspring and findings in females are still sparse. 9,10 Several preclinical studies indicate that the schizophrenia-related behaviors of adult offspring after MIA involve interactions between inflammatory events and prenatal brain development. 12 Thus, pharmacological treatments with anti-inflammatory properties during pregnancy might be effective to prevent or minimize the impact of MIA in offspring. Here, we hypothesize that, among different potential targets, the peroxisome proliferator-activated receptors alpha (PPARα) might possess a therapeutic potential.
PPARα are members of a family of nuclear receptor transcription factors that have been shown to play essential roles in diseases associated with inflammatory processes. 13 Growing evidence suggests that gene expression modulated by PPARα might mitigate the inflammatory component that occurs in psychiatric diseases, such as depression 14 and schizophrenia. 15 Additionally, an association with the gene encoding for PPARα (PPARA) was found in patients with schizophrenia 16 and the expression of these nuclear receptor genes was also downregulated in hair-follicle cells from schizophrenia patients. 17 On these bases, the first aim of our study was to investigate the impact of MIA on schizophrenia-related behavior and dopamine transmission in rat female offspring versus males both at preweaning age and during young adulthood. In order to test the hypothesis of sex differences in the outcome of MIA, we designed this investigation as a follow-up of our previous study, whereby, consistent with previous literature, we found that male offspring from Poly I:C-treated dams show marked deficits in prepulse inhibition of startle reflex (PPI) and in dopamine transmission, such as higher baseline levels of dopamine in the nucleus accumbens (NAc) and reduced number and firing rate of spontaneously active VTA dopamine neurons. 18 Although nigrostriatal dopamine transmission might be compromised in schizophrenia, in our study we focused on the VTA subregion as the vast majority of studies on animal models of schizophrenia rely on the examination of the mesolimbic dopamine pathway, and chose to analyze PPI, a behavior that is dependent on mesolimbic dopamine, as an index of psychotic-like behavioral abnormalities.
The second aim of our study was to investigate whether fenofibrate, a PPARα agonist clinically used as lipid-lowering medication, is able, when administered during pregnancy, to mitigate the effects of MIA in the offspring.
F I G U R E 1 MIA impacts on VTA dopamine neuron activity in male but not female offspring. (A) Diagram representing the experimental protocol. Dams were fed with standard or 0.2% fenofibrate-enriched diet from gestational day (GD) 8 to GD 18. At GD 15, a single iv injection of Poly I:C (4 mg/kg) or vehicle (sterile pyrogen-free saline) was administered. Offspring underwent different experiments according to their postnatal age: patch-clamp recordings were carried out between postnatal day (PND) 12-20, prepulse inhibition of sensorimotor gating was tested at PND 60-70, and in vivo electrophysiology experiments were performed at PND 70-90. (B) Representative localization of recording sites of VTA putative dopamine neurons in vehicle (white dots for males and white squares for females) and Poly I:C (black dots for males and black squares for females) offspring, as verified by histological sections. RN, red nucleus; IP, interpeduncular nucleus, SN, substantia nigra pars reticulata. (C) Poly I:C offspring showed a reduced number of spontaneously active dopamine neurons. (D) Two-way ANOVA yielded a significant interaction between sex and treatment for firing rate of dopamine cells. Hence, when compared with controls, the mean firing rate of VTA dopamine cells was decreased in males, as showed by the scatter plot. Burst duration (F) and mean spikes per burst (G) are impaired by MIA only in male offspring (two-way ANOVA showed a significant interaction between sex and treatment). No significant difference was found for the percentage of spikes in burst (E) between sex or treatment. Superimposed colored diamonds show the averages for each individual rat. (H, I, J) Comparison of VTA dopamine cells from Poly I:C and controls in male and female offspring during preweaning age ex vivo. The graph shows individual firing rates of VTA dopamine neurons (H) recorded from male and female offspring from vehicle-treated and Poly I:C-treated dams. Each circle represents the mean firing rate of a 3-min recording. Representative traces of action potential are displayed on top (calibration bars: 50 ms and 100 pA). No changes in paired-pulse ratio (EPSC2/ EPSC1) of AMPA EPSCs are observed between Poly I:C and vehicle female offspring (I), whereas a paired-pulse facilitation is found in Poly I:C male rats as compared to controls. Representative traces of recordings from vehicle (top) and Poly I:C (bottom) rats are shown above graphs (calibration bars: 50 ms and 100 pA). Similarly, a paired-pulse facilitation (IPSC2/IPSC1) has been observed in GABAA IPSCs of Poly I:C male offspring compared to vehicle rats, but not in females (J). Representative GABAA IPSCs from VTA dopamine neurons recorded in vehicle (top) and Poly I:C (bottom) rats are shown above graphs. N values are indicated in Table 1. Horizontal black lines represent means. Statistical analysis was conducted with two-way ANOVA (sex and treatment as factors, see Table 1) and Sidak's multiple comparison test. Asterisks on graphs represent the results of the Sidak's multiple comparison test: *P < 0.05, **P < 0.01, ***P < 0.001  food and water available ad libitum. We made all efforts to minimize animal discomfort and to reduce the number of animals used.

| Prenatal treatment
Female Sprague Dawley rats (Envigo, Italy) were mated at the age of 3 months. The first day after the copulation was defined as gestational day 1 (GD 1). Dams were randomly assigned to two experimental groups: The first group received a control diet for the whole pregnancy, whereas the second group received a diet enriched with the PPARα agonist fenofibrate (0.2% w/w) ad libitum from GD 8 to GD 18. MIA was induced at GD 15, following the procedure described by Zuckerman et al. 2 Dams were anesthetized with isoflurane 2% and a single dose of Poly I:C (4.0 mg/ kg, iv) (InvivoGen, San Diego, CA, USA) or an equivalent volume of endotoxin-free saline solution was administered in the lateral vein of the tail (see Figure 1A for a schematic representation of protocol Subsequently, rats were randomly assigned to the experimental procedures and care was taken to avoid assigning more than three animals from the same litter to the same experimental group. In fact, for the experiments described here a total of 33 dams were utilized (15 were treated with vehicle and 18 with Poly I:C).
Each rat underwent only one experimental procedure, with the exception of those tested for prepulse inhibition, which were subsequently (following a recovery of at least 10 days) assigned to in vivo electrophysiology experiments.
Prepulse inhibition experiments were carried out in rats at postnatal day (PND) 60 to 70, whereas in vivo electrophysiology experiments were carried out at PND 70-90. This age window, which corresponds to the late adolescence or young adulthood in humans, was selected as it is the most vulnerable age for the onset of schizophrenia. 8 Moreover, studies on the ontogeny of MIA-induced deficits showed that these are evident at PND 70. 19,20 Ex vivo electrophysiological experiments were carried out before weaning (PND [12][13][14][15][16][17][18][19][20] to the aim of detect early abnormalities in dopamine neuron activity and synaptic properties.

| Ex vivo and in vivo electrophysiological experiments
Whole-cell patch-clamp recordings from rat VTA dopamine cells were as described previously. 21,22 Extracellular single-unit recordings from VTA DA neurons in anesthetized rats were as described previously. 18,21,23 See Method S1 for more details.

| Prepulse inhibition (PPI) of startle reflex
Startle and PPI were performed as previously described by Frau et al 24 with slight modifications. See Method S1 for more details.

| Statistical analysis
Statistical analysis is described in detail as Method S1.
This weight loss indicates that Poly I:C treatment induced a flu-like syndrome in treated rats. However, Poly I:C treatment did not affect litter size (controls: 11.3 ± 1.2 pups, n = 13; Poly I:C: 12.0 ± 0.9 pups, n = 16, P = 0.64, Student's t test). Poly I:C treatment did not significantly affect abortion rate when compared with controls, as two dams from the control group (2/15) were excluded from the study as they interrupted their pregnancy, whereas two from the Poly I:C group (2/18) were excluded as offspring were not growing normally. Offspring were randomly selected for the experimental procedures, taking care that no more than three animals (one male and two females or two males and one female) from the same litter were assigned to the same experimental group or procedure (see Tables 1 and 2 for the number of litters used for each experimental group and procedure).
We previously showed that MIA disrupts behavior and affects dopamine transmission in male offspring. 18 To investigate whether these effects are sex-specific, we carried out, both in female and male offspring, in vivo and ex vivo electrophysiological recordings from VTA dopamine neurons and analysis of sensorimotor gating functions. The comparison between male and female offspring was carried out only in the group that received the control diet.

| MIA disrupts activity of VTA dopamine cells in male but not female offspring: in vivo electrophysiology
We first determined whether estrous cycle influenced spontaneous dopamine neuron activity in female animals. Controls and Poly I:C female rats underwent vaginal smears before electrophysiological experiments and were staged according to their estrous cycle in estrus, proestrus, metestrus, and diestrus ( Figure S1). The latter two groups were pooled in diestrus. By computing all electrophysiological parameter analyzed (cells/track, firing rate, burst firing) of dopamine cells according to the estrous stage (estrus, diestrus, and proestrus) and treatment (controls vs. Poly I:C), statistical analysis did not reveal any significant difference among estrous stages (Table   S1). Therefore, data from all female rats were pooled independently of their estrous stage. Ns, not significant; nd, not determined. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 (two-way ANOVA); TA B L E 1 This table shows the effect of vehicle or Poly I:C administration during pregnancy in male and female offspring. Statistical analysis was conducted with two-way ANOVA (sex and Poly I:C treatment as factors) or Student's t test offspring, and n = 12 vehicle-treated and n = 21 Poly I:C-treated female offspring.
As several neurons were recorded from individual rats and then grouped considering each cell as an independent replicate, a twoway ANCOVA was carried out with sex and treatment as factors and individual subjects as covariate, to exclude that differences among individual rats had significant effects. The results indicated that individual subjects had no significant effect overall (effects of subjects analyzed with two-way ANCOVA, firing rate P = 0.84; percentage of spikes in bursts P = 0.9; mean burst duration P = 0.65; number of spikes in bursts P = 0.66). Significance levels yielded by two-way ANCOVA for factors or interaction between factors were the same as those obtained with two-way ANOVA as indicated in Table 1.

| Effect of MIA on VTA dopamine cells ex vivo in male and female offspring
Several studies have reported that Poly I:C administration during pregnancy leads to psychotic-like behavioral abnormalities at adulthood without affecting rat behavior at preweaning age. 20,25,26 Considered that, our next aim was to assess whether VTA dopamine  Figure 1H), and an interaction between sex and treatment (F (1, 45) = 7.12, P < 0.05). Post hoc analysis (Sidak's multiple comparisons test) disclosed that male offspring displayed a marked reduction in the mean spontaneous firing rate of VTA dopamine cells recorded from rats exposed to Poly I:C, when compared with control males (Figure 1H Table 1). Properties of GABA synapses (IPSCs2/IPSCs1 ratio) differed between treatments (F (1, 31) = 17.29; P < 0.01) but not between sexes, and the interaction between main factors was significant (F (1, 31) = 4.18, P < 0.05) ( Figure 1J).
Post hoc analysis showed that properties of GABA synapses were different in male Poly I:C rats as compared to control rats. In fact, we observed a paired-pulse facilitation in dopamine cells recorded from Poly I:C male rats. On the other hand, IPSCs2/IPSCs1 ratio was similar in female Poly I:C rats when compared with control females ( Figure 1J).
Increases in the paired-pulse ratio in males are predictive of either a reduced probability of neurotransmitter release, 28 or a reduced function of postsynaptic receptors, or a combination of these.
Therefore, we examined spontaneous miniature AMPA EPSCs (ie, mEPSCs) and GABA A IPSCs (ie, mIPSCs) to detect changes in AMPA and GABA A receptor function, number, or both. In Poly I:C rats, we found a reduction in frequency of both mEPSCs (t (8) = 6.95, P < 0.0001, Student's t test; Table 1) and mIPSCs (t (10) = 2.83, P < 0.001, Student's t test; Table 1). Furthermore, no significant difference has been observed in the amplitude of mEPSCs (t (8) =0.04, P > 0.05, Student's t test; Table 1) and mIPSCs (t (10) = 0.08, P > 0.05, Student's t test; Table 1). Altogether, the paired-pulse protocol and the decreased frequency of spontaneous miniature events indicate a reduced probability of glutamate and GABA release in the VTA of Poly I:C rats. In fact, a reduction in frequency but not in amplitude is considered to reflect a presynaptic decreased probability of transmitter release.

| Effect of MIA on prepulse inhibition of startle reflex in male and female offspring
PPI provides an operational measure of sensorimotor gating, a neurological process that filters irrelevant from salient information. PPI deficits are found in schizophrenia patients, and in rodents are induced by the administration of dopaminergic agonists and reversed by benchmark antipsychotics. 29,30 Therefore, we investigated whether MIA affected PPI in female and male offspring.

| The PPARα agonist fenofibrate protects male offspring from MIA-induced alterations
As several detrimental effects induced by MIA were only evident in males, we evaluated the potential efficacy of a pharmacological treatment with a PPARα agonist to prevent neurodevelopmental aberrations solely in male rats. By measuring daily food intake, the oral intake of the PPARα agonist fenofibrate by dams was quantified in a subset of experimental subjects, resulting 119 ± 11 mg/kg per day (range 64-196, n = 13).
The timeline of the experiments, as well as the age of the animals, was exactly as stated in the section above and illustrated in  Figure 3A,B, Table 2) and for the firing rate of recorded dopamine cells (F (1,429) = 6.286, P < 0.05; Figure 3C, Table 2). Post hoc analysis  Table 2.
In light of these findings, we next investigated whether the pro- test showed that only Poly I:C rats of mothers fed with a control diet displayed an enhanced EPSCs2/EPSCs1 ratio and, therefore, a paired-pulse facilitation (Table 2, Figure 3H). On the other hand, a main effect of fenofibrate (F (1, 36) = 4.721, P < 0.05), but not of interaction or Poly I:C, was detected for GABA A -mediated IPSCs ( Figure 3I).
Finally, we investigated whether fenofibrate during pregnancy prevents PPI disruption in males. Two-way ANOVA analysis showed an interaction between Poly I:C and fenofibrate treatments (F (1,82) = 4.36, P < 0.05, Table 2, Figure 3J). Post hoc test (Sidak's multiple comparisons test) revealed that Poly I:C had a significant effect to decrease PPI only in the offspring from rats fed with control diet ( Table 2, Figure 3J), whereas no difference was found between offspring from fenofibrate-treated animals, indicating that fenofibrate was able to prevent PPI deficits induced by Poly I:C exposure ( Figure 3J).

| D ISCUSS I ON
Our findings indicate that MIA, evoked by maternal exposure to Poly I:C, induces detrimental effects in offspring that are more severe in males than females, and that activation of the nuclear receptor transcription factors PPARα might represent a pharmacological strategy to prevent such an outcome.
In line with gender differences in the prevalence of many neuropsychiatric disorders, such as early onset schizophrenia 8 and autism, 31 we found that dopaminergic system of female offspring was less vulnerable to MIA, when compared to male littermates.
Specifically, male littermates showed a reduction in spontaneous activity of VTA dopamine cells and a paired-pulse facilitation of inhibitory and excitatory synapses onto dopamine cells. Effect of MIA on dopamine system was consistently demonstrated by previous studies. In particular, Vuillermot et al 19 provided evidence that exposure F I G U R E 2 Effect of sex and Poly I:C treatment on startle reflex parameters. In both sexes, startle amplitude (A), latency to peak (B), startle habituation (C), and PPI (D) were strongly affected by sex (two-way ANOVA, see Table 1), but no significant effect of MIA was detected as a main factor, as evoked by exposure of dams to Poly I:C, nor interaction between sex and treatment. N values are indicated in Table 1. Horizontal black lines represent means. Statistical analysis was conducted with two-way ANOVA (sex and treatment as factors, see Table 1). Prepulses are indicated by the intensity corresponding to decibels above background noise. AU, arbitrary units; % IBR, percent inter-block ratio to Poly I:C during early/middle gestation in mice leads to a complex pattern of age-dependent structural abnormalities in mesoaccumbal and nigrostriatal dopamine systems. Thus, MIA might compromise the integrity of the developing fetal dopamine system, suggesting that prenatal exposure to immunological insults induces deficits in dopaminergic development. 19 In fact, at the time of gestational Poly I:C exposure (GD 14 to GD 17), an inflammatory cytokine production might represent a window of higher vulnerability for the dopamine system, as studies have detected the maximal rise in dopamine neurons in the developing mesencephalon, that decreases again soon after (GD17 to GD 21). 32 Interestingly, the density of TH-positive fibers in projection regions such as the striatum presents a strong sex dimorphism, being higher in female rats. 33 Additionally, male fetuses have been shown to possess lower mesencephalic TH immunoreactivity, 34  In our recent report, 18 we showed that MIA ensuing Poly I:C exposure during gestation disrupts PPI in male adult rats. Here, when analyzing sex differences in the vulnerability of PPI response to MIA, we did not find an interaction between prenatal  Translational studies might be facilitated by the fact that strategies to achieve PPARα activation, beside fibrates, include supplements of specific dietary lipids that might be safer during pregnancy. The timing for optimal treatment with PPARα agonists is yet to be established and needs to be pondered considering safety issues of medications or of dietary lipids and the potential risks for the offspring, especially in vulnerable individuals.

ACK N OWLED G M ENTS
We thank Dr. Fabrizio Sanna, Dr. Pierluigi Saba, Marta Tuveri, and Dr. Barbara Tuveri for their skillful assistance. We gratefully acknowledge Sardinia Regional Government for the financial support

CO N FLI C T O F I NTE R E S T
The authors declare no conflict of interest.