Embryonic and adult synaptic proteome perturbations after maternal immune activation: Identification of persistent changes relevant for early intervention

Maternal infections during pregnancy pose an increased risk for neurodevelopmental psychiatric disorders (NPDs) in the offspring. Here, we examined age- and sex-dependent dynamic changes of the hippocampal synaptic proteome after maternal immune activation (MIA) in embryonic and adult mice. Adult male and female MIA offspring exhibited social deficits and sex-specific depression-like behaviours, among others, validating the model. Furthermore, we observed dose-, age-, and sex-dependent synaptic proteome differences. Analysis of the embryonic synaptic proteome implicates sphingolipid and ketoacid metabolism pathway disruptions during neurodevelopment for NPD-pertinent sequelae. In the embryonic hippocampus, prenatal immune activation also led to changes in neuronal guidance, glycosphingolipid metabolism important for signalling and myelination, and post-translational modification of proteins that regulate intercellular interaction and developmental timing. In adulthood, the observed changes in synaptoneurosomes revealed a dynamic shift toward transmembrane trafficking, intracellular signalling cascades, and hormone-mediated metabolism. Importantly, 68 of the proteins with differential abundance in the embryonic brains of MIA offspring were also altered in adulthood, 75% of which retained their directionality. These proteins are involved in synaptic organisation, neurotransmitter receptor regulation, and the vesicle cycle. A cluster of persistently upregulated proteins, including AKT3, PAK1/3, PPP3CA, formed a functional network enriched in the embryonic brain that is involved in cellular responses to environmental stimuli. To infer a link between the overlapping protein alterations and cognitive and psychiatric traits, we probed human phenome-wise association study data for cognitive and psychiatric phenotypes and all, but PORCN were significantly associated with the investigated domains. Our data provide insights into the dynamic effects of an early prenatal immune activation on developing and mature hippocampi and highlights targets for early intervention in individuals exposed to such immune challenges.


50
Recent insights from large-scale systematic studies identify psychiatric disorders as one of the leading 51 causes of the global burden of disease, and their prevalence as not only age-but also sex-and gender-52 specific 1,2 . Since growing evidence about the pleiotropic nature of mental illness and the major role of 53 disrupted brain development therein emerged 3-7 , neuropsychiatric developmental disorders (NPD) have 54 increasingly occupied a more central role in psychiatric research. A viral infection during pregnancy is an 55 established risk factor for NPDs, including schizophrenia (SCZ), autism spectrum disorder (ASD), and 56 attention-deficit/hyperactivity disorder (ADHD), in the offspring [8][9][10][11] . Neuroinflammation, especially 57 during gestation, has gained more prominence in recent years as developmental and social deficits have 58 been observed in infants after a maternal infection during pregnancy 12,13 . Animal and cell models of

92
Most recent investigations have utilised transcriptomic methods to explore gene expression changes 93 after a prenatal immunostimulation 49 , but sex-specific findings have remained inconsistent and sporadic.

94
However, given the prominence of sex differences in mental disorders, this issue is of high relevance. So 95 far, available studies have examined either hypothesis-derived genes and pathways [50][51][52][53][54][55] or cell-or 96 region-specific RNA transcripts as described above. Importantly, RNA and protein levels often do not 97 correlate 56 and the brain, specifically, exhibits a large number of distinct protein expression profiles not 98 detectable at the RNA level, including for neurotransmission-relevant vesicle transport genes 57 . To the best of our knowledge, MIA-evoked changes at the functional proteomic level have not been assessed 100 at separate ages in a sex-specific manner 58 .

101
To address these knowledge gaps, we investigated and compared the proteomic changes in embryonic 102 (gestational day E18) and adult hippocampal synaptoneurosomes of MIA mouse offspring of both sexes, 103 using mass spectrometry. MIA was induced in wild-type pregnant dams (E9) using the viral mimetic 104 polyinosinic:polycytidylic acid [poly(I:C)] injection at two different doses during early gestation to mimic 105 an immune challenge in the late first trimester in humans 59,60 . Moreover, the intervention encompasses 106 a developmental window, during which increased forebrain neurogenesis, initial neuronal migration, and 107 early immune cell colonisation of the brain occur 10,14,46,61 . The model was validated via NPD-relevant 108 behavioural endophenotypes in adult male and female offspring, which then allowed direct comparison 109 of MIA-induced hippocampal synaptic proteome alterations in a sex-and age-dependent manner.

110
Importantly, we aimed to identify persistent proteome changes that could be relevant targeted-111 intervention strategies for individuals upon early diagnosis or at high risk after MIA.

127
MIA was induced on E9 of the putative pregnancy according to Meyer et al., 2005 63 Table 1). More detailed information on mating, immune challenge procedure, and

166
The resulting detected protein lists were then cleaned by removing histones and keratins considered 167 impurities (n=20 in adult, 29 in embryonic data set), and used for all subsequent analyses. All data sets

192
Proteomic results were analysed for differential abundance using an unpaired t-test and multiple-testing

210
Adult MIA mice display deficits in NPD-related behavioural paradigms

211
To confirm the validity of our prenatal immune activation model, we generated mouse offspring of both

222
No differences were observed in the open-field test, which was used to assess general locomotor activity 223 (Supplementary Figure S3A).

224
Additionally, the low and high viral mimetic dosages led to divergent spatial novelty preference 225 outcomes, increasing preference in the high-versus low-dose MIA male offspring ( Figure 2D).

226
Interestingly, the high-dose treatment group lead to social novelty preference deficits ( Figure 2F) in both 227 sexes, in the absence of significant social interaction differences ( Figure 2E). Furthermore, depressive-228 like behaviours in high-dose offspring manifested as self-neglect in females, as assessed in the nest 229 building test ( Figure 2G), and anhedonia in males ( Figure 2H), assessed in the sucrose preference test.

230
No significant behavioural deficits were observed in the FST (Supplementary Figure S3E).

231
The ASR was differentially influenced by poly(I:C) dosage in females, so that higher dose MIA-exposed 232 offspring showed an increased startle response (Supplementary Figure S3D

238
Taken together, these findings substantiate the model in our laboratory and further support its relevance 239 for NPD-associated behaviours, while revealing sex-and dose-specific effects.

308
To further investigate these common proteins, we performed a hierarchical cluster analysis ( Figure 4D), 309 which confirmed the similarity among the embryonic samples but a clear separation from those 310 observed in the adult HPC. In adult MIA offspring, the pool appears to be most strongly driven by protein 311 changes in the females, while the embryonic pool resembles more closely the male proteomic profile 312 ( Figure 4D). The interrogated proteins formed three major clusters, namely mostly down-or upregulated 313 at both ages, and those overwhelmingly decreased in embryos but increased in adults. Most probed 314 proteins were upregulated at both time-points.

315
Subsequently, we interrogated protein-protein-interactions for the common protein dataset in a publicly

325
Finally, we aimed to investigate the extent of association between the revealed MIA-induced, long-term 326 synaptic changes with NPDs and their comorbidities. Therefore, we analysed PheWAS data for the genes 327 encoding the 68 proteins that overlapped between the embryonic and adult proteomes. We determined 328 that all but one of these proteins (porcupine, PORCN) were significantly associated with psychiatric or 329 cognitive traits (Supplementary Table S8

DISCUSSION
In the current study, we examined and directly compared dynamic and persistent proteomic changes in 347 embryonic and adult gestationally immune-challenged offspring of both sexes. Here, we demonstrate 348 sex-specific effects in adult, but not embryonic, SN HPC proteome of MIA-exposed mice. Importantly, a 349 subset of 68 proteins which were consistently and age-independently affected by MIA in mature and 350 prenatal HPC SN were associated with NPD-related phenotypes, and environmental stimuli. These

401
Detection of a larger number of proteins for the embryonic dataset (n=6 435 vs 3 511 for adults) 402 corresponds to higher amounts of tissue in pooled samples as well as of analysed peptides, the immature 403 state of the synapse influencing the SN composition and isolation, or marginally higher sensitivity of 404 mass spectrometry equipment used. To explore the comparability of data further, we randomly selected 405 five proteins only detected in adults (5.1% of detected protein set), for which or whose paralogs data 406 was available in the Allen Brain Atlas for the developing mouse brain 94 . We compared the expression 407 profiles between the late embryonic (E15.5-18.5) and postnatal (P14-28) brain. For the five probed 408 proteins (GRIN2A, KCNJ6, LRRTM1/2, MOBP, SYT1/10) corresponding gene expression in the 409 telencephalic vesicle, giving rise to the HPC, was significantly higher after birth. This suggests that the 410 differences in the numbers of identified proteins might be rooted in physiological, dynamic changes in 411 expression during brain tissue development and maturation.

412
The proteomic alterations observed in the present study are significantly associated with psychiatric 413 disorders and contain a cluster of upregulated proteins that build an enriched functional network.

466
In conclusion, this investigation of the HPC synaptic proteome provides insights into the dynamic effects 467 of an early prenatal immune challenge on the developing and mature brain. The observed dynamic 468 alterations at the synapse were age-specific and varied by sex or immunostimulant dose in the adult and 469 embryonic brain, respectively. Simultaneously, we discovered a smaller but highly relevant subset of 470 embryonic SN changes caused by MIA that persist into adulthood, which could represent targeted 471 approaches in high-risk or early diagnosis individuals. The provided evidence infers a causal link between 472 altered synaptic proteins, neurodevelopmental pathways, and related psychiatric endophenotypes.

473
Supplementary information is available at MP's website.

DATA AVAILABILITY
The mass spectrometry proteomics data have been deposited to the ProteomeXChange Consortium via