TLR-7 and 8 recognize viral single stranded RNA. Together, TLRs activate transcription factors nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) and activator protein 1 (AP-1) via a TIR-domain-containing adapter-inducing interferon-β (TRIF)-dependent pathway to induce release of pro-inflammatory cytokines such as interleukins IL-1b, IL-6, IL-8 and tumor necrosis factor TNF-a. The TRIF dependent pathway also leads to apoptosis and release of anti-viral interferons by interferon regulatory factor (IRF)-3 activation via a mitogen-activated protein kinase (MAPK)-dependent pathway. Rig-1 recognize 5’ triphosphate RNA in the cytosol. Also in the cytosol, NOD-like receptor family member NOD-, LRR- and pyrin domain-containing 3 (NLRP3) is a multiprotein inflammasome complex which is recruited by the condition of a few factors during infection with the human influenza virus. Clearly, influenza virus triggers an innate immune response at many sites. Other models of MIA rely on PAMPs that have a more precise target. PolyI:C, a synthetic double-stranded RNA composed of inosine and cytidine, either organized randomly or as inosine strands hybridized with cytidine strands, is recognized by TLR-3. Lipopolysaccharides, a bacterial endotoxin, is recognized by membrane receptors TLR-4 of granulocytes, B-lymphocytes and monocytes and engages TRIF-dependent and myeloid differentiation primary response gene 88 (MyD88)-dependent pathways to activate NF-kB and AP-1, also leading to the release of inflammatory cytokines. Injecting immunogens such as poly(I:C) or LPS allow more precision than injecting influenza virus, with immune activation lasting 1 to 2 days only (Cunningham et al., 2007; Urs Meyer, 2014). These MIA model all seem to rely on the secretion of IL-6 since IL-6 KO models are protected from prenatal immune activation. This agrees with evidence showing human patients diagnosed with schizophrenia and autism have altered blood levels of cytokine, especially pro-inflammatory IL-6 (Masi et al., 2015; Miller et al., 2011).IL-6 activates regulatory helper T cells 17 (Th17) which maintain mucosal barriers. In disease, Th17 have been linked to depression. Here, they have been recognized as an important downstream element of the MIA paradigm. IL-17, a major cytokine secreted by th17. Testing for ASD-relevant behaviours – ultrasonic vocalization, sociability and repetitive behaviour – Choi et al. determined that genetically knocking out retinoic acid receptor–related orphan nuclear receptor gamma t (RORγt)-dependent effector T lymphocytes such as Th17 or blocking the effector cytokine IL-17a with antibody ameliorated behaviours in MIA mice prenatally treated with poly(I:C) on E12.5. Injecting any of IL-6 or IL-17, is enough to induce MIA-induced phenotypes (Choi et al., 2016; Estes et al., 2016; S. E. P. Smith et al., 2007). Enhancing anti-inflammatory IL-10 cytokine expression genetically can normalize behavioural abnormalities in MIA model mice in PPI, LI, OF, and glutamatergic locomotion tests (U. Meyer et al., 2008). Despite understanding that proinflammatory cytokines must play a role in the etiology of MIA-induced neuropsychiatric conditions, little is understood about the mechanism by which these effects are manifested.
Yet, there exist many potential pathways. For example, immunogen injections performed E9-E10 coincide with the conversion of neuroepithelial cells into radial glial cells and the onset of neurogenesis (Deverman et al., 2009; Hatakeyama, 2004). It is also known cytokines can have many effects on the developing brain. For example, IL-1B can inhibit cell proliferation and neurogenesis (Crampton et al., 2012). In the literature, MIA models for ASD activate maternal immune system later, while schizophrenia models favor E9.5 (Malkova et al., 2012). E9.5 coincides with formation of the neural tube in mice and is also when microglia start colonizing the brain (Semple et al., 2013). Schizophrenia related symptoms such as sensorimotor gating deficits are generally observed when prenatal injections of immunogens are performed around E9.5 or E12.5 whereas E16 and E17 injections manifest increased seizure susceptibility, reduced social skills, and cognitive deficits, symptoms more closely related to ASD. (Ozawa et al., 2006; Pineda et al., 2013; Reisinger et al., 2015).
While both immune and stress models of gestational environment produce elevated corticosterone levels and both IL-6 and corticosterone can cross the placenta and blood-brain barrier, double hit experiments where both MIA and juvenile stress were performed did not demonstrate interaction (Babri et al., 2014; Brummelte et al., 2010; Dahlgren et al., 2006; Threlkeld et al., 2010; N. Yee et al., 2011). Yet, other fruitful studies make use of double-hit models, pairing the MIA model with genetic models of SZ or ASD such as DISC1 alpha 7 nicotinic acetylcholine receptor (Abazyan et al., 2010; Careaga et al., 2017; Lipina et al., 2013; Wu et al., 2015) and NRG1 (which encodes Neurgulin-1) (Leary et al., 2014), NR4A2 (for Nurr1) and TSC2 which all synergize to create greater phenotypes than independently. (Estes et al., 2016)
MIA Translational power
The maternal immune activation has been broadly studied using methods implicating diverse dose sizes, strains, and developmental ages (Babri et al., 2014; Urs Meyer, 2014). Yet, there is no consensus on a standard of methodology for this model and it is therefore difficult to compare different results. It is important to consider the validity of this model (Urs Meyer et al., 2007). Validity is evaluated in three parts. Construct validity requires a model to involve a known risk factor of the disease being modeled. Secondly, a model with good face validity will induce phenotypes comparable to known signs and symptoms of the disease. Finally, predictive validity refers to the ability of the model to predict treatment outcomes from model to human disease. It is estimated by testing if known treatment methods for the disease are also efficient in the model (Careaga et al., 2017).
The maternal immune activation model of neuropsychiatric disorders has construct validity because of the evidence supporting correlation between schizophrenia and autism cases and immune activation during gestational period (Urs Meyer et al., 2012; Reisinger et al., 2015). Though the model has construct validity, it is a single-risk model and can therefore not be expected to recreate the whole disease. Another issue with animal models of neuropsychiatric disorder arise from the fact that these disorders, as far as they are genetic disorders, do not arise from one gene mutation but a various combinations of numerous gene mutations and how they interact in an environment of human cells and human proteins (McCarroll et al., 2013).
Face validity is also present because of the numerous behavioural and structural deficits present in the MIA model which are relevant to neuropsychiatric disorders (Urs Meyer et al., 2007). A challenge in creating valid animal models is the heterogeneity of neuropsychiatric disorders. Two individuals with the same diagnosis may have little symptoms in common (for DSM diagnosis, they need only express a specific number of symptoms from a list). Many of these symptoms cannot be measured objectively (ex. Hallucinations, delusions, specific emotions) which are subjective experiences that can only be communicated by language). Abnormal social behavior, motivation, working memory, emotion, and executive function can be measured in animal models but the extent of their own translational power must be kept in mind. (Nestler et al., 2013)