Maternal infection is a risk factor for neurodevelopmental disorders. Mouse models of maternal immune activation (MIA) may elucidate infection’s impact on brain development and function. Here, general guidelines and a procedure are provided to produce reliably resilient and susceptible offspring exposed to MIA.
Maternal immune activation (MIA) during pregnancy is consistently linked to increased risk of neurodevelopmental and neuropsychiatric disorders in offspring. Animal models of MIA are used to test causality, investigate mechanisms, and develop diagnostics and treatments for these disorders. Despite their widespread use, many MIA models suffer from a lack of reproducibility and almost all ignore two important aspects of this risk factor: (i) many offspring are resilient to MIA, and (ii) susceptible offspring can exhibit distinct combinations of phenotypes. To increase reproducibility and model both susceptibility and resilience to MIA, the baseline immunoreactivity (BIR) of female mice before pregnancy is used to predict which pregnancies will result in either resilient offspring or offspring with defined behavioral and molecular abnormalities after exposure to MIA. Here, a detailed method of inducing MIA via intraperitoneal (i.p.) injection of the double stranded RNA (dsRNA) viral mimic poly(I:C) at 12.5 days of gestation is provided. This method induces an acute inflammatory response in the dam, which results in perturbations in brain development in mice that map onto similarly impacted domains in human psychiatric and neurodevelopmental disorders (NDDs).
Epidemiological evidence links maternal infection to increased risk of psychiatric and NDDs, including schizophrenia (SZ) and autism spectrum disorder (ASD)1,2,3,4,5,6,7. The MIA mouse model was developed to test causality and the mechanistic role of MIA in the etiology of these disorders, as well as to identify molecular biomarkers and develop both diagnostic and therapeutic tools4,6. Despite the utility of this model and its increasing popularity, there is considerable variability in MIA induction protocols within the field, making it difficult to compare results across studies and replicate findings8,9. In addition, most iterations of the model do not investigate two important translational aspects of MIA: (i) many offspring are resilient to MIA, and (ii) susceptible offspring can exhibit distinct combinations of phenotypes8.
To generate a reproducible MIA model, investigators should report at least one quantitative measure of the magnitude of MIA induced in dams. To induce MIA during gestation, our lab performs intraperitoneal (i.p.) injections of the double stranded RNA viral mimic polyinositic: polycytidilic acid [poly(I:C)]. Poly(I:C) induces an immune cascade similar to influenza viruses as it is recognized by toll-like receptor 3 (TLR3)10. As a result, poly(I:C) activates the acute phase response that results in rapid elevation of proinflammatory cytokines8,11,12. Previous studies have demonstrated that the elevation of proinflammatory cytokines, including interleukin-6 (IL-6), is necessary to produce behavioral abnormalities and neuropathology in offspring as a result of MIA11,12,13. Thus, the level of IL-6 in maternal serum collected during its peak at 2.5 h following poly(I:C) injection is a compelling quantitative measure of MIA that can be used to compare results across laboratories within the field.
In order to generate an MIA model that addresses the translationally essential elements of resilience and susceptibility with a single induction protocol8,14, researchers can combine typical induction approaches with characterization of the dam's baseline immunoreactivity (BIR) before pregnancy8. Recently, it was discovered that virgin female C57BL/6 mice show a wide range of IL-6 responses to a low-dose exposure to poly(I:C) before pregnancy8. It is only a subset of these females that go on to produce susceptible offspring, and only at certain magnitudes of immune activation as dictated by the combination of BIR and poly(I:C) dose8. MIA induces phenotypes in an inverted U pattern; offspring show the greatest behavioral and molecular aberrations when dams are moderately immunoreactive, and the magnitude of maternal inflammation reaches, but does not exceed, a critical range8. Here, a detailed method of how to reliably create both resilient and susceptible offspring with divergent behavioral phenotypes as a result of mid-gestational injection of poly(I:C) is provided.
All protocols are performed under the approval of the University of California-Davis Institutional Animal Care and Use Committee (IACUC).
1. Animal preparation
2. Poly(I:C) lot testing and preparation
3. Baseline immunoreactivity (BIR) testing
NOTE: Figure 1 shows the schematic of the steps. Use a different molecular weight poly(I:C) for BIR testing as compared to gestational to lower the likelihood of adaptive immune responses to the compound.
4. Tail bleed method for blood collection
NOTE: To avoid use of potentially immunomodulatory sedatives, use the tail bleed method of blood collection.
Figure 1. The timeline for testing virgin females' baseline immunoreactivity and mating. Order mice to arrive at 7 weeks old and allow to acclimate to facility for 1 week. Inject animals with 5 mg/kg of poly(I:C) and 2.5 h later draw blood. Allow blood to clot overnight, then centrifuge at 3,768 x g, 4 °C for 8 min. Collect serum and assess relative IL-6 levels via ELISA or Multiplex. At 9 weeks old, set up mating pairs. Created using BioRender.com Please click here to view a larger version of this figure.
5. Weight based method for mating and gestational E12.5 injection
NOTE: Figure 2 shows the schematic of the steps. Two methods can be used to set up mating pairs and determine the E12.5 time point. The first, timed-mating, is described elsewhere22. Weight-based calculations can also be used to assess for an E12.5 pregnancy23. The benefit of this approach is that it allows time-locking of the dam's age at mating, decreasing variability in the immune response. This procedure is used here.
Figure 2. MIA induction. MIA induction requires assessment of pregnancy, i.p. injection of poly(I:C), and litter checks to ensure correct timing of maternal inflammation. After assessing gestational day either via timed mating or the weight-gain method, deliver an i.p. injection of poly(I:C) at E12.5. Collect a blood sample at 2.5 h after injection to confirm immune activation and determine level of IL-6 activation. Litters will be born at approximately E18.5-E20.5. Created using BioRender.com Please click here to view a larger version of this figure.
6. Investigation of alterations in behavior in adult MIA and control offspring (optional)
Not all animals exposed to 30 mg/kg of poly(I:C) at E12.5 produce offspring with consistent behavioral abnormalities8,31. Though both 30 mg/kg and 40 mg/kg of poly(I:C) reliably produce sickness behaviors in dams, including decreased activity levels, hypothermic responses, and weight loss, and also cause significant elevations in IL-6, only a subset of litters exposed to MIA will go on to develop behavioral abnormalities in domains similar to those observed in human psychiatric and NDDs (Figure 3A–C)8. A lower dose of 20 mg/kg of poly(I:C) also induces sickness behavior and weight loss, but in contrast to higher doses it does not consistently produce IL-6 responses high enough in magnitude to induce behavioral aberrations in offspring even though the IL-6 responses are elevated well above those from dams injected with saline (Figure 3D)8.
Figure 3. Different doses of poly(I:C) lead to differential effects in dams. (A) Dams exposed to 20 mg/kg, 30 mg/kg, or 40 mg/kg of poly(I:C) experienced decreased activity in a subjective scale (one-way ANOVA; P < 0.0001). (B) Dams exposed only to 30 mg/kg of poly(I:C) showed significantly altered temperature in the form of a hypothermic response (one-way ANOVA; F3,35 = 4.289, P < 0.05). (C) Both 30 mg/kg of poly(I:C) and 40 mg/kg of poly(I:C) induced significant weight loss (one-way ANOVA; F7,187 = 26.93, P < 0.0001) and (D) showed elevated IL-6 levels above the threshold required to induce behavioral alterations (one-way ANOVA; F3,35 = 25.54, P < 0.0001). (E) Baseline immunoreactivity in isogenic female C57BL/6J animals is highly variable, and categorizing females BIR into low, medium, and high groups allows researchers to predict which offspring are most likely to be susceptible to the impact of MIA. Bars represent mean ± SEM. This figure has been modified from Estes et al.8. Please click here to view a larger version of this figure.
Unexpectedly, virgin female C57BL/6 mice exhibit quite variable baseline immunoreactivity (BIR) to a low dose of poly(I:C) (5 mg/kg of poly(I:C)) before pregnancy even though they are isogenic, and this variability is not associated with weight (Figure 3E, Supplementary Figure 1)8. Dams injected with 5 mg/kg of poly(I:C) before pregnancy whose IL-6 responses are in the middle 50% (medium BIR dams) produce adult male offspring with alterations in STAT3, MEF2, and tyrosine hydroxylase protein levels in P0 striatal tissue (Figure 4C–E)8. Male offspring of medium BIR dams exposed to 30 mg/kg of poly(I:C) also exhibit decreased synapse density, and elevated major histocompatibility complex I (MHCI) in dissociated neuronal culture (Figure 4A,B)8. Dams injected with 5 mg/kg of poly(I:C) before pregnancy whose IL-6 responses are in the middle 50% (medium BIR dams) reliably produce adult male offspring with elevated repetitive behaviors and decreased exploratory behavior when exposed to 30 mg/kg of poly(I:C) at E12.5 (Figure 5A–F)8.
Conversely, mice from the high BIR group (with IL-6 levels in the top 25% when exposed to 5 mg/kg of poly(I:C) before pregnancy) reliably produce offspring with no repetitive behavior changes following MIA. However, male offspring of these high BIR dams do exhibit elevated exploratory behavior following MIA (Figure 5D)8. Together, these results indicate that MIA can cause differential outcomes in offspring, depending on the dam's BIR8.
Figure 4. An intermediate dose of poly(I:C) and BIR lead to the greatest outcomes in MIA models. (A) Cortical neurons from offspring exposed to mid-gestational maternal immune activation showed significantly increased MHCI presentation only when dams were given 30 mg/kg of poly(I:C) (one-way ANOVA; F3,19 = 5.156, P < 0.01). (B) In contrast, all dosages (20 mg/kg, 30 mg/kg, and 40 mg/kg) resulted in significantly decreased synapse density in dissociated neuronal culture (one-way ANOVA; F3,43 = 11.01, P < 0.0001). (C–E) P0 striatal western blots show elevated STAT3, MEF2A, and TH, only in animals whose mothers had medium BIRs and were exposed to 30 mg/kg of poly(I:C) (One-way ANOVA; MEF2A: F3,24 = 3.968, P < 0.05; STAT3: F3,24 = 6.401, P < 0.01; TH: F3,24 = 3.668, P < 0.05). Bars represent mean ± SEM. This figure has been modified from Estes et al.8. Please click here to view a larger version of this figure.
Susceptible animals in the medium BIR 30 mg/kg and high BIR 30 mg/kg groups can not only be compared to controls, but also to resilient animals. Injection of medium BIR dams with an even higher dose of 40 mg/kg of poly(I:C) produces offspring with no significant alterations in behavior identified using the assays employed to date (Figure 5A–F)8. This suggests an inverted U relationship between immune activation and susceptibility to MIA.
Figure 5. Male offspring from dams exposed to an intermediate dose of poly(I:C) exhibit the greatest alterations in behavior. (A–F) Male offspring from dams exposed to 30 mg/kg of poly(I:C) (Nested one-way ANOVA; F3,27 = 8.775; Low: P = 0.0427; Medium: P = 0.0062; High: (P = 0.9568) but not 20 mg/kg or 40 mg/kg of poly(I:C) show alterations in repetitive grooming and exploratory rearing behavior. Additionally, animals in the 30 mg/kg poly(I:C) treatment group show disparate forms of susceptibility, and male offspring of medium BIR mothers show increased repetitive behavior and decreased exploration, while male offspring of high BIR mothers show no alteration in repetitive behavior, but had increased exploratory behavior (A,D; Nested one-way ANOVA; F3,15 = 9.407, Low: P = 0.4910; Medium: P < 0.001; High: P = 0.0117). Offspring exposed to 20 mg/kg of poly(I:C) did not appear to meet the threshold of immune activation required to alter neuronal development since they showed no alterations in the behaviors tested, while offspring exposed to 40 mg/kg of poly(I:C) were also mostly resilient to its effects (B,C,E,F). Bars represent mean ± SEM. This figure has been modified from Estes et al.8. Please click here to view a larger version of this figure.
Supplementary Figure 1. Baseline immunoreactivity is not correlated with animal weight. Virgin female mice exhibit a large range of IL-6 responses to 5 mg/kg of Poly(I:C) injected before pregnancy in a weight-independent manner, R2 = 0.0086, P = 0.9. Please click here to download this File.
Maternal infection alters the course of brain development in humans and in both rodents and nonhuman primates4,5,7. Here, a procedure to induce MIA in mice at a mid-gestational time point using poly(I:C) is outlined. This method incorporates assessment of BIR before pregnancy, which increases reproducibility and offers the chance to mechanistically investigate mechanisms that lead to resilience and susceptibility of offspring to MIA8. After MIA, dams from the medium BIR group (with IL-6 levels in the middle 50%) reliably generate adult offspring with alterations in repetitive behaviors, alterations in MHCI levels on neurons from newborn offspring as determined by immunocytochemistry, and elevated levels of striatal tyrosine hydroxylase, MEF2, and STAT3 protein from newborn offspring as determined by Western blot8.
The use of MIA as an environmental model confers increased translational relevance as it meets the criteria for a disease model: construct, face, and predictive validity7. However, as with any environmental model, great care must be taken to minimize external variables. Many factors such as vendor, poly(I:C) lot, timing of injection, age of dams, and even the caging system can alter the impact of MIA on offspring8,9,39. As previously reported, the potency of poly(I:C) is inconsistent between manufacturers, lots, and even bottles within a lot due to high variability in the dsRNA concentrations and molecular weights8,40. Because this variability can increase heterogeneity in the maternal immune response, it is critical that labs determine the effective dose for each lot to maintain maximum reproducibility in observable phenotypes. For example, it has been noted that Charles River mice exposed to MIA produce consistent BIR and dose-dependent phenotypes in offspring, and mice from Taconic may also be impacted in a similar manner with some differences across treatment groups8. Additionally, it is vital that researchers standardize husbandry practices and keep detailed records to increase reproducibility of the model. The publication authored by Kentner et al. outlined the many details that should be included in experimental reports and can also function as a checklist for researchers as they finalize their protocols9.
BIR is assessed using relative serum interleukin-6 (IL-6) levels from virgin female mice. Dividing those mice into three groups (low, medium, and high) reveals reproducible resilient and susceptible models8. Because BIR is a matter of relative concentration of IL-6, it is not crucial to rigorously test the high molecular weight poly(I:C) potency as is necessary with the mixed molecular weight poly(I:C) used to induce maternal immune activation during gestation. BIR is a relatively new measure that may not reduce all variable results.
The immune responses of dams during their first exposure to gestational doses of poly(I:C) may differ from their response during subsequent pregnancies and exposures. To this end, using virgin females reduces the potential for variability that alterations in immune response resulting from multiple pregnancies could present. The weight-based method of pregnancy time point estimation is necessary because mice often do not get pregnant within the first 24 h of mating.
It is important to note that there are statistical challenges with this model. Because MIA is induced in the dams, the offspring are unable to be randomized into treatment conditions. Thus, each litter must be considered an n of 19,41,42, and individuals within that litter should be averaged to create each data point. The most appropriate statistical design for this data therefore utilizes nested analyses8. A minimum of six litters per group (BIR x dose) is needed to reliably detect alterations in behavioral and molecular measures. Significant sex differences have been noted extensively in the MIA literature, and thus sexes should never be pooled in analyses8,9,43,44.
BIR is a relatively new predictive tool, and it has yet to be defined in terms of mechanistic impact. It remains unknown whether BIR is associated with specific gestational immune responses, however the IL-6 response of mice before pregnancy is not equivalent to their response during pregnancy8. BIR therefore represents a correlative predictive measure, and more research is ongoing to determine its origins.
Despite the variability inherent to the MIA model, no other environmental model of neuropsychiatric disorders and NDDs to date can provide the same level of translational relevance. Preparation and extensive pilot testing are necessary to produce consistency in the MIA model, but the robustness of phenotypic results makes up for this initial investment. Ultimately, MIA animal models offer unparalleled potential to investigate a single risk factor that creates divergent and distinct clusters of behavioral and molecular alterations in offspring, similar to those observed in human populations.
The authors have nothing to disclose.
We thank Dr. Myka Estes for her persistence in addressing variability in the mouse MIA model and all of the contributors in Estes et al.8 for their work that led to the development of the methods protocol described here. The research reported here was supported by NIMH 2P50 MH106438-06 (A.K.M.) and NIMH T32MH112507 (K.P.).
0.9% NaCl physiological endotoxin free saline | Sigma-Aldrich | 7647-14-5 | Control and vehicle for Poly(I:C) |
35mm petri dish | Thomas Scientific | 1219Z45 | Used to hold oil during tail bleed |
7.5% TGX gels | Bio-rad | 4561084 | Optional |
Ancare Nestlets | Fisher Scientific | NC9365966 | Optional |
anti-β-tubulin | Millipore | MAB3408 | Optional |
Bio-Plex Pro Mouse Cytokine Standards Group I | Bio-rad | 171I50001 | |
Bio-Plex Pro Reagent Kit with Flat Plate | Bio-rad | 171304070M | |
Bovine Serum Albumin | ThermoFisher | 23209 | Optional |
Centrifuge | Eppendorf | 5810R | Optional |
Covidien Monoject 1/2 mL Insulin Syringe with 28G x 1/2 in. Needle | Spectrum | 552-58457-083 | |
Dithiothreitol | Sigma-Aldrich | D9779-10G | Optional |
Environmental enrichment | Bio-serv | K3327 and K3322 | Optional |
Ethovision | Noldus | Ethovision | Optional |
Fluorsecent-tagged seondary ntibodies | Li-cor | 925-32213 and 925-68072 | Optional |
Food-grade edible oil (like olive, canola or grapeseed) | Various vendors | Use to lubricate tail during tail bleeds | |
HBSS | ThermoFisher | 14060040 | Optional |
High molecular weight polyinositic:polycytidilic acid | Invivogen | #tlrl-pic-5 | Used to establish females' BIR |
Humane Mouse Restrainer | AIMS | 1000 | Used to restrain mouse during tail bleeds |
Image Studio Software | Licor | 5.2 | Optional |
Laemmli buffer | Bio-rad | 1610737EDU | Optional |
Luminex200 | ThermoFisher | APX10031 | |
Microvette CB300 300μl Serum capillary tube | Sarstedt | 16.440.100 | |
Mixed molecular weight polyinositic:polycytidilic acid | Sigma-Aldrich | #P0913 | Gestational induction of MIA |
monoclonal anti-MEF2A | AbCam | ab76063 | Optional |
monoclonal anti-STAT3 | Cell signaling | 12640S | Optional |
Observer | Noldus | Observer | Optional |
Odyssey blocking buffer (TBS) | Li-cor | 927-50003 | Optional |
Odyssey CLx imaging system | Li-cor | 9140 | Optional |
Omnipure PBS | Millipore | 65054L | Optional |
Pierce BCA Protein Assay Kit | ThermoFisher | 23227 | Optional |
polyclonal anti_TH | Pel-Freez | P4101-150 | Optional |
PVDF membrane | Bio-rad | 162-0177 | Optional |
Qsonica Sonicator Q500 | Fisher Scientific | 15-338-282 | Optional |
Quick blood stopper | Petco | 17140 | |
Seal-Rite 1.5 ml microcentrifuge tube, natural non-sterile | USA Scientific | 1615-5500 | |
Soldering stand | Amazon | B08Y12QC73 | Used to hold capillary tube during tail bleeds |
Sunflower seeds | Bio-serv | S5137-1 | Use to increase breeding efficiency |
The Bio-Plex Pro Mouse IL-6 set, | Bio-rad | 171G5007M | |
Tris base | Fisher Scientific | BP152-1 | Optional |
Tween 20 | Bio-rad | 23209 | Optional |