Schizophrenia Environmental Risk Factors Act Synergistically in Mice
12 March 2013. A new mouse study links two environmental risk factors for schizophrenia and other psychiatric disorders—prenatal maternal infection and early life trauma—and finds a synergistic effect on behavior, neurochemistry, and neuroimmunology. Led by Urs Meyer of the Swiss Federal Institute of Technology and published March 1 in Science, the study supports a “two-hit” model of schizophrenia, which posits that genetic or environmental factors that disrupt early brain development result in vulnerability to a later environmental hit that produces symptom onset (Maynard et al., 2001).
The idea that mothers who develop infections while pregnant are more likely to have offspring with schizophrenia has been around for several decades, first supported by the finding in a Finnish cohort that fetuses in their second trimester during the 1957 Asian influenza epidemic had an elevated risk of developing schizophrenia later in life (Mednick et al., 1988; see SRF related news story). The risk for schizophrenia extends beyond viral infections to include bacteria and parasites as well, suggesting that the mother’s inflammatory response to infection is more critical to the alteration of fetal brain development than the actual infection itself (Brown and Derkits, 2010). Early life trauma, including abuse and death of a parent during childhood or adolescence, has also been associated with risk for schizophrenia and psychosis later in life (see SRF related news story; SRF news story).
The risks of maternal infection and trauma extend beyond schizophrenia, and have also been implicated in other developmental psychiatric illnesses such as bipolar disorder and autism (Tsuchiya et al., 2003; Herbert, 2010). However, the rather small risk that each factor alone contributes to the illnesses suggests that two or more factors may combine to produce symptom onset, with the first factor producing a vulnerability to subsequent hits.
In the current study, first author Sandra Giovanoli and colleagues examined this hypothesis using a mouse model. The researchers investigated the consequences of prenatal immune activation with or without subsequent stress during puberty. To elicit maternal immune activation, they mimicked viral infection with administration of polyriboinosinic-polyribocytidylic acid, or poly(I:C), to pregnant dams. Offspring were then allowed to develop normally, or were exposed to a series of five stressors such as an electric foot shock and water deprivation during puberty (postnatal days 30-40).
Immune activation + stress = behavioral alterations
Behavioral testing of adult animals revealed several separate and combined effects of maternal immune activation and peripubertal stress. Anxiety-like behavior, as assessed using the elevated plus maze, was similarly increased in both groups of stress-exposed animals (with and without prenatal immune activation), suggesting that the prenatal immunological manipulation did not affect anxiety-like behavior. Stress exposure and immune activation also exerted separate effects on a conditioned active avoidance paradigm, with the absence of a significant latent inhibition effect, indicating impaired associative learning, observed in all groups except the control animals not exposed to stress.
In contrast, synergistic effects of prenatal immune activation and peripubertal stress were observed on two schizophrenia-relevant behaviors. Neither immune activation nor stress exposure alone produced a sensorimotor gating deficiency (measured by prepulse inhibition of the acoustic startle reflex) or behavioral hypersensitivity to the psychotomimetic drugs amphetamine and MK-801 (measured by locomotor activity). However, when the two environmental factors were combined, sensorimotor gating impairments and hyperactivity after drug exposure were observed.
In contrast to the findings in adult animals, with the exception of anxiety-like behavior, none of the behaviors were altered during puberty, suggesting that the effects of immune activation and stress are dependent on maturational processes that occur during or after puberty. In addition, when the researchers delayed the application of stress until later in adolescence (postnatal days 5-60), they did not find an interaction with prenatal immune activation, indicating that a precise timing of the stress was necessary to produce the synergistic effects.
The observed behavioral abnormalities appear to be independent of changes in the hypothalamus-pituitary-adrenal (HPA) stress-response system, since levels of the primary HPA axis hormone corticosterone were unaffected. However, several neurochemical abnormalities were observed after prenatal immune activation and/or peripubertal stress. Giovanoli and colleagues found elevated dopamine levels in the nucleus accumbens after prenatal immune activation, which were not affected by stress exposure. In contrast, stress exposure produced a decrease in the medial prefrontal cortex levels of serotonin that was independent of prenatal history. Dopamine levels in the hippocampus were elevated only after exposure to both environmental factors, pointing to a role for combined exposure in alterations of the classical schizophrenia neurotransmitter.
Interestingly, another recent study has found alterations in the glutamatergic system, another key player in schizophrenia, in mice born to mothers exposed to poly(I:C) (Holloway et al., 2013). A reduction in mGluR2 receptors in frontal cortex was observed, suggesting that the glutamatergic system may be an interesting target to examine in future studies investigating the synergy between maternal immune activation and stress. Holloway and colleagues also observed elevated levels of 5HT2A receptors in offspring of poly(I:C)-exposed mice.
Because prenatal immune activation and chronic stress have both individually been associated with subsequent immune alterations in the brain (Hsiao et al., 2012; Frank et al., 2007), Meyer and colleagues next examined neuroimmunological responses to their maternal immunological activation and peripubertal stress paradigms in two relevant brain regions—the hippocampus and the prefrontal cortex—as well as in a control region, secondary motor cortex, which is insensitive to immunological changes after stress.
Immune activation and stress, singly or in combination, had little impact on microglial cells (key mediators of inflammatory responses in the nervous system) in adult animals, but led to a dramatic increase in peripubertal animals’ vulnerability to stress-induced neuroimmunological changes. CD68 and CD11b, markers of activated microglia, were increased in the hippocampus and prefrontal cortex, but not secondary motor cortex. These changes were also accompanied by increased levels of the proinflammatory cytokines IL-1β and TNF-α, although levels of corticosterone were unchanged. Finally, mRNA levels of CD200, CD200R, and CD47, effectors of neuron-microglia inhibitory signaling, were also altered in the hippocampus and prefrontal cortex after combined exposure to prenatal immune activation and stress.
Future studies are needed to determine how well the findings in mice translate to humans, although the current data are consistent with synergistic interactions between two known environmental risk factors for developmental psychiatric illnesses. The authors write that their findings suggest that “prenatal adversities … can thus function as a ‘disease primer’ that increases the offspring’s vulnerability to the detrimental neuropathological effects of subsequent stress exposure during peripubertal life.”—Allison A. Curley
Giovanoli S, Engler H, Engler A, Richetto J, Voget M, Willi R, Winter C, Riva MA, Mortensen PB, Schedlowski M, Meyer U. Stress in puberty unmasks latent neuropathological consequences of prenatal immune activation in mice. Science . 2013 Mar 1 ; 339(6123):1095-9. Abstract
Comments on News and Primary Papers
Primary Papers: Stress in puberty unmasks latent neuropathological consequences of prenatal immune activation in mice.Comment by: Patricio O'Donnell, SRF Advisor
Submitted 7 March 2013
Posted 7 March 2013
The “two-hit” hypothesis for neuropsychiatric disorders, although popular, has seldom been directly tested using animal models. Giovanoli et al. provide compelling evidence of synergistic interactions among prenatal immune activation and pubertal social stress in driving abnormal adult behaviors. This is a refreshing use of animal models, quite welcome at a time when the field is plagued with useless considerations about “validity” in rodent models of psychiatric disorders. Questioning animal models based on perceived validity issues not only does not advance the field, but also blocks avenues of progress, and animal research related to psychiatric disorders should abandon psychiatry-irrelevant concepts such as “validity.” Animal models are tools, like reagents, that can be used to test specific hypotheses that relate to etiology or pathophysiology of mental disorders (O’Donnell, 2013). The group of Urs Meyer used maternal immune activation in mice as a tool to establish a latent pathological condition that per se produces mild behavioral deficits. Adolescent stress, on the other hand, has also been shown to produce mild adult deficits. Combining both approaches results in adult animals with more strongly affected behavior and neurochemistry. The study should be used as an example of how to cast animal models in a useful light.
The article by Giovanoli et al. highlights altered neuroimmune responses in the adolescent brain following the prenatal manipulation. Proinflammatory cytokines are elevated in the hippocampus of young mice, but only in those stressed as adolescents. These data indicate that animals without a predisposition can handle the immune activation induced by stress, but not those with a compromise induced by maternal immune activation. The link to psychiatry is tantalizing. The notion of immune factors and oxidative stress playing a role in major psychiatric disorders has been around for some time; only recently there has been a surge in animal testing of these possibilities. The findings reported in Science open the way to the exploration of novel therapeutics that may target immune factors and/or microglia activation for psychiatric conditions.
O’Donnell P. How can animal models be better utilized? In: Schizophrenia: Evolution and Synthesis. S.M. Silverstein, B. Moghaddam, T. Wykes, Eds. Strungman Forum Reports, vol 13. Cambridge, MA: MIT Press (2013).
View all comments by Patricio O'Donnell
Primary Papers: Stress in puberty unmasks latent neuropathological consequences of prenatal immune activation in mice.
Comment by: Lindsay Hayes, Akira Sawa (SRF Advisor)
Submitted 8 March 2013
Posted 8 March 2013
In a time when gene-environment interactions are the hot topic, Giovanoli et al. pushed the envelope a little further by providing evidence that multiple environmental hits may be an additional mechanism mediating the neuropathology of some developmental psychiatric disorders—a model for environment-environment interactions. In the current study, the authors used a sub-threshold dose of maternal immune activation (MIA) in combination with juvenile stress to uncover a synergistic effect mediating a worsened behavioral outcome, suggesting MIA and juvenile stress may utilize a common biological pathway. In addition, they revealed that juvenile stress was only effective during a critical period to mediate the compounded effect, thus highlighting the need for a greater understanding of the mechanisms underlying biological susceptibilities in adolescence.
Urs Meyer, and his group, are leading figures in the MIA model used to study developmental psychiatric disorders. MIA at various stages during prenatal development can lead to stereotypic behavioral changes in the offspring (Meyer et al., 2009). While there is not a clear mechanism mediating the phenotype of MIA, this paper showed that MIA leaves the offspring vulnerable to any secondary insult, be it genetic or, in this case, environmental.
A mechanism the authors propose to mediate the behavioral changes is through microglia cell number and activation status, including data on normal astrocytes and endothelial cells. Stress caused increased cell numbers of Iba1+ microglia and an increased cell size, suggesting an amoeboid/activated morphology in control and MIA offspring. But the combination of MIA and stress led to increased expression of CD68, a functional marker for activated microglia, and production of proinflammatory cytokines, suggesting that MIA induced susceptibility and, when combined with stress, provoked microglial overactivation. However, these changes were mostly observed in the hippocampus at postnatal day (P) 41-45, before the emergence of the behavioral phenotypes at P70. Together, these data suggest that the MIA made the microglia susceptible to activation, and the addition of juvenile stress reactivated the microglia, causing proliferation in the hippocampus and activation in the hippocampus and prefrontal cortex. This potentially produced a remodeling of the neural circuits mediating the behavioral phenotype observed at P70. The authors also showed changes in dopamine and serotonin levels in the brain; however, it is unclear if microglia play a role in the neurotransmitter regulation. Alternatively, we recently reported that adolescent stress, in combination with genetic susceptibility, showed changes in extracellular dopamine through epigenetic modification of mesocortical dopamine neurons (Niwa et al., 2013). In any case, microglia are long-lived in vivo, and the residual increased numbers of microglia in the hippocampus remained at P70 but returned to a resting state by showing normal size and CD68 expression. The effect on microglia was different based on brain region (hippocampus vs. prefrontal cortex), emphasizing a need for further microglia subcategorization and identification of molecular markers that may cause microglia in different circuits to behave differently to stimulation. These distinctions are critical for future drug design studies to target specific circuits for defined symptoms in disease.
Several questions still remain, such as the mechanisms for how microglia are interacting with the neural circuits and how MIA makes the microglia susceptible to further overactivation. A new trend in uncovering the role of microglia in specific behaviors was observed in several previous studies by further rescuing the behaviors with bone marrow transplantation following lethal irradiation (Chen et al., 2010; Derecki et al., 2012; Hsiao et al., 2012). This strategy is a good next setup to solidify the role of overactivated microglia in these specific behaviors.
Meyer U, Feldon J, Fatemi SH (2009) In-vivo rodent models for the experimental investigation of prenatal immune activation effects in neurodevelopmental brain disorders. Neurosci Biobehav Rev 33:1061-1079. Abstract
Niwa M, Jaaro-Peled H, Tankou S, Seshadri S, Hikida T, Matsumoto Y, Cascella NG, Kano S, Ozaki N, Nabeshima T, Sawa A (2013) Adolescent stress-induced epigenetic control of dopaminergic neurons via glucocorticoids. Science 339:335-339. Abstract
Chen SK, Tvrdik P, Peden E, Cho S, Wu S, Spangrude G, Capecchi MR (2010) Hematopoietic origin of pathological grooming in Hoxb8 mutant mice. Cell 141:775-785. Abstract
Derecki NC, Cronk JC, Lu Z, Xu E, Abbott SB, Guyenet PG, Kipnis J (2012) Wild-type microglia arrest pathology in a mouse model of Rett syndrome. Nature 484:105-109. Abstract
Hsiao EY, McBride SW, Chow J, Mazmanian SK, Patterson PH (2012) Modeling an autism risk factor in mice leads to permanent immune dysregulation. Proc Natl Acad Sci USA 109:12776-12781. Abstract
View all comments by Lindsay Hayes
View all comments by Akira Sawa
Primary Papers: Stress in puberty unmasks latent neuropathological consequences of prenatal immune activation in mice.
Comment by: John McGrath, SRF Advisor
Submitted 8 March 2013
Posted 12 March 2013
I recommend this paper
This is a smart paper. It extends the solid body of research demonstrating the impact of prenatal maternal immune activation (the first hit) on brain development and later brain functioning. The impact of stress (the second hit) compounds or amplifies the resultant phenotype. It is now clear that trauma is associated with an increased risk of schizophrenia, but why do only a small percent of individuals exposed to stressful trauma develop schizophrenia (vs. developing more common disorders such as depression or anxiety)? These new models help us unravel the complex pathways that can disrupt optimal brain function.
Can I also echo Patricio O'Donnell's comment about the need to adopt a modern and pragmatic perspective on the utility of animal models for understanding complex neuropsychiatric disorders? Considering how little we understand about brain development, any animal experiments that uncover pathways between 1) candidate exposures or genes linked to schizophrenia, versus 2) any altered brain outcome (from molecular to behavioral) should be of intense interest to the field. This type of research can inform neuroscience discovery, and feed back into more refined research questions. The new paper from Urs Meyer and colleagues is a perfect example of this. Based on clues linking prenatal infection and risk of schizophrenia, animal models using immune activation (like poly(I:C) or LPS) have made crucial discoveries about the crosstalk between biological systems first identified in immune function, and brain development.
The schizophrenia research community should not passively wait for the rest of neuroscience to work out the instructions for building healthy brains. We should take the lead based on the clues that we observe clinically or from epidemiology, and use the tools of modern neuroscience to catalyze discoveries (McGrath and Richards, 2009).
McGrath JJ, Richard LJ. (2009) Why schizophrenia epidemiology needs developmental neurobiology – and vice versa. Schizophrenia Bulletin 35(3):577-81. Abstract
View all comments by John McGrath
Comments on Related News
Related News: Evidence Mounts for the Maternal Inflammation Hypothesis of SchizophreniaComment by: Chris Carter
Submitted 16 July 2014
Posted 16 July 2014
I recommend the Primary Papers
If the development of schizophrenia depends upon maternal infection, perhaps a large number of susceptibility genes (possibly related to infection susceptibility and the immune system) would be concentrated in maternal alleles, with other maternal/paternal genes contributing later in life, and in different ways, for the offspring.
Have there been , or are there programmed, any genome-wide association studies on non-schizophrenic mothers of schizophrenic patients? A comparison of maternal/paternal genetic donation might perhaps be a useful means of dissecting out the various pathways leading to disease.
View all comments by Chris Carter