Family studies have provided some evidences for a higher genetic risk for autism in siblings compared to the general population (Holt et al. 2010) . Early findings suggested that mothers of children with ASD share HLA haplotypes with their children more often than the typically developing mother–child pairs. In addition, these children showed a higher rate of these haplotypes when compared with the healthy population (Gregg et al. 2008) . HLA haplotypes, considered among the strongest predictors of risk for autoimmune conditions, have also been associated with ASD (Fig. 9.1).

Studies in children and adults with autism have indicated other various candidate genes, many of them related to the immune system. These genes include the null allele of C4B in class III region, MHC haplotypes B44-SC30-DR4, HLA-DR β1, DR4, and DR13, all reinforcing also the factors lined as guttering in the development of autoimmunity (Daniels et al. 1995; Lee et al. 2006; Torres et al. 2006; Warren et al. 1995) . Genes regulating signal pathways related to the development of both neural and immune systems have also been referred to as altered in ASD: macrophage migration inhibitory factor (MIF) (Grigorenko et al. 2008) , MET encoding tyrosine kinase, (Correll et al. 2004) , the serine and threonine kinase C gene PRKCB1 (Lintas et al. 2009) , PTEN (Herman et al. 2007; Butler et al. 2005), and the reelin gene RELN (Garbett et al. 2008; Pardo and Eberhart 2007 ; Fig. 9.1).

As mentioned above, not only the adaptive but also the innate immune system response appears altered in ASD. These abnormal responses in children with autism could result from prenatal infection or exposure to toxic substances, and/or they could have a genetic component.

Various viral and bacterial infections during pregnancy have been linked to a higher risk for autism. More than 40 years ago, a higher risk for autism was reported in the offspring of maternal rubella infection cases (Chess 1971) . Since then, prenatal exposure to different virus have been linked to increased risk for ASD, including herpes simplex virus (DeLong et al. 1981; Ghaziuddin et al. 1992; Gillberg 1986; Greer et al. 1989) , cytomegalovirus (Ivarsson et al. 1990; Markowitz 1983; Stubbs 1978; Yamashita et al. 2003) , measles (Ring et al. 1997; Singh and Jensen 2003) , and viral meningitis (Barak et al. 1999; Ring et al. 1997) . Moreover, a larger study showed an increase in the prevalence of autism in children born to mothers hospitalized for viral infections during the first trimester of pregnancy or for bacterial infections during the second trimester (Atladottir et al. 2010) . These evidences support the idea that the maternal response to viral infection can be a risk factor for autism (Ciaranello and Ciaranello 1995) .

How can different pathogenic infections, all result in an increased risk for autism? Evidence from experimental animal models show that prenatal and early life inflammation can alter behavior later in life (Coe and Lubach 2005; Meyer et al. 2006, 2007, 2009; Depino 2006; see also Chap. 8). Therefore, different infectious agents, autoimmune diseases, and stressful events could elicit the expression of immune factors affecting brain development and resulting in the behavioral symptoms (Fig. 9.2).

Interestingly, fetal inflammatory response syndrome (FIRS) is a condition in which, despite an absence of cultivable microorganisms, neonates have very high circulating levels of inflammatory cytokines, such as IL-1, IL-6, IL-8, and TNF-α (Madsen-Bouterse et al. 2010) . Studies in animal models have shown that viral infection of the placenta triggers a fetal inflammatory response similar to the one observed in FIRS, even though the virus is not able to reach the fetus. In the case of human FIRS, these cytokines have been shown to affect the CNS and the circulatory system (Deverman and Patterson 2009) . Fetal morphological abnormalities in the animals have also been reported, probably caused by fetal proinflammatory cytokines such as IL-1, TNF-α, MCP-1, MIP-1β, and IFN-γ, and it has been suggested that beyond morphological effects on the fetal brain, the presence of FIRS increases the risk of autism, schizophrenia, neurosensorial deficits, and psychosis. Interestingly, all these factors induced at the neonatal period can predispose to diseases in adulthood (Romero et al. 2007) . Due to this, we propose that an inflammatory response in the placenta, altering the cytokine balance in the fetus, may affect the normal development of the fetal immune system leading to anomalous responses during childhood or later in life. In all these cases the parasite may never reach the placenta, but the inflammatory process at this level may affect the normal fetal development.

Recently, it has been hypothesized that the presence of gastrointestinal inflammation makes a child with a genetic predisposition for ASD more prone to express the autistic phenotype, or that it increases the severity of the autistic behavior. A recent meta-analysis found that the reported prevalence of gastrointestinal symptoms in individuals with ASD can range from 9 to 91 % (Buie et al. 2010; Coury et al. 2012) . The variability observed could be mainly due to the lack of appropriate (nonrelated) control subjects, but the data is however consistent with a high prevalence of gastrointestinal symptoms and disorders in ASD individuals.

Gastrointestinal symptoms include abdominal pain or discomfort, constipation, diarrhea, and bloating (Afzal et al. 2003; Horvath and Perman 2002; Buie et al. 2010) . In addition, autism severity shows a strong correlation with gastrointestinal disturbances (Adams et al. 2011) , since studies in the field have mainly investigated immune dysfunction and the connection between the gut microbiome and the brain.

Autoimmune diseases, viral infection, gastrointestinal disturbances, and other proposed etiological and pathophysiological pathways in ASD share a common result: Changes in the levels of cytokines produced either in the periphery and/or within the brain. Hence, cytokines could be considered as the main molecular effectors acting on the brain to alter development and behavior.

A neuroimmune hypothesis on the role of cytokines in autism can be that: (1) proinflammatory cytokines arising from maternal inflammation, infection or allergy during pregnancy, cross the placenta, enter the fetal circulation and pass through the fetal BBB; (2) an excess of proinflammatory cytokines released into the fetal brain cause aberrant neurogenesis, alters synaptogenesis and neuronal function; and (3) altered brain development results in the behavioral symptoms, including reduced sociability, stereotypic and repetitive behavior, hyperactivity, attention deficit, anxiety, and hypersensitivity to environmental stimuli.

After an immune response is elicited, mechanisms are triggered to limit this response either in time or in magnitude. When these mechanisms fail, autoimmune or other pathological results can derive and affect the organism. So, as important to mount an effective immune response is to end it when the stimulus is eliminated. To this purpose, anti-inflammatory cytokines and endocrine factors are a key to limit the immune response, and their failure can lead to long-lasting immune diseases.

The analysis of cytokine profiles in autism pathology reinforces the implication of an immune deregulation in this disorder. The plasma of autistic patients presents increased levels of proinflammatory cytokines such as IL-1β, IL-6, and TNF-α (Ashwood et al. 2011a; Malik et al. 2011) , and reduced levels of anti-inflammatory cytokines such as TGF-β1 (Ashwood et al. 2008a). In addition, peripheral blood mononuclear cells (PBMC) from subjects with ASD show greater lipopolysaccharide-stimulated production of IL-1β (Enstrom et al. 2010b) , TNF-α, and IL-6 (Jyonouchi et al. 2008) than those from controls. Moreover, PBMC from ASD children stimulated with phytohemagglutinin have increased activation of both Th2 and Th1 cytokines (Molloy et al. 2006) . Finally, a multiplex assay for cytokines and chemokines in plasma from individuals with high-functioning ASD and matched controls showed increased plasma concentrations of IL-1β, IL-1RA, IL-5, IL-8, IL-12(p70), IL-13, IL-17, and GRO-α (Suzuki et al. 2011) . On the other hand, Onore and colleagues showed normal levels of IL-17 and decreased IL-23 in autistic children (Onore et al. 2009) . Another study revealed significant differences to IL-1β, IL-12(p70), IL-12(p40), and IL-17 in relationship to electrophysiological events, identified at EEG register and focused at temporal region (Robinson et al. unpublished results).

Altered brain cytokine levels have also been reported in ASD patients. Postmortem studies have shown high levels of TGF-β, IL-6, and MCP-1, as well as activated microglia in the brain of ASD subjects compared to controls (Vargas et al. 2005) . Moreover, increased expression of cytokines has also been reported in cerebrospinal fluid of ASD patients (Chez et al. 2007; Garbett et al. 2008; Li et al. 2009) .

Peripheral cytokines can not only alter immune functions but they can also affect different behaviors (Konsman et al. 2002) . Cytokines can reach the brain directly, or through activation of different communication pathways, including the gut–brain and immune–brain pathways.

During an allergic process, immunoglobulins and neuropeptides (e.g., substance P, NGF, VIP, and neurotensin) activate mastoid cells and basophils to release cytokines, which can in turn trigger enteric neurons. As the mast cell–neurons interaction occurs at the gut level, an allergic reaction at this level might influence behavior following the action of enteric neurons, which are triggered by mast cell to convey information through the vagal and spinal afferent pathways into the CNS. A hypothesis emphasizing the effect of allergy-linked inflammatory mediators in autism underlines that molecules like IL-6 and histamine promote gene transcription of estradiol and retinoic acid. This results in an overexposure of the fetus to retinoic acid and estradiol, which has been described as causing brain abnormalities similar to those observed in ASD, such as cerebellar malformations, abnormalities of dopaminergic system, and cranial nerve abnormalities (Garay et al. 2012) . Moreover, fetal overexposure to estradiol can result in anxiety, hyperactivity and attention deficit, stereotype and repetitive movement, and motor deficits (de Theije et al. 2011) .

For centuries, the structured system of the BBB, the lack of conventional lymphatic drainage, and the low response to alloantigen inoculated in the CNS, were arguments to consider this compartment as a privileged site. Nevertheless, today it is well known that there is no such privilege, but rather the main differences of the immune response in the brain, regarding the periphery, are in the kinetics and the degree of regulation of the different stages of the immune response at this level, which will take place under the functional control of the immune response and influence of cellular mediators acting in a regulated quantitative and temporally balanced sequence. Moreover, numerous cytokines are able to cross the BBB, including IL-1β, IL-6, and TNF-α (McLay et al. 2000; Prat et al. 2001) , whose plasma levels are enhanced in autistic patients (Ashwood et al. 2011b; Malik et al. 2011) . After reaching the brain, specific cytokines can affect brain development and function. For example, TGF-β involved in diverse aspects of development (Ashwood et al. 2008b; Letterio and Roberts 1998) . Levels of TGF-β have been found to be reduced in the plasma of autistic children and adults, in correlation with more severe behavioral scores (Ashwood et al. 2008b), while higher levels of TGF-β were described in postmortem brain and cerebrospinal fluid samples in ASD (Vargas et al. 2005) . This apparent contradiction could result from different roles of this cytokine in the brain and in the periphery, and/or result from different temporal roles of this cytokine during development. Indeed, using an animal model, it was shown that the overexpression of TGF-β1 in the brain can have opposite effects on behavior, dependant on the age of the animals (Depino et al. 2011) .

Macrophage inhibitory factor (MIF), other proinflammatory immune regulator that is constitutively expressed in the brain tissue, has an important influence on neural and endocrine systems (Grigorenko et al. 2008) . Recently, high plasma levels of MIF were observed in ASD individuals with severe behavioral symptoms, and genotyping studies linked MIF to ASD and showed the presence of two polymorphisms in its promoter region (Grigorenko et al. 2008).

In addition to others functions in the immune system, TNF-αα and IL-1β regulate brain processes such as neuron plasticity, long-term potentiation (LTP) and associative learning, and memory (more information in Chap. 7) (Depino et al. 2004) .

In the brain, cytokines can also act on neuroglial cells to induce neuroinflammation . Several factors released by astrocytes might be important for the induction and maintenance of the BBB, as manifested by the appearance of endothelial tight junctions in cells unsheathed by astrocytic endfeet. In the healthy CNS, astrocytes and microglia play important roles in neuronal function and homeostasis, as they are both fundamentally involved in cortical organization, neuroaxonal guidance, and synaptic transmission (Bilbo and Schwarz 2009) .

Astrocytes and microglia, produce neurotrophic factors, cytokines and chemokines (Bilbo and Schwarz 2009), regulate the integrity of the BBB (Prat et al. 2001) , and maintain brain homeostasis. Conversely, activated astrocytes and microglia can induce neuronal and synaptic changes, modifying CNS homeostasis, and contributing to neuronal dysfunction during disease. In postmortem brains of autistic patients, enhanced activation of astrocytes and microglia was observed in the subcortical white matter of the midfrontal gyrus, in the anterior cingulate gyrus, in the granular cell layer, in the Purkinje cell layer, and in the white matter of the cerebellum (Pardo et al. 2005; Vargas et al. 2005) . Similar changes have been observed in animal models of autism (Lucchina and Depino 2014; more information in Chap. 8).

Altered brain development can result in long-term changes in neuronal function, evidenced by altered neurotransmitter production and receptor expression. In addition, an alternative to the hypothesis of cytokines acting on the brain to induce behavioral changes is to propose that changes in brain development can result in persistent alterations of neurons, which can in turn modulate immune function. Differential changes in neurotransmitter systems have been referred in autism, including serotonin (5-HT), dopamine, noradrenaline (norepinephrine), gamma-aminobutryic acid (GABA), glutamate, and neuropeptides suggesting that a dysregulation in neurotransmission could result in altered development. Molecules like oxytocin and vasopressin are involved in immune tolerance, serotonin exerts suppressive or proliferative responses on T cells , and the vasoactive intestinal peptide (VIP) has an immunomodulatory action susceptible to be implicated in ASD.

It has been also reported that cytokines can affect enteric glial cells, although their effects on mood, sleep, nutritional intake, exploratory behavior, and social interactions do not necessarily imply that they cause neuronal damage or death (Bernardo et al. 2004) . Systemic cytokine administration of IFN-α, IL-2, or TNF-α can cause increases in noradrenergic, dopaminergic, and serotonergic metabolism in the hypothalamus, hippocampus, and nuleus accumbens (Mohankumar et al. 1991; Shintani et al. 1993; Merali et al. 1997) .

Neurotransmission at cholinergic synapses is essential for functional aspects of cognition, reward, motor activity and analgesia, and a dysregulation of cholinergic neurotransmission has been associated with autism. Differences in the regional expression of specific nicotinic acetylcholinergic receptor (nAChR) subunits in postmortem adult brains from autistic and non-autistic individuals have been reported (Ray et al. 2005) . A decrease in the density of [3H]epibatidine binding sites within the parietal cortex, cerebellum, and thalamus of autistic brains was correlated to decreases in the levels of α3, α4, and β2 subunits (Martin-Ruiz et al. 2004; Lee et al. 2002) . Although without a total understanding on the impact of this finding in autism pathology, acetylcholine receptor (AChR) expression raises important questions on whether deficits in nAChR neurotransmission identified in at least a subset of autistic children might confer a particularly high sensitivity to the adverse effects of organophosphates and/or neonicotinoid exposure. Indeed, allelic variants of the gene that encodes for a metabolic enzyme for the detoxification of organophosphates have been associated with autism in North America, and the ASD-related variant has lower metabolic activity in vitro (D’Amelio et al. 2005) .

GABA receptors (GABAR) primarily mediate inhibitory responses, but also mediate excitatory transmission (Owens and Kriegstein 2002) . This temporally regulated effect of GABA has been proposed to underlie a key role of this neurotransmitter in conveying information of the environment into a developing brain (Ben-Ari 2002) . A failure in proper regulation of the chloride gradient can result in an imbalance between excitation and inhibition within developing neural circuits, a phenomenon that is proposed to underlie autistic symptoms (Snijders et al. 2013) . Changes in GABA levels have been found in platelets of autistic children (Rolf et al. 1993) . Moreover, the density of GABAR is diminished in the hippocampus of autistic individuals (Blatt et al. 2001) , and both glutamic decarboxylases (GAD) 65 and GAD67 are reduced in the autistic parietal and cerebellar cortex (Fatemi et al. 2002) . Differential levels of specific GABAA and GABAB receptors also suggest a role of GABA in autism (Schmitz et al. 2005; Fatemi et al. 2009a; Fatemi et al. 2009b) .

Central catecholamine imbalance has been reported as a major factor in the behavioral phenotype in autism, and a hypofunction of dopamine (DA) and norepinephrine systems has been also suggested in autism. Attention to a role of DA in autism was drawn after the observation that some DA blockers (i.e., antipsychotics) can treat some aspects of autism (Arnold et al. 2010) . However, the vast majority of studies measuring the levels of DA and its metabolites in the blood or the CSF of autistic individuals, have failed to detect differences with controls as was reviewed by Lam et al. (2006) .

The concepts of autoimmunity and neuroimmune dysfunction in neurodevelopmental disorders are current important areas of research. However, some precautions are required in interpreting the information at this time of knowledge, since much of the evidences are not rigorous and most are aimed at identifying potential markers, with only preliminary evidences on the mechanisms underlying their involvement in the pathology of ASD.

Many of the experimental data mentioned in our review only suggest or support a primary role of failure of immune functions in the pathogenesis of this particular neuropsychiatric disorder, which does not imply that such a primary role could have been formally proved. We consider that some mechanisms underlying these links could be discovered using the recently developed and validated animal models (see Chap. 8).

This understanding may contribute to designing better protocols of treatment, useful to recover from neuronal dysfunction related to social interaction, verbal and nonverbal communication, as well as the repetitive, stereotypical, and overly restrictive behaviors present in ASD.

In the context of immunomodulatory therapy in ASD, it has been suggested that around 70 % of the children affected undergo complementary and alternative medicinal therapies to treat their symptoms (Levy et al. 2003; Wong and Smith 2006) , including the use of probiotics, anti-infectives, and dietary augmentation (Golnik and Ireland 2009) . The clinical trials aimed at testing therapies that modulate the immune response in children with ASD have addressed mainly the use of steroids, intravenous immunoglobulin G, antibiotics, and vitamin D (Cannell 2008; Gupta 1999; Plioplys 1998; Pradhan et al. 1999; Sandler et al. 2000; Dalakas 1997; Shenoy et al. 2000; Matarazzo 2002; Bradstreet et al. 2007; Buitelaar et al. 1990; Geier and Geier 2005) . Understanding the biological effects of these treatments could probably provide a better background toward more effective conventional and alternative therapies.

Although it is well known that autism affects primarily the brain function (mainly social functioning and cognition), it is also known that it follows the compromise of other organs and systems. So, published evidences have identified widespread changes in the immune system of children with autism, both at systemic and cellular levels. Brain specimens from autistic subjects exhibit signs of active, ongoing inflammation, as well as alterations in gene pathways associated with immune signaling and immune function. Moreover, many genetic studies have also indicated a link between autism and relevant genes to both the nervous system and the immune system, which can in turn affect function in both systems. Together, these reports suggest that autism may in fact be a systemic disorder with connections to abnormal immune responses in and outside the brain. Consequently, the immune system dysfunction may represent a novel target for treatment. Hence, a better understanding of the involvement of the immune response in autism, and of how early brain development is altered, is crucial for a better and earlier intervention of the disease and has therapeutic implications. In our opinion, this chapter is a modest contribution to these propositions.