Interleukin-18 modulation in autism spectrum disorders

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by repetitive and stereotypic behaviors and impairment in social communications. In most cases, ASD is clinically diagnosed during the first 3 years of age and is a lifelong condition for most [1]. The incidence of autism has dramatically increased in recent decades rising from 2–5/10,000 children to 60–100/10,000 [2]. One out of 88 children has been identified with ASD, which marks a 78 % increase since the first report in 2007. A recent systematic review established that evidence for a pro-inflammatory state is stronger for autism spectrum disorders compared to other children and adolescent neuropsychiatric disorders [3]. The etiology of autism is practically still unknown although numerous studies have identified a number of genes and environmental factors related to the development of the disease. Alternatively, the involvement of chronic neuroinflammation has been ascertained [4‐7]. Here, dysregulated immune mediators act by altering the normal development of the nervous system leading in particular to the upregulation of inflammatory cytokines in the ASD brain, probably due to altered blood–brain barrier functions [4]. Cytokines have been reported to influence the development of neuronal and glial cells as well as behavioral phenotypes. A bulk of studies showed that multiple members of the large family of cytokines are essential for proper brain development and for synaptic plasticity and responses to injury [5]. Cytokines are produced by neurons, astrocytes as well as microglia, and abnormalities in their levels were reported in association with neurodevelopmental disorders. Interleukin (IL)-6 elevation in the brain, dependent on glia activation, was related to impaired neuroanatomical structures and altered synaptic plasticity. Some cytokines like IL-1β and tumor necrosis factor-α (TNF-α) induce neurotoxicity through elevated glutamate production that results in neuronal excitotoxic death [8]. In a previous study, we showed that cytokines IL-1β, IL-6, IL-12, TNF-α, and IL-23 were significantly increased in the blood serum of ASD patients [9]. The chronic alterations in the inflammatory and immunological responses in patients with autism suggest that this can constitute an endophenotype for ASD. Peripheral cytokines are known to affect different behaviors including sickness and depression and are increased in the brain of subjects with Alzheimer’s disease [10, 11]. Since both the neuroinflammatory processes and the increased immune response observed in ASD would comprise high levels of cytokines in the brain, these proteins could affect behavior [12].

Our interest focuses in the present paper on IL-18, a member of the IL-1 family of cytokines, synthesized as an inactive precursor requiring processing by caspase-1 to be activated; IL-18 is associated to several inflammatory disorders influencing both cellular and humoral immunity [13‐15]. The activity of IL-18 is balanced by the presence of a high affinity, naturally occurring IL-18 binding protein (IL-18BP). In humans, increased disease severity can be associated with an imbalance of IL-18 to IL-18BP such that the levels of free IL-18 are elevated in the circulation [15]. IL-18 synthesis was demonstrated in different brain regions, mainly at the level of activated microglia; moreover, IL-18 was shown to be increased in the brains of AD patients [16], and in addition, IL-18 was shown to increase amyloid-β production by human neuron-like cells [17] affecting amyloid precursor protein (APP) processing and therefore, Aβ production [18]. Finally, it is know that autism patients exhibit increased amounts of APP in their brain [19].

The aim of the present work is to evaluate IL-18 serum levels in autism patients compared to healthy controls and in the murine experimental model of autism, the Reeler mice, compared to wild-type controls. Furthermore, we investigated the expression of IL-18 in the brain sections obtained from individuals affected by tuberous sclerosis with autistic behavior, mimicking different features of ASD subjects or by inflammatory diseases, compared to normal subjects. IL-18 brain expression was investigated in Reeler and wild-type mice as well.

Table 1 summarizes clinical characteristics of ASD patients included in the study, describing in detail their intolerances and allergies, their language development, the presence of delay in babbling, CARS score, comorbidities and other relevant clinical data, and family history.

In the present study, focus was placed on the presence of autoimmune disorder, since the other conditions are present in almost comparable levels in the group of healthy patients: maternal autoimmune diseases include psoriasis (four mothers), rheumatoid arthritis, celiac disease, and Takayasu’s arteritis. In one child, we found paternal familiarity for psoriasis. Another child’ sister suffers from rheumatoid arthritis.

Figure 1 shows the analysis of IL-18 levels in the sera of autism patients and in age- and sex-matched healthy controls. We observed significant differences in the levels of IL-18 present in the sera of autistic patients regardless of age, compared to the values measured in the sera of healthy individuals. Indeed, in patients aged below 10 years, the IL-18 stands at a mean value of 600 pg/mL and goes down to 400 pg /mL in children above the age of 10. The lowest values were observed in patients with severe autism (CARS ≥37).

In order to check whether a similar difference was also present in the brain, we evaluated the expression of IL-18 by immunohistochemistry in the brain sections (Limbic regions/Frontotemporal cortex + tuber) obtained from patients with encephalitis, tuberous sclerosis, or controls. Tuberous sclerosis is a multi-system genetic disease where 25 to 61 % of affected individuals meet the diagnostic criteria for autism with an even higher proportion showing features of a broader pervasive developmental disorder. As shown in Fig. 2, IL-18 was localized within reactive astrocytes in the hippocampus of patients with medically intractable temporal lobe epilepsy as well as in reactive astrocytes of patients with viral encephalitis and in reactive astrocytes and giant cells in cortical tubers of TSC patients. No reaction (below detection level) or weak positivity at the level of hippocampal astrocytes was present in normal brains.

We then decided to analyze the expression of IL-18 in the brain of Reeler mice, an experimental model of autism, characterized by a mutation within Reelin, a glycoprotein of the extracellular matrix that plays a key role in migration and positioning of neurons, thus bearing a fundamental neurodevelopmental role in the laminar and columnar organization of the cortex [22, 23]. When levels of Reelin are reduced by 50 %, as in heterozygous mice compared to wild-type ones, several subtle neuroanatomical and behavioral abnormalities are detectable [20]. We studied the presence of IL-18 and IL-18BP in the brain of heterozygous Reeler mice as well as in wild type: results are shown in Figs. 3 and 4 and in Table 2. Several neurons of Reeler heterozygous mice brain were stained, as well as astrocytes, microglia, and giant cells, as shown in Figs. 3 and 4. We quantified the reaction by gray density analysis obtained by NIS-Element AR program and by counting positive cells in five different fields/sections. Reeler brain sections resulted to have higher IL-18 as well as IL-18BP associated positivity (the intensity of the reaction was 10 % bigger; the number of positive cells was tripled for IL-18 and more than doubled for IL-18BP compared to wild-type brain sections). Moreover, we analyzed the presence of IL-18 and IL-18BP in brain homogenates by Western blots. As shown in Fig. 5, immunoblots revealed increased expression of both IL-18 and IL-18BP in the brain of Reeler heterozygous mice compared to wild type. In addition, we measured IL-18 and IL-18 BP levels by ELISA in plasma of both mice strains. The data indicated that IL-18 is reduced in Reeler mice compared to wild type (p = 00.0556), whereas there was not a significant difference in the IL-18BP plasma levels from Reeler or wild-type mice (Fig. 6).

Finally, we analyzed BDNF levels in the sera from autism patient cohort. We grouped the patients according to the severity of autism, and as shown in Fig. 7, all three patient groups showed significantly higher levels of BDNF compared to the healthy control group (t > 3.34, p < 0.01), with higher amounts detected in patients with severe autism (CARS ≥37): severe autism mean = 1170 pg/ml; standard deviation = 372.7; healthy control sample means = 460.4 pg/mL; and standard deviation = 146.78. Patients with severe autism showed IL-18 serum levels lower than the other groups of patients and much lower than the healthy controls: severe autism mean 113.17 pg/mL; standard deviation 23.79; healthy controls mean 529.9 pg/mL; standard deviation = 274.1 (Figs. 1 and 7). Therefore, patients with severe autism (CARS ≥37) show low levels of IL-18 but high levels of BDNF.

Cytokine imbalance was claimed to impact neural activity and mediates behavioral aspects of autism spectrum disorders [24]. ASD is very heterogeneous from a clinical point of view and is equally very heterogeneous on the genetic aspect [25], although a large number of genetic variants associated with ASD converge toward a core set of dysregulated biological processes involving abnormal immune responses. Gene expression studies support the concept that immune and inflammatory genes are upregulated in ASD, a phenomenon observed both in the brain and in peripheral tissues [26]. Furthermore, it is now established that cells belonging to the immune system, such as microglia, are critical to maintaining normal synaptic plasticity and that alterations of these cells may have an impact on neuronal functions [27], and in particular, the cytokines produced by activated microglia have been demonstrated to affect neuronal synaptic function [28]. Previous results showed that neuroinflammation, with increased amounts of IL-1β, IL-6, and TNF-α, takes place in the brains of ASD subjects [29]. We have shown in a previous study that altered levels of IL-1β, IL-6, IL-12, TNF-α, IL-23, and BDNF occur in the sera of autistic patients [9]. The present results indicate that there is also a dysregulation of IL-18 levels, both in the serum obtained from the same group of patients previously under study as well as in brains of TSC affected subjects with an inverse relationship, though, as a matter of fact, we observed IL-18 decrease in the serum and increase in the brain of both ASD patients and mouse model. The largest decrease was detected in the sera of patients with severe autism (CARS ≥37). It is not easy to explain why quantitatively different levels have been found in the brain compared to the sera of individuals with autism. Previous studies reported that plasma levels of IL-18 were significantly elevated in Alzheimer’s disease compared to healthy controls, but that also levels of IL-18 transcript and proteins were increased in frontal lobe of AD patients compared to healthy age-matched controls [16], even if AD-severe patients showed low levels compared to AD-mild patients [30]. El-Ansary and coworkers reported that autistic children exhibited significantly lower serum concentrations of both Aβ (1–40) and Aβ (1–42) compared to control subjects, despite the higher amounts of Aβ found in the brain of the subjects with autism [31]. Moreover, they detected unexpected lower concentrations of caspase-3, TNF-α, and IL-6 in plasma of autistic patients compared to age- and gender-matching controls [32] despite being known from other studies that the same cytokines were increased in the brain of autistic subjects [33]. The gene encoding for IL-18 and its receptor have distinct promoter regions and then alternatively spliced transcripts giving to these molecules the potential to be produced in tissue-/cell-specific way in response to different stimuli [34, 35]. Furthermore, the splice variants of IL-18Rβ chain have been proposed to be the soluble negative regulator of IL-18 action [36, 37].

IL-18 is physiologically produced in different regions of the brain, including the hippocampus, the hypothalamus, and the cerebral cortex, where several isoforms of its receptor have been identified [13]. It has been hypothesized that IL-18 may have a role in central nervous system development and is involved in synaptic plasticity, in glutamate release and enhanced postsynaptic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor responses as well as in fear memory and spatial learning [13, 38]. IL-18 is also induced during physical/emotional stress responses [34, 39] and high levels of IL-18 are detected in several neuropathological conditions such as during microbial infections, following trauma, stroke, or ischemia. Our results, showing higher number of IL-18-positive cells in the brain of viral encephalitis patients confirm these results. In addition, intrathecal injection of IL-18 is able to induce behavioral, morphological, and biochemical changes similar to that detected after nerve injury [40]. Human IL-18 serum levels are elevated in patients with multiple sclerosis [41], Alzheimer’s disease, vascular dementia, and mild cognitive impairment [42]. Previous studies suggested that brain high levels of IL-18 may induce motor and cognitive dysfunctions, impairing learning and memory by acting as an attenuator of long-term potentiation [40, 43]. Our finding is in agreement with previous results detecting increased levels of IL-18 in the brain of an autism experimental mouse model consisting of an inbred strain with behavioral deficits similar to those found in children with autism [44].

In addition, IL-18 can enhance the production of toxic inflammatory molecules such as interferon (IFN)-γ and IL-1β [45, 46], and recent experimental and clinical studies have proven the close connection between the rise of pro-inflammatory cytokines, glucocorticoids, and behavioral changes, such as those associated with anxiety and depression [47, 48]. In this connection, the pro-inflammatory cytokines induce an altered serotonergic function by increasing the conversion of tryptophan to kynurenine. The decrease in the synthesis of serotonin in the brain leads to the formation of neurotoxins such as quinolinic acid and n-methyl-d-aspartate (NMDA) receptor agonist and contributes to increasing apoptotic events in astrocytes, oligodendrocytes, and neurons, exacerbating mood and oxidant status [49]. The diminished serotonin content in the brain of autistic patients was already revealed by positron emission tomography neuroimaging using a serotonin precursor [50] and was related to language and sensory dysfunctions observed in autism [51] as confirmed by the worsening of stereotyped movements observed in autistic children following acute tryptophan depletion and subsequent reduction of serotonin [52]. In fact, it was suggested that autism might be a disorder of serotonin metabolism. Pro-inflammatory cytokines, including IFN-β and IFN-γ, have been shown to reduce the availability of tryptophan, which is required for 5-hydroxytryptamine synthesis through activation of indoleamine-2,3-dioxygenase (IDO), an initiator of kynurenine pathway. IL-18 can enhance production of toxic inflammatory molecules such as IFN-γ [43] and IL-1β [46],which may lead to a vicious cycle where inflammatory processes contribute to different aspects of neurodegeneration. Furthermore, IL-18 belongs to the family of pro-inflammatory cytokines IL-1 and determines an activation signal on neurons and glia increasing both the synaptic release of glutamate and the expression of its postsynaptic AMPA receptor. IL-1β inhibits the removal of glutamate by astrocytes thus causing an excess of this excitatory neurotransmitter that causes neurotoxicity [53].

Further studies are needed to clarify the cause that led to the increase of IL-18 in the brain of patients with autism and its downregulation in sera. We are aware that the number of subjects analyzed is quite small due to the difficulties of collecting samples from autistic children, especially for the postmortem tissues from autistic patients; however, our data represent the first attempt to investigate the role of IL-18 in ASD, and the small sample size seem appropriate for the exploratory aim of this work. Furthermore, increasing the number of cases examined will clarify whether the decrease of IL-18 in sera can be considered a biomarker of the disease and if this measure in combination with other markers, for example, increased levels of BDNF may be included in a diagnostic panel. In addition, the evaluation of SNPs at the level of IL-18 gene or the existence of splice variants for the beta chain of IL-18 receptor proposed to be the soluble negative regulator of IL-18 action may give important information for the better understanding of the mechanisms underlying IL-18 dysregulation.

Immune dysfunction is present in autism patients. IL-18 is reduced in sera but increased in the brain of patients with tuberous sclerosis with autism. An IL-18 increase was detected also in Reeler brains, mainly at the level of neurons and glial cells; the higher amount of IL-18 was paralleled by a quite similar increase in the amount of IL-18BP. On the contrary, reduced levels of IL-18 were measured in plasma of Reeler mice compared to wild-type mice, whereas no significant variation of IL-18BP was observed. Our data suggest that a chronic neuroinflammation is present in autism affected subjects, including IL-18 dysregulation. The present study might open new scenarios for the comprehension of molecular pathways of the disease.