Association of social and cognitive impairment and biomarkers in autism spectrum disorders

Autism spectrum disorders (ASDs) are a behaviorally defined group of disorders characterized by impaired social skills, impaired language, and restricted areas of interest [1]. Additional features may include poor eye contact [2], repetitive behavior [3], sensory modulatory dysfunction [4], and varying levels of cognition and motor disturbances [5, 6]. Clinical signs are usually present at the age of 3 years, but prospective studies of infants at risk have demonstrated that deficits in social responsiveness, communication, and play could be present early, at the age of 6–12 months [7]. However, the condition is often missed and not diagnosed until later in a child’s life, especially when the condition is mild or even moderate in severity. There has been increasing interest in developing effective interventions for young children with autism since the evidence suggests that early intervention programs are indeed beneficial for children with autism, often improving developmental functioning and decreasing maladaptive behaviors and symptom severity and can also improve outcomes in later years for many individuals [8].

Abnormalities in neurotransmitter systems have frequently been reported in autism. Clinical observations include altered levels of various neurotransmitters compared to controls, including alterations in dopamine metabolism [9] and dysregulations of serotonergic systems [10–13]. Expression of several types of gamma aminobutyric acid (GABA) receptors is altered in the brains of subjects with autism, with levels being significantly reduced in autism compared to controls [14–16]. In addition, impaired vasopressin/oxytocin metabolism was also reported in patients with autism compared to controls [17]. Moreover, studies have revealed increased levels of pro-inflammatory cytokines in brain, cerebrospinal fluid (CSF), and blood from children with ASDs [18].

Recently, there has been increasing interest in the role of chronic inflammation in neurological disorders. The progressive increase of clinical and experimental research into inflammation has been initiated by knowledge that inflammation-induced changes in brain disorders are not limited to specific neurotransmitter abnormalities but reflect multifunctional changes in the oxidative/antioxidant status, and immune, endocrine, and neurotransmitter circuits in the brain [19]. Neuroinflammation was first identified in a study of post-mortem samples from 11 individuals with autism aged 5–44, in which activated astrocytes and microglial cells together with abnormal inflammatory cytokines were found [20, 21]. Additional support for the presence of tissue neuroinflammatory-based changes in brains of people with autism comes from the various studies showing reduction in Purkinje cell numbers, possibly due to inflammation-induced oxidative stress leading to an increased excitation/inhibition ratio that could potentially be due to imbalanced glutamatergic and GABAergic systems [13, 14]. Although it is accepted that neuroinflammation can alter neurotransmitter function and its transporters [22] producing cognitive, behavioral, and psychiatric impairments, and that excess central nervous system (CNS) glutamate levels can induce neurotoxicity-mediated inflammation [23–26], the interaction between these two pathologic mechanisms have not been clarified in patients with autism.

A number of studies have shown that inflammatory cytokines, including tumor necrosis factor (TNF)-α, interferon (IFN)-γ, IL6, IL8, and IL12, are elevated in blood mononuclear cells, plasma, serum, CSF, and brain of autistic subjects [27–30]. Among these cytokines, IFN-γ orchestrates the trafficking of specific immune cells to sites of inflammation through up-regulating protein mediators’ expression such as IFN-γ-induced protein-16 (IFI16). Additionally, TNF-α synergistically regulates the expression of these in adhesion and chemokine molecules [31].

This work is an attempt to understand the interaction between GABA, serotonin, dopamine, and oxytocin as important neurotransmitters and IFI16 as an inflammatory mediator that has not been extensively studied in ASD compared to other cytokines. Moreover, a biochemical correlate to the social responsiveness scale (SRS) and the Childhood Autism Rating Scale (CARS) as measures of social and cognitive impairments in patients with autism was screened among the measured parameters.

Plasma as a complex body fluid containing proteins, peptides, lipids and metabolites reflects the physiological activity and pathology in various organs, including the CNS. In humans, about 500 mL of CSF is absorbed into the blood daily, making blood a suitable source of neurodegenerative or neurodevelopmental disease biomarkers [34].

Although neurotransmitter assessment can be a useful tool in any clinical practice, immune system and nervous system activity must be considered and examined as a single system functioning in parallel. It is well established that neurological and immunological abnormalities exist in autistic individuals; however, the relationship between neural and immune function has not been emphasized enough.

Recently, GABA, as the major inhibitory transmitter in the CNS, has been implicated in the pathophysiology of ASDs [35]. Table 1 and Figure 1A demonstrate the significantly higher level of plasma GABA in patients with autism compared to controls with a possibility of concomitant low brain GABA levels. This suggestion is supported by many previous studies in which GABAergic inhibitory problems, reduced GABAA-receptor binding in hippocampus, neocortex and cerebellum, and protein levels of several GABA receptor subunits have been reported in autism [36]. The association between elevated plasma GABA and severity of autism (CARS and SRS) can be easily seen in Table 1. Severe autistics show higher plasma GABA levels compared to mild-moderate patients. This could be supported by the previous studies which show that male mice that have been socially isolated for more than 4 weeks demonstrate a reduced responsiveness of GABA_A receptors to the administrations of GABA-mimetic drugs [37]. Moreover, knockout mice lacking the GABRB3 gene display seizures, hypersensitive behavior, learning and memory deficits, and poor motor skills as features common to the Angelman syndrome, a disorder with some similarities to autism [38]. Additionally, the association between elevated GABA levels and impaired cognition reported in the present study could be supported by a recent study in which treatment with GABA_A receptor antagonist pentylenetetrazole elicited long-lasting (>1 week) normalization of cognitive function in young and aged mice [39].

As serotonin has a wide range of effects on normal physiological function including circadian rhythms, appetite, mood, sleep, anxiety, cognition, and memory formation, its impaired level could be easily related to many autistic features such as sleep disturbances and social and cognition impairment [40]. In the present study, platelet-free plasma serotonin levels were significantly lower in patients with autism compared to control subjects (Table 1 and Figure 1B). This does not necessarily contradict with the hyperserotonemia previously reported in whole blood or platelet-rich plasma of patients with autism. The observed lower level can be explained on the basis that serotonin was measured in platelet-free plasma as the major source of blood serotonin. The obtained results are in good agreement with those of Spivak et al. [41] and Mostafa and Al-Ayadhi [42], who reported low levels of serotonin in platelet poor plasma and serum, respectively, in patients with autism compared to age-matched controls. Based on the fact that 99% of circulating serotonin is contained in platelets as a result of the function of the serotonin transporters [43], the hyperserotonemia previously reported in autistic whole blood or platelet-rich plasma and the recorded low serotonin in platelet-free plasma used in the present study, could be the result of i) increased serotonin production, or ii) increased expression of the transporter on platelet surface, or both. This is consistent with the previous work of Marazziti et al. [44], in which a significantly higher density of serotonin transporters per binding site was recorded in autistic compared to control subjects. In the present study, the lack of biochemical correlation between low serotonin levels in platelet-free plasma and severity of autism (SRS and CARS), could ascertain the role of platelet-related hyperserotonemia in the pathogenicity of this disorder [45]. This suggestion is in agreement with the previous work of Kuperman et al. [46], who recorded that although there was no significant correlation between platelet-rich serotonin concentration and autism behavior checklist scores, four individual items on the checklist that reflect behavioral abnormality were significantly associated with serotonin concentration. The relationship between serotonin levels and the behavioral symptoms in ASDs is of particular interest because it focuses on symptoms (e.g., repetitive and sensory circumscribed interests) that may respond to serotonergic medications.

Dopamine is a catecholamine synthesized from the dietary amino acid tyrosine. Once released from the neuron, central dopamine is broken down into homovanillic acid (HVA) and 3,4-dihydroxyphenylacetic acid. It is well documented that the dopaminergic system has been associated with speech and communication skills which are impaired in autism [47]. Approximately 50% of subjects with autism exhibited significantly elevated levels of HVA in CSF [48]. In the present study, Table 1 and Figure 1C demonstrate decreased plasma dopamine levels in patients with autism compared to control subjects. This is not consistent with previous studies in which assessment of circulating dopamine suggest possible elevations of whole-blood or plasma dopamine levels in children with ASD [49], with no consistent differences in the level of the dopamine metabolite HVA [50]. While increased levels have been associated with increased motoric and tic-like behavior, decreased dopamine levels have been linked to impaired communication and information processing [51]. The low level of plasma dopamine observed in the present study could be due to increased urinary excretion of HVA observed in children with autism [9]. Low plasma dopamine could be attributed to a remarkably higher rate of dopamine breakdown in Saudi individuals with autism compared to control subjects. The controversies between the present work and the previous work of El-Ansary et al. [12] on the same Saudi population could be solved and accepted through considering the pharmacologic effects of dopamine antagonistic drugs. While haloperidol and risperidone, which are used to treat aggressive symptoms in individuals with autistic aggressive symptoms and behaviors, were found to induce hyperactivity in patients with autism [52], administration of secretin (a peptide hormone) is associated with increases in CSFHVA levels and improvements in communication and reciprocal social interaction patterns [53]. This could help to conclude that alteration of dopamine homeostasis may confer risk of autism and this could, in turn, explain the absence of association between dopamine levels and the severity of SRS and CARS recorded in the present study (Table 1 and Figure 1C); in other words, either elevated or decreased levels are involved in the etiology of autism. This can be attributed to the differences in the distribution and abundance of dopamine receptors, reuptake transporters, and auto-receptors in the basal ganglia and prefrontal cortex involved in cognition function. A working hypothesis states that dopamine in these two brain areas regulates the balance between two functionally opponent processes (stability versus flexibility) and proposes that there is neurochemical reciprocity between dopamine in the prefrontal cortex and dopamine in the basal ganglia, with increases and decreases in prefrontal dopamine being associated, respectively, with decreases and increases in terms of dopamine in the basal ganglia [54, 55]. This suggestion is consistent with a most recent study in which de novo missense mutation in the human dopamine transporter gene is identified as a risk factor for autism. The dopamine transporter is a presynaptic membrane protein that regulates dopaminergic tone in the CNS by mediating the high-affinity reuptake of synaptically released dopamine, making it a crucial regulator of dopamine homeostasis [55].

Another explanation which can be related to the role of dopamine in early brain development and etiology of autism is the conversion of dopamine into norepinephrine, which is catalyzed by dopamine β-hydroxylase [56] an enzyme controlled largely by a single gene, DβH. Robinson et al. [57] suggested that lower levels of DβH activity results in prenatal conditions that contribute to autism. Although they could not identify a mechanism, the authors raise the possibility that prenatal exposure to excessive dopamine levels could cause later down-regulation of dopamine production or dopamine sensitivity, with lasting effects throughout development. This could explain the reported low level of dopamine in the Saudi individuals with autism who participated in the present study.

Given that oxytocin is involved in the regulation of repetitive and affiliative behaviors, and that these are key features of autism, it is believed that oxytocin may play a role in ASD. Additionally, oxytocin is released into blood and within distinct brain regions in response to stressful and social stimuli, and has been shown to have an antidepressant-like effect in animal studies. Clinical reports suggest oxytocin to be a promising drug for psychiatric diseases such as depression, anxiety disorders, schizophrenia, and ASD. Oxytocin may also have therapeutic potential in the treatment of major depressive disorders, even though oxytocin administered into blood does not readily cross the blood-brain barrier [58].

Table 1 and Figure 1D show that autistic children have lower plasma oxytocin compared to age- and gender-matched control subjects. This is in agreement with several previous studies which demonstrate remarkable lower oxytocin levels in autistics compared to control subjects and suggest that oxytocin may play a role in the symptoms of ASDs [59–61]. Oxytocin has been implicated in the regulation of behavior in animals, but has not yet been examined in depth in autistic children. A profound impairment in social recognition in oxytocin receptor-knockout mice has been shown, indicating an important role played by these peptides in social and affective disorders, including autism and anxiety disorders [62]. The animals behaved normally, except they could not learn to recognize other mice or recognize their mother’s smell, though their sense of smell was normal. A single dose of oxytocin into the brain, however, cured the mice. This information could support the significant positive correlation recorded in the present study between lower levels of oxytocin in Saudi autistic subjects and the severity of the SRS score. This correlation is consistent with the previous work of Hoge et al. [63] and El-Masry et al. [64], which provide preliminary support for a link between social anxiety severity and plasma oxytocin. In a review by Lukas and Neumann [62], the regulatory capacity of oxytocin to modulate social behaviors in various rodent species implies a high translational potential. It also suggests that the brain oxytocin and arginine vasopressin systems are promising pharmaco therapeutic targets to improve social behavior and to reverse social deficits. Genetic evidence also supports the role of oxytocin in the pathophysiology of autism reported in the present study and the correlation presented between oxytocin and severity of autism. In a study by Yrigollen et al. [65], a link was found between the autism phenotype and both the oxytocin gene (OXT, located at 20p13), and the oxytocin receptor gene (OXTR, located at 3p26).

IFNs are a family of cytokines [66] which includes Type I (IFN-α and IFN-β) and Type II (IFN-γ) IFNs, among others [67, 68]. Binding of an IFN (α, β, or γ) to the corresponding cell surface receptor results in activation of a family of tyrosine kinases and activation of signal transducer and activator of transcription (STAT) proteins [67]. Importantly, transcriptional activation of IFN stimulated genes by the activated STATs results in induction of IFN-inducible proteins such as IFI16. Recently, the role of IFI16 in recruiting inflammatory monocytes after Herpes simplex virus type 1 was ascertained [68].

Table 1 and Figure 1E demonstrate the significant increased levels of plasma IFI16 in children with autism compared to controls and significant correlation between high levels of IFI16 and severity of CARS and SRS. This means that concentrations of IFI16 in plasma of severe autistic subjects in CARS and SRS were remarkably higher than mild-moderate patients in both scores. This is consistent with the previous work of Garbett et al. [69], in which expression profiling of the superior temporal gyrus of six autistic subjects and matched controls revealed increased transcript levels of many immune system-related genes; IFI16 was one of the most significantly elevated transcripts. Overall, these expression patterns were found to be associated with the late recovery phase of autoimmune brain disorders. The remarkable elevation of IFI16 in Saudi individuals with autism as subjects of the present study could be supported through consideration of the recent work of El-Ansary and Al-Ayadhi [30], in which they recorded high significant increases of IFN-γ as inducer to IFI16. In addition, elevation of IFI16 in patients with autism compared to control subjects could also ascertain the role of autoimmunity in the pathophysiology of autism as, interestingly, anti-IFI16 antibodies are present in autoimmune diseases such as systemic lupus erythematosus. Hence, IFI16 may have a role not only in anti-viral responses, but also in pathogenic autoimmune responses [70].

Recently, the pro-apoptotic activity of IFI16 was reported by Gugliesi et al. [71]. They proved that IFI16 contributes in the regulation of apoptosis triggered by IFN Type I and that overexpression of IFI16 in primary endothelial cells is sufficient to induce apoptosis through the simultaneous activation of caspase 2 and caspase 3, which in turn is mediated by the NF-kB complex, since the inactivation of NF-kB was found to prevent IFI16-induced apoptosis. This information, together with the recorded elevation of IFI16 in patients with autism compared to control subjects, could be supported by the recent work of El-Ansary et al. [28] and El-Ansary Al-Ayadhi [30], in which they recorded elevation of caspase 3 and caspase 7 as pro-apoptotic markers in plasma of Saudi autistic children. The proposed relationship between NF-kB and the proinflammatory and proapoptotic effects of IFI16 could be supported by the work of Lam et al. [72], in which they suggested that obstetric complications in the form of perinatal hypoxia may lead to inflammation and apoptosis, and may contribute to an increased risk of neurodevelopmental disorders such as autism through increased expression of transcription factors, NF-κB, inflammatory mediators, IL-6, and the apoptotic protein caspase3. In contrast, this proposed relationship is not consistent with the recent finding of Malik et al. [73], which suggests that the NF-κB signaling pathway is not dysregulated in the brain of autistic subjects and thus may not be significantly involved in the processes of abnormal inflammatory responses suggested in the autistic brain.

Based on the fact that the predictiveness curve is better if it is farther away from the prevalence line and useless if it is close to the prevalence line, the predictiveness curves of the four measured parameters (Figure 2A–E), varies significantly from the baseline risk depending on whether GABA, serotonin, dopamine, oxytocin, and IFI16 concentrations were low or very high. This shows their usefulness as predictive biomarkers. This is supported by the high sensitivity and specificity recorded through ROC analysis (Table 2).

In conclusion, based on the obtained and discussed data, this study highlights the role of the investigated parameters representing inflammation, autoimmunity, and neurochemistry signaling pathways in the etiology and severity of ASD as an increasingly prevalent disorder. GABA, oxytocin, and IFI16 can be used as biochemical correlates to social and cognitive impairment in autism. In addition, the significant positive correlation between IFI16 and GABA together with the negative correlations between this inflammatory marker and serotonin, dopamine, and oxytocin (Figure 3) can support the role of inflammation in the neurochemical dysregulation recorded in patients with autism. Although initial reports look promising, the clinical application of the recorded biomarkers remains a vision for the future, since crucial questions have to be addressed first before these biomarkers find their way into clinical practice. Until the clinical use of the biomarkers become a reality, this study could help in the early detection of infants flagged at risk for autism and subsequently may likewise benefit from early behavioral intervention, making use of operant conditioning models, i.e., positive and negative reinforcement, to modify undesired behaviors.