Abnormal brain growth and development in autism is suggested by imaging and head circumference studies (Courchesne et al. 2001) and stereological Purkinje cell analysis (Whitney et al. 2008). It is logical to assume that etiology of autism is related to disrupted brain development, and abnormal course of various developmental events. Many of these processes, such as cell proliferation, migration, differentiation and synaptogenesis are dependent on neurotrophins (McAllister et al. 1999).

The term neurotrophin generally refers to six structurally related factors: brain-derived neurotrophic factor (BDNF), nerve growth factor (NGF) , neurotrophin-3 (NT-3) , neurotrophin-4 (NT-4), neurotrophin-5 (NT-5) and neurotrophin-6 (NT-6). During brain development individual cell populations in different brain regions undergo chronological transition from cell proliferation, cell migration, differentiation, synaptogenesis and elimination of cell excess; neurotrophins play important functions in all of these processes (Riikonen and Vanhala 1999; Nelson et al. 2001, 2006; Miyazaki et al. 2004; Connolly et al. 2006; Hashimoto et al. 2006). Moreover, disruption of synapse formation, stabilization and function, observed in autism (Zoghbi 2003), may be brought about by abnormal neurotrophin expression (Zagrebelsky and Korte 2014). Arguments favor the hypothesis that altered expression of neurotrophins early during the development contributes to the autistic pathology (Nelson et al. 2001). Furthermore, altered neurotrophin expression may be brought about by the environmental triggers as discussed below.

Animal data suggest a brain region-specific developmental expression of individual neurotrophins (Das et al. 2001). In rat cerebellum NT-3 is most abundantly expressed during embryonic development with a peak at postnatal day 1 and a subsequent decrease (Das et al. 2001). In monkey cerebellum the level of NT-3 was also higher during embryonic stages and decreased in adulthood, while the level of NT-4/5 increased during both embryonic and postnatal development, gradually declining with age (Takumi et al. 2005). If a similar developmental profile occurs in the human cerebellum, then the persistent elevation in NT-3 could upset the balance of neurotrophic factors and affect brain growth and development. Specifically, over expression of NT-3 could contribute to the cerebellar overgrowth observed in some autistic cases (Courchesne et al. 2001).

Evidence in support of the environmentally-derived disruption of neurotrophin expression comes from several lines of observations such as increased levels of hippocampal BDNF in response to lead (Chao et al. 2007), increased expression of NGF in C6 glioma cells following exposure to polychlorinated biphenyls (PCBs; Gurley et al. 2007), or increased production of NGF by fibroblasts in response to methyl mercury (Söderström and Ebendal 1995).

Recent studies have indicated that maternal infection during fetal brain development may be one of the risk factors for autism. Furthermore, it has been observed that maternal infection during pregnancy results in decreased amniotic fluid levels of NGF and BDNF (Marx et al. 1999). Study of the effect of the Borna disease virus (BDV) infection on the hippocampal neurons showed inhibition of BDNF-induced ERK 1/2 phosphorylation, BDNF-induced expression of synaptic vesicle proteins and severely impaired synaptogenesis and synaptic organization (Hans et al. 2004). Thus viral infection can interfere with neurotrophin-regulated neuroplasticity and neuronal connectivity. More recently, using lipopolysaccharide (LPS)-exposed rat model of infection we have shown elevation in cerebellar NT-3 levels in pups of LPS exposed dams (Xu et al. 2013a). Furthermore using the same model we have observed a significant elevation in BDNF gene expression (Xu et al. 2013b).

Neurotrophins, and specifically, BDNF and NGF are important for the survival, maintenance and regeneration of specific neuronal populations in the adult brain and their depletion has been linked to neurodegenerative diseases such as Parkinson’s, Alzheimer’s disease (AD) and Huntington’s disease. BDNF and NT-3, implicated in neuronal growth and synapse activity, decrease in different brain regions of patients with AD (Corbett et al. 2013).

Studies in transgenic mice indicated that NGF changes induced activation of amyloidogenesis and tau processing and provided link between neurotrophin signaling deficit and AD neurodegeneration (Cattaneo and Callisano 2012). Furthermore, experiments in neuronal culture revealed anti-amyloidogenic action of NGF/TrkA and suggested the importance of homeostatic balance between the neurotrophins and their receptors (Cattaneo and Callisano 2012). The expression of p75NTR receptor is increased in the AD hippocampus and Aβ-bound p75NTR triggers cell death (Armato et al. 2013).

It is important to realize that neurotrophin levels in normal controls are not constant but alter according to genetically determined trajectories. In humans, BDNF increases with age, while NT-3 and NT4/5 concentrations decrease (Nelson et al. 2006). The level of NT-3 decreases with age in the rodent brain (Das et al. 2001). Consequently, if a similar developmental down regulation occurs in the human cerebellum, then prolonged and persistent elevation of NT-3 in autism suggested by our data (Sajdel-Sulkowska et al. 2009, 2011) may have profound consequences on neuronal growth and synapse formation. Furthermore, while in healthy controls serum BDNF increases with age decreasing slightly in adulthood, it is abnormally low in children with autism suggesting a delay in normal age-dependent increase in BDNF (Katoh-Semba et al. 2007).

In addition to the trophic function, NT-3 may also under certain conditions exacerbate oxidative stress (Bates et al. 2002) as a part of a response to hypoxia (Pasarica et al. 2005) affecting survival of Purkinje cells in vitro (Morrison and Mason 1998). In that respect, NT-3 behaves like NO playing an important role in neuronal differentiation by nitrative modification of specific proteins (Cappelletti et al. 2003). However overproduction of nitric oxide leads to peroxynitrite formation as a source of neurotoxicity (Vargas et al. 2004). Increase in peroxynitrite results in nitration of free- and protein-bound tyrosine residues leading to 3-nitrotyrosine (3-NT) formation and protein modification as observed in a number of human pathologies.

It has been suggested that the negative action of NT-3 may be related to altered glutamate receptors (Behrens et al. 1999). Thus it is of interest that glutamatergic system abnormalities (Yip et al. 2007) have been implicated in selective Purkinje cell decrease in autism (Rout and Dhossche 2008).

Interestingly, BDNF has been implicated in mediating several of the effects of estrogen in hippocampus; in turn estrogen regulates BDNF as well as trkB and p75NTR of the mossy fiber pathway with altered hippocampal BDNF levels potentially affecting hippocampal functions (Harte-Hargrove et al. 2013). Estrogen and the neurotrophin ligand-receptor complexes have been implicated in etiology of diverse types of neurological or psychiatric disorders (Harte-Hargrove et al. 2013). This estrogen-BDNF interaction may contribute to the sex-dependent differences in brain development.

Most of the data on neurotrophin levels in autism are derived from the analysis of blood levels (Kozlovskaia et al. 2000; Nelson et al. 2001, 2006; Miyazaki et al. 2004; Tsai 2005; Connolly et al. 2006; Nelson et al. 2006; Katoh-Semba et al. 2007; Croen et al. 2008; Tostes et al. 2012). In this context, it is important to realize that these do not necessarily reflect the brain neurotrophin levels because of the blood–brain-barrier. A few of the results derived from the direct measurement of the brain neurotrophin levels suggest an independent regulation of the two pools.

The hypothesis of early hyperactivity of BDNF in autism (Tsai 2005) is supported by increased levels of BDNF in the serum (Nelson et al. 2001) and may be associated with early brain outgrowth (Courchesne et al. 2001). BDNF levels were also elevated in newborn sera of children with autism (Connolly et al. 2006), but so were elevated autoantibodies suggestive of interaction between immune system and BDNF (Connolly et al. 2006). Others observed increased blood BDNF levels in young adults with autism (Miyazaki et al. 2004). Furthermore, elevated BDNF levels were also observed in basal forebrain of young adults with autism (Perry et al. 2001); NGF levels remained normal and NT-3 levels were not measured in this study. However, in adult male patients with autism the levels of BDNF were reduced as compared to normal controls (Hashimoto et al. 2006).

In healthy controls the serum BDNF concentrations increased over the first several years, then decreased after reaching the adult level. In autism, the serum levels of BDNF were lower in children 0–9 years compared to teenagers and adults indicating a delayed increase of this neurotrophin with development (Katoh-Semba et al. 2007). On the other hand population-based case control studies of archived maternal mid-pregnancy and neonatal blood did not show changes in BDNF levels in autism (Croen et al. 2008). BDNF expression in the peripheral blood lymphocytes of the drug naïve autism patients was found to be significantly higher than in the control group suggestive of a possible pathologic role of BDNF on the serotonergic system (Nishimura et al. 2007). Additionally, an increased level of BDNF in the basal forebrain in autism has also been reported (Nelson et al. 2006).

An association between BDNF gene polymorphism and autism was found in Chinese population (Cheng et al. 2009). More recently we have observed abnormal expression of brain derived neurotrophic factor (BDNF) gene in autism (Khan et al. 2014). Study of rs6265(BDNF) as a genetic marker of anxiety, ADHD and tics found BDNF genotype marginally significant for social phobia in children with ASD (Gadow et al. 2009)

Elevation of blood serum levels of NGF in early autism was reported (Kozlovskaia et al. 2000), but the analysis of CSF showed normal levels of NGF in children with autism (Riikonen and Vanhala 1999). The analysis of basal forebrain in autism showed no changes in NGF in teenage cases (Nelson et al. 2006).

It has been previously reported that blood NT-3 levels were significantly lower in autistic neonates (Nelson et al. 2006); others supported this observation and reported plasma levels of NT-3 that were significantly lower in children with autism (Tostes et al. 2012). NT-3 is specifically involved in neuronal differentiation (Ghosh and Greenberg 1995) neurite fasciculation (Segal et al. 1995) and axonal targeting. In the prenatal system, overexpression of NT-3 affects the formation of specific synapses. Our data indicating increased NT-3 expression in brain tissue derived from a subset of autistic cases including older donors suggest an abnormal synapse formation in autism (Sajdel-Sulkowska et al. 2009). Specifically, we reported an increase in NT-3 in cerebellar hemispheres, dorsolateral prefrontal cortex, Wernicke’s area and cingulate gyrus suggesting brain region specific NT-3 changes (Sajdel-Sulkowska et al. 2011).

Neurotrophin 4/5 (NT-4/5) were also elevated in autism; furthermore that elevation was observed in peripheral blood in the first days of life (Nelson et al. 2001).

Thus several lines of evidence point out to altered neurotrophin levels in autism; complementary changes in related trophic factors have also been recently reviewed (Jockschat and Miche 2011).

Consistent neuropathological findings in autism include an early increase in brain size (Courchesne et al. 2001; Sparks et al. 2002), decrease in Purkinje cells, smaller neuronal size and decreased dendritic branching in the cerebellum (Bailey et al. 1998), hippocampus and amygdala (Kemper and Bauman 1998; 2002) and reduced neuronal and dendritic pruning (Ben Bashat et al. 2007; Palmen et al. 2005). Additionally, reduced connectivity in autistic brain regions is associated with impaired social cognition as documented by diffusion tensor imaging of white matter structure (Barnea-Goraly et al. 2004). Decrease in executive function in autism has been suggested by functional magnetic resonance imaging studies (Dawson et al. 2002). However, the link between neurotrophin abnormalities and autistic symptoms, is supported by a very small number of animal studies and human observations.

The hypothesis that early BDNF hyperactivity may play a role in autism etiology is supported by an association between the early increase in BDNF levels in both serum and brain tissue in autistic children and early brain outgrowth in autism (Tsai 2005).

We reported an increase in NT-3 in cerebellar hemispheres, dorsolateral prefrontal cortex, Wernicke’s area and cingulate gyrus suggesting brain region specific NT-3 changes in autism (Sajdel-Sulkowska et al. 2011). Significantly, these particular brain regions are functionally altered in autism.

Importantly, variants of BDNF genes are linked to the decline in executive function with aging (Erickson et al. 2008; Raz et al. 2009) and animal data suggest that this effect may be mediated through a BDNF effect on synaptic plasticity in the prefrontal cortex (Sakata et al. 2009).

Further neurotrophin-behavioral link is provided by a transgenic mouse model deficient in the BDNF that exhibited diminished brain circuitry and selective deficit in social- and anxiety-related behaviors (Sadakata et al. 2012). Independently, a knockout mouse model of fragile X syndrome with a reduction in BDNF expression showed a deficit in cognitive skills (Uutela et al. 2012). However, at this point, the link between neurotrophin abnormalities and autistic pathology awaits further clinical and experimental confirmation.

Neurotrophins are synthesized as pro-neurotrophins and are then converted to the active neurotrophins . Both forms interact with the neurotrophin receptors, p75 and a member of the tyrosine kinase family (Trk), located on the target cells. The interaction of neurotrophins with the p75 receptor is of low affinity type and nonspecific. Binding to the Trk receptors is of high affinity type and specific for individual neurotrophins; BDNF binds to TrkA, B, and C, NT-3 binds to TrkC and NT-4/5 binds to TrkB receptors (Jockschat and Miche 2011).