Asymmetric development of the nervous system

Abstract

The human nervous system consists of seemingly symmetric left and right halves. However, closer observation of the brain reveals anatomical and functional lateralization. Defects in brain asymmetry correlate with several neurological disorders, yet our understanding of the mechanisms used to establish lateralization in the human central nervous system is extremely limited. Here, we review left-right asymmetries within the nervous system of humans and several model organisms, including rodents, zebrafish, chickens, Xenopus, Drosophila, and the nematode Caenorhabditis elegans. Comparing and contrasting mechanisms used to develop left-right asymmetry in the nervous system can provide insight into how the human brain is lateralized.

Introduction

The central nervous system, consisting of the brain and spinal cord, is one of the most complex and intricate systems of the human body, and is responsible for all our perceptions, thoughts, and behaviors. The human brain is divided into left and right hemispheres, connected by a large bundle of nerve fibers called the corpus callosum (Choudhury et al., 1965) (Fig. 1). The hemispheres of the brain control the contralateral sides of the body (Fig. 1). The left and right hemispheres of the human brain show subtle, yet striking, anatomical asymmetries and functional lateralization (Gazzaniga et al., 1965; Gazzaniga and Sperry, 1967; Geschwind and Levitsky, 1968; Gazzaniga and Smylie, 1983; Van Essen, 2005; Illingworth and Bishop, 2009). Several studies have implicated anatomical brain asymmetry in cognitive performance and social behavior. Defects or lack of asymmetry of the brain has been associated with several neurological diseases, including dyslexia, schizophrenia, bipolar disorder, and autism (Falkai et al., 1992; Hugdahl et al., 1998; Klar, 1999; Herbert et al., 2005; Ravichandran et al., 2017). However, our understanding of how human brain asymmetry is established is extremely limited. Studies of human brain asymmetry have been restricted to non-invasive and observational methods.

Figure 1. Laterality in the human nervous system.

The left and right hemispheres of the human brain are connected by the corpus callosum. The hemispheres display functional differences and control contralateral sides of the human body. Broca’s and Wernicke’s areas are language centers located in the left hemisphere of the majority of individuals. L, left; R, right; A, anterior; P, posterior.

Central nervous systems are found in all vertebrates, as well as in some invertebrates. The brain consists of neurons and glia with synaptic connections and functions that are evolutionarily conserved among species. Thus, knowledge gained from model organisms may provide insight into our own brain development and function. In this review, we discuss our current knowledge of laterality in the human nervous system. Additionally, we discuss studies on brain asymmetries in other model organisms, including rodents, zebrafish, chickens, Xenopus, Drosophila, and Caenorhabditis elegans (C. elegans). We delve into two types of asymmetries: directional and stochastic. Directional asymmetry occurs when features are stereotypically located on either the left or right side of the body across a population. In contrast, stochastic asymmetry exhibits attributes that are randomly assigned to either the left or right side. Together, the studies from these various model organisms will greatly enhance our understanding of how the human nervous system is lateralized.

Laterality in the human nervous system

Structural and functional asymmetry in the human brain has been observed since the 1860s, with Paul Broca’s observation of a lesion in the left hemisphere in an autopsy of a patient who had a language disability (Broca, 1865). The patient was only able to communicate a single word. The location of his lesion in the frontal lobe of the dominant hemisphere has since been deemed Broca’s area, responsible for the production of speech (Fig. 1). In 1874, Karl Wernicke observed a lesion at an area, now referred to as Wernicke’s area, in the left hemisphere of a patient unable to comprehend language (Fig. 1) (Wernicke, 1874). However, people with damage to these areas on the right side usually did not have any language problems, indicating functional lateralization of language comprehension (Broca, 1865).

Much of what we know about the left and right hemispheres comes from split-brain experiments, performed by Roger Sperry and Michael Gazzaniga in the 1960s, in patients who suffered from epilepsy and had the corpus callosum cut to prevent the spread of seizures from one hemisphere to the other (Gazzaniga et al., 1965; Gazzaniga and Sperry, 1967). In the majority of people, the left hemisphere may be more dominant for calculations, math, and logical abilities, in addition to language production and comprehension skills, while the right hemisphere is dominant for visual-spatial and facial recognition (Gazzaniga and Smylie, 1983; Riss, 1984) (Fig. 1). Here, we discuss studies of asymmetry in the human nervous system related to language and handedness.

Asymmetry of language in the cerebral cortex

With the advent of modern techniques in neuroimaging, studies have confirmed Broca’s and Wernicke’s observations that language is lateralized to the left hemisphere of the brain in the majority of the human population. Over 90% of right-handed individuals show language localization to the left hemisphere, as do over 70% of left-handed people (Knecht et al., 2000). Research into the differences between the two halves of the brain has focused on studying anatomical differences. Studies have found that structural cerebral asymmetry may correlate with language. The planum temporale, a language center located within the Sylvian fissure that forms Wernicke’s area, has been reported to be larger in the left hemisphere in the majority of people (Geschwind and Levitsky, 1968). In addition, imaging of the Sylvian fissure, which separates the frontal and parietal lobes from the temporal lobe, was shown to have greater depth in the left hemisphere than the right hemisphere (Van Essen, 2005), whereas the right hemisphere has a sharper slope as compared to the left (Rubens et al., 1976). Another study showed that dyslexic patients display less asymmetry in the cerebrum than individuals without dyslexia, as measured by functional transcranial Doppler ultrasound (Illingworth and Bishop, 2009). However, many imaging studies on the asymmetry of the brain reported conflicting results, which may be a result of differences in measurement criteria. It is unclear whether disruption of asymmetry causes language disorders, if a common genetic trait leads to both language disorders and impaired cerebral lateralization, or if perhaps language disabilities result in reduced brain asymmetry (Eckert and Leonard, 2000; Bishop, 2013).

Conflicting studies have emerged using microarrays to determine whether gene expression is lateralized in the human cortex, with two studies reporting no asymmetric expression (Hawrylycz et al., 2012; Pletikos et al., 2014). However, another report reanalyzed the data from these two studies, focusing on cerebral regions implicated in language: the superior temporal gyrus and Heschl’s gyrus (Karlebach and Francks, 2015). The study found lateralization of gene sets that include those for cell surface receptors, G protein-coupled receptors, calcium-binding protein, dopamine receptors, serotonin receptors, and voltage-gated channels (Karlebach and Francks, 2015).

Handedness as a result of laterality in the nervous system

Over 90% of the human population is more skilled with the right hand, which is controlled by the left side of the brain (Sun and Walsh, 2006) (Fig. 1). Hand preference was found in developing embryos and fetuses, as thumb sucking with a preferred hand was observed before birth (Hepper et al., 1998). As with research of asymmetric function of language, imaging studies have been carried out to investigate any structural differences between the left and right hemispheres of the brain that may correlate with handedness. An MRI study found that the area of the cortex that controls the right hand was larger than that of the left hand in right-handed individuals, and vice versa (Soros et al., 1999). In addition, another imaging study using magnetic resonance morphometry concluded that right-handed people have a deeper left central sulcus, which divides the frontal and parietal lobes, than left-handed people, and vice versa (Amunts et al., 1996). However, other imaging studies report no brain asymmetries observed in relation to handedness (Good et al., 2001).

Right-handedness highly correlates with brain laterality, as 97% of right-handed individuals exhibit left hemisphere dominance (Klar, 2004). 91.6% of individuals with right-handedness have clockwise hair whorl, but in left-handed people, the direction of the hair whorl is randomized (Klar, 2004). These correlations suggest that handedness and certain anatomical traits may be linked.

A few genetic studies have been performed to determine the molecular mechanisms used to establish preferred hand use. A genome-wide linkage scan identified a haplotype upstream of the leucine-rich repeat transmembrane neuronal 1 (LRRTM1) that is associated with handedness in dyslexic siblings, while no association was observed between handedness and reading ability in this study (Francks et al., 2007). In addition, methylation of the LRRTM1 promoter is associated with mixed-handedness (Leach et al., 2014). Furthermore, a genome-wide associated study implicated the proprotein convertase/subtilisin/kexin type 6 (PCSK6) gene in genetic basis for handedness in dyslexic individuals (Scerri et al., 2011). Other molecules are also implicated in hand preference. The transcription factor LIM domain only 4 (LMO4) is asymmetrically expressed in the right perisylvian cortex of human fetuses (Sun et al., 2005). Individual difference in CAG-repeat lengths of the androgen receptor (AR) gene was associated with variation in handedness (Medland et al., 2005; Hampson and Sankar, 2012; Arning et al., 2015).

A recent study hypothesized that handedness develops in the fetal spinal cord, as opposed to the cerebral cortex, since fetuses were found to display arm preference prior to linking of the motor cortex with the spinal cord, at 8 weeks post conception (Ocklenburg et al., 2017). Thus, this study examined asymmetric mRNA expression, miRNA expression, and DNA methylation in different segments of the spinal cord in human fetuses. The left spinal cord showed higher expression of brain-derived neurotrophic factor (BDNF), platelet-derived growth factor (PDGF), collagen, and RNA polymerase complex (Ocklenburg et al., 2017). Genes asymmetrically expressed in the right spinal cord included forkhead box P2 (FoxP2) as well as those involved in cell cycle, cellular component biogenesis, transferase activity, transcription factor activity, system development, reproductive process, and cell proliferation (Ocklenburg et al., 2017). In addition, miRNAs, including those involved in the TGF-β signaling pathway, were also found to be asymmetrically expressed. Furthermore, asymmetric DNA methylation was observed in the fetal spinal cord (Ocklenburg et al., 2017). Although genetics is likely to play a role in development of asymmetry in the human nervous system, differences in hand preference in monozygotic twins suggest additional non-genetic mechanisms of laterality (Gurd et al., 2006).

Other reported asymmetries in the human nervous system

Other functions are reportedly asymmetrically localized in the human nervous system, including attention, memory, and processing of emotions (Habib et al., 2003; Demaree et al., 2005; Duecker et al., 2013). In addition, some neurological disorders, such as migraine headaches, Rasmussen encephalitis, and cerebral palsy, randomly affect one side of the brain, suggesting asymmetry of variable brain functions (Ophoff et al., 1996; Bien et al., 2005; Gale et al., 2015).

Research has been carried out to investigate genes that may be responsible for human cerebral asymmetry. It was found that certain genes are asymmetrically expressed in the human fetal cortex, including genes used for signal transduction and regulation of gene expression (Sun et al., 2006). It is thought that gender and hormones can also influence asymmetry of different parts of the human cerebral cortex (Shaywitz et al., 1995; Wisniewski, 1998).

Laterality in the rodent nervous system

Structural and functional asymmetry has been reported in the nervous system in rodents. The rat has been shown to have a thicker cerebral cortex and hippocampus in the right hemisphere than the left in males (Diamond et al., 1983). Recent studies have started to elucidate the molecular basis of brain asymmetry in rodents.

Kawakami et al. reported that N-methyl D aspartate (NMDA) GluRε2 (NR2B) receptor subunits are asymmetrically distributed between the left and right hippocampus in the adult mouse (Kawakami et al., 2003). The asymmetrical distribution of ε2 subunits seems to result in differential properties of NMDA receptors and synaptic plasticity between the left and right hippocampus. These findings suggest that brain asymmetry can be also established at microscopic levels of neurons and synapses in addition to macroscopic levels of hemispheres (Kawakami et al., 2003).

Rodents also display an equivalent of human handedness, as they prefer to use a particular paw for tasks (Glick and Ross, 1981; Waters and Denenberg, 1994). A recent study has begun to elucidate the role a gene may play in establishing paw preference. As in fetal humans, asymmetric expression of the transcription factor LMO4 was also observed during brain development of the mouse strain tested (Sun et al., 2005). This asymmetry appeared to be random, as the expression area of Lmo4 in the anterior cortex was larger in the left or right hemisphere (Sun et al., 2005). Knockdown of Lmo4 in the right hemisphere of mice resulted in reduced neurogenesis and axonal projection in the right hemisphere as compared to the left (Li et al., 2013). Mice with knockdown of Lmo4 in the right hemisphere also showed altered paw preference. While control mice used their left and right front paws equally and swam either clockwise or counterclockwise equally, mice with Lmo4 knocked down in the right hemisphere preferred to use their right paw and preferred to swim in a counterclockwise manner (Li et al., 2013). Similarly, knockdown of Lmo4 in the left hemisphere resulted in mice preferring to use their left paws (Li et al., 2013). This study may suggest a role for the transcription factor LMO4 in regulating paw preference in mice. However, unilateral knockdown of several factors may result in laterality defects, and may not indicate a direct role in development of asymmetry.

Asymmetric development of the nervous system in zebrafish

The epithalamus of the zebrafish is responsible for the animal’s fight or flight response, as well as for sensing light and odors in the larval stage. It consists of the pineal complex and bilateral habenular nuclei (habenula), and develops prominent left-right asymmetry at both anatomical and functional levels (Butler and Hodos, 1996; Liang et al., 2000; Concha and Wilson, 2001; Sagasti, 2007; Taylor et al., 2010). The pineal complex is made up of a medial pineal organ and a left-lateral parapineal organ (Fig. 2A) (Liang et al., 2000; Concha and Wilson, 2001). In addition, the subnuclei of the habenula show asymmetry in size, connectivity, and gene expression (Concha and Wilson, 2001). The habenula on each side is subdivided into medial and lateral subnuclei. The right medial subnucleus is larger than the right lateral subnucleus, while the left lateral subnucleus is bigger than the left medial subnucleus (Fig. 2A).

Figure 2. Lateralization of the nervous system in zebrafish and chicken.

(A) Asymmetry of the epithalamus and lateralization of eye use in zebrafish. Nodal from the lateral plate mesoderm induces its own expression in the left habenula. FGF and Nodal act together to ensure directional asymmetry of the epithalamus. L, left; R, right; A, anterior; P, posterior; LPM, lateral plate mesoderm; Hb, habenula; LHb, lateral habenula (green); MHb, medial habenula (blue); Po, pineal organ (purple); Pp, parapineal organ (purple).

(B) Lateralization of eye use in chicken. L, left; R, right.

Nodal and FGF signals act together to ensure directional asymmetry of the epithalamus. Initially, Nodal expression in the lateral plate mesoderm provides a directional cue that activates its own expression in the left epithalamus (Fig. 2A) (Roussigne et al., 2009). Then, Fgf8 signals in habenular precursor cells attract the parapineal to the left, which further elaborates asymmetric development of the habenula (Regan et al., 2009). In addition, the T-box transcription factor TBX2b is required for asymmetric placement of the parapineal on the left side (Snelson et al., 2008). Mutations that block the Nodal signaling pathway randomize epithalamic asymmetry (Gamse et al., 2003; Facchin et al., 2009). Furthermore, absent or bilateral expression of Nodal reduces the number of the left habenular precursor cells, suggesting that differential levels of Nodal between the left and right sides are required for asymmetric enlargement of the left habenula (Roussigne et al., 2009). Thus, coordinated communication between the two sides of the epithalamus is required for proper brain asymmetry. Lateral inhibition by the Notch pathway may play a role in this communication process, as hyperactivation of Notch signaling results in symmetric neurogenesis independent of Nodal signaling (Aizawa et al., 2007).

The habenula of larval zebrafish is responsible for sensory responses to both odor and light. Most sensory neurons responding to olfactory stimuli are located on the right habenula, while those responding to visual stimuli are lateralized on the left (Dreosti et al., 2014). 2–5% of zebrafish have spontaneous reversal of epithalamic asymmetries (Concha et al., 2000; Liang et al., 2000). Animals with reversed brain asymmetry show reversal of lateralized functions of the habenular neurons. In addition, double-right-brained zebrafish lose response to visual stimuli and double-left-brained zebrafish lose response to olfactory stimuli (Dreosti et al., 2014). Light preference in larval zebrafish is mediated by the left dorsal habenula. The habenula receives visual inputs from an asymmetrical pathway consisting of a subset of bilateral retinal ganglion cells (RGC) and eminentia thalami (Zhang et al., 2017). Neurons in eminentia thalami receive projections from a subset of RGCs from both eyes, and preferentially innervate the left dorsal habenula (Zhang et al., 2017).

The zebrafish dorsal habenula exhibits pronounced asymmetry in its connectivity with the interpeduncular nucleus. Genetic inactivation of the lateral subnucleus of the dorsal habenula resulted in behaviors biased towards freezing rather than the normal flight response to a conditioned fear stimulus (Agetsuma et al., 2010). This suggests that the asymmetric dorsal habenula-interpeduncular nucleus pathway may be important for the modulation of fear behaviors. Additionally, when the habenula asymmetry is either reversed or symmetrical, the zebrafish shows an increase in fear and anxiety (Facchin et al., 2015). The zebrafish with abnormal right parapineal position was bolder in predator inspection compared to wild-type zebrafish with the parapineal organ on the left side (Dadda et al., 2010). Furthermore, social conflict resolution in zebrafish is mediated by lateral and medial subregions of the dorsal habenula that antagonistically control the outcome of conflict. Silencing the medial dorsal habenula increased the individual chance of winning a fight, while silencing the lateral dorsal habenula reduced the chance of winning (Chou et al., 2016). Taken together, these asymmetric organs antagonistically control the threshold of fight versus flight. Defects in asymmetry of the dorsal habenula alter fight or flight behavior, depending on the stimuli.

Eye usage laterality in zebrafish, chicken, and Xenopus

Eye use has been shown to be lateralized in zebrafish, where the left eye is used to determine the novelty of the environment, including reflection of self, and the right eye is associated with biting (Fig. 2A) (Miklosi et al., 1997; Miklosi and Andrew, 1999; Sovrano and Andrew, 2006). Similarly, eye use dominance has been observed in domestic chickens for attack, social cognition, and sexual behaviors (Fig. 2B) (Rogers et al., 1985; Dharmaretnam and Andrew, 1994; Andrew et al., 2000; Vallortigara et al., 2001; Daisley et al., 2009).

In both zebrafish and chickens, the lateralized behaviors of eye usage are modulated by light (Rogers, 2008; Budaev and Andrew, 2009). In chickens, the right eye is exposed to light in ovo during development, which is responsible for the right eye dominance to select and sustain an approach to a target (Andrew et al., 2000). If the left eye is exposed, the asymmetry is reversed (Budaev and Andrew, 2009). Similarly, light exposure is required for asymmetric eye usage in zebrafish, as asymmetries are abolished if the zebrafish develops in the dark (Budaev and Andrew, 2009).

In Xenopus, physiological left-right asymmetry has been demonstrated (Pai et al., 2012). A study observed that right eye hyperpolarization occurs before hyperpolarization of the left eye in 61% of Xenopus. This physiological asymmetry may be mediated by K_ATP ion channels, as SUR1, a K_ATP channel subunit transcript is asymmetrically expressed in the right eye during development (Pai et al., 2012). This observation is consistent with those found in other systems where ion channels are known to be involved in neural asymmetry decisions (Alqadah et al., 2016b).

Brain asymmetry in Drosophila melanogaster

The asymmetrical body (AB) structure, detected with an antibody against the neural protein fasciclin II, is found in the right side of the Drosophila brain (Pascual et al., 2004). However, the AB structure is present bilaterally in a small proportion of flies. The flies with the AB structure on both sides of the brain have normal learning and short-term memory, but are defective in long-term memory compared to flies with lateralization of the AB (Pascual et al., 2004). In addition, the AB is innervated exclusively by neurons in the right side of the brain in most of the fly brains examined, while innervation of the AB is detected bilaterally but more extensively on the right side in a small population of flies (Jenett et al., 2012). In aging humans, there is a correlation between loss of cognitive function, including memory, with loss of hemispheric asymmetry (Cabeza, 2002). Age-related memory impairment also occurs in Drosophila (Tonoki and Davis, 2012; Tonoki and Davis, 2015). Although it is not known if the AB structure has any role in age-related memory loss, further studies may provide insight into the relationship of brain asymmetry and cognitive function, including memory.

Drosophila shows strong lateralization of odor tracking during flight. While bilateral inputs are required for gradient odor sensing, the left antenna contributes disproportionately more to odor tracking than the right, suggesting sensory lateralization (Duistermars et al., 2009). However, the mechanism of this lateralization has not been elucidated. Interestingly, in humans, the right nostril performed better in odor discrimination and furthermore, the right hemisphere had greater neural activity during odor stimulation, suggesting lateralization of odor discrimination and sensation (Zatorre et al., 1992; Yousem et al., 1997; Sobel et al., 1998). There may be commonalities in the molecular mechanisms of odor sensory lateralization between flies and humans. Further investigation into the mechanism utilized by Drosophila may provide information on lateralization in humans.

Neuronal asymmetry in C. elegans

Of the 302 neurons in the C. elegans hermaphrodite nervous system, there are currently only two known neuronal pairs that display left-right asymmetry at both molecular and functional levels: ASE (amphid neuron, single ciliated ending) taste neurons and AWC (amphid wing C) olfactory neurons (Yu et al., 1997; Troemel et al., 1999; Pierce-Shimomura et al., 2001; Wes and Bargmann, 2001). The cell bodies of both neuron pairs reside in the lateral ganglion of the head, with one subtype of the pair located on the left side of the body and the other subtype on the right (Fig. 3 inset) (White et al., 1986). The axons of AWC and ASE neurons enter the nerve ring, the brain of the worm, through the ventral ganglion (Fig. 3 inset) (White et al., 1986). While ASE neurons display directional asymmetry, AWC neurons show stochastic asymmetry. Directional ASE asymmetry and stochastic AWC asymmetry are established through distinct mechanisms (Sagasti, 2007; Taylor et al., 2010; Alqadah et al., 2013; Alqadah et al., 2014; Hobert, 2014; Hsieh et al., 2014; Alqadah et al., 2016c; Hsieh et al., 2017).

Figure 3. Establishment and maintenance of stochastic AWC asymmetry in C. elegans.

(A) Top: Intercellular calcium signaling through NSY-5 gap junctions coordinates the AWC^ON/AWC^OFF decision. Calcium signaling in non-AWC cells of the NSY-5 gap junction network promotes or inhibits the AWC^ON subtype. Bottom: In the default AWC^OFF cell (right), calcium influx through voltage-gated calcium channels activates CaMKII to trigger a MAPK cascade and the expression of AWC^OFF genes. In the induced AWC^ON cell (left), NSY-5 gap junctions, SLO BK potassium channels, and NSY-4 claudins suppress calcium channel-mediated signaling to promote the expression of AWC^ON genes. AWC^ON/AWC^OFF orientation here is displayed as left/right, but can also occur in the opposite orientation. Red, active AWC^ON-promoting molecules; green, active AWC^OFF-promoting molecules; grey, less active or inactive molecules. (B) Maintenance of AWC asymmetry. Red, active AWC^ON-maintaining molecules; green, active AWC^OFF-maintaining molecules; grey, less active or inactive molecules. (Inset) A diagram of a worm head that demonstrates anatomical position of AWC and ASE neuron pairs. Anterior is to the top.

Stochastic AWC asymmetry

The left and right AWC olfactory neurons asymmetrically differentiate into two subtypes, default AWC^OFF and induced AWC^ON. The two AWC neurons express different sets of genes and exhibit distinct functions (Troemel et al., 1999). The AWC^OFF subtype expresses the G protein-coupled receptor (GPCR) gene srsx-3 and senses the attractive odorant 2,3-pentanedione, and the AWC^ON subtype expresses the GPCR gene str-2 and senses the attractive odorant 2-butanone (Troemel et al., 1999; Wes and Bargmann, 2001; Bauer Huang et al., 2007). This AWC^OFF/AWC^ON subtype decision is determined during late embryogenesis and is maintained throughout life (Troemel et al., 1999; Chuang and Bargmann, 2005; Lesch et al., 2009; Lesch and Bargmann, 2010; Alqadah et al., 2016b).

The decision to specify the AWC^ON and AWC^OFF subtypes is stochastic yet coordinated (Troemel et al., 1999). Wild-type C. elegans have one AWC^ON cell and one AWC^OFF cell, each with approximately equal chances to be either on the left or right side. Although the cell bodies of the two AWC neurons reside on opposite sides of the head, their axons communicate with one another directly in the nerve ring through chemical synapses (Fig. 3) (White et al., 1986). Like brain asymmetry in zebrafish, intercellular communication is required for AWC asymmetry (Troemel et al., 1999; Chuang et al., 2007; Schumacher et al., 2012). This process is reminiscent of Delta-Notch signaling events that also occur during development, wherein a group of equivalent cells use lateral signaling to generate distinct cell fates (Cau and Blader, 2009). However, the Delta-Notch signaling pathway does not seem to regulate stochastic AWC asymmetry (Troemel et al., 1999). This implies that different mechanisms are used for establishing and maintaining AWC asymmetry.

A calcium-regulated MAP kinase cascade specifies the default AWC^OFF subtype

The default AWC^OFF subtype is specified by a calcium-regulated protein kinase cascade (Fig. 3A, bottom). In this signaling pathway, calcium influxes through the UNC-2 N-type or EGL-19 L-type pore-forming α1 subunit, and the UNC-36 α2δ regulatory subunit of voltage-gated calcium channels, which activates the UNC-43 calcium/calmodulin-dependent protein kinase II (CaMKII). UNC-43 and its target, the NSY-1 MAPKKK (ASK1), are brought together by the TIR-1 (Sarm1) adaptor protein at postsynaptic sites of AWC axons (Sagasti et al., 2001; Chuang and Bargmann, 2005; Chang et al., 2011). Upon activation, NSY-1 phosphorylates its downstream effector SEK-1 MAPKK to propagate the p38 MAP kinase pathway that suppresses str-2 expression and promotes srsx-3 expression (Sagasti et al., 2001; Tanaka-Hino et al., 2002; Pagano et al., 2015). Due to its localization to the postsynaptic region, the UNC-43(CaMKII)/TIR-1(Sarm1)/NSY-1(MAPKKK) complex can convey the MAP kinase cascade signaling towards the nucleus upon calcium stimulation (Chuang and Bargmann, 2005). Then, at the nucleus, the transcription factor DIE-1 cell autonomously promotes AWC^OFF (Alqadah et al., 2014; Cochella et al., 2014).

Intercellular calcium signaling in a NSY-5 gap junction network establishes stochastic AWC asymmetry

Genetic experiments suggest that the AWC with a relatively higher calcium level remains as the default AWC^OFF subtype, while the AWC with a relatively lower calcium level becomes the induced AWC^ON subtype (Schumacher et al., 2012). Induction of the AWC^ON subtype requires a balance between cell autonomous and non-cell autonomous inputs, as well as coordinated positive and negative feedback communication between AWC and non-AWC neurons (Chuang et al., 2007; Schumacher et al., 2012).

During late embryogenesis, NSY-5 gap junctions establish a transient neural network that connects both AWC neurons with other sensory neurons and interneurons (Fig. 3A, top) (Chuang and Bargmann, 2005; Chuang et al., 2007). Intercellular calcium signaling between AWC and non-AWC neurons across the NSY-5 network promotes the AWC^ON subtype. In addition, calcium signaling in other neurons of the network confers different side biases towards AWC^ON (Fig. 3A, top) (Schumacher et al., 2012).

Downstream of NSY-5 in AWC, the SLO-1 and SLO-2 voltage- and calcium-activated BK potassium channels antagonize UNC-2 and EGL-19 calcium channel-mediated signaling to promote AWC^ON (Alqadah et al., 2016b). Consequently, the downstream UNC-43 (CaMKII)/TIR-1 (Sarm1)/NSY-1 (MAPKKK) cascade is deactivated as well, suppressing the AWC^OFF subtype (Alqadah et al., 2016b).

NSY-4, a four-transmembrane domain claudin protein, acts in parallel with NSY-5 to downregulate calcium signaling in the pre-AWC^ON neuron (Vanhoven et al., 2006; Chuang et al., 2007). The AWC left (AWCL) and AWC right (AWCR) cells have distinct intrinsic biases for promoting AWC^ON: nsy-4 has a bias in AWCL and nsy-5 has a bias in AWCR (Chuang et al., 2007). Downstream of nsy-4 and nsy-5, the AWC^OFF fate is also inhibited at the post-transcriptional level by the microRNA (miRNA) mir-71, which targets tir-1. Downregulation of tir-1 effectively uncouples UNC-43 (CaMKII) and NSY-1 (MAPKKK), thus inhibiting their downstream signaling pathways (Hsieh et al., 2012; Alqadah et al., 2013). In addition, the highly conserved HMG-box transcription factor SOX-2 also promotes the AWC^ON subtype (Alqadah et al., 2015; Alqadah et al., 2016a).

Maintenance of AWC asymmetry

While the AWC^ON/AWC^OFF subtype decision is established using transient signaling during late embryogenesis, various mechanisms are used to maintain this decision throughout life (Troemel et al., 1999; Chuang and Bargmann, 2005; Chuang et al., 2007; Lesch et al., 2009; Lesch and Bargmann, 2010). The maintenance of AWC asymmetry requires olfactory cyclic guanosine monophosphate (cGMP) signaling, transcriptional regulation, and TGF-β signaling (Troemel et al., 1999; Lesch et al., 2009; Lesch and Bargmann, 2010). Consequently, it is thought that AWC asymmetry is not an immutable decision; rather, it is perpetuated throughout adulthood by ongoing sensory activity and environmental conditions.

The olfactory cGMP transduction pathway required to maintain both AWC^ON and AWC^OFF subtypes consists of the guanylyl cyclases (GCs) ODR-1 and DAF-11, the G_α protein ODR-3, and the cGMP-dependent kinase EGL-4 (Fig. 3B) (Troemel et al., 1999). In addition to these genes, tax-2 and tax-4, encoding cGMP-gated ion channels, are specific to maintaining the AWC^OFF subtype (Fig. 3B) (Troemel et al., 1999; Lesch and Bargmann, 2010).

Along with olfactory activity, there is transcriptional regulation of AWC asymmetry maintenance. The transcription factor NSY-7 is asymmetrically expressed in AWC^ON neurons, predominantly in early larval stages (Lesch and Bargmann, 2010). NSY-7 has been implicated in inhibiting promoter regions of known AWC^OFF-specific genes to reinforce the AWC^ON state (Fig. 3B) (Lesch et al., 2009). Similarly, the transcription factor HMBX-1 suppresses srsx-3 expression in AWC^ON cells, but is active predominantly in adult animals (Lesch and Bargmann, 2010).

TGF-β signaling is also shown to maintain AWC asymmetry. DAF-7, a TGF-β ligand, is known to bind the TGF-β type I receptor DAF-1, and is involved in the regulation of chemoreceptor expression in multiple olfactory neurons (Nolan et al., 2002). Interaction between daf-7 and daf-1 is required to maintain srsx-3 expression and the AWC^OFF phenotype. In addition, the dauer hormone inhibits daf-7 to repress srsx-3 expression, suggesting that environmental cues can regulate AWC asymmetry as well (Fig. 3B) (Lesch and Bargmann, 2010).

Directional ASE asymmetry

Unlike AWC neurons, in which asymmetry is stochastic, ASE neurons develop directional asymmetry. The ASE left (ASEL) neuron stereotypically expresses a set of genes distinct from those expressed in the ASE right (ASER) neuron, consistently across the C. elegans population (Yu et al., 1997; Ortiz et al., 2006). Although both ASE neurons function to detect salts, the ASEL and ASER neurons respond to distinct salt ions, as ASEL responds to sodium, lithium, and magnesium, and ASER responds to chloride, potassium, and bromide (Pierce-Shimomura et al., 2001; Ortiz et al., 2009). ASEL and ASER also exhibit morphological asymmetry, as the ASER soma is larger than that of ASEL (Goldsmith et al., 2010).

The ASE neurons use mechanisms largely distinct from the AWC neurons in developing neuronal asymmetry, with the exception of the C2H2 zinc finger transcription factor DIE-1, which acts in both AWC and ASE asymmetry pathways (Alqadah et al., 2014; Cochella et al., 2014). Unlike stochastic AWC asymmetry, directional ASE asymmetry is independent of intercellular communication and is dependent on cell lineages. A “prime and boost” model has been described for the mechanism used to establish asymmetry in the ASE neuronal pair (Cochella and Hobert, 2012). The asymmetry begins at the 4-cell stage of the embryo, where the precursor cell of the ASER neuron (ABp) has activated Notch receptor GLP-1, whereas the ASEL precursor cell (ABa) does not (Fig. 4) (Priess, 2005; Poole and Hobert, 2006). Several cell divisions before the ASEL neuron is born, a pair of redundant T-box transcription factors, TBX-37 and TBX-38, are expressed in an ancestor cell of ASEL. These transcription factors act on a cis-regulatory element deemed the “primer element” of a miRNA locus called lsy-6 (Cochella and Hobert, 2012). The T-box transcription factors result in de-compaction of the chromatin of the lsy-6 locus (Cochella and Hobert, 2012). This leads to a chromatin status that is “primed” for transcriptional activation (Fig. 4) (Cochella and Hobert, 2012). The priming of lsy-6 chromatin is “remembered” several cell divisions later in the mother cell of the ASEL neuron. In this cell, the zinc finger transcription factor CHE-1 binds to a cis-regulatory element in the lsy-6 locus called the “booster” element, and activates lsy-6, “boosting” its levels (Fig. 4) (Cochella and Hobert, 2012). This leads to sufficient lsy-6 miRNA in the ASEL neuron to down-regulate COG-1, a transcription factor responsible for ASER identity, and allows for expression of the ASEL-promoting transcription factor DIE-1 (Fig. 4) (Johnston and Hobert, 2003; Didiano and Hobert, 2008).

Figure 4. Establishment of directional ASE asymmetry in C. elegans.

Priming and boosting of lsy-6 miRNA throughout the ASEL cell lineage leads to the ASEL identity. At the 4-cell stage, activated Notch receptor in the ASER precursor inhibits priming and boosting events of lsy-6 miRNA, leading to the ASER identity. TF, transcription factor; purple, active ASER-promoting molecules; blue, active ASEL-promoting molecules; grey, less active or inactive molecules; circles represent cells.

Activation of Notch in the ASER precursor cell at the 4-cell stage results in repression of the T-box transcription factors TBX-37 and TBX-38 (Fig. 4) (Poole and Hobert, 2006). Therefore, “priming” of the lsy-6 chromatin by these transcription factors does not occur, and the CHE-1 transcription factor does not boost the levels of lsy-6. Therefore, in the ASER neuron, lsy-6 does not suppress COG-1, allowing for the transcription factor to promote the ASER cell fate (Cochella and Hobert, 2012). COG-1 and mir-273 down-regulate the ASEL-promoting transcription factor DIE-1, in a bistable feedback loop (Fig. 4) (Chang et al., 2004; Johnston et al., 2005). Loss-of-function mutations in the molecular players of the ASE asymmetry pathway result in both ASE neurons adopting the ASEL or ASER cell fate (Hobert, 2014).

Additional asymmetries in C. elegans

Ventral nerve cord asymmetry

In C. elegans, the ventral nerve cord is a structure similar to the spinal cord of vertebrates. It is composed of two bundles of axons that run along the left and right sides of the animal (White et al., 1986; Aurelio et al., 2002; Hobert and Bulow, 2003; Johnston and Hobert, 2003). The ventral nerve cord displays asymmetry, as the axon bundle on the right side is larger than that on the left side (White et al., 1976). Several genes have been implicated in the asymmetric development and maintenance of the ventral nerve cord, including nid-1/Nidogen, zig-3, zig-4/2-Ig Domain protein, and uncharacterized ast-4, ast-7, sax-5, and sax-9 genes (Zallen et al., 1999; Kim and Wadsworth, 2000; Aurelio et al., 2002; Hutter et al., 2005). A percentage of nid-1 loss-of-function mutants display a symmetric ventral nerve cord, as a larger number of axons project to the left bundle, and a lower number populate the right fascicle than in wild type animals (Kim and Wadsworth, 2000). This suggests a role for nid-1 in left-right asymmetry of the ventral nerve cord. nid-1 encodes a homologue of Nidogen, which is a component of the basement membrane. nid-1 localizes to the basement membrane in body wall muscles of C. elegans, however it does not appear to play a role in basement membrane assembly (Kim and Wadsworth, 2000). nid-1 may facilitate interactions between axons of the ventral nerve cord with the basement membrane of body wall muscles to ensure correct axon positioning (Kim and Wadsworth, 2000).

zig-3 and zig-4 are secreted immunoglobulin proteins that act together to maintain correct axon position in the ventral nerve cord (Aurelio et al., 2002; Benard et al., 2009). Loss of zig-3 and zig-4 results in a phenotype in which axons of the ventral nerve cord exhibit flipping over to the opposite fascicle. These proteins may function to signal receptors for keeping axons in place, or may be involved in anchoring axons to a particular fascicle (Aurelio et al., 2002; Benard et al., 2009).

In all of ast-4, ast-7, sax-5, and sax-9 mutants, the ventral nerve cord displays a symmetrical pattern. These mutants were identified from forward genetic screens for axon positioning defects (Zallen et al., 1999; Hutter et al., 2005). Molecular characterization of these genes will provide insight into mechanisms involved in left-right patterning of the ventral nerve cord.

Neuroblast Migration Asymmetry

C. elegans also displays asymmetric cell migration patterns in different pairs of blast cells (Aurelio et al., 2002). Two pairs of Q cells, QL and QR, divide to give rise to different classes of neurons (Sulston and Horvitz, 1977). These cells appear bilaterally symmetric in terms of cell position and lineage. However, their migration patterns are asymmetric, as the QL cell migrates posteriorly, while the QR cell migrates anteriorly. The homeobox gene mab-5 is asymmetrically expressed in QL, and is required for the posterior migration of that cell (Salser and Kenyon, 1992). Asymmetric expression of mab-5 requires elg-20/Wingless, and QL and QR have different threshold responses to egl-20 (Whangbo and Kenyon, 1999). It has been shown that the proneural gene lin-32 works together with the transcription factor HAM-1 in Q cell migration (Zhu et al., 2014). The pairs of P1/P2 and P11/P12 blast cells also display asymmetric migration across the left-right axis, as they rotate asymmetrically in a biased fashion. The P11/P12 asymmetry requires the Y blast cell as well as lin-12/Notch (Greenwald, 1998; Delattre and Felix, 2001).

Relationship between brain and visceral asymmetries

In humans, left-right asymmetry is observed in placement of visceral organs. For example, the heart is found on the left side of the body, while the liver is located on the right. A natural question arises as to whether the same genes used in left-right patterning of visceral organs are also used to establish brain laterality. Patients with a condition called situs inversus, in which visceral organ asymmetry is reversed (e.g. heart is found on the right side of the body, instead of the left), have been studied to determine whether reversal of asymmetry in the brain is also observed. Different studies reveal conflicting results, as one study found that patients with this condition had some reversal of certain anatomical cerebral asymmetry, yet still displayed language processing in the left hemisphere of the brain and were right-handed (Kennedy et al., 1999). In contrast, another study determined that language localization is altered in the brains of situs inversus patients (Ihara et al., 2010). The number of individuals tested in these studies is quite low, thus the interpretations from these results are conflicting. Therefore, zebrafish with situs inversus were studied to gain insight into the relationship between visceral and brain asymmetry. In a study, mutagenized zebrafish lines with situs inversus showed concordant reversal of visceral organs and neuroanatomical asymmetries. However, not all lateralized behaviors were reversed (Barth et al., 2005; Facchin et al., 2009). In addition, a variety of lateralized behaviors were not altered in zebrafish with randomized asymmetry of the epithalamus (Barth et al., 2005; Facchin et al., 2009).

A study by Blackiston & Levin investigated whether cognitive behaviors are influenced by defects in body asymmetries from heterotaxia, a condition in which visceral organ asymmetry is randomized, and situs inversus, where organ asymmetry is reversed, using Xenopus as a model (Blackiston and Levin, 2013). Induction of random heterotaxia and situs inversus in Xenopus tadpoles resulted in opposite preferences in swimming direction, as wild-type animals preferred to swim in a clockwise manner, while tadpoles with situs inversus largely swam in an anti-clockwise direction. Furthermore, in a shock treatment experiment, a significant delay in associative learning was observed in heterotaxic and situs inversus animals. This study suggests that body asymmetries may influence cognition. However, the exact mediation of this communication is not yet fully understood and requires further study. Taken together, these observations suggest that brain lateralization may not be the only factor leading to behavioral lateralization.

In C. elegans, embryos display invariant left-right asymmetry that is first evident during early embryogenesis and persists throughout development (Wood, 1991). A loss-of-function mutation in gpa-16, encoding a G protein alpha subunit, results in randomization of embryonic left-right asymmetry, leading to the reversal of visceral asymmetries (Bergmann et al., 2003). In addition, the authors of this study described unpublished data that gpa-16 mutants also display reversal of anatomical asymmetries of the nervous system, including Q cell migration and ventral nerve cord organization (Bergmann et al., 2003). In gpa-16 mutants, directional ASE asymmetry is also reversed (Poole and Hobert, 2006). These results suggest that neuronal asymmetry is linked to visceral laterality in C. elegans.

Conclusions and Perspectives

Here, we review studies on left-right asymmetry in the nervous system of humans and several model organisms. There are several similarities between the model organisms in establishing asymmetry. Intercellular communication is used in both zebrafish and C. elegans to establish asymmetry. Zebrafish utilize Nodal and Notch signaling, while AWC neurons use calcium signaling to communicate with a network of neurons through claudins and gap junctions (Table 1). Notch signaling and TBX transcription factors are required for both zebrafish and C. elegans ASE asymmetry (Table 1).

Table 1.

Comparison of mechanisms of brain asymmetry

	Human	Rodent	Zebrafish	Drosophila	C. elegans
Types of asymmetry known to date	Directional 1	Stochastic2	Directional3	Directional4	Stochastic and directional5
Molecular to functional correlates of asymmetry identified	No	Yes6	Yes7	Yes8	Yes (AWC and ASE)9
Intercellular communication required	Not determined	Not determined	Nodal and Notch signaling10	Not determined	Gap junction calcium signaling, claudins (AWC)11
Cell lineage-dependent	Not determined	Not determined	Not determined	Not determined	Yes (ASE)12
Notch signaling required	Not determined	Not determined	Yes13	Not determined.	Yes (ASE)14
TBX transcription factors used	Not determined	Not determined	Yes15	Not determined	Yes (ASE)16
Maintenance of asymmetry	Not determined	Not determined	Not determined	Not determined	Transcriptional regulation, olfactory cGMP signaling, TGF-β signaling17

¹(Gazzaniga et al., 1965; Knecht et al., 2000)

²(Glick and Ross, 1981; Waters and Denenberg, 1994)

³(Concha and Wilson, 2001)

⁴(Pascual et al., 2004; Jenett et al., 2012)

⁵(Yu et al., 1997; Troemel et al., 1999)

⁶(Li et al., 2013)

⁷(Gamse et al., 2003; Barth et al., 2005; Snelson et al., 2008; Facchin et al., 2009; Regan et al., 2009; Roussigne et al., 2009; Agetsuma et al., 2010; Dadda et al., 2010; Dreosti et al., 2014; Facchin et al., 2015)

⁸(Pascual et al., 2004)

⁹(Yu et al., 1997; Troemel et al., 1999; Pierce-Shimomura et al., 2001; Wes and Bargmann, 2001)

¹⁰(Aizawa et al., 2007; Roussigne et al., 2009)

¹¹(Troemel et al., 1999; Vanhoven et al., 2006; Chuang et al., 2007; Schumacher et al., 2012)

¹²(Cochella and Hobert, 2012; Alqadah et al., 2015)

¹³(Aizawa et al., 2007)

¹⁴(Cochella and Hobert, 2012)

¹⁵(Snelson et al., 2008)

¹⁶(Poole and Hobert, 2006)

¹⁷(Troemel et al., 1999; Nolan et al., 2002; Lesch et al., 2009; Lesch and Bargmann, 2010)

The loss of directional asymmetry uncovering the presence of stochastic asymmetry seems to be a common theme within zebrafish, Xenopus, and humans. While C. elegans use both types of asymmetries, stochastic asymmetry is normally overshadowed by directional asymmetry in human, zebrafish, and Xenopus. Zebrafish epithalamic asymmetry is directional, however, the asymmetry becomes stochastic when Nodal signaling is lost. In addition, directional asymmetry of Xenopus eye induction becomes randomized when left-right patterning of viscera is perturbed (Pai et al., 2012). In humans, approximately half of patients with Karagener’s syndrome exhibit situs inversus, in which the organs are located in reverse asymmetry (Hirokawa et al., 2009). Further studies may shed light into the mechanism and evolutionary aspects of the directional and stochastic neural asymmetries. As genetic and signaling pathways have been identified in the development of asymmetry of the nervous system in C. elegans and zebrafish, it would be interesting to test whether the same molecules are also involved in asymmetry in rodents.

Additionally, chirality is involved in left-right asymmetric development in neurons across the animal kingdom (Lobikin et al., 2012; Inaki et al., 2016). This is a phenomenon in which neurites have been observed to grow clockwise in vitro (Heacock and Agranoff, 1977). Disruption of left-handed chirality of microtubules causes a defect in AWC asymmetry in C. elegans (Lobikin et al., 2012).

While quite a few imaging studies have been performed in humans to identify anatomical asymmetries within the brain, many contradict each other, likely due to conflicts in measurement criteria. Therefore, it would be beneficial to establish a consensus for measurement criteria of anatomical asymmetry. In addition, gene expression studies have identified genes that are expressed differentially in the left and right hemispheres. However, to the best of our knowledge, functional correlates of these molecules have yet to be identified in humans. This will be key to understanding development of asymmetry in the human nervous system, and will further clarify whether diseases associated with increased symmetry have a genetic cause that leads to both symmetry phenotypes and a disease state. Model organisms are beginning to pave the way to provide a window into genetic pathways that may be involved in human brain asymmetry.

In rodents, studies have begun to identify the molecular mechanism behind paw preference using gene expression data gained from human studies. A recent study that identified genes asymmetrically expressed in the human spinal cord would be interesting to follow up on in rodent models to determine whether mutating any of the genes identified may result in functional differences in paw preference.

From studies in mutant zebrafish, in which anatomical asymmetry of the nervous system was reversed, it was found that some behaviors associated with lateralization were not reversed. This suggests that brain lateralization may not be the only factor resulting in lateralization of function. It would be interesting to determine the molecular mechanisms used in development of behavioral asymmetry that is independent of the anatomical asymmetry.

C. elegans remains one of the few model organisms in which clear molecular to functional correlates of brain asymmetry have been identified. Two forms of asymmetry have been extensively studied in the C. elegans nervous system: stochastic and directional. Interestingly, many of the molecules found to be asymmetrically expressed in the human fetal spinal cord and the human cortex are also used to establish brain asymmetry in C. elegans, such as voltage-gated channels, G protein-coupled receptors, miRNAs, and players in the TGF-β signaling pathway. This suggests that perhaps the pathways used to develop brain asymmetry in C. elegans are conserved in humans. Several players have been identified in establishing stochastic AWC asymmetry in C. elegans. However, the event used to break symmetry remains a mystery and warrants further study. Notch signaling early in the C. elegans lineage triggers a chromatin status change several cell divisions later, which is “remembered” further down the lineage to establish directional asymmetry in the pair of ASE taste neurons. It would be interesting to further dissect the mechanisms used in this intriguing process.

Left-right asymmetry is also observed in placement of visceral organs, including the heart, spleen, and liver. In vertebrates, different molecular mechanisms are proposed to establish left-right asymmetry of visceral organs. One mechanism involves the use of cilial movement to drive leftward flow of extraembryonic fluid at the ventral node during gastrulation. This leads to increased levels of intracellular calcium and triggers the expression of Nodal on the left side of the ventral node (Hirokawa et al., 2006; Raya and Izpisua Belmonte, 2006; Babu and Roy, 2013). Recently, a BMP-driven pathway has been shown to promote expression of the transcription factor Prrx1a, an epithelial-mesenchymal transition inducer, in zebrafish (Ocaña et al., 2017). Prrx1a is asymmetrically expressed in the right lateral plate mesoderm, and specifies the right side of the body for proper heart looping. A similar BMP-driven mechanism was also observed in chickens and mice in this study. This suggests that vertebrates use Nodal- and BMP-driven pathways that mutually repress each other on opposite sides to establish left-right asymmetry. Another mechanism proposes the involvement of tubulin subunits in establishing left-right asymmetry independent of cilia (Lobikin et al., 2012). Gap junctions and tight junctions have also been implicated in left-right patterning of the body plan in vertebrates (Levin, 2006; Oviedo and Levin, 2007). Tubulin, Nodal, gap junctions, tight junctions, and calcium signaling have been shown to also play roles in development of left-right asymmetry of the nervous system. In zebrafish, Nodal induces its own expression in the left epithalamus (Roussigne et al., 2009). In C. elegans, disruption of microtubules, gap junctions, tight junction protein claudins, and calcium signaling results in AWC asymmetry defects (Troemel et al., 1999; Vanhoven et al., 2006; Bauer Huang et al., 2007; Chuang et al., 2007; Chang et al., 2011; Lobikin et al., 2012; Schumacher et al., 2012). Therefore, there are commonalities between mechanisms used in left-right patterning of visceral organs and the nervous system.

Consolidating knowledge from different animal models can help identify the mechanisms used to establish left-right asymmetry in the nervous system of humans, and may lead to discovery of treatments of diseases associated with lack of asymmetry in the brain, as well as diseases that appear unilaterally, such as migraines, Rasmussen encephalitis, and cerebral palsy.