Variation in taste ganglion neuron morphology: insights into taste function and plasticity

Abstract

Chemical information from food is transduced by cells in the taste bud (taste-transducing cells) and carried to the brain by peripheral taste ganglion neurons. These neurons are thought to act simply as cables without any transformation of the signal or circuitry between the taste-transducing cells and the neurons. However, these neurons vary in structure, particularly in the extent of their peripheral axon branching. Such structural differences would be expected to underlie differences in the number of taste-transducing cells providing convergent information to these neurons. However, axon branching may vary over time and morphological differences between neurons might also reflect neuron plasticity. Because taste-transducing cells die and are replaced, the morphology of neurons may change as they form connections with new cells within the taste bud. Modern genetic approaches may permit investigations of the complex relationship among gustatory neuron morphology, circuitry, and function. This review discusses potential relationships among peripheral taste neuron morphology, function, and plasticity to help advance our understanding of taste system function and dysfunction.

Introduction

Classifying neurons allows the characterization of distinct neuronal features within specific types and enables comparisons of these neuronal features across experiments [1]. Where as neurons have traditionally been classified based on their morphology [2], such classifications have more recently been based on a combination of morphology, function, and gene expression [1]. However, this combination of features has not yet been applied to peripheral taste ganglion neurons, which carry information concerning the chemical content of food from taste transducing cells within the taste bud to the brain. These neurons are often thought of as simple cables whose function solely reflects that of the receptor cells they innervate [3] and were historically classified into types based only on their functional characteristics [3–5]. When a limited number of stimuli at mid-range concentrations are used to examine neuron and taste-transducing cell function, peripheral taste neurons indeed appear to have a similar breadth of tuning as the taste-transducing cells that they innervate [6]. As a result, little attention has focused on any potential circuitry that exists at peripheral connections between sensory neurons and taste-transducing cells.

Recently, the genetic profiles of gustatory sensory ganglion neurons have been revealed [5,7,8^•]. In a separate study, morphologies of the peripheral axons of individual taste neurons have been characterized [9^•]. Rather than existing as simple cables, the peripheral axons of taste neurons vary in structure and thus are likely to differ in the extent of convergent information they transmit to the hindbrain. To date, no studies have used genetic expression to examine the relationship between neuron morphology and function in peripheral taste neurons. Differences in neuron morphology often give rise to circuitry features that reflect differences in function [10]. Because the array of stimuli used for any given functional experiment is limited, understanding the structure of a neuron type can inform the design of functional studies, resulting in a more complete view of the role of particular neuron types [11]. For example, the morphological observation that neurons with a specific type of ending that terminates on one side of a hair follicle led to the finding that this neuron type is direction-selective [12]. Thus, understanding differences in the structure and circuitry of gustatory sensory ganglion neurons and their peripheral axons could facilitate new hypotheses and functional investigations that may not have otherwise been considered.

For gustatory neurons, the relationship between structure and function is complicated by the possibility that gustatory neuron structure is plastic. That is, the structure of taste neurons likely changes over time. Taste neurons connect to taste-transducing cells residing in taste buds. These taste-transducing cells have a limited lifespan and are constantly replaced [13]. Within this dynamic environment, some morphological features of peripheral taste neurons may change as connections are broken and reformed, while other structural characteristics may remain relatively stable. Although this plasticity might make the task of relating structure to function more difficult, understanding how the morphology of peripheral taste neurons influences their function could provide tremendous insight, particularly in situations where molecular, developmental, disease, or treatment manipulations alter both function and morphology in predictable ways. The goal of this review is to consider the complicated relationship between peripheral gustatory neuron morphology and function in order to help advance our understanding of taste system function and dysfunction.

Do peripheral taste neurons exist as distinct morphological types?

The branching complexity of the peripheral axons of gustatory neurons can vary considerably [9^•]. This variation in morphology produces large differences in the number of taste-transducing cells that could provide input to neurons. Peripheral gustatory neuron morphology varies along a continuum, such that distinct morphological types are not obvious. However, this does not mean that functionally or genetically defined types of gustatory neurons do not differ in morphology, as some may be more heavily branched and innervate a larger number of taste-transducing cells than others. Consistently, some functional neuron types are more narrowly tuned, whereas others are more broadly tuned [6,14,15], yet the anatomical or circuitry underpinnings of this difference are unclear. One possibility is that broadly tuned neurons are more heavily branched. That is, a neuron with more terminal arbors will likely innervate more taste transducing cells. Approximately one-fourth of taste-transducing cells within fungiform taste buds are capable of responding to more than one stimulus [16]. Therefore, the greater the number of terminal arbors a taste sensory neuron has, the higher the probability that the neuron receives input from taste-transducing cells responding to more than one stimulus. This ought to be true even if individual neurons do not receive input from more than one type of taste-transducing cell (i.e., Type I, Type II, or Type III cells). For example, if a subset of Type III cells responds to both sour and another stimulus quality (i.e., bitter, sweet or umami) [16–18,19^•], then as the number of Type III cells contributing to the response increases the probability the neuron will innervate at least one cell responding to multiple stimuli also increases. Assuming that all responses in taste-transducing cells drive a neuronal response, increasing the number of more broadly tuned Type III taste cells that the neuron innervates would also be expected to broaden the tuning properties of the neurons. Consistent with this possibility, the receptive field size of a geniculate ganglion neuron is positively correlated with the number of different chemical stimuli to which that neuron responds [20].

Interestingly, variations in tuning properties exist across taste neurons that respond best to different qualities (i.e., sweet, salty, sour). Specifically, neurons that respond best to sour stimuli are typically described as more broadly tuned than those responding best to other stimuli [15,21,22]. Neurons contacting Type III (sour-transducing) are more heavily branched than neurons contacting sweet/bitter-transducing or umami-transducing cells [9^•]. Also, sour-transducing cells contain a larger subpopulation of that are broadly tuned [17,18,19^•]. Thus, both neuron morphology and the type of transducing cell innervated may act synergistically to produce a more broadly tuned neuron. If so, it may be important to determine functional contribution of broader tuning, rather than assuming it is noise.

Consistent with the idea that neurons vary in their breadth of tuning, neurons responding to the sodium-specific (i.e., amiloride-sensitive) component of a salt stimulus are described as more narrowly tuned [23]. Interestingly, these neurons refine their receptive field size during development [24]. Specifically, neurons responsive to NaCl prune their receptive fields to a greater extent than those responding to NH₄Cl. This process could be activity-dependent and likely results in more sparsely branched neurons that innervate fewer taste buds, rendering them more narrowly tuned. It is noteworthy that this refinement occurs in only some types of neurons. Thus, different functionally and/or genetically defined neuron types likely also differ in morphology.

In addition to variation across neuron types, there may be variability in peripheral axon morphology within populations of neurons responding to specific taste qualities (i.e., sweet, bitter, sour). These morphological differences would allow a subset of neurons to receive input from a larger number of taste-transducing cells within a taste neuron typing that responds to a single stimulus. Differences in convergence of input from the same type of transducing cells onto separate neurons could result in differences in thresholds or ranges of concentrations to which particular neurons respond (Figure 1). For example, a neuron innervating eight sour-transducing cells [25,26^•] may have a lower threshold for citric acid than a neuron that innervates only two sour-transducing cells. Consistent with this idea, stimulation of independent areas of a neuron’s receptive field with the same stimulus enhances the neuronal response [27]. Because only some peripheral taste neurons branch to innervate multiple taste-transducing cells, only these neurons may exhibit an enhanced response [9^•]. Thus, differences in the extent of branching across a population of neurons responding to the same stimulus could produce variations in taste thresholds and intensity ranges. Whereas taste quality coding has been examined extensively for peripheral taste neurons [3,4,28], intensity coding has received little attention. In fact, much of what we know about differences between neurons in taste quality are determined by a single concentration of each stimulus. The unspoken assumption is that all neurons responding to a given stimulus respond to that stimulus at all concentrations. However, consistent with the idea that the same neurons may not respond to all concentrations of a given stimulus, stimulus intensity impacts taste quality coding [14,21,29]. Thus, morphological variation within a type of taste neurons responding to similar taste qualities could give rise to an intensity code, such that the number of neurons responding increases as the stimulus concentration increases (Figure 1). If so, stimulus intensity would be represented by both an increase in firing rate for individual neurons, but also an increase in the number of neurons responding across concentrations (graded coding [30]).

Figure 1.

The peripheral axons of taste ganglion neurons exist along a continuum of branching complexity. Heavily branched neurons innervate more taste transducing cells (lower left-hand corners of a–b) than axons that are sparsely branched (upper right-hand corners of a–b). (a) Only a few neurons are predicted to respond to low taste stimulus concentrations. We predict that neurons innervating many taste transducing cells of the same type will require a lower concentration of a taste stimulus to reach threshold for activation (lower left-hand corner a). (b) A larger number of neurons are predicted to respond at higher stimulus concentrations. Neurons that branch sparsely may require a higher taste stimulus concentration to reach their threshold for activation (upper right-hand corner of B). If taste neurons differ in the concentration of stimulus required for activation, then as the stimulus concentration increases (X-axis) more neurons will respond to the stimulus (Y-axis). All of the taste-transducing cells depicted represent Type III cells, which tend to transduce sour stimuli. Blue cells represent the cells responding exclusively to sour stimuli whereas purple cells represent the sour transducing cells that also respond to additional taste qualities. As neurons innervate a greater number of Type III cells, the probability increases that it will connect with cells that respond to more than one taste stimulus. Thus, one consequence of increases in branching could be an increased breath of tuning.

Do variations in neuron morphology reflect plasticity?

A unique feature of taste-transducing cells is that they have a limited lifespan and are constantly renewed [13,31]. As a result, peripheral axons of taste neurons must continually locate and form functional connections with new taste-transducing cells throughout adulthood. These connections are in two forms: conventional chemical synapses and synapse-like connections. The latter consist of CALMH1 + CALMH3 channels that secrete ATP, which is detected by post-synaptic neurons [32,33,34^•]. For, the purposes of this review we will refer to both types of these connections as synapses. Although several studies have investigated the molecular mechanisms underlying the formation of new synapses between taste-transducing cells and gustatory neurons [35,36], no studies have examined the cellular processes or actions required for a gustatory neuron to come into contact with an appropriate taste-transducing cell. One of these cellular processes could be a change in neuron branching.

The peripheral axons of sensory neurons likely undergo morphological changes when connecting with new transducing cells during cell turnover. That is, once a synapse with a transducing cell is lost, an individual axon may need to retract an old branch or grow a new branch, possibly this new branch follows specific cues toward a specific type of taste-transducing cell. Alternatively, the same axon may be influenced to retract from other types of taste bud cells. Once an axon contacts a taste-transducing cell of an appropriate type, it may initiate the formation of a synapse or synapse-like connection. Thus, multiple cellular processes and numerous molecular cues are likely required for synapse formation between taste-transducing cells and gustatory nerve arbors. One complication is that many molecular factors that regulate synapse formation and connectivity also regulate axon guidance, growth, retraction, and branching. Disruption of any of these cellular processes could ultimately impact the ability of a nerve arbor to connect with a taste-transducing cell, which may in turn disrupt synapse formation and/or neuron function. As a consequence, it is not possible to pinpoint the cellular mechanisms driving the formation of new connections between a new taste transducing cell and a nerve arbor by measuring taste function [35] or total taste bud innervation [36]. We could speculate that some released factors, such as brain-derived neurotrophic factor (BDNF) and semaphorins, regulate branching [37–39]. Indeed, during development BDNF and semaphorins exert antagonistic effects on taste neurons [40]. Sema3a and Sema3F inhibit gustatory axon growth and are repulsive factors in taste neuron axon guidance [41,42]. If a factor (e.g., Sema3A) normally inhibits branching, then removing that factor would be expected to increase branching. Increased branching could in turn permit a neuron to connect with a larger number of taste-transducing cells, potentially making the neuron more broadly tuned [35]. This regulation of axon growth, branching, and/or the formation of a functional connection could be neuron type-specific. Ultimately, the continual rewiring of the peripheral taste system likely requires a complex combination of cellular processes and underlying molecular factors. Currently, the list of molecular factors potentially involved includes growth factors [36,43], semaphorins [35,44], ephrins [45], cadherins [5], and protocadherins [46^•]. Determining the precise role of any of these factors requires a combination of detailed functional and morphological analyses.

Given that dynamic changes in branching may occur when afferent axons innervate a taste-transducing cell, it is unclear to what extent the varying morphological complexity of the peripheral axon of a gustatory ganglion neuron represents merely a snapshot of a continuous process. Some branch points of the peripheral axon may remain stable, whereas others may be plastic. An axon of a taste neuron has branch points in the tongue musculature and lamina propria that determine the number of separate terminal arbors that innervate taste buds (Figure 2a). Each peripheral terminal arbor could have branch points, and individual arbors vary in complexity. Changes in branching may be limited, occurring only at terminal branch ends, or the entire peripheral axon could remodel over time (Figure 2b). If differences in morphology underlie differences in convergence, then the degree of plasticity in the system might determine the range of functional change occurring within a given neuron over time [47].

Figure 2.

(a) Sagittal section of the tongue including two peripheral chorda tympani axons enter the tongue at the dorsal midline. After entry, individual nerve axons branch in the musculature of the tongue (dark blue dots), in the lamina propria or papilla beneath the taste bud (light blue dots), and/or within the taste bud, referred to as the terminal arbor (yellow dots). (b) The degree to which plasticity alters peripheral axon branching is unclear. As innervated receptor cells undergo apoptosis and the innervating branch retracts, it is possible that neurons that innervate receptor cells in 1–2 taste buds remain sparsely branched (top) and that neurons with many arbors remain complex (bottom) throughout the animal’s life. Alternatively, it is also possible, that plasticity allows individual neurons to vary across this entire continuum of branching complexity.

Conclusions

In general, variation in neuron morphology is indicative of circuitry characteristics, specifically the amount of convergent information a neuron receives. The taste system is no exception; variation in peripheral axon morphologies of gustatory neurons likely dictates the number of taste-transducing cells that are innervated by individual ganglion neurons. These differences likely underlie some aspects of function. Two possibilities are that differences in the degree of convergence could influence both the breadth of tuning and sensitivity of a given neuron to a specific stimulus. Variations in peripheral gustatory neuron morphology likely depend on too many factors for these neurons to exist as distinct morphological types-based solely on stereotypical patterns of branching. However, it is likely that functionally and/or genetically defined neuron types differ in morphology. This possibility is consistent with several functional studies demonstrating that some neuron types are more broadly tuned than others. Furthermore, a consistent but limited spectrum/range in neuron morphology within a functionally defined type may underlie an intensity code that has not yet been determined. Lastly, many developmental and molecular manipulations likely alter neuron morphology. If differences in neuron morphology underlie differences in function, then it is likely that functional changes occur downstream of morphological changes. Thus, consideration of how neuron type or molecular manipulations impact neuron morphology may be important for understanding how these factors influence taste function.