|Home | About | Journals | Submit | Contact Us | Français|
The rapidly adapting (RA) low threshold mechanoreceptors respond to movement of the skin and vibration, and are critical for the perception of texture and shape. In this issue of Neuron, two papers demonstrate that early-born Ret+ sensory neurons are RA mechanoreceptors, whose peripheral nerve terminals are associated with Meissner corpuscles, the longitudinal lanceolate endings, and the Pacinian corpuscles. The studies further show that Ret signaling is essential for the development of these mechanoreceptors.
In mammals, sensory neurons in the dorsal root ganglia (DRG) and trigeminal ganglia detect a wide variety of mechanical stimuli, and they are critical for the perception of touch, body positions and pain. In this preview, we focus on low threshold mechanoreceptors whose peripheral terminals are associated with specialized end-organs in the skin or the deep tissues (Albrecht, 2008; Lewin and Moshourab, 2004; Montaño et al., 2009). Based on adaptation rates in response to sustained mechanical stimuli, these mechanoreceptors are divided into rapidly adapting (RA) and slowly adapting (SA) subtypes. In mice, RA mechanoreceptors terminate as Meissner corpuscles in the dermal papillae, the longitudinal lanceolate endings in hair follicles, and Pacinian corpuscles in the joints and the periostea of bones; these neurons respond to movement of the skin and vibration, and are therefore critical for the perception of texture, shape and vibration. SA mechanoreceptors innervate the Merkel discs and Ruffini corpuscles, and respond to static indentation stretch.
Despite extensive morphological and electrophysiological characterizations, developmental ontogeny and molecular identities of RA versus SA mechanoreceptors are still poorly studied. A few studies have implicated a crucial role for neurotrophic factors in the development of specific subtypes of RA or SA mechanoreceptors. For example, the neurotrophin NT3 is required to maintain Merkel cells and nerve terminals from SA mechanoreceptors (Airaksinen et al., 1996). In another example, BDNF-mediated TrkB signaling is critical for the formation of Meissner corpuscles, but is dispensable or plays a minor role for the development of other RA mechanoreceptors (Montaño et al., 2009). In this issue of Neuron, Luo et al. and Bourane et al. not only expand upon this knowledge by demonstrating the involvement of the Ret signaling pathway in the development of these mechanoreceptors, but they also molecularly and anatomically identify a distinct set of neurons—the “early Ret+ neurons”—as the prospective RA mechanoreceptors (Bourane et al., 2009; Luo et al., 2009).
The Ret tyrosine kinase is a signaling receptor for the glial derived neurotrophic factor (GDNF) family of ligands, which includes four members: GDNF, neurturin, artemin, and persephin. The binding of these ligands to Ret requires a GPI-anchored coreceptor (GFRα1–4). Previous studies have revealed two distinct waves of Ret+ neurons in the developing DRG (Kramer et al., 2006; Luo et al., 2007; Molliver et al., 1997). Most Ret+ neurons emerge after E13.5 to form a subset of nociceptors that later become nonpeptidergic nociceptors. However, a small population of neurons (about 5% of DRG neurons) express Ret at E10.5–11.5, and until now the functional identity of these “early Ret+ neurons” was unknown. Using elegant genetic and molecular tools, Luo et al. and Bourane et al. now show that these early Ret+ neurons develop into RA mechanoreceptors, and Ret is critical for their development.
At the molecular level, Bourane et al. show that early Ret+ neurons coexpress GFRα2 (also shown by Luo et al., this issue) and MafA, a basic leucine-zipper class transcription factor. Previous and current studies show that early Ret+ neurons initially coexpress TrkB at E10.5–E11.5 (Kramer et al., 2006). TrkB+ neurons are formed during an early wave of sensory neurogenesis and their development is dependent on the neuronal determination gene Neurogenin2 (Ngn2). Consistently, Bourane et al. show that early Ret+ neurons fail to form in Ngn2 mutant mice. By E14, the Arber group showed that TrkB expression is extinguished in most early Ret+ neurons (Kramer et al., 2006). Others have reported a second wave of TrkB+ and TrkC+ neurons emerging at perinatal stages (Nakamura et al., 2008). Consistently, in newly born mice about 50% of Ret+;GFRα2+;Mafa+ neurons “regain” TrkB and another 20% coexpress TrkC. As a result, early Ret+;GFRα2+;MafA+ neurons are segregated into three distinct subtypes: i) Ret+, ii) Ret+; TrkB+, and iii) Ret+; TrkC+ neurons.
To determine if early Ret+ neurons represent specific classes of mechanoreceptors, Bourane et al. and Luo et al. used different approaches to examine peripheral and central projections of these neurons. Luo et al. used genetic manipulations to mark early Ret+ neurons. To do this, they first made RetERT2 knock-in mice, in which the expression of an inducible form of Cre recombinase (CreERT2) is driven from the Ret promoter. They then crossed RetERT2 mice with neuron-specific Tauf(mGFP) reporter mice to generate RetERT2; Tauf(mGFP) mice. Injection of 4-hydroxytamoxifen (4-HT) from E10.5 to E12.5 (the time of the first wave of Ret expression) led to exclusive expression of a membrane associated form of green fluorescent protein (GFP) in early Ret+ neurons. This genetic marking technique allowed Luo et al. to examine the relationship between GFP+ fibers and specific mechanosensory end-organs. They found that GFP+ fibers terminated as Meissner corpuscles in the dermal papillae of the footpad, longitudinal lanceolate endings associated with guard hair follicles, and Pacinian corpuscles in the periosteum of the fibula. Strikingly, GFP+ fibers did not innervate the Merkel cells in hairy or glabrous skin. These data suggest that early Ret+ neurons belong to RA, but not SA, mechanoreceptors.
The marking of early Ret+ neurons also allows for direct visualization of the central projections of these neurons. Mechanoreceptors send two sets of collateral branches: one innervating the deep lamina of the dorsal horn of the spinal cord and another ascending ipsilaterally within the dorsal column to the gracile and cuneate nuclei (DCN). Consistently, GFP+ fibers innervated lamina III–V of the dorsal horn in RetERT2; Tauf(mGFP) mice. More interestingly, GFP+ fibers showed specific termination zones within the DCN, including i) most of the gracile nucleus (with the exception of a small dorsal-rostral zone) and ii) a small medial-rostral zone within the cuneate nucleus. Remarkably, this zonal organization nearly matches the projection patterns of RA mechanoreceptors revealed by electrophysiological recording (Dykes, 1982), further supporting the conclusion that early Ret+ neurons belong to a group of RA mechanoreceptors.
Bourane et al. took a different approach to characterize early Ret+ neurons, and reached almost the same conclusion. They took advantage of the finding that early Ret+ neurons express a high level of the neurofilament NF200 as well as GFRα2. By performing a series of double immunostaining experiments, they first show that centrally, Ret+;NF200+ fibers innervate lamina III/IV of the dorsal horn, and peripherally, they terminate as longitudinal lanceolate endings in the guard hair follicles. They also found that Ret+ fibers do not innervate the Merkel cells in the glabrous skin. All of these observations are consistent with the findings by Luo et al. One notable exception is that Bourane et al. reported innervation of Ret+ fibers to the Merkel cells within the touch dome of the guard hair follicle, while Luo et al. failed to see such innervation by genetically marked early born Ret+ neurons. It should be pointed out that the triple staining of GFRα2, NF200 and CK20 (a marker for Merkel cells) done by Bourane et al. showed that the fibers innervating the Merkel cells exhibited a weak staining for GFRα2. Luo et al. also show that a large subset of late-born Ret+ neurons expresses GFRα2, albeit at a lower level in comparison with that in early Ret+ neurons. Thus, one possibility is that the Ret+ fibers observed by Bourane et al. in the Merkel disc might belong to Ret+ neurons emerging after E12.5. Accordingly, 4-HT injection from E10.5 to E12.5 in RetERT2;Tauf(mGFP) mice would then fail to label these neurons. Indeed, it has already been shown that one subset of late Ret+ neurons marked by the expression of the G-protein coupled receptor Mrgprb4 innervates exclusively hair follicles, although these fibers are NF200-negative (Liu et al., 2007).
After demonstrating that early Ret+ neurons predominantly belong to RA mechanoreceptors, both groups next investigated the role of Ret signaling in controlling the development of these mechanoreceptors. Both groups crossed the floxed Ret mutant mice (Retf) with Wnt1Cre mice, with the resulting conditional homozygous mutant mice referred to as RetcKO mice. In these mice, Ret was removed from the neural crest (and the dorsal neural tube), including sensory precursor cells. They first examined the expression of molecular markers, and found that Ret signaling is partially required to establish MafA expression and later is required to maintain the expression of MafA and GFRα2. Luo et al. then showed that neuronal survival was largely unaffected in RetcKO mice, although Bourane et al. detected a small reduction of TrkB+, TrkC+ and NF200+ neurons.
They next examined the development of peripheral endings of RA mechanoreceptors in these knockout mice. Bourane et al. found that longitudinal lanceolate endings are completely absent in RetcKO mice. Because other types of sensory fibers still innervated this area, Luo et al. failed to reveal this dramatic mutant phenotype when the hairy skin was sectioned parallel to the surface. Luo et al. then show a complete loss of the Pacinian corpuscles in RetcKO mice. The loss of Pacinian corpuscles was also observed in GFRα2 or NRTN mutant mice, indicating that NRTN-mediated signaling is essential for the development of Pacinian corpuscles. Meissner corpuscles in the dermal papillae of the footpad are present in normal numbers, but they are disorganized or morphologically underdeveloped. These studies strongly suggest that Ret signaling is essential for proper development of all three classes of RA mechanoreceptors. Bourane et al. additionally found that the development of touch domes surrounding the guard hair follicles was partially impaired in RetcKO mice, as suggested by a reduction of Merkel cells and NF200+ fibers. As discussed above, it needs to be determined if Ret signaling in late Ret+ neurons, rather than in early Ret+ neurons, contributes to the proper development of this class of SA mechanoreceptors.
Both groups show that Ret signaling is also required for central projections by mechanoreceptors. Bourane et al. showed a marked reduction of NF200+ terminals in the lamina III/IV of the dorsal spinal cord, a region normally innervated by mechanoreceptors. Most mechanoreceptors also express the synaptic glutamate transporter VGLUT1, and Luo et al. found that VGLUT1+ fibers in the dorsal horn as well as in the DCN were markedly reduced in RetCKO mice.
To more specifically analyze central projections by early born Ret+ neurons, Luo et al. again used a genetic marking system to label fibers from prospective Ret null neurons. To do this, they took advantage of the availability of conditional Retf(CFP) knock-in mice, in which a Cre-mediated recombination event will activate cyan fluorescent protein (CFP) expression from a null Ret allele. After crossing with RetERT2 mice, the resulting RetERT2;Retf(CFP) mice were treated with 4-HT at E11.5–E12.5 to make CFP+ Ret null neurons. Moreover, by reducing the dosage of 4-HT, they were able to generate and analyze a select set of CFP+ mutant neurons. By this strategy, Luo et al. found that in the absence of Ret signaling, the central axons of mechanoreceptors were still able to bifurcate in the dorsal funiculus and project properly in both rostral and caudal directions. However, the third-order collateral branches failed to emanate from the longitudinal projections. As a result of this, no CFP+ null fibers were detected in the deep lamina of the dorsal horn. Thus, Ret signaling plays an essential role in the development of RA mechanoreceptors, including both morphological development of peripheral end-organs and proper axonal innervation to central targets.
One important conclusion from the current and previous studies is that the development of each class of mechanoreceptors is dependent on a myriad of neurotrophic factors. For example, BDNF-mediated TrkB signaling plays a major role in the formation of peripheral Meissner corpuscles (Montaño et al., 2009; also concluded here from the analysis of TrkB conditional knockouts by Luo et al.), whereas Ret signaling is shown here more critical in controlling central projections by these mechanoreceptors. Meanwhile, a given neurotrophic factor can have quite distinct roles in controlling the development of different classes of sensory neurons. For example, while Ret signaling plays a minor role in the formation of Meissner corpuscles, it is absolutely essential for the formation of the Pacinian corpuscles and the longitudinal lanceolate endings. Similarly, Ret signaling is necessary for early-born RA mechanoreceptors, but dispensable for late-born nonpeptidergic nociceptors, to innervate the dorsal spinal cord (Lindfors et al., 2006; Luo et al., 2007).
These exciting findings raise a number of new questions. First, while early-born Ret+ neurons likely represent RA mechanoreceptors, the molecular identity of SA mechanoreceptors remains unknown or controversial. Bourane et al. showed a defect in the development of the hair touch dome in Ret mutant mice, suggesting a possible role of Ret signaling in controlling the development of SA mechanoreceptors. However, as suggested above, it remains unclear if early or late Ret+ neurons are involved. Second, how are early Ret+ neurons further segregated into distinct subtypes of RA mechanoreceptors? Bourane et al. have divided early Ret+;MafA+ neurons into three groups, i) Ret+, ii) Ret+;TrkB+, and iii) Ret+;TrkC+. Based on the prominent role of TrkB signaling in controlling the formation of Meisnner corpuscles, it is tempting to speculate that the Ret+;TrkB+ subset of neurons might represent this class of RA mechanoreceptors. It will also be interesting to determine if TrkC signaling in the Ret+;TrkC+ subset of neurons is required for the development of Pacinian corpuscles versus longitudinal lanceolate endings. Third, the intrinsic transcription factors that control mechanoreceptor subtype specification also need further characterization. MafA is expressed in all early Ret+ neurons and its expression is dependent on Ret signaling. Mutant analyses show that MafA is required to establish the Ret+ versus Ret+;TrkB+ subset of neurons by activating Ret and repressing TrkB. However, it is not known how the presence of MafA in the Ret+;TrkB+ subset of neurons fails to suppress TrkB. Nor is it known if MafA is required to establish other molecular or anatomical features of RA mechanoreceptors.
Luo et al. and Bourane et al. have identified the early Ret+ neurons as RA mechanoreceptors, and demonstrated that Ret signaling is critical for the development of these neurons. The genetic marking of these mechanoreceptors by the expression of GFP should open up a range of future studies. For example, the electrophysiological properties of these labeled neurons can now be measured to directly prove that these early Ret+ neurons are RA mechanoreceptors. The expression of GFP should also allow for the isolation of these neurons for future gene expression profiling analyses. To date, the ion channels that respond to mechanical stimuli remain to be identified. The prospective identification and detailed molecular characterization of these RA mechanoreceptors will certainly help to identify these still elusive mechanical ion channels.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.