By making a detailed and quantitative comparison of six, hind-leg-innervating, peripheral nerves in six species (four genera) from the African mole-rat family Bathyergidae, we conclude that naked mole-rats are unique among these bathyergids (and other mammals examined to date) in having a low C:A-fiber ratio in the cutaneous saphenous and sural nerves, whereas no such C-fiber deficit was observed in tibial, common peroneal, medial gastrocnemius, or lateral gastrocnemius nerves, which provide innervation to skeletal muscle.
Low C:A-fiber ratio in naked mole-rat saphenous and sural nerves
Unmyelinated cutaneous C-fibers perform different functions in mammals, ranging from being polymodal nociceptors to thermoreceptors and even low-threshold mechanoreceptors (Olausson et al.,2010; Li et al.,2011; Wende et al.,2012). The ability to detect noxious stimuli is fundamental for an organism's survival and has presumably been the selection pressure behind the evolution of an elaborate repertoire of nociceptors in mammals to carry out this task (Kavaliers,1988; Walters,1996; Smith and Lewin,2009). C-fiber nociceptors usually outnumber A-fibers (predominantly non-nociceptors) in nerves innervating the skin, including the saphenous nerve (Scadding,1980; Alpsan and Lal,1980; Lynn,1984; Jancso et al.,1985; Carter and Lisney,1987; Illanes et al.,1990; Milenkovic et al.,2007; Wetzel et al.,2007; Park et al.,2008) and sural nerve (Ochoa and Mair,1969; Schwab et al.,1984; Jenq and Coggeshall,1984a, b, 1985a, b; Peyronnard et al.,1986). We have previously shown that, in comparison with other rodents, the naked mole-rat saphenous nerve has a C-fiber deficit, resulting in a very low C:A-fiber ratio (Park et al.,2008), which is confirmed in this study (Fig. 2A). By conducting a comparative study with other members of the bathyergid family, we can now conclude that this phenomenon is species specific; all other Bathyergidae species that we examined had significantly higher C:A-fiber ratios in the saphenous nerve (2.5–3.7:1). Comparing calculated body surface area to A-/C-fiber counts supports the hypothesis that the low C:A-fiber ratio in naked mole-rat saphenous nerves is due to a loss of C-fibers rather than an increase in A-fiber numbers (Fig. 1G). In addition, naked mole-rats are slightly larger than mice, and, although one would therefore expect higher total fiber counts in the naked mole-rat, this is not the case compared with the mice used in this study (A-fibers 412 ± 29 vs. 751 ± 20, and C-fibers 682 ± 55 vs. 3,289 ± 115 in the saphenous nerve), which again supports the hypothesis that the low C:A-fiber ratio observed in naked mole-rat cutaneous nerves is due to a C-fiber deficit as opposed to more A-fibers.
By examining branches of the sciatic nerve, which innervate different tissues (sural, largely skin; common peroneal and tibial, skin and muscle; medial gastrocnemius and lateral gastrocnemius, muscle; Schmalbruch,1986; Swett et al.,1991; Lewin and McMahon,1991a, b), we observed that the C-fiber deficit in naked mole-rats appears to be restricted to cutaneous nerves: naked mole-rats had a significantly lower C:A-fiber ratio in the sural nerve (1.4:1, compared with 2.9–3.3:1 for other species; Fig. 2B), whereas C:A-fiber ratios in the common peroneal, tibial, medial gastrocnemius, and lateral gastrocnemius nerves were more similar across all species (Fig. 2C–F) and more similar to those ratios previously observed in the rat (Jenq et al.,1984, 1987; Jenq and Coggeshall,1984a, b, 1985a, b; Schmalbruch,1986). An exception was the significantly lower C:A-fiber ratio in the common peroneal nerve compared with Damaraland mole-rats (Fig. 2C). The common peroneal nerve innervates both skin and muscle (Schmalbruch,1986), and it is possible that, in species with particularly high common peroneal C:A-fiber ratios, a larger proportion of the common peroneal axons innervates skin than in other species.
In humans, C-fiber innervation of the skin is very dense; receptive fields overlap, and this leads to spatial summation of noxious stimuli, which may well aid high-resolution stimulus localization (Jørum et al.,1989; Ochoa and Torebjörk,1989; Koltzenburg et al.,1993; Schmidt et al.,1997). Therefore, it might be expected, based on the C-fiber deficit observed in naked mole-rats, that naked mole-rats have a hypofunctional nociceptive system. Systematic examination of the nociceptive system in the naked mole-rat demonstrated that the animals have normal nocifensive responses to heat and mechanical stimuli but that they fail to respond behaviorally to certain chemical stimuli: acid, capsaicin, and histamine (Park et al.,2008; Smith et al.,2010; Brand et al.,2010). Isolated sensory neurons are, however, responsive to both capsaicin and histamine, and it is thought that unusual connectivity in the spinal cord and an endogenous lack of neuropeptides in C-fibers may explain the lack of nocifensive and scratching behavior evoked by these substances (Park et al.,2008; Smith et al.,2010). Interestingly, acid fails to activate naked mole-rat C-fiber nociceptors (Park et al.,2008), and, although one explanation for the C-fiber deficit would be that naked mole-rats lack acid-sensitive nociceptors, it has now been shown that acid insensitivity most likely is due to a variant of the voltage-gated sodium channel (NaV) NaV1.7, which is hypersensitive to acid block (Smith et al.,2011).
It should also be noted that some unmyelinated fibers are autonomic sympathetic fibers. In the rat, the contribution of sympathetic fibers has been measured to be sural 35%, common peroneal 27%, and tibial 40% of unmyelinated fibers (Schmalbruch,1986). Therefore, naked mole-rats might have lost sympathetic fibers as opposed to C-fiber afferent nociceptors. However, a mixture of the afferent and efferent C-fiber loss more likely is due to the dramatic C-fiber loss in cutaneous nerves, which could not be accounted for by loss of sympathetic efferent fibers alone. Interestingly, in congenital insensitivity to pain with anhidrosis (hereditary sensory and autonomic neuropathy 4, HSAN4), mutations in tyrosine receptor kinase A (TrkA), the receptor for nerve growth factor (NGF), result in subjects presenting with a total loss of C-fibers and lack of nociception, and they do not sweat due to hypotrophic, uninnervated sweat glands (Indo,2009). Moreover, a newly identified loss of function NGF mutation also results in a lack of nociception and anhidrosis (Carvalho et al.,2011), whereas a second NGF mutation, which causes a lack of nociception without anhidrosis (Einarsdottir et al.,2004), is proposed to be hypofunctional. Mice, in which either NGF or TrkA has been ablated, also show a severe C-fiber loss and lack normal nocifensive responses (Smeyne et al.,1994; Crowley et al.,1994). The fact that naked mole-rats have a hypofunctional nociceptive system (Park et al.,2008) as well as lacking sweat glands (Tucker,1981) raises the possibility that hypofunctional NGF-TrkA signaling might underlie the C-fiber deficit. However, if such hypofunctional NFG-TrkA signaling exists, then the effect must be restricted to cutaneous sensory afferents.
Another possible explanation for the cutaneous C-fiber deficit might simply be the lack of hair follicles, which are normally innervated by both A- and C-fibers (Fundin et al.,1997; Park et al.,2003). Therefore, the profound absence of hair follicles may result in an absence of C-fibers because they lack their normal cutaneous target. There are very few nonaquatic mammals lacking body hair; the hairless bat Cheiromeles torquatus being one example, but these animals do have hair on their undersides (Stephen Rossiter, Queen Mary University of London; personal communication). We therefore made use of a transgenic mouse model that completely lacks hair to model the situation in the naked mole-rat. In β-cat LOF mice, hair follicle stem cells fail to differentiate into follicular keratinocytes, producing a progressive hair loss resulting from a lack of hair follicles (Huelsken et al.2001). Interestingly, in the saphenous nerve, but not in the tibial nerve, we observed a decrease in the C:A-fiber ratio and total C-fiber number in adult β-cat LOF mice compared with wild-type littermates. These results suggest that a lack of hair follicles results in a mild cutaneous C-fiber deficit, but the fact that the saphenous C:A-fiber ratio in β-cat LOF mice is still more than double that of the naked mole-rat (3.81 vs. 1.69) suggests that mechanisms other than hair follicle loss are needed to explain the cutaneous C-fiber paucity in naked mole-rats. However, it should be noted that skin from β-cat LOF mice does not perfectly model the skin of naked mole-rats, most importantly because, whereas β-cat LOF mice lose hair follicles over time, naked mole-rat skin never contains hair follicles (with the exception of guard hairs and whiskers), so it may well be that the C-fibers that develop in the β-cat LOF mice are not dependent on hair follicles for their continuing survival. Nevertheless, it is striking that there is a small but significant loss of C-fibers but not A-fibers in the saphenous nerve of β-cat LOF mice. This observation is consistent with the idea that C-fibers that innervate hair follicles (Li et al.,2011; Wende et al.,2012) depend on the follicle for continued survival but that classical A-fiber mechanoreceptors do not.
In addition to certain A- and C-fibers being activated by noxious thermal stimuli, others are thermosensitive across a range that might be considered non-noxious (Iggo,1969). A thorough investigation of naked mole-rat A- and C-fiber thermosensitivity has not yet been described, but observations of huddling behavior and movement to warmer (heated) areas of cages suggests that naked mole-rats have thermoreceptors (our personal observations). Consequently, it seems unlikely that the C-fiber paucity in naked mole-rats reflects a loss of thermoreceptors, but only further investigation can fully answer this question.
A-fiber characteristics in Bathyergidae
The three largest branches of the sciatic nerve are the tibial, common peroneal, and sural, and in the rat the number of A-fibers and C-fibers follows the order tibial > common peroneal > sural (Schmalbruch,1986). We found the same pattern across all six species, with the exception of the silvery mole-rat, in which there were more C-fibers in the sural nerve than in the common peroneal nerve (Tables 1, 2). It has also been documented in rat that lateral gastrocnemius nerves contain more A- and C-fibers than medial gastrocnemius nerves (Jenq and Coggeshall,1985a, b), and this was also observed here for all six species studied (Table 2).
In keeping with the presence of large-diameter motor neurons and type Ia sensory afferents in nerves innervating muscle (Sherrington,1894; Boyd and Davey,1968), we observed that A-fibers present in common peroneal, tibial, medial gastrocnemius, and lateral gastrocnemius nerves generally had larger diameters than A-fibers in saphenous and sural nerves, those fibers of the medial gastrocnemius nerve being the largest in every species (Fig. 3). The trend of muscle-innervating nerves containing larger-diameter A-fibers is not, however, fully apparent in the naked or Damaraland mole-rat: saphenous A-fibers do have smaller diameters than all other nerves, but sural nerves have diameters similar to those of both common peroneal and tibial nerves (Fig. 3, Tables 1, 2). In the rat sural nerve, ∼10% of myelinated fibers are motorneurons, innervating muscles in the foot (Peyronnard and Charron,1982), but it is possible that, in naked and Damaraland mole-rats, the percentage of sural A-fibers, which innervate muscle, is higher than in other species, giving rise to a larger average A-fiber diameter.
With respect to g-ratios, it has been calculated for myelinated nerves that the optimal ratio for conduction of current from one node to the next is 0.6 (Rushton,1951). Although higher average g-ratios have been observed in various A-fibers across different species (Williams and Chalupa,1983; Guy et al.,1989; Fraher and O'sullivan,2000), in the sciatic nerves in rodents g-ratios have been found to be closer to the theoretical optimum of 0.6 (Schwab et al.,1984; Sterne et al.,1997; Willem et al.,2006), but in other species they can be higher, for example, 0.8 in the European common frog Rana temporaria (Friede et al.,1985). In this study, we found that, as in other rodents, g-ratios in all branches of the sciatic nerve examined were ∼0.6 (Tables 1, 2). Size frequency distributions of A-fiber axon diameters in the saphenous nerve and other nerves showed a bimodal distribution in all mole-rat species (Fig. 4 for the saphenous nerve; other data not shown), which is likely reflective of Aα/β and Aδ fiber types, as has been previously observed in other rodents (Scadding,1980; Lynn,1984; Schwab et al.,1984; Schmalbruch,1986).
C-fiber characteristics in Bathyergidae
Similarly to A-fibers, C-fiber diameter was generally positively correlated with species size (Fig. 5, Tables 1, 2). Furthermore, as with A-fibers, there was a trend for muscle innervating nerves to have larger C-fiber diameters (Fig. 5, Tables 1, 2). For the rat, Schmalbruch (1986) found there to be only a 0.03-μm difference in the means for C-fiber diameter in sural, common peroneal, and tibial nerves. In the present study, the range in the mean C-fiber axon diameter from sural, common peroneal, and tibial nerves was from 0.06 μm in the Damaraland mole-rat to 0.12 μm in the silvery mole-rat (Tables 1, 2). When it has been investigated in other rodent species, the C-fiber axon diameter distribution has been observed to be unimodal in saphenous (Scadding,1980; Alpsan and Lal,1980; Lynn,1984; Illanes et al.,1990), sural (Ochoa and Mair,1969; Schwab et al.,1984; Schmalbruch,1986; Hoffmeister et al.,1991), and common peroneal and tibial nerves (Schmalbruch,1986). We could confirm in every nerve, from all species examined, that there was a unimodal distribution for C-fiber diameter.
Although we observed a C-fiber deficit in naked mole-rat saphenous and sural nerves, we did not observe any difference in the number of C-fibers per Remak bundle (Fig. 7). This would suggest that factors known to be involved in normal Remak bundle formation, such as neuregulin-1 (NRG-1), function normally in naked mole-rats (Taveggia et al.,2005; Willem et al.,2006). Indeed, naked mole-rats apparently express high levels of NRG-1 in the nervous system throughout their normal life span (Edrey et al.,2012). Therefore, we can state that naked mole-rat C-fibers appear morphologically normal compared with those of the other species examined. It has long been known that NGF levels are lower in muscle tissues than they are in the skin (Korsching and Thoenen,1983; Shelton and Reichardt,1984; Lewin et al.,1992). It is thus possible that interference with NGF signaling in naked mole-rats might bring about a selective reduction of C-fiber in cutaneous nerves. However, the hypothesis that the C-fiber deficit observed in naked mole-rat cutaneous nerves is due to hypofunctional NGF-TrkA signaling should be tested more directly. One approach would be to clone and characterize the naked-mole rat NGF and TrkA genes to examine whether these proteins function differently compared with those of other rodent species. However, a detailed examination of the development of the sensory innervation of the skin in naked mole-rats, as has been conducted in the mouse (Crowley et al.,1994; Smeyne et al.,1994; Lechner et al.,2009), would be very difficult, given the eusocial nature of this species and the very long gestation time (∼75 days). It is still also possible that genes involved in the differentiation of sensory neuron lineages are also altered in the naked mole-rat in a way that leads to a selective C-fiber deficit in cutaneous nerves (Marmigère and Ernfors,2007).