2011 touch gives new life

ORIGINAL ARTICLE

Touch gives new life: mechanosensation modulatesspinal cord adult neurogenesisR Shechter1, K Baruch1, M Schwartz1,3 and A Rolls2,3

1Department of Neurobiology, The Weizmann Institute of Science, Rehovot, Israel and 2Department of Psychiatry andBehavioral Sciences, Stanford University, Stanford, CA, USA

The ability to respond to a wide range of novel touch sensations and to habituate uponrepeated exposures is fundamental for effective sensation. In this study we identified adultspinal cord neurogenesis as a potential novel player in the mechanism of tactile sensation. Wedemonstrate that a single exposure to a novel mechanosensory stimulus induced immediateproliferation of progenitor cells in the spinal dorsal horn, whereas repeated exposures to thesame stimulus induced neuronal differentiation and survival. Most of the newly formedneurons differentiated toward a GABAergic fate. This touch-induced neurogenesis reflectedthe novelty of the stimuli, its diversity, as well as stimulus duration. Introducing adultneurogenesis as a potential mechanism of response to a novel stimulus and for habituation torepeated sensory exposures opens up potential new directions in treating hypersensitivity,pain and other mechanosensory disorders.Molecular Psychiatry (2011) 16, 342352; doi:10.1038/mp.2010.116; published online 16 November 2010

Keywords: adult neurogenesis; enriched environment; habituation; mechanosensory; spinalcord; touch

Introduction

Our world is surrounded by various mechanosensoryinputs, which are captured by our largest sensoryorganthe skin. We are constantly exposed to bothnovel and familiar mechanosensory inputs such asshaking hands, putting on a piece of clothing orwalking barefoot on the grass. These stimuli vary intheir duration, novelty and their functional signifi-cance; some are persistent stimuli, like the touch ofthe grass under ones feet, whereas others are moreacute, like a bug suddenly crawling on ones leg.Obviously, these signals require different levels ofexcitability or habituation by the nervous system.Habituation in sensory systems is a key to effective

perception; among other features, it allows bettersignal to noise detection. Many of the habituationmechanisms involve desensitization at the level of thesingle sensory cell, as in the case of the eyes rod cells,along with higher-level processes and system plasti-city. In the mechanosensory system, a large body ofevidence supports the existence of plasticity andhabituation process at the single cell level, mostly inthe form of receptor sensitivity.1 In this study weintroduce a new level of plasticity in the mechan-

osensory modality in the form of adult neurogenesis,manifested in the dorsal horn of the spinal cord.Recent studies have identified neurogenesis in the

adult spinal cord, both in healthy and in diseasedtissues.28 Although the origin of the newly formedcells and their potential functional role is stillunknown, adult neurogenesis of the spinal cord isreported to be mostly confined to the dorsal horn.2,4

As mechanosensory information travels from the skinto the sensory cortex through specialized interneur-ons in this region,911 and based on the induction ofneurogenesis in another sensory system, the olfactorybulb,1216 we hypothesized that adult spinal cordneurogenesis may be involved in the response totouch and to mechanosensory inputs.By exposing mice to novel mechanosensory stimuli

for varying periods of time and of different diversity, wefound that novel sensory stimuli induce progenitor cellproliferation in the sensory dorsal pathway of the adultspinal cord, and that the duration of the stimulationand its diversity affect their subsequent g-aminobutyricacid (GABA)ergic neuronal differentiation and survi-val. Thus, our results are of potential functionalrelevance to neurogenesis in the dorsal pathway ofadult spinal cord, with likely implications to plasticityand habituation in mechanosensation and touch.

Materials and methods

Animals

C57BL/6J mice were supplied by the Animal BreedingCenter of the Weizmann Institute of Science. Animals

Received 31 December 2009; revised 29 August 2010; accepted 22September 2010; published online 16 November 2010

Correspondence: Professor M Schwartz, Department of Neuro-biology, The Weizmann Institute of Science, Herzl Street, Rehovot76100, Israel.E-mail: [email protected] authors contributed equally to this work.

Molecular Psychiatry (2011) 16, 342352& 2011 Macmillan Publishers Limited All rights reserved 1359-4184/11

www.nature.com/mp

were housed from birth under standard housingconditions. For all experiments, 8-week-old malemice were used. Mice were handled according tothe regulations formulated by the Weizmann Insti-tutes Animal Care and Use Committee.

Sensory-enriched environment (SEE)Mice were kept in cages covered with standardbedding. For the SEE sessions, mice were moved intonew cages, the bottoms of which were covered withdifferent types of textures: sandpapers comprisingdifferent sizes of embedded particles (coarse (p20),medium (p60) and fine (p150)), gravel and sponge.Each group was exposed to a different combinationof these textures: cages in the group defined asmultiple comprised sections with each of thesedifferent textures, whereas cages in the uniformgroup were covered with a single sensory texture,p60 sandpaper (Figure 1a). Mice in the altered groupwere placed each day in a cage covered with a differentsingle textured material, either the sandpapers orgravel or sponge. The duration of exposure to theSEE is indicated in text. In general, for single exposure,mice were placed in the SEE for 2h; for repeatedexposures, mice were exposed to SEE for 2h, twice aday with an 8-h interval for 7 consecutive days; and inthe group of continuous exposure, animals werepermanently housed in the SEE cages. As socialbehavior and animal handling has been previouslyshown to affect neurogenesis, at least in the hippo-campus,17 in all the described experiments, theanimals were maintained in fixed groups of fiveanimals and housed in routine housing conditions(cage size: 32 1612.5cm). Water and food werefreely available in their standard positions. The controlgroup was housed under similar conditions as the SEEgroup, and for the training session mice were placed ina new cage, covered with standard bedding.

BrdU administration5-Bromo-2-deoxyuridine (BrdU; Sigma-Aldrich,St Louis, MO, USA) was dissolved by sonicationin phosphate-buffered saline and injected intraper-itoneally (75mgkg1 body weight) twice daily for 7days immediately following the SEE exposure, or 14days before the SEE exposures (as indicated in thetext); animals were killed 14 or 28 days after the firstSEE exposure. For identification of the stage follow-ing exposure in which the new cells formed, the micewere treated with a single administration of BrdU(300mgkg1 body weight) at different time pointsfollowing the exposure, and analyzed 8h after theinjection. BrdU dose was chosen based on previousstudies that suggested that X300mgkg1 is necessaryto obtain efficient labeling of most of the S-phasecells;18,19 such labeling is crucial for comparingbetween two different groups. This dose was pre-viously shown to be nontoxic,4,20 and was previouslyused to measure cell proliferation in mice. Indeed, inour experiments, no behavioral or physiological sideeffects were noticed.

ImmunohistochemistryFollowing phosphate-buffered saline perfusion,spinal cords were postfixed with Bouins fixative(Sigma-Aldrich) for 48h and then embedded inparaffin. Paraffin sections, 6-mm thick, were usedthroughout the study. The paraffin was removed bysuccessive rinsing of slides with a gradient of xyleneand ethanol. Following antigen retrieval by heating,the slides were blocked with blocking solution(20% horse serum with 0.3% Triton) and incubatedfor 48h with specified combinations of the followingprimary antibodies. For BrdU staining, the slideswere incubated in 2N HCl at 37 1C for 30minfollowing microwave treatment. The following pri-mary antibodies were used: rat anti-BrdU (1:100;Oxford Biotechnology, Oxford, UK), goat anti-DCX(1:100; Santa Cruz Biotechnology, Santa Cruz, CA,USA), rabbit anti-GABA (1:500; Sigma-Aldrich),rabbit anti-GAD65/67 (1:100; Abcam, Cambridge,MA, USA), rabbit anti-calretinin (1:1000; Chemicon,Temecula, CA, USA), rabbit anti-calbindin (1:200;Cell Signaling Technology, Beverly, MA, USA), rabbitanti-Sox-2 (1:150; Abcam), rabbit anti-NSE (1:50;Millipore, Billerica, MA, USA), mouse anti-HuC/D(1:50; Molecular Probes, Eugene, OR, USA) andrabbit anti-NG2 (1:150; Millipore). After rinsing inphosphate-buffered saline, sections were incubatedfor 1h with the appropriate secondary antibodies(1:200; Jackson Immunoresearch Laboratories, Cam-bridgeshire, UK). For nuclear labeling, Hoechst 33 342staining (1:2000; Molecular Probes) was done beforemounting and covering.

QuantificationFor microscopic analysis, a fluorescence microscope(E800; Nikon, Tokyo, Japan) or laser-scanning con-focal microscope (Carl Zeiss MicroImaging GmbH,Jena, Germany) was used. An observer, blind to theidentity of the samples, counted the number oflabeled cells from a total of 18 coronal spinal cordsections per mouse, taken from six different locationsseparated by 0.5 cm along the spinal cord. To obtainan estimate of the number of labeled cells per mm3

volume, the average number of cells counted in theselected sections (average surface area = 1mm2, thick-ness = 0.006mm) was multiplied by 166.66. Hoechststaining was routinely used for nuclear labeling,which served to verify quantification of the cells.Cell counting was performed using high-magnifica-tion imaging (objective 40). In order to avoidoverestimation because of counting fragments of cellsthat spanned several sections, only cells that had anintact morphology and a nucleus that was > 5 mm indiameter were counted. Confocal Z-sectioning wasperformed in order to verify double labeling. Thefluorescence microscope was equipped with a digitalcamera (DXM 1200F; Nikon) and with either a 20NA 0.50 or 40 NA 0.75 objective lens (Plan Fluor;Nikon). The confocal microscope was equipped withLSM 510 laser scanning (three lasers: Ar 488, HeNe543 and HeNe 633) and with a 40 oil-immersion

Touch modulates spinal neurogenesisR Shechter et al

343

Molecular Psychiatry

NA 1.3 Plan Neofluor objective lens. Recordings weremade on postfixed tissues at 24 1C using acquisitionsoftware (ACT-1 (Nikon); or LSM (Carl Zeiss)).

Statistical analysis

Data were analyzed using Students t-test or factorialanalysis of variance followed by Fishers least


344


significant difference procedure. Error bars represents.e.m. Statistical data that are given in the textrepresent averages.e.m.

Results

Single exposure to novel tactile stimuli inducesimmediate cell proliferation, which depends on thediversity and intensity of the sensory stimuliBased on our hypothesis that neurogenesis may beinvolved in mechanosensation, we first examinedwhether a novel tactile stimulation would inducechanges in adult spinal cord neurogenesis. We there-fore placed adult mice in a SEE consisting of cagescovered with novel textures (sandpaper, gravel orsponge; Figure 1a) for 2h. Immediately before theSEE exposure and for an additional 6 subsequent days,the mice were injected with BrdU (75mgkg1) twice aday to label proliferating cells. Mice were analyzed 14days after exposure to SEE. Using immunohistochem-ical analysis, we found that a 2-h exposure to SEEresulted in a 2.5-fold increase in the number of cellslabeled with BrdU in the dorsal horn (n=35 pergroup; Figures 1b and c), indicating cellular prolifera-tion. To determine whether the degree of proliferationreflects the intensity and diversity of the novel stimuli,we compared two different SEE paradigms; uniform,composed of the single novel sensory texture (sandpa-per), and multiple, comprising several different noveltextures (Figure 1a). The exposure to multiple sensoryinputs induced more intense cell proliferation (1.4-fold increase, multiple vs uniform; Figure 1c), suggest-ing that cell proliferation not only reflected the noveltyof the sensory stimuli, but also its diversity.In order to gain a better insight into the nature of

this cell-renewal process, we aimed to identify thespecific time frame in which proliferation took placein response to the SEE. As in the above-describedexperiment, the mice were analyzed only at a singletime point, 14 days after the SEE. To this end, werepeated the experiment and administered a single

dose of BrdU (300mgkg1), immediately, 24h or 7days after the exposure to a single touch stimulus. Ineach case, the animals were analyzed for cell renewal8h following the BrdU injection. We found that theexposure to a novel touch sensation elicited animmediate proliferative response, which was detectedas early as 2h after the exposure, and dramaticallydeclined thereafter (n=35 per group; Figure 1d).To determine whether these proliferating cells

expressed cellular markers associated with neuralprogenitor cells, we co-labeled the cells with Sox-2, atranscription factor required for stem-cell mainte-nance in the central nervous system (Figure 1e).13 Asmall fraction (81%) of the BrdU cells was alsofound to be positive for this progenitor cell marker.Although the exposure to SEE did not affect thepercentage of this population, the numbers ofBrdU/Sox2 cells were increased following thesensory stimulus (15532 versus 27932, controland SEE, respectively; P

Repeated exposure to the same stimuli inhibits cellproliferation and induces neuronal differentiation ofnewly formed cells

To examine whether repeated exposures to the SEEwould induce a form of habituation, we introducedmice repeatedly to the same touch stimulus. Micewere exposed to the SEE in four different regimes:

single exposure (as in the previous experiments),repeated exposures (two SEE sessions a day separatedby an 8-h interval, and repeated for 7 consecutivedays), continuous exposure to a uniform texture(animals were permanently housed in the SEE cages),and continuous exposure to a stimulus that wasaltered on a daily basis (the SEE housing cages were

Figure 2 Repeated exposures to sensory stimulation induce neuronal differentiation. (a, left panel) The experimentalparadigm. (a, right panel) The quantification of 5-bromo-2-deoxyuridine (BrdU)-labeled cells in the gray matter of the spinalcord as a function of the duration/frequency of exposure to sensory-enriched environment (SEE); SEEsingle is single exposureto a uniform texture; SEERepeated is repeated exposures, with equal intervals between exposures, to a uniform texture;SEEContinuous is housing of mice in SEE cases of uniform texture; or a texture that was rotated daily (Continuousaltered). BrdU(75mgkg1) was administrated twice a day for 7 consecutive days, starting from the first SEE (n=35 per group; factorialanalysis of variance (ANOVA); F=11.75; P=0.0001). (b) Quantification of doublecortin (DCX)-labeled cells in the dorsal graymatter of the spinal cord of the control and sensory-enriched mice (n=48 per group; factorial ANOVA; F=12.15; P=0.0009followed by Fishers test; *P

altered, with daily introduction of novel stimuli;altered; Figure 2a, left panel). We found that incontrast to the increased proliferative responseachieved by increasing the number of concurrentsignals (uniform vs multiple sensory stimuli;Figure 1c), repeated exposures to a single texturehad no effect on proliferation, and continuousexposure to the same SEE even inhibited it byB50% (n=35 per group; Figure 2a, right panel).However, by altering the sensory stimuli daily for themice housed in the SEE cages, the inhibitory effect onthe proliferation was attenuated (n=35 per group;Figure 2a, right panel).In other neurogenic niches, it was shown that a

stimulus can also induce differentiation, especially ifit is presented after the proliferation; therefore, weconsidered the possibility that the lack of prolifera-tion might be because of increased differentiation. Toexamine if differentiation indeed took place, weexamined the expression of the immature neuronalmarker DCX in the dorsal horn of the spinal cord.Repeated exposures to SEE increased the number ofcells that expressed DCX (n=48 per group; Figures2be). Combined analysis with NG2 showed anincrease in the numbers of DCX/NG2 cells(1143129 vs 4267532, control vs SEE repeated,respectively, P=0.002). DCX/NG2 cells were sug-gested to be committed to a neuronal lineage derivedfrom the NG2/DCX cells,22 a population that alsoincreased following single SEE. Moreover, the frac-tion of this population (DCX/NG2) out of the DCX

cells increased following the repeated exposures(433 vs 723%, control vs SEE repeated, respec-tively, P=0.001).To characterize the nature of the signal that induces

this differentiation, we used the same paradigm asabove and introduced mice to repeated exposures ofdifferent novelties (Figure 2f) including cages coveredwith: (1) multiple tactile textures (sandpapers ofdifferent roughness, gravel and sponge; multiple);(2) a single tactile texture (medium roughnesssandpaper, uniform); and (3) alternating textureson a daily basis (one of the sandpapers, gravel orsponge; altered). Although the number of exposuresand the time spent in the SEE in all groups weresimilar, significant differences were found in thenumbers of newly formed neurons under the differentconditions. Thus, mice that were exposed repeatedlyto either a single stimulus (uniform) or to a cagecontaining a multiple but fixed set of textures(multiple) showed a similar increase in differentia-tion relative to mice maintained in standard housing(n=5 per group; Figure 2g). Interestingly, the groupthat was exposed to repeated stimuli, the nature ofwhich was changed on a daily basis (altered),exhibited a significantly greater increase in neuronaldifferentiation (n=5 per group; Figure 2g). Thus,this finding suggested that the novelty of themechanosensory stimulus is an important factor indetermining the number of differentiating newlyformed neurons.

Taking into consideration the dynamic nature ofthis process, we aimed to determine whether pro-longed exposure induced both proliferation andsubsequent differentiation of the same proliferatingcells, or whether it induced differentiation of thepre-existing pool of proliferating cells. To this end, wefirst created a new labeled pool of cells by injection ofBrdU (75mgkg1; for 7 consecutive days, twice a day)to naive animals. The mice were exposed to repeatedSEE 14 days later. We found a greater than twofoldincrease in the BrdU-labeled neurons (DCX/BrdU

cells) in the group prelabeled with BrdU comparedwith the group injected with BrdU concomitantlywith the SEE exposure (DCX/BrdU cells out ofDCX ; 10.32.1 vs 4.51.0%, respectively; n=4 to 5per group). These findings thus indicated thatrepeated sensory stimuli induced neuronal differen-tiation from both newly proliferated progenitors andpre-existing ones.

The newly formed cells are mostly localized in the graymatter of the lumbar segments of the dorsal hornTaken together, our results identified spinal cordadult neurogenesis as a novel aspect of touchsensation plasticity, highly sensitive to the natureand duration of the stimuli. Brief exposure to a novelstimulus induced immediate cell proliferation in thedorsal horn of the spinal cord, whereas prolongedexposure affected the differentiation of the newlyformed cells into immature (DCX ) neurons. Todetermine whether these newly formed immatureneurons were preferentially formed at specific loca-tions, we spatially characterized these proliferationand differentiation events. Analyzing the cells 8 hafter exposure to a single novel stimulus revealed thatthe majority of the proliferating (BrdU ) cells werelocated in the gray matter (n=5; Figure 3a). Similarly,analyzing the cells 14 days after sensory stimulirevealed that the immature DCX neurons producedfollowing repeated SEE exposures were alsomainly located in the gray matter (n=5; Figure 3b)of the dorsal horn (n=5; Figure 3c). Repeated SEEexposures specifically increased the numbers ofimmature neurons in this area (n=5 per group;Figures 3d and e) as well as in the central canal(n=5 per group; Figures 3f and g), which is known tomaintain neural progenitor cells.2,5 Analysis of con-secutive sections along the anteriorposterior axisrevealed that the increase in neurogenesis was moststrongly evident in sections derived from the lumbarsegments (n=5; Figure 3h).

Most of the newly formed cells differentiate intoGABAergic immature neuronsAlthough the role of adult neurogenesis in the centralnervous system is still not clear, identifying itsspecific neuronal differentiation might shed somelight on their potential role in the central nervoussystem. We hypothesized that at least some portion ofthe newly formed cells described above eventuallydifferentiate into a GABAergic fate. This was sup-


347


ported by several lines of evidence. First, in theolfactory system, where neurogenesis continuouslytakes place, most of the immature neurons that reachthe olfactory bulb differentiate into neurons expres-sing GABAergic granules.23 Second, recent reportssuggested that progenitor cells expressing NG2 anddifferentiating into neurons appear to differentiateinto GABAergic interneurons.24,25 Finally, a previousreport suggested the presence of immature GABAergicneurons in the dorsal horn of the spinal cord.4

To test our hypothesis and examine the cellular fateof the newly formed cells, we first verified that thesecells indeed show markers of fully committedneurons. We used human neuronal protein (HuC/D)and neuronal-specific enolase (NSE), the markers thatappear only in neurons. Analyzing cells 28 days afterthe first BrdU injection (75mgkg1, administratedtwice a day for 7 consecutive days), which corre-sponds to 22 days after the last exposure to the novelsensory stimuli, enabled us to detect 55 co-labeledcells in 1mm3 that were positive for both BrdU andHuC/D or NSE (Figures 4a and b). Further analysisrevealed BrdU cells that coexpressed GABA and theenzyme responsible for production of this neuro-transmitter, glutamic acid decarboxylase (GAD65/67)(Figures 4c and d). Among the BrdU cells, 534.5%expressed GABA. Some of these newly formed

cells were also co-labeled with the calcium-bindingprotein, calretinin (Figure 4e), but we could notdetect any cell that was co-labeled with calbindin(not shown), potentially indicating a specializedGABAergic differentiation state, or specific matura-tion phase.To test whether sensory exposure affects the

survival rate of the newly formed cells, we evaluatedthe number of surviving BrdU-positive cells 28 daysafter the first SEE exposure (following a single300mgkg1 injection of BrdU given immediatelybefore the first exposure to SEE). Similar to the effectof exposure to novel odors on neurogenesis in theolfactory bulb, we found that repeated exposures tothe same stimuli, unlike a single exposure, led to anoverall twofold increase in the number of survivingnewly formed cells (n=6 per group; Figure 4f). Toquantify the number of newly formed GABAergiccells, we repeated the above-described experimentwhile using a protocol of repeated BrdU injections(75mgkg1; 7 consecutive days), which improves thedetection of such cells. Repeated exposures resultedin an increased number of newly formed (BrdU )GABAergic cells (n=6 per group; Figure 4g). Thesefindings thus indicate that repeated exposure tosensory stimuli supports the survival of the newlyformed GABAergic immature neurons.

Figure 3 The newly formed cells are mostly localized in the gray matter of the dorsal horn of the lumbar segments.(a) Quantification of the number of 5-bromo-2-deoxyuridine (BrdU)-labeled cells in the dorsal region (gray matter (black bars)and white matter (white bars)) of the spinal cord of mice exposed to a single session of SEEmultiple and injected with a singleinjection of BrdU (300mgkg1) immediately before the sensory-enriched environment (SEE; immediate). No changes wereseen in the number of proliferating cells in the white matter. Enhanced proliferation was attributed to proliferating cells inthe gray matter. (be) Numbers of newly formed neurons (b) in the gray vs white matter, and (c) in the dorsal vs ventral hornsin the SEERepeated mice (n=5; Students t-test; (b) ***P=0.007; (c) **P=0.01). Relative distribution of doublecortin-positive(DCX ) cells (d) in the gray vs white matter, and (e) in the dorsal vs ventral horns in mice that were exposed to SEERepeated,and in control mice (n=5 per group; Students t-test; (d) *P=0.01; (e) **P=0.003). (f) DCX-positive (red; marked by fullarrow) and BrdU-positive (green; marked by arrow head) cells in the central canal of naive mice, and mice that were exposedto SEERepeated (scale = 10 mm; magnification 40). (g) Number of BrdU- and DCX-positive cells in the central canal of bothgroups (mean numbers of cells per 1mm cord length; Students t-test; n=5 per group; BrdU: P=0.7; DCX: *P=0.01).(h) Quantitative analysis of DCX -labeled cells in the dorsal gray matter in coronal sections taken from the cervical, thoracicand lumbar areas of the spinal cords (n=5; factorial analysis of variance (ANOVA); F=3.802; P=0.01; followed by Fisherstest, *P

Discussion

Our data suggest spinal cord adult neurogenesis asa novel player in touch sensation, responsiveto different aspects of the mechanosensory input,including the novelty of the stimulus, its duration andits diversity. We showed that single exposure tonovel mechanosensory stimulus resulted in an in-crease in the number of proliferating cells in thedorsal horn of the adult rodent spinal cord incorrelation with the diversity of the new stimulipresented (higher diversity resulted in a greaternumber of proliferating cells). The proliferativeresponse of the progenitor cells was found to be animmediate one, and was evident as early as 2hfollowing exposure to the novel stimuli (the effectdramatically declined thereafter). Furthermore, re-peated exposures to the same stimulus inducedneuronal differentiation and survival of the progeni-tor cells, possibly reflecting a mechanism of habitua-tion (Figure 5).In contrast to the proliferation phase, which was

found to be a reflection of both the novelty anddiversity of the stimuli, the extent of neuronaldifferentiation depended mainly on the novelty ofthe stimuli (repeated exposures to various novelstimuli presented together (multiple) had a weakereffect on the differentiation compared with present-

ing the same number of novel stimuli on a rotatingbasis (altered). Notably, the cellular fate of thesecells was mostly GABAergic. Our observation thatprolonged exposures to the same stimuli increasedthe number of GABAergic cells might reflect ahabituation or de-sensitization processes related tothese stimuli. This possibility is especially attractivein light of the fact that in the superficial lamina of thedorsal spinal cord, many inhibitory interneuronsexpress and respond to GABA.26

The origin of the newly formed cells is not clear. Itwas first suggested by Horner et al.2 that cellscontinuously proliferate at the outer circle of theadult spinal cord. Further characterization of thisniche by others revealed that these cells routinelydifferentiate to immature neurons and are predomi-nately found in the dorsal gray matter.4 The exposureto novel stimuli results in an immediate proliferationof cells, the majority of which are NG2 positive. Asreviewed by Nishiyama et al.,27 the lineage progres-sion of NG2 cells, which are the largest pool ofpostnatal proliferative progenitors, is a generallycontroversial issue. NG2-expressing cells, originallysuggested to serve only as oligodendrocyte precur-sors, are now considered to be a heterogeneouspopulation. Several recent studies have suggestedthat at least some of these assumed oligodendrocyteprecursors retain the ability to generate interneurons

Figure 4 The newly formed cells differentiate into g-aminobutyric acid (GABA)ergic neurons. Animals were subjected torepeated exposures to sensory-enriched environment (SEE), and analyzed 28 days after first exposure. 5-Bromo-2-deoxyuridine (BrdU) was administered at the time of SEE exposure. (ae) Double labeling for BrdU and (a) human neuronalprotein (HuC/D); (b) neuronal-specific enolase (NSE); (c) GAD65/67; (d) GABA; and (e) calretinin (CR) (scale: (a, b) 10mm;(ce) 50 mm; boxed: 10mm; all at magnification 40). Double-positive cells are marked by full arrows, whereas arrowheadsindicate BrdU cells that do not express the other tested neuronal marker. Confocal Z-axis scan of single cells positive for theindicated markers are provided. (f) Number of BrdU cells in the dorsal horn (following single injection of 300mgkg1 BrdU,administrated once parallel to the first sensory stimuli; n=6 per group). (g) Quantification of BrdU/GABA cells in thedorsal horn 28 days following the first SEE exposure (BrdU was administrated parallel to the SEE sessions at a dose of75mgkg1), (n=6 per group; Students t-test; *P

and, thus, are suggested to have properties of multi-potential progenitors. Although the lineage progres-sion of NG2 cells toward a neuronal fate is stillunder debate, evidence from in vitro and in vivoexperiments increasingly support such a possibility. Itwas demonstrated that a sizeable fraction of postnatalNG2 proteoglycan-expressing progenitor cells in thehippocampus are proliferative precursors, whoseprogeny appear to differentiate into GABAergicneurons capable of propagating action potentialsand displaying functional synaptic inputs.24 More-over, in adult nonconstitutively active neurogenicniches, such as the neocortex, it has been shown thata fraction of NG2-expressing cells in the adult brainexpress neuronal markers, suggesting that the sourceof the newly generated GABAergic cells are NG2

progenitors.21,24 However, under normal conditions,the vast majority of NG2-positive cells in the adultneocortex that incorporate BrdU remain NG2 positivemany weeks after the BrdU injection, and theirdifferentiation to immature neurons is a relativelyrare event. Tamura et al.22 reported that such NG2-expressing multipotent progenitor cells are also DCX

positive, and that in the rare process of theirdifferentiation to neuronal committed cells, theydownregulate NG2 and maintain the expression ofDCX. In line with this suggested model of fatedecision, in the current study we found that theproliferative population in the spinal cord inresponse to SEE stimuli is mostly NG2/DCX andrepeated stimulations increase the fraction of theneuronal committed DCX/NG2 cells. Accordingly,the sensory-sensitive progenitor population in thespinal cord, described here, could represent a novelniche from which such neuronal differentiation canoccur.Most of these newly formed cells die within 28 days

(in agreement with the cell fate in other neurogenicniches, especially the dentate gyrus of the hippocam-pus where most of the newly formed cells die within4 weeks of proliferation28). However, of the remainingcells, B50% express markers associated with aGABAergic cell fate. Tamura et al.22 suggestedGABAergic differentiation as the reasonable fate ofthe rare DCX/NG2 sub-population, which werereferred to as immature neurons, although such amaturation fate was not detected in the nonstimulatedneocortex. Interestingly, Tamura et al.22 suggested thatthe low differentiation rate into GABAergic fate maybe an outcome of an environment with minimalstimulation. In apparent correlation, our observationsof increased survival of newly formed GABAergiccells following repeated sensory stimuli might con-stitute a response to a strongly stimulatory environ-ment. Further support to this claim comes from ourobservation that sensory enrichment affects therelative proportions of this newly characterizedsub-population, with an increased fraction of theDCX/NG2 subset. Further studies are needed to testwhether other enriched environmental conditions,such as social activities or exercise, can affect the fateof this rare population in other neurogenic niches.GABAergic neuronal differentiation is one of the

hallmarks that characterize adult neurogenesis in theolfactory system.12,29 Interestingly, olfactory bulbGABAergic neurogenesis is enhanced by novel odors.In this study we suggest that in analogy to olfactoryneurogenesis, which is associated with odor discri-mination30,31 and with mating behavior,32 neurogen-esis in the dorsal horn of the adult spinal cord isrelated to the ability to discriminate and/or habituatedifferent mechanosensory stimuli. In analogy to theeffects of exposure to novel odors,30,31 SEE prolongedthe survival of the newly formed GABAergic cells inthe spinal cord. The role of the newly formedGABAergic immature neurons is, as yet, unclear.Unlike their olfactory counterparts, which wereshown to be capable of differentiating to maturecircuit-integrated interneurons, we did not observeany mature neurons such as NeuN/BrdU co-labeled cells after 4 weeks. It is possible that in thespinal cord neurogenic niche, the newly generatedGABAergic immature neurons serve a transientneuromodulatory role, which supports a mechanism

Figure 5 Schematic representation of spinal cord adultneurogenesis in the sensory-enriched environment (SEE)model. Single exposure to a novel mechanosensory stimu-lus induces immediate proliferation from an apparentreservoir of progenitor cells (DCX/NG2 ; Sox2 ) in theadult spinal cord. In contrast, repeated exposure to the samestimulus induces neuronal differentiation and survival ofthe new cells, mostly into immature (DCX , NG2, NeuN,Calbindin) GABAergic (GABA , GAD 65/67 ) neurons(NSE , HuC/D , Calretinin ), and inhibits their additionalproliferation.


350


in which the niche plasticity undergoes fine tuningthat is regulated both temporally and spatially.The similarity between the neurogenic response to

novel odors and the neurogenic response to themechanosensory stimulation described in the presentstudy might reflect a more general phenomenon ofplasticity in sensory organs, although with differentspatial and temporal dynamics. Such a mechanismmay perhaps be of greater relevance to sensations thatare subject to background noise in the sensory inputand those that involve continuous and uninterruptedinput, such as hearing (background noise that iscontinually present), olfaction (a lingering smell inthe room) and mechanosensation (the sensationderived from extended contact with everyday objects,such as clothes); these sensory organs are likely toutilize adult neurogenesis as part of their plasticitymechanism. This is in contrast to the fast dynamics ofstimulus presentation in vision, which would benefitless from the use of neurogenesis in the encodingprocess.Therapeutically, the formation of new, mostly

GABAergic, neurons in the dorsal horn of the adultspinal cord, where nociceptive fibers terminate, mayalso be relevant for pain transduction, especially forallodynia, a syndrome characterized by sensations ofpain in response to non-painful stimuli, and waspreviously shown to be ameliorated by transplanta-tion of stem cells.33 Such pathology may potentiallyinvolve a maladaptive neurogenic process in thisregion of the spinal cord. Other conditions, whichinvolve a mechanosensory component, are alsopotential subjects of intervention. Further under-standing of this neurogenic process, responsive tomechanosensory inputs, may open new opportunitiesfor noninvasive therapy.

Conflict of interest

The authors declare no conflict of interest.

Acknowledgments

We thank Shelley Schwarzbaum for editing themanuscript. MS holds the Maurice and Ilse KatzProfessorial Chair in Neuroimmunology. This workwas supported in part by the High Q foundation andby IsrALS (to MS).

References

1 Bounoutas A, Chalfie M. Touch sensitivity in Caenorhabditiselegans. Pflugers Arch 2007; 454: 691702.

2 Horner PJ, Power AE, Kempermann G, Kuhn HG, Palmer TD,Winkler J et al. Proliferation and differentiation of progenitor cellsthroughout the intact adult rat spinal cord. J Neurosci 2000; 20:22182228.

3 Ke Y, Chi L, Xu R, Luo C, Gozal D, Liu R. Early response ofendogenous adult neural progenitor cells to acute spinal cordinjury in mice. Stem Cells 2006; 24: 10111019.

4 Shechter R, Ziv Y, Schwartz M. New GABAergic interneuronssupported by myelin-specific T cells are formed in intact adultspinal cord. Stem Cells 2007; 25: 22772282.

5 Shihabuddin LS, Horner PJ, Ray J, Gage FH. Adult spinal cordstem cells generate neurons after transplantation in the adultdentate gyrus. J Neurosci 2000; 20: 87278735.

6 Vessal M, Aycock A, Garton MT, Ciferri M, Darian-Smith C. Adultneurogenesis in primate and rodent spinal cord: comparinga cervical dorsal rhizotomy with a dorsal column transection.Eur J Neurosci 2007; 26: 27772794.

7 Weiss S, Dunne C, Hewson J, Wohl C, Wheatley M, Peterson ACet al. Multipotent CNS stem cells are present in the adultmammalian spinal cord and ventricular neuroaxis. J Neurosci1996; 16: 75997609.

8 Yang H, Lu P, McKay HM, Bernot T, Keirstead H, Steward O et al.Endogenous neurogenesis replaces oligodendrocytes andastrocytes after primate spinal cord injury. J Neurosci 2006; 26:21572166.

9 Brodin L, Christenson J, Grillner S. Single sensory neuronsactivate excitatory amino acid receptors in the lamprey spinalcord. Neurosci Lett 1987; 75: 7579.

10 Dykes RW, Craig AD. Control of size and excitabilityof mechanosensory receptive fields in dorsal column nucleiby homolateral dorsal horn neurons. J Neurophysiol 1998; 80:120129.

11 Schneider SP. Mechanosensory afferent input and neuronal firingproperties in rodent spinal laminae III-V: re-examination ofrelationships with analysis of responses to static and time-varyingstimuli. Brain Res 2005; 1034: 7189.

12 Akiba Y, Sasaki H, Huerta PT, Estevez AG, Baker H, Cave JW.gamma-Aminobutyric acid-mediated regulation of the activity-dependent olfactory bulb dopaminergic phenotype. J Neurosci Res2009; 87: 22112221.

13 Favaro R, Valotta M, Ferri AL, Latorre E, Mariani J, Giachino Cet al. Hippocampal development and neural stem cell main-tenance require Sox2-dependent regulation of Shh. Nat Neurosci2009; 12: 12481256.

14 Ma DK, Kim WR, Ming GL, Song H. Activity-dependent extrinsicregulation of adult olfactory bulb and hippocampal neurogenesis.Ann NYAcad Sci 2009; 1170: 664673.

15 Ortega-Perez I, Murray K, Lledo PM. The how and why of adultneurogenesis. J Mol Histol 2007; 38: 555562.

16 Gould E. How widespread is adult neurogenesis in mammals? NatRev Neurosci 2007; 8: 481488.

17 Mitra R, Sundlass K, Parker KJ, Schatzberg AF, Lyons DM. Socialstress-related behavior affects hippocampal cell proliferation inmice. Physiol Behav 2006; 89: 123127.

18 Cameron HA, McKay RD. Adult neurogenesis produces a largepool of new granule cells in the dentate gyrus. J Comp Neurol2001; 435: 406417.

19 Cooper-Kuhn CM, Kuhn HG. Is it all DNA repair? Methodologicalconsiderations for detecting neurogenesis in the adult brain. BrainRes Dev Brain Res 2002; 134: 1321.

20 Bauer S, Patterson PH. The cell cycle-apoptosis connectionrevisited in the adult brain. J Cell Biol 2005; 171: 641650.

21 Cameron HA, Dayer AG. New interneurons in the adultneocortex: small, sparse, but significant? Biol Psychiatry 2008;63: 650655.

22 Tamura Y, Kataoka Y, Cui Y, Takamori Y, Watanabe Y, Yamada H.Multi-directional differentiation of doublecortin- and NG2-immu-nopositive progenitor cells in the adult rat neocortex in vivo.Eur J Neurosci 2007; 25: 34893498.

23 Betarbet R, Zigova T, Bakay RA, Luskin MB. Dopaminergic andGABAergic interneurons of the olfactory bulb are derived fromthe neonatal subventricular zone. Int J Dev Neurosci 1996; 14:921930.

24 Belachew S, Chittajallu R, Aguirre AA, Yuan X, Kirby M,Anderson S et al. Postnatal NG2 proteoglycan-expressing pro-genitor cells are intrinsically multipotent and generate functionalneurons. J Cell Biol 2003; 161: 169186.

25 Dayer AG, Cleaver KM, Abouantoun T, Cameron HA. NewGABAergic interneurons in the adult neocortex and striatumare generated from different precursors. J Cell Biol 2005; 168:415427.


351


26 Seagrove LC, Suzuki R, Dickenson AH. Electrophysiologicalcharacterisations of rat lamina I dorsal horn neurones and theinvolvement of excitatory amino acid receptors. Pain 2004; 108: 7687.

27 Nishiyama A, Komitova M, Suzuki R, Zhu X. Polydendrocytes(NG2 cells): multifunctional cells with lineage plasticity. Nat RevNeurosci 2009; 10: 922.

28 Kempermann G, Gast D, Kronenberg G, Yamaguchi M, Gage FH.Early determination and long-term persistence of adult-generatednew neurons in the hippocampus of mice. Development 2003; 130:391399.

29 Lledo PM, Lazarini F. Neuronal replacement in microcircuits ofthe adult olfactory system. C R Biol 2007; 330: 510520.

30 Alonso M, Viollet C, Gabellec MM, Meas-Yedid V, Olivo-Marin JC,Lledo PM. Olfactory discrimination learning increases the survival

of adult-born neurons in the olfactory bulb. J Neurosci 2006; 26:1050810513.

31 Rochefort C, Gheusi G, Vincent JD, Lledo PM. Enriched odorexposure increases the number of newborn neurons in the adultolfactory bulb and improves odor memory. J Neurosci 2002; 22:26792689.

32 Mak GK, Enwere EK, Gregg C, Pakarainen T, Poutanen M,Huhtaniemi I et al. Male pheromone-stimulated neurogenesis inthe adult female brain: possible role in mating behavior. NatNeurosci 2007; 10: 10031011.

33 Hofstetter CP, Holmstrom NA, Lilja JA, Schweinhardt P, Hao J,Spenger C et al. Allodynia limits the usefulness of intraspinalneural stem cell grafts; directed differentiation improves outcome.Nat Neurosci 2005; 8: 346353.


352


Touch gives new life: mechanosensation modulates spinal cord adult neurogenesisIntroductionMaterials and methodsAnimalsSensory-enriched environment (SEE)BrdU administrationImmunohistochemistryQuantificationStatistical analysis

ResultsSingle exposure to novel tactile stimuli induces immediate cell proliferation, which depends on the diversity and intensity of the sensory stimuli

Figure 1 Single exposure to mechanosensory stimulation induces proliferation in the dorsal horn of the adult spinal cord.Repeated exposure to the same stimuli inhibits cell proliferation and induces neuronal differentiation of newly formed cells

Figure 2 Repeated exposures to sensory stimulation induce neuronal differentiation.The newly formed cells are mostly localized in the gray matter of the lumbar segments of the dorsal hornMost of the newly formed cells differentiate into GABAergic immature neurons

Figure 3 The newly formed cells are mostly localized in the gray matter of the dorsal horn of the lumbar segments.DiscussionFigure 4 The newly formed cells differentiate into gamma-aminobutyric acid (GABA)ergic neurons.Figure 5 Schematic representation of spinal cord adult neurogenesis in the sensory-enriched environment (SEE) model.Conflict of interestAcknowledgmentsReferences