Distinct Local and Global Functions of Aβ Low-Threshold Mechanoreceptors in Mechanical Pain Transmission

The roles of Aβ low-threshold mechanoreceptors (LTMRs) in transmitting mechanical hyperalgesia and in alleviating chronic pain have been of great interest but remain contentious. Here we utilized intersectional genetic tools, optogenetics, and high-speed imaging to specifically examine functions of SplitCre labeled Aβ-LTMRs in this regard. Genetic ablation of SplitCre-Aβ-LTMRs increased mechanical pain but not thermosensation in both acute and chronic inflammatory pain conditions, indicating their modality-specific role in gating mechanical pain transmission. Local optogenetic activation of SplitCre-Aβ-LTMRs triggered nociception after tissue inflammation, whereas their broad activation at the dorsal column still alleviated mechanical hypersensitivity of chronic inflammation. Taking all data into consideration, we propose a new model, in which Aβ-LTMRs play distinctive local and global roles in transmitting and alleviating mechanical hyperalgesia of chronic pain, respectively. Our model suggests a new strategy of global activation plus local inhibition of Aβ-LTMRs for treating mechanical hyperalgesia.


Introduction
Chronic pain is a devastating disorder of the nervous system, affecting more than 30% people worldwide (Cohen et al., 2021). A prominent symptom of chronic pain is mechanical hyperalgesia (Jensen and Finnerup, 2014; Price and Gold, 2018), or increased pain triggered by mechanical stimuli. At present, effective treatment for mechanical hyperalgesia is limited, re ecting a knowledge gap in its underlying mechanisms.
Aβ-LTMRs are large-diameter, highly-myelinated, and fast-conducting primary somatosensory neurons, which normally mediate tactile, discriminative touch, and vibration sensation. Aβ-LTMRs are further divided into rapidly adapting (RA) and slowly adapting (SA) types based on their ring patterns in response to a sustained mechanical stimulus. At baseline conditions, co-activation of touch-and painsensing neurons inhibits the ow of nociceptive information (Arcourt et al., 2017), as proposed in the "gate control theory of pain" (Melzack and Wall, 1965).
Functions of Aβ low-threshold mechanoreceptors (LTMRs) in chronic pain have been of great interest but are controversial. On one hand, several studies suggested that Aβ-LTMRs may mediate mechanical hyperalgesia of chronic pain. Spinal nerve ligation elicited increased irregular rings of rat RA Aβ-LTMRs (Liu et al., 2000;Na et al., 1993). When activities of Aβ-LTMRs were masked using nerve block, mechanical hyperalgesia was abolished in patients with nerve injury (Campbell et al., 1988). Spinal nerve ligation-induced mechanical allodynia in rats was abolished after destroying the dorsal column, where ascending axons of Aβ-LTMRs project through (Sun et al., 2001). Moreover, pharmacologically blocking of Aβ ber activity relieved neuropathic pain-induced mechanical allodynia in mice (Xu et al., 2015). On the other hand, some studies suggested that Aβ-LTMRs inhibit pain transmission even in chronic conditions. Rubbing or massaging of a painful area of the skin, which presumably activates Aβ-LTMRs, attenuated pain in humans (Mancini et al., 2014). Optogenetic co-activation of nociceptors and MafA + Aβ-LTMRs, as mentioned above, reduced acute pain behaviors (Arcourt et al., 2017). The spinal cord stimulator (SCS) and transcutaneous electrical nerve stimulation (TENS), which was developed based on the "gate control theory of pain" (Melzack and Wall, 1965) to target ascending axons of Aβ-LTMRs in the dorsal column (Shealy et al., 1967), are effective for treating various chronic pain conditions, including neuropathic pain (Gibson et al., 2019). Additional real-life therapy procedures, including massage therapy (Crawford and Caterina, 2020), transcutaneous electrical stimulation (Vance et al., 2012), and electroacupuncture (Duan-Mu et al., 2021), presumably involve the activation of Aβ-LTMRs for their bene cial effects.
When comparing different studies and their results, two main factors, which likely contribute to the diverse functions of Aβ-LTMRs in transmitting mechanical hyperalgesia or in its alleviation, were identi ed. The rst consideration is the speci city of manipulating Aβ-LTMRs and the functional readouts. Though pharmacological and mechanical/electrical stimuli can inhibit or activate Aβ-LTMRs, other types of nerve bers are likely simultaneously affected as well. For genetic and optogenetic manipulations, few available mouse genetic tools are selective or show high preference for Aβ-LTMRs (Dhandapani et al., 2018;Xu et al., 2015). Moreover, it has been challenging to differentiate touch and nociception associated re ex behavioral responses in animal studies. The second consideration is whether Aβ-LTMRs are manipulated locally at the area affected by chronic pain or in a broad manner including the unaffected Aβ-LTMRs, which may generate different sensory and behavioral outcomes.
To provide novel insight into this question, we utilized tools with improved speci city, including intersectional mouse genetics, opto-tagged electrophysiological recordings, and optogenetic activation of Aβ-LTMRs, and we performed a battery of behavioral tests to probe mouse mechanical and thermal sensations when Aβ-LTMRs were manipulated locally or globally. We also took advantage of high-speed imaging method that our lab previously established (Abdus-Saboor et al., 2019) to differentiate touchrelated non-nociceptive behavior from nocifensive behaviors. Here we found that mice with global ablation of Split Cre -Aβ-LTMRs showed increased mechanical hyperalgesia in the chronic in ammatory pain condition, suggesting that Aβ-LTMRs function to inhibit mechanical hyperalgesia in chronic pain. Thermosensation was not affected by Aβ-LTMRs manipulation, indicating that these afferents functioned in a modality-speci c manner. In reverse experiments, local optical activation of Split Cre -Aβ-LTMRs at the hind paw triggered nociception in chronic in ammatory and neuropathic pain models. This suggests that activities of locally affected Aβ-LTMRs lead to nociception, likely contributing to mechanical hyperalgesia of chronic pain. Finally, global activation of Aβ-LTMRs at the dorsal column alleviated mechanical hyperalgesia in a chronic in ammatory pain model. Together, our results establish a new model that Aβ-LTMRs play globally inhibitory but locally promoting roles for mechanical hyperalgesia. Our model suggests a new strategy, global activation plus local inhibition of Aβ-LTMRs, for treating mechanical hyperalgesia of chronic pain.
Immunostaining of lumbar spinal cord sections showed that GFP + central terminals did not innervate layers I/II, labeled by CGRP and IB4 staining (Fig. 1F), but were enriched in deeper dorsal horn laminae (III-V), indicated by VGLUT1 + staining (Fig. 1G) Immunostaining of plantar skin sections showed that GFP + peripheral axons were NF200 + (Fig. 1H), and they innervate a large percentage of S100 + Meissner's corpuscles (a type of RA Aβ-LTMRs, Fig. 1I). Few  Figures S1K-N). Together, the histological data suggest that this intersectional genetic strategy induce ReaChR expression preferentially in Aβ-LTMRs. In addition, RA Aβ-LTMRs are preferentially labeled in the glabrous skin, but RA and some SA Aβ-LTMRs are labeled in the hairy skin of these mice.
To con rm electrophysiological properties of ReaChR-labeled sensory afferents, we conducted optotagged electrophysiological recordings. We rst performed patch-clamp recordings at the Node of Ranvier of ReaChR/YFP + ber in isolated saphenous nerves (Fig. 1K). Action potentials were evoked by electrical stimulation at the distal site of the nerves (Fig. 1L), and conduction velocities were determined. The conduction velocity of ReaChR/YFP + bers was ~ 20 m/s (n = 17), in the range of Aβ bers (Fig. 1M).
Next, we used an ex-vivo skin-nerve preparation with both hind paw glabrous skin and paw hairy skin and performed single-unit recordings to determine mechanical threshold and ring patterns of ReaChR/YFP + bers. ReaChR/YFP + units were rst identi ed by their responses to orange (605 nm) LED light stimuli.
Mechanical stimuli by an indentator were then applied to the same receptive eld as light stimuli (Fig. 1N-O). A total of 7 light-sensitive saphenous nerve recordings and 9 light-sensitive tibia nerve recordings were recorded from the hairy and the glabrous skin, respectively. 4 units were RA (Fig. 1P), and 3 units were SA (Fig. 1Q), in the hairy skin, while 8 units were RA, and 1 unit was SA (Fig. 1R) in glabrous skin.
Mechanical stimulation-response curves of ReaChR + units showed typical RA and SA ring patterns (Fig. 1S). The mechanical thresholds of light sensitive units in the hairy and glabrous skin were similar, 1.4 ± 0.5 mN (n = 11) from RA units and 3.9 ± 1.0 (n = 4) from SA units (Fig. 1T), which were within the threshold of LTMRs. Taken together, the electrophysiological data con rm that this intersectional genetic strategy speci cally label cutaneous Aβ-LTMRs in both glabrous and hairy skin and that most of them are RA Aβ-LTMRs in the hind paw glabrous skin.
Ablation of Aβ-LTMRs by DTA treatment of Split Cre -Aβ TauDTR mice Next, we crossed Split Cre ;Advil FlpO mice to a Cre-dependent double-reporter (Tau ds-DTRf/f ;Ai9 tdTomatof/f ) mouse line (Duan et al., 2014), containing Tau ds-DTR (the human diphtheria toxin receptor was driven by a pan neuronal Tau promoter) and Ai9 tdTomato . The quadruple progeny was regarded as Split Cre -Aβ TauDTR mouse ( Fig. 2A). Similar to ReaChR2, the expression of tdTomato highly overlapped with large diameter neuronal marker NF200 (86.7% of tdTomato + neurons), to a lesser extent with PV (12.13%), but rarely overlapped with nociceptive markers CGRP (4.02%) and IB4 (3.03%) ( Figure S2A To ablate Split Cre -Aβ-LTMRs, six-week-old Split Cre -Aβ TauDTR mice were intraperitoneally injected with either DTA or water (vehicle control) daily for 1 week. Two weeks after the last injection, mice were sacri ced to con rm the ablation e ciency. The numbers of tdTomato + DRG neurons in whole-mount lumbar DRGs (Figs. 2B & C) were quanti ed, and a 45.3% decrease of tdTomato + DRG neurons was found after DTA treatment (p < 0.0001) (Fig. 2D). In addition, the mean uorescence intensity of tdTomato + central terminals in the deeper dorsal horn of the lumbar spinal cord of DTA-treated mice decreased 73% (p < 0.0001) (Figs. 2E-G). Moreover, a decrease of 81% in tdTomato + Meissiner's corpuscles-like structures in dermal papilla per foot-pad section was observed (Figs. 2H-J) (p < 0.0001). The total number of tdTomato + bers innervating the dermal papilla also showed signi cant reduction following DTA treatment (vehicle = 7.57 ± 0.76 nerves/dermal papilla section, DTA = 2.38 ± 0.68 nerves/dermal papilla section, p = 0.0022) ( Figure S2O). It is notable that the ablation phenotype of peripheral and central axons is much severer than that of the cell body, suggesting peripheral and central axon retracting in the remaining Split Cre -Aβ-LTMRs. Since afferents could not function without normal peripheral or central connectivity, we considered that ~ 80% of Split Cre -Aβ-LTMRs were functionally disrupted using this genetic and pharmacological strategy.
Ablation of Split Cre -Aβ-LTMRs reduced gentle touch sensation but increased mechanical nociception in the glabrous skin To investigate the requirement of Aβ-LTMRs in acute and chronic pain sensations, we conducted a battery of mouse behavior assays with control and ablated mice. We rst examined general locomotor behavior of mice with the open eld test ( Figure S3A). Ablation of Split Cre -Aβ-LTMRs did not signi cantly alter the time spent in peripheral (vehicle = 1098 ± 16.7 s vs DTA = 1075 ± 15.7 s, p = 0.38) and central zones (vehicle = 103 ± 16.8 s vs DTA = 125.3 ± 15.8 s, p = 0.328). There was also no signi cant difference between the two groups in the total distance travelled (vehicle = 62.32 ± 3.1 m vs DTA = 64.05 ± 4.0 m, p = 0.72). These results suggest that ablation of Split Cre -Aβ-LTMRs do not signi cantly change mouse general locomotor functions or generate obvious anxiety-associated behaviors. Next, we tested mechanosensitivity of mouse hind paws of control and ablated mice (Fig. 3A). Ablation of Split Cre -Aβ-LTMRs did not alter 50% paw-withdrawal mechanical threshold (PWT, static mechanosensitivity, vehicle = 0.99 ± 0.005 g, DTA = 0.98 ± 0.007 g, p > 0.99, n = 9 mice) (Fig. 3B) measured by the von Frey hair (VFH) test. A likely reason for no change of 50% PWT in the ablated mice is that other types of mechanosensory afferents in the hind paw, such as SA Aβ-LTMRs and MrgprD + afferents, are largely spared by this genetic strategy. On the other hand, sensitivity to dynamic gentle touch, measured as the percentage paw withdrawal in response to by a dynamic cotton swab (Fig. 3C), was signi cantly attenuated in the ablated mice (vehicle = 91.11 ± 4.84% vs DTA = 51.11 ± 8.88%, p = 0.0014, n = 9 mice). The ablated mice also spent a signi cantly longer time attempting to remove a sticky tape attached to the paw, which generated small amounts of mechanical forces (vehicle = 87.6 ± 14.33 s vs DTA = 198.1 ± 32.74 s, p = 0.0063, n = 10 mice) ( Fig. 3D and supplementary video 1). Together, these results suggest that the ablated mice are less sensitive to dynamic gentle mechanical forces. Moreover, when mice were tested using a chamber with different oor textures, hook (rough) vs loop (smooth) ( Figure S3E), the texture preference was altered in the ablated mice. Compared to control mice, which showed no signi cant preference for the oor textures (Hook = 386.5 ± 23.38 vs Loop = 478.4 ± 56.54 seconds, p = 0.505), the ablated mice preferred to stay in the smoother loop surface compartment (Hook = 323.9 ± 36.95 vs Loop = 559.9 ± 26.59 seconds, p = 0.001, n = 8 (vehicle) and 7 (DTA) mice) (Fig. 3E). In short, the behavioral de cits in dynamic gentle mechanical forces and tactile discrimination of DTA-treated mice supported signi cant functional disruption of Aβ-LTMRs in these mice.
Next, we tested mechanical nociception at the paw. In response to an alligator clip at the plantar skin, licking episodes, which re ect mouse nociception, signi cantly increased in the ablated mice (vehicle = 4.6 ± 1.01 episodes/1 min. vs DTA = 13.8 ± 1.98 episodes/1 min., p = 0.0002, n = 10 mice in each group, . This result suggests that Aβ-LTMRs function to inhibit mechanical pain sensation and that disruption of Aβ-LTMR functions results in mechanical hyperalgesia. In contrast, no difference in thermal nociception, tested by static hot plate ( Figure S3B), dynamic hot plate ( Figure  S3C), and dry ice test ( Figure S3D) (see also supplementary table S1), was found between the two groups, suggesting that Aβ-LTMRs inhibits nociception in a modality-speci c manner.
Moreover, we tested mechanical and heat hyperalgesia in a CFA-induced in ammatory pain model. In contrast to the baseline condition, the ablated mice showed signi cant reductions in 50% PWT (Fig. 3G), indicating an increased mechanical nociception. The PWT signi cantly decreased in DTA-treated mice at 2h (vehicle = 0.54 ± 0.12 g vs DTA = 0.05 ± 0.14 g, p = 0.0149), 14th day (vehicle = 0.18 ± 0.06 g vs DTA = 0.037 ± 0.007 g, p = 0.0427), 21st day (vehicle = 0.38 ± 0.12 g vs DTA = 0.049 ± 0.015 g, p = 0.0012), and 28th day (vehicle = 0.694 ± 0.086 g vs DTA = 0.052 ± 0.018 g, p = 0.0012). The differences at days 1, 3, and 7 were not signi cant due to the oor effect (the PWTs of control mice were already close to 0). Interestingly, the ablated mice showed no signi cant change in thermal hyperalgesia (Fig. 3H), which further supported the modality-speci c function of Aβ-LTMRs. Together, these behavior results suggest that in the glabrous skin, disruption of Split Cre -Aβ-LTMRs reduces the dynamic gentle-touch sensitivity and speci cally dis-inhibits mechanical pain transmission at baseline and chronic in ammatory conditions.
Ablation of Split Cre -Aβ-LTMRs altered gentle touch sensation and mechanical nociception in the hairy skin The hairy skin contains Aβ-LTMRs as well as C-and Aδ-LTMRs ( Figure S4A) (Li et al., 2011). Thus, behavior outcomes from the hairy skin of ablated mice could be more complicated than the glabrous skin. We conducted a similar sticky tape test, which generated gentle mechanical stimuli at the de-haired back skin. The ablated mice displayed signi cantly increased scratching behavior in response to the tape (vehicle = 0.181 ± 0.181 scratch bouts / 5 min vs DTA = 11 ± 3.964 scratch bouts / 5 min, p = 0.0027, n = 11 and 10 mice in vehicle and DTA groups, respectively) ( Figure S4B), but no difference in back-attending episodes between the vehicle and ablated groups (vehicle = 17.64 ± 3.8 attending episodes / 5 min vs DTA = 19.2 ± 3.552 attending episodes / 5 min, p = 0.743) was observed ( Figure S4C). These results suggest that the ablated mice show increased responses to gentle mechanical forces at the hairy skin, which might be mediated by dis-inhibited C-and/or Aδ-LTMRs. In addition, in response to the application of an alligator clip at the nape of the neck, the ablated mice displayed signi cantly increased number of attending episodes (vehicle = 8.45 ± 1.60 vs DTA = 17.4 ± 2.16 attending episodes / 1 min, p = 0.0074, n = 11 and 10 mice in vehicle and DTA groups, respectively) ( Figure S4D and supplementary video 3), indicating an increased mechanical pain sensation. No difference in heat nociception between the groups was found, tested by tail-immersion assay at different temperatures (supplementary table S1). These data indicate that Split Cre -Aβ-LTMRs are required to "gate" both gentle touch and mechanical pain transmission in the hairy skin.
To differentiate nocifensive or non-nocifensive paw withdrawal re exes, we used a "pain score" system (paw guarding, paw shaking, jumping, and eye grimace) that we previously established with high-speed imaging (Abdus-Saboor et al., 2019). Laser at all intensities triggered paw withdrawal re ex (Fig. 4B), but paw withdraw re exes evoked by 5 mW blue laser did not show nocifensive features (pain score: 0 ± 0 (5 mW), 0.2 ± 0.08 (10 mW), and 0.74 ± 0.13 (20 mW)) ( To determine changes in ring properties of Split Cre -ReaChR + Aβ-LTMRs in chronic pain condition, we performed opto-tagged recording of these afferents in saline-or CFA-treated mice at post-injection Day 7. Mechanically-evoked RA impulses were recorded from light-sensitive (ReaChR/YFP tagged) afferent bers that innervate the glabrous skin of hind paws (Fig. 5A). We detected RA impulses of light-sensitive afferent bers in both saline-(n = 7) and CFA-injected (n = 5) mice. The number of impulses increased proportionally with higher mechanical force in CFA-treated groups, which was signi cant at 30 (p < 0.05) and 80 (p < 0.001) mN force (Fig. 5B). Similarly, we detected tissue indentation induced SA impulses in light-sensitive afferent bers of saline-(n = 1) and CFA-injected (n = 6) mice. Though the number of Split Cre -ReaChR + Aβ afferents displaying SA property in control mice (n = 1) is too low for a meaningful statistical comparison, an obviously higher number of Split Cre -ReaChR + Aβ afferents displaying SA property after CFA treatment (n = 6) was observed (Fig. 5C).
The percentages of opto-tagged afferents showing RA and SA responses were quanti ed. In the glabrous skin of the control mice, few light-sensitive afferents were SAs (n = 1), and most of them were RAs (n = 7), similar to the result of naïve animals (Fig. 1R). However, with CFA-induced in ammation, the percentage of RA afferents (n = 5) decreased while that of SA (n = 6) increased signi cantly (p = 0.05) (Fig. 5E).
Mechanical thresholds of RA afferents in control and CFA groups did not alter signi cantly (Fig. 5F).
Mechanical thresholds of SA afferents in both groups were also similar ( Fig. 5G). Overall, these data suggest that in chronic in ammation, the electrophysiological properties of many RA Aβ-LTMRs became SA-like, which could be one of the underlying mechanisms for triggering stronger local activity and inducing mechanical hyperalgesia in the CFA model.
To determine whether ReaChR + LTMRs increased ring was due to an alteration in peripheral end organ innervation, we performed foot pad section immunostaining for ReaChR + afferents, Merkel cells ( Figures  . This anatomical feature allows an Aβ-LTMR to "gate" nociceptors (inter-modal crosstalk) or other Aβ-LTMRs residing in 3 to 4 dermatomes away (inter-somatotopic crosstalk) (Fig. 7). Thus, even though locally affected Aβ-LTMRs are sensitized to trigger nociception in in ammation or nerve injury regions, the overall effect of all Aβ-LTMRs, most of which are unaffected, is still to inhibit mechanical pain transmission.
If our model is correct, then the global activation of Aβ-LTMRs, even in chronic pain, theoretically should still inhibit mechanical hyperalgesia. To test this idea, we implanted light-cannula above the T11 dorsal column of the spinal cord of Split Cre -Aβ ReaChR mice (Fig. 6A). Similar to the peripheral experiments, we rst tested blue laser light at four different intensities, 0.5 mW, 3 mW, 5 mW, and 10 mW, at the baseline condition. During light stimuli (15 minutes total, 30 second on (10 Hz) and 1 minute off, 10 cycles), all intensities triggered some spontaneous behaviors (supplementary table S3). These behaviors calmed down to the resting state when light was off. We then did VFH test to measure 50% PWT at different time points afterwards (Fig. 6A), using blue laser at the low (0.5 mW) or high (10 mW) intensities (Fig. 6B). We averaged PWT of both paws for quanti cation as the light cannula was implanted at the dorsal column and would affect both sides of the spinal cord. 0.5 mW blue laser spinal cord stimuli did not signi cantly alter the hind PWT (supplementary table S3 To further discern whether the activated DH neurons are excitatory or inhibitors, we performed RNAScope in situ hybridization of cFos, Vglut2 and Vgat (Figs. 6D and S6M-P). Around 80% cFos + neurons were VGAT+ (Fig. 6E). Together, our results suggested that dorsal column stimuli of Split Cre -ReaChR + Aβ-LTMRs would preferentially activate DH inhibitory interneurons in a broad range of spinal cord levels.

Discussion
In this study, we combined intersectional mouse genetics, opto-tagged electrophysiology recordings, and high-speed imaging behavioral assays to clarify roles of Aβ-LTMRs in transmitting and alleviating mechanical hyperalgesia. Our results revealed that both global ablation and local activation of Aβ-LTMRs promoted mechanical hyperalgesia, whereas their global activation alleviated it. Therefore, we propose a new model (Fig. 7), which integrates the inter-modal crosstalk between the nociceptive pathway and Aβ-LTMRs and the inter-somatotopic crosstalk among Aβ-LTMRs at different spinal segments, to explain the complicated phenotypes of Aβ-LTMRs in mechanical hyperalgesia. Our model suggests that the global activation plus local inhibition of Aβ-LTMRs would be an effective strategy for treating mechanical hyperalgesia.

Complicated function of Aβ-LTMRs in chronic pain
A half century after the introduction of Gate Control Theory (Melzack and Wall, 1965), the exact functions of Aβ mechanoreceptors in chronic pain are still not fully resolved. The primary function of these afferents is mediating discriminative touch and tactile/vibration sensation. However, these neurons also play important roles in "gating" or inhibiting other somatosensory pathways to generate the appropriate sensation. The fast conduction velocity of Aβ-LTMRs put them in a great position for this role. In response to a mechanical stimulus, signals of Aβ-LTMRs arrive at the dorsal spinal cord rst and "set up" the "gate" for other mechanosensory pathways with slower conduction velocities (Aδ-and C-LTMRs and high-threshold mechanoreceptors). Thus, activating Aβ-LTMRs has been a commonly sorted strategy for treating chronic pain (Caylor et  Here we utilized intersectional genetics to selectively ablate Aβ-LTMRs or optogenetically activate them either locally at the affected site (peripheral) or globally at the dorsal column (central) and examined behavioral outcomes in baseline and chronic pain conditions. Interestingly, activities of locally affected Aβ-LTMR S in chronic in ammation triggered nociception (Fig. 4), whereas global activities of Aβ-LTMR S , many of which were unaffected, could still inhibit transmission of mechanical pain in both acute and chronic pain conditions in a modality-speci c manner (Figs. 3 & 6). This clear distinction of "local" vs "global" functions of Aβ-LTMR S in mechanical pain transmission, as explained by our model (Fig. 7), will help to consolidate the existing data in this topic and provide a theoretic reference for the design of new treatment strategies.
Our ndings are somewhat different from two previous studies in this topic but using different genetic tools for ablation or optogenetic activation (Dhandapani et al., 2018, Chamessian et al., 2019. The rst study found that DTA ablation of Trkb + Aβ-LTMRs attenuates mechanical allodynia in a neuropathic pain (SNI) but not in ammatory pain (CFA) models (Dhandapani et al., 2018). A main difference between ours and this study is the genetic tools, which manipulated largely non-overlapping sensory afferents. We used Split-Cre allele that preferentially labeled Aβ-LTMRs in both glabrous and hairy skin, whereas the other study used TrkB-CreERT2 mice, which preferentially label Aδ-LTMRs (D-hair) in the hairy skin and a few Aβ-LTMRs in the glabrous skin. Here we found that a high percentage (~ 75%) of paw Meissner corpuscles were innervated by Split-Cre labeled afferents (Figs. 1, 2, S1, S2), ~ 80% these innervation was lost in the ablated mice (Fig. 2), and obvious de cits of DTA-ablated mice in dynamic gentle mechanical force and tactile sensation (Fig. 3). In contrast, the other study showed almost no Meissner corpuscle innervation and found no mechanosensory de cits from paws of ablated mice at the baseline condition.
Thus, global DTA ablation using TrkB-CreERT2 and Split-Cre lines would disrupt largely different populations of sensory afferents and thus generated different behavioral outcomes. The second study used a VGlut1-Cre mouse line, which should cover most Aβ-LTMRs, but failed to detect behavior changes when optogenetically activating VGlut1 Cre -Aβ-LTMRs in a neuropathic pain model (Chamessian et al., 2019). As the authors discussed in that paper, it is hard to differentiate between touch-like and pain-like pain re ex behaviors without a high-speed camera. This is what we used in this study, which improved resolution of behavior assays.
Targeting Aβ-LTMRs for treating mechanical hyperalgesia and chronic pain Multiple existing techniques, presumed to involve the manipulation of Aβ-LTMRs, are used in practice for treating chronic pain. The best known one is the SCS, which was developed based on the "gate control theory" and was designed to target Aβ-LTMR axons projecting through the dorsal column (Shealy et al., 1967). The SCS is effective for treating different chronic pain conditions, refractory chronic pain, and even those failed available pharmacological approaches (Gilbert et al., 2022;Sun et al., 2022). Since several types of axons, besides those of Aβ-LTMRs, project through the dorsal column, and can simultaneously be activated by SCS, whether stimulating Aβ-LTMR axons alone in the dorsal column is su cient for alleviating chronic pain has remained untested. Our study provides direct supportive evidence for the chronic pain-alleviating effect of activating Aβ-LTMRs at the dorsal column (Fig. 6)  Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request. Tissue was cleared in 100% BABB and mounted on a slide (with little BABB). Four drops of grease were put at four corners around the tissue and a cover glass was put over it with gentle pressure so that the coverslip sticks with the mounting tissue/grease.

Image Acquisition
Images were taken using a Leica SP5II confocal microscope ( uorescent). Image processing and gure generation were performed in Fiji-ImageJ and Adobe Illustrator.
Ex vivo skin-nerve preparation Split Cre ;Advil FlpO ;Rosa ReaChRf/+ mice of both male and female sexes were used, and most of them (31 mice) were aged 6 to 15 weeks and 4 were over 20 weeks. In one set of experiments, naïve animals were used, and in another set of experiments, animals were randomly assigned into saline and CFA groups. In saline group, each animal was injected with 10 µl saline into both hind paws. In CFA group, each animal was injected with 10 µl CFA (5 µg/10 µL) into both hind paws. 4-7 days after the injection of saline or CFA, animals were anesthetized with 5% iso urane and then sacri ced by decapitation. The saphenous nerves with their innervated hairy skin of the hind paws or the tibial nerves with their innervated glabrous skin of the hind paws were dissected out from the animals. The skin-nerve preparation was then placed in a Sylgard Silicone-coated bottom of a 100-mm recording chamber that contained the Krebs bath solution described below. The fat, muscle and connective tissues on the nerves and the skin were carefully removed with a pair of forceps. The skin was a xed to the bottom of the chamber by tissue pins with epidermis side facing down, and the nerve bundle was a xed by a tissue anchor in the same recording chamber. The recording chamber was then mounted on the stage of the Olympus BX51WI upright microscope. The skin-nerve preparation was superfused with a normal Krebs bath solution that contained (in mM): 117 NaCl, 3.5 KCl, 2.5 CaCl2, 1.2 MgCl2, 1.2 NaH2PO4, 25 NaHCO3, and 11 glucose (pH 7.3 and osmolarity 325 mOsm) and was saturated with 95% O2 and 5% CO2. The Krebs bath solution in the recording chamber was maintained at 24℃during experiments recordings. To facilitate the pressureclamped single-ber recordings, the cutting end of the nerve bundle was brie y exposed to an enzyme solution that contained 0.1% dispase II and 0.1% collagenase in Krebs solution for 30 to 60 s, and the enzyme was then washed off by the continuous perfusion of the normal Krebs solution.

Pressure-clamped single-ber recordings
The pressure-clamped single-ber recording was performed in a manner similar to our previous studies (Sonekatsu et al., 2020). In brief, recording electrodes were made by thin-walled borosilicate glass tubing without lament (inner diameter 1.12 mm, outer diameter 1.5 mm). They were fabricated using a P-97 Either Diphtheria Toxin (20 µg/kg, Sigma-Aldrich, USA) or the water (vehicle) was injected (intraperitoneal) into Split Cre ;Advil FlpO ;Tau ds−DTR f/+ ;Ai9 tdTomatof/+ mice at the age of 6-week old for 7 days. Two weeks after the last injection, some mice were sacri ced and perfused. Their tissues (spinal cord, DRGs, and skin) were collected for histological characterization to determine e ciency of DTA-induced ablation. The remaining mice were used for different behavioral tests from one month after the last DTA or vehicle injection.

CFA-induced in ammatory pain model
The CFA-induced in ammatory pain model was generated as previously described (Beattie et al., 2022). Brie y, a mouse with the desired genotype was anesthetized by iso urane, and the plantar surface of one paw was sterilized. 10 µL of CFA emulsion (Sigma, F5881) was slowly injected into the plantar skin surface (from the lower middle walking footpad at the ventral surface towards the plantar area) using a 0.5 mL insulin syringe. Injection site was hold using gentle thumb pressure to avoid leakage of CFA for ~ 10 s. The success of model was con rmed by measuring the paw-thickness by a digital electronic caliper (#30087-00, Fine Science Tools) (1 day post injection) and mechanical sensitivity using von Frey laments.

Medial Plantar Nerve Ligation (MPNL)-induced neuropathic pain model
The MPNL model was generated as previously published (Sant'Anna et al., 2016). Brie y, a mouse with the desired genotype was anesthetized using iso urane. After sterilizing the paw skin, the medial surface of the ankle of one leg was incised (0.5 cm) using #11 blade to expose the medial plantar nerve. One ligation was performed with a 4-0 catgut suture (Ethicon). The skin was sealed. After waking up, mice were returned into the home cage.

Behavioral Assays von Frey Hair Assay
To test the static touch sensitivity, mice were habituated for 2 days in the behavioral room. Under Plexiglas chambers (11.5 x 4.5 x 4 cm), they were kept over a perforated wire mesh platform (Ugo Basile, Italy, Part #37450-045) which had 5 x 5 mm gaps, for 1 h before starting the experiment. Paw withdrawal threshold was assessed using the up-down method (Chaplan et al., 1994) as described previously (Cui et al., 2022). Brie y, 8 calibrated and logarithmically spaced von Frey mono laments (bending forces: 0.008, 0.02, 0.04, 0.07, 0.16, 0.4, 0.6 and 1 g; Stoelting, Wood Dale, IL) were used. These were applied tangentially to the plantar surface for ~ 3-4 s with enough force to cause a slight buckling of the lament. First, the middle lament (0.16 g) was applied to the hind paw. If the mouse showed a withdrawal, an incrementally lower lament was applied. If there was no response, an incrementally higher lament was applied. Trials were separated by at least 2-3 minutes to avoid sensitization and learning. A positive response was characterized as a rapid withdrawal of the paw away from the stimulus ber within 3-4 s (Sudden withdrawal, icking or licking of the tested paw). Testing was continued until four laments were applied (heavier or lighter, depending on the exact lament size to which the last response occurred) after the rst one that produced a withdrawal. The nal value of 50% withdrawal threshold was calculated by using pseudo-log calculator in Excel which uses the following equation (Chaplan et al., 1994): 50% withdrawal threshold (g) = (10[Xf + kd])/10,000 Where Xf = value (in log units) of the nal von Frey lament used; k = tabular value for the pattern of positive/negative value and d = mean difference (in log units) between stimuli.
The observer was blind to the genotype and the drug-treatment while performing the behavioral experiments.

Cotton Swab Assay
After performing von Frey testing, on the consequent day, gentle dynamic touch-evoked sensitivity was measured as described previously (Ranade et al., 2014). Mice were placed under transparent Plexiglas chambers on an elevated wire-mesh platform oor. The oor consists of mesh-like grids that are accessible from below due to small gaps of ~ 5×5 mm. Mice were habituated 1 h daily for 2 days on this setup in the behavior room and allowed to acclimate for 1 hour before testing. A cotton swab from a cotton applicator (Puritan 25-806 1WC) was manually pulled so that it was "puffed out" to ~ 3X the original size. When the mouse was at rest, a constant sweeping motion from heal towards the toes was used underneath the mouse paw. Mice were recorded using high-speed imaging camera (500 & 1000 fps). Their paw withdrawal response and other spontaneous behaviors were analyzed. Five trials were performed with ~ 5 minutes interval between each sweep. Both hind paws were used randomly. The videos were saved in an external hard drive and later analyzed for the number of hind paw withdrawals out of 5 times as a percentage (%) response for each mouse and averaged.

Tape Removal Assay
This assay was performed over glabrous skin (plantar surface) of the mice slightly modi ed from the (Liu et al., 2018). Mouse was habituated in a transparent plexiglas rectangular chamber over an elevated perforated wire-mesh platform for 45 minutes. Mouse was removed from the enclosure, and a 9.5 mm diameter, red, circular adhesive Microtube Tough-Spots label (RPI 247106R) was attached to the plantar surface of the hind paw. The mouse was returned to the chamber immediately for video recording. The behavior of mice was recorded using a HD web camera (C922 Pro HD Stream Webcam). The time that each mouse took in removing the tape was measured. 5 minute was the cutoff time for the experiment.

Texture-Preference Assay
A chamber with two compartments separated by a smaller mid-compartment was designed. The two big chambers' oors were covered with loop (Soft side) or hook (Rough side) of Velcro tape, respectively.
Also, each chambers had distinctive visual cue patterns (white and black stripes vs white checkered duct tapes). Middle chamber separating these two chambers was completely black, and its oor was made to the level of the two other chambers using plane thick white cardboard. Experiment was done in dim light.
Daily for 3 days, each mouse was given restricted habituation of 20 minutes in each chamber. On 4th day, each mouse was allowed to move freely in all chambers for 20 minutes and videotaped from the upside.
These videos were analyzed using AnyMaze software for quantifying the total time spent in each chamber.

Plantar-pinch Assay
Mice were habituated in a transparent cylindrical glass chamber (Diameter 10 cm) for 30 minutes. An alligator clip (Amazon, "Generic Micro Steel Toothless Alligator Test Clips 5AMP"), which produces 340 g force, was applied to the ventral skin surface between the footpads. The animals were returned into the chamber, and their behaviors were video recorded using a high-speed imaging camera (80-200 fps) for 60 s. Behaviors, such as licking, wiping, shaking and scratching episodes, were quanti ed. Those animals with the clip somehow removed before 60 s cutoff were not considered in the study.

Hargreaves Plantar Assay
Mice were placed in opaque rectangular Plexiglas chambers over the glass surface of the Hargreaves apparatus (UCSD Instruments, San Diego, CA, USA), and allowed to acclimate for ~ 45 minutes. Mice urine or feces, if any, were removed and the hind paw and the glass surface were cleaned and dried using wipes immediately. A thermostat was used to assess the constant temperature of the glass plate (∼22°C ± 1). The heat stimulus was applied from a bulb beneath the glass, to the middle of the plantar surface of the hind paw. Either a brisk paw withdrawal ( ick) or licking in response to the thermal stimuli was considered as a positive response. Cutoff time was 20 s. Three trials with a gap of 10 minutes were taken and averaged as the paw withdrawal latency (in second).

Open Field Activity Assay
Mice were habituated for two days in the behavior room in their cage for 45 minutes in the dim light condition. On the third day, mouse was directly kept in the center of the arena and the behavior was recorded using a webcam from the upside. The experimenter was not present in the room while recording the behavior to avoid observer-related behavioral changes. The total recording time was 20 minutes.
Before starting experiment in the next mouse, the surface of arena was cleaned using 40% alcohol and allowed to dry for 10 minutes. The videos were analyzed using AnyMaze software to measure the time spent in the Central and peripheral zone as well as the overall distance traveled.

Static Hotplate Assay
For the static hot plate test, mice were placed on top of a hot plate (IITC, Life Science), preset to 50 ± 0.5°C covered by a transparent plexiglas chamber. The behavior of the mouse was recorded using a highspeed imaging camera (200 fps). Videos were analyzed, and the latency to lick or ick the hind paw or jumping was quanti ed. Three trials at intervals of at least 15 min were taken and the average score for each mouse was obtained. To avoid tissue injury of the mice, a cutoff of 30 s was set.

Incremental Hot Plate Assay
The incremental hot plate test was carried out on the same apparatus using different settings, with at least 24 h rest from the static hotplate assay. The initial temperature was set to 28°C and increased by 6°C / min towards a nal temperature of 55°C (Alshahrani et al., 2012). The temperature when the rst hind paw lick occurred was recorded. If no hind paw lick was observed, the test was terminated at 55°C.
Three trials at intervals of at-least 15 min were taken and the average temperature for each mouse was obtained.

Dry-ice Assay
Mice were placed in Plexiglas chambers on a 2.5 mm thick elevated glass plate and allowed to habituate for at least 45 min. When the mouse was completely at rest, a dry ice pellet (1 cm diameter) was applied to the lower glass surface underneath the hind paw of the animal, and the withdrawal ( ick, lick, or both) latency was measured using a stopwatch. Each hind paw was tested randomly for three times with at least 15 min interval in between two consecutive trials. To avoid frost-induced tissue injury, the cutoff latency was set to 10 s.

Tape Response Assay
This assay was performed over hairy (back) skin with slight modi cations (Dhandapani et al., 2018).
Mice were habituated in a transparent Plexiglas rectangular chamber over an elevated perforated wiremesh platform for 45 minutes. They were removed, and a small piece of laboratory tape (~ 3 cm × 1 cm) was placed gently on the bottom center of the mouse's back. Mice were video recorded for a duration of 5 min by a web or high-speed camera? The total number of scratching bouts, wipes and other behaviors in response to the tape were quanti ed.

Nape-pinch Assay
Mice were anesthetized using iso urane, and the nape of the neck (~ 2 cm 2 area) was shaved using an electric trimmer. After 2-3 days, the mice were habituated for 15 minutes in a glass chamber of ~ 10 cm diameter. An alligator clip producing 340 g force was applied to the shaved nape skin fold. The animal was placed back into the chamber and video recorded using high-speed camera (80-200 fps) for 60 s.
The videos were analyzed to quantify the scratch bouts, headshakes, bilateral wipes, and attending duration/episodes.

Tail-immersion Assay
A mouse to be tested was restrained in a plastic 50 mL screw capped conical centrifuge tubes with several holes in the tube wall so that mouse can breathe normally. A 0.5 cm 2 opening was cut in the cap to allow the tail access to the water bath. Mice were habituated in the tube for 30 minutes in 2 days. On 3rd day, by holding the tube horizontal, the distal part of the tail (~ 4 cm) was submerged in the temperature-controlled water bath. Sudden tail-ick was used as a response sign. 3 33210A)), was shined upon the plantar surface below the wire-mesh space (the distance between paw and the laser cord-outlet tip is ~ 2 mm). Laser intensity was measured using a Digital Optical Power Meter with a 9.5 mm aperture (ThorLabs, PM100A). Each mouse was tested for 5 trials with an inter-trial interval of 5 min. The percentage of trials showing paw withdrawal response, such as the paw utter, ick, lick or the brisk withdrawal of the paw, was quanti ed. The paw-withdrawal latency (s) in response to the plantar optogenetic stimulation was also quanti ed and averaged from 5 trials.
For optical stimuli of the tail, when the mouse was not grooming and its tail was at complete rest, the mid of the tail was stimulated by shining a blue laser from ~ 2 mm distance below the wire-mesh. Each mouse was tested for 5 trials with an inter-trial interval of 5 min. The latency to ick the tail and the percentage of trails showing withdrawal response were quanti ed and averaged from 5 trials.
For optical stimuli of the nape and back, mice were shaved at these regions using an electric trimmer and then habituated for 3 days under transparent plexiglas chambers. Nape or back of these mice were stimulated from upside of the chamber by a blue laser when they were still. Any sudden bodymovement/shaking or avoidance behaviors of the mice was considered as a positive response. Each mouse was tested for 5 trials with an inter-trial interval of 5 min. The latency to ick the tail and the percentage of trails showing withdrawal response were quanti ed and averaged from 5 trials.
The pain score was quanti ed as previously described (Abdus-Saboor et al., 2019). Brie y, four individual behavior features: orbital tightening, hind paw shake, hind paw guarding, and jumping were considered as "pain" related behaviors, and 1 score was given for each behavior shown (0 minimum and 4 maximum score) for a testing trial. The nal pain score was averaged from all ve trials.

Conditioned Place Preference Assay
A chamber with two compartments separated by a smaller central compartment was designed. The two big chambers' walls had distinctive visual cue patterns (white and black stripes vs white checkered duct tapes). The apparatus was placed over the elevated wire-mesh platform. Each mouse was allowed 20 minutes to explore the three-chambered apparatus with no optogenetic stimulation for 3 days. The mouse was excluded from the study if it showed more than 60% preference to a particular chamber at this stage. From days 4-6, whenever mouse entered a speci c chamber (White strips pattern on the wall), a blue laser (5 mW, 10 Hz) was shined on the plantar surface of the treated hind paw (either saline-or CFA-treated) from below the wire mesh platform till the mouse leaves this chamber and reach to the other chamber. To avoid any prospective optic cable and hand movement-related non-speci c response, the optic cable (held in the hand of the experimenter) was also moved under the platform with the movement of the mouse but the light was only shined when the mouse entered the stimulation chamber. This was repeated for 20 minutes from day 4-6. On day 7, we tested these mice for place preference where each mouse was allowed 20 minutes to freely move about the three-chambered apparatus without blue laser stimulation. The activity was recorded using a webcam from the upside and scored later using AnyMaze software. Percent post-stimulation change was calculated as percent time in blue light chamber after training minus percent time in blue light chamber before training.

Spinal light cannula implantation
A light cannula (Ceramic, 1.25 mm diameter, 200 µm optic core, and length ~ 0.25 mm, Thor Labs) was implanted in T11 region of the mouse spinal cord as described previously (Christensen et al., 2016).
Brie y, under iso urane anesthesia, the hairs of the back region of a mouse were removed, and the mouse was placed in a stereotax (Model 940, KOPF instruments, Tujunga, CA, USA). After sterilization of the surgical site, a local anesthetic was administered prior to the incision. For implantation at ~ T11 region, a 1 cm long incision was made starting caudal of the peak of the dorsal hump, extending approximately 0.5 cm rostral and 0.5 cm caudal from the initial incision site. White tendons were cut from both sides of spinal column and the vertebral column was exposed by clearing tissue from the transverse processes without damaging spinal cord and nerves. T11 vertebra was xed from both sides using the spinal adapters (#51690, Stoelting Co., Wood Dale, IL). Surface connective tissue was removed from the T11 vertebra and the adjacent rostral and caudal vertebrae by gentle scrubbing using the tip of sterilized cotton swabs. With a ne tipped burr drill (0.5 mm in diameter), the bone and dura mater were punctured.
A light cannula was prepared to implant. The hole surface was dried using the cotton swabs. Using the other side tip of the cotton applicator, a very little amount of glue (Krazy®Glue, Part #963257) was applied around burr-hole before lowering the cannula into place. After lowering the cannula through the hole, and once the cannula attaches to the spinal cord dorsal surface rmly, dental cement was applied around the outside of the vertebra to stabilize the cannula. Spinal xation bars were removed after drying of the cement. The skin was sutured, followed by administration of systematic analgesics. Mouse was put on a warm pad for recovery. After surgery, each mouse was housed individually to avoid accidental removal of the light-cannula by a cage mate. Implanted mice were used for behavior assays 2 weeks after implantation.
Central/Spinal optogenetics Two weeks after the spinal light-cannula implantation into Split Cre -Aβ ReaChR mice, the light cannula was attached to a rotating ber cannula connected to the blue laser. These mice were habituated on an elevated perforated wire-mesh platform daily for 1 h in a transparent circular Plexiglas chamber (7 cm diameter and 30 cm height) for 3 days. On 4th day, in the similar settings, after 45 min habituation, mice were stimulated by a blue laser of lower (0.5 mW) or higher intensity (10 mW Quanti cation, statistical analysis, software, and data presentation Electrophysiological data were analyzed using Clamp t 10 (Molecular Devices, Sunnyvale, CA, USA).
Data were collected from 17 male and 14 female animals and were aggregated for data analysis. To con rm that impulses evoked by LED light and mechanical indenter are generated from the same ber, the amplitudes and shapes of the impulses evoked by both LED light and mechanical indenter at the same receptive eld were compared, and the data were included only when mechanically evoked impulses matched the light-evoked impulses. Conduction velocity was calculated by the distance between stimulation site and recording site divided by the time latency for eliciting an impulse following electrical stimulation.
All data shown in column and line graphs represent Mean ± SEM, unless otherwise mentioned. Sample sizes and statistical methods are de ned in respective gure legends. Counting of cell-number was performed in FIJI software (NIH). All data were analyzed, and graphs were originally created using GraphPad Prism Version 9.4.1 and further modi ed for publication purpose in Adobe Illustrator Version 27. Some cartoons in the gures were made using BioRender.com. Figure 1 Histological and electrophysiological characterizations of Split Cre -ReaChR mice con rmed that Split Cre preferentially recombined in Aβ-LTMRs.