The butterfly effect and its electrical mechanism: Non-linear non-reciprocate sawtooth subthreshold oscillations deactivate nociceptors

Abstract

Across mammalian evolution, chronic pain has no adaptive value, and in the wild, there’s no evidence of its existence. Since rodents are often used to model chronic pain in humans, the question of how the peripheral somatosensory system of these animals responds to injury becomes critical to our overall translatability efforts. Over a decade of intensive work on this question has led to the discovery of the primordial systemic process that protects the mammalian peripheral somatosensory system against uncontrolled hyperexcitability, as well as its underlying electrical mechanism. Named the “butterfly effect,” this two-stage process enables the appropriate animal behavioral response to injury (first stage) while evading pathology by deactivating hyperactive nociceptive neurons (second stage). This deactivation process involves the generation of subthreshold membrane sawtooth oscillations, which, rather than producing ectopic discharges, lead the cells to a quiescent state. The complex nature of this phenomenon challenges any simplistic approach to modeling and translating animal pain physiology directly into human pain pathology.

Introduction

For over a decade, the complex set of cellular events that follows the occurrence of a peripheral insult has been carefully studied in the primary somatosensory system of rodents. These studies have encompassed the full spectrum of injuries, ranging from potential (acute) ¹ to degenerative (chronic) ² animal models, affecting the function and sensory trajectories of mechanosensory neurons broadly simplified into two primary modalities (tactile: low-threshold mechanoreceptors (LT) and nociceptive: high-threshold mechanoreceptors (HT) neurons innervating dorsal root (DRG) and trigeminal (TG) ganglia dermatomes). From this intensive work, one constant phenomenon, present in every study performed to date, has emerged, known internally to our laboratory as the “butterfly effect.” Its picturesque name reflects two facts: 1. the shape of the sensibility trajectories of the somatosensory afferents’ plastic response to injury, similar to closing/opening of the wings of the insect (Figure 1), and 2. the origin of the seminal ideas on mechanosensory information encoding in insects (noctuid moths’ tympanic organ)^3,4 that have unexpected fundamental similarities with the process of mammalian multimodal peripheral plasticity.

Figure 1.

Butterfly effect, mechanosensory components’ response to different types of injuries in different dermatomes of two Rodentia species (rats and mice). (a and b) Schematic of the nociceptive (high threshold mechanoreceptors (HT) A- and C-nociceptors combined, violet, up-triangle) and non-nociceptive (low threshold mechanoreceptors (LT), black, circle) response of male Sprague-Dawley (SD) rats to nerve injury (L5 partial spinal nerve ligation (pSNL)). Notice the afferents’ overall mechanical threshold (MT) trajectories matching the shape of a butterfly as they pass from week 0 (W0) to week 8 (W8) and the concurrent changes in the active afferents’ distributions (percentage). B: Baseline, D: Deactivation, R: Resolution. Values are medians extracted from Boada et al. ¹³ (c and d) Effect of different types of injury on the MT and distribution of mechanosensory afferents of male C57BL/6J (C57) and B6.Cg Foxn1nu/J (Fox1) mice. Notice to progressive MT afferents’ proximity and changes in their distribution depending on injury magnitude and dermatome (L4 DRG vs TG). Values are medians presented with the MT cellular range (colored boxes) extracted from: Paw incision, ⁶ UVB irradiation, ⁷ PNI, ² IMP. ¹⁰ PNI: MOC2-perineural invasion model on the sciatic nerve. IMP: Bioluminescent human breast cancer cell line (MDA-MB-231^LUC+) naturally implanted on the animal’s mandible.

Consistent across various peripheral stimulation protocols, pain models, dermatomes, tissues, and rodent species, the “butterfly effect” is observed as a two-staged process.^1,2,5,6–12 This process involves the systemic injury-induced peripheral response of cellular entities at opposite sides along the mechanical sensibility spectrum and can be summarized as follows. First, a systemic rapid reaction occurs in which both mechanosensitive components (tactile and nociceptive) become compromised by the peripheral injury and react in opposite directions, losing (LT: desensitization) and gaining (HT: sensitization) sensitivity, respectively (Figure 1(a) and (b)). Second, the progressive loss of active afferents leads to a greatly diminished peripheral signal (deactivation) as nociceptive hypersensitivity reaches its maximum, depending on the magnitude of the injury and the dermatome (Figure 1(c) and (d)). The latter appears to be tightly correlated with the recovery (resolution) of the animal, ¹³ and this represents an important paradox, as it indicates a reduction in the density of active nociceptive afferents that are supposed to encode the occurring injury (Figure 1(a) and (c)).

A partial answer to the meaning and mechanisms behind activity-induced nociceptive deactivation was provided by the accidental observation that spontaneous activity (SA) triggers, over time, the occurrence of nociceptive-specific harmonic electrical disturbances (oscillations) on the cell’s membrane potential (Em). Rhythmic fluctuations in the membrane potential of excitable cells, such as neurons, are not new. ¹⁴ In single neurons, these oscillations may occur below the neuronal firing threshold (subthreshold). ¹⁵ In the central nervous system (CNS), it has been described as a mechanism for implementing intrinsic memory and establishing preferred input/output relationships,^16,17 thereby contributing to neuronal excitability and, in some cases, synchronized cellular discharges and bursting.^18,19

In the peripheral nervous system (PNS), subthreshold oscillations are rarely observed, and when present, they have been associated with changes in both normal and abnormal electrogenesis of primary sensory neurons. ²⁰ In general, these electrogenic changes have been described as occurring in only a handful of cells (5%) that can produce small (~3 mV peak-to-peak) reciprocating (symmetric) subthreshold oscillations, which may lead to ectopic spiking. ²¹ Although these descriptions have been associated with cells having myelinated fibers (A-fibers) and narrow action potentials (agreeable with a non-nociceptive electrical profile), ²² some authors have speculated on the role of these oscillations in maintaining or enhancing peripheral pronociceptive electrical activity after injury. ²³

Conversely, neither the shape nor the amplitude of the observed oscillations in nociceptive afferents was consistent with this description. Furthermore, rather than producing ectopic discharges, their occurrence led the cells to an unexpected dormant state. During this state, although still electrically excitable, the cells were unable to encode either thermal or mechanical stimulation, which led us to theorize that this electrical disturbance was, perhaps, a fundamental mechanism behind the “butterfly effect.”

Therefore, to clarify the function and properties of these oscillations, the current report also includes a study to evaluate the relations between nociceptive cellular activation, the genesis of cellular membrane oscillations, and their potential as the source of cellular electrical deactivation. To reduce variability and likely data ambiguity, the study focused on afferents with a clear nociceptive profile (C-polymodal nociceptors (CPNs) that belong to the C-HT nociceptive subpopulation but also respond to thermal stimulus). ²⁴ These cells were targeted due to their distinctive electrical signature (high input resistance, S-type action potential), their mechanical threshold, and cellular densities comparable to other nociceptive cellular types (A- and C-high threshold mechanonociceptors (HT)) innervating the T11 dermatome. In addition, these cells exhibit a robust response to thermal pronociceptive stimulation, which facilitates their controlled cellular activation (with cold and heat) without inducing tissue damage. Furthermore, they are also highly susceptible to developing SA discharges, ²⁵ a key hypothetical aspect of the cellular response directly linked to the “butterfly effect.”

Thus, the current report aims to present this theory in the context of our current challenges (and failures) to develop new non-opioid analgesics (using rodent pain models) and to provide a mechanism that may explain some of the most startling cellular events of the “butterfly effect.”

Methods

The present in vivo electrophysiological experiments were conducted on young adult males Swiss Webster mice (Swiss OF-1), ranging in age from postnatal day (P) 28 to P72 and weighing between 25 and 40 g. The Institutional Animal Care and Use Committee of the University of Wyoming and Wake Forest University Baptist Medical Center approved all procedures used in the present experiments.

Electrophysiology

The in vivo mouse thoracic preparation used in the present experiments has been described in detail. ²⁴ Briefly, the animals were anesthetized (ketamine 90%, xylazine 10%, mg/kg, ip), intubated, and ventilated (Minivent, Harvard Apparatus, Holliston, MA). The T11 ganglia were surgically exposed by laminectomy, and the preparation was moved to the recording chamber.

DRG cells were impaled with quartz electrodes (80–150 MΩ), filled with 1 M potassium acetate. Following, the T11 dermatome was explored (with a brush; Figure 2(a)) to locate the mechanical receptor field (RF) of the afferent, the cellular basal properties (CBP) established by injecting current pulses (SAP ¹ : IcP, 500 ms, ≤0.1 nA, 0.01 steps) followed by a current depolarizing ramp (RAMP ¹ : IcR: ~3 s 0–0.08 nA). This characterization included active and passive membrane electrical properties (SAP). The mechanical threshold (MT) and response dynamic (rapid-adapting (RA) or slow-adapting (SA)) were measured by the use of calibrated von Frey filaments (Stoelting, Wood Dale, IL), followed by thermal stimulation (TP) always in the same sequence (Cold pulse (32–0°C, 20 s) followed by a recovery (0–32°C, approx. 30 s) followed by two consecutive heat pulses (32–50°C, 10 s), separated by 30 s), The afferent conduction velocity (CV) was established by stimulating the afferent RF with a bipolar electrode (0.5 Hz) (eP), recoding the absolute minimal intensity to activate the neuron, and dividing the latency by the distance between the cellular RF and the ganglia. The cellular impalement was preserved for 45 min–3 h, and the presence of oscillations was re-evaluated by CBP at 45 and 90 min after (SAP ² and RAMP ² ; Figure 2(b)). Cells that lost more than 10 mV Em during the whole recording or presented less an Em ≥ to 30 mV were eliminated from the current study.

Figure 2.

Single-cell CPN response, development of non-reciprocal subthreshold oscillations, and its physiological consequences over time. (a) Schematic of the T11 in vivo model of physiological intracellular recording. (b) Study and activation protocol (CBP: cellular basal properties, RF: receptor field, SAP: cellular somatic properties, MT: mechanical threshold, CV: conduction velocity, eP: electrical stimulation pulse, TP: thermal pulse (cold and heat), IcP: intracellular current pulse, Ramp: intracellular current ramp). (c) Typical C-polymodal nociceptor spike (black) presented with the first derivative of the voltage (red). The arrow indicates the inflection point produced by Ca²⁺ conductance. (d) and (e) C-polymodal nociceptor pre-activation SAP response to IcP (500 ms, 0.02 nA), subthreshold (SubT, black) and suprathreshold (SupraT, blue) and Ramp (2 s, 0 to approx. 0.08 nA). (f) C-polymodal nociceptor response to thermal pulses (cold: 32–0°C, 20 s; heat: 32–50°C, 10 s²) and the concurrent development of spontaneous activity (SA). (g) and (h) C-polymodal nociceptor post-activation SAP response (45 and 90 min after TP stimulation) and membrane oscillatory activity. Data is presented with a detailed section of the membrane oscillatory response to IcP SubT and SupraT pulses and Ramp (black boxes, italics I–IV) and peak-to-peak absolute values (red) (voltage (mV) and max frequency (Hz)). (i) Relation between max depolarization peak and max hyperpolarization peak (mv) and its rate (mV/ms) during IcP-induced membrane oscillations 90 min after TP stimulation. (j) example of the effect of superfusion (bath) temperature on the spontaneous membrane oscillations (no SAP, 90 min after TP; different cell). The data is presented with the maximum frequency (Hz) and oscillation peak-to-peak values (mV, P-P) at three different temperatures (2 min exposure). Scale bars: (e) 1 s, (f) 10 s, (h) 1 s and 0.2 s, (j) 100 ms.

Statistical analysis

Before analysis, parametric assumptions were evaluated for all variables using histograms and descriptive statistics. For normally distributed measurements, means ± SD are used. For measurements not normally distributed, descriptive statistics are reported using medians (and range). Standard parametric (Student’s t-test, one-way ANOVA, linear regression) or nonparametric (Mann-Whitney U-test, Kruskall-Wallis) tests were used, depending on normality. Statistical tests were carried out using multiple packages (OriginPro 9.5, Northampton, MA; InStat/Prism, San Diego, CA).

Results

During this study, information on 45 CPNs’ afferents was collected. Of these afferents, 12/45 were exposed to only cold, whereas 33/45 used consecutive cold and heat (2X) stimulation to achieve RF acute pronociceptive activation.

CPNs’ basal electrical profile, conduction velocity, and sensibility

All the cells (45/45) showed an electrical profile (CBP) consistent with a nociceptive afferent (passive properties means: Em: −59 ± 1.4 mV, Ir: 190 ± 12 MΩ, τ: 2.9 ± 0.3 ms; active properties means: Spike amplitude: 73 ± 2.0 mV, Spike duration (D50): 1.5 ± 0.06, AHP amplitude: 17.3 ± 0.9, AHP duration (AHP50); 10.9 ± 0.5) with an inflection (Ca²⁺ hump) on the repolarizing phase (Figure 2(c)). In all cases, the consequence of depolarizing the cellular membrane, either by discrete current pulses (SAP) or constant current ramps (RAMP), was the cellular discharge (above rheobase). The magnitude of this discharge (# of AP per stimulus) correlated with the duration of the stimulus (tonic response) and the amount of injected current (Ic). No changes in the cellular electrical profile (e.g. Em) were detected either to subthreshold (SubT IcP) or suprathreshold (SupraT IcP) current intensities (Figure 2(d) and (e)). While their mechanical threshold was one order of magnitude above tactile afferents’ normal sensibility (MT: median 3.3 mN (range: 1.26–5.3 mN)), suprathreshold activation did not produce lasting spontaneous activity (SA). In the same manner, cold stimulation (CT mean: 8 ± 2°C) elicited a discrete cellular response failing to sensitize or to produce lasting SA, with only an occasional (6/45) paroxysmal discharge (16 ± 3°C) as the temperature returned to basal (32°C; Figure 1(f)). However, heat stimulation induced a robust response (HT mean: 44.6 ± 0.9°C), followed by sensitization (HT² mean: 35.7 ± 0.9°C) and low-frequency SA (~2 Hz) long-lasting (~60 min, max 3 h; Figure 1(f)). Their fiber conduction velocity (CV) was below the 1.2 m/s C-fiber cutoff (rodents; CV: mean 0.7 ± 0.04).

CPN’s electrical response to heat activation and SA

After stimulation (thermal), the cells were left undisturbed for over 30 min, and their Em was continuously monitored to ensure the stability of the cellular impalement. Cold-activated CPNs (12/12) neither developed SA, changed their MT, nor exhibited any detectable form of Em electrical instability.

By contrast, heat-activated cells (33/45) gradually develop Em instability and small-amplitude subthreshold oscillations (~15 min, 5 mV Peak-Peak (P-P)). As the oscillations developed, SA diminished (~0.2 Hz). After ~45 min, the use of IcP (SubT and SupraT) showed a medium amplitude subthreshold oscillation (amplitude P-P: 7.1 ± 3.5 mV (range: 3–14 mV), IFmax: 34.7 ± 4.2 Hz (range: 29–43 Hz); Figure 2(g) left; neither MT nor RAMP was evaluated to prevent cellular overactivation). Although the oscillations were evident at Em, the cells continued discharging (SA) until they finally ceased after ~90 min (only 2/33 cells continued for over 180 min). At this time point, the use of IcP (SubT and SupraT) showed a high amplitude subthreshold oscillation (amplitude P-P: 11.0 ± 2.5 mV (range: 2–23 mV), IFmax: 53.0 ± 4.0 Hz (range: 33–71 Hz); Figure 2(g) right). While the cells were still excitable and capable of producing somatic APs, none showed a still-active mechanical RF (complete mechanical desensitization). SupraT IcP-induced cellular discharge produces a momentary decline in the cellular oscillatory activity. Due to this decline, it was possible to establish a basal value for the Em (0 mV of deviation from the Em; Figure 2(g), inserts I to II, black trace), process the signal for noise extraction (red trace, Savitzky-Golay method, Figure 2(g) and (h)) and to calculate the membrane voltage above and below the Em during the SupraT IcP-induced oscillation (Figure 2(g), inserts I and II). RAMP (IcR: ~0.04 to ~0.08 nA) produced a non-linear increase in the oscillations’ amplitude (75%) and frequency (90%). Further increments in the IcR (>0.08 nA) reduced both amplitude and frequency of the oscillations. In either case, the IcR was unable to lead to cellular discharge (Figure 2(h) and inserts III and IV).

As shown in Figure 2(i), during these oscillations, the depolarization has 30% less amplitude (left panel) at a rate six times slower (right panel) than the hyperpolarization at both SubT and SupraT IcP’s. Consequently, the generated membrane oscillations were consistent with a non-linear (Ic vs IFmax at both 45 and 90 min), non-reciprocate (unbalanced toward hyperpolarization) sawtooth waveform and were temperature-dependent (Figure 2(j)).

Discussion

Neuronal oscillations and synchronization in the CNS have been extensively reported. This electrical activity has been linked to fundamental cognitive functions, including information transfer, ²⁶ memory, ²⁷ motor control, ²⁸ and perception. ²⁹ However, on the PNS, reports are outdated, scarce, and scattered. These reports primarily focus on the role of small, sinusoidal (harmonic) voltage oscillations on the membrane of putative tactile afferents (A-fibers) as the driving force behind ectopic spike activity in DRG primary sensory neurons.^30–34

The current report reveals an entirely different type of oscillation that affects the functionality and activity of physiologically identified polymodal nociceptors (CPNs). Our results indicate that these oscillations are activity-dependent, temperature-dependent, and voltage-dependent. Due to their shape (non-reciprocate toward hyperpolarization sawtooth), high amplitude (~11 mV, over three times the P-P amplitude reported on PNS cells),^20–23 and apparent non-linearity between the injected current (IcT/IcR) and its magnitude (amplitude/IFmax), rather than promoting ectopic spike activity in sensory neurons, these oscillations acted as a brake to nociceptive electrical hyperactivity. These results enable us to speculate on the potential of these oscillations as a mammalian control process to prevent spinal circuitry hyperactivation, functionally linked to the “butterfly effect” nociceptive deactivation response, and therefore, a primordial defense mechanism to avoid or mitigate the development of chronic pain syndromes in rodents.

Precedent and function of subthreshold oscillations on primary sensory neurons

Although oscillatory activity is a common feature of the CNS’s normal functions, in general terms, it has been described mainly as a harmonic (linear) dynamic process, asymmetric toward depolarization, ³⁵ primarily useful for temporally linking neuronal assemblies and synchronizing functions. ³⁶ More rarely observed, sawtooth oscillations are often linked to inhibitory mechanisms, ³⁷ associated with interneuron activity (network inhibitory interactions). ³⁸

As indicated above, this type of electrical activity has been rarely observed in the PNS.^20–23 It has been interpreted as a somewhat speculative cellular function due to the absence of a clear correlation with cellular modality. This is not the case with the current report. To our knowledge, this is the first time sawtooth oscillations have been reported on the PNS. Furthermore, we have correlated these microscopic rhythmic fluctuations in cellular Em with the afferent modality (CPN) and observed that this is a consequence of cellular activation (and SA generation). Although focused on the polymodal nociceptive population, similar oscillations have been observed in hypersensitive A-HT and C-HT (Boada MD, personal observation), depending on cellular input resistance and the magnitude of their evoked activity (hyperactivation).

The observation that, unless hyperactivated, cells will not develop any electrical response is consistent with our results using only cold stimulation. It potentially extends to the general nociceptive population (A-HT and C-HT) in experimental situations in which these cells have reached a hyperexcitable state.^1,2,5,6–12 Due to its amplitude, shape, and frequency, we concur with the general interpretation of its occurrence as an inhibitory electrical mechanism in both CNS^37,38 and PSN. However, any follow-up study aiming to corroborate the universality of these effects to other dermatomes and peripheral ganglia (e.g. TG), must consider the relationship between the oscillatory response (voltage peak to peak), the cellular input resistance (interpreted mainly as a measure of cellular diameter) and cellular hyperactivity (e.g. SA), expecting variability on this parameter (and its observability) assuming similar membrane subthreshold currents densities, something yet to be established.

Physiological function of the “butterfly effect”

The observation that this inhibitory electrical mechanism appears to occur only in nociceptive afferents and only as a consequence of their hyperactivation is both puzzling and provocative. Although nociception and pain are considered distinct, pain from injury cannot happen without nociception. ³⁹ It is broadly accepted that nociceptors are plastic; they sensitize (MT reduction) and rarely return to normal after an injury has occurred.^40,41 Thus, it seems logical to ask why such a cellular type with a clear protective function has a “breaking mechanism.”

From an evolutionary biology standpoint, ⁴² only those traits that increase the fitness of an individual or a population have adaptative value. ⁴³ In this way, acute pain, an essential defense mechanism, has been extraordinarily preserved across animal phyla (vertebrate and invertebrate). However, chronic pain, which compromises survivability, has no adaptive value, ⁴⁴ and there is virtually no evidence of its occurrence in other mammals beyond farm and companion animals and modern humans. ⁴⁵

Pain research has traditionally relied on animal models, mainly rodents, assuming they effectively recreate human chronic pathologies. ⁴⁶ Arguably, these models have been very useful in understanding specific aspects of pain pathologies and advancing disease-specific questions.^47–49 However, it is now clear that these models, along with the concurrent behavioral measures, have failed to lead to the translation of basic science into effective new analgesics.^50,51

This failure, largely attributed to modulatory factors (e.g. sex, genotype, and social communication), ⁵¹ has overlooked the obvious yet unacknowledged possibility that rodents may not be capable of developing chronic pain (as perceived by humans) (e.g. non-evoked, disabling pain). In the wild and throughout mammalian evolution, it is difficult to imagine a situation in which chronic, unbearable pain and disability would not lead to social isolation and compromise a rodent’s survival. Thus, although speculative, it is not unreasonable to attribute to the “butterfly effect” (and its electrical mechanism) a protective role with adaptive value, which leads to the preservation of the peripheral system’s functionality (first stage) while reducing the signal’s magnitude by decreasing the percentage of active afferents (second stage), thereby avoiding incapacity and, in the wild, death.

In humans, there is evidence of the first stage of the “butterfly effect” but not the second. It’s not uncommon to observe that a pain pathology is preceded by or concurrent with local or regional numbness (first stage). Furthermore, the latest has been proposed as an accurate predictor of disease worsening and potential chronic pain.^52,53 In the absence of evidence of any reduction in the percentage of injury-induced sensitized nociceptive afferents, and once the sensitization process has reached its maximum (leading to unrelieved spontaneous activity), it seems logical to predict that the undiminished peripheral nociceptive signal could easily overwhelm the spinal circuitry (central sensitization), leading, unlike in rodents, to the development of full pain pathology.

The electrical mechanism presented in the current report appears to be ideal for explaining, at least in part, the observed nociceptive deactivation process (second stage) that occurs in rodents, and its absence may be crucial in humans. However, several pieces of information must be obtained to demonstrate this hypothesis. First, the molecular entities responsible for this sawtooth oscillation must be identified and their presence linked to the cellular modality (nociceptors). This is a non-trivial initial step that can be performed in vitro using a combination of classical nociceptive markers and single-cell patch clamp recordings, provided that the physiological temperature and cellular activation requirements are considered. Second, when identified, these ionic conductances must be down-modulated (by pharmacological or molecular means) to correlate their dysfunction with a potential increase in the peripheral nociceptive signal (e.g. increased pain-related spontaneous behavior). Third, the presence (or lack thereof) of these molecular entities (active subthreshold ionic channels responsible for the sawtooth nociceptive oscillation) must be demonstrated in human DRG/TG neurons under identical experimental conditions.

Potential consequences of the “butterfly effect” on spinal cord integration

We are all aware of sensory thresholds, adaptation processes, and nonlinearity, which represent departures of perception from regular variations of magnitude in a physical stimulus. ⁵⁴ From a peripheral perspective and unlike other monomodal sensory systems (e.g. visual), the somatosensory system must deal with a multimodal-rich environment (e.g. mechanical and thermal) and the ambiguity generated by these multiple energy sources. ⁵⁵ Centrally, the integration process needs to optimize information transmission and interpretation, dealing with the interference of the “noise” always present in the form of a general background of spontaneous activity in the integration areas of the spinal cord. ⁵⁶ Preserving the informative integrity of the signal (modality) in the presence of these peripheral and central backgrounds requires reducing unwieldy complexity by canceling out some of the information and possibilities of its treatment. ⁵⁷ Mammalian evolution and survivability depend dramatically on the individual’s ability to rapidly identify the nature of the stimuli they detect. This recognition process doesn’t rely only on the output level attained, but, more importantly, it depends on the signal-to-noise ratio. ⁵⁸

For decades, the field of pain research has led to the belief that receptor-cell specificity is the fundamental means of signal identity recognition in the peripheral sensory system. ⁵⁹ This concept implies that stimulus encoding is based on the magnitude of the energy applied, activating molecular complexes that respond specifically to that amount and type of energy (e.g. TRPV1). ⁶⁰ Although somewhat valid for monomodal stimuli (e.g. thermal), this concept becomes contradictory when the stimulus activates cellular populations with opposite functions (tactile vs nociceptive), particularly in situations in which their mechanical activation threshold is no longer the discriminative factor (e.g. injury-induced nociceptive sensitization) or it has been altered by multimodal stimulation. ¹

The simplest way to eliminate contradictory information is to modulate the sensitivity of receptors in conflict. This method is used effectively in all complex living systems.^3,4,54 Therefore, it is unsurprising that the mammalian somatosensory system has a similar mechanism (the first-stage butterfly effect) to reduce ambiguity, prevent encoding errors, and ensure correct spinal integration and appropriate reflexive responses. ⁶¹ However, a second layer of complexity surrounds this mechanism and its impact on spinal processing. This layer involves the actual role, if any, of the tactile afferents that we know arrive in the superficial dorsal horn (SDH), ²⁵ the primary spinal cord area for integrating nociceptive signaling.

Thus, two mutually exclusive alternatives surround the function of the tactile afferents on the SDH: A. Under specific conditions (e.g. significant injury), tactile afferents have an excitatory influence on the circuit and, therefore, are likely responsible for the phenomenon of tactile allodynia, ⁶² or B. Their influence is inhibitory, and thus, sensitized nociceptive afferents are the driving force of this phenomenon. ⁶³ This question is not trivial, and the correct answer is directly linked to the modulation of the circuit input, the precise neuroanatomical organization of the afferents arriving at this area, the role of central descending modulatory projections, and the intrinsic spinal processing of both information sources (central and peripheral). In the absence of data about the physiological integration process at the SDH and based exclusively on the peripheral input to the circuit, our observations support the latter (B) for four fundamental reasons: (1) the divergent response of the sensibility of the mechanosensory afferents moving toward each other, (2) the overlap of their sensibilities as the effects of the injury reached their maximum (~1 mN = 0.102 gf), ¹⁰ (3) the catastrophic failure and recovery of the tactile afferents in parallel to the injury and its resolution, ¹³ and (4) the plastic changes in the size of the receptor field of both types of afferents matching the sensibility changes (tactile RF reduction, nociceptive RF expansion), ⁸ sometimes orders of magnitude below and above (respectively) the original cellular RF size prior the injury.

Conclusions

As the pressure to find a new non-opioid analgesic increases, so does the need to understand the boundaries and limitations of our models with as much precision as possible. Based on the evolutionary history of the mammalian lineage, it’s not surprising that rodents have a protective electrical mechanism (probably one among many others, such as the PDH mechanism ⁵ ) to prevent or mitigate the development of chronic pain. However, based on the existence of such mechanisms, and under the educated assumption that they may be suppressed in humans, the translatability of our preclinical models becomes significantly compromised.

This does not mean that our preclinical pain models are useless – quite the opposite, as illustrated by the current report. However, by the same account, it seems also clear that we do not fully understand the mammalian somatosensory system or its functional capabilities. Functional capabilities that extend far beyond the putative role of merely “static sensors,” suitable to be molecularly dissected in a Petri dish, as some authors have suggested. ⁶⁴ As demonstrated by the “butterfly effect” and its nociceptive oscillatory mechanism, the rodents’ somatosensory system adapts, thrives, and responds to our disruptive efforts. To uncover how much of this response is missing or suppressed in our species is of critical importance to enhance the efficacy of our translation efforts and the effectiveness of our clinical trials.