Somato–Autonomic Reflexes of Acupuncture

Abstract

Objective: Acupuncture, as an important part of Traditional Chinese Medicine, has been practiced for thousands of years in China and now all over the world, but the underlying neuroanatomical basis is still poorly understood. This article explores how acupuncture drives autonomic reflexes and why the widely used Streitberger sham-needling control should be revisited.

Method: This article summarizes modern studies, suggesting that functional connections between somatic tissues and internal organs may be explained via somato–autonomic reflexes.

Results: Modern studies have revealed a few organizational rules regarding how acupuncture drives distinct somatosensory autonomic pathways, including acupoint selectivity and intensity dependence. Activation of these autonomic pathways modulates various body physiologic functions, such as gastrointestinal motility and systemic inflammation. Meanwhile, extensive anatomical and functional characterization of the somatosensory system raises a question about the widely used Streitberger sham-needling control. Specifically, the skin epidermis and hair follicles contain mechanically sensitive afferents, whose activation by this sham stimulation could modulate pain and the autonomic nervous system.

Conclusions: A deeper understanding of the underlying neuroanatomical basis of acupuncture is crucial for optimizing stimulation parameters and designing proper sham-controls to demonstrate and improve the efficacy and the safety of using this modality to treat human conditions.

Introduction

Modern randomized clinical trials have demonstrated the efficacy of acupuncture for treating a range of conditions, including gastrointestinal (GI) disorders, stress urinary incontinence, and others.^1–5 Animal studies have also shown that acupuncture stimulation can modulate systemic inflammation powerfully.^6–11 These somato–visceral and immune effects of acupuncture reflect one of the core ideas of Traditional Chinese Medicine (TCM) and acupuncture practice. That is, stimulation at specific somatic tissues (acupoints) can modulate internal organ physiology distantly. The ancient meridian-channel theory provides a conceptual framework of somato–internal organ connections, but the physical presence of such channels has not yet been supported by modern anatomical studies, although acupoints appear to be located along the fascial tissues enriched with nerves, vascular/lymphatic vessels, and immune cells.^12–14

Method

This mini-review summarizes modern studies, suggesting that functional connections between somatic tissues and internal organs may be explained via somato–autonomic reflexes.

Results/Discussion

Acupoint Selectivity for Driving Somato–Autonomic Pathways

Somatosensory neurons are located in the dorsal-root ganglia and trigeminal ganglia, and are crucial for humans and animals to sense touch, temperature, pain, itch, and body positions. The autonomic nervous system (ANS) is divided into parasympathetic and sympathetic branches, which jointly modulate whole-body physiology. A functional connection between the somatosensory and ANS' has been recognized for a while. Since the 1970s, pinching or electroacupuncture stimulation (ES) at different body regions has been reported to drive different autonomic pathways associated with GI-motility control.^15,16 Strong pinching or ES at the abdominal region, but not at forelimb or hindlimb regions, can drive GI sympathetic pathways to inhibit motility, via both segmental spinal reflexes and supraspinal circuits.^15,16 Meanwhile, stimulation at limb regions can drive parasympathetic reflexes.^15,16

Anatomically, there are spinal ascending-projection neurons that innervate the nucleus of solitary tract (NTS) in the medulla oblongata, and NTS neurons, in turn, send synaptic connections to the dorsal motor nuclei of the vagus, where cholinergic parasympathetic vagal efferent neurons are located.^17,18 Sato and others found that pinching and high-intensity ES can drive vagal reflexes to promote GI motility, and, strikingly, that such parasympathetic reflexes can be evoked from forelimb/hindlimb acupoints, such as ST 36 (Stomach 36; Zusanli) but not from abdominal acupoints, such as ST 25 (Stomach 25; Tianshu), which are exactly opposite to GI sympathetic reflexes that can be evoked selectively from the abdominal region.^15,16 These early studies illustrate clearly a degree of somatotopic organization, in terms of driving distinct autonomic pathways from different acupoints or body regions that, positively or negatively, modulate GI motility. These preclinical findings might also explain why acupuncture can be used effectively to treat various GI disorders.^1–4

The presence of somatotopic organization was also demonstrated for ES to drive autonomic pathways that are capable of modulating severe systemic inflammation or cytokine storms. Cytokine storms, indicated by excessive release of a group of proinflammatory cytokines, contribute to high fatality rates for patients suffering from severe bacterial or viral infections.^19,20 In 2000, Tracey and colleagues first reported that electrical stimulation of cervical cholinergic vagal efferents could suppress systemic inflammation.^21,22 Subsequent studies suggested that vagal efferent activation can activate splenic sympathetic neurons located in the celiac ganglia, whose subsequent release of noradrenaline in the spleen can activate cholinergic T-cells; acetylcholine released from cholinergic T-cells can, in turn, suppress synthesis and release of proinflammatory cytokines from splenic macrophage cells.^21,22 However, studies from other laboratories suggested that activation of splenic sympathetic neurons following vagal-nerve stimulation might be mediated via activation of vagal afferents and subsequent descending control of sympathetic neurons from the hindbrain, rather than via vagal efferents.^23,24

Yet, inspired by the discovery of the cholinergic anti-inflammatory pathway, several studies showed that ES at limb-area acupoints could suppress systemic inflammation, mainly or partly via activation of vagal efferents.^6–10 In particular, Ulloa and colleagues showed that ES at the hindlimb ST 36 acupoint could drive a vagal–adrenal axis, which leads to dopamine release and subsequent suppression of systemic inflammation.²² Most recently, Liu et al. found that chromaffin cells associated with systemic inflammation modulation can be marked by the expression of the neuropeptide Y (NPY).¹¹ Using genetic tools to create mice with selective removal of NPY⁺ chromaffin cells, Liu et al. demonstrated that this vagal–adrenal axis can be evoked by low-intensity ES (0.5 mA, 50 μs pulse, 10 Hz, for 15 minutes) at the hindlimb ST 36 acupoint,¹¹ which is different from the requirement of higher stimulation intensity to drive gastric vagal reflexes.²⁵

In any event, both types of vagal reflexes can be activated by ES at hindlimb acupoints but not at abdominal acupoints. In contrast with this shared somatotopic organization in driving different vagal reflexes, the requirement for driving distinct sympathetic pathways is different. While GI sympathetic pathways can be activated selectively by pinching or ES at abdominal acupoints, the splenic sympathetic pathway can be activated from all body regions tested, including both abdominal and hindlimb acupoints.¹⁵ A key direction for future studies is to explain the underlying neural anatomical basis for acupoint-dependent or -independent driving of various autonomic pathways.

Intensity Dependence for Driving Distinct Autonomic Pathways

Somatosensory neurons are molecularly and functionally heterogeneous. For example, sensory axons are of distinct axon diameters, myelination degrees, and ion channel profiles, all of which can make an impact on activation thresholds of sensory afferents in response to electric stimulation.^26–28 Accordingly, different stimulation intensities will presumably activate different spectrums of sensory afferents, with electrophysiologically lower-intensity stimulation activating larger myelinated fibers while a higher intensity is needed to activate thinly myelinated and unmyelinated fibers.^26–28 Activation of distinct sensory afferents might, in turn, drive distinct autonomic reflexes. Indeed, a recent study showed that, while low-intensity ES (0.5 mA) at the hindlimb ST 36 acupoint is sufficient to drive the vagal–adrenal axis, higher stimulation intensities (1–3 mA), either at ST 36 or at the abdominal ST 25 acupoint, are needed to drive spinal sympathetic reflexes, including activation of the splenic sympathetic pathway by ES at ST 25.¹¹ This intensity dependence echoes earlier findings by Sato showing that low-intensity mechanical stimulation at abdominal regions can suppress sympathetic reflexes via a supraspinal mechanism, whereas high-intensity pinching can drive both splenic and gastric sympathetic pathways via segmental and supraspinal reflexes.¹⁵

Future studies are needed to determine sensory afferents that are activated by 0.5 mA versus 1–3 mA at the same ST 36 acupoint or at different acupoints, and determine how their activation drives vagal–adrenal reflexes versus spinal–sympathetic reflexes.

Disease State–Dependent Modulation of Pain and Systemic Inflammation

Acupuncture is generally considered to be a safe practice for treating human conditions. Herein, 2 recently reported studies highlight disease stage–dependent safety issues associated with acupuncture practice. The first study was reported by Zucker and colleagues on treatment of fibromyalgia by manual acupuncture.²⁹ These researchers divided the patients into 2 groups based on the degrees of pressure-pain tenderness. For patients with low-level pressure-pain tenderness, verum acupuncture at known acupoints induced better pain relief in comparison with sham acupuncture at non-acupoints. In contrast, in patients with sensitized (high-level) pressure-pain tenderness, verum acupuncture increased pain intensity, whereas sham acupuncture paradoxically relieved their pain.²⁹

In a more-recent study, Liu et al. also reported disease state–dependent modulation of systemic inflammation by high-intensity ES (1–3 mA).¹¹ As mentioned above, such high-intensity stimulation can activate splenic noradrenergic sympathetic neurons, at least for ES at ST 25. If 3 mA ES is performed right before the injection of the bacterial endotoxin (lipopolysaccharides; LPS), activation of splenic noradrenergic neurons can suppress LPS-induced cytokine storms effectively, via activation of β2 adrenergic receptors in splenic cells. In contrast, if 3 mA ES is performed after LPS-induced cytokine storms have reached their peaks, this high-intensity stimulation will make inflammation worse and animals' survival rates will be reduced. This drastic switch is due to LPS-mediated induction of α2A adrenergic receptors in splenic immune cells,¹¹ whose activation promotes inflammation.^30,31 A blockage of α2A adrenergic receptors can enable 3 mA ES to regain the ability to suppress ongoing systemic inflammation.¹¹

In contrast with disease state–dependent modulation by high-intensity ES, low-intensity ES (0.5 mA) at the ST 36 acupoint—which drives the vagal–adrenal axis selectively—can be used to both prevent and treat LPS-induced systemic inflammation, thereby operating in a disease state–independent manner.¹¹ Thus, acupuncture-stimulation parameters need to be adjusted for different disease states to produce beneficial effects and prevent detrimental ones. These studies also highlighted a rarely appreciated safety issue associated with acupuncture practice.

A Revisit to Sham Acupuncture Controls

Researchers doing human clinical trials often face a difficult situation: verum and sham acupuncture stimulations produce similar beneficial effects in comparison with conventional treatments. There are 2 commonly used sham controls: (1) skin-penetrating acupuncture at non-acupoints or (2) skin-nonpenetrating stimulation at acupoints, such as the Streitberger sham-needling control.³² However, from the neurobiologic point of view, there is no real inert sham control. For example, there is a dense nerve network within the skin epidermis and hair follicles throughout the body. This network includes a large group of mechanically sensitive sensory neurons: (1) low-threshold Aβ/Aδ mechanoreceptors for tactile perception; (2) Aδ nociceptors for sharp or sharp pain perception; (3) mechanically sensitive, unmyelinated C-fiber polymodal nociceptors that respond to light punctate noxious mechanical stimuli; (4) silent mechanoreceptors that gain mechanical sensitivity after tissue injuries, and (5) C-fiber low-threshold mechanoreceptors that produce pleasant tactile sensations.^33,34

As such, Streitberger's sham-needle stimulation could activate a range of mechanoreceptors. Depending upon what biologic problems are studied, such activation could be problematic. For example, for pain modulation, activation of low-threshold Aβ mechanoreceptors could produce analgesic effects, via activation of spinal inhibitory neurons, according to the classic gate-control theory of pain,³⁵ although with loss of gate control in pathologic conditions, activation of Aβ mechanoreceptors could activate pain pathways, leading to bidirectional modulation of pain by tactile stimuli.^36,37 In humans, in response to skin pinching stimulation, firing of mechanically sensitive C-fiber polymodal nociceptors (plus other sensory afferents) produces pressure or sharp perception; pain perception arises only after these neurons stop firing.^38,39 This anticorrelation between the firing of C-fiber polymodal nociceptors and pain perception raises a possibility that activation of these nociceptors might produce antipain effects, which might happen in response to non-penetrating Streitberger sham-needling stimulation given that these neurons normally respond to light noxious punctate pressure evoked by von Frey filaments.⁴⁰

Activation of C-fiber low-threshold mechanoreceptors, which produce pleasant tactile perceptions,^33,34 could also contribute to analgesic effects, given that pain is considered to be the opposite of pleasure or reward.⁴¹

Furthermore, activation of cutaneous sensory fibers could modulate conditions by influencing the ANS. For example, activation of Aβ mechanoreceptors has been implicated in suppressing splenic and adrenal sympathetic reflexes, via spinal–supraspinal pathways,¹⁵ and their activation could have an impact on those forms of pain that are partially maintained by tonic firing of sympathetic neurons.⁴² Thus, use of the Streitberger sham needling as an “inert sham” control could be scientifically invalid. To date, proper designing of a sham control might represent one of biggest challenges in clinical studies. A possible solution will come after a deeper understanding of the studied problems is reached. For example, for diseases whose treatment requires activation of parasympathetic neurons—which can be evoked from limb acupoints such as ST 36—abdominal stimulation could be used as a sham control, given that vagal reflexes cannot be evoked from the abdominal region.

Conclusions

Modern neuroanatomical studies have started to reveal that acupuncture stimulation can drive various somatosensory autonomic pathways in an acupoint- and intensity-dependent manner. Activation of these autonomic pathways can dynamically—even bidirectionally—modulate disease progression. These lines of progress will eventually help acupuncture researchers optimize stimulation parameters and design proper sham controls in order to improve research on the efficacy and safety of acupuncture practice.

Author Disclosure Statement

No financial conflicts of interest exist.

Funding Information

Work done in the current author's laboratory was supported by National Institutes of Health grants (No. R01AT010629), the Harvard/Massachusetts Institute of Technology Joint Research Program in Basic Neuroscience, and a Wellcome Trust grant (No. 200183/Z/15/Z).