In this study, we show how the use of electroacupuncture (EA) at specific points stimulates mesenchymal stem cell (MSC) release into peripheral blood through the activation of the nervous system. EA could be used to aid tissue repair through increasing the levels of circulating MSC. Moreover, MSC can be harvested directly from the blood of EA‐treated humans and animals and expanded ex vivo. Thus, EA may be a low cost, low risk method for MSC harvest for autologous stem cell therapy.

While both anatomical characteristics and local mediators of signaling are associated with specific acupoints, the mechanism responsible for the beneficial systemic effects and healing associated with acupuncture still lacks understanding. In this study, we sought to understand the central and peripheral nervous system response to EA using two sets of points (1: LI‐4, LI‐11, GV‐14, Bai‐hui ; 2:ST‐36, LIV‐3, GV‐14 and Bai‐hui ; Supporting Information Fig. S1) associated with successful treatment of arthritis and asthma and with modulation of immunity 5 .

The primary mechanism implicated in the anti‐nociceptive effect of acupuncture involves release of opioid peptides in the central nervous system (CNS) in response to the long‐lasting activation of ascending sensory tracks during the stimulation 1 . Adenosine has also been implicated, and interfering with adenosine metabolism prolonged the clinical benefit of acupuncture 4 .

Acupuncture, one of the oldest medical therapies, mediates its therapeutic effect through the insertion of needles into specific points in the body called acupoints 1 . Electroacupuncture (EA) combines traditional acupuncture with modern electrotherapy as a means to enhance the stimulation at the acupoints. Acupoints are located in areas of decreased electrical resistance and increased electrical conductivity in the body attributed to both neural and vascular elements in the dermis or hypodermis 2 . Histological studies reveal that acupoints are located in areas with high densities of free nerve endings, arterioles, lymphatics, and mast cells 3 . Certain acupoints have been identified to correspond to specific neural structures, such as superficial nerves and nerve plexuses 3 .

Results

EA Induces Activation of Hypothalamic Regions of the Brain in Both Rats and Humans The systemic beneficial effects of EA may be centrally driven; therefore, we sought to determine the relationship between the hypothalamus and other brain structures, which is termed “connectivity.” The hypothalamus plays a critical role as a primary homeostatic center in the brain and contains neurons with important projections to other limbic sites and sympathetic nuclei directly communicating with the periphery. To determine whether the hypothalamus was involved in the EA response, blood oxygen level‐dependent (BOLD) functional magnetic resonance imaging (fMRI) was performed in anesthetized male Sprague‐Dawley rats (n = 6) during a single EA session. Connectivity was derived from four time points: baseline, 0–8 minutes during EA, 9–22 minutes, and immediately post‐EA. Seed regions included the anterior, posterior, and lateral hypothalamus. EA‐stimulation produced changes in rats in the strength of functional connectivity within the hypothalamus and between the hypothalamus and adjacent brain regions, such as the amygdala (Fig. 1A), compared with baseline and the post‐EA period. A representative example of the increase in signal (over time) in the PVN of a rat receiving acupuncture is shown in Figure 1B. Arterial spin labeling fMRI was performed in healthy human subjects (n = 6) during a single EA session. Connectivity was derived from four time points: baseline, 0–8 minutes during EA, 8–16 minutes during EA, and immediately post‐EA. EA stimulation produced changes in humans in the strength of functional connectivity within the hypothalamus and between the hypothalamus and adjacent brain regions (Fig. 2) compared with baseline and the post‐EA period. Thus, rats and humans exhibited a similar response in terms of the connectivity observed. Figure 1 Open in figure viewerPowerPoint Electroacupuncture (EA) stimulation increases hypothalamic functional connectivity in rats. (A): Rat brains were monitored through functional magnetic resonance imaging during administration of EA. Functional connectivity increased within the hypothalamus and between the hypothalamus and adjacent brain region with progression of treatment (n = 6). (B): A representative example of the increase in signal (over time) in the PVN of a rat receiving acupuncture (n = 1). Abbreviations: EA, electroacupuncture; PVN, paraventricular nucleus. Figure 2 Open in figure viewerPowerPoint Electroacupuncture (EA) stimulation increases hypothalamic functional connectivity in humans. Brains from normal subjects were monitored using functional magnetic resonance imaging before, during, and after EA. Functional connectivity increased within the hypothalamus and between the hypothalamus and adjacent brain regions with progression of treatment (n = 6). As these regions have been associated with hematopoietic stem cell mobilization 6, peripheral blood was examined in the rat before and following EA for evidence of mesenchymal stem cell (MSC) mobilization. A 313% increase in the circulating MSC population, Lin‐CD90+CD44HI cells, was detected in the blood of EA‐treated rats 2 hours post‐EA compared with baseline (Fig. 3A, 3B). Figure 3 Open in figure viewerPowerPoint Electroacupuncture (EA) stimulation induced mesenchymal stem cell (MSC) mobilization. (A): Rat peripheral blood MSC were increased (p = .0063) after EA. Circulating MSC were defined as Lin‐(CD45‐CD31‐erythroid‐CD11b‐) cells that were positive for CD44 and CD90. Gated cells increased post treatment (n = 11 for baseline and 4 hours, n = 9 for 2 hours). (B): Representative flow charts for rat Lin‐ cells are shown at baseline and 2 hours samples. (C): The percentage of human peripheral blood MSC increased in post EA‐treatment (p = .006, n = 6). (D): The percentage of circulating MSCs from adipose tissue (AD‐MSC) is significantly elevated 2 hours post EA‐treatment (p = .033, n = 4). (E, F): EA‐mobilized MSCs were expanded in vitro. After undergoing adipogenesis differentiation, EA‐mobilized MSCs developed fat deposits as seen by Oil Red staining, which were not seen in the undifferentiated control cells (G, H). Magnification bars = (E, G): 100 µm; (F, H): 50 µm. Abbreviations: ASC, adipose stem cells; EA, electroacupuncture; MNC, mononuclear cell.

Characterization of Human EA Mobilized MSC as Adipocyte‐Derived MSC The availability of human antibodies to fully characterize the populations of cells mobilized by EA lead us to next examine the populations of cells mobilized by EA in the human subjects that underwent fMRI. Due to limitations regarding the fMRI equipment used, only points GV‐14, LI‐11, and LI‐14 were able to be used in this experiment. We observed an increase in MSC in the peripheral blood 2 hours following EA in all subjects (Fig. 3C), while total lymphocyte numbers did not change. To further characterize the MSC population and potentially determine whether they were released from adipose stores, the levels of CD34+CD45‐CD31‐ cells were examined. Two hours following completion of EA, CD34+CD45‐CD31‐ cells were increased in four of six individuals (Fig. 3D) and the response was proportional to the body mass index of the subjects as previously noted 7. Two individuals with body mass index less than 18.5 did not demonstrate an increase in this population. Human EA mobilized MSC were expanded in vitro and underwent adipogenic differentiation. Adipogenesis potential was confirmed by the cells’ ability to form lipid droplets as established through oil red O staining (Fig. 3E–3H). Because hypothalamic activation (Fig. 1B) preceded the mobilization of circulating MSC, it was proposed that the EA‐induced connectivity changes contributed to the subsequent release of these cells into the peripheral blood. Exogenous administration of epinephrine or dopamine in rats resulted in a similar increase of Lin‐CD90HICD44+ cell population into the circulation (Supporting Information Fig. S2A–S2C), supporting the role of CNS in mobilization of MSC and in EA activating these key CNS centers.

EA‐Induced Uncoupling Protein 1 Expression Promotes the Browning of White Adipose Tissue in Rats To confirm that EA can activate the sympathetic nervous system (SNS), we asked whether EA was associated with browning of white adipose tissue (WAT). The “inducible/recruitable” brown‐like (beige a.k.a bright) adipocytes arise in WAT and are functionally indistinguishable from adipocytes in the canonical (constitutive) brown adipose tissue 8, 9. The function of brown and beige adipocytes is muscle‐independent thermogenic energy dissipation, which relies on the function of uncoupling protein 1 (UCP1) 10. Following 14 days of EA (every other day) in rats, UCP1 immunofluorescence of inguinal subcutaneous WAT was increased (Fig. 4A), indicating a greater number of brown adipocytes, compared with control (sham EA) (Fig. 4B). Consistent with the notion that SNS signaling mainly activates subcutaneous WAT, no changes were detected in intraperitoneal WAT. Figure 4 Open in figure viewerPowerPoint Electroacupuncture (EA) increases sympathetic activation leading to browning of white adipose tissue and the effects of EA can be duplicated by pharmacological disinhibition of hypothalamus (sympathetic activation). (A, B): UCP1 immunofluorescence (red) detectable in inguinal subcutaneous adipose tissue (blue: adipocytes nuclei) from animals that underwent EA treatment (A) but not in control (B). EA resulted in an increase in beige adipocytes (n = 4). Magnification bars = 50 µm. (C): The effects of EA can be duplicated by pharmacological disinhibition of hypothalamus. Rats underwent injection of either vehicle, 30 pmol or 50 pmol/100 nL of the GABA A receptor antagonist bicuculline methiodide. Sites of injections are represented on a coronal section from the tuberal hypothalamus from a Standard Steroetaxic Atlas of the Rat brain 11. Colored circles indicate injection sites (black, orange, and red represent vehicle, 30 pmol and 50 pmol, respectively). (D): Representative photomicrograph showing an injection site from one rat. Magnification bar = 1 mm. (E): There was a significant increase (p = .027) in Lin‐CD90HICD44+ cells in the plasma 4 hours post injection (n = 6). Data presented as means ± SEM. Abbreviations: BMI, bicuculline methiodide; DMN, dorsomedial hypothalamic nucleus; f, fornix; MNC, mononuclear cells; mt, mammillothalamic tract; PeF, perifornical hypothalamus; PH, posterior hypothalamic nucleus; VMN, ventromedial hypothalamic nucleus; 3V, third ventricle.

Pharmacological Disinhibition of the Dorsomedial Regions of the Tuberal Hypothalamus Mobilizes Lin‐CD90HICD44+ Cells into the Circulation To confirm that activation of the SNS can mediate release of MSC, we performed stereotaxic injections with the GABA A receptor antagonist bicuculline methiodide (BMI) in the dorsomedial regions of the tuberal hypothalamus to disinhibit these regions. Histological verification of injection sites is indicated in the illustration (Fig. 4C). The exact location of all injection sites are shown on coronal sections from a Standard Stereotaxic Atlas of the Rat Brain 11 (Fig. 4C) and a photomicrograph that shows the representative injection site (Fig. 4D). In rats, BMI at the dose of 50 pmol increased the percentage of circulating Lin‐CD90HICD44+ cells (F (2, 14) = 6.7, p = .027) at 4 hours post injection of BMI (Fig. 4E) in the absence of EA. These results demonstrate that circulating Lin‐CD90HICD44+ cells are mobilized during hypothalamic disinhibition and EA. However, the delay in release (4 hours) compared with EA (2 hours) may suggest an alternative mechanism of release.

EA‐Treated Rodents Exhibit Reduced Mechanical Hyperalgesia, Increased Serum Interleukin‐10 Levels and Enhanced Tissue Remodeling following Partial Achilles Tendon Rupture To address the possible anti‐inflammatory and analgesic effects of EA‐mobilization of circulating MSC 7, we analyzed the contribution of an EA treatment paradigm on injury‐induced hyperalgesia in rats in the presence and absence of a β‐blocker, propranolol. Treatments included EA (i.e., applied to the forelimb LI‐4, LI‐11, GV‐14, and Bai‐hui, and hindlimb ST36, LIV‐3, GV‐14 and Bai‐hui immune points in horses), sham EA (i.e., applied to skin not associated with an acupoint) or EA plus propranolol (i.e., potential inhibitor for sympathetic effects of EA). EA or sham EA was administered every other day for 2 weeks following injury. Using sham EA applied to nonimmune acupoints or EA plus propranolol, nociceptive behavior elicited by von Frey mechanical stimulation did not change over the time course in the hind paw ipsilateral to the injury (Supporting Information Fig. S3). In contrast, mechanical hyperalgesia (assessed at both day 7 and 14) was considerably decreased (i.e., able to tolerate more pressure) in injured rodents subjected to EA application at LI‐4, LI‐11, GV‐14, and Bai‐hui (Fig. 5A). At early stages of tendon repair, the granulation tissues mainly synthesize type III collagen, while at later stages of healing, intrinsic fibroblasts produce type I collagen, whose fibers are orientated longitudinally to replace type III collagen. At 14 days post‐injury, EA sham‐treated or EA plus propranolol did not change type I collagen content, while type I collagen was significantly enhanced by EA (Fig. 5B); there was no change in type III collagen across treatments (Fig. 5C). Taken together, these data suggest that EA may enhance the replacement of thinner and immature type III collagen fibers with mature type I collagen fibers in the injured tendon 12, thereby supporting a better quality of regeneration and tissue reorganization. Given that EA can attenuate mechanical hyperalgesia and enhance tissue reorganization, we examined possible alterations of the anti‐inflammatory cytokine, interleukin (IL)‐10, in blood plasma. Previous evidence suggested that EA following surgical trauma contributed to increased levels of IL‐10 by T‐cells 13 and focal microinjections of IL‐10 diminish mechanical hyperalgesia 14. The results herein demonstrate that tendon injury in rodents subjected to sham EA or EA combined with propranolol fail to produce detectable changes in plasma IL‐10. However, injured rodents treated with EA significantly increased IL‐10 in plasma. These results indicate that EA increases production of the endogenous anti‐inflammatory cytokine IL‐10 (Fig. 5D). Figure 5 Open in figure viewerPowerPoint Electroacupuncture (EA)‐treated rodents exhibit reduced mechanical hyperalgesia, enhanced tissue remodeling, and increased serum IL‐10 levels following partial Achilles tendon rupture. (A): Effects of EA application on mechanical allodynia in rats at 7 and 14 days after partial tendon rupture in the right hind leg. Mechanical hypersensitivity was determined by measuring the change in weight‐bearing forces on the affected limb. Behavioral changes in the hind paw tactile threshold (in millinewtons, mN) were observed in the hindpaw ipsilateral to the tendon injury 18 hours after EA treatment, EA sham treatment or EA treatment with propranolol. (*, p < .01 versus baseline. EA: n = 9; EA: sham n = 7; EA + propranolol: n = 10). (B): EA increased type I collagen content in injured tendons in rats at 14 days after unilateral Achilles tendon partial tenotomy. In EA‐treated animals (n = 6), type‐I collagen content was 24% greater in injured tendons than EA sham treated tendons (n = 7; p < .05) and 28% greater in injured tendons than EA + propranolol treated tendons (n = 8; p < .02). (C): In contrast, there was no difference in type‐III collagen content between injured EA and EA sham or EA + propranolol treated tendons (p = .67). (D): IL‐10 serum levels were also elevated in EA‐treated rats compared with EA sham and EA + propanolol animals (p = .0041). Data presented as means ± SEM. Abbreviations: EA, electroacupuncture; IL‐10, interleukin‐10.

EA Performed over Immune Points in Pirt‐GCaMP3 Mice Rapidly Activates Primary Sensory Neurons We next examined mice to confirm that EA stimulation of acupoints (LI‐4, LI‐11, and GV‐14 and Bai-hui) resulted in mobilization of MSC. At 4 hours post EA, murine MSC as defined by Lin‐PDGFRa+Sca‐1+ cells were significantly increased (p = .01) in peripheral blood, a response that was markedly reduced when mice were pretreated with propranolol (Supporting Information Fig. S4). Immune acupoints are present on both the front and hind limbs of mice and are differentially used based on their ease of access in the particular species undergoing EA. Thus, we used acupoints in both front limbs (LI‐4, LI‐11) and hind limbs (ST‐36 and LIV‐3). To distinguish the afferent stimulation of the hypothalamus by peripheral acupoints, we treated Pirt‐GCaMP3 mice with either a noxious hind paw pinch (100g force) or EA directed at the hind limb acupoints (Fig. 6A–6F) [15]. Hind paw stimulation evoked robust and transient increases in 10–20 neurons per DRG in naïve mice, on average 12.7 ± 3.0 per DRG (Fig. 6A, 6B) of which nearly all were small diameter neurons (<20 µm; Supporting Information Fig. S5; Supporting Information Video 1) suggestive of pain fibers. Placement of acupuncture needles alone did not elicit activity in sensory neurons (Fig. 6C, 6D; Supporting Information Video 2). We noticed a striking pattern of neuronal activation in the DRG following EA stimulation; many activated neurons were medium to large diameter neurons (medium [20–25 µm]; large [>25 µm] with an average of 6.1 ± 4.0 neurons per ganglia; (Fig. 6E, 6F; Supporting Information Video 3) suggesting activation of touch fibers.