in vivo 19: 495-500 (2005)

Electroacupuncture and Moxibustion Influence the Lipopolysaccharide-induced TNF-· Production by Macrophages

ERI AOKI, TAKAKO KASAHARA, HISAKO HAGIWARA, MIKAKO SUNAGA, NAOKO HISAMITSU and TADASHI HISAMITSU

Department of Physiology, School of , Showa University, 1-5-8, Hatanodai, Shinagawa-ku, Tokyo, 142-8555, Japan

Abstract. The purpose of this study was to evaluate the performed in animals. It is thought that when EA and Mox are modifying effects of electroacupuncture (EA) and moxibustion applied to specific body areas, for instance the fourth governor (Mox) through the communication networks of the neuro- vessel (GV4) point, which is on the post meridian immune system using the two-step bacterial stimulation method. line, between the second and third lumbar processes, they may When EA or Mox treatment was implemented immediately modulate immunological functions and maintain mechanisms following injection of the Streptococcus pyogenes preparation by which the body defends itself against disease. The authors OK-432, a significant suppression of the tumor necrosis factor previously reported that, in tests conducted on mice, when EA (TNF)· production by the peritoneal macrophages induced by was administered at the treatment point corresponding to the lipopolysaccharide (LPS 1 Ìg/ml) was observed. When the GV4 point in humans, the delayed type of hypersensitivity to stimulation with EA or Mox was reduced in volume, the degree 2,4,6-trinitro-chlorobenzene (TNCB) was either enhanced or of suppression decreased correspondingly. Naloxone antagonized suppressed, depending on the timing of treatment to some extent the suppression of TNF· induced by EA, but did administration (1). We also reported that EA, administered a not compete with the suppression of TNF· release induced by single time at an acupuncture point corresponding to the GV4 Mox. These results suggest that activated macrophages are an point in humans, immediately prior to the challenge of TNCB, important target of the immuno-suppressive effects of EA and suppressed ear swelling that had been induced by the delayed Mox and that Ì-opioid receptor-mediated mechanisms are type of hypersensitivity, through a central opioidergic responsible, to some extent, for the suppressive effect of EA, mechanism (2, 3). In another report, we indicated that a single although Mox may not be dependent on these mechanisms. administration of Mox at an acupuncture point corresponding to the GV4 point in humans affected the production/release of Acupuncture (electroacupuncture; EA) and moxibustion lipopolysaccharide (LPS)-induced cytokines such as the (Mox) have come into use as alternative treatments in the cytotoxic factor (TNF) and interferon in mouse sera (4). management of diseases. EA is a mechanical somato-sensory In this study, we demonstrated that EA and Mox stimulatory technique, while Mox is a thermal somato-sensory modulated the reactivity against LPS of macrophages that stimulatory technique. professionals employing were infiltrated following treatment with the Streptococcus acupuncture and moxibustion use appropriate acupuncture pyogenes preparation OK-432 (5), modifying the ability of points in treating various diseases, which include excessive the macrophages to produce cytokines. We also showed that inflammation and abnormalities of the immune system. Many the effects of EA and Mox were similar, but that they had people have experienced acupuncture which is becoming, different mechanisms of action. however slowly, accepted by Western practising physicians. Furthermore, acupuncture experiments have also been Materials and Methods

Mice. Specific pathogen-free, male C57BL/6 mice were obtained from Nihon SLC Co., Ltd. (Hamamatsu, Japan). The mice were Correspondence to: Tadashi Hisamitsu, M.D., Ph.D., Professor and eight weeks old and were kept in plastic cages, given food and Chair, Department of Physiology, School of Medicine, Showa water ad libitum and maintained in a temperature-controlled room University, 1-5-8, Hatanodai, Shinagawa-ku, Tokyo, 142-8555, (23±1ÆC) with an LD of 12:12. Japan. Tel:+81-03-3784-8110, e-mail: [email protected] Electroacupuncture (EA) treatment. Each mouse was held in a Key Words: Peritoneal macrophage, tumor necrosis factor ·, LPS, mouse holder (KN-330, Natsume Seisakusho Co., Ltd., Tokyo, electroacupuncture, moxibustion, OK-432. Japan) and EA with low-frequency mechanical stimulation (1 Hz,

0258-851X/2005 $2.00+.40 495 in vivo 19: 495-500 (2005)

500 Ìs or 50 Ìs in duration) was applied at a point on the post- meridian line, between the second and third lumbar processes, anatomically equivalent to the GV4 acupuncture point in humans. A wave 500 Ìs in duration was generated from a pulse generator (NC-707, Kimura Med. Ins. Co., Ltd., Tokyo, Japan) and applied to each EA-15 group for 15 minutes. For the weak pulse electroacupuncture (PA)-15 group, a wave 50 Ìs in duration was generated from a pulse generator (PG 701, Suzuku Iryoki, Tokyo, Japan) and applied for 15 minutes. In order to standardize conditions, the mice were normally treated with EA or PA between 1:00 and 3:00 p.m.

Moxibustion (Mox) treatment. Thermal stimulation using moxa (moxa grass products) was also applied to a point anatomically equivalent to the GV4 acupuncture point in humans. Small pieces of moxa, turned in the shape of a cone by hand (moxa cones), were placed on the skin at the acupuncture point and the moxa was burned. One cone comprising two mg of moxa was applied each time for a total of five times on each mouse in the Mox-10 group, while a cone of two mg of moxa was applied once on each mouse in the Mox-2 group. In order to standardize conditions, the mice were Figure 1. The course of the TNF· level induced by LPS from peritoneal normally treated with moxa between 1:00 and 3:00 p.m. infiltrated macrophages. Each point represents mean±SD obtained from 6 distinct animals for each time point. Animal priming and collection of peritoneal cells. To induce in vivo priming (primary immunization) and in vitro induction, we used the two-step stimulation method (6). As the priming agent, we used a preparation consisting of an inactivated and lyophilized low- virulence strain of Streptococcus pyogenes (OK-432, provided by Measurement of tumor necrosis factor (TNF)·. The TNF· level of Chugai Pharmaceutical Co., Ltd., Tokyo, Japan). Peritoneal cells cell-free supernatants was measured using an enzyme-linked were primed through an i.p. injection at a dose of 0.1 mg of dry immunosorbent assay kit (Mouse ELISA TNF·, Endogen, Inc., streptococci for each mouse, suspended in an isotonic sodium MA, USA). chloride solution (Otsuka Pharmaceuticals, Tokyo, Japan). Immediately following immunization, EA or Mox was administered Systemic injection of naloxone. Five minutes prior to treatment with a single time. After 48 hours, the peritoneal cells that had EA and Mox, each mouse was injected i.p. with naloxone infiltrated into the peritoneal cavity were harvested from the cavity hydrochloride (Sigma Chemical Co.), in the form of an isotonic using ice-cold Hanks' Balanced Salt Solution (HBSS). After being sodium chloride solution (Otsuka Medicine, Tokyo, Japan) at a washed twice with cold HBSS, the cells were suspended in RPMI dose of 10mg / kg. 1640 (NIPRO, Osaka, Japan) supplemented with heat-inactivated 10% fetal bovine serum (FBS, GIBCO, Grand Island, NY, USA) Statistical analysis. Data were expressed as the mean±standard and 10 mM Hepes. This medium was used for the experiment. The deviation (SD). The statistical significance of the difference washed abdominal cavity cells from each mouse were counted using between two groups was determined using the Student's t-test. The a cell counter, and the survival rate was determined using statistical significance of the differences among three groups was the trypan blue method of exclusion. determined using an analysis of variance followed by Fisher's LSD test. The results were considered significantly different if p<0.05. Isolation of macrophages and induction of cytokine production. Macrophages were suspended in the medium at a density of 1x106/ml. Results The cells were plated in the flat-bottomed wells (100 Ìl/well) of 96-well culture plates (Iwaki Glass, Chiba, Japan) and cultured in a Cytological analysis of peritoneal cell suspensions. Cytological humidified, 5% CO2 atmosphere at 37ÆC. Cells were left to adhere for 2 hours, and non-adherent cells were removed by washing. analysis was performed on peritoneal cell suspensions. The Using latex minute-particle phagocytosis and Diff Quik stain, it number of peritoneal cells in the control group was 9.8±2.0 was found that more than 70% of the remaining monolayer x 106 cells/mouse (n=27), and in the EA (EA-15, PA-15) consisted of macrophages. After washing three times with HBSS, group was 10.3±1.6 x 106 (n=21). For the Mox (Mox-10, the macrophage monolayers were cultivated for 1 to 48 hours in Mox-2) group, the number of peritoneal cells was 9.2±0.5 x medium with 1 Ìg/ml of lipopolysaccharide (LPS, Escherichia coli 106 (n=21). There were no significant differences among 026 : B6, Sigma Chemical Co., St. Louis, MO, USA). the three groups. Following incubation, the culture plates were centrifuged and the cell-free culture supernatants were used to determine the The mean percentages of macrophages in the peritoneal cytokines that had been released. These samples were sterilized cells were 73±2.0 for the control group, 74±3.0 for the EA using a filter and were stored at –30ÆC until assayed. group and 71±3.0% for the Mox group. The infiltration of

496 Aoki et al: Acupuncture and Moxibustion Control LPS-induced Macrophage TNF·

Figure 2. Suppressive effects of EA-15 and Mox-10 on LPS-induced TNF· level in peritoneal infiltrated macrophages. Column and vertical lines expressed as percentage of control value at 4h indicate mean±SD for 6 mice. ***p<0.001: significant difference vs corresponding control.

cells into the peritoneal cavity during the immunization Following Mox (Mox-10), the TNF· level at 4 hours was process with the i.p. injection of the streptococcal about 55% of the control group (Figures 2, 3). While in preparation OK-432 was not affected by EA or Mox. Mox-2 treatment, the TNF· level at 4 hours was higher (about 80% of the control) than that of the Mox-10 group Effects of EA and Mox on TNF· applied at an acupuncture (Figure 3). It is shown that the dose of moxa used in the point corresponding to GV4. It was not possible to detect the Mox treatment closely correlated to the level of suppression TNF· level over time in the macrophage culture of the macrophage function. supernatant when LPS was not added. When 1 Ìg/ml of LPS was added, however, it became possible to detect TNF· Effect of naloxone on suppression of TNF· levels induced by approximately one hour later, with a maximum level of EA or Mox. The suppression of the TNF· production at 4 3,861±213 pg/106 macrophages being detected 4 hours after hours following EA-15 treatment was incompletely LPS exposure (Figure 1). antagonized by prior administration of a single i.p. injection The time-course of the level of TNF· induced by LPS in of naloxone (Figure 4). We also observed suppression of the the EA and the Mox groups approximated that of the level TNF· level change 4 hours after Mox treatment, but there induced by the LPS in the control group. The EA and Mox was no antagonism in response to prior injection of a single treatments did not affect the time-course of the TNF· level dose of naloxone (Figure 4). (Figure 2). The TNF· level at 4 hours in the EA-15 group, after Discussion treatment with a continuous wave duration level of EA500 Ìs, was about 52% that of the control group (Figures 2, 3). It The effects of EA and Mox with respect to the production was also found that the TNF· level of the group treated of TNF· in response to immunized peritoneal macrophages with a continuous wave duration of EA 50 Ìs (PA-15) was were investigated. The method used for this study was the suppressed, but it was significantly less than that of the two-step bacterial (OK-432 and LPS) stimulation method. EA-15 group (Figure 3). The wave duration of the EA The Streptococcus pyogenes detoxicated pharmaceutical treatment closely correlated with the suppression of the preparation OK-432 is an active Biological Response macrophage function. Modifier (BRM) (5).

497 in vivo 19: 495-500 (2005)

Figure 3. Suppressive effects of EA-15, PA-15, Mox-10 and Mox-2 on LPS-induced TNF· production in peritoneal infiltrated macrophages. Column and vertical lines expressed as percentage of control value at 4h Figure 4. Partially antagonizing effect of naloxone (10 mg/kg, i.p.inj.) to indicate mean±SD for 6 mice. ***p<0.001 vs 4h control. **p<0.01 vs suppressive effect of EA on TNF· production. Column and vertical lines 4h control. indicate mean±SD for 6 mice. ***p<0.001.

I.p. injections of OK-432 brought about an inflammatory No effects were observed after EA or Mox on the number microenvironment in the abdominal cavity. Cross-talk of the of cells infiltrating the abdominal cavity and on the neuro-endocrine-immune system is supposed to occur percentage of macrophages. The percentage of involving the peritoneal macrophages (7). Circulating macrophages was also confirmed using flow cytometric monocytes infiltrated the abdominal cavity in response to the analysis (data not shown). Therefore, we assumed that the i.p. injection of OK-432. The peritoneal macrophages were change in the TNF· production by the macrophages in the activated by OK-432 over a period of 24~72 hours. mice treated with EA and Mox, using a point equivalent to The abdominal cavity was infiltrated first by neutrocytes, GV4 in humans as a stimulus point, was not caused by a then by macrophages, and then by lymphocytes. We change in the number of macrophages, but rather was in collected peritoneal cells 48 hours subsequent to the response to a change in the TNF· production. injection, which was the point at which macrophage In our experiment, it was shown that the LPS-induced infiltration reached its peak. TNF· produced by the macrophages of mice treated with Next, macrophages were treated in vitro with LPS to EA-15 and Mox-10 decreased to about 50% of that of the induce a more strongly activated state in which pro- corresponding control group. We then considered the inflammatory immune mediators such as the cytokine, question of how the immunization process of the TNF·, were released in extremely large amounts. This macrophages was affected by EA and Mox treatment. We did experiment is appropriate for testing macrophage function, not examine the differences between various acupuncture since macrophages begin to synthesize cytokines such as points in this study; rather, we compared the degree of TNF· only when immunized macrophages are induced with stimulus at the acupuncture point, which was the equivalent LPS in vitro (5). to the GV4 point in humans. When EA was used, the wave In this experiment, we applied EA or Mox immediately duration was 500 Ìs (EA-15) and 50 Ìs (PA-15). With Mox, after i.p. injection of OK-432. Subsequently, LPS-induced we used 10 mg (Mox-10) and 2 mg (Mox-2) of moxa. TNF· was measured to see whether or not the In recent years, a greater understanding has been gained immunization process caused any inhibitory effects. If the of the theory of the communication networks of the neuro- EA or Mox modified the cross-talk of the neuro-endocrine- endocrine-immune system and how peripheral stimulatory immune system, one would expect a change in the cytokine procedures, such as EA and Mox, affect those production in the peritoneal macrophages. communication networks. The hypothesis of the action

498 Aoki et al: Acupuncture and Moxibustion Control LPS-induced Macrophage TNF· mechanisms of EA and Mox have been grouped and which catecholamines are produced through the HANS demonstrated under the collective name of "acupuncture". axis and the ‚-adrenergic mechanisms may be strongly Acupuncture (EA and Mox) is being used to treat various enhanced by EA and Mox. conditions apart from pain control (8-10). 3. Parasympathetic system and acetylcholine: The Various types of information taken in by living organisms autonomic nervous system, through the vagus nerve, can are controlled by the hypothalamus, and pass through the also inhibit the release of macrophage TNF. Acetylcholine hypothalamic-pituitary-adrenal (HPA) axis and hypothalamic- released from vagus nerve endings is bound to the nicotinic autonomic nervous system (HANS) axis to the endocrine and acetylcholine receptor ·7 subunit on the surface of neuronic tissues to maintain homeostasis (11, 12). macrophage and inhibits the release of TNF (19). Endorphinergic, monoaminergic, cholinergic transmitters and Consequently, it is possible that the mechanism by which endocrinological factors are involved in the course of that acetylcholine is released through the vagus nerve, as well as transmission route (10). the mechanism activating the nicotinic acetylcholine It still has not been clarified how the production of receptor, may be strongly influenced by EA and Mox. various inflammatory factors by the macrophages is 4. Endogenous opioids: Numerous reports have been controlled. With respect to TNF· production, when LPS published stating that endogenous opioids (endorphins) binds to CD14 and the Toll-like receptor 4 (TLR-4) of a and exogenous opioidergic work as biological macrophage, a signal is transmitted to the nucleus and then response modifiers (20). In this study, we demonstrated a transcription factor NFÎ B is activated. It is thought that that suppression of the TNF· level following EA-15 the activated NFÎ B induces gene expression so that treatment was antagonized to some extent by naloxone, inflammatory cytokines such as TNF· are produced (13, 14). although incompletely. It has also previously been There are various reports concerning endogenous factors reported that the EA-evoked delayed-type hyper- that control TNF· production. These theories are classified sensitivity suppression was blocked in a dose-dependent into four categories, in accordance with the hypothesis of manner by systemic pretreatment with naloxone, the action mechanisms of EA and Mox, as follows; indicating that opioid receptor-mediated mechanisms are 1. HPA axis and glucocorticoids: Glucocorticoids down- involved in this immune response (3). Therefore, the regulate the immune system and repress TNF· genetic mechanism by which endogenous opioids are produced, transcription activation (15). In experiments that we have as well as the opioid receptor-mediated mechanisms, may conducted on mice, a 200% increase in adrenocorticotropic be strongly enhanced by EA. hormone (ACTH) was observed in the plasma 15 minutes In summary, we investigated the process by which after the EA-15 treatment. Additionally, there was a 250% macrophage activation, using the streptococcal preparation increase in ACTH 30 to 60 minutes after Mox-10 treatment OK-432 as the immunizing agent, was modified after EA (unpublished data). or Mox treatment by examining the LPS-induced TNF· We earlier reported the results of a study comparing EA production in vitro. The neuro-endocrine-immune system and morphine in delayed type hypersensitivity in mice, has many feedback responses in the body as a whole. EA showing that the HPA axis became involved based on the and Mox are thought to modify these feedback responses. timing of the biological response of the suppression of For both EA and Mox, when treatment was delayed type hypersensitivity resulting from EA-15 administered at an acupuncture point corresponding to treatment (3, 16). Consequently, there may be a possibility the GV4 point in humans, macrophages were influenced that the production of glucocorticoids through the HPA axis in a way that the LPS reactivity was suppressed by 50%. may be strongly enhanced by EA and Mox. The production of TNF· is thought to be suppressed 2. HPAS axis and catecholamines: Numerous reports when the degree of macrophage activation is suppressed have been published concerning the endogenous by 50%. However, this does not necessarily mean that the catecholamines that control TNF· production. It is body’s defenses against infection are suppressed thought that the sympathetic nervous system modulates completely, as they are, for instance, when an anti-TNF· immune response. It has been reported that adrenaline antibody reagent is used for treatment. Therefore, side- regulates the TNF· production in sepsis (17). LPS- effects such as increased susceptibility to infection are induced TNF· production in vivo is said to be mediated not likely to occur. by the release of endogenous catecholamines. In that With respect to morbid conditions in which chronic TNF· process, ‚-adrenoceptors may act as the suppressor, while production is seen, meaning diseases such as rheumatoid ·-adrenoceptors may act as the enhancer (18). The arthritis and auto-allergic diseases, treatment using EA or ligands of ‚-adrenoceptors increase intracellular cAMP Mox is a means of achieving normalization of the homeostasis levels, while the ligands of ·-2 adrenoceptors decrease maintenance mechanism(21, 22). Further research is necessary them. Therefore, it is conceivable that the mechanism by to elucidate the mechanisms of action of EA and Mox.

499 in vivo 19: 495-500 (2005)

References 13 Hambleton J, MacMahon M and DeFranco AL: Activation of raf-1 and mitogen-activated protein kinase in murine 1 Kasahara T, Y, Wang Y, Sakurai Y and Oguchi K: A macrophages partially mimics lipopolysaccharide-induced pharmacological approach to oriental medicine; a comparative signaling events. J Exp Med 182: 147-154, 1995. study of the effect of moxibustion, electroacupuncture and 14 Sweet MJ and Hume DA: Endotoxin signal transduction in -hozai ohrengedoku-toh on delayed type hypersensitivity macrophages. J Leukoc Biol 60: 8-26, 1996. in mice. Showa Med Ass 49: 488-495, 1990. 15 Auphan N, DiDonato J A, Rosette C, Helmberg A and Karin 2 Kasahara T, Wu Y, Wang Y, Sakurai Y and Oguchi K: M: Immunosuppression by glucocorticoids: inhibition of NF-Î B Suppressive effect of acupuncture on delayed type activity through induction of IÎ B synthesis. Science 270: 286- hypersensitivity to trinitrochlorobenzene and involvement of 290, 1995. opiate receptors. Int J Immunopharmac 14: 661-665, 1992. 16 Kasahara T, Sakurai Y, Amemiya M, Oguchi K and 3 Kasahara T, Wu Y, Wang Y, Sakurai Y and Oguchi K: Hisamitsu T: Suppressive effects of central opioids on delayed Involvement of central opioidergic and nonopioidergic type hypersensitivity to trinitrochlorobenzene: comparative neuroendocrine systems in the suppressive effect of study with morphine and electroacupuncture. In Vivo 9: 177- acupuncture on delayed type hypersensitivity in mice. Int J 182, 1995. Immunopharmac 15: 501-508, 1993. 17 Severn A, Rapson NT, Hunter CA and Liew FY: Regulation of 4 Kasahara T, Wu Y, Wang Y, Sakurai Y and Oguchi K: tumor necrosis factor production by adrenaline and ‚- Modulation of lipopolysaccharide-induced cytotoxic factor and adrenergic agonists. J Immunol 148: 3441-3445, 1992. interferon production by moxibustion in mice. In Vivo 4: 289- 18 Elenkov IJ, Hasco G, Kovacs KJ and Vizi ES: Modulation of 292, 1990. lipopolysaccharide-induced tumor necrosis factor · production 5 Yamamoto A, Naganuma M, Usami H, Sugawara Y, Watanabr by selective ·- and ‚-adrenergic drugs in mice. J Neuroimmun N, Niitu Y and Urushizaki I: Release of tumor necrosis factor 61: 123-131, 1995. (TNF) into mouse peritoneal fluids by OK-432, a streptococcal 19 Wang H, Yu M, Ochani M, Amella CA, Tanovic M, Susarla S, preparation. Immunopharmacology 11: 79-86, 1986. Li JH, Wang H, Yang H, Ulloa L, Al-Abed Y, Czura CJ and 6 Kawashima T and Arai S: Laboratory Methods in Lymphocyte Tracey KJ: Nicotinic acetylcholine receptor ·7 subunit is an Studies (Yata J and Fujiwara M, eds). Chugai Igaku Co., pp. essential regulator of inflammation. Nature 421: 384-388, 2003. 665-671, 1994. 20 Carr D JJ and Serou M: Exogenous and endogenous opioids as 7 Besedovsky HO, delRey AE and Sorkin E: What do the biological response modifiers. Immunopharmacology 31: 59-71, immune system and brain know about each other? Immunol 1995. Today 4: 342-346, 1983. 21 Fang J-Q, Aoki E, Seto A, Yu Y, Kasahara T and Hisamitsu T: 8 Sato A, Sato Y and Uchida S: Reflex modulation of visceral Influence of moxibustion on collagen-induced arthritis in mice. functions by acupuncture-like stimulation in anesthetized rats. In Vivo 12: 421-426, 1998. In: Acupuncture –Is There a Physiological Basis? (Sato A, Li P 22 Fang J-Q, Aoki E, Yu Y, Sohma T, Kasahara T and Hisamitsu and Campbell J L, eds). Elsevier, Tokyo, pp. 111-124, 2002. T: inhibitory effect of electroacupuncture on murine collagen 9 Hisamitsu T, Kasahara T, Umezawa T, Ishino T and Hisamitsu arthritis and its possible mechanisms. In Vivo 13: 311-318,1999. N: In: Acupuncture – Is There a Physiological Basis? (Sato A, Li P and Campbell J L, eds). Elsevier, Tokyo, pp. 125-131, 2002. 10 Cho ZH, Wong EK and Fallon JH: Hypotheses of acupuncture mechanisms. In: Neuro-acupuncture, Volume 1: Q-puncture, Inc., CA, USA. pp.121-131, 2001. 11 Chrousos GP: The hypothalamic-pituitary-adrenal axis and immune-mediated inflammation. N Engl J Med 332: 1351- 1362, 1995. 12 Straub RH and Cutolo M: Involvement of the hypothalamic- pituitary-adrenal/gonadal axis and the peripheral nervous Received December 30, 2004 system in rheumatoid arthritis. Arthritis Rheum 44: 493-507, Revised March 18, 2005 2001. Accepted March 22, 2005

500