The Endocannabinoid System: An Osteopathic Perspective John M. McPartland, DO

 

 

This is an informative Article by JAOA
(Journal of the American Osteopathic Association)
October 2008 Exerts follow

The present review provides an update on endocannabinoid
basic science and clinical studies and proposes a new
model to describe reciprocal interactions between somatic
dysfunction and the endocannabinoid system. The endocannabinoid
system consists of cannabinoid receptors,
endogenous ligands, and ligand-metabolizing enzymes. The
system exemplifies the osteopathic principle that the body
possesses self-regulatory mechanisms that are self-healing
in nature. Enhancing endocannabinoid activity has broad
therapeutic potential, including the treatment of patients
with somatic dysfunction, chronic pain, and neurodegenerative
diseases as well as inflammatory conditions, bowel
dysfunctions, and psychological disorders. Blockade of the
endocannabinoid system with drugs such as rimonabant
and taranabant may oppose self-healing mechanisms and
elicit adverse effects. Osteopathic physicians wield several
tools that can augment endocannabinoid activity, including
lifestyle modifications, pharmaceutical approaches, and
osteopathic manipulative treatment.
J Am Osteopath Assoc. 2008;108:586-600
The endocannabinoid system was discovered long after the
endorphin system, which was indirectly detected in 1801
when morphine sulfate was isolated from opium. Morphine’s
mechanism of action remained a mystery until the opioid 
receptor was identified. That discovery begged the question:
Why do humans express a receptor for an opium poppy
(Papavera somniferum) plant compound? Scientists quickly
identified endorphins and enkephalins, which are endogenous
compounds mimicked by the plant compound.1
In 1897, Andrew Taylor Still, MD, DO,2 the founder of
osteopathic medicine, famously stated, “Man should study
and use the drugs compounded in his own body.” Still
hypothesized that osteopathic manipulative treatment (OMT)
stimulated endogenous compounds that promoted homeo –
stasis and healing. Not long after the discovery of endogenous
opioids in 1975, JAOA—The Journal of the American Osteopathic
Association published a supplement dedicated to endorphins
and enkephalins.3 However, the initial enthusiasm
dampened after seven subsequent studies4,5 showed no effects
of OMT or chiropractic manipulation on serum levels of
these compounds.
Since then, research—particularly osteopathic medical
research—has redirected its attention from the endorphin
system to the endocannabinoid system.4-10 A search on the
National Library of Medicine’s PubMed database of endorphin
in the 1992 literature, the year endocannabinoids were discovered,
returns 596 citations, whereas endocannabinoid yields
only two citations. In a search limited to 2007, endorphin produces
122 citations, whereas endocannabinoid generates 480 citations.
The primary purpose of the current article is to review
the expanding endocannabinoid literature beginning with
exogenous compounds—Cannabis and plant cannabinoids—
and then shift to the endogenous system, highlighting embryology
and development, neuroprotection, autonomics and
immunity, inflammation, apoptosis, hunger and feeding,
and nociception and pain.
Because the literature is so voluminous—more than
10,000 cannabinoid citations in PubMed—the present article
is not a systematic review. Review articles were considered
the preferred source and are referenced throughout to enable
continued education.
In addition, the present review seeks to draw parallels
between the endocannabinoid system and the principles of
osteopathic medicine.11 The endocannabinoid system is a
homeostatic mechanism that exemplifies the key osteopathic
concepts of mind-body unity11 on a molecular level. A new
neuroimmunologic model describes reciprocal interactions
between the endocannabinoid system and the osteopathic
concept of somatic dysfunction. Finally, the current review
will describe how osteopathic physicians may enhance endocannabinoid
function in their patients.
The Endocannabinoid System: An Osteopathic Perspective
John M. McPartland, DO
From the Department of Osteopathic Manipulative Medicine at Michigan State
University College of Osteopathic Medicine in East Lansing.
This work began under the aegis of a Unitec AssocProf research grant
(School of Osteopathy, Unitec New Zealand). A lecture based on this
manuscript was presented at the 2007 meeting of the American Academy of
Osteopathy in Birmingham, Ala, on March 27, 2006. The author serves as a
scientific advisor for the Cannabinoid Research Institute, a research division
of GW Pharmaceuticals.
Address correspondence to John M. McPartland, DO, 53 Washington St
Ext, Middlebury, VT 05753-1288.
E-mail: mcpruitt@verizon.net
Submitted June 6, 2007; final revision received February 20, 2008; accepted
March 3, 2008.
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
JAOA • Vol 108 • No 10 • October 2008 • 587
receptor agonist ([3H]CP55,940) bound specifically to a receptor
located in neuron cell membranes.21 Two years later, Matsuda
and Bonner cloned the gene for the receptor, which translated
into a chain of 472 amino acids that weave back and
forth across the cell membrane seven times.21 This weaving
structure is the same as those of G protein–coupled receptors
(GPCRs), which are named for their G (guanine nucleotide
binding) proteins and function as intracellular “molecular
switches.” Well-known GPCRs include opioid, dopamine,
serotonin, and -adrenergic receptors.
Each GPCR possesses a unique binding pocket with an
affinity for specific ligands, like a lock-and-key mechanism.
When a ligand docks in the binding pocket, it may initiate
any of the following changes, depending on the agonist:
 Full agonists—These agonists, such as morphine and -
endorphin, maximally activate receptors.
 Partial agonists—These agonists submaximally activate
receptors, often less than an endogenous ligand. For example,
the -blocker pindolol occupies the norepinephrine binding
site but exerts much less activity.
 Neutral antagonists—Agonists in this category, such as
naloxone hydrochloride and alprenolol, dock at receptors but
do not activate them.
 Inverse agonists—These agonists, such as metoprolol succinate,
metoprolol tartrate, prazosin hydrochloride, cimetidine
and haloperidol deactivate receptors by suppressing
their spontaneous activity.
A ligand docking in a GPCR distorts the shape of the
GPCR’s transmembrane weave of amino acids (a “conformational
change”), thereby altering the intracellular side of the
receptor and its interface with the G protein. If the ligand is an
agonist, the G protein decouples from the receptor and couples
to an ion channel (eg, K, Ca2) or enzymes (eg, adenylate
cyclase), causing a ‘’signal cascade’’ that governs cell behavior.21
The cannabinoid receptor can activate different G protein
subtypes. For example, subtype Go couples to ion channels,
Gi inhibits adenylate cyclase, and Gs stimulates adenylate
cyclase.21 The deciding factor is the agonist because various
agonists preferentially direct the receptor toward the different
G protein subtypes.21 This “agonist trafficking” may explain
why different strains of cannabis produce different psychoactive
effects.22 Agonist trafficking alters the traditional
lock-and-key metaphor. In this setting, an assortment of keys
unlock the same lock, but the door opens into different rooms.
Cannabinoid receptors are the most common GPCRs in
the brain, but they are unevenly distributed. High densities are
found in the basal ganglia, which is composed of the globus
pallidus, substantia nigra, and striatum (comprising the caudate
nucleus and putamen); hippocampus; cerebral cortex;
cerebellum; and amygdaloid nucleus.23 Receptor distribution
accounts for the well-known effects of cannabis on short-term
memory, cognition, mood and emotion, motor function, and
Exogenous Cannabinoids
The plant Cannabis sativa, the source of cannabis (ie, marijuana,
hashish), is native to central Asia. Several indigenous
“traditional medicine” systems (eg, Ayurvedic medicine,
Tibetan medicine, traditional Chinese medicine) evolved and
make use of this substance.12
In the 1830s, a physician serving the British Crown in
India conducted a series of laboratory experiments, including
animal studies and clinical trials, to determine the safety and
efficacy of cannabis. He subsequently introduced cannabis to
England and was knighted by Queen Victoria.13
Cannabis entered The Dispensatory of the United States of
America in 1854.14 Although cannabis met a variety of applications,
its primary use was for the alleviation of pain and
spasticity. In fact, Sir William Osler considered cannabis the
“most satisfactory remedy” for migraine.15
In the late 19th and early 20th centuries, cannabis was
manufactured by a number of pharmaceutical companies in the
United States and was dispensed as an orally administered
fluid extract. When coupled with variable product potency,
unreliable sources of supply, poor storage stability, and evidence
that the liquid form was erratically absorbed by the
gut, fluid extracts soon fell out of favor.13 Its decline in popularity
was further hastened by new and inexpensive synthetic
medicines such as aspirin. Finally, after growing concern of
“reefer madness,” cannabis was prohibited in 1937 despite
vigorous opposition by the American Medical Association.16
In fact, many leaders in the allopathic medical profession continue
to support the medicinal use of cannabis.17
Whereas morphine forms a water-soluble salt easily isolated
from Papavera somniferum, the active ingredients in
cannabis are lipophilic and resist crystallization. In 1964, after
foiling scientists for 150 years, Raphael Mechoulam isolated 9-
tetrahydrocannabinol (THC) and cannabidiol.12 More than 70
separate 21-carbon terpenophenols unique to cannabis, collectively
called the cannabinoids, have been identified by
Mechoulam,12 Pertwee,18 and others.
Animal studies of THC began immediately after its discovery.
By 1975, the first phase 3 clinical trials were published.
12,18 Dronabinol, a synthetic THC, was approved as a
schedule II drug in 1986 and was moved to schedule III in
1999.17 Its indications include nausea and vomiting associated
with cancer chemotherapy as well as appetite and weight loss
in patients with AIDS.19 Nabilone, a THC analog with the
same indications, was approved by the US Food and Drug
Administration in 1985 but was not marketed in the United
States until 2006.20
Receptors and Signal Transduction
Discovery of the -receptor launched a search for cannabinoid
receptors. However, the search was stymied by THC
because of its nonspecific (ie, indiscriminate) binding.
In 1988, Howlett et al developed synthetic, water-soluble
THC analogs showing that a radiolabelled cannabinoid
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
588 • JAOA • Vol 108 • No 10 • October 2008
nociception. Unlike -receptors, cannabinoid receptors are
virtually absent in brainstem cardiorespiratory centers, which
probably account for the lack of lethal effects from cannabis
overdose.23
Cannabinoid receptors are downregulated and desensitized
when exposed to high doses of THC. This effect results
in “drug tolerance,” which occurs at varying rates and magnitudes
in different brain regions. For example, it occurs faster
and greater in the hippocampus, which regulates memory,
compared with the basal ganglia, which mediates euphoric
effect.24 This difference may explain why memory loss
decreases among frequent cannabis users, but its euphoric
effects remain.25
Curiously, cannabinoid receptors may activate G proteins
in the absence of THC or endogenous cannabinoid compounds.
21 This “constitutive activity” is difficult to separate
from “endogenous tone” experimentally, where tonic release
of endocannabinoids activates the receptors. Constitutive
activity has also been measured in other GPCRs, such as the
angiotensin receptor, which is stretched into an active conformation
by hydrostatic pressure in blood vessels.26
A second cannabinoid receptor, CB2, was discovered in
1993. While the first receptor, CB1, is principally located in the
nervous system, CB2 is primarily associated with cells governing
immune function: leukocytes, splenocytes, and
microglia.21 The genes for CB1 and CB2 are paralogs (ie, genes
separated by a gene-duplication event), with orthologs (ie,
genes separated by speciational events) in all vertebrate species
investigated to date.27
In fact, orthologs of cannabinoid receptors have been
identified in primitive organisms, such as nematodes and sea
squirts, which suggests cannabinoid receptors evolved 600 million
years ago.28 Therefore, from an evolutionary perspective,
human CB1 is under strong purifying selection, whereas CB2
is under reduced functional restraint. In fact, the mutation
rate of CB2 is four times higher than CB1.29
Endogenous Cannabinoids
Humans likely did not evolve receptors for a Cannabis compound.
Indeed, the cannabinoid receptor evolved long before
cannabis, which is not more than 34 million years old.22 The first
endogenous cannabinoid, anandamide (AEA), was discovered
by Mechoulam in 1992—nearly 30 years after he discovered
THC. The discovery of 2-arachidonoylglycerol (2-AG)
occurred shortly thereafter.
Both the AEA and 2-AG endocannabinoids are metabolites
of arachidonic acid. They do not resemble THC but
nonetheless fit the CB1 and CB2 binding pockets. Therefore, the
effects of THC, AEA, and 2-AG substantially overlap, activating
the same receptors.30 However, THC is a partial agonist
and may block 2-AG’s full agonist activity in some situations.
31-33
Unlike classic neurotransmitters, AEA and 2-AG are not
stored in vesicles. Instead, they are synthesized and released
“on demand” from precursor phospholipids within the cell
membrane. The AEA endocannabinoid is cleaved from its
precursor phospholipid by the enzymes N-acyl phosphatidylethanolamine
phospholipase D and alpha-beta hydrolase
4, while 2-AG is cleaved from its precursor diacylglycerol
(DAG) by two DAG lipase enzymes, DAGL and DAGL.
After release into the synapse, AEA and 2-AG activate CB1.
Thereafter, several other catalytic enzymes break down AEA
and 2-AG.18,34 Several agents that block catalytic enzymes—
specifically fatty acid amide hydrolase (FAAH) and monoacylglycerol
lipase (MAGL)—have been described, which prolong
AEA and 2-AG synaptic activity, analogous to a serotonin
uptake inhibitor. Pharmacologists are searching for inhibitors
of the other endocannabinoid enzymes as well.18 Although
few endogenous inverse agonists are known for any receptors,
one inverse agonist called hemopressin was discovered for CB1
and has exhibited surprising analgesic properties.35
Within the central nervous system, the endocannabinoid
system acts as a negative feedback mechanism to dampen
synaptic release of classic neurotransmitters. A simplified
example is presented in Figure 1. Persistent activation of a
nerve—in this case, a sensory C-fiber nociceptor—causes excessive
release of glutamate from its central terminal, which
synapses in the dorsal horn. Excessive glutamate causes an
upregulation of glutamate receptors in the postsynaptic cell (in
this case, a wide dynamic range neuron). Persistent nociception
and upregulated glutamate receptors lead to a form of
neural plasticity known as central sensitization.36,37 Neural plasticity
is characterized by the sprouting and pruning of synapses,
changes in dendritic spine density, and changes in neurotransmitter
pathways. It gives rise to all types of adaptive
learning, including the subcortical events leading to a facilitated
spinal segment37 as well as the conscious act of gaining a new
skill or the unconscious acquisition of a new emotional
response.38
Central sensitization elicits a homeostatic response by the
endocannabinoid system: upregulated glutamate receptors in
the post-synaptic cell lead to an influx of Ca2 (Figure 1A),
which causes DAGL enzymes in the post-synaptic cell to
synthesize 2-AG (Figure 1B). The 2-AG endocannabinoid moves
retrograde (ie, opposite the direction of glutamate) across the
synapse to CB1, located on the presynaptic neuron. The activated
CB1 closes presynaptic Ca2 channels, which then halts
glutamate vesicle release.31 This “retrograde signaling” mechanism
is termed depolarization-induced suppression of excitation
(DSE) and enables the postsynaptic cell to control its own
incoming synaptic traffic.
Alternatively, if endocannabinoids transiently attenuate
the release of an inhibitory neurotransmitter such as -
aminobutyric acid, the mechanism is termed depolarizationinduced
suppression of inhibition (DSI).31 As a ubiquitous phenomenon,
DSI modulates neurotransmission in the
hippocampus, cerebellum, basal ganglia, cerebral cortex, and
amygdaloid nucleus.31
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
JAOA • Vol 108 • No 10 • October 2008 • 589
tion and synaptogenesis. Lastly, axon guidance is shaped by
axon growth cones, and endocannabinoids are part of the
molecular “soup” that guides growth cones to their destinations.
43
Adult neurogenesis is regulated by many of these “embryonic”
mechanisms and primarily arises in neural stem cells
within the subependymal layer lining the cerebral ventricles
and the dentate gyrus of the hippocampus. Neural stem cells
in both of these brain regions express CB1,44 and neurogenesis
by these cells is driven by the endocannabinoid system.45
Considering the prominence of the endocannabinoid
system in embryogenesis, its equal importance in adult neurogenesis
brings to mind a quote by James Jealous, DO:
The formative and regenerative forces that organize embryological
development are present throughout our life span […].
In other words, the forces of embryogenesis become the forces
of healing after birth.46
Although DSE is less common and the data suggest CB1
occurs both pre- and postsynaptically in the dorsal horn,36
evidence indicates that DSE dampens nociception at the spinal
level.39 The endocannabinoid system also controls other forms
of neural plasticity such as long-term depression, which is
caused by a sustained decrease in glutamate release from
presynaptic cells.40
Embryology and Development
For endocannabinoid receptors to survive 600 million years,
they must serve evolutionarily important functions. The CB1
receptors have been detected in mouse embryos as early as the
second day of gestation.41 Blastocysts express CB1, CB2, and
FAAH, and blastocyst implantation into the endometrium
requires suitable levels of AEA.41 In fact, the endocannabinoid
system organizes a broad array of developmental processes
in the embryonic brain. Proliferation and differentiation
of neural stem cells are shaped by extracellular cues provided
by endocannabinoids.42 Indeed, once stem cells commit to
neuronogenesis, endocannabinoids regulate neuronal migra-
McPartland • Review Article
REVIEW ARTICLE
A B
Figure 1. Retrograde transmission at the dorsal horn. Nociceptor action potentials open calcium (Ca2) channels in the presynaptic axon terminal,
which cause vesicles of glutamate to release into the synaptic cleft. (A) Excessive glutamate release causes upregulation of glutamate
receptors in the postsynaptic cell, which open Ca2 channels. (B) Postsynaptic calcium influx stimulates diacylglycerol lipase enzymes to synthesize
2-AG, which diffuses across the synapse to the presynaptic cell and activates CB1, which closes presynaptic Ca2 channels, arresting the
release of glutamate vesicles. Abbreviations: AMPA, alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor; eCBS, endocannabinoid
system; mGLU, metabotropic glutamate receptor; NMDA, N-methyl-D-aspartic acid receptor. Reprinted with permission from the
Journal of Bodywork and Movement Therapies.10 Copyright Elsevier, 2008.
Downloaded From: http://jaoa.org/ on 07/19/2017
590 • JAOA • Vol 108 • No 10 • October 2008
Neuroprotection
The anti-inflammatory, antioxidant, and antispasmodic properties
of THC and cannabidiol in cannabis reduce the symptoms
and slow the progression of multiple sclerosis. These
properties also benefit patients with Huntington disease and
amyotrophic lateral sclerosis.47 Dampening of glutamate excitotoxicity
prevents epileptic attacks and limits infarct size poststroke.
Both AEA and 2-AG thwart Alzheimer disease by
blocking microglial activation and -amyloid plaque formation.
They also prevent Parkinson disease symptoms by rebalancing
neural activity in the striatum.47
Although controversial, the endocannabinoid system has
been associated with psychotic disorders. Individuals with
schizophrenia have elevated levels of AEA in their cerebrospinal
fluid, but the elevated levels are negatively correlated
with psychotic symptoms.48 This association suggests
that abnormal activation of postsynaptic D2 receptors trigger
the release of AEA and retrograde signaling via CB1, thus
homeo statically attenuating dopamine release. High doses of
THC may therefore provoke psychiatric illness in susceptible
individuals by desensitizing CB1 receptors and diminishing retrograde
signaling.48 Alternatively, cannabidiol, which is not
psychoactive, shows promise as an antipsychotic agent.49
Autonomic Function and Immunity
Endocannabinoids and THC affect autonomic outflow through
the peripheral and central nervous systems. The endocannabinoid
system reduces elevated parasympathetic activity,
providing the antiemetic effects of cannabinoids.18
In rodent studies, activation of myocardial CB1 caused
vagally mediated biphasic effects in heart rate and cardiac
contractility, while activation of CB1 in vascular tissues leads
to vasodilation.50
Cannabinoids provide antihypertensive benefits in
humans and a protective role in myocardial ischemia has been
suggested in rodent studies.50 However, human in vitro studies
have implicated the autonomic effects of THC and endocannabinoids
in hypotension associated with hemorrhagic
and endotoxic shock. The autonomic effects have also been
described in human clinical studies on advanced liver cirrhosis.
50
Sympathetic nerve terminals contain CB1, and activation
of these receptors has been shown to inhibit norepinephrine
release and dampen sympathetically mediated pain.50 The
endocannabinoid system modulates the sympathetically driven
hypothalamic-pituitary-adrenocortical (HPA) axis as well as the
hypothalamic-locus coeruleus-norepinephrine (HLN) axis.
Psychologic stress induces the secretion of corticotropinreleasing
hormone, which activates the HPA and HLN axes
and results in corticosteroid and norepinephrine release, respectively.
The endocannabinoid system blocks HPA and HLN
axis activation, though doses of THC may cause the reverse
response and increase corticotropin-releasing hormone and
anxiety.50
In addition, HPA axis activation hinders the immune
response. Cannabinoids are immunomodulators—not simply
immunosuppressors, as they were characterized in the
1970s.51,52 Cannabinoids suppress production of T-helper 1
(TH1) cytokines such as interleukin (IL) 2, immune interferon
(INF-), and tumor necrosis factor  (TNF-). On the other
hand, cannabinoids increase secretion of TH2 cytokines (eg, IL-
4, IL-5, IL-10). Other subsets of lymphocytes, including B cells
(eg, MZ, B1a) and natural killer cells, require endocannabinoids
and CB2 to function properly.51,52 Cannabis, Echinacea, and
other plant products that stimulate resistance to infection and
fatigue have been described as “adaptogens”—natural products
that work “osteopathically” by enhancing health rather
than fighting disease.53 The alkylamides in Echinacea potently
agonize CB2 and stimulate phagocytosis.54 The lack of psychoactivity
caused by Echinacea can be attributed to the relative
lack of CB2 in the brain.
Inflammation and Connective Tissues
More than 4000 years ago, the Chinese physician Shen Nung recommended
cannabis for rheumatic pains.12 More recently,
patients with myofascial pain and arthritis are among those
who most often use cannabis medicinally.13 Activation of CB2
suppresses proinflammatory cytokines such as IL-1 and
TNF- while increasing anti-inflammatory cytokines such as
IL-4 and IL-10.51,52 Although THC has well-known anti-inflammatory
properties, cannabidiol also provides clinical improvement
in arthritis via a cannabinoid receptor–independent mechanism.
18 Many connective tissue–related cells express CB1, CB2,
and endocannabinoid-metabolizing enzymes such as fibroblasts,
myofibroblasts, chondrocytes, and synoviocytes.10
The endocannabinoid system alters fibroblast “focal adhesions,”
by which fibroblasts link the extracellular collegen
matrix to their intracellular cytoskeleton—the mechanism of
fascial remodeling. Cannabinoids prevent cartilage destruction
such as proteoglycan degradation and collagen breakdown by
inhibiting chrondrocyte expression of cytokines and metalloproteinase
enzymes.10 In addition, the tonic release of endocannabinoids
is upregulated in rats that have experimentally
induced osteoarthritis, thereby providing endogenous pain
relief.55
The endocannabinoid system has been shown to attenuate
allergic contact dermatitis in rodent studies—THC decreases
allergic inflammation whereas CB1-blocking agents exacerbate
the condition.56 The endocannabinoid system likewise
protects against Crohn disease—a TH1-mediated inflammatory
bowel condition—and also perhaps ulcerative colitis.6,51,52
The endocannabinoid system dampens the inflammatory component
of atherosclerosis in animal studies via CB2 receptors
expressed by macrophages within atherosclerotic plaques.18
Lastly, the endocannabinoid system is essential for the maintenance
of normal bone mass: CB2 agonists enhance osteoblast
activity and inhibit osteoclast activity, therefore offering a
potential treatment option for osteoporosis.57
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
JAOA • Vol 108 • No 10 • October 2008 • 591
phates), leukocytes (eg, histamines, prostaglandins,
leukotrienes, proinflammatory cytokines), leukocyte-activated
platelets (eg, 5-hydroxytryptamine), neighboring autonomic
nerves (eg, norepinephrine), and the nociceptor itself (ie, substance
P and calcitonin gene-related peptide). Activators and
sensitizers cause peripheral sensitization, including hyperalgesia
and allodynia. Peripheral sensitization elicits a homeostatic
response by the endocannabinoid system: CB1 signaling
decreases the release of activators and sensitizers around the
site of tissue injury and opens K channels in the nociceptor
cell membrane, so the nerve becomes hyperpolarized and less
likely to fire.36
In addition, CB2 signaling decreases the release of activators
and sensitizers from neighboring mast cells and
macrophages (Figure 2).36 Functioning of the endocannabinoid
system at the peripheral terminal of the nociceptor provides
the “first line of defense against pain.”60
In the dorsal horn, the nociceptor synapses with a nociceptive-
specific neuron or a wide dynamic range neuron. Normally,
the nociceptor action potential arrives at the dorsal
horn and releases glutamate and substance P into the synaptic
cleft. These neurotransmitters bind to their respective receptors
in the postsynaptic cell. Cell activation initiates another action
potential that ascends to the brain. Abnormal persistent release
of glutamate upregulates glutamate receptors in the postsynaptic
cell, as illustrated in Figure 1. As previously described,
this causes an influx of Ca2 in the post-synaptic cell and leads
to central sensitization, “wind-up,” or “dorsal horn memory.”37
However, Ca2 influx elicits endocannabinoid synthesis,
followed by retrograde signaling, and quickly shuts down
presynaptic glutamate release. In other words, the endocannabinoid
system induces “dorsal horn memory loss” and
Apoptosis
With the exception of cannabis smoke, cannabinoids are anticarcinogenic.
They have been found to induce apoptosis in
cancer cells via a CB1-mediated ceramide-caspase pathway,
thereby inhibiting tumor growth in breast, prostate, and lung
carcinomas as well as gliomas, melanomas, lymphomas, and
other cancers.58 In normal, nontransformed cells, endocannabinoids
actually promote cell survival via the extracellular
signal-regulated kinase (ERK) pathway.58
Reducing apoptosis in normal cells is one mechanism by
which cannabinoids act as neuroprotectants.47 Cannabinoids
also suppress tumor angiogenesis.58 However, one contrary
study59 reported that the CB2 gene serves as a retrovirus insertion
site (ie, a proto-oncogene) involved in leukemic transformation.
Nociception and Pain
Research on the effects of endocannabinoids on nociception has
focused on the following four areas:
▫ peripheral terminals of nociceptors
▫ the dorsal horn
▫ the descending pain inhibitory pathway
▫ supratentorial sites
Peripheral terminals of C-fiber nociceptors contain receptors
for “activators” and “sensitizers” (Figure 2). Activators
trigger an action potential in the nerve while sensitizers
decrease the nerve’s activation threshold, so it fires with less
activation.
Activators and sensitizers are released from damaged
tissue (eg, K and H ions, bradykinins, adenosine triphos-
McPartland • Review Article
REVIEW ARTICLE
Figure 2. Polymodal C-fiber nociceptor
with an enlarged view of its distal terminal,
its cell body in the dorsal root
ganglion, and central terminal in the
dorsal horn. Several nociceptor activators
(Roman) and sensitizers (italic) are
illustrated with their corresponding
receptors named by gene symbols. The
distal terminal also expresses CB1 and
two ion channels (G protein–coupled
Kir3 [GIRK] and sensory neuron sodium
[Nav 1.8]) regulated by the receptor. A
sympathetic postganglionic neuron and
lymphocyte expressing CB2 are located
nearby the distal terminal. Abbreviations:
5-HT, 5-hydroxytryptamine; ATP,
adenosine triphosphate; NGF, nerve
growth factor. Reprinted with permission
from the Journal of Bodywork and
Movement Therapies.10 Copyright Elsevier,
2008.
Downloaded From: http://jaoa.org/ on 07/19/2017
592 • JAOA • Vol 108 • No 10 • October 2008
short-circuits central sensitization.36
The “descending pain inhibitory pathway” in rodents
and humans is activated by the perception of pain in the brain.
The pathway descends through the periaqueductal gray (PAG)
and periventricular gray in the midbrain to the nucleus raphe
magnus in the rostroventral medulla, and down to the dorsal
horn of the spinal cord.
Endorphins, endocannabinoids, serotonin, norepinephrine,
and adenosine play important roles in this pathway. Endocannabinoids
and CB1 are found in high concentration in the
PAG, periventricular gray, nucleus raphe magnus, and dorsal
horn, where they suppress GABA-releasing interneurons that
inhibit neurons in the descending pathway.36
The coordinated release of endocannabinoids in this
pathway mediates a rodent model of “stress-induced analgesia,”
61 the well-known phenomenon in which individuals are
less responsive to pain immediately after encountering an
environmental stressor (eg, soldiers wounded in action may not
feel pain during the battle). The endocannabinoid and endorphin
systems colocalize within the descending pathway. This
process indicates how endocannabinoids and THC can work
synergistically with morphine to provide a “morphine-sparing
effect.”36 A rodent study62 also showed that activated CB2
receptors stimulated the release of -endorphins.
Nociceptive input into supratentorial sites, such as the
cortex and limbic structures, registers as pain and suffering.
Rodent and human studies show that endocannabinoids
squelch amygdala-based aversive memories and fear conditioning.
63,64 Painful memories, fear, and anxiety are factors
that turn chronic pain into chronic suffering. Thus, the endocannabinoid
system may benefit hospice patients and those
unable to extinguish painful memory (eg, patients with posttraumatic
stress disorder).63
Defacilitation of the amygdyla also rebalances the autonomic
system and boosts “off-cell” activity in descending pain
inhibitory pathways in rodents.65 Cannabis imparts supratentorial
“cannabimimetic” effects, such as anxiolysis, alleviation
of suffering, increased sense of well-being, and even
euphoria.66 Although AEA and 2-AG have not been injected
into human subjects, the augmentation of endocannabinoids
using OMT has caused them to feel “high, happy, light-headed,
and hungry.”5 Similarly, exercise-induced “runner’s high”
correlated with augmented levels of serum AEA.67
Hunger and Feeding
Marijuana-enhanced hunger and feeding (“the munchies”) is
a behavior that teleologically begins in utero. Blastocyst implantation
has been characterized as organisms’ first suckling function—
the blastocyst actively orients itself so it implants “head
first” into the endometrium, followed by its uptake of nutrients68—
and blastocyst implantation requires a functional endocannabinoid
system.41 At the other end of gestation, newborn
mice given rimonabant, a drug that blocks CB1, will stop suckling
and die.69
The endocannabinoid system modulates cell metabolism
via ghrelin, leptin, orexin, and adiponectin signaling pathways.
70 Obesity leads to excessive production of endocannabinoids
by adipocytes, which drives CB1 into a feed-forward
dysfunction, contributing to metabolic syndrome.70
A pharmaceutical corporation recently sought approval
of rimonabant for the treatment of obesity. However, the US
Food and Drug Administration rejected it, in part, because
26% of subjects who took rimonabant in clinical studies
reported depressed mood, irritability, agitation, anxiety,
insomnia, headache, or other adverse psychiatric effects.71
Given the many beneficial roles of the endocannabinoid system,
it should be no surprise that rimonabant unmasked previously
silent multiple sclerosis and seizure disorders and doubled
the risk for suicidality.71
Mood disorders call into question the external validity
(or generalizability) of rimonabant clinical trials because
patients with depression were excluded from the studies, yet
psychiatric illness was the leading reason subjects dropped
out of the studies.72 In standard clinical practice, approximately
50% of patients seeking treatment for obesity also suffer
from depression.73
Complete blockade of CB1 might approximate the phenotype
expressed by genetically engineered “CB1 knockout
mice.” Mice lacking CB1 have increased morbidity, premature
mortality, age-related neuron loss, and show greater,
epilepsy, and anhedonia. They also exhibit aggressive, anxiogenic-
like, and depressive-like behavior, as well as a fear of
newness.74,75
Compared with mice, humans have a higher “endocannabinoid
tone.”76 Rimonabant blocks endocannabinoid
tone at CB1 and spontaneous CB1 constitutional activity because
it works as a full inverse agonist. As suggested in one study,22
sustaining a baseline endocannabinoid “hedonic tone” in the
mesolimbic system may enable humans to maintain personal
optimism and productivity in the face of chronic societal stress
and an ultimately unrewarding consumer culture. Thus, systemic
CB1 blockade may not help a psychologically stressed
obese patient who does not exercise and whose diet is high in
refined sugars, white flour, and trans fatty acids, and lacks
fiber, vegetables, and omega-3 fatty acids.
Rimonabant is currently approved in Europe, and similar
drugs, including taranabant, surinabant, CP-945598, and SLV-
319, are in development (see www.clinicaltrials.gov). Although
the risk-benefit profile of systemic CB1 full inverse agonists may
prohibit their use for chronic conditions such as obesity and
drug or alcohol dependence, they could serve in the treatment
of acute endocannabinoid dysregulation, such as hepatic
cirrhosis, hemorrhagic or endotoxic shock, cardiac reperfusion
injury, and doxorubicin-induced cardiotoxicity.50
The use of a CB1 partial agonist may prevent adverse
psychiatric effects while simultaneously acting as a partial
antagonist by blocking the binding of endocannabinoids at
CB1. Partial agonists serve best as partial antagonists when
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
JAOA • Vol 108 • No 10 • October 2008 • 593
Role of Osteopathic Medicine
Osteopathic principles and practice (OPP) were established
by Still in the late 19th century. The endocannabinoid system
may reflect OPP on a molecular level. Leaders in the osteopathic
medical profession proposed four tenets of OPP,11
which are presented below, accompanied by a description of
their relationship to the endocannabinoid system.
1. A person is the product of dynamic interaction between
body, mind, and spirit—This holistic principle is exemplified
by cannabinoid receptors, which span the field of psychoneuroimmunology.
Taken together, CB1, CB2, and their
endocannabinoid ligands represent a microcosm of mindbody
medicine.
2. An inherent property of this dynamic interaction is the
capacity of the individual for the maintenance of health and
recovery from disease—This self-regulatory capacity can
be rephrased as the maintenance of homeostasis. The endocannabinoid
system’s capacity to maintain homeostasis has
been cited many times in this review, and dozens of additional
citations can be found in the exhaustive review by
Pacher et al.50
The endocannabinoid literature rarely mentions allostasis,
a process by which homeostasis adapts to environmental
stress. However, it appears in osteopathic medical literature
and may become costly in the scenario of chronic stress
(“allostatic load”).37 Many studies cited in the present review,
especially concerning HPA axis activation and the immune
response, indicate that the endocannabinoid system promotes
allostasis as well as homeostasis.
3. Many forces, both intrinsic and extrinsic to the person,
can challenge this inherent capacity and contribute to the
onset of illness—Correspondingly, the endocannabinoid
system is challenged by intrinsic forces (changes in its structure
and genetic expression) and extrinsic forces (changes in
its function by unhealthy lifestyles). A corollary to this tenet
is the principle that structure and function are interrelated
at all levels. Because CB1 and CB2 express different molecular
structures, they exert different molecular functions.
4. The musculoskeletal system significantly influences the
individual’s ability to restore this inherent capacity and
therefore to resist disease processes—The endocannabinoid
system is expressed by the musculoskeletal system.10,70
Harnessing it has broad therapeutic potential for treating
degenerative and inflammatory conditions of the musculoskeletal
system. The fact that insulin resistance resides in
skeletal muscles also implicates the endocannabinoid system
in its relationship to cardiometabolic risk.70
Reciprocal Interactions
A focus on somatic dysfunction, formerly referred to as the
“osteopathic lesion,” dates to the founding of osteopathic
medicine. Somatic dysfunction is defined as “impaired or
altered function of related components of the somatic system:
levels of endogenous agonists are elevated—the exact scenario
seen with endocannabinoids and obesity.70 Partial agonists
are currently used in cardiology (eg, pindolol) and psychiatry
(eg, aripiprazole, buspirone hydrochloride), with fewer
adverse effects than antagonists or inverse agonists and without
compromising clinical efficacy. In animal studies, two partial
agonists have been shown to reverse obesity: the 5-HT6 ligand
E-683777 and a Chinese herbal formula with an “undisclosed
herb” that acts as CB1.78
Partial agonists often act as pleiotropic drugs, also known
as “selectively nonselective drugs.” These “magic shotguns”
interact with several molecular targets and provide superior
therapeutic effects and adverse effect profiles compared with
the action of a selective, single “magic bullet.”79,80
Biological Oscillators
The endocannabinoid system alters every biological oscillator
or pacemaker cell investigated to date, beginning with somite
formation in the embryo. The segmentation clock converts
oscillating Hox genes into spatial somite patterns.81 Fibroblast
growth factor regulates the Hox clock,81 and evidence suggests
that it uses endocannabinoid signaling.30,82,83 Many endocannabinoid-
altered oscillators change the rhythms of tissue
movement, such as:
 Cardiac pulse rate and contractility—Endocannabinoids
have been reported to cause dose-related biphasic effects in
rodents and humans.50
 Thoracic respiration—Endocannabinoids cause little change
in thoracic respiration, though intravenous injection of highdose
synthetic CB1 agonists in urethane-anaesthetized rats
decreased the respiration rate.50
 Gastrointestinal motility and peristalsis—Slowing of rate
and rhythm and gastrointestinal secretions in rodents and
humans.50
Human consciousness represents the rhythmic entrainment
of synchronously firing neurons.84 In five regions of the
human brain,23 such neurons are particularly enriched with
CB1:
 The hippocampus is a source of theta and gamma band
oscillations and is the neural substrate responsible for declarative
memory.85
 Striatal tissues contribute to the “beat frequency” model of
time perception.86
 The cerebellum plays a role in rhythm production and selfpaced-
behaviors, and cannabis accelerates the cerebellar
clock.87
 The suprachiasmatic nucleus is responsible for controlling
circadian rhythms.88,89
 The pineal gland produces melatonin and 2-AG in a circadian
rhythm driven by the suprachiasmatic nucleus and
regulated in part by CB2 in animal studies.90
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
594 • JAOA • Vol 108 • No 10 • October 2008
skeletal, arthrodial, and myofascial structures and related vascular,
lymphatic, and neural elements.”37 A new model hypothesizing
reciprocal interactions between somatic dysfunctions
and the endocannabinoid system is presented in the following
paragraphs.
In 1910, Still91 wrote that somatic dysfunction in peripheral
tissues “produce[s] pressure and obstruct[s] the normal discharge
of nerve and blood supply. Sometimes we find them
squeezed so closely as to produce adhesive inflammation.”
The inflammatory basis of somatic dysfunction was
demonstrated experimentally by Burns 100 years ago and was
subsequently confirmed in humans by Denslow.37,92,93 Inflammation
arises from three sources: (1) the release of inflammatory
mediators by injured tissue, (2) cytokines released by
leuckocytes that migrate into the area, and (3) inflammatory
neuropeptides (substance P and calcitonin gene-related peptide)
released from the peripheral terminals of nociceptors.37
All three sources are diminished by the endocannabinoid
system, as discussed previously36,51,52,60 and as illustrated in
Figure 2. However, the endocannabinoid system requires the
presence of CB1 in the peripheral terminal of the nociceptor
because CB1 receptors are synthesized in the dorsal root ganglion
of nociceptors and are carried by axoplasmic flow to
peripheral sites.94 By obstructing axoplasmic flow and cellular
trafficking of CB1, the pathophysiology of somatic dysfunction
perpetuates itself.
Still2 further alluded to axoplasmic flow and attributed the
cause of dysfunction to “partial or complete failure of the
nerves to properly conduct the fluids of life.” Korr95 demonstrated
mechanical derangement of axoplasmic flow in rabbits
and noted the following:
“Deformations of nerves and roots, such as compression,
stretching, angulation, and torsion, that are known to occur
all too commonly in the human being […] are subject to
manipulative amelioration and correction.
A rat study demonstrated that a suture loop ligated
around the sciatic nerve caused damming of CB1 proximal to
the suture loop (Figure 2).94 The suture loop may be analogous
to myofascial barriers that restrict axoplasmic flow, such
as in piriformis contracture, carpal tunnel syndrome, or thoracic
outlet restriction. Osteopathic physicians frequently use
OMT to treat nerves restricted by mechanical compression96,97
and conceivably restore the axoplasmic transport of CB1.
Central neural mechanisms also perpetuate somatic dysfunction.
Still91 presciently focused on the dorsal horn, stating
“the lesion is a sclerosis of the posterior root-zones of the
spinal cord.” Denslow described the “facilitated segment” as
a sustained neural reflex with motor and autonomic components,
whose sensory element was proprioceptive92 or nociceptive.
93
Willard37 updated and extended the nociceptive model by
describing segmental facilitation as a form of central sensitization
driven by excessive glutamate release. The endocannabinoid
system again comes into play here by reducing
central sensitization, as previously discussed,36 and by dampening
sympathetically mediated pain50—thereby thwarting
the primary “organizers” of somatic dysfunction.92
Chiropractic researchers98 implicated glutamate release
from nociceptors as the source of long-term potentiation in
spinal cord neurons. They noted that long-term potentiation
can be reversed by long-term depression and that spinal
manipulation imparts long-term depression by an uncertain
mechanism.98 However, these researchers98 did not mention
that the endocannabinoid system induces long-term depression,
which is caused by a sustained decrease in glutamate
release from presynaptic cells.40 It has been proposed that
high levels of CB1 expressed in the dorsolateral funiculus of the
spinal cord are positioned to influence viscerosomatic reflexes
as well.36
Central sensitization is analogous to the upregulation of
acetycholine (ACh) transmission at the motor endplate (ie,
the neuromuscular junction), which may be the pathologic
process underlying myofascial trigger points.8 As my colleague
David G. Simmons, MD, and I previously hypothesized,
8 CB1 in motor end plates dampen ACh release and perhaps
play a role in preventing or treating myofascial trigger
points.8 Our hypothesis has been supported by two new
animal studies99,100 showing that CB1 activation in motor end
plates dampens ACh release.
It is important to note that cannabinoids suppress pain
responses rather than all somatosensory input because cannabinoids
scarcely alter nonnociceptive neurons in the spinal cord
and thalamus.36 To wit, the endocannabinoid system inhibits
persistent nociception (ie, inflammatory pain, neurogenic pain,
chronic pain) more efficiently than acutely evoked nociception.
This difference may explain the anecdotal observation that
cannabis provides little relief at the dentist’s office.13
Beginning with William Garner Sutherland, DO, in 1899,
many researchers have made hypotheses regarding the primary
respiratory mechanism, known to osteopathic physicians
and other clinicians as the cranial rhythmic impulse
(CRI).101 The CRI is an oscillatory phenomenon that remains
poorly understood and is somewhat difficult to relate.
Acupuncturists face a similar situation when asked to describe
“chi.” The CRI is usually attributed to the well-documented
oscillatory secretion of cerebrospinal fluid (CSF) by ependymal
cells that line the choroid plexus and cerebral ventricles.101
Alternatively, the CRI may represent a palpable harmonic frequency,
a summation of several biological oscillations,
including CSF pulsations, Traube-Hering waves, cardiac pulse,
and diaphragmatic respiration.102,103
Given the impact of the endocannabinoid system on many
biological oscillators, it is easy to speculate that the endocannabinoid
system modulates the CRI. Human CSF is awash
with endocannabinoids.48,104 Cells lining the rodent and human
ventricular system express CB1 and endocannabinoid
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
JAOA • Vol 108 • No 10 • October 2008 • 595
tion, dietary supplements, and drug and alcohol restraint.
Exercising on a treadmill or a stationary bike increased AEA
levels circulating in the bloodstream.67,115 Chronic stress downregulated
CB1 expression in rodents,116 so stress-reduction
programs may offer a potential enhancement of the endocannabinoid
system.
Dietary inclusion of fish oils with DHA (docosahexaenoate
22:6 -3) and other polyunsaturated fatty acids increased AEA
and 2-AG levels in the brain.117,118 Oral administration of Lactobacillus
upregulated CB2 in intestinal epithelial cells and
relieved irritable bowel syndrome.119 Acute ethanol ingestion
decreased AEA and 2-AG in most regions of the brain,120 and
chronic ethanol abuse downregulated CB1 expression.121
A study of isopropyl dodecylfluorophosphonate, a
homolog of organophosphate pesticides, showed that it inhibited
both FAAH and MAGL. The inhibition caused 10-fold
elevations in AEA and 2-AG in mice, leading to
cannabimimetic behavioral effects.122 However, when incautious
adolescents sprayed cannabis with organophosphate
pesticides, the adulteration did not enhance cannabimimetic
effects. Instead, it overwhelmed the patients with cholinergic
adverse effects.123
Preclinical studies (animal models and human in vitro
assays) indicate that indomethacin, ibuprofen, and other nonsteroidal
anti-inflammatory drugs (NSAIDs) inhibit the endocannabinoid
catabolic enzymes COX2 and FAAH,124 so
NSAIDs prolong the activity of 2-AG and AEA. In animal
models, coadministration of NSAIDs with endocannabinoids
had a synergistic effect.125 This may explain why NSAIDs
sometimes cause sedation and other unexpected psychotropic
effects in patients. Acetaminophen is deacetylated and conjugated
with arachidonic acid into N-arachidonoylphenolamine,
a compound that activates CB1.126,127
Dexamethasone potently upregulates CB1 in rodents.128
The tricyclic antidepressant desipramine increases CB1 densities
in the brain of rodents129 while fluoxetine decreases CB1
expression in rodents.130 Diazepam and endocannabinoids
produce synergistic anxiolytic effects in mice, leading
researchers to propose that enhancement of endocannabinoid
function increases the effectiveness of diazepam.131 Valproate
sodium, an anticonvulsant and mood-stabilizing drug, upregulates
CB1 in rodents—a newly discovered mechanism of
action.132
Cannabidiol and THC may widen their own therapeutic
windows by increasing AEA levels.61,133-135 Low, subtherapeutic
doses of THC markedly potentiate the antinociception
imparted by endogenous cannabinoids.61 Surprisingly, THC
upregulates CB1 expression when administered acutely.136 It
may also cause post-translational modifications in CB1 that
stabilize the receptor in a constitutively active conformation.
Hypothetically, CB1 may remain constitutively active long
after THC has been metabolized and excreted.22 Similar adaptations
arise in -receptors after chronic exposure to plant
opiates.137
enzymes,23,44,105,106 which modulate the rhythmic production
of CSF in rodents,107 control endocannabinoid levels in rodent
CSF,105 and even provide restraint of suture ossification.57
An osteopathic procedure that alters the CRI, known as
the compression of the fourth cerebral ventricle (CV-4), may
transiently increase hydrostatic pressure in the cerebral ventricular
system in humans and cats.101,108 Conceivably, the
increased hydrostatic pressure could trigger CB1-constitutive
activity.48 As described in a previous study,26 the angiotensin
receptor is stretched into constitutive activity by hydrostatic
pressure in a Flexercell apparatus (Flexcell International Corp,
Hillsborough, NC).26 Correspondingly, equiaxial stretching
of fibroblasts in an identical apparatus caused a doubling of CB1
expression.10 The Flexercell management of fibroblasts has
provided an in vitro model of osteopathic manipulation.109
Activation of CB1 may explain many CV-4 effects, such as
relaxation and drowsiness, decreased sleep latency, and
decreased sympathetic nerve activity.110
Lastly, it should be noted that enhancing the endocannabinoid
system improves cardiovascular circulation.50,111
This may be one mechanism by which OMT improves health—
“the rule of the artery is supreme.”2
Enhancing the Endocannabinoid System
New research with endocannabinoid enzyme inhibitors provides
a proof-of-principle for the concept that enhancing endocannabinoid
signaling is a beneficial therapeutic strategy.112
Inhibitors of FAAH, an enzyme that breaks down AEA, and
MAGL, an enzyme that breaks down 2-AG, are anxiolytic
and antidepressant and block nociception, yet the inhibitors do
not impart psychoactive effects characteristic of direct CB1
agonists such as THC.112
The remainder of this review focuses on three ways to
enhance endocannabinoid function: (1) lifestyle modifications,
(2) pharmaceutical approaches, and (3) OMT. An increasing
number of conditions have been characterized as “endocannabinoid
deficiency syndromes,” including posttraumatic
stress disorder, chronic anxiety, migraine, Parkinson syndrome,
and irritable bowel syndrome.104,113 Fibromyalgia may
also involve endocannabinoid deficiency. During a normal
menstrual cycle, AEA decreases during the luteal phase (circa
day 21) as a result of the progesterone-induced upregulation
of FAAH.114
In a study of healthy women with normal menstrual
cycles, the luteal phase corresponded with hypersensitivity
to algometer-induced pressure at fibromyalgia tender points.
Several subjects “changed” fibromyalgia diagnosis during the
course of a menstrual cycle, fulfilling the tender point criterion—
defined as tenderness elicited by 4 kg of pressure at 11
of 18 pressure points—during the AEA-deficient luteal phase
or menstrual phase, but never during the AEA-rich follicular
phase.114
Studies suggest endocannabinoid deficiency may be rectified
by lifestyle modifications, including exercise, stress reduc-
McPartland • Review Article
REVIEW ARTICLE
Downloaded From: http://jaoa.org/ on 07/19/2017
596 •