To determine if peripheral neuropathology exists among the innervation of cutaneous arterioles and arteriole–venule shunts (AVS) in fibromyalgia (FM) patients.
To determine if peripheral neuropathology exists among the innervation of cutaneous arterioles and arteriole–venule shunts (AVS) in fibromyalgia (FM) patients.
Cutaneous arterioles and AVS receive a convergence of vasoconstrictive sympathetic innervation, and vasodilatory small-fiber sensory innervation. Given our previous findings of peripheral pathologies in chronic pain conditions, we hypothesized that this vascular location may be a potential site of pathology and/or serotonergic and norepinephrine reuptake inhibitors (SNRI) drug action.
Twenty-four female FM patients and nine female healthy control subjects were enrolled for study, with 14 additional female control subjects included from previous studies. AVS were identified in hypothenar skin biopsies from 18/24 FM patient and 14/23 control subjects.
Multimolecular immunocytochemistry to assess different types of cutaneous innervation in 3 mm skin biopsies from glabrous hypothenar and trapezius regions.
AVS had significantly increased innervation among FM patients. The excessive innervation consisted of a greater proportion of vasodilatory sensory fibers, compared with vasoconstrictive sympathetic fibers. In contrast, sensory and sympathetic innervation to arterioles remained normal. Importantly, the sensory fibers express α2C receptors, indicating that the sympathetic innervation exerts an inhibitory modulation of sensory activity.
The excessive sensory innervation to the glabrous skin AVS is a likely source of severe pain and tenderness in the hands of FM patients. Importantly, glabrous AVS regulate blood flow to the skin in humans for thermoregulation and to other tissues such as skeletal muscle during periods of increased metabolic demand. Therefore, blood flow dysregulation as a result of excessive innervation to AVS would likely contribute to the widespread deep pain and fatigue of FM. SNRI compounds may provide partial therapeutic benefit by enhancing the impact of sympathetically mediated inhibitory modulation of the excess sensory innervation.
Fibromyalgia (FM) is a common chronic widespread pain condition, afflicting 2–5% of the US population, with females being affected at well over twice the rate of males [1-3]. FM is a complex multifaceted disorder characterized by chronic widespread pain, allodynia, hyperalgesia, sensory sensitivities (visual, auditory), severe fatigue, sleeplessness, cognitive dysfunction, and endocrine abnormalities [4-7]. Disease modulators include psychologic and biologic stressors, diet, and genetic predispositions involving serotonin, dopamine, and/or norepinephrine/noradrenaline (NA) [1, 2, 8, 9]. Current Food and Drug Administration (FDA)-approved treatments include serotonergic and norepinephrine reuptake inhibitors (SNRI), and a voltage-gated calcium channel subunit (alpha 2-delta) ligand, pregabalin. However, at best, these compounds provide temporary partial relief in only a portion of FM patients, and have dose limiting side effects [7, 10-12].
The prevailing hypothesis of FM primarily postulates a sensitization of pain pathways in the central nervous system (CNS) accompanied by evidence of increased excitatory neurotransmitters and inflammatory cytokines in the cerebrospinal fluid (CSF) [5, 13-17]. The source and maintenance of the CNS sensitization is unknown. To date, a peripheral nervous system (PNS) and/or target cell pathology has not been identified in FM patients, although evidence of small fiber neuropathy and cutaneous compartment pathologies have been detected in skin biopsies among several other chronic pain conditions [18-26]. Increasing evidence indicates that the excessive fatigue and widespread deep pain associated with FM is caused by peripheral tissue ischemia and hyperactivation of deep tissue nociceptors by anaerobic metabolites and inflammatory cytokines, a process that would drive and maintain CNS sensitization [1, 27-29]. The source of the ischemia is assumed to arise from excessive sympathetically mediated vasoconstriction, which can also be exacerbated by stress [9, 30-33].
Diagnosis of FM includes hypersensitivity to pressure applied to at least 11 out of 18 specific tender points, although more recent guidelines have placed less emphasis on tender point counts [34, 35]. FM pain is widespread with mechanical allodynia, thermal sensitivity, and especially tenderness over most areas of the body. This includes the hands, which are not a diagnostic FM tender point site specifically, but have been used for extensive testing of pain thresholds and blood flow in FM patients [29, 36-39]. Mechanical allodynia in response to pressure occurs rapidly as a “first pain” perception in most FM patients. Tests using heat stimulation demonstrate that heat hyperalgesia occurs as a delayed “second pain” response, forming the rationale for postulating a “windup” mechanism contributing to central sensitization [17, 38, 40]. However, cold conditions particularly aggravate FM pain and vasospasms , and cold dysesthesia and hyperalgesia occur more frequently in FM subjects, whereas heat hyperalgesia typically occurs without dysesthesia and only as a subset among those with cold hyperalgesia [36, 37, 42].
Although SNRI are mostly assumed to act on CNS mechanisms, our integrated multimolecular immunocytochemical analyses of cutaneous innervation in mammals ranging from rodents to primates, including humans, has documented that arterioles and arteriole–venule shunts (AVS) located deep in the dermis of glabrous skin are innervated by a convergence of dense sympathetic and sensory innervation that could be a peripheral target of SNRI [19, 43-47]. In particular, the arterioles and AVS have a well-known, dense noradrenergic sympathetic innervation, whose role in vasoconstriction has been extensively researched, but also a lesser known, dense sensory innervation composed of several varieties of immunochemically distinct C and Aδ fibers.
We have previously hypothesized a functional role in chronic pain conditions of small-caliber sensory innervation to the cutaneous vasculature [19, 43, 45]. However, although nearly all sensory innervation to cutaneous vasculature contains substance P (SP) and calcitonin gene-related peptide (CGRP), which are potent vasodilators and are implicated in inflammatory pain mechanisms [48-53], this topic has received only little research attention. Thus, the sensory innervation likely plays a role in vascular-related sensory feedback to regulate sympathetic activity, but may also be implicated in a direct effector role in vasodilation and potential inflammation among FM patients [48-50, 52, 54, 55]. Our previous multimolecular assessments demonstrated that the sensory fibers express the α2C receptor, indicating that the sympathetic innervation may exert a local inhibitory effect on the sensory activity [56-58]. We recently documented that this innervation may also contribute to a variety of conscious tactile sensations , but it undoubtedly works in concert with the sympathetic innervation for purposes of vasoregulation, especially the important thermoregulatory function of the glabrous skin in humans. Given the extreme sensitivity to cold conditions and stress, the deep pain sensation of FM patients, and the convergence and immunochemical properties of the sensory and sympathetic innervation, which could be a therapeutic target of SNRI, we hypothesized that FM patients might manifest a pathology involving the innervation to the cutaneous arterioles and AVS located deep in the dermis.
In order to assess the neurochemical and morphological characteristics of cutaneous innervation in FM compared with control subjects, we conducted multimolecular immunofluorescence analyses of 3 mm skin biopsies taken from the diagnostic trapezius tender point location and from the glabrous hypothenar skin where there is a high concentration of AVS [19, 43, 45, 47]. Skin biopsies were collected from 24 female FM-diagnosed patients and 23 age-matched female control subjects. Only females were studied here based on the increased incidence of FM [3, 8], which may be due to gender-related differences in vascular innervation [59-61]. We discovered excessive innervation of AVS in the palmar hypothenar skin of the FM patients which was skewed toward an over-representation of peptidergic sensory innervation and an underrepresentation of NA sympathetic innervation. This AVS neurovascular pathology is consistent with FM symptoms of pain and tenderness, exacerbation of these symptoms by cold conditions, and the potential beneficial effects of SNRI agents. Moreover, this pathology may have wide-ranging implications impacting blood flow throughout the body which may contribute to ischemia and widespread symptoms of deep tissue pain and fatigue in FM [2, 30].
The research protocol was approved by the Albany Medical Center Institutional Review Board (IRB), and all subjects gave written informed consent. All subjects were administered a medical evaluation to verify the absence of exclusionary diseases and signed appropriate IRB-approved consent prior to enrollment in the study. Skin biopsy collection was performed using established procedures and appropriate monitoring. No adverse events or serious adverse events attributed to the testing or biopsy procedures were documented for this study.
This investigation was performed to assess cutaneous pathology and the effect of treatment on patients diagnosed with FM. The reported data are compiled from two cohorts, FM patient and control subjects, as described in Table 1. Cohort 1 consisted of 24 female adults diagnosed with FM based from the 1990 American College of Rheumatology (ACR) criteria, including widespread pain present for at least 3 months, pain in at least 11 of 18 specific tender point sites, and 22/24 had scores greater than 40 on the 100-point visual analog scale (VAS) [6, 34]. After eligibility determination and informed consent, FM subjects underwent a washout of excluded (FM specific) medications and a 1-week baseline. Cohort 2 was composed of 23 age- and gender-matched control subjects, 9 enrolled for this study and 14 additional gender-matched control subject hypothenar biopsies from other ongoing IRB-approved studies. Comorbidities known to produce muscle pain, unstable medical conditions, and severe or uncontrolled depression were exclusion criteria for each group. Each enrolled subject was evaluated by standardized pain questionnaires, general and neurologic exam, tender point survey, and computerized quantitative sensory testing (QST) (Medoc, Durham, NC, USA). The QST was conducted on the dominate trapezius tender point site, a common location for intense FM pain. The full unblinding of the QST data is awaiting completion of the second portion of the ongoing study (FM response to medication).
|Age (Years)||FM Intensity Score||Visual Analog Scale (mm)||AVS Innervation area (mma) per Tunica Media Profile||Total Innervation Area (mma) of Largest AVS||CGRP/PGP Ratio||NF/PGP Ratio||NPY/CGRP Ratio|
|T-test (FM vs CTL)|
Immediately following the clinical assessments, a single 3 mm skin punch biopsy was collected under local lidocaine anesthetic from two distinct locations: 1) the trapezius thoracic tender point site utilized for QST testing, and 2) the palmar hypothenar glabrous skin compartment (halfway between proximal to distal extent of the compartment and about 1 cm from the medial margin). This pair of skin biopsies (one trapezius, one hypothenar) was collected from all 24 FM patients and the nine healthy controls specifically enrolled for this study. Only a hypothenar biopsy was collected under identical procedures from the additional 14 control subjects. The hypothenar biopsy site has a high probability of sampling cutaneous vasculature, including specifically AVS [18, 43, 62]. This skin biopsy location has been utilized routinely by Integrated Tissue Dynamics, LLC (Intidyn, Rensselaer, NY, USA) as part of an integrated multimolecular Chemomorphometric Analysis (CMA) platform, directed at profiling a wide variety of C, Aδ, and Aβ axon endings and terminal cell compartments [19, 43, 45, 62]. A rich variety of large and small-caliber sensory endings and vascular structures are located within the hypothenar region, whereas these are sparse in back skin and rarely found in 3 mm biopsies obtained from that location. Immediately after removal, the biopsies were fixed by immersion in 4% paraformaldehyde (in 0.1 M phosphate-buffered saline; pH 7.4) for 4 hours at 4°C, then rinsed and stored in PBS at 4°C until processing. Biopsies were then cryoprotected overnight in 30% sucrose in PBS, frozen, and cryostat sectioned at 14 μm thickness perpendicular to the epidermal surface. The sections were thaw mounted in serial order, alternating across at least 20 slides such that each slide contained numerous sections from equally spaced intervals throughout the biopsy.
Biopsy specimen were processed following established protocols for integrated multimolecular immunofluorescence assessments [18, 24, 26, 43, 62] and were evaluated utilizing the CMA platform developed by Intidyn. This procedure enables the use of a wide variety of immunolabel (IL) combinations designed to elucidate different functional and morphological varieties of cutaneous innervation. Based on our previous research, we had identified that arterioles and AVS are innervated by four major subtypes, all of which express immunolabeling for the pan-neuronal enzyme protein gene product (PGP)9.5:
Therefore, immunofluorescence in this study was assessed using combinations of rabbit antihuman PGP9.5 (1:800; UltraClone LTD, Isle of Wight, UK), rabbit anti-CGRP (1:800; Millipore, Billerica, MA, USA), sheep anti-CGRP (1:800, AbCam, Cambridge, UK), mouse anti-200 kD NF protein (1:800, Sigma, St. Louis, MO, USA), sheep anti-NPY (1:800, Millipore). Based on preliminary assessments that CGRP-positive innervation expresses immunoreactivity for α2C receptors, a rabbit anti-α2C (1:400, Neuromics, Minneapolis, MN, USA) was also used. A separate slide of sections from each biopsy was processed for immunolabeling for each of the following: anti-PGP alone, CGRP/PGP, NF/PGP, CGRP/NF, NPY/CGRP, α2C/CGRP, and α2C/NPY. Double label combinations used the appropriate species of secondary antibodies conjugated with either Cy3 or Alexa488 (Jackson ImmunoResearch, West Grove, PA, USA; Molecular Probes, Eugene, OR, USA). Slides were preincubated in 1% bovine serum albumin and 0.3% Triton X-100 in PBS (PBS-TB) for 30 minutes and then incubated with primary antibodies diluted in PBS-TB overnight in a humid atmosphere at 4°C. Slides were then rinsed in excess PBS for 30 minutes and incubated for 2 hours at room temperature with the appropriate secondary antibodies diluted in PBS-TB. Following secondary antibody incubation, the sections were rinsed for 30 minutes in PBS and coverslipped under 90% glycerol in PBS. A number of unprocessed slides were archived under glycerol for future retrospective analyses with additional biomarkers.
Images were captured utilizing an Olympus DSU spinning disk confocal BX51-WI base microscope equipped with a Hamamatsu ER (Olympus America, Center Valley, PA, USA), DVC high-speed and Optronics Microfire cameras (Optronics, Goleta, CA, USA), 10 position LEP motorized filter wheel (Ludl Electronic Products Inc., Hawthorne, NY, USA), 3 axis motorized stage system, linear focus encoder, and Vistek vibration isolation platform (Vistek Inc, Tempe, AZ, USA), or an Olympus Optical Provis AX70 microscope (Olympus America) equipped with a high-resolution three-color camera (DKC-ST5; Sony, Montvale, NJ, USA). Both systems are equipped with conventional fluorescence filters (Cy3: 528–553 nm excitation, 590–650 nm emission; Cy2/Alexa 488: 460–500 nm excitation, 510–560 nm emission) and are linked to computers interfaced with NeuroLucida (MBF Bioscience, Essex, VT, USA) quantitative microscopy software, and Photoshop (Adobe Systems, San Jose, CA, USA) or Northern Eclipse (Empix Imaging, Mississauga, ON, Canada) image capture and processing software. For each IL, camera settings were kept the same for all image captures across all biopsies. Immunolabeling and innervation were quantified for AVS, arterioles, venules, and epidermal nerve fiber counts using NeuroLucida and Photoshop software. Quantitative results were compared between the FM patients and the healthy control subjects by Student's t-test or Wilcoxon Rank Sum test, with significance set at P < 0.05.
Quantification of the dense, compact innervation of arterioles and AVS consisted of: 1) the area occupied by the innervation located around the perimeter of each arteriole and AVS profile within each section, and 2) the relative proportions of different types of innervation affiliated with each profile. To determine the area of innervation, each arteriole or AVS profile was located in a slide of sections processed only for PGP immunofluorescence for each biopsy. NeuroLucida routines were used to trace the contour and calculate the enclosed area of the vessel lumen, the smooth muscle vessel wall (tunica media), and the immunolabeled innervation in the surrounding tunica adventitia.
To determine the relative proportions of different types of arteriole and AVS innervation, different slides of sections from each biopsy were used that were double labeled using different pairs of primary antibodies detected with appropriate species of Cy3 (red) or Alexa 488 (green) conjugated secondary antibodies. The relative proportions for the following double label combinations were quantified:
Because inter-subject quantification of IL intensity is limited by several variables (i.e., individual antibody mixes for each processing day and different biopsies), the relative proportions of label luminescence was assessed for each arteriole and AVS profile in every section on each slide for each biopsy. The areas occupied by the innervation for each arteriole and AVS profile were circumscribed using Photoshop tools and the average luminescence of the circumscribed area was determined for the red and green channels. The average luminescence values for the red and green channels within the circumscribed area were then utilized to create relative proportion for each double label combination. Resulting proportions were subsequently averaged from all the double-labeled arterioles and AVS profiles in each biopsy. As the luminescence is a function of both the density and intensity of each labeled pixel, this approach provides a relative comparison between the proportions of different arteriole and AVS innervation types within control subject and FM patient biopsies.
Unlike arterioles and AVS, cutaneous venules have sparse innervation. This innervation was quantified by using NeuroLucida routines to trace the outer contour of the venous wall and then to map the location of nerve fiber along the contour. Innervation density was expressed as the number of profiles per unit length of each outer venule contour.
The epidermal innervation density was quantified in the hypothenar and thoracic biopsies of the nine control subjects and nine age-matched FM patients. Using NeuroLucida software, complete montages of three PGP9.5 immunolabeled sections taken at approximately one-third intervals through each biopsy were captured, and the total number of intraepidermal nerve fiber profiles that were observed to be in contact with the basement membrane were determined. The values were subsequently divided by the length of the epidermis to determine the intraepidermal nerve fiber densities (IEFD) for each biopsy.
The FM subjects ranged in age from 24 to 74 (average age 47 years) and the control subjects ranged from 22 to 70 years old (average age 51 years), and was not significantly different (Table 1, P = 0.30). The average ages of the 18 FM patients and 14 control subjects utilized for AVS analysis were 49 and 48 years, respectively. At initial screening visits, the patient average FM intensity score (FIS) and VAS were 6.0 ± 2.0 and 60.0 ± 18.0, respectively. This is consistent with moderate to severe pain. Control subjects reported no pain; FIS and VAS were not recorded.
As seen in sections from hypothenar skin biopsies, profiles of arterioles, venules, and AVS have various morphologies depending upon how they are cut by the plane of sectioning (Figures 1 and 2). Arterioles ramify throughout the deep dermis and appear in at least some sections of all biopsies as individual profiles with mostly uniform circular to oval shapes of varying lengths (solid arrows in Figures 1 and 2). Arterioles often have an open lumen surrounded by a uniformly thick smooth muscle tunica media, enveloped by a tunica adventitia composed of collagen bundles, and which contains dense innervation of both sympathetic and sensory origin [19, 45]. Venules typically have irregular contours with a relatively thin tunica media, an open irregular lumen, and very sparse innervation discontinuously spaced around the perimeter of the tunica media (open arrows in Figure 1). Arteriole and venule profiles were encountered in the hypothenar biopsies from 21 of the 23 control subject biopsies and from 21 of the 24 FM patient biopsies.
AVS occur at discreet locations within the vascular tree as short, tortuous, direct connections between an arteriole and a venule, where they function as valves to shunt blood flow either into or to bypass terminal capillary networks for thermoregulatory mechanisms. Due to their tortuous shape, a section passing through an AVS can result in several circular, oval, or s-shaped vessel profiles (asterisks in Figures 1 and 2). The tunica media of each AVS profile is much thicker than that of an arteriole profile, the lumen is typically more constricted, and the tunica adventitia is thicker and more densely packed with innervation. AVS profiles were obtained in the hypothenar biopsies from 14 of the 23 control subjects and 18 of the 24 FM patients. In certain planes of sectioning, an arteriole can be seen entering and connecting to the AVS (Figures 1A,I) and a venule can be seen leaving the AVS (Figure 1H).
The deep dermis of thoracic biopsies had few vascular profiles, and those present were far smaller than those in the hypothenar biopsies. Presumptive arterioles had a thin perimeter of innervation and presumptive venules had little if any innervation. There were no AVS profiles.
Anti-PGP9.5 labels for an enzyme highly enriched in neuroedocrine cells which is normally expressed in all known peripheral innervation. PGP9.5 immunolabeling reveals the totality of the vascular innervation (Figure 1), which can be subtyped through immunolabeling for other specific antigens. Normal sympathetic innervation of cutaneous arterioles consists of unmyelinated noradrenergic fibers that have been previously shown to express TH and to IL for anti-NPY. The NPY-positive sympathetic innervation is concentrated at the perimeter of the tunica media (Figure 2G). Normal sensory innervation of arterioles is more diffusely distributed in the tunica adventitia and consists predominately of presumptive C fibers that label with anti-CGRP, but not with anti-NF (Figure 2A,E,G). A smaller contingent of sensory innervation consists of presumptive Aδ fibers that label with anti-NF, and which have previously been shown to IL for anti-MBP (Figure 2C). A minority of the Aδ innervation colabels with anti-CGRP (Figure 2). The control AVS profiles were innervated by a similar mix of sympathetic and sensory axons, but which was at least 2× greater in quantity compared with an innervated arteriole. Importantly, double labeling with anti-α2C (NA receptor) and anti-CGRP (Figure 2I,J) demonstrated that α2C was largely coexpressed on CGRP positive C fibers. Double labeling for anti-α2C and anti-NPY revealed that the α2C-expressing peptidergic sensory C fibers are closely intermingled with the NPY-expressing sympathetic innervation.
AVS profiles from FM patients were dramatically larger in size compared with those in age-matched control subjects (Figure 1). As a reflection of their larger size, AVS were present in the hypothenar palm biopsies of 18 out of 24 FM patient biopsies (75%) compared with 14 out of 23 control subject biopsies (61%), and AVS profiles were encountered ∼3× more frequently among the individual biopsy sections from FM patients compared with control subjects. The total average area occupied by complete AVS profiles among FM patients was significantly increased compared with control subjects (Figure 3A, P < 0.001). Importantly, serial reconstructions revealed that there were not increased numbers of AVS per biopsy in FM patients compared with control subjects (Figure 3B, P = 0.11), indicating that the AVS had increased in size.
To assess what components were contributing to the increased size of the FM AVS, NeuroLucida mapping routines were applied to PGP9.5 labeled sections from every FM patient and control subject biopsy that contained an AVS. Profiles were mapped to determine: 1) the area of the tunica media of each AVS profile (Figure 3C); 2) the number of tunica media profiles (with lumen) per AVS (Figure 3D); and 3) the area containing the innervation affiliated with each AVS profile (Figure 3E,F). These metrics were also determined for the arteriole profiles within sections from each hypothenar biopsy that contained arteriole profiles (Figure 4A–C).
No significant difference was detected in the average cross-sectional areas of the tunica media profiles encountered among the AVS from control subject compared with FM patient biopsies (Figure 3C, P = 0.83), indicating that the smooth muscle had not hypertrophied in the FM patients. Similarly, no significant difference was found among the cross-sectional areas of arterioles from control subject compared with FM patient biopsies (Figure 4A). To determine whether the AVS had became more tortuous, the number of tunica media profiles associated with each sectioned AVS was assessed. FM patient AVS did show a slight but significant increase in tortuosity compared with control biopsies (Figure 3D, P = 0.04).
Quantification of the area occupied by the innervation to each AVS revealed a highly significant increase among FM patients compared with control subjects. The results demonstrate that the overall average innervation area of AVS profiles were two times greater in FM patient than control subject biopsies (Figure 3E: 20 × 103 μm  vs 10 × 103 μm , respectively, P < 0.001). Furthermore, the total area occupied by innervation from the single largest AVS profile identified from each biopsy demonstrated an innervation area that was four times greater in the FM patient compared with control subject biopsies (Figure 3F: 87 × 103 μm  vs 22 × 103 μm , respectively, P < 0.001). There were no significant differences in the innervation area surrounding the arterioles (Figure 4B); however, when arteriole innervation area was normalized by tunica media size, a slight but significant increase in arteriole innervation among the FM patient biopsies was detected (Figure 4C, P = 0.011). There were no significant differences in the innervation area surrounding venules (Figure 4D) from FM patient compared with control subject biopsies.
Additionally, there was only very limited overlap between the ranges of AVS innervation areas from FM patient compared with control subject biopsies, regardless of age (Figure 5A). Furthermore, the innervation area remained similar across age in the control subjects, whereas the exuberant innervation areas varied dramatically across the FM patient cohort. Importantly, at every age-matched comparison, the AVS innervation areas from FM patient biopsies were always significantly greater than control biopsies (Figure 5A).
Initial qualitative assessments of double label combinations indicated that peptidergic (CGRP-containing) sensory axons disproportionately contributed to the increased AVS innervation (Figure 2A,B). Therefore, a quantitative assessment was also conducted to determine whether the increased AVS innervation observed in FM patients involved a differential shift in the functional type of axons, as assessed by IL combinations for CGRP, NF, NPY, and PGP. These quantitative analyses were conducted on all biopsies that contained AVS profiles. The results demonstrate that the CGRP-positive innervation was disproportionately represented in FM patient AVS profiles compared with control subject AVS profiles (Figure 3G, P < 0.0001). By contrast, the average proportion of NF-IL to PGP-IL luminescence (Figure 2C,D) was virtually identical between the FM patient and control subject AVS profiles (Figure 3H, P = 1.04). Results from CGRP-IL and NF-IL revealed that only a small proportion of the total Aδ fiber innervation among AVS profiles was peptidergic (CGRP positive) from either the control subject or FM patient biopsies (Figure 2E,F). Furthermore, a comparison of NPY-IL and CGRP-IL combinations (Figure 2G,H) revealed that the proportion of NPY-expressing sympathetic innervation was significantly decreased relative to the CGRP expressing sensory innervation in the FM patient AVS profiles (Figure 3I, P < 0.0001). Taken altogether, these results demonstrate that the sensory peptidergic C-fiber innervation has increased to a greater degree than the Aδ fiber innervation, and has resulted in a significantly increased ratio of CGRP-expressing vasodilatory sensory innervation to noradrenergic vasoconstrictive sympathetic innervation among the FM patient AVS profiles compared with control subjects (Figure 5B).
A routine subjective assessment of arterioles and venules among the thoracic skin biopsies from FM patients and control subjects did not reveal any potential differences between the vascular profiles or innervation.
Epidermal innervation density was assessed in the hypothenar and thoracic biopsies from the nine control subjects specifically recruited for this study, and in nine age-matched FM patients. No differences were detected between the IEFD among hypothenar palmar control subject and FM patient biopsies (Figure 4E). In contrast, the thoracic back skin of FM patients did show a significant (∼20%) decrease in IEFD compared with control subjects (Figure 4F, P = 0.03), consistent with observations of decreased IEFD in other chronic pain skin conditions, and recently also demonstrated in FM [18, 63, 64]. Interestingly, based upon the counting procedures utilized, the overall IEFD counts were lower in the palmar skin compared with the thoracic skin. Total counts of epidermal endings, which included PGP-positive ending profiles not found to be in contact with the basement membrane, also did not reveal any difference between FM patient and control subjects for the thoracic back skin or hypothenar palmar skin (data not shown).
A major finding in this study revealed that the AVS in palmar glabrous skin biopsies of FM patients had excessive sympathetic and sensory innervation compared with normal control subjects, whereas innervation to arterioles and venules appeared normal. The CGRP-positive vasodilatory sensory innervation appeared significantly increased compared with the NYP-positive vasoconstrictive sympathetic innervation among the AVS. AVS have an especially thick tunica media and form short, valve-like connections directly between arterioles and venules, where they presumably function in concert with resistance arterioles to provide dynamic control of blood flow to the superficial capillaries [19, 43, 45, 47, 65, 66]. In human and nonhuman primates, the glabrous skin of the hands and feet has a particularly high surface to volume ratio with a high capillary density and plays a major role in thermoregulation and the maintenance of core body temperature as well as the apportionment of blood to other organs such as skeletal muscle during periods of high metabolic demand. Therefore, the particularly excessive sensory innervation of the AVS in the palmar skin of FM patients may contribute to extreme tenderness and pain in the hands, but may also contribute to the chronic widespread deep pain, fatigue, sleep disturbances, and cognitive dysfunction associated with FM.
Chronic pain research is focused primarily on two well-defined mechanisms which are not mutually exclusive. The first involves aberrant hypersensitivity and hyperactivity of the primary sensory neurons (nociceptors) in the PNS that are the inputs to central pain pathways, i.e., “irritable nociceptors,” while the second involves hyperresponsivity of pain pathway CNS neurons, particularly dorsal horn or thalamus, i.e., “central sensitization” [19, 45]. Hypersensitivity and hyperactivity of primary sensory neurons has particularly been implicated in chronic pain associated with such neuropathic conditions as postherpetic neuralgia (PHN), complex regional pain syndrome (CRPS), painful diabetic neuropathy, and erythromelalgia, where pathologies have been documented among the small-caliber innervation to the epidermis and upper dermis, and signaling mechanisms of keratinocytes [18, 19, 24, 26, 62, 67]. The pain symptoms associated with these conditions typically include combinations of thermal, chemical, and/or mechanical hyperalgesia and allodynia, as well as spontaneous burning pain perceived as originating primarily from the skin surface and localized to a specific location of the body. By contrast, FM has been attributed primarily to a central sensitization disorder due to the widespread diffuse deep pain, lack of direct evidence of a PNS pathology, and the partial therapeutic benefit of SNRI and calcium channel modulators that are regarded as acting primarily in the CNS. The diffuse deep tissue pain is presumed to originate from sensitized CNS responsiveness to deep tissue afferents with some indications that such afferents may also be hyperactive due to an unknown source of deep tissue ischemia.
Our study examined the innervation in biopsies from two sites: 1) thoracic hairy skin overlying the trapezius area tender point used to diagnose FM; and 2) the hypothenar glabrous skin of the hand. The hands in FM patients are typically among the most painful locations, although they not included as specific diagnostic tender point likely because pain in the hands also occurs in a variety of other disorders [1, 18, 34, 35, 68-71]. As such, the hands have been used in many studies designed to assess altered tactile sensation and changes in various pain thresholds in FM patients [17, 29, 36, 42, 72, 73]. In FM, a “first pain” mechanical pressure allodynia has been observed followed by a delayed “second pain” heat hyperalgesia. The immediate pressure allodynia is attributed to hypersensitivity among unidentified cutaneous mechanoreceptors, whereas the delayed heat hyperalgesia is attributed to sensitization of neurons in the CNS [5, 6, 10, 42, 74, 75]. However, the source and maintenance of the hypothesized central sensitization is unknown [4, 5, 10, 76, 77].
A reduction of small-caliber nociceptor epidermal innervation has been detected in a variety of chronic painful neuropathic conditions such as CRPS, diabetic neuropathy, and PHN [18, 20-22, 25, 67, 70, 78-80]. The seeming paradox of chronic pain associated with reduced nociceptor (pain) innervation has been partially explained by physiological evidence of hypersensitivity among the remaining innervation, which could be exacerbated by increased purinergic signaling and CGRP upregulation among epidermal keratinocytes [24, 26, 45, 81, 82]. However, reduced epidermal innervation does not always manifest chronic pain, and chronic pain is not always accompanied by reduced epidermal innervation. Recently, a significant loss of epidermal fibers from nonglabrous forearm and leg skin of FM patients has now been reported . Consistent with this observation, we detected 20% less nerve fibers entering the epidermis in the thoracic biopsies of FM patients compared with normal subject biopsies. We did not observe a difference among the epidermal innervation of hypothenar palmar skin biopsies. Although the loss of epidermal endings, accompanied by hyperexcitability of remaining endings might contribute to widespread pain, the decrease observed among the FM patients was not as severe as documented in other painful conditions, and furthermore, the loss of epidermal innervation is not intuitively consistent with the deep pain characteristic of FM patients.
We chose to biopsy the glabrous hypothenar skin to specifically investigate the innervation to arterioles and AVS as a potential site of FM related pathology in the PNS, particularly as these targets are located deep in the skin and are a major site of convergence for dense sympathetic and sensory innervation. The arterioles and AVS are innervated by dense NA sympathetic innervation, and also several molecularly distinct subtypes of small-caliber sensory innervation [19, 43, 46, 47, 62, 83]. The vast majority is peptidergic C-fiber innervation containing CGRP and SP, while a smaller, albeit substantial, contingent of the sensory innervation is composed of Aδ fibers, of which subsets can express or lack CGRP and SP. More recently, it has been established that subsets among the C and Aδ fibers express various combinations of ligand-gated receptors and ion channels, such as TrpV1, TrpA1, ASIC3, and/or H3R, each of which are implicated in mechanoreceptive, metaboreceptive, thermoreceptive, and various nociceptive capabilities attributed to small fiber afferents [18, 44, 84, 85].
To date, the function of the multiple varieties of arteriole and AVS sensory innervation has received little attention. The literature on the neural vasoregulation of the glabrous skin has focused almost entirely on sympathetically mediated vasoconstriction caused by NA and NPY activation of the smooth muscle tunica media [86-88]. The dilation of the arterioles and AVS is widely regarded as a passive consequence of reduced sympathetic activity or the dilatory action of compounds such as NO released from endothelial cells ; however, there is little mention of sensory feedback mechanisms operating among microvasculature to regulate blood flow. It is well established that CGRP and SP are potent vasodilators and that heat stimulation of peptidergic sensory endings in the skin can elicit a local “axon-reflex” effector release of CGRP and SP from collateral branches terminating among capillaries and precapillary arterioles [50, 86, 90-97]. Presumably, the extensive peptidergic sensory innervation to the arterioles and AVS serves an important vasodilatory role as well, particularly as the arteriole smooth muscle has also been shown to express appropriate peptidergic receptors . We recently reported that cutaneous vascular sensory innervation may also contribute to a variety of conscious mechanical and thermal tactile sensations, as well as pain . Aδ fibers innervating arterioles and AVS may particularly include low threshold mechanoreceptors whose primary role is to detect stretch and/or tension in the vessel walls in response to blood pressure .
Importantly, we found that the peptidergic sensory innervation to the arterioles and AVS coexpress α2C receptor immunoreactivity. Previous studies have shown that NA activation of an α2 receptor mechanism can inhibit the release of CGRP and SP . Therefore, sympathetic release of NA may mediate vasoconstriction not only by direct constriction of the tunica media smooth muscle, but also by blocking the vasodilatory peptide release from vascular afferents [56, 58, 100, 101].
Our multilabeling analyses revealed that nearly all of the AVS obtained in hypothenar biopsies from FM patients had excessive innervation. The excessive innervation included Aδ fibers (defined as NF-positive), but was significantly skewed toward an overrepresentation of CGRP-expressing innervation, which consists mostly of C fibers (defined as NF-negative, NPY-negative), and an underrepresentation of NA sympathetic innervation [19, 43, 47, 83]. Furthermore, even though AVS branch directly from the arterioles, only a slight excess in arteriole innervation was detected among the FM patients, and there were no indications that the proportions of arteriole sensory or sympathetic innervation differed between FM patient and control subjects. Taken together, these data indicate that FM patients have a neurovascular pathology in the palms of the hands predominantly increasing the peptidergic sensory innervation of the AVS.
Although Aδ fibers are not disproportionately represented in the AVS of FM patients, they are excessive compared with those in the AVS of control subjects and, therefore, may contribute to the symptoms of fast pain pressure allodynia elicited from the palms of FM patients. Furthermore, heat thermal functionality is presumably among the peptidergic C fibers which are known to express TrpV1, and prolonged exposure to capsaicin diminishes heat-induced vasodilatory and consequent inflammatory responses [18, 44, 88]. Thus, the increased peptidergic C-fiber innervation among the FM patient AVS could be a source of the observed heat hyperalgesia, although the relative delay and windup of the heat hyperalgesia has been interpreted as an indicator of central sensitization [10, 16, 17, 39, 74-76]. The central terminals of the excessive peptidergic innervation of the AVS may contribute to abnormally high levels of vasodilatory peptides previously documented in the CSF of FM patients .
Of particular note, relative to the excess sensory innervation of the AVS, is the extreme sensitivity of the FM patients to cold stimuli and the exacerbation of FM symptoms especially by cold weather [36, 41, 42, 68, 69, 103-105]. While heat stimulation has most often been used for pain assessments of palmar skin in FM patients, heat hyperalgesia elicited by thermal testing of palmar skin has been detected in only a subset of FM patients, whereas cold pain hyperalgesia is more widely prevalent among FM patients, including those with heat hyperalgesia . Moreover, unlike heat, cold stimulation not only elicits pain but also is more likely to be associated with dysesthesias such as “pins and needles” and paradoxical heat pain. Although the neural regulation of blood flow, especially the mechanism of vasodilation in glabrous skin, is not well understood, there is evidence that sympathetically mediated vasoconstriction of arterioles is coupled with an independent vasodilation of AVS involving CGRP and SP release from the sensory innervation under cold external conditions in order to reduce heat loss [65, 91, 92, 97, 106-108]. It is exclusively the AVS sensory innervation that appears excessive in the FM patients and may account for the increasing severity of FM symptoms under cold conditions.
Dysregulation of blood flow in the glabrous skin in humans, as a consequence of a “positive” neuropathy and AVS dysfunction, has the potential to compromise regulation of blood flow in other areas of the body. Blood flow and temperature of the skin and skeletal muscle are interdependent with inverse relationships occurring under a variety of metabolic conditions . Thus, in humans, the relatively limited glabrous skin of the hands and feet and in some areas of the face and head such as the checks, nose, and ears, all specifically contain AVS, and have been shown to play a disproportionately major role in thermoregulation, given the high surface to volume ratio of these areas of the body [27, 32, 66, 92, 93, 107, 110-113]. Under heat stress conditions, an estimated 60% of cardiac output is distributed to the skin , and a particularly high proportion of this blood flow is distributed to the glabrous skin, as indicated by an estimated 6 to 1 ratio of blood flow to the palmar glabrous vs the dorsal hairy skin of the hands [95, 113]. Even under moderate thermal conditions, the glabrous hand and foot skin receives a blood flow volume that far exceeds the metabolic demand, and can also withstand prolonged periods of extremely low blood flow without severe consequences. Therefore, the glabrous skin also functions as a strategic reserve that can be diverted to other tissues during transient phases of high metabolic demand. Thus, blood flow to the palms of the hands is not only reduced under cold conditions to conserve heat, but is also reduced during the onset of exercise to increase the blood supply to skeletal muscle. When prolonged exercise begins to increase core temperature, blood flow becomes rerouted especially to the glabrous skin to dissipate heat. Conceivably, the neuropathology of the AVS could result in insufficient blood flow and ischemia in deep tissues like skeletal muscles, which may contribute to the widespread deep pain and fatigue of FM, and cause compromised circadian blood flow regulation thereby impacting sleep and cognition [28, 114-118].
Of potential importance to the AVS neuropathology in FM, is the particularly high occurrence of diagnosis among females, possibly pointing to a hormonal role in impacting vascular innervation, which is manifested by well-documented gender differences in normal vasoregulation [59-61, 118-120]. Our data on gender differences among glabrous skin innervation in humans indicate that females normally have twice the sensory innervation to arterioles and AVS as males, whereas the sympathetic innervation is similar. Small-caliber neurovascular sensory innervation appears to be estrogen-dependent , which potentially underlies these gender differences and may predispose females to developing the excessive sensory innervation of the AVS observed in the FM patients. Studies are underway to determine whether a similar AVS neuropathology also occurs in male FM.
The authors wish to thank Mrs. Margaret Czerwinski, RN and Mrs. Marilyn Dockum, BS for biopsy collection and tissue processing, and Dr. Dean Kellogg for critical comments and assistance in interpreting and integrating vascular literature. The authors of this study have been funded, in part, through investigator-initiated grant support from Eli Lilly and Forest Laboratories to conduct basic research on FM. This pathology investigation is not intended to be reported to the FDA as a well-controlled human study in support of a new indication for any marketed drug, nor is this study intended to support any significant change in current FM product labeling or serve as advertising for any currently marketed drug. This research study provided strict institutional review and informed consent. The authors declare no perceived or real conflicts of interest in performing this research.