Obesity-Associated Neuropathy: Recent Preclinical Studies and Proposed Mechanisms

Abstract

Significance:

The prevalence of metabolic syndrome (MetS) and associated obesity has increased in recent years, affecting millions worldwide. One of the most common complications of obesity is damage to the peripheral nerve system, referred to as neuropathy. The lack of disease-modifying therapy for this complication is largely due to a poor understanding of the complex neurobiology underlying neuropathy. Recent preclinical studies suggest that in addition to glucotoxic events, other mechanisms, including lipid signaling, microbiome, or inflammation, may be viable targets to prevent nerve damage and neuropathic pain in obesity.

Recent Advances:

Clinical and preclinical studies using diet-induced obesity rodent models have identified novel interventions that improve neuropathy. Notably, mechanistic studies suggest that lipid, calcium signaling, and inflammation are converging pathways.

Critical Issues:

In this review, we focus on interventions and their mechanisms that are shown to ameliorate neuropathy in MetS obese models, including: (i) inhibition of a sensory neuron population, (ii), modification of dietary components, (iii) activation of nuclear and mitochondrial lipid pathways, (iv) exercise, and (v) modulation of gut microbiome composition and their metabolites.

Future Directions:

These past years, novel research increased our knowledge about neuropathy in obesity and discovered the involvement of nonglucose signaling. More studies are necessary to uncover the interplay between complex metabolic pathways in the peripheral nerve system of obese individuals. Further mechanistic studies in preclinical models and humans are crucial to create single- or multitarget interventions for this complex disease implying complex metabolic phenotyping. Antioxid. Redox Signal. 37, 597–612.

Introduction

Metabolic syndrome

According to the National Cholesterol Education Program Adult Treatment Panel III from the National Heart, Lung, and Blood Institute, metabolic syndrome (MetS) comprises the following: (i) abdominal obesity (defined as waist circumference >35 inches for men and 40 inches for women); (ii) dyslipidemia (high triglycerides and low high-density lipoprotein [HDL]); (iii) hypertension; (iv) insulin resistance; (v) proinflammatory state; and (vi) prothrombotic state (Fig. 1) (49). In 2017, 22% of adults in the United States met the MetS criteria (114) and its prevalence has been increasing in the past years.

FIG. 1.

Metabolic syndrome. To be diagnosed with MetS, a patient must present with three or more of the following or currently be under treatment for them: (i) insulin resistance (fasting glucose ≥100 mg/dL), (ii) hypertension (blood pressure ≥130/85 mmHg), (iii) hypertriglyceridemia (≥150 mg/dL), (iv) low HDL (<40 mg/dL in men; <50 mg/dL in women), and (v) abdominal obesity (waist circumference ≥102 cm in men; >88 cm in women). Obesity is defined as a BMI >30 kg/m² in both men and women. Individuals with MetS tend to persist in a prothrombotic and proinflammatory state that leaves them at high risk of a wide range of complications, including but not limited to cardiovascular disease, type 2 diabetes, nonalcoholic steatohepatitis, and portal vein thrombosis. As shown, obesity and dyslipidemia have been associated with neuropathy. BMI, body mass index; HDL, high-density lipoprotein; MetS, metabolic syndrome.

This increase has been accompanied by a rise in analgesic use, mostly due to a high prevalence of chronic pain and neuropathies reported in obese individuals (116). Current chronic pain therapies mainly focus on alleviating the symptoms and are not disease-modifying therapies. Anticonvulsants (pregabalin and gabapentin), serotonin and noradrenaline reuptake inhibitors, and tricyclic antidepressants are among the most common recommendations for pain management in obesity (34). Of note, although obesity and MetS are very often linked, several subsets of lean people may meet the criteria for MetS.

For example, recent studies demonstrated a prevalence of MetS of 61.6% in the obese group, 33.2% in the overweight group, and 8.6% in the normal-weight group (103). However, obesity, dyslipidemia, and insulin/glucose signaling are often comorbidities and drive neuropathy in MetS subjects. Recent studies discussed in this review revealed that MetS may drive neuropathy progression by modifying novel pathways that are not directly linked to glucose pathways, but rather lipid.

MetS-associated neuropathy and pain

MetS-associated neuropathy

Recent investigations reveal that obesity, diabetes, and dyslipidemia are the factors that correlate with peripheral neuropathy progression in MetS subjects (19, 21, 23, 26, 111, 112). Neuropathy affects 10%–40% of people with obesity, representing a large portion of the American population. In peripheral neuropathy, autonomic, motor, and sensory peripheral nerves are damaged (5). Many peripheral neuropathies have been identified to date and are classified according to the number of nerves affected: (i) mononeuropathy—only one nerve is damaged, (ii) mononeuropathy multiplex—two or more nerves, and (iii) polyneuropathy—when many nerves are injured (52).

The symptoms vary according to the type of nerve affected: neuropathy of autonomic nerves typically leads to innervated organ dysfunction, that is, urinary incontinence or cardiac arrhythmias, while motor neuropathies result in muscle weakness and atrophy. Sensory neuropathies are known to cause neuropathic pain or numbness (5, 52). In addition, neuropathies can be either acute or chronic depending on the disease time course (52). Among the chronic polyneuropathies, diabetic polyneuropathy is the most prevalent as it accounts for up to 50% of all polyneuropathies (53). Diabetic neuropathy is a direct consequence of chronic diabetes in 50% of cases and commonly presents as distal symmetric polyneuropathy (34). Twenty to 60% of affected individuals report neuropathic pain as one of the main symptoms (17, 53).

Current methods in diagnosis reveal damage of multiple fiber types

Diagnosis of neuropathy is usually based on standardized examination and questionnaires. Due to its complex etiology and unclear pathophysiology, MetS-associated neuropathy is a diagnosis of exclusion (93, 97). Initially, the medical history of the patient and pain distribution is considered. Then, quantitative sensory tests are performed, followed by confirmatory tests when appropriate (97).

During bedside examinations, neuropathy is assessed by clinician-administered questionnaires, such as the Utah Early Neuropathy Scale (105) and Michigan Neuropathy Screening Instrument (MNSI) (36), in combination with a physical examination. Patients undergo testing to evaluate sensory loss, vibration sensation, and pain, among other features (97). Lastly, confirmatory tests are applied to assess neurological dysfunction.

These tests are rarely used in clinical settings because of their costs and because their sensitivity and reproducibility are debated within clinicians, however, they are valuable tools for clinical research (34). The two most used tests in clinics are as follows: (i) nerve conductance studies (NCS) and (ii) intraepidermal nerve fiber (IENF) density (17, 34, 93). NCS are used to evaluate abnormalities in large motor and sensory fibers. A stimulus is applied and the response amplitude and conduction velocity, among other functions, are recorded.

Patients with neuropathy usually demonstrate lower amplitude, and slower conduction velocities (5, 17, 34). However, some patients present with normal NCS, in which case the small fibers' presence needs to be investigated. The gold standard assessment of small-fiber neuropathy is IENF density analysis by skin biopsy (5, 34, 73, 117), and patients with diabetic neuropathy generally show loss of intraepidermal fibers. Small-fiber damage is also examined by quantitative thermal sensory tests that are less invasive than skin biopsies and offer reliable assessment of fiber function.

MetS-associated neuropathy, glycemia, and lipid profiles

Interestingly, ∼40% of peripheral neuropathies are idiopathic. Idiopathic polyneuropathies are linked to a higher risk of prediabetes and MetS (5). A large body of evidence indicates that components of MetS may be linked to peripheral neuropathy onset and progression, including metabolic and complex lipid pathways (16, 20, 26, 34, 87, 111, 112, 114). For example, a recent study using lipidomics and metabolomics found that specific subclasses of metabolites distinguished obese individuals by peripheral neuropathy status independent of the glycemia (98).

A recent study from Callaghan et al. showed that people with obesity, including normoglycemic participants, had a higher prevalence of neuropathy when compared with lean control individuals (24). Moreover, the obese participants from this study had severe indicators of neuropathy, as observed by alterations in NCS and IENF analysis (24). Obesity preferentially injures small- and medium-sized nerve fibers as is also observed in obese rodent models (18, 21). Alternative mechanisms, in addition to hyperglycemia, are suspected to be involved in peripheral neuropathy progression (29, 44, 62).

Dyslipidemia is a specific component of MetS to highly correlate with peripheral neuropathy (45). Dyslipidemia can be characterized as elevated blood levels of triglycerides (higher than 150 mg/dL) and altered levels of circulating lipoproteins, such as increased low-density lipoprotein (LDL) and decreased HDL (<50 mg/dL for women and <40 mg/dL for men) (Fig. 1) (45, 49, 50). Although some studies have demonstrated an association between LDL/HDL levels and neuropathy (45), others have failed to replicate these results (22, 24). Similarly, elevated blood triglycerides have been linked to neuropathy (45). Specifically, a particular study demonstrated a correlation between fiber loss and hypertriglyceridemia (121), while other groups failed to demonstrate the same association between neuropathy and high triglycerides (22, 24).

A recent report by Fridman et al. demonstrated that atypical 1-deoxy-sphingolipids may be linked to and serve as biomarkers of peripheral neuropathy in obese individuals with or without type 2 diabetes mellitus (T2DM) (40). This study found that 1-deoxy-dehydroceramide levels were higher in obese diabetic individuals with neuropathy when compared with healthy lean controls (40). Importantly, there was a negative correlation between leg IENF density and 1-deoxy-dehydroceramides in people with obesity, irrespective of glycemic levels. Causes of neuropathy likely engage multiple cross-talking pathways involving glycemia and lipids in MetS since dyslipidemia and hyperglycemia are frequent comorbidities.

MetS-associated pain

Pain, recently redefined as “an unpleasant sensory and emotional experience associated with, or resembling that associated with, actual or potential tissue damage” by the International Association for the Study of Pain (IASP) (94), is a critical conserved protection mechanism that aims at evading tissue damage, inflammation, and disease. However, acute pain can become maladaptive due to central or peripheral sensitization and lead to chronic pain states (94, 100). Nociceptive pain is defined by IASP as “pain that arises from actual or threatened damage to non-neural tissue and is due to the activation of nociceptors.” On the contrary, neuropathic pain is “pain caused by a lesion or disease of the somatosensory nervous system.”

These definitions are contrasted by the fact that nociceptive pain indicates a normally functioning somatosensory system, while neuropathic pain denotes abnormal activity in the nociceptive pathway that is unrelated to nociceptor stimulation (94).

Obesity is known to be one of the risk factors in some nociceptive pain conditions (e.g., migraine). For example, trigeminal nerve activation causes the release of neuropeptides such as calcitonin gene-related peptide (CGRP) and induces vasodilation resulting in local immune cell recruitment and pain, which is aggravated by obesity (9).

Animal studies have demonstrated aberrant ion channel function in obesity and type II diabetic neuropathy. Voltage-gated sodium channels (VGSCs), shaker-type potassium channels (Kv), and T-type calcium channels, for instance, have been implicated in the disease pathophysiology (34, 72). Increased expression of VGSCs can lead to increased nociceptor excitability and exacerbated pain (34). Diminished expression of Kv in myelinated neurons also causes hyperexcitability (34). Pharmacological inhibition of calcium channels Ca_v3.2 ameliorated neuropathic pain (72).

While few recent findings suggest that obesity is a risk factor for neuropathic pain (28, 58), the relationship between neuropathic pain and obesity is not clear. Electrophysiological examinations demonstrated that sensory action potentials are impaired in obese subjects (28). In one recent study, people with obesity showed high prevalence and severity of neuropathic pain despite normal mental health status (58).

Murine Preclinical Studies

Animal models

Animal models can offer insight about disease onset and progression; Biessels et al. published an in-depth review of the animal models suitable for studying diabetic and obesity-induced neuropathy (8).

Some genetic models represent models of obesity and include the leptin receptor-deficient mouse (db/db) and the leptin0deficient mouse (ob/ob). These animals develop insulin resistance and glucose intolerance alongside impaired nerve conduction velocity, decreased cutaneous nerve fiber densities, and altered pain behavior, such as thermal hypoalgesia starting at 12 weeks of age and tactile allodynia at 21 weeks of age (8, 89).

These models have been pivotal to build our understanding in the development and progression of diabetic peripheral neuropathy (89), they are also cost effective, and their neuropathy phenotype is also robust. However, the leptin receptor is broadly expressed, including in the peripheral nervous system (PNS), and leptin pathways are altered in the nervous systems of these genetic models of obesity. Furthermore, human MetS and associated obesity are rarely linked to genetic mutation in leptin pathways, thus confounding interpretation. Animal models of MetS using special diets have been increasingly used in neuropathy studies.

Diet-induced obesity (DIO) animal models mainly include a high-fat diet (HFD 45%–60% of calories from fat) as reviewed in O'Brien et al. (89) and a Western diet (WD) that has the particularity of containing high fat, and also high carbohydrates and cholesterol mirroring the nutrient content of westernized countries (Fig. 2).

FIG. 2.

WD-fed mouse model, metabolism, and behaviors. WD-fed mice consume a diet composed of 42% fat, 34% sucrose, and 0.2% cholesterol (Envigo TD88137). Compared with regular/normal chow-fed mice, these animals display increased body weight, glucose intolerance, insulin resistance, hypertriglyceridemia, and hypercholesterolemia. After 6–10 weeks of feeding, they present reduced intraepidermal nerve fiber density, mechanical allodynia, as well as hyperalgesia or hypoalgesia in response to thermal stimulation (results vary between laboratories). WD, Western diet.

The WD (42% fat, 0.2% cholesterol, and 34% sucrose)-fed mouse model developed MetS features and obesity (insulin resistance, glucose intolerance, and dyslipidemia) and exhibited indices of neuropathic pain (10, 41–43, 64, 79). The mice are defined below as obesity models with MetS indices since mice do not develop MetS or diabetes per se. Early thermal hyperalgesia that progresses to hypoalgesia, tactile allodynia, and reduced nerve conduction velocities are all early hallmarks of WD-fed mice. Furthermore, WD-fed mice display a reduction of IENF density (10, 64), indicating the loss of small sensory nerve fiber terminals, as detailed in the studies below.

Preclinical studies, novel targets, and mechanisms

Exercise

As detailed above, exercise and weight management are the two lifestyle interventions that have proven to be the most efficient at delaying and/or improving neuropathy in individuals with prediabetes (110). Wright's group demonstrated the beneficial effects of exercise in an animal model of obesity-induced neuropathy (46). Animals fed an HFD, but allowed to exercise, showed improvement in mechanical hypersensitivity when compared with an HFD control group (46). In addition to behavioral changes, exercise also led to axonal phenotypic alterations; HFD-fed mice that exercised did not display an increase in nociceptive peptidergic fibers known to be involved in neuropathic pain observed in nonexercised obese mice (46).

Dietary reversal

In a study by Hinder et al., HFD-fed mice were subjected to a dietary reversal paradigm to investigate neuropathic indices (55). The dietary reversal paradigm consisted of animals being fed normal chow (NC) for 4 weeks following 16 weeks on HFD. Animals that underwent dietary reversal showed an improvement in metabolic parameters, such as decreased body weight and enhanced insulin sensitivity, concurrent with normalized neuropathy as observed by restored sciatic and sural nerve function (55).

Ketogenic diet

Another study done by Cooper et al. evaluated the effects of a ketogenic diet on PNS function (25). Feeding mice a ketogenic diet for 12 weeks prevented the development of mechanical allodynia seen in sedentary HFD-fed animals, despite leading to an increase in energy intake when compared with NC and HFD sedentary controls (25). Furthermore, the authors found that 4 weeks of ketogenic diet following 8 weeks on HFD reversed mechanical allodynia phenotype, in addition to increasing IENF density (25). Lastly, ex vivo and in vitro experiments revealed an increase in neurite outgrowth when dorsal root ganglia (DRG) neurons were (i) cultured from ketogenic diet-fed mice or (ii) supplemented with ketone bodies (25), suggesting that ketone bodies may act directly on sensory neurons to promote axonal regrowth.

Monounsaturated fatty acid-rich diet

As mentioned above, human data have indicated that dyslipidemia is one of the main drivers of peripheral neuropathy in obese individuals (45, 121). Several groups have thus focused on the role of dietary lipids as one of the underlying causes of neuropathy and neuropathic pain in prediabetes, type 1 diabetes mellitus, and T2DM (16, 34). O'Brien et al. identified elevated triglycerides derived from saturated fatty acids in the nerves of DIO mice (86). The animals fed an HFD also developed neuropathy indices, as previously shown (55, 86).

In another study, Rumora et al. demonstrated that HFD-induced neuropathy can be reversed by a monounsaturated fatty acid (MUFA)-rich diet (99). The authors determined that different fatty acids had specific effects on neuropathy indices, DRG mitochondria motility and function, as well as in neuronal apoptosis and lipid droplet formation (99). The two main findings of the study were that 8 weeks of MUFA, following 16 weeks on HFD (i) restored sural and sciatic nerve function and (ii) increased IENF density. In addition to the functional and structural analyses, this study provided evidence that a saturated fatty acid, namely palmitate, altered axonal mitochondria motility in DRG neurons, which is rescued by an MUFA, oleate. Lastly, oleate prevented apoptosis in DRG neurons previously exposed to palmitate, likely by rescuing adenosine triphosphate (ATP) levels (99).

These data suggest that neuropathic features, such as nerve function, can be restored by a diet rich in MUFA, such as oleate, in lieu of an HFD rich in saturated fatty acid (99).

A recent study by Boyd et al. showed that mice fed an HFD rich in omega-6 polyunsaturated fatty acids (PUFAs) displayed a nociceptive hypersensitivity phenotype (14). Mechanistically, the authors showed that an omega-6 PUFA diet loaded the plasma membrane of lumbar DRG neurons with linoleic and arachidonic acids—precursors of proinflammatory mediators, such as prostaglandins, via phospholipase cleavage (14). Similar to pharmacological inhibition of phospholipase, dietary reversal to a diet rich in omega-3 PUFAs ameliorated neuropathic pain induced by omega-6 PUFAs (14), as observed by improved thermal thresholds and IENF densities (14). These data implicate omega-6 PUFAs as an underlying cause of neuropathy in obese settings.

Fecal microbiome transplantation

Recent publications and our own data indicate that gut microbiota, potentially butyrate producers, may ameliorate neuropathic pain in obesity (10). It was hypothesized that modulation of gut microbiome could play a role in neuropathy and neuropathic pain in WD-fed mice. To test this hypothesis, WD-fed mice were subjected to antibiotic treatment, followed by a transplantation of feces from lean mice. Improvement in mechanical allodynia and thermal hyperalgesia thresholds was observed.

These changes were accompanied by alterations in transcripts and pathways involved in Ca²⁺ handling, pain behavior, and immune response. Upregulated levels of Ryanodine receptor 2 (Ryr2) mRNA levels were detected in DRGs of WD-fed mice, but interestingly, downregulated after fecal microbiome transplantation (FMT). It was observed that circulating butyrate correlated with improved mechanical allodynia, decreased expression of inflammatory markers in the DRGs of obese mice, and transcriptional changes via upregulating histone deacetylase 2 (HDAC2) mRNA (10).

Data published suggest that microbiome reshaping may alleviate pain by changing circulating short-chain fatty acids (SCFAs; e.g., butyrate). Some SCFAs have been demonstrated to cross the blood/brain barrier in rats (90) and in humans (4), possibly due to high expression of their transporter, monocarboxylate transporters (MCTs) in blood cells and astrocytes (4, 27). MCTs are also expressed in the PNS, especially MCT1 in DRGs, Schwann cells (SC), and perineurial cells (35, 83). In a model of crush injury, MCT1-deficient mice showed delayed nerve regeneration, providing evidence that SCFAs, which can be used by and shuttled between cells of the PNS, are essential for axonal regrowth (83).

In cells, SCFAs can be used as fuel, but can also act as signaling molecules through two main mechanisms: (i) by directly activating G-coupled protein receptors (GPCRs) and (ii) by inhibiting HDACs. Free fatty acid receptors 2 and 3 (FFAR2 and FFAR3, respectively) have been the two-most well-studied SCFA receptors and they are largely expressed throughout the body. Lastly, FFAR2 and FFAR3 can heterodimerize and produce a completely distinct signaling pathway (69). The main features of the heterodimer are increased intracellular Ca²⁺ ([Ca²⁺]_i) and decreased cyclic adenosine monophosphate (cAMP) generation (69). Both [Ca²⁺]_i and cAMP are important second messengers that can be involved in peripheral neuropathy (38). In addition to SCFAs, ketone bodies have been shown to activate FFARs.

Kimura et al. demonstrated that β-hydroxybutyrate inhibited FFAR3 in sympathetic neurons (69). Overall, data suggest that SCFAs may impact either [Ca²⁺]_i balance, possibly via FFAR, or PNS inflammation by acting directly on resident immune cells or via neuroimmune communication.

SCFAs are known to regulate gene transcription via their HDAC inhibition activity. HDACs are enzymes that catalyze the removal of acetyl groups bound to N-tails of histones. Epigenetic mechanisms likely regulate pain sensitization since they can be long-lasting modifications (85). Thus, this epigenetic regulatory function of SCFAs may be crucial in preventing and/or delaying MetS-associated neuropathy and neuropathic pain (Fig. 3).

FIG. 3.

Butyrate may improve neuropathy via pleiotropic mechanisms. Fecal transplantation from lean to obese mice increases circulating short-chain fatty acids including butyrate. A report suggests that butyrate may be the gut-derived metabolite that improves neuropathy in obesity. Beneficial interventions, including exercise and dietary modifications, for instance, a ketogenic diet or a diet rich in fibers and monosaturated fatty acids, are known to modulate gut microbiome composition and may also have their beneficial effect through butyrate. Based on current literature, the proposed pathways are (i) a direct effect of butyrate on its G-coupled receptors, FFARs located at the surface of sensory neurons, and/or immune cells of the PNS; (ii) via HDAC inhibition; and/or (iii) by cellular import through MCT1. By acting through FFARs, butyrate can alter intracellular calcium levels, since these receptors are GPCRs linked to G_αi and/or G_αq. If linked to G_αi, butyrate binding to FFARs can inhibit AC and thus decrease cAMP production from ATP. Differently, when linked to G_αq, butyrate leads to the formation of IP₃ and consequently the release of calcium from ER. Butyrate can also modify VCCG (Cav) activity via FFAR coupling. Butyrate can also lead to transcriptional changes by inhibiting HDACs; this inhibitory function alters chromatin structure, allowing for enhanced gene transcription. Lastly, butyrate can be transported into neurons, immune cells, and SC, via MCT1, and therefore be used as metabolites, modifying mitochondrial activity. AC, adenylyl cyclase; ATP, adenosine triphosphate; cAMP, cyclic adenosine monophosphate; ER, endoplasmic reticulum; FFARs, free fatty acid receptors; GPCRs, G-coupled protein receptors; HDAC, histone deacetylase; IP₃, inositol 1,4,5-triphosphate; MCT1, monocarboxylate transporter 1; PNS, peripheral nervous system; SC, Schwann cells; VCCG, voltage-gated calcium channels.

Silencing the DRG neurons expressing Na_v1.8

Tetrodotoxin-resistant Na_v1.8 channels are members of the VGSC family. They are selectively expressed on the surface of sensory neurons within the DRG and trigeminal and nodose ganglia, and characterize more than 85% of nociceptors (51, 115). Given the electrophysiology channel properties, such as the slow rate of inactivation, neurons expressing Na_v1.8 produce action potential of long durations, with Na_v1.8 currents contributing to neuronal excitability (6, 51) (Fig. 4).

FIG. 4.

Recent beneficial interventions and proposed mechanisms. WD is concomitant with alteration in (i) neuronal hyperexcitability/calcium signaling (Ca²⁺), (ii) ER stress, (iii) nuclear receptor activation (LXR or PPAR), (iv) impaired mitochondrial function and motility, (v) disruption of neuron-SC communication, (vi) altered immune/neuron communication, and (vii) ion channel dysfunction. Inhibition of Na_v1.8 nociceptors leads to decreased neuronal excitability and thus amelioration of neuropathic pain. Inhibition of CXCR4/SDF-1 signaling prevented neuropathy indices and decreased neuronal calcium level. Activation of LXRs and PPARγ pathways using agonists has been shown to increase the expression of target genes listed and to improve mechanical allodynia. LXR activation (i) decreases ER stress, (ii) may modify neuron/SC communication by increasing interaction between NRG1 in sensory neurons with epidermal growth factor receptors (ErbB) located on SC, and (iii) may alter lipid membrane composition via lpcat3/abca1 regulation. Monosaturated fatty acids, in lieu of saturated ones, restore nerve function likely by improving the mitochondrial function and motility along the peripheral process. RXR, abca1, lpcat3, cd36, VCCG (Ca_v), TRPV. abca1, ATP-binding cassette subfamily A member 1; cd36, cluster of differentiation 36; CXCR4, C-X-C motif chemokine receptor 4; lpcat3, lysophosphatidylcholine acyltransferase 3; LXRs, liver X receptors; NRG1, neuregulin 1 type III; PPAR, peroxisome proliferator-activated receptor; RXR, retinoid X receptor; SDF-1, stromal cell-derived factor 1; TRPV1, transient receptor potential cation channel subfamily V member 1.

Neuropathy and neuropathic pain have been shown in both, humans and animal models, to be driven by neuronal hyperexcitability (3, 7, 91). Moreover, Na_v1.8 channel mutations are found in patients experiencing peripheral neuropathy and have been linked to a hyperexcitable phenotype of DRG neurons in humans and rodents (33, 51). Na_v1.8 nociceptors represent 75% of DRG neuronal population (64, 104), including more than 90% of nociceptors, and some Aβ and C-mechanoreceptors (104).

Menichella et al. showed that WD-fed mice developed neuropathic pain beginning at 6 weeks of WD feeding as observed by the lower mechanical paw withdrawal thresholds when compared with NC-fed animals (64, 81). Interestingly, small-fiber degeneration is preceded by mechanical allodynia and starts after 10 weeks of WD (64). The authors monitored [Ca²⁺]_i concentrations specifically in Na_v1.8 nociceptors by utilizing a Ca²⁺ indicator mouse line, GCaMP (64). Briefly, Ca²⁺ transients were measured in response to capsaicin and potassium in DRG explants from mice fed the WD for 2 or 8 weeks.

Ca²⁺ imaging revealed that Na_v1.8 sensory neurons from mice fed the WD for 8 weeks responded to lower concentrations of capsaicin or high potassium when compared with control or mice fed the WD for only 2 weeks, demonstrating DRG neuronal hyperexcitability with prolonged WD feeding (64). The authors then hypothesized that decreasing neuronal hyperexcitability in Na_v1.8 nociceptors could prevent neuropathic pain and IENF loss. To test their hypothesis, they chemogenetically inhibited Na_v1.8 nociceptors using designer receptors exclusively activated by designer drugs (DREADDs) (64).

Remarkably, when Na_v1.8 nociceptors were acutely chemogenetically silenced, mechanical allodynia was reversed in WD-fed mice, with thresholds returning to baseline 4 h after nociceptors' inhibition (64). Prolonged Na_v1.8 DRG neuron inhibition led to the prevention and reversal of IENF loss observed in WD-fed mice (64).

Lastly, the authors tested whether Na_v1.8 inhibition could reverse neuropathy indices after they were established. To test this hypothesis, they inhibited Na_v1.8 sensory neurons following 10 weeks of WD when mice already displayed mechanical allodynia and IENF loss. Four weeks of inhibition completely reversed the behavioral and fiber loss phenotypes observed in obese mice (64). They also proceeded with chemogenetic activation of GPCRs in Nav1.8-positive neurons and observed an increase in neuronal excitability. These results indicated that Na_v1.8 DRG neuronal hyperexcitability is a probable cause of mechanical allodynia and skin fiber degeneration in WD-fed mice.

This is an important study identifying a neuronal population and suggesting that modulating Ca²⁺ homeostasis using GPCRs is a potentially new avenue to alleviate or delay pain and neuropathy (64).

Nuclear lipid receptor activation

Liver X receptor

Transcriptomic analyses revealed that dysregulation of genes associated with lipid metabolism and adipogenesis is conserved in humans and obese mouse models of T2DM, consistent with the potential involvement of lipids in neuropathy (60, 66, 78, 86, 92). To investigate shared pathways between humans and mice, Hur's group used a microarray approach combined with the development of a transcriptional network using nerve samples from patients with diabetic neuropathy and from rodent models of type 2 diabetic neuropathy. Several differentially regulated genes were identified, and centrality analysis revealed liver X receptor/retinoid X receptor (LXR/RXR) activation to be the main conserved pathway (78).

Alterations of LXR/RXR transcription levels seemed to occur differently in genetic models of T2DM. In earlier stages, LXR/RXR was shown to be downregulated in db/db mice, while at later stages of neuropathy, this pathway was upregulated. ob/ob showed consistent upregulation throughout the disease course (60, 66, 78, 86, 92). Nonetheless, the fact that this specific pathway is dysregulated across human samples and murine models indicated a potential involvement of LXR/RXR in peripheral neuropathy. LXR was deleted from sensory neurons and exacerbated allodynia in obesity models (42).

This study also showed that LXR activation using GW3965 treatment (after 10 weeks of WD [early allodynia]) decreased endoplasmic reticulum (ER) stress in the sensory neurons of WD-fed mice and delayed the development of allodynia (42).

LXRs are nuclear receptors known to play a major role in cholesterol efflux, lipid metabolism, and inflammation.

Many cellular stressors, including dyslipidemia and saturated fatty acids, lead to ER overload with unfolded and misfolded proteins and trigger the unfolded protein response (UPR) (61, 88, 96). During UPR, transcripts and proteins involved in (i) decreasing protein shuttling to ER, (ii) increasing chaperones to improve protein folding, and (iii) increasing protein degradation is upregulated to restore ER function (61, 88). Activation of LXR canonical genes by agonist treatment (GW3965) decreases the expression of transcripts involved in the UPR, such as CHOP and ATF3, in neurons treated with palmitate (42).

Another study showed that LXRs may ameliorate obesity-induced neuropathy through a neuron-SC communication mechanism (41).

Finally, multiple laboratories identified that LXR regulates lysophosphatidylcholine acyltransferase 3 (LPCAT3) in addition to ATP-binding cassette subfamily A member 1 (ABCA1) in many cell types including neurons (42). Interestingly, PUFAs in phospholipids greatly affect the physical properties of membranes. It has been shown that LPCAT3 regulates membrane arachidonate levels and membrane properties. ABCA1 also regulates cholesterol efflux and impacts membrane composition and lipid raft contents (54, 80).

Interestingly, recent studies discovered that manipulating lipid raft or membrane cholesterol represented a novel approach to alleviate pain (2, 57). It would be important to evaluate whether LXR activation dynamically regulates plasma membrane phospholipids and cholesterol composition in the PNS of obese mice.

LXR activation is concomitant with liver production of triglycerides, and thus, research developing the PNS delivery system of LXR agonist would be needed to its activation in clinical studies.

Peroxisome proliferator-activated receptor

Peroxisome proliferator-activated receptors (PPARs) are nuclear transcription factors that, similar to LXRs, heterodimerize with RXR, bind to PPAR-responsive element, and activate downstream target genes. There are three PPAR isoforms, PPARα, PPARβ, and PPARγ. They are widely yet distinctly expressed and are known to regulate metabolism and homeostasis (1). Due to alternative splicing, PPARγ is present in two isoforms, PPARγ1 and PPARγ2, and differs in tissue expression: PPARγ1 is expressed throughout the body and PPARγ2 is mainly present in adipocytes (1). PPARγ is a major lipid metabolism regulator, but also has tissue-specific functions, for example, by modulating macrophage polarization into M1 or M2 phenotypes (1).

In the PNS, PPARγ has been implicated in neuropathy progression in patients with T2DM (66). In a study led by Junguk Hur et al., human sural nerve analyses revealed that PPARγ targets, including CD36, were downregulated in progressors when compared with nonprogressor individuals (66).

Feldman's group performed a preclinical study using pioglitazone, a PPARγ agonist (56). They observed that the use of pioglitazone had positive effects in enhancing peripheral neuropathy in leptin receptor-deficient (db/db) mice (56). The authors demonstrated that 11 weeks of pioglitazone treatment improved hind paw latencies, but did not alter sciatic or sural nerve conductance (56). These results suggest that pioglitazone prevented small-fiber dysfunction, without affecting large nerve fiber properties. However, pioglitazone improved glycemic levels. Thus, the phenotypic changes observed may be a result of improved hyperglycemia.

Calcium Imbalance and Neural Inflammation at the Crossroad of Preclinical Interventions

Calcium homeostasis

Intracellular Ca²⁺ homeostasis is a common altered pathway in most of the interventions mentioned above (Fig. 5). As previously detailed, we and others have published that DRG sensory neurons display ER stress, mitochondrial dysfunction, and ion channel dysregulation following WD feeding (42, 64, 79), all ultimately altering the Ca²⁺ cytoplasmic concentrations.

FIG. 5.

Intracellular calcium: intersection among altered pathways. Intracellular calcium ([Ca²⁺]_i) is shared between altered pathways in obesity-induced neuropathy and most of the beneficial interventions. For example, disruption of mitochondrial function and of lipid nuclear receptor expression and activity, as well as dysfunctional ER and channels (ion and GPCRs), converges to promote dysregulated [Ca²⁺]_i. Elevated [Ca²⁺]_i lead to neuronal hyperexcitability, and an increase in firing frequency and spontaneous firing. The rise in [Ca²⁺]_i can also alter the secretion of Ca²⁺-mediated vesicle release of neuropeptides associated with pain states (CGRP). Neuropeptide secretion is one of the main avenues for neuroimmune communication, and increases in CGRP levels within the PNS lead to immune cell recruitment, activation, and polarization. Voltage-gate calcium channel (Ca_v); voltage-gated sodium channel 1.8 (Na_v1.8); AC; cAMP; ATP; PKA; PI3K; SERCA2B; RYR2; VCCG; TRPM8; the TRPV1. CGRP, calcitonin gene-related peptide; PI3K, phosphoinositide 3-kinase; PKA, protein kinase A; RYR2, ryanodine receptor 2; SERCA2B, sarco/endoplasmic reticulum Ca²⁺-ATPase 2b; TRPM8, transient receptor potential cation channel subfamily M member 8.

Activation of Gα_i-coupled receptor downstream pathways in Na_v1.8 neurons using chemogenetic approaches reversed painful neuropathy of WD-fed mice (64). These data strongly suggest that activating Gα_i-coupled receptors (and thus decreasing Ca²⁺ levels) may represent an approach to ameliorate neuropathic pain in obesity.

While molecular mechanisms of Ca²⁺ balance are multifaceted, impaired ER Ca²⁺ handling plays a critical role in pain and neuropathy associated with type II diabetes and likely obesity (42, 59, 71, 118). It has been shown that cytosolic [Ca²⁺] ([Ca²⁺]_Cyt) is increased in sensory neurons isolated from rodents with diabetes (64), whereas ER Ca²⁺ load is significantly reduced. These changes in Ca²⁺ homeostasis were likely to be mediated by a decrease in ER Ca²⁺ pump and sarco/endoplasmic reticulum Ca²⁺-ATPase 2b (SERCA) expression (118), and an increase in transient receptor potential cation channel subfamily V member 1 (TRPV1) and M member 8 (TRPM8) channels' activity (38).

In a recent study, we also reported an alteration of Ca²⁺ homeostasis in sensory neurons isolated from mice DRGs after 14 weeks of WD (10). These alterations were associated with an increase in Ryr2-dependent Ca²⁺ release (10). In these experiments, Ca²⁺ transients were evoked by the Ryr2 agonist caffeine after culturing DRG sensory neurons. Moreover, WD led to a significant increase in Ryr2 expression (10). As a result of ER Ca²⁺ overload and Ryr2-mediated Ca²⁺ leak, resting [Ca²⁺]_Cyt was significantly increased in WD sensory neurons.

The discrepancy between our findings and the previously published work reviewed in Fernyhough and Calcutt (38) may reflect the different consequences of obesity versus diabetes. We also found that FMT of WD-fed mice restored ER Ca²⁺ load, Ryr2 expression, and Ryr2-mediated Ca²⁺ release in sensory neurons to a level of NC-fed mice (10).

GPCRs may also indirectly alter SERCA-mediated Ca²⁺ uptake and Ryr-mediated Ca²⁺ release. Due to their important role in Ca²⁺ signaling, SERCA and Ryr2 activity is tightly regulated. Phosphorylation regulates many ion transporters and channels, particularly Ryr2 (13, 65). Protein kinase A (PKA) downstream GPCRs may modulate Ryr2 by phosphorylating the channel at serine 2808 (S2808) (12). Activation of the cAMP-PKA pathway can cause ER Ca²⁺ overload, increasing ER Ca²⁺ leak and [Ca²⁺]_Cyt. Neuropathies, pain, and obesity were previously associated with increased PKA activity due to activation of adenylate cyclase (37, 74, 77, 119, 120).

Lastly, Ca²⁺ is known to mediate neuropeptide-containing vesicle release following neuronal depolarization (63, 67). Altered [Ca²⁺]_i could lead to an increase in neuropeptide secretion by sensory neurons in the spinal cord. Interestingly, obese subjects have been found to have elevated circulating levels of CGRP (95, 122). Immune cell recruitment to the PNS following peptide secretion by sensory neurons may initiate a cascade of events that contribute to proper nerve tissue repair in a nerve injury model of pain (76).

Inflammation

Previous reports have shown that immunoregulatory transcripts are differentially regulated in humans and in animal models of neuropathy (60, 66, 92). Elzinga et al. identified changes in Toll-like receptor (TLR) signaling pathway in the mentioned data sets and, subsequently, deleted TLR2/4 in mice (32). Whole-body knockout (KO) of TLR2/4 protected obese mice from the early thermal hypoalgesia seen in HFD-fed controls, but it did not affect IENF density or nerve conductance velocities (NCV) (32). This protection is only temporary, however, since TLR2/4 KO mice developed indices of neuropathy at later stages, such as worsening of thermal hypoalgesia and NCV, similar to HFD wild-type animals (32).

Activation of the proinflammatory pathway mediated by cyclooxygenase (COX) had been implicated in the pathophysiology of peripheral neuropathy in a model of diabetic neuropathy (68). Work by Kellogg et al. demonstrated that mice lacking COX2 were protected from streptozocin-induced neuropathic indices (68). COX2-deficient mice had improved nerve function and higher IENF densities when compared with wild-type diabetic littermates (68). These results suggest that COX2 is an important mediator of diabetic neuropathy that can be selectively blocked, resulting in attenuation of disease progression (68).

Remarkably, deletion of C-X-C motif chemokine receptor 4 (CXCR4) chemokine receptors reversed neuropathic pain in obese mice (as illustrated in Fig. 4). Menichella et al. demonstrated that loss of CXCR4 signaling in Na_v1.8 nociceptors prevented mechanical allodynia and IENF loss seen in obese mice (64). Furthermore, the authors identified that stromal cell-derived factor 1 (SDF-1) (also known as CXCL12 chemokine) was able to increase neuronal [Ca²⁺]_i in DRG cultures and Na_v1.8 sensory neurons from WD-fed animals (64, 81).

Our recent data suggest that immune cells may be implicated in early obesity-induced neuropathy (10). Immune cells' transcripts and pathways were differentially regulated in DRG and sciatic nerve (SN) between NC-fed, lean controls and WD-fed animals (10). In addition, obese animals subjected to FMT (transplantation of feces from lean mice) demonstrated improvement of mechanical allodynia, thermal hyperalgesia, and IENF when compared with obese controls, alongside differential immune cell profile within the PNS (10). FMT-treated obese animals had a higher percent of anti-inflammatory M2 macrophages within the DRGs, suggesting that this cell phenotype may underlie some of FMT-induced amelioration in neuropathic indices in obese mice (10).

While more investigations are necessary, data suggest that immune cells (resident or recruited) are involved in axonal regeneration and/or delay of skin fiber loss when obese mice were subjected to FMT (10).

Clinical Studies

Lifestyle interventions

A few clinical studies have shown that aerobic exercise ameliorates neuropathy indices in obese individuals. Singleton et al. demonstrated that weekly supervised aerobic training increased IENF densities when compared with control subjects (106). Moreover, patients who underwent the exercise paradigm did not show changes in body weight or other MetS components, suggesting that the benefits of exercise on neuropathy are unrelated to improvement in weight, glucose, and/or MetS (106). In another study by the same group, the authors evaluated small-fiber regeneration capacity induced by exercise following capsaicin axotomy in patients with MetS, with and without T2DM, and without diagnosed neuropathy (108).

Exercise, aimed at normalizing body mass index or decreasing initial body weight by 7%, improved the cutaneous regeneration rate measured by IENF density in patients with MetS (108). In addition, patients with exercise-induced improved fasting glucose and glycated hemoglobin observed higher rates of IENF regeneration following capsaicin axotomy (108).

A pilot study by Kluding et al. was one of the first reports to demonstrate the beneficial effects of exercise in diabetic neuropathy patients (70). Seventeen individuals with diagnosed diabetic neuropathy were subjected to a 10-week exercise paradigm, and outcomes were measure pre- and postintervention (70). Strength training and aerobic exercise improved severity of neuropathy symptoms according to the pain scale and MNSI questionnaire utilized, in addition to increasing IENF density in the skin (70).

Participants of the long-term weight loss clinical trial—Look AHEAD study—had improved diabetic neuropathy symptoms based on the MNSI assessment (47). The changes in pain scores were associated with weight loss, and improvements in glycated hemoglobin and serum lipidomics (47). However, exercise and low-calorie diets are known to positively regulate hypothalamic neurocircuitry of feeding behaviors, meaning that MetS subjects undergoing diet and exercise regimen increase their food intake and are, with time, resistant to body weight loss making this approach challenging (109, 110). Other beneficial interventions and their underlying mechanisms need to be further tested to identify new targets potentially disease modifying.

Bariatric surgery

Bariatric surgery is a promising strategy to treat obesity and complications (48). Achieving long-term weight loss via surgery has been recently shown as more effective than treatments and lifestyle modifications (102). A systematic review and meta-analysis showed that bariatric surgery improved microvascular complications of type II diabetes, encompassing diabetic neuropathy, nephropathy, and retinopathy (102). A prospective study revealed improved pain scores following gastric bypass by patients with type II diabetic neuropathy (113). Currently, an ongoing trial (NCT03617185) is assessing the effects of exercise and bariatric surgery on distal polyneuropathy (113).

Fecal microbiome transplantation

Recent evidence in humans and rodent models suggests that the gut microbiota play a critical role in several types of neuropathic pain such as chemotherapy-induced peripheral neuropathy and fibromyalgia (82, 101). MetS is known to modify gut microbes in humans and rodents (11, 30, 31, 39, 75, 84), including those responsible for the production of SCFA butyrate, acetate, and propionate. One recent clinical case found that FMT from a healthy lean donor to a person with obesity relieved early type 2 diabetic neuropathy, independent of glucose management (15). While many questions remain, this observation suggests that gut microbiota may be linked to diabetes-associated pain in humans.

Conclusion and Open Questions

Recent advances helped to better understand the molecular mechanisms underlying MetS-associated neuropathy and neuropathic pain and guided more thorough approaches that may lead to personalized modifying therapy in the future. PNS lipid homeostasis appears to be one driver in the onset and progression of neuropathy in normoglycemic MetS.

Notably, novel preclinical research is focused on targeting fatty acids or cholesterol in the PNS. In MetS and obesity models, alteration of lipid profiles and glycemia is always concomitant, independent glucose pathways are difficult to rule out, and likely cross talk with lipid signaling. Nevertheless, it is still not clear whether defect of nuclear receptor activation, diet, microbiome, or a sedentary lifestyle is the cause of peripheral neuropathy. Targeting pathways downstream of these interventions may be of interest to improve neuropathic pain.

While studies have clearly shown that these interventions modify cellular signaling in PNS cells, including sensory neurons, SC, or immune cells, neuropathy amelioration might be also due to indirect effects on body weight, lipid profile, or insulin sensitivity. Cellular Ca²⁺ homeostasis appears to be one of the converging pathways impacted by all the beneficial interventions, and thus, comprehensively defining cell-specific changes in Ca²⁺ homeostasis (including ER-Ca²⁺ release and uptake, Ca²⁺ channels, and GPCRs' downstream pathways, among others) may be of interest to better understand the disease and find specific targets.

Most of the studies focused on DRG and SN, however, the spinal cord, receiving sensory neuron synapses (also originating from the gut) and containing immune cells, would need further investigation since neuroinflammation has been identified recently in many studies. The release of neuropeptides (e.g., CGRP) and the neuroimmune communication that occurs in the spinal horn represent an area of study to potentially target early in MetS (Fig. 6). More mechanistic and translational studies to further understand the pathways delineated in this review will hopefully allow for the identification of new target(s) to ameliorate or to block the development of neuropathy in obese patients.

FIG. 6.

Changes in CNS in addition to peripheral nerve system may represent other causes of neuropathy and pain in MetS. A majority of the studies have focused on the PNS and showed that MetS modifies DRG neurons together with their peripheral processes and associated cells. Primary afferents synapse to secondary order neurons in the dorsal horn of the spinal cord that carry the pain information to the higher orders of the brain. In addition, many studies showed that dyslipidemia and MetS alter the CNS, including the spinal cord and discrete brain areas such as the hypothalamus, but also the somatosensory cortex. Chronic pain is a sensitized state, either peripherally or centrally, that can induce lasting changes within the CNS. Thus, there is a need for better defining changes in the CNS of MetS mouse models subjected to the described beneficial interventions. CNS, central nervous system; DRG, dorsal root ganglia.

Abbreviations Used

ABCA1

ATP-binding cassette subfamily A member 1

AC

adenylyl cyclase

ATF3

activating transcription factor 3

ATP

adenosine triphosphate

BMI

body mass index

[Ca²⁺]_i

intracellular Ca²⁺

cAMP

cyclic adenosine monophosphate

CD36

cluster of differentiation 36

CGRP

calcitonin gene-related peptide

CHOP

C/EBP homologous protein

CNS

central nervous system

COX

cyclooxygenase

CXCR4

C-X-C motif chemokine receptor 4

DIO

diet-induced obesity

DREADDs

designer receptors exclusively activated by designer drugs

DRG

dorsal root ganglia

ER

endoplasmic reticulum

FFARs

free fatty acid receptors

FMT

fecal microbiome transplantation

GPCRs

G-coupled protein receptors

HDAC

histone deacetylase

HDL

high-density lipoprotein

HFD

high-fat diet

IASP

International Association for the Study of Pain

IENF

intraepidermal nerve fiber

IP₃

inositol 1,4,5-triphosphate

KO

knockout

Kv

shaker-type potassium channels

LDL

low-density lipoprotein

LPCAT3

lysophosphatidylcholine acyltransferase 3

LXR

liver X receptor

MCTs

monocarboxylate transporters

MetS

metabolic syndrome

MNSI

Michigan Neuropathy Screening Instrument

MUFA

monounsaturated fatty acid

NC

normal chow

NCS

nerve conductance studies

NCV

nerve conductance velocities

NRG1

neuregulin 1 type III

PI3K

phosphoinositide 3-kinase

PKA

protein kinase A

PNS

peripheral nervous system

PPARs

peroxisome proliferator-activated receptors

PUFAs

polyunsaturated fatty acids

RXR

retinoid X receptor

Ryr2

ryanodine receptor 2

SC

Schwann cells

SCFA

short-chain fatty acids

SDF-1

stromal cell-derived factor 1

SERCA

sarco/endoplasmic reticulum Ca²⁺-ATPase 2b

SN

sciatic nerve

T2DM

type 2 diabetes mellitus

TLR

Toll-like receptor

TRPM8

transient receptor potential cation channel subfamily M member 8

TRPV1

transient receptor potential cation channel subfamily V member 1

UPR

unfolded protein response

VCCG

voltage-gated calcium channels

VGSCs

voltage-gated sodium channels

WD

Western diet

Authors' Contributions

The authors confirm contribution to the article as follows: R.H.B., S.K., and V.M.A. conceived, drafted, revised, and finalized the article. All authors reviewed and approved the final version of the article.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The authors are thankful to their funding sources DiaComp Pilot & Feasibility project, 20AU4112 and NIDDK 5R01DK117404-02, to V.M.A.