Exosomes as emerging mediators and therapeutic vehicles in painful diabetic neuropathy: From pathophysiology to clinical translation

Table 1 Summary of pathophysiological mechanisms in painful diabetic neuropathy

Category	Mechanism/component	Pathways and molecular mediators	Pathological consequences	Functional/clinical correlates	Synthetic insight: Cross-mechanism interaction and therapeutic implication
Peripheral mechanisms of nerve injury	Polyol pathway activation	Aldose reductase (glucose → sorbitol); sorbitol dehydrogenase (sorbitol → fructose); NADPH depletion; NADH generation	Osmotic stress; myo-inositol depletion; impaired Na+/K+-ATPase; reduced glutathione; mitochondrial superoxide production	Increased oxidative stress; axonal dysfunction; reduced antioxidant capacity	Represents the earliest measurable metabolic insult; NADPH depletion directly links polyol pathway activation to downstream mitochondrial dysfunction (row below) by impairing glutathione regeneration. Exosomal cargo from metabolically stressed neurons under polyol pathway activation is the proposed upstream trigger for Schwann cell exosomal reprogramming (Table 2)
	AGEs	Non-enzymatic glycation by methylglyoxal; RAGE activation; NF-κB signaling	Direct structural modification of myelin and cytoskeletal proteins; accelerated protein degradation; chronic sterile neuroinflammation	Progressive protein dysfunction; sustained inflammatory state; Schwann cell, endothelial cell, and macrophage activation	AGE/RAGE activation provides the mechanistic bridge between chronic hyperglycemia and sustained neuroinflammation; RAGE activation on Schwann cells directly reprograms their exosomal secretome (Table 2) toward a pro-inflammatory, neurotrophic-depleted profile, linking metabolic and inflammatory pathological arms
	Mitochondrial dysfunction	Impaired oxidative phosphorylation; reduced ATP; increased ROS; PGC-1α suppression; excessive mitochondrial fission	Reduced neuronal viability; impaired axonal transport; intrinsic apoptosis pathway activation; nitration/inactivation of mitochondrial enzymes	Energy failure in DRG neurons; progressive distal axonopathy; impaired nerve regeneration	Mitochondrial dysfunction in DRG neurons is both a downstream consequence of polyol/AGE pathology and an independent amplifier; impaired axonal transport reduces retrograde neurotrophic signaling to Schwann cells, accelerating their phenotypic reprogramming and exosomal cargo shift. MSC-derived exosomes carrying HSP70/90 and SIRT1 specifically target this node (Table 5)
	Axonal degeneration and demyelination	Progressive IENFD loss; Schwann cell phenotypic transition	Release of MMPs; pro-inflammatory cytokine secretion; altered Schwann cell exosomal cargo (pro-inflammatory miRNAs, depleted neurotrophic factors)	IENFD loss correlates with pain severity and diabetes duration; predominant vulnerability of small unmyelinated C-fibers and thinly myelinated A-delta fibers	IENFD reduction is the structural readout integrating all upstream metabolic and inflammatory insults; its correlation with exosomal miR-21-5p and NfL levels (Table 3) positions it as the primary composite endpoint bridging mechanistic and diagnostic domains, and its reversibility in response to MSC-derived exosome therapy (35%-80% recovery; Table 5) makes it the most relevant disease-modification endpoint for future clinical trials
	Peripheral nociceptor sensitization	Voltage-gated sodium channels (Nav1.7, Nav1.8, Nav1.9); TRP channels (TRPV1, TRPA1); HCN channels; potassium channel subtypes	Lowered action potential threshold; increased spontaneous ectopic discharge from injured axons and DRG; channel upregulation and sensitization in neuroinflammation/oxidative stress	Thermal hyperalgesia; mechanical allodynia; spontaneous burning pain	Nav1.7 upregulation and TRPV1 sensitization are directly regulated by Schwann cell-derived exosomal let-7 downregulation and ceramide/sphingosine-1-phosphate delivery (Table 2), mechanistically connecting the glial exosomal secretome to nociceptor hyperexcitability; exosomal let-7a mimic and anti-TRPV1 siRNA therapies (Table 5) specifically target this node without systemic cardiac conduction side effects
Neuroinflammation	Macrophages	Transition from M2-like to M1-like phenotype; RAGE activation; TLR4 signaling	The iNOS upregulation; enhanced TNF-α, IL-1β, IL-6, IL-18; increased matrix metalloproteinase secretion	Degradation of extracellular matrix; Schwann cell dysfunction; axonal degeneration; nociceptor sensitization	M1 macrophage polarization and Schwann cell pathology form a bidirectional feed-forward loop mediated by exosomal miR-155/SOCS1/JAK-STAT signaling (Table 2); this loop, once established, is self-sustaining independently of glycemic level, which may explain why PDN symptoms persist after glycemic control is achieved
	Schwann cells	Hyperglycemia-induced phenotypic reprogramming; p38 MAPK activation	Reduced NGF, NT-3, BDNF; upregulation of MCP-1 and TNF-α; altered exosomal repertoire (pro-inflammatory miRNAs, depleted neurotrophic cargo)	Impaired neurotrophic support for small-diameter sensory fibers; direct nociceptor sensitization; exosome-mediated propagation of inflammatory signals	Schwann cell exosomal cargo remodeling is the central pathogenic hub connecting vascular injury (endothelial exosomal dysfunction), metabolic stress (AGE/mitochondrial pathway), macrophage activation, and peripheral nociceptor sensitization; it is therefore also the primary target of MSC-derived exosome therapy (Table 5), which restores the physiological Schwann cell exosomal phenotype
	SGCs	Reactive phenotype in response to peripheral nerve injury and inflammatory signals	Increased GFAP, connexin 43; IL-1β, TNF-α production	Paracrine inflammatory microenvironment within DRG; lowered excitability threshold of sensory neuron somata; promoted ectopic discharge; sustained peripheral sensitization	SGC activation amplifies peripheral sensitization by lowering DRG neuron excitability thresholds via connexin 43-mediated gap junctions; this mechanism is currently underrepresented in exosomal biomarker panels (Tables 3, 4 and 7) and represents a significant gap in the PDN exosome literature
	Endoneurial vasculature	Dysregulation of vasa nervorum; blood-nerve barrier disruption	Altered exosomal cargo (pro-coagulant factors); impaired tight junction integrity	Facilitated macrophage infiltration; entry of systemic inflammatory mediators into endoneurial compartment	Vascular injury is the earliest initiating node in the pathological cascade (Table 2), upstream of Schwann cell and macrophage pathology; adipose-derived MSC exosomes with pro-angiogenic VEGF/miR-132-3p cargo (Table 5) specifically target this node, but are expected to only partially attenuate downstream pathology occurring through vasculature-independent routes
Central sensitization	Spinal dorsal horn synaptic plasticity	Wind-up and long-term potentiation at C-fiber synapses; NMDA receptor activation; substance P; CGRP; neurokinin-1 receptor	Calcium influx; post-synaptic kinase activation; immediate early gene expression; synaptic structural remodeling	Sustained enhancement of synaptic efficacy; transition to sensitized central state	Spinal LTP at C-fiber synapses has been directly demonstrated in STZ-diabetic rodent models; however, the contribution of peripherally derived pathological exosomes to this process (Table 2) remains inferential and has not been directly demonstrated in PDN models – distinguishing it from the better-evidenced peripheral mechanisms above
	Spinal microglia	Activation via ATP (P2X4, P2X7), fractalkine (CX3CR1), CSF-1; BDNF release	BDNF → TrkB → KCC2 (reduce) → converts GABAergic inhibition into net excitation	Loss of inhibitory tone; enhanced nociceptive transmission; central disinhibition	Microglial BDNF/KCC2 suppression is the most directly evidenced central sensitization mechanism in PDN models and the most tractable central exosomal therapeutic target; IL-4-stimulated macrophage exosomes enriched with IL-10, TGF-β, and miR-124 restore KCC2 expression in diabetic rodent spinal cord (Table 5)
	Spinal astrocytes	Activation; downregulation of GLT-1, GLAST	Reduced GLT-1/GLAST expression; glutamate spillover; D-serine and TNF-α release; extrasynaptic NMDA receptor activation	Potentiated synaptic transmission; sustained central sensitization	Astrocytic GLT-1/GLAST downregulation by exosomal miRNAs has been demonstrated in general neuroinflammatory models but not specifically in PDN spinal cord tissue; this mechanistic inference requires direct PDN-model confirmation before astrocyte-targeting exosomal therapies can be rationally designed
	Descending pain modulatory system	Impaired descending inhibition (reduced noradrenaline, reduced alpha-2 adrenergic receptors); enhanced descending facilitation (increased 5-HT on 5-HT3 receptors)	Imbalance between descending inhibition and facilitation; permissive central environment for pain amplification	Allodynia and hyperalgesia; pain signals amplified rather than gated; largely refractory to peripherally targeted analgesics	Descending modulatory pathway dysfunction explains the frequent inadequacy of peripherally targeted therapies in PDN and creates the strongest mechanistic rationale for intranasal exosomal delivery strategies capable of accessing supraspinal pain-modulatory circuits; however, direct exosomal evidence for this pathway in PDN is currently absent

PDN: Painful diabetic neuropathy; NADPH: Nicotinamide adenine dinucleotide phosphate; Na⁺/K⁺-ATPase: Sodium-potassium adenosine triphosphatase; NADH: Nicotinamide adenine dinucleotide (reduced form); AGEs: Advanced glycation end-products; RAGE: Receptor for advanced glycation end products; NF-κB: Nuclear factor kappa B; ATP: Adenosine triphosphate; ROS: Reactive oxygen species; PGC-1α: Peroxisome proliferator-activated receptor gamma coactivator 1-alpha; IENFD: Intraepidermal nerve fiber density; DRG: Dorsal root ganglion; Nav1.7, Nav1.8, Nav1.9: Voltage-gated sodium channel subtypes 1.7, 1.8, and 1.9; TRPV1: Transient receptor potential vanilloid 1; TRPA1: Transient receptor potential ankyrin 1; HCN: Hyperpolarization-activated cyclic nucleotide-gated channels; iNOS: Inducible nitric oxide synthase; TNF-α: Tumor necrosis factor-alpha; IL-1β: Interleukin-1 beta; IL-6: Interleukin-6; IL-18: Interleukin-18; TLR4: Toll-like receptor 4; NGF: Nerve growth factor; NT-3: Neurotrophin-3; BDNF: Brain-derived neurotrophic factor; MCP-1: Monocyte chemoattractant protein-1; p38 MAPK: P38 mitogen-activated protein kinase; GFAP: Glial fibrillary acidic protein; NMDA: N-methyl-D-aspartate; CGRP: Calcitonin gene-related peptide; P2X4, P2X7: Purinergic receptor subtypes P2X4 and P2X7; CX3CR1: CX3C chemokine receptor 1; CSF-1: Colony-stimulating factor 1; TrkB: Tropomyosin receptor kinase B; KCC2: Potassium-chloride cotransporter 2; GABA: Gamma-aminobutyric acid; GLT-1: Glutamate transporters-1; GLAST: Glutamate aspartate transporter; 5-HT: Serotonin (5-hydroxytryptamine); 5-HT₃: Serotonin receptor subtype 3; miRNAs: MicroRNAs; MSC: Mesenchymal stem cell; NfL: Neurofilament light chain; HSP70: Heat shock protein 70; MMPs: Matrix metalloproteinases; SGCs: Satellite glial cells; SOCS1: Suppressor of cytokine signaling 1; siRNA: Small interfering RNA; VEGF: Vascular endothelial growth factor; JAK-STAT: Janus kinase-signal transducer and activator of transcription; STZ: Streptozotocin; LTP: Long-term potentiation; M1: Classically activated macrophage.

Table 2 These changes promote pathological remodeling of Schwann cell-derived exosomal cargo, enhance immune cell activation, and ultimately amplify neuroinflammation, axonal degeneration, and peripheral sensitization during Painful diabetic neuropathy progression

Exosome source	Condition/stimulus	Altered exosomal cargo	Pathophysiological consequence	Mechanism of action	Temporal sequence/position in cascade
Hyperglycemia-exposed cells (general)	Chronic hyperglycemia, oxidative stress	Increased nSMase2, increased ESCRT components	Exosomal hypersecretion; increased burden of pathologically conditioned vesicles	Ceramide-dependent and ESCRT-dependent pathway activation	Foundational/upstream node – establishes the generalized hypersecretory state permissive for all downstream cell-specific cargo remodeling below
Endothelial cells (vasa nervorum)	Hyperglycemic injury	Increased pro-coagulant phosphatidylserine, tissue factor, von Willebrand factor; oxidized LDL, asymmetric dimethylarginine; miRNAs targeting eNOS	Hypercoagulable endoneurial milieu; impaired NO bioavailability → vasoconstriction; pericyte apoptosis; blood-nerve barrier disruption	Transfer of pro-apoptotic proteins and miRNAs targeting claudin/occludin	Early/initiating node – precedes measurable Schwann cell cargo shift and macrophage accumulation in diabetic rodent models; proposed upstream trigger of Schwann cell and macrophage pathology (below) via endoneurial hypoxia and barrier breakdown
Schwann cells (physiological)	Normoglycemia	NGF, BDNF, NT-3, GDNF; pyruvate, lactate; myelin proteins/Lipids	Axonal trophic support, metabolic substrate delivery, myelin maintenance	Exosomal transfer of neurotrophins and metabolic fuels to axons	Baseline/reference state – homeostatic phenotype contrasted against the hyperglycemic Schwann cell shift below
Schwann cells (hyperglycemic)	High glucose exposure	Reduced NGF, BDNF, HSP70; increased cleaved caspase-3, Bax, oxidized proteins; increased miR-21-3p, miR-28-3p, miR-34a	Loss of neuroprotection; axonal degeneration; intraepidermal nerve fiber loss; mechanical allodynia, thermal hyperalgesia	Delivery of pro-apoptotic cargo; targeting pro-survival pathways in axons	Secondary/amplified node – driven directly by hyperglycemia and further intensified by upstream endothelial-derived hypoxia and barrier breakdown (rate-amplifying cofactor, not obligate trigger)
Schwann cells (hyperglycemic)	High glucose exposure	Increased miR-let-7 (downregulation of HMGA2 repressor); ceramide, sphingosine-1-phosphate	Nav1.7 and TRPV1 upregulation in sensory neurons; peripheral nociceptor hyperexcitability	MiRNA transfer targeting channel repressors; p38 MAPK/ERK activation → channel phosphorylation/translocation	Secondary/parallel node — co-occurs with the apoptotic cargo shift above; nociceptor-sensitizing arm of the same hyperglycemia-driven Schwann cell reprogramming
Classically activated macrophages (M1)	Diabetic endoneurial microenvironment	TNF-α, IL-1β, IL-6; miR-155, miR-146a-3p, miR-21-5p; PGE2, LTB4	Sustained inflammatory responsiveness; positive feedback loop with Schwann cells; TRPV1 sensitization	The miR-155 → SOCS1 suppression → JAK-STAT amplification; delivery of lipid mediators to nociceptor terminals	Tertiary/reinforcing node – activated via endothelial barrier breakdown (DAMP exposure); feeds back onto Schwann cells (above) via miR-155/SOCS1/JAK-STAT, closing a self-amplifying loop
CD4+ T helper cells	Diabetic nerve infiltration	IFN-γ, TRAIL, FasL	Schwann cell apoptosis; impaired regenerative capacity	Induction of death receptor-mediated apoptosis in Schwann cells	Concurrent immune-effector node – acts in parallel with macrophage activation; compounds Schwann cell injury initiated upstream
CD8+ cytotoxic T cells	Diabetic nerve infiltration	Perforin, granzyme B	DRG neuron apoptosis	Perforin-mediated membrane poration and granzyme B-induced apoptosis	Concurrent immune-effector node – parallels CD4+ T cell action; extends peripheral cascade to DRG neuronal compartment
Regulatory T cells	Diabetes-associated dysfunction	Reduced TGF-β, IL-10	Loss of anti-inflammatory exosomal signaling; removal of brake on neuroinflammation	Failure to suppress excessive immune activation	Permissive/disinhibiting node – loss of this regulatory checkpoint allows the macrophage/T-cell cascade above to proceed unchecked
Primary afferent central terminals	Hyperexcitable afferents (spinal dorsal horn)	Substance P, CGRP precursors; miRNAs targeting GABAergic interneuron function	Synaptic potentiation; reduced inhibitory tone; central sensitization	Transfer of pro-nociceptive cargo to second-order projection neurons and interneurons	Central propagation node – peripheral sensitization (Schwann cell/nociceptor arm above) is transmitted centrally via this route, initiating the spinal cascade below
Activated spinal microglia	Diabetic spinal cord	BDNF, IL-1β, miR-let-7c	Reduced KCC2 expression; loss of GABAergic inhibition; enhanced NMDA receptor-mediated transmission	BDNF-mediated downregulation of KCC2; impaired chloride extrusion	Central reinforcing node – activated by primary afferent input above; central counterpart to the peripheral macrophage feedback loop
Resting spinal microglia	Normoglycemia	Anti-inflammatory cargo; neuroprotective factors	Maintained synaptic homeostasis and inhibitory tone	Preservation of GABAergic function and KCC2 expression	Baseline/reference state – homeostatic phenotype contrasted against activated spinal microglia above
Reactive spinal astrocytes	Diabetic spinal cord	The miRNAs targeting GLT-1 and GLAST	Reduced glutamate clearance; glutamate spillover; sustained NMDA receptor activation	The miRNAs-mediated downregulation of astrocyte glutamate transporters	Central parallel node – co-occurs with microglial activation above; jointly sustains spinal central sensitization
Systemic circulation (to CNS)	Diabetes-compromised BBB	Peripherally derived pathological exosomes (e.g., from endothelial cells, immune cells)	Central sensitization; maladaptive CNS circuit remodeling	Delivery of pro-inflammatory and pro-nociceptive cargo to spinal and supraspinal centers	Terminal/integrative node – closes the peripheral-to-central loop by directly importing the vascular- and immune-conditioned exosome pool (rows 2, 6-9) into the CNS, independent of the synaptic relay route above

ESCRT: Endosomal Sorting Complex Required for Transport; LDL: Low-density lipoprotein; eNOS: Endothelial nitric oxide synthase; NO: Nitric oxide; NGF: Nerve growth factor; BDNF: Brain-derived neurotrophic factor; NT-3: Neurotrophin-3; GDNF: Glial cell line-derived neurotrophic factor; HSP70: Heat shock protein 70; Bax: Bcl-2-associated X protein; HMGA2: High mobility group AT-hook 2; TRPV1: Transient receptor potential vanilloid 1; p38 MAPK: P38 mitogen-activated protein kinase; ERK: Extracellular signal-regulated kinase; TNF-α: Tumor necrosis factor-alpha; IL: Interleukin; PGE2: Prostaglandin E2; LTB4: Leukotriene B4; SOCS1: Suppressor of cytokine signaling 1; JAK-STAT: Janus kinase-signal transducer and activator of transcription; IFN-γ: Interferon-gamma; TRAIL: Tumor necrosis factor-related apoptosis-inducing ligand; DRG: Dorsal root ganglion; TGF-β: Transforming growth factor-β; CGRP: Calcitonin gene-related peptide; GABA: Gamma-aminobutyric acid; KCC2: Potassium-chloride cotransporter 2; NMDA: N-methyl-D-aspartate; GLT-1: Glutamate transporters-1; GLAST: Glutamate aspartate transporter; CNS: Central nervous system; DAMP: Damage-associated molecular pattern; miRNAs: MicroRNAs; nSMase2: Neutral sphingomyelinase 2; FasL: Fas ligand; M1: Classically activated macrophage; BBB: Blood-brain barrier.

Table 3 Exosomal diagnostic biomarkers in painful diabetic neuropathy

Biomarker category	Specific biomarker	Expression change in PDN	Mechanistic/clinical correlate	Diagnostic utility	Validation status and key specificity limitation
MiRNAs	MiR-21-5p	Upregulated	Targets PTEN (reduced) and PDCD4 (reduced) → enhanced PI3K-Akt survival signaling	Correlates positively with neuropathic pain scores, negatively with IENFD	Limited human PDN cohort data only; not PDN-specific – also upregulated in diabetic nephropathy, retinopathy, cardiomyopathy, and chemotherapy-induced neuropathy; no comparative cohort study has distinguished PDN from these conditions using this marker alone
	MiR-146a-5p	Upregulated	Negative feedback regulator of TLR/IL-1R signaling (targets IRAK1, TRAF6); parallels increased IL-6 and TNF-α	Dynamic indicator of neuroinflammatory activity; correlates with disease duration and nerve conduction deficits	Supporting evidence derived from a DPN cohort using a ceRNA framework, not a pain-stratified PDN population; also elevated in diabetic nephropathy (a frequent PDN comorbidity), representing a critical confound for clinical diagnostic use without tissue-source enrichment
	Let-7a-5p, let-7b-5p	Downregulated	Derepresses HMGA2 → increased Nav1.7 transcription; derepresses IL-6Rα and STAT3 → increased IL-6 signaling	Molecular readout of peripheral nociceptor sensitization and neuroinflammation	Evidence is primarily preclinical (rodent DRG/spinal cord); no published human PDN cohort has confirmed circulating exosomal let-7 downregulation with clinical pain correlates; it is also dysregulated in diabetic retinopathy and hereditary neuropathies, limiting current specificity
	MiR-9-5p	Reduced	Impairs post-transcriptional suppression of NF-κB signaling in Schwann cells	Reflects Schwann cell pro-inflammatory phenotype	Mechanistic evidence only; no published PDN-specific human biomarker cohort data; lowest validation tier among listed miRNAs
	MiR-29 family	Downregulated	Promotes extracellular matrix accumulation and endoneurial fibrosis	Correlates with structural nerve remodeling	Mechanistic evidence only; endoneurial fibrosis correlation is inferred from general fibrosis biology rather than PDN-specific cohort data
	MiR-132-3p	Reduced	Derepresses RhoGAP32 → compromises axonal branching and synaptic plasticity	Indicates impaired axonal regeneration capacity	Dual role: Pathogenic exosomal cargo in PDN (Table 2) and therapeutic cargo in MSC-derived exosomes (Table 5); this context-dependency must be considered when interpreting circulating levels as a biomarker
	MiR-223-3p	Elevated	Derived from M1-polarized macrophages	Marker of endoneurial macrophage pro-inflammatory activation	Not PDN-specific; elevated in systemic inflammatory states; utility as a PDN biomarker depends on demonstrating neural-compartment specificity via L1CAM or MOG immunoaffinity enrichment (Table 4)
Proteins	NfL	Elevated (exosomal fraction)	Axonal cytoskeletal damage; exosomal NfL correlates more strongly with neuropathy severity than total plasma NfL	Indicator of axonal degeneration; correlates with nerve conduction studies and IENFD	Best-validated protein biomarker in this table; however, NfL elevation is not PDN-specific (also elevated in AD, MS, ALS); the exosomal fraction advantage over total plasma NfL has not yet been validated in large independent PDN cohorts
	HSP70	Reduced	Neuroprotective chaperone normally enriched in healthy Schwann cell exosomes	Inversely correlates with neuropathic pain severity; reflects depletion of neuroprotective Schwann cell cargo	Reduction is consistent with hyperglycemia-induced Schwann cell dysfunction; relatively more PDN-contextual than miRNA markers since HSP70 enrichment is a Schwann cell-specific exosomal feature, though formal specificity studies across PDN comorbidities are lacking
	GAP-43	Variable (higher in regenerative phenotype)	Expressed in regenerating axonal growth cones	Prognostic indicator of residual regenerative capacity; may guide timing of neuroprotective interventions	Prognostic rather than diagnostic marker; variable expression pattern limits use as a single threshold biomarker; most useful as part of a composite panel with NfL (degeneration) and HSP70 (Schwann cell support) to characterize individual patient disease trajectory
	MBP fragments, PMP22	Elevated	Released from degenerating myelin sheaths; packaged into exosomes by activated macrophages	Correlates with slowing of motor and sensory nerve conduction velocities; marker of demyelinating neuropathy	Relevant specifically to large-fiber and demyelinating PDN subtypes; complementary to NfL (small-fiber/axonal marker); combined NfL + MBP/PMP22 panel could enable fiber-type-specific PDN subtyping, which has direct implications for therapeutic target selection

PDN: Painful diabetic neuropathy; miRNAs: MicroRNAs; MSC: Mesenchymal stem cell; NfL: Neurofilament light chain; HSP70: Heat shock protein 70; PMP22: Peripheral myelin protein 22; L1CAM: L1 cell adhesion molecule; MOG: Myelin oligodendrocyte glycoprotein; PTEN: Phosphatase and tensin homolog; PDCD4: Programmed cell death protein 4; PI3K-Akt: Phosphatidylinositol 3-kinase-protein kinase B; IENFD: Intraepidermal nerve fiber density; TLR4: Toll-like receptor 4; IL-1R: Interleukin-1 receptor; IRAK1: Interleukin-1 receptor-associated kinase 1; TRAF6: Tumor necrosis factor receptor-associated factor 6; TNF-α: Tumor necrosis factor-alpha; IL: Interleukin; DPN: Diabetic peripheral neuropathy; HMGA2: High mobility group AT-hook 2; STAT3: Signal transducer and activator of transcription 3; DRG: Dorsal root ganglion; NF-κB: Nuclear factor kappa B; RhoGAP32: Rho GTPase activating protein 32; M1: Classically activated macrophage; MS: Multiple sclerosis; ALS: Amyotrophic lateral sclerosis; GAP-43: Growth-associated protein 43; AD: Alzheimer’s disease.

Table 4 Summary of tissue-specific signatures

Tissue-specific signatures	Source	Method to enhance specificity	Key cargo detected	Clinical/anatomical window
Plasma/serum (neural-derived)	Immunoaffinity capture (L1CAM for neurons, MOG for Schwann cells)	Neuron- or Schwann cell-enriched exosomal cargo	Enables selective capture of neural exosomes from total circulating pool; improves diagnostic discrimination	Highest current clinical priority – accessible, scalable, and partially validated in adjacent conditions (TBI, AD); L1CAM enrichment has not yet been validated in PDN cohorts specifically, representing the most actionable near-term research gap
Skin blister fluid/dermal interstitial fluid	Direct sampling of affected area	TRPV1, substance P, pro-inflammatory miRNAs	Anatomically proximal to small fiber pathology; more powerful discrimination from pain-free diabetic controls than plasma	Anatomically ideal for PDN given distal lower extremity predilection; however, standardized blister induction and exosome isolation protocols are not yet established, limiting multi-site reproducibility
CSF	Lumbar puncture (research use)	BDNF, Iba1 fragments, miRNAs associated with NMDA receptor upregulation	Direct access to CNS compartment; reflects central sensitization state; distinguishes central vs peripheral pain mechanisms	Highest mechanistic specificity for central sensitization but invasiveness limits routine clinical use; most appropriate for research studies specifically designed to characterize spinal vs peripheral exosomal pathology in PDN
Urine	Non-invasive collection	The miR-21-5p (increased), let-7 family (reduced)	Practical advantages; partially mirror plasma signatures; requires validation for PDN vs diabetic nephropathy discrimination	Critical confound diabetic nephropathy independently alters urinary exosomal composition and co-exists in a significant proportion of PDN patients; urinary exosomal miRNA signatures cannot be attributed specifically to neuropathic pathology without concurrent nephropathy exclusion or renal-source depletion strategies

PDN: Painful diabetic neuropathy; miRNAs: MicroRNAs; L1CAM: L1 cell adhesion molecule; MOG: Myelin oligodendrocyte glycoprotein; TBI: Traumatic brain injury; TRPV1: Transient receptor potential vanilloid 1; CSF: Cerebrospinal fluid; BDNF: Brain-derived neurotrophic factor; NMDA: N-methyl-D-aspartate; CNS: Central nervous system; AD: Alzheimer’s disease.

Table 5 Exosomes as therapeutic vehicles in painful diabetic neuropathy

Category	Agent/approach	Mechanism of action/key cargo	Therapeutic outcomes/advantages	Evidence density/quality
Therapeutic rationale	Exosomes (general)	CD47 “don’t eat me” signal; natural barrier traversal; multi-molecular cargo delivery; surface engineerability	Bypasses limitations of single-agent drugs; non-genomic; immunologically well-tolerated; protects labile cargo	Conceptual/mechanistic rationale – not outcome-specific
Leading platform	MSC-derived exosomes	Cargo: BDNF, NGF, NT-3, GDNF, VEGF, HGF, HSP70/90, miR-146a, miR-223, let-7 family, miR-132-3p, miR-21-5p. Action: PI3K-Akt/MAPK-ERK activation; TLR4/NF-κB suppression; macrophage M2 polarization	Improves nerve conduction velocity (30%-60%), IENFD (35%-80%), and behavioral pain thresholds (40%-70%); reduces endoneurial TNF-α, IL-1β, iNOS; promotes structural regeneration	High – multi-study preclinical evidence base using consistent, composite functional/electrophysiological/behavioral outcome batteries
Preconditioning strategies	Hypoxic	Enriches HIF-1α products (VEGF, EPO, miR-210)	Enhanced endoneurial vascular restoration	Moderate – mechanistic enrichment data; outcome reported qualitatively
	Inflammatory	Augments anti-inflammatory miR-146a and let-7b content.	Enhanced anti-inflammatory effect	Moderate – mechanistic enrichment data; outcome reported qualitatively
	Neurotrophic	Increases neurotrophic protein cargo	Enhanced DRG neuroprotection	Moderate – mechanistic enrichment data; outcome reported qualitatively
Engineering for targeted delivery	Surface engineering	RVG peptide fusion (targets neuronal nAChRs); PEGylation	Increased DRG/peripheral nerve accumulation; extended circulation half-life (2-4 x)	Moderate-high – quantified pharmacokinetic data (half-life), but limited PDN-specific functional endpoint data
	Cargo engineering	Electroporation (miRNA/siRNA); passive incubation (curcumin); parent cell genetic engineering (lactadherin C1C2 fusion)	Overcomes bioavailability limits (e.g., curcumin); enables scalable, enriched cargo production	Moderate – proof-of-concept delivery data; outcome reported qualitatively
Mechanistically targeted cargo	let-7a mimic	Suppresses HMGA2-driven Nav1.7 transcription	Reduces ectopic discharge and mechanical allodynia without cardiac side effects	Moderate – single-mechanism preclinical study; qualitative behavioral outcome
	Anti-TRPV1 siRNA	RVG-targeted knockdown of TRPV1	Sustained reduction of thermal hyperalgesia from a single injection	Moderate – single-mechanism preclinical study; qualitative behavioral outcome
	MiR-146a-5p mimic	Reduces NF-κB activation and pro-inflammatory cytokines	Decreased endoneurial macrophage infiltration; improved nerve conduction and IENFD	Moderate – qualitative directional outcome; not consolidated into a quantified range
	Combined Cargo (let-7b-5p + miR-146a-5p)	Synergistic targeting of peripheral sensitization and neuroinflammation	Outperforms single-cargo preparations; demonstrates multiplex therapeutic advantage	Moderate – comparative (vs single-cargo) data; outcome reported qualitatively
	IL-4-stimulated macrophage exosomes	Enriched in IL-10, TGF-β, miR-124	Suppresses microglial activation; restores KCC2 expression; reverses central sensitization	Moderate – proof-of-concept central-mechanism study; qualitative outcome
	Exosomal curcumin	Suppresses NF-κB activation	Superior anti-inflammatory efficacy vs free curcumin	Moderate – direct comparative study; qualitative outcome
Alternative cell sources	Healthy Schwann cell-derived exosomes	Full physiological complement of neurotrophic cargo	Restores axonal mitochondrial function and IENFD more effectively than recombinant factors	Low-moderate – limited to comparison against single recombinant factors; no consolidated quantitative range or multi-domain (functional + electrophysiological + behavioral) outcome battery equivalent to MSC-derived exosome data
	Plant-derived nanoparticles (ginger)	Enriched in phosphatidic acids and 6-gingerol	Attenuates NF-κB signaling and mechanical hyperalgesia; oral delivery possible; scalable and non-immunogenic	Low – limited to a single identified preclinical study; no IENFD, nerve conduction velocity, or quantified behavioral threshold data available; should be regarded as an early proof-of-concept signal rather than a validated efficacy estimate

MSC: Mesenchymal stem cell; BDNF: Brain-derived neurotrophic factor; NGF: Nerve growth factor; NT-3: Neurotrophin-3; GDNF: Glial cell line-derived neurotrophic factor; VEGF: Vascular endothelial growth factor; HGF: Hepatocyte growth factor; HSP: Heat shock protein; PI3K-Akt: Phosphatidylinositol 3-kinase-protein kinase B; MAPK-ERK: Mitogen-activated protein kinase-extracellular signal-regulated kinase; TLR4: Toll-like receptor 4; NF-κB: Nuclear factor kappa B; IENFD: Intraepidermal nerve fiber density; TNF-α: Tumor necrosis factor-alpha; IL: Interleukin; iNOS: Inducible nitric oxide synthase; HIF-1α: Hypoxia-inducible factor 1-alpha; DRG: Dorsal root ganglion; RVG: Rabies virus glycoprotein; nAChRs: Nicotinic acetylcholine receptors; PEGylation: Polyethylene glycol conjugation; miRNAs: MicroRNAs; siRNA: Small interfering RNA; HMGA2: High mobility group AT-hook 2; TRPV1: Transient receptor potential vanilloid 1; KCC2: Potassium-chloride cotransporter 2; TGF-β: Transforming growth factor-β; PDN: Painful diabetic neuropathy; EPO: Erythropoietin.

Table 6 Summary risk-of-bias assessment for key preclinical studies contributing to Table 5 efficacy ranges (Systematic Review Centre for Laboratory animal Experimentation tool) and the human randomized controlled trial (Cochrane Risk of Bias 2.0)

Ref.	Study type	Model/population	Sequence generation	Baseline characteristics/allocation concealment	Random housing/blinding of personnel	Blinded outcome assessment	Incomplete outcome data	Selective outcome reporting	Overall RoB
Fan et al[13], 2020	Preclinical (SYRCLE)	STZ-induced diabetic mouse; MSC-derived exosomes; IV administration	Unclear	Unclear	Unclear	High (behavioral endpoints unblinded)	Low	Unclear	High
Shan et al[12], 2022	Preclinical (SYRCLE)	STZ-induced diabetic mouse; SIRT1-enriched bone marrow MSC exosomes	Unclear	Unclear	Unclear	High (unblinded outcome assessment reported)	Low	Unclear	High
Jia et al[23], 2018	Preclinical (SYRCLE)	STZ-induced diabetic rat; high glucose-conditioned Schwann cell exosomes	Unclear	Unclear	Unclear	High (behavioral endpoints)	Low	Low	High
Lu et al[24], 2025	Systematic review/meta-analysis (SYRCLE applied to included studies)	STZ-induced diabetic rodents; cell-derived exosomes	Unclear (majority of included studies)	Unclear (majority)	Unclear (majority)	High (majority of included studies)	Low	Low	High (majority of included studies)
Karkhaneh Yousefi et al[25], 2025	Human RCT (Cochrane RoB 2.0)	Diabetic foot ulcer patients; Wharton’s jelly MSC-derived exosomes	Low	Low	Some concerns (personnel not blinded to treatment)	Some concerns (outcome assessors partially blinded)	Low	Low	Some concerns

STZ: Streptozotocin; IV: Intravenous; MSC: Mesenchymal stem cell; RCT: Randomized controlled trial; SYRCLE: Systematic Review Centre for Laboratory animal Experimentation; RoB: Risk of Bias.

Table 7 Summary of methodological considerations

Methodological considerations: Isolation method	Advantages	Limitations	Recommendation	Synthetic insight: PDN-specific consideration
Differential ultracentrifugation	Reference standard; widely used	Co-pellets aggregates, lipoproteins, other vesicle subtypes; operator-dependent	Requires standardized protocols (rotor type, g-force, time)	Most widely used in PDN biomarker studies to date, enabling cross-study comparison; however, co-pelleting of lipoproteins is a particular concern in diabetic patients, who frequently have dyslipidemia that elevates plasma lipoprotein levels
Size exclusion chromatography	Superior purity; maintains vesicle integrity	Reduced yield; requires specialized equipment	Suitable for protein biomarker studies	Preferred for NfL and protein biomarker studies in PDN given superior removal of lipoprotein contamination; recommended when exosomal protein rather than miRNA cargo is the primary analytical target
Polymer-based precipitation	Simple; scalable	Substantial non-vesicular contamination (protein aggregates, lipoproteins)	Caution advised for protein cargo analysis	Particularly unreliable in diabetic plasma due to elevated lipoprotein and acute-phase protein levels; not recommended for PDN biomarker discovery or validation studies
Immunoaffinity capture	High specificity for vesicle subpopulations	Requires validated antibodies; limited by incomplete knowledge of surface proteomes	Preferred for cell-type-specific exosome analysis	Critical enabling technology for PDN specificity: L1CAM capture for neuronal and MOG capture for Schwann cell-derived exosomes directly addresses the neural-signal dilution problem in plasma and is the recommended approach for PDN biomarker validation cohorts

NfL: Neurofilament light chain; L1CAM: L1 cell adhesion molecule; MOG: Myelin oligodendrocyte glycoprotein; PDN: Painful diabetic neuropathy; miRNAs: MicroRNAs.