Microglia and different types of cell death after spinal cord injury

Across SCI models, NLRP3 and autophagy are tightly coupled and bidirectionally regulated. Inhibiting NLRP3 (MCC950 or knockdown) elevates mTOR signaling, reduces autophagic flux (↓LC3-II/Beclin-1, ↑p62), limits autophagic cell death, promotes an M2-like microglial phenotype, and improves locomotion; rapamycin reverses these effects, defining a functional NLRP3–mTOR–autophagy axis in vivo and in BV-2 microglia (Tian et al., 2025). In contrast, targeted autophagy induction can eliminate NLRP3 itself. Cannabinoid receptor-2 (CB2) activation (JWH-133) triggers AMPK/ULK1-dependent autophagy (↑microtubule-associated protein 1 light chain 3 beta (LC3B), ↓p62), drives K48-linked ubiquitination, and routes NLRP3 to autophagosomes, yielding M2 polarization, reduced neuroinflammation/demyelination, and better recovery (Jiang et al., 2022). Mitochondria-directed strategies—urolithin A and ginkgolide B—boost putative kinase 1 (PINK1)-Parkin mitophagy, clear damaged mitochondria, lower mitochondrial ROS (mtROS), and block ASC specks, caspase-1, IL-1β, and GSDMD cleavage; 3-methyladenine (3-MA), bafilomycin a1 (BafA1), or Parkin knockdown abrogate these benefits, implicating mitophagy as a key brake on NLRP3-driven pyroptosis and pain (Chen et al., 2025; Liang et al., 2024). Zinc confers dual autophagy-centric control by suppressing long non-coding ribonucleic acid (lncRNA) XIST/upregulating miR-374a-5p to reduce NLRP3 expression and by enhancing ubiquitin-mediated autophagic degradation; autophagy blockade reinstates inflammasome activity (Zhao et al., 2022; Lin et al., 2021). MicroRNA (miRNA) regulation offers another layer of control. miR-99b-3p targets matrix metallopeptidase 13 (MMP13) to increase autophagic flux, repress NLRP3, and attenuate caspase-1/GSDMD-dependent pyroptosis, easing neuropathic pain (Gao X. et al., 2024). Photobiomodulation downregulates Toll-like receptor 2 (TLR2), normalizes autophagy markers, and dampens NLRP3–caspase-1–IL-1β signaling with motor gains (Zuo et al., 2025).

Together, these data indicate that benefit derives not from unidirectional changes in autophagy but from restoring an optimal autophagy set-point that restrains pathological inflammasome signaling without triggering autophagic cell death.

Limitations across studies include reliance on rodent models and short-term outcomes, limited cell-specific genetics and safety data, frequent use of bulk LC3-II/p62 without rigorous flux assays, and focus on single pathways despite crosstalk between mTOR/AMPK, ubiquitination, and mitophagy.

The PI3K/protein kinase b (AKT)/mTOR axis sits at the center of nutrient sensing and therefore at the core of autophagy regulation. In nutrient-replete cells, class-I PI3K activates AKT, which in turn activates mTORC1 (Figure 3); mTORC1 then phosphorylates the ULK1 complex, blocking autophagy initiation. Energy stress, AMPK activation, growth-factor withdrawal, or upstream inhibition of PI3K/AKT all relieve this brake and trigger autophagic flux, thereby maintaining metabolic homeostasis. Parallel modulators such as brain-derived neurotrophic factor (BDNF) (via tropomyosin receptor kinase B (TrkB)/AKT) and g protein-coupled receptor kinase 2 (GRK2) [which constrains p38-mitogen activated protein kinase (MAPK)] fine-tune the pathway in neural tissue, while AMPK provides an energy-status override that can suppress mTORC1 even when PI3K is active (Zhang HP. et al., 2024; Xu et al., 2023).

Several lines of evidence show that sustained PI3K/AKT/mTOR activity blocks autophagy in spinal cord macrophages and thereby aggravates secondary injury.

Lipid-engorged "foamy" macrophages that accumulate after traumatic SCI depend on PI3K activity to phagocytose myelin debris while simultaneously suppressing autophagic clearance of lipid droplets; Torin-2 (a PI3K/mTOR inhibitor) reverses the phenotype, whereas rapamycin (mTORC1-specific) does not, underscoring a PI3K-dependent, mTOR-independent step in this context (Ryan CB. et al., 2024).

Bioinformatic mapping of immune subsets in injured cords identifies mTOR as a hub gene in proinflammatory M1 macrophages; increased PI3K-AKT-mTOR signalling coincides with low LC3B-II and high p62, perpetuating cytokine output (TNF-α, IL-1β) (Ju et al., 2025).

In ischemic SCI, PI3K/AKT phosphorylation falls transiently, yet further inhibition (mild hypothermia or rapamycin) is required to push autophagy high enough to curb microglial over-activation and neuroinflammation (Li J. et al., 2022).

Astragaloside IV dampens mTORC1, restores neuronal autophagic flux, and switches microglia toward an M2 phenotype, together improving functional recovery (Lin et al., 2020).

Inhibiting casein-kinase-2 with Dimethylamino-4,5,6,7-tetrabromo-1H-benzimidazole elevates Beclin-1/LC3II, lowers p62, and forces M1 to M2 microglial repolarization; blocking autophagy with 3-MA abolishes these effects, proving autophagy is the mechanistic bridge (Liu et al., 2023).

Microglia-derived exosomes loaded with resveratrol enter the spinal cord, raise LC3B and Beclin-1, and repress apoptosis via PI3K activation; 3-MA negates the benefit, confirming pathway dependence (Fan et al., 2020).

Small extracellular vesicles (sEVs) from M2 bone-marrow macrophages deliver miR-421-3p to neurons, directly targeting mTOR and thereby restoring autophagic flux and reducing cell death (Wang et al., 2020).

Voluntary treadmill running suppresses spinal BDNF, releases the AKT/mTOR brake, expands autophagic flux in microglia, favors M2 polarization, and lessens neuropathic pain (Bai et al., 2022).

High-voltage pulsed radiofrequency (HVPRF) applied to dorsal root ganglia reinstates GRK2, inhibits p38, restores autophagic flux in microglia, and alleviates both pain and depressive-like behavior; GRK2 knock-down reverses these gains (Xu X. et al., 2024).

Separate HVPRF work shows an autophagy-dependent drop in TNF-α and rise in IL-10 within dorsal horn microglia, correlating with structural nerve repair (Chen et al., 2024).

Electroacupuncture (EA) activates AMPK, suppresses mTOR, clears p62 in dorsal-root-ganglion macrophages, and produces potent analgesia (Xu et al., 2022).

Zinc drives Signal Transducer and Activator of Transcription 3 (STAT3)/Forkhead box O3a (FoxO3a)-dependent mitophagy, restores mitochondrial ATP output, and reduces ROS in microglia, indirectly easing secondary damage (Cui et al., 2025).

Collectively, these studies shows that PI3K/AKT/mTOR pathway operates as a molecular regulator on autophagy after SCI. Hyper-activation of PI3K or mTOR sustains lipid accumulation, inflammatory cytokine release, and neuropathic pain, whereas targeted inhibition (pharmacological, genetic, or physical) re-engages autophagic clearance and tilts immune cells toward a reparative M2 profile. Importantly, several interventions (Torin-2, AMPK activation, BDNF suppression, GRK2 restoration) achieve benefits that rapamycin alone does not, highlighting PI3K- or upstream-specific checkpoints and cross-talk with energy-sensing modules.

However, most data come from rodent models. Human verification is urgently needed. The temporal window for safe mTOR inhibition versus essential anabolic repair remains undefined. Few studies address long-term outcomes beyond acute or sub-acute phases. The differential roles of resident microglia versus infiltrating macrophages are not consistently dissected, leaving cell-type specificity uncertain.

Comparative studies using conditional knock-outs or cell-specific drug delivery could clarify lineage-specific requirements. Combining PI3K/mTOR modulation with mitophagy boosters or metabolic support may further refine therapeutic windows. Finally, dose-controlled clinical trials of pathway modulators such as metformin, alongside biomarker-guided imaging of autophagic flux, are warranted to translate these promising pre-clinical findings into neuro-restorative strategies for patients with SCI.

Autophagy functions as a hub that tunes microglial and vascular responses after SCI. On the inflammatory axis, excess lipid-peroxidation product 4-hydroxy-2-nonenal (4-HNE) (in aldose-reductase deficiency) covalently modifies IKK—especially IKKα—diverting it to p62-mediated autophagolysosomal degradation, thereby suppressing canonical NF-κB signalling and limiting secondary damage; at sub-toxic levels, 4-HNE transiently activates NF-κB (Han et al., 2022). Restoring lysosomal competence with progranulin (↑LAMP1 (lysosome-associated membrane protein 1), cathepsin D) re-opens autophagic flux (↑Beclin-1, ↑LC3-II/I, ↓p62), promotes an M2 shift, reduces NF-κB–dependent cytokines, and elevates IL-10 (Shi et al., 2022).

Hypoxia/ischemia imposes a biphasic program: brief hypoxia-inducible factor-1α (HIF-1α) stabilization upregulates Beclin-1/LC3 to curb early TNF-α/IL-1β, whereas sustained HIF-1α overdrives flux and precipitates autophagy-dependent death; late HIF-1α dampening rescues viability but re-elevates cytokines (Wang et al., 2017). After ischemia-reperfusion (I/R), an <24 h "early-autophagy" window is protective, while by 72 h persistent Beclin-1/LC3-II with reduced Bcl-2 signals autophagic cell death (Fang et al., 2016). Super-resolution structured-illumination microscopy adds spatial detail to this temporal story. lysosomes shift from subplasmalemmal clusters at 48 h to a perinuclear configuration by day 7, aligning trafficking with a transition from phagocytosis to bona fide autophagy (Fumagalli et al., 2019). HIF-1α-driven estrogen-related receptor-α (ERRα) further tempers p38-MAPK/NF-κB, narrows Beclin-1/LC3 windows while restoring p62, and promotes Fibronectin type III domain–containing protein 5 (FNDC5)-mediated BDNF release, coupling metabolic control to anti-inflammatory signalling and survival (Deng et al., 2022).

Myelin-derived lipids push microglia into a lipid-droplet–accumulating state with sluggish phagocytosis; miR-223 silences cathepsin B to enhance lipophagy, accelerate droplet clearance, and suppress IL-1β (Ma et al., 2024). Conversely, CD36 upregulation blocks AMPK and TFEB, stalling autophagosome biogenesis and lysosome formation; CD36 inhibition restores lipophagy, limits pyroptosis, and improves locomotion (Wang BN. et al., 2025). Systemically, M1-skewed macrophage/microglia deliver nitric oxide (NO) to Leydig cells, shutting down AMPK/ULK1-S555 and activating mTOR/ULK1-S757, freezing lipophagy and causing hypogonadism; Scavenging NO (cPTIO) or ULK1 agonism reverses this (Zhuang et al., 2024).

Epigenetic control shapes these programs: single-cell maps place autophagy genes (cathepsin D, CD68, CD81, Tyrobp) in phagocytic MG2 microglia, while histone deacetylase 3 (HDAC3) in proliferative MG4 coordinates autophagy-immune networks; HDAC3 inhibition re-routes trajectories and dampens these transcripts, implicating chromatin-level regulation of autophagy and inflammation (Wahane et al., 2021). In microvascular endothelial cells, injury-induced histone deacetylase 6 (HDAC6) elevation impairs autophagosome–lysosome fusion; Tubastatin-A restores flux, enhances myelin clearance, and reduces inflammation (Wu et al., 2024a).

Cell-cycle control intersects via cyclin-dependent kinase 1 (CDK1). Elevated in chronic SCI, it drives autophagy and microglial activation; knockdown lowers LC3/Beclin-1, curbs apoptosis, and is neuroprotective (Nie et al., 2022).

Vascular-immune crosstalk also hinges on autophagy: M1 microglia upregulate lymphatic vessel endothelial hyaluronan receptor-1 (LYVE-1)/podoplanin and secrete vascular endothelial growth factor-C (VEGF-C), which feeds back via vascular endothelial growth factor receptor-3 to reinforce M1 polarization while suppressing autophagy (Xian et al., 2025); endothelial autophagy clears opsonized myelin via ATG5-dependent machinery but triggers IL-6/monocyte chemoattractant protein-1 (MCP-1)/VEGF release, TGF-β1-driven endothelial-to-mesenchymal transition, and scarring (Zhou et al., 2019); and iron overload (IRP1↑ (Iron Regulatory Protein 1), transferrin-receptor-1–mediated import, ferritin↓) activates NF-κB/autophagy/cytokine cascades that injure neurons and drive pain, mitigated by chelation, NOS blockade, or microglial silencing (Meng et al., 2017).

Taken together, therapeutic benefit derives not from uniformly raising or lowering autophagy, but from tuning its timing, location, and metabolic context to restrain NF-κB–driven inflammation, avoid autophagic death, and coordinate vascular remodeling after SCI.

A growing pharmacological strategy for SCI focuses on restoring the microglial autophagy–inflammation balance. Drugs that either inhibit the canonical mTOR brake on autophagy, activate upstream energy sensors, or modulate chemokine or cell-cycle signalling have all shown the capacity to tilt microglia away from a chronically proinflammatory state and thereby limit secondary neurodegeneration.

Across two independent rat SCI studies, systemic or intrathecal rapamycin consistently suppressed mTORC1 (↓p-p70S6K), lowered the autophagy cargo adaptor p62/sequestosome-1, and increased Beclin-1 and LC3-II accumulation, indicating restored autophagic flux. Concomitantly, microglia/macrophage infiltration (CD11b⁺/CD14⁺) and TNF-α secretion fell, while neuronal survival and axon regrowth improved. A modest rise in phosphorylated AKT further supported pro-survival signalling. Histology confirmed abundant autophagic vacuoles within phagocytic cells and reduced cavitation, identifying rapamycin as a dual autophagy enhancer and inflammation dampener in the injured cord (Firat et al., 2021; Goldshmit et al., 2015).

Metformin drove a phenotypic switch from M1 to M2 microglia, boosted myelin-debris phagocytosis and preserved white matter. Mechanistically, AMPK activation and secondary mTOR inhibition re-established autophagosome–lysosome fusion; blocking autophagy with 3-MA abolished these benefits, underscoring autophagy as the primary driver of metformin's neuroprotection (Wu et al., 2022).

Trehalose prevented the accumulation of lipid-laden "foamy" macrophages whose defective autophagy perpetuates scarring. By up-regulating TFEB, it coordinately induced autophagy- and lysosome-related genes, restored lipid clearance, curtailed proinflammatory cytokines, reduced fibrotic borders and improved motor scores. Chloroquine negated these gains, confirming TFEB-dependent autophagy as the key target (Ma et al., 2025).

The C-C chemokine receptor type 1 antagonist, BX471, dampened microglial activation and TNF-α/IL-1β release while simultaneously increasing Beclin-1 and LC3-II and lowering p70S6K (70-kDa ribosomal protein S6 kinase) and p62, thereby re-establishing autophagic flux and limiting neuronal apoptosis (Hasan et al., 2025).

In a chronic SCI model, heightened CDK1 expression paralleled excessive autophagy and microgliosis. CDK1 shRNA curtailed LC3/Beclin-1 accumulation and Iba1 levels, suggesting that aberrant CDK1 activity sustains both autophagic stress and inflammatory activation long after the initial insult (Nie et al., 2022).

Collectively, these studies highlight autophagy restoration as a unifying mechanism by which diverse small molecules blunt microglia-driven secondary injury. Rapamycin and metformin target the canonical AMPK/mTOR axis and therefore possess predictable metabolic side-effects but robust efficacy across time points. Trehalose and BX471 reveal alternative entry points, lysosomal biogenesis and chemokine signalling, broadening the therapeutic palette and potentially reducing systemic immunosuppression. CDK1 inhibition, while mechanistically intriguing, still lacks pharmacological tractability and a clear safety profile.

Major limitations remain. All data are pre-clinical, with short follow-up windows and variable dosing paradigms. Most studies assess single-agent therapy, despite the multifactorial nature of SCI. Autophagy read-outs are largely surrogate (LC3 puncta, p62), not dynamic flux measurements in vivo. Future work should standardize injury models, incorporate sex and age variables, employ real-time autophagy reporters, and test rational drug combinations to determine whether convergent autophagy modulation translates into durable neurological recovery.

A growing body of work shows that plant-derived small molecules can fine-tune microglial autophagy after SCI, thereby curbing neuroinflammation, limiting secondary neuronal loss and, ultimately, improving functional outcome. Although their chemical scaffolds differ, most of these agents converge on two interconnected nodes highlighting common points of therapeutic leverage. Astragaloside IV, fisetin and salidroside all enhance autophagic flux by dampening mTORC1 activity, yet each does so through a slightly different upstream switch.

Astragaloside IV lowers p-mTOR and p-p70S6K without altering p-AKT, implicating AMPK as the likely brake on mTORC1. The net result is M2-skewed microglia, reduced cytokine release and concomitant protection of neurons from apoptosis, translating into better locomotor scores after SCI (Lin et al., 2020).

Fisetin directly activates AMPK, which in turn represses mTOR, restores autophagic flux and suppresses TNF-α/IL-6 output from M1-polarized microglia. This dual action relieves neuronal apoptotic pressure and accelerates behavioral recovery (Liu Y. et al., 2024).

Salidroside also recruits AMPK to block mTORC1, tipping microglia toward an M2 phenotype while curbing iNOS and cyclooxygenase-2 (COX-2) expression. Enhanced autophagy appears to be the linchpin linking salidroside's anti-inflammatory and neuroprotective effects (Wang et al., 2018).

Clearing dysfunctional mitochondria is another recurring strategy. Ginkgolide B activates the PINK1–Parkin pathway, promoting mitophagy in microglia, lowering mtROS, preventing NLRP3 inflammasome activation and, consequently, easing neuropathic pain. Genetic or pharmacological blockade of Parkin abolishes its benefit, underscoring mitophagy as the critical driver (Liang et al., 2024).

Urolithin A, a gut-derived ellagitannin metabolite, similarly boosts PINK1-Parkin-dependent mitophagy, thereby dampening ROS-mediated NLRP3 priming, Caspase-1 cleavage and GSDMD-driven pyroptosis. The protection disappears when autophagy is blocked by 3-MA, again pinpointing mitophagy as the therapeutic nexus (Chen et al., 2025).

Some nutraceuticals modulate autophagy indirectly by correcting upstream inflammatory circuits. Resolvin D2 interrupts NF-κB-induced miR-155 upregulation, restores PTEN, inhibits AKT/mTOR and re-opens autophagy flux. With the block removed, microglia revert to an M2-like state and neuropathic pain is eased in SCI models (Yang L. et al., 2023).

Leveraging microglia themselves as drug carriers can amplify phytochemical efficacy. Microglia-derived exosomes loaded with resveratrol cross the injured cord, stimulate neuronal autophagy through PI3K signalling and suppress apoptosis, producing marked motor recovery. Inhibition of PI3K or autophagy neutralizes these gains, confirming pathway specificity (Fan et al., 2020).

Collectively, these studies reinforce the centrality of autophagy, especially AMPK–mTOR regulation and PINK1-Parkin-mediated mitophagy, in dictating microglial phenotype and downstream neuro-inflammatory tone after SCI. The convergence of chemically diverse phytochemicals on a limited set of intracellular hubs strengthens the rationale for targeting autophagy but also raises several caveats. First, most data derive from single-agent, short-term rodent experiments; the durability of benefit and potential off-target metabolic effects in larger animals or humans remain unknown. Second, because nearly every compound alters the AMPK/mTOR axis, disentangling autophagy-specific actions from broader metabolic consequences (e.g., altered glycolysis or lipid utilization) will require more nuanced readouts. Third, the favorable results with exosome-packaged resveratrol highlight the importance of delivery and bioavailability—an under-explored variable for many phytochemicals that suffer from poor solubility or rapid clearance. Finally, head-to-head comparisons and combination trials are lacking; such work is essential to determine whether these agents offer additive or redundant benefits. In summary, natural products constitute a promising, mechanistically coherent toolkit for re-programming microglial autophagy after SCI, but their translation will hinge on rigorous pharmacokinetic optimisation, long-term safety studies and careful pathway deconvolution.

Microglia are among the first responders to SCI, yet they rapidly succumb to mitochondrial fragmentation, oxidative stress, and energy failure. Restoring mitochondrial quality control has therefore become a key therapeutic goal, and zinc has emerged as the most extensively studied trace element in this context (Cui et al., 2025).

In a recent mouse SCI model, systemic zinc supplementation reinstated mitophagy and preserved mitochondrial membrane potential by activating a STAT3/FoxO3a/Superoxide Dismutase 2 (SOD2) axis. Zinc curtailed FoxO3a phosphorylation, allowing its nuclear import and the transcription of antioxidant (SOD2) and fusion/biogenesis genes (peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC-1α), optic atrophy protein 1 (OPA1), Mitofusin 2), while simultaneously suppressing the Fission Protein Dynamin-related Protein 1. The net effect was reduced mtROS, improved ATP production, and better locomotor recovery (Cui et al., 2025).

A complementary study showed that zinc preferentially expands a vascular endothelial growth factor A (VEGF-A)-secreting microglial subcluster (MG4). The released VEGF-A engages PI3K/AKT/Bcl-2 signaling in both microglia and endothelial cells, fostering endothelial proliferation, attenuating microglial apoptosis, and alleviating local hypoxia. Consequently, vascular regeneration and axonal sparing are enhanced, highlighting a dual mitochondrial-vascular mechanism by which zinc remodels the post-traumatic niche (Deng et al., 2023).

Beyond dietary Zinc, methylprednisolone—still the only FDA-approved acute anti-inflammatory for SCI—can harness endogenous zinc to induce Beclin-1–dependent autophagic death of over-activated microglia. The glucocorticoid upregulates the zinc importer Zrt/Irt-like Protein 8 (ZIP8), raises the labile‐zinc pool, suppresses NF-κB, and triggers a protective wave of autophagy. Chelation with Tetrakis (2-pyridulmethy-1) Ethylenediamine (TPEN) abolishes these benefits, underscoring zinc's central role in the drug's neuroprotective profile (Li et al., 2016).

Collectively, these studies converge on zinc as a pleiotropic modulator that couples mitochondrial repair, angiogenesis, and immunomodulation after SCI. By acting both as a micronutrient and as a mobilizable second messenger, zinc targets canonical (STAT3/FoxO3a) and non-canonical (PI3K/AKT/Bcl-2, Beclin-1) autophagy checkpoints, while simultaneously shaping the vascular microenvironment. However, several gaps remain. First, the dose-response window between beneficial mitophagy and detrimental microglial loss is still ill-defined. Second, most evidence derives from young male rodents; age-, sex-, and comorbidity-dependent zinc homeostasis warrants systematic study. Finally, cross-talk with other trace elements (e.g., copper, iron, manganese) and with zinc transporters in peripheral immune cells is largely unexplored. Addressing these issues will be critical before "zinc-centric" therapies can be translated into precision medicine for human SCI.

Hypoxia-preconditioned adipose-derived Mesenchymal Stem Cell (MSC) exosomes deliver circ-Astn1 into injured spinal cords, where the Circular RNA sequesters miR-138-5p and thereby de-represses ATG7. The Circular RNA ASTN1/miR-138-5p/ATG7 axis re-engages autophagy in microglia, blunts inflammation-linked apoptosis, and translates into superior neurological recovery; genetic disruption of any node in the axis abolishes the benefit (Shao et al., 2025).

Bone-marrow MSC exosomes rich in miR-21a-5p silence the E3 ubiquitin ligase PELI1, a brake on autophagy. Restored autophagic flux suppresses NLRP3 inflammasome activation, limits microglial pyroptosis, and accelerates motor recovery after SCI (Gu et al., 2024).

MSC-derived EVs carrying miR-99b-3p mitigate neuropathic pain by dampening PI3K/AKT/mTOR signalling—an autophagy-inhibitory axis. Upregulation of LC3-II and Beclin-1, coupled with p62 clearance, curbs LPS-driven microglial activation and inflammation in vivo, easing pain behaviors (Gao et al., 2023).

Olfactory-mucosa MSCs foster a switch from proinflammatory M1 to reparative M2 microglia through paracrine, exosome-mediated cues. Autophagy appears permissive for this shift, as reflected by reciprocal NOS2/ARG1 expression and enhanced axonal regrowth in SCI models (Wang X. et al., 2024).

Peripheral macrophage-derived exosomes likewise promote M2 polarization by shutting down PI3K/AKT/mTOR in recipient microglia, heightening autophagy, dampening cytokine output, and preserving neurons—together improving functional outcome (Zhang et al., 2021).

sEVs from alternatively activated (M2) bone-marrow macrophages deliver miR-421-3p to neurons; by directly targeting mTOR, the miRNA boosts autophagic flux and reduces apoptosis, translating into substantial locomotor improvement post-SCI (Wang et al., 2020).

Neural stem-cell (NSC)-derived sEVs provide the scaffold protein 14-3-3τ, which binds Beclin-1 to nucleate autophagosomes. Heightened autophagy lowers neuronal apoptosis and glial inflammation; over-expression or knock-down of 14-3-3τ, respectively, magnifies or blunts these gains (Rong et al., 2019a).

A complementary NSC-sEV study showed that broad autophagy activation—evidenced by LC3B and Beclin-1 upregulation with p62 loss—suppresses NO release and proinflammatory cytokines in microglia; the protective effect disappears when autophagy is blocked with 3-MA (Rong et al., 2019b).

Finally, microglia can be coaxed to release resveratrol-loaded exosomes that cross the blood-spinal cord barrier (BSCB). The phytochemical cargo activates PI3K-dependent autophagy in spinal neurons, curtails apoptosis, and markedly improves hind-limb function; PI3K inhibition negates these effects (Fan et al., 2020).

Across diverse cell sources—mesenchymal, neural, and immune—extracellular vesicles converge on a single therapeutic motif which is reinstating autophagy to reprogram neuroinflammation and promote tissue preservation after SCI. Most studies identify a miRNA or protein "switch" (miR-138-5p, miR-21a-5p, miR-99b-3p, miR-421-3p, 14-3-3τ, resveratrol) that ultimately suppresses the PI3K/AKT/mTOR axis or liberates core autophagy genes (ATG7, Beclin-1). Functional gains are robust in rodent models, yet several gaps remain. Direct head-to-head comparisons of EVs sources and loading strategies are lacking. Dosing, biodistribution, and manufacturing reproducibility are seldom addressed. long-term safety and off-target autophagy induction in non-neural tissues remain unknown. Future work should standardize EVs production, dissect combinatorial cargos, and advance toward large-animal and, ultimately, clinical testing to clarify whether autophagy-centric EVs therapy can fulfil its translational promise in human SCI.

Targeted delivery of genes, non-coding RNAs, and trophic proteins is emerging as a precise way to recalibrate microglial autophagy-inflammation crosstalk after SCI. By fine-tuning key signalling nodes these approaches dampen secondary neuroinflammation and neuropathic pain while sparing (or even rescuing) neighboring neurons.

miR-15a restores autophagy by silencing AKT3. In the chronic constriction injury (CCI) model, miR-15a is markedly downregulated, a change that coincides with heightened cytokine release and stalled autophagic flux in dorsal-horn microglia. Intrathecal delivery of a miR-15a agomir reverses this deficit. AKT3 expression falls, Beclin-1 and LC3-II rise, p62 declines, and pain behaviors improve. These data place the miR-15a/AKT3 axis upstream of microglial autophagy and identify it as a tractable target for neuropathic pain (Cai et al., 2020).

miR-99b-3p suppresses NLRP3-driven pyroptosis via MMP13 inhibition. CCI also triggers NLRP3 inflammasome activation, caspase-1 cleavage, and IL-1β release. Over-expressing miR-99b-3p directly represses MMP13, thereby enhancing autophagic flux (↑LC3-II, ↓p62). Restored autophagy, in turn, dampens inflammasome priming and reduces neuropathic pain behaviors, highlighting a feed-forward loop in which miR-99b-3p curtails pyroptosis by promoting autophagy (Gao X. et al., 2024).

Glial cell line–derived neurotrophic factor (GDNF) gene therapy blocks p38-MAPK/Protein Kinase C (PKC) signalling to quell microglial activation. GDNF plasmid delivery mitigates CCI-induced microgliosis, down-regulating Beclin-1, IL-1β, IL-6, and MMP-9. Mechanistically, GDNF suppresses p38-MAPK and PKC pathways, leading to reduced autophagy initiation in over-activated microglia, lower neuronal apoptosis, and attenuated mechanical allodynia (Chou et al., 2014).

Collectively, these studies converge on a unifying theme. Over-activated microglia in SCI sit at a tipping point where too little autophagy sustains inflammasome signalling and chronic pain, whereas carefully re-engaging autophagic flux normalizes the inflammatory milieu. Both miRNA replacement (miR-15a, miR-99b-3p) and trophic-factor gene delivery (GDNF gene) succeed by hitting distinct molecular switches (AKT3, MMP13, p38/PKC) yet ultimately funnel their effects through the same autophagy-centered checkpoint. While encouraging, important caveats remain—small sample sizes, single-sex cohorts, and limited long-term read-outs leave efficacy and safety unanswered for chronic phases of SCI. Future work should compare combinatorial or sequential delivery of miRNAs and neurotrophic genes, test off-target effects in larger animal models, and couple behavioral improvements to unbiased single-cell 'omics to confirm microglia-specific reprogramming.

Physical and neuromodulatory interventions use non-pharmacological energy to re-engage the autophagy pathway in microglia or peripheral macrophages, thereby dampening neuroinflammation, limiting secondary tissue loss, and reducing neuropathic pain after SCI. A unifying theme across the studies below is that restoring autophagic flux realigns the balance between pro- and anti-inflammatory signaling.

In a spared-nerve-injury model, 85 V HVPRF restored autophagic flux in dorsal-horn microglia, lowered TNF-α, raised IL-10, repaired ganglion ultrastructure, and produced robust analgesia. The data point to a TNF-α/IL-10 neuro-immune shift driven by enhanced formation and clearance of autophagosomes and autolysosomes (Chen et al., 2024).

Using the same modality, Xu et al. showed that nerve injury suppresses GRK2, leading to p38 MAPK over-activation and stalled autophagic flux. HVPRF reinstated GRK2, curtailed p-p38, reinvigorated autophagy, and simultaneously relieved mechanical allodynia and depression-like behavior; GRK2 knock-down abrogated all benefits, identifying GRK2 as an obligatory switch in the autophagy-pain loop (Xu X. et al., 2024).

Zou et al. found that SCI elevates TLR2, suppresses autophagy (LC3, Beclin-1 ⇣, p62 ⇡), and activates the TLR2/NLRP3/Caspase-1/IL-1β cascade. Red/near-infra-red photobiomodulation downregulated TLR2, restored autophagic markers, curtailed inflammasome activity, and improved hind-limb recovery, implicating TLR2 as a photomodulatable hub linking innate immunity to autophagy (Zuo et al., 2025).

Spared-nerve injury stalled autophagy in dorsal root ganglia macrophages (p62 ⇡, LC3-II, Beclin-1 altered). EA phosphorylated AMPK, inhibited mTOR, normalized autophagic flux, and lessened pain hypersensitivity—highlighting macrophage-rather-than-neuron autophagy as the decisive target of this ancient neuromodulatory technique (Xu et al., 2022).

Running exercise lowered spinal BDNF, thereby quenching AKT/mTOR activity, boosting microglial autophagy, and driving an M1 to M2 phenotypic shift. Autophagy inhibition reversed the analgesic benefit, confirming that exercise-induced autophagy is causal rather than correlative in mitigating neuropathic pain (Bai et al., 2022).

Collectively, physical/neuromodulatory strategies converge on the re-activation of microglial or macrophage autophagy as a mechanistic fulcrum for neuro-protection and pain relief after SCI. Despite diverse energy sources studies documented that impaired autophagic flux is a common pathological denominator following SCI or peripheral nerve injury. Restoration of that flux consistently shifts cytokine output toward an anti-inflammatory profile (IL-10, M2 markers) and attenuates canonical inflammasome signaling (TNF-α, IL-1β, p38 MAPK, TLR2–NLRP3). Up-stream gating molecules differ by modality—GRK2 for HVPRF, TLR2 for photobiomodulation, AMPK for EA, BDNF for exercise—yet all funnel through the mTOR axis, underscoring the pathway's therapeutic tractability. Future works should integrate multimodal stimulation with precise autophagy reporters, expand to large-animal models, and align dosing regimens with feasible clinical protocols to move these promising biophysical interventions from bench to bedside.

Moreover, zein-based nanoparticles co-delivering celastrol, metformin, and everolimus restore autophagy in BV2 microglial cells by inhibiting mTOR and enhancing mitochondrial function. Celastrol also activates the nuclear receptor subfamily four group A member 1 (NR4A1; Nur77) pathway, inducing translocation to mitochondria and promoting mitophagy. These effects suppress inflammatory cytokine release (e.g., TNF-α, IL-6), reduce oxidative stress, and protect neurons. Enhanced autophagic activity improves spinal cord tissue repair and motor function in rats, highlighting autophagy modulation in microglia as a therapeutic avenue for SCI (Wang J. et al., 2024).

A growing body of evidence shows that micro- and macro-non-coding RNAs orchestrate the inflammasome machinery that drives microglial pyroptosis after SCI. Most of the work to date converges on a simple paradigm. miRNAs tend to brake pyroptosis, whereas longer non-coding species (lncRNAs and circRNAs) often release that brake—although exceptions exist. Below, the principal studies are arranged to emphasize this continuum from miRNA-mediated restraint to lncRNA/circRNAs-mediated activation.

Resveratrol restores locomotor function after SCI in rats largely by boosting microglial miR-124-3p. The miRNA silences Death-Associated Protein Kinase 1 (DAPK1), uncoupling NLRP3 from its activating scaffold and suppressing Caspase-1/GSDMD cleavage. Genetic loss of DAPK1 or forced miR-124-3p expression phenocopies the drug, cutting IL-1β release and rescuing cell viability (Li D. et al., 2025).

MSC–derived EVs loaded with miR-22 deliver the miRNA directly into microglia. miR-22 pairs with the 3′-UTR of GSDMD, blunting pore formation and the IL-1β/IL-18/TNF-α surge, thereby accelerating neurological recovery beyond that achieved by unmodified EVs (Sheng et al., 2021).

Regulatory T cells (Treg) secrete exosomes enriched in miR-709. Once inside microglia, miR-709 downregulates NF-κB activating protein (NKAP), damping inflammasome priming. Depleting Tregs worsens motor scores, whereas adoptive transfer or purified exosomes reverses the deficit, highlighting a physiological checkpoint on pyroptosis (Xiong et al., 2022).

Switching focus from microglia to their neuronal neighbors, exosomes from M2-polarized microglia ferry miR-672-5p into injured neurons. The target is AIM2, whose repression collapses the AIM2/ASC/Caspase-1 axis, limits neuronal pyroptosis, and supports axon regrowth—an indirect yet important microglia-to-neuron relay (Zhou et al., 2022).

Induced pluripotent stem cell–derived neural stem cells (NSCs) release exosomes rich in let-7b-5p. In microglia, the miRNA represses Leucine-Rich Repeat Inherited Protein 3 (LRIG3), upstream of NLRP3, curtailing GSDMD cleavage and cytokine release. Animals receiving either the cells or isolated exosomes show comparable gains in motor function (Liu J. et al., 2024).

In contrast, SCI upregulates the lncRNA F630028O10Rik via Toll-like receptor 4 (TLR4)/Signal Transducer and Activator of Transcription 1 (STAT1) signalling. Acting as a competing endogenous RNA, it sponges miR-1231-5p, de-repressing Collagen Type I Alpha 1 Chain (Col1a1) and activating the PI3K/AKT cascade to fuel microglial pyroptosis. Blocking any node in this axis diminishes cell death and improves outcome (Xu et al., 2020).

The circular RNA circ0000381 sequesters miR-423-3p, indirectly boosting NLRP3 expression. Most data indicate that this escalates inflammasome activation and cytokine release; paradoxically, circ0000381 knock-down was reported to increase pyroptosis in one model, pointing to cell-type or temporal complexities that remain unresolved (Zhang Y. et al., 2024).

Collectively, five independent reports converge on a common protective motif. Exogenous or endogenous miRNAs (miR-124-3p, -22, -709, -672-5p, let-7b-5p) target distinct but convergent nodes (DAPK1, GSDMD, NKAP, AIM2, LRIG3) to dampen the NLRP3 or AIM2 axis. Delivery vehicles range from small molecules that upregulate endogenous miRNA to designer stem-cell or immune-cell exosomes—highlighting translatable platforms already in pre-clinical use.

Conversely, longer non-coding species appear to lift this brake, with lncRNA F630028O10Rik and circ0000381 both relieving repression of pyroptosis-promoting genes. Whether this reflects a general pro-pyroptotic role for lncRNAs/circRNAs or merely injury-timed expression spikes is still uncertain, underscored by the contradictory data on circ0000381.

However, most studies rely on single-sex rodent models and acute time-points. Network cross-talk among multiple ncRNAs is largely unexplored. Off-target and dose-limiting toxicities of exosome therapies remain to be defined. Addressing these issues will determine whether ncRNA-based modulation of microglial pyroptosis can move from bench to bedside.

Transcription factors and chromatin-modifying enzymes orchestrate the intensity and duration of microglial pyroptosis after SCI. By modulating the expression, activation, or degradation of key inflammasome components, they tilt the balance between neuro-destructive and reparative responses.

SCI rapidly induces the transcription factor CCAAT/Enhancer-Binding Protein Beta (C/EBPβ) in microglia. C/EBPβ binds to the Fc Gamma Receptor I (Fcgr1) promoter, drives its expression, and thereby facilitates NLRP3-inflammasome priming. The ensuing rise in caspase-1 activity and GSDMD cleavage releases IL-1β, IL-18, IL-6 and TNF-α, which propagate neuronal apoptosis and hinder functional recovery. Genetic knock-down of either C/EBPβ or Fcgr1, or pharmacological blockade of caspase-1 with VX-765, breaks this feed-forward loop and improves behavioral outcomes, identifying the C/EBPβ–Fcgr1 axis as a tractable transcriptional target (Li J. et al., 2025).

In a chronic constriction injury model, activated spinal microglia upregulate the cytoplasmic de-acetylase HDAC6. HDAC6 enhances NF-κB nuclear translocation, boosts NLRP3 expression, and accelerates the caspase-1/GSDMD pathway, culminating in IL-18 release and pyroptotic cell death. Selective HDAC6 inhibition with ACY-1215 dampens microglial activation, lessens mechanical allodynia, and highlights epigenetic acetylation status as a switch that gates pyroptotic pain after SCI (Sun et al., 2023).

Conversely, the antioxidant transcription factor nuclear factor erythroid 2-related factor 2 (Nrf2) is suppressed after SCI. Restoring Nrf2 upregulates miR-146a, a microRNA that directly targets the GSDMD mRNA 3′-UTR. Reduced GSDMD translation lowers caspase-1 activation and attenuates IL-1β and IL-18 release. Both in vitro and in vivo, Nrf2 over-expression preserves microglial viability and limits tissue loss, positioning the Nrf2/miR-146a/GSDMD axis as an endogenous anti-pyroptotic circuit (Zhang et al., 2023).

The E3 ubiquitin ligase Tripartite Motif Containing 32 (TRIM32), downregulated after SCI, affords similar protection when re-introduced. TRIM32 binds the NLRP3 co-factor NIMA-related kinase 7 (NEK7) and catalyzes its K64-linked ubiquitination, targeting NEK7 for proteasomal degradation. With less NEK7 available to license NLRP3 oligomerization, caspase-1 and GSDMD activation fall, limiting IL-1β/IL-18 efflux and tissue damage. Thus, TRIM32 links post-translational ubiquitin editing to inflammasome restraint (Yu et al., 2025).

Together, these studies converge on a unifying theme. Transcriptional and epigenetic checkpoints dictate whether microglia intensify or diffuse pyroptotic signalling after SCI. C/EBPβ and HDAC6 act as pro-pyroptotic "accelerators," respectively amplifying Fc-receptor transcription and de-acetylating NF-κB to sustain NLRP3 expression. In contrast, Nrf2 and TRIM32 function as "brakes," suppressing GSDMD translation and NEK7 availability. All four pathways ultimately modulate the same caspase-1/GSDMD pore-forming axis, underscoring its centrality.

Despite clear therapeutic promise, the evidence remains limited to rodent or cellular models, with short observation windows and little insight into temporal cross-talk among these regulators. Whether simultaneous modulation of multiple nodes is synergistic, redundant, or detrimental is unknown, as is the impact on other microglial death modes (e.g., ferroptosis, necroptosis). Future work should integrate time-resolved multi-omics and conditional knockout approaches to map the hierarchy of these checkpoints in vivo and to test combination strategies that fine-tune microglial pyroptosis for optimal neurological recovery.

After SCI, microglia switch on the NLRP3-caspase-1–GSDMD axis, causing pyroptotic rupture and a surge of IL-1β/IL-18 that fuels secondary tissue loss. A growing body of work shows that apparently unrelated enzymatic or chaperone systems converge on this same pathway and can either dampen or amplify it.

The ecto-enzyme CD73 (ecto-5′-nucleotidase) exemplifies a brake that operates at the cell surface. SCI-induced hypoxia first raises HIF-1α, which transcriptionally upregulates CD73 on microglia. Once expressed, CD73 converts extracellular AMP to adenosine; adenosine engages Adenosine A2B receptor (A2B) receptors and activates PI3K/AKT signalling, which inactivates the transcription factor Forkhead box O1 (FoxO1). FoxO1 would otherwise drive GSDMD gene expression, so its inhibition keeps the NLRP3–GSDMD axis in check. In a positive feedback loop, A2B–p38 MAPK signalling feeds back to sustain HIF-1α and therefore CD73 itself, consolidating suppression of pyroptosis and improving histological and functional outcomes after SCI (Xu et al., 2021).

Ceramide synthase-5 (CerS5) has the opposite effect by rewiring lipid metabolism. CerS5 preferentially generates C16-ceramide, which upregulates Phospholipase A2 Group VII (Pla2g7) and primes NLRP3. Down-stream caspase-1 cleavage of GSDMD produces membrane pores, heightening microglial pyroptosis, neuronal loss and behavioral deficits. Genetic or pharmacological blockade of CerS5 lowers C16-ceramide, silences Pla2g7–NLRP3 signalling and markedly improves recovery, positioning CerS5 as a metabolic accelerator of the pyroptotic response (Zhang et al., 2025).

The inducible heat-shock protein HSPA1A (Hsp70-1A) acts as an intracellular chaperone-mediated shield. HSPA1A upregulates the MAPK phosphatase Dual-Specificity Phosphatase 1 (DUSP1), which dephosphorylates p38, c-Jun N-terminal kinase (JNK) and Extracellular Signal-Regulated Kinases 1 and 2 (ERK1/2). Dampening these stress kinases prevents full assembly of the NLRP3 inflammasome, blocks caspase-1/GSDMD activation and curtails IL-1β/IL-18 releases. In rat SCI models, adeno-associated-virus–driven HSPA1A significantly limits microglial pyroptosis and reduces neuroinflammation, translating into better locomotor scores (He et al., 2025).

Glutathione peroxidase-3 (GPx3) brings the antioxidant arm into play. By detoxifying reactive oxygen species, GPx3 interrupts the ROS-driven activation of Interleukin-1 Receptor-Associated Kinase 4 (IRAK4), a kinase that normally propels the IRAK4/ROS/NLRP3 feed-forward loop. Direct binding of GPx3 to IRAK4 destabilizes the kinase, further quenching inflammasome activity and GSDMD cleavage. Viral over-expression of GPx3 in microglia therefore lessens pyroptosis, dampens cytokine outflow and accelerates motor recovery after SCI (Liu et al., 2025).

CD73 and GPx3 represent extracellular and antioxidant brakes, respectively; HSPA1A provides a chaperone-based checkpoint inside the MAPK network; CerS5 functions as a metabolic accelerator. Together they map an interconnected landscape in which purinergic, lipid, oxidative-stress and protein-folding cues all feed into the same inflammatory death pathway.

While each intervention yielded robust structural and behavioral rescue in rodent models, several gaps remain. The studies rely heavily on single-gene manipulations; systems-level analyses are needed to rule out off-target effects and to capture crosstalk between the modules. Timing differs as CD73 is most effective early (hypoxia phase), whereas CerS5 and GPx3 peak later. Except for CD73, translation-relevant delivery routes (systemic, intrathecal) have not been rigorously tested. Addressing these issues will clarify whether a combination strategy—simultaneously boosting brakes and blocking accelerators—can offer synergistic protection in human SCI.

After SCI, damage-associated molecular patterns such as biglycan engage TLR4 on microglia. The receptor signals through Janus Kinase 2 (JAK2)/STAT1, driving transcription of the RNA helicase DEAD-box Helicase 3 X-Linked (DDX3X). DDX3X in turn facilitates assembly of the NLRP3 inflammasome, culminating in caspase-1 activation, GSDMD cleavage and pyroptotic death. Genetic or pharmacological inhibition of TLR4, JAK2/STAT1 or DDX3X lessens pyroptosis and improves locomotor recovery, identifying this linear cascade as a tractable therapeutic axis (Wang et al., 2022).

TLR4/STAT1 signalling also initiates a non-coding-RNA circuit. STAT1 transactivates the LncRNA F630028O10Rik, which sponges miR-1231-5p. Loss of miR-1231-5p derepresses its target Col1a1, thereby engaging the PI3K/AKT pathway and further priming the NLRP3–caspase-1– GSDMD inflammasome. Interrupting any node—TLR4, STAT1, the lncRNA itself, or PI3K/AKT—dampens pyroptosis and yields better functional outcomes after SCI, underscoring the multilayered control that a single PRR can exert through coding and non-coding effectors (Xu et al., 2020).

In sum, while TLR4-centred signalling clearly shapes inflammasome priming after SCI, a more integrated appraisal of the wider Pattern-Recognition Receptor landscape and of chronic-phase pathology will be essential before these insights can be converted into clinically viable interventions.

A growing body of work shows that plant-derived polyphenols can restrain microglial pyroptosis after SCI by converging on a few recurrent targets. Below, the main studies are arranged from upstream regulators to metabolic fine-tuning to provide a clearer mechanistic thread.

Resveratrol increases miR-124-3p, which in turn suppresses DAPK1. Because DAPK1 promotes NLRP3 assembly, its downregulation breaks the NLRP3/caspase-1/GSDMD axis, curbing pyroptosis and improving locomotor recovery in vivo (Li D. et al., 2025).

Taxifolin counters oxidative stress and dampens NLRP3, ASC, caspase-1 and GSDMD by activating the PI3K-AKT survival pathway. The dual antioxidant/anti-pyroptotic action supports axonal regrowth and functional restoration after SCI (Hu et al., 2023).

Kaempferol lowers NADPH oxidase 4-derived ROS, thereby shutting down p38/JNK-MAPK and NF-κB signalling. This cascade prevents NLRP3 activation, reduces IL-1β/IL-18 release and limits microglial pyroptosis, translating into better behavioral scores (Liu Z. et al., 2021).

Cynarin triggers the Nrf2 antioxidant axis, suppressing ROS and secondarily blocking the NLRP3/caspase-1/GSDMD route. Reduced cytokine release and tissue sparing lead to significant gains in locomotor outcomes (Zhang B. et al., 2024).

Celastrol, a triterpenoid often grouped with polyphenols, directly inhibits NF-κB, limiting NLRP3 priming and favoring an M2 (repair-oriented) microglial phenotype. Lower pyroptotic protein levels coincide with smaller lesions and improved motor recovery (Dai et al., 2019).

Urolithin A, a gut-derived ellagitannin metabolite, enhances mitophagy, stabilizes mitochondrial function and cuts mtROS. The intact mitochondria no longer feed NLRP3 activation, keeping caspase-1 and GSDMD in check and facilitating functional repair—an effect abolished by autophagy blockade (Chen et al., 2025).

Lycium barbarum glycopeptide (LbGp) boosts the Fatty Acid Desaturase 1 and 2 in microglia, increasing docosahexaenoic acid (DHA) production. DHA then suppresses MAPK/NF-κB and NLRP3 signalling, limiting pyroptosis and enhancing motor outcomes in SCI models (Jiang et al., 2023).

Collectively, these studies converge on the NLRP3/caspase-1/GSDMD axis as the main regulator of microglial pyroptosis, yet they intervene at distinct checkpoints. miRNA regulation (resveratrol), canonical survival signalling (taxifolin), redox balance (kaempferol, cynarin), transcriptional priming (celastrol), organelle quality control (urolithin A) and lipid-mediated immunomodulation (LbGp). The consistency of behavioral benefits across compounds strengthens the overall case. However, several limitations remain. Most agents ultimately dampen NLRP3; whether combinations yield additive benefits or plateau remains untested. Few studies address pharmacokinetics or delayed treatment windows, critical for clinical translation. Microglia are the stated target, yet off-target effects on neurons, astrocytes and peripheral immune cells are seldom quantified. Urolithin A and LbGp underscore microbiota or dietary lipid involvement, but causal links in human SCI are unknown.

Microglial pyroptosis becomes a decisive secondary insult after SCI, and several stem-cell–derived secretomes now converge on this target. Collectively, these approaches work by ferrying regulatory non-coding RNAs or other paracrine factors that quench inflammasome activity, blunt cytokine release and create a permissive milieu for axonal regrowth.

The simplest intervention is the whole conditioned medium. Human dental-pulp stem-cell secretome dampens the canonical NLRP3-caspase-1-IL-1β axis in LPS-primed BV2 microglia and in contused rats, translating to reduced neuroinflammation, enhanced axonal sprouting and better locomotor scores (Liu T. et al., 2024).

Going a step further, exosomes from bone-marrow MSCs were loaded with miR-22. Because miR-22 directly targets GSDMD, the executioner of pyroptosis, the engineered vesicles curtailed pore formation and cytokine release, markedly improving behavioral recovery in rats (Sheng et al., 2021).

Olfactory mucosa MSC exosomes traffic the lncRNA RMRP into microglia. RMRP complexes with Eukaryotic Translation Initiation Factor 4A3 (EIF4A3) to stabilize Sirtuin 1 mRNA; elevated Sirtuin 1 then represses NLRP3, caspase-1 and GSDMD, easing motor deficits in mice (Wang C. et al., 2025).

Induced-pluripotent NSCs and their vesicles shuttle let-7b-5p, which silences the adaptor LRIG3, again throttling NLRP3/caspase-1/GSDMD signalling. The result is preserved myelin, denser axonal tracts and superior motor coordination (Liu J. et al., 2024).

Pyroptosis peaks around day 7 post-injury, but endogenous Treg cells temper this surge. Their exosomes deliver miR-709, which knocks down NKAP and thereby the NF-κB feed-forward loop; genetic ablation of Tregs or miR-709 exacerbates, and supplementation ameliorates, sensorimotor decline (Xiong et al., 2022).

Finally, anti-inflammatory M2 microglia lend support to neighboring neurons: their vesicles are enriched in miR-672-5p, an inhibitor of the AIM2/ASC/caspase-1 pathway. Loss of miR-672-5p nullifies the vesicles' protection, confirming a neuron-centered pyroptosis checkpoint distinct from NLRP3 (Zhou et al., 2022).

Across diverse cell sources (dental pulp, mesenchyme, olfactory mucosa, induced pluripotent cells, Tregs and even homeostatic M2 microglia) the common therapeutic denominator is delivery of RNA cargo that throttles one of three pressure points in the pyroptotic cascade: (i) inflammasome activation (NLRP3 or AIM2), (ii) execution (GSDMD) or (iii) transcriptional priming (NF-κB/NKAP). These studies consistently report reduced cytokine storms and improved locomotion, reinforcing pyroptosis as a tractable target.

Yet several gaps remain. All work is confined to acute or subacute rodent models; chronic lesions, large mammals and female subjects are largely absent. Cargo heterogeneity, isolation protocols and dosing schedules are rarely standardized, hampering cross-study comparisons. Long-term biodistribution, off-target immunomodulation and potential crosstalk with other death axes (ferroptosis, necroptosis) are also untested. Finally, while most strategies harness non-coding RNAs, only one exploits direct GSDMD inhibition—suggesting unexplored space for combinatorial or small-molecule co-packaging. Overall, stem-cell–and immune-cell–derived secretomes offer a multi-modal, low-immunogenic platform to subdue microglial pyroptosis, but rigorous optimization and head-to-head validation are essential before clinical translation.

Segmental (Jiaji) EA delivered soon after SCI reverses microglia-driven pyroptosis and accelerates locomotor recovery. In rat models, EA suppressed the entire NLRP3/caspase-1/GSDMD axis, lowering IL-1β and IL-18 release and shrinking the population of GSDMD-positive microglia in the injured cord. Administration of an NLRP3 agonist partially restored pyroptotic signalling and motor deficits, confirming that EA's benefit is tightly coupled to inflammasome inhibition rather than to general neurostimulation (Mei et al., 2024).

Extending those findings to chronic neuropathic pain, a second study showed that EA applied at hind-limb acupoints likewise downregulated NLRP3, cleaved caspase-1, N-GSDMD and downstream cytokines—but only in microglia, not neurons or astrocytes. Pharmacological blockade of NLRP3 with MCC950 reproduced EA's antiallodynic effect and provided no additional benefit when combined, underscoring the specificity of the pathway (Kui et al., 2025).

Intermittent metabolic stress also curbs microglial pyroptosis. Every-other-day Fasting elevated circulating bile acids that engage the G-protein-coupled bile acid receptor 1 (TGR5), a known negative regulator of NLRP3. Activation of TGR5 dampened microglial activation, reduced caspase-1 cleavage and GSDMD pore formation, and, in turn, blunted IL-1β/IL-18 release, limiting secondary tissue loss and improving behavioral outcomes after SCI (Song et al., 2024).

Conversely, exposure to advanced oxidation protein products—biomarkers of systemic oxidative stress—worsened SCI pathology by driving NADPH oxidase-derived ROS production. ROS activated the MAPK/NF-κB signalling cascade, primed the NLRP3 inflammasome and precipitated full-blown microglial pyroptosis. The NADPH oxidase inhibitor apocynin broke this feed-forward loop, attenuating ROS accumulation, suppressing inflammasome formation and improving histological as well as functional outcomes (Liu et al., 2020).

Across diverse pre-clinical settings, both physical (like EA) and metabolic (intermittent fasting, redox manipulation) interventions converge on a single molecular hub—the NLRP3/caspase-1/GSDMD inflammasome—highlighting its centrality in microglial pyroptosis after SCI. EA provides tunable, region-specific suppression of this axis without systemic metabolic perturbation, whereas Every-other-day Fasting offers a non-invasive systemic strategy that leverages endogenous bile-acid signalling. The advanced oxidation protein products (AOPPs) study usefully reminds us that background oxidative stress can push the same pathway in the opposite direction, and that pharmacological antioxidants such as apocynin may be necessary adjuncts.

Strengths of the section include mechanistic depth (each study maps a clear causal chain) and functional read-outs that tie molecular changes to motor or pain phenotypes. Limitations are equally clear: all data are rodent-based, head-to-head comparisons of intervention timing and dosing are missing, and none of the studies examine long-term safety or combinatorial effects with standard neuro-rehabilitation. Future work should (i) validate these approaches in large-animal or humanized models, (ii) dissect potential cross-talk with Ferroptotic and necroptotic pathways, and (iii) explore whether physical and metabolic strategies can be synergistically layered to achieve durable neuroprotection without off-target immune suppression.

Lupenone is a triterpenoid with anti-inflammatory properties and significantly improves SCI outcomes by targeting microglial pyroptosis. In SCI, microglia become overactivated (M1 phenotype), triggering inflammation through the NF-κB/NLRP3/caspase-1/GSDMD axis, leading to pyroptosis. Lupenone inhibits NF-κB activation (via p65/IκBα modulation), reducing NLRP3 inflammasome activation, IL-1β/IL-18 releases, and promoting M2 microglial polarization. This dual modulation of pyroptosis and phenotype shift leads to neuroprotection and improved motor recovery (Li F. et al., 2024).

Kanglexin is a novel synthetic anthraquinone, ameliorates SCI by modulating neuroinflammation in microglial cells. Kanglexin inhibits pyroptosis by blocking the NLRP3 inflammasome pathway. It enhances Protein Kinase A (PKA) phosphorylation, which suppresses NF-κB activation and reduces transcription of inflammasome components. Kanglexin also promotes microglial polarization from proinflammatory M1 to anti-inflammatory M2 phenotypes, aiding neurorepair. Overall, Kanglexin protects neural tissue post-SCI by suppressing pyroptosis and shifting microglial phenotype via the PKA/NF-κB signaling axis (Yan et al., 2025).

In spinal cord ischemia/reperfusion (SCIR) injury, pyroptosis alongside apoptosis and necroptosis, contributes to neural death. Melatonin, a hormone synthesized in the pineal gland with antioxidant and anti-inflammatory properties, reduces microglial pyroptosis via melatonin receptor (MT1/MT2)-mediated signaling. Specifically, melatonin inhibits NLRP3, caspase-1, and GSDMD activation, reducing IL-1β release in microglia. Luzindole, a melatonin receptor antagonist, negates this effect, confirming receptor involvement. Melatonin thereby improves neurological outcomes by suppressing PANoptosis—including microglial pyroptosis—in SCIR rats (Xie et al., 2024a).

After SCI, activated microglia accumulate iron and release NO that destabilizes neuronal iron homeostasis by up-regulating IRP1, Divalent Metal Transporter 1 (DMT1) and Transferrin Receptor 1 while down-regulating ferritin. The ensuing neuronal iron overload—together with mitochondrial shrinkage and altered Prostaglandin-Endoperoxide Synthase 2 (PTGS2)/Gpx4 expression—triggers neuronal ferroptosis and limits motor recovery. Iron chelators (like deferoxamine), ROS scavengers (like N-acetylcysteine) and the canonical ferroptosis inhibitor ferrostatin-1 all attenuate this cascade, underscoring the central role of microglia-driven iron dysregulation in SCI pathology (Feng et al., 2021).

Ferroptotic microglia themselves amplify inflammation. Their lipid-peroxidation-induced mtROS activate PKR-like endoplasmic reticulum kinase (PERK) signalling, whereas NF-κB drives IL-1β and IL-6 transcription. Together, these events markedly upregulate IL-23, fostering differentiation of pathogenic T helper 17 (Th17) cells that exacerbate tissue damage. Dual inhibition of ferroptosis and IL-23 therefore improves functional recovery in experimental models (Zhao et al., 2024).

TNF-α sits at the core of an iron–TNF–ferroptosis–inflammation axis. Iron overload raises microglial TNF-α, which in turn potentiates lipid peroxidation (4-HNE), depletes glutathione and pushes cells toward an iNOS⁺/CD86⁺ M1 phenotype. Genetic deletion of TNF-α—or pharmacological zinc supplementation—breaks this feed-forward loop, diminishes ferroptosis and shifts microglia toward an Arg1⁺/CD206⁺ reparative state, improving neuronal survival and locomotion (Zeng et al., 2024).

Glycosylated lysosomal membrane protein (GLMP) is an intrinsic brake on ferroptosis. SCI lowers GLMP expression; loss-of-function increases microglial iron deposition, raises COX-2 and acyl-CoA synthetase long-chain family member 4 (ACSL4), damages mitochondria and worsens behavioral outcomes. Conversely, preserving GLMP limits ferroptosis and supports tissue repair, highlighting lysosomal iron handling as a therapeutic entry point (Ouyang et al., 2025).

Epigenetic control further modulates the Ferroptotic response. Protein arginine methyl-transferase-8 (PRMT8) catalyzes Histone 3 Lysine 4 (H3K4) trimethylation to boost GDNF expression. PRMT8 over-expression reduces iron, 4-HNE and Heme Oxygenase 1 (HO-1), curtails ferroptosis and promotes Arg1-dominated M2 microglial polarization, collectively protecting the injured cord (Zou et al., 2023).

Systemic metabolic cues intersect with local Ferroptotic signalling. SCI-induced oxidative stress increases ApoB-100/SORT1 expression, elevating ox-LDL, driving M1 activation and aggravating neuronal ferroptosis. Silencing ApoB-100/SORT1 restores GPX4 and Transferrin Receptor 1, lowers ROS and skews microglia away from an inflammatory phenotype. Intriguingly, M1 macrophages can transiently suppress ferroptosis in vitro yet promote it in vivo, illustrating the context-dependence of immune–metabolic cross-talk (Wu et al., 2024b).

Collectively, these studies converge on a unifying theme. Iron-laden, lipid-peroxidized microglia are both victims and amplifiers of ferroptosis after SCI. Up-stream regulators range from classical cytokines (TNF-α) and stress pathways (PERK, NF-κB) to lysosomal (GLMP), epigenetic (PRMT8) and metabolic (ApoB-100/SORT1) checkpoints, all feeding into a self-propagating loop of iron accumulation, oxidative damage and inflammatory polarization.

Two conceptual tensions remain. First, microglia can simultaneously incite neuronal ferroptosis and undergo it themselves, but whether they die or survive long enough to orchestrate chronic inflammation is still unclear. Second, the in-vitro versus in-vivo discrepancy in M1-mediated ferroptosis points to stage-specific or micro-environmental modifiers that are largely unmapped.

Therapeutically, single-target interventions (iron chelators, ferrostatin-1) show partial benefit, whereas combined strategies—modulating iron handling (GLMP), lipid metabolism (ApoB-100/SORT1) and inflammatory signalling (TNF-α, IL-23)—offer a more compelling rationale. Future work should therefore focus on the following factors: (i) temporal mapping of ferroptosis relative to microglial phenotypic shifts; (ii) multi-omic integration to resolve micro-environmental heterogeneity; and (iii) combinatorial or stage-tailored treatments that can interrupt the ferroptosis–inflammation feedback with greater precision.

After SCI, microglia engulf erythrocytic debris and, through HO-1 induction and NCOA4-dependent ferritinophagy, accumulate catalytically active iron that precipitates glutathione (GSH) depletion, lipid peroxidation, and Ferroptotic death. The lipophilic antioxidant UAMC-3203 interrupts this cascade: by restoring GSH and GPX4 activity it limits lipid peroxidation in microglia/macrophage populations, curtails secondary tissue loss, and accelerates locomotor recovery in mice (Ryan F. et al., 2024).

Ferrostatin-1, the prototypical ferroptosis blocker, likewise reduces ROS and iron overload after SCI. Consequent downregulation of the ferroptosis markers iron-responsive element-binding protein 2 (IREB2) and PTGS2 in oligodendrocytes and microglia preserves myelin integrity, dampens gliosis, and improves hind-limb function in rat models (Ge et al., 2022).

A second study with the more water-soluble analog UAMC-3203 confirms and extends these findings: activation of the Nrf2/HO-1 antioxidant axis normalizes GPX4, and GSH levels, while simultaneously re-balancing the inflammatory milieu—suppressing TNF-α and IL-1β and boosting IL-4 and IL-10. This dual ferroptosis-modulating and immunoregulatory action protects neurons and further enhances behavioral recovery (Kan et al., 2024).

Punicalagin, a pomegranate-derived polyphenol, targets the same Nrf2/solute carrier family 7 member 11 (SLC7A11)/GPX4 axis but from a natural-compound vantage. By preventing iron-dependent lipid peroxidation in neurons and driving microglia toward the reparative M2 phenotype, it synchronously attenuates oxidative stress and fosters axonal regrowth, resulting in superior functional outcomes (Wang W. et al., 2025).

Albiflorin, isolated from Paeonia lactiflora, inhibits lysine-specific demethylase-1 (LSD1). Because LSD1 upregulation promotes ferroptosis, its suppression by Albiflorin lowers lipid peroxidation and iron accumulation in microglia, dampens proinflammatory cytokine release, and supports motor recovery (Zhang L. et al., 2024).

Gallic acid, another plant phenolic, mitigates chronic post-SCI pain and depression by restraining P2X purinoceptor 7 (P2X7)-driven TNF-α/TNF-α–converting enzyme (TACE; ADAM17)/NF-κB/STAT3 signalling. The downstream result is reduced ROS, malondialdehyde, and iron load, restored GSH/GPX4, preserved mitochondrial morphology, and ultimately less microglial ferroptosis and neuroinflammation (Yang R. et al., 2023).

Across small synthetic inhibitors (Ferrostatin-1, UAMC-3203) and phytochemicals (punicalagin, Albiflorin, gallic acid) a convergent therapeutic principle emerges: preventing iron-catalyzed lipid peroxidation while re-establishing the antioxidant (GSH–GPX4) axis halts microglial ferroptosis and its inflammatory sequelae, translating into robust neuro-protection and functional improvement. Notably, UAMC-3203's superior solubility and its additional immunomodulatory effects position it as a leading candidate for clinical translation. Natural compounds broaden the armamentarium and—by engaging epigenetic (like LSD1) or purinergic (like P2X7) nodes—highlight novel ancillary targets. Yet most studies rely on acute rodent models; long-term efficacy, optimum dosing, and combinatorial potential with existing neuro-rehabilitative strategies remain open questions demanding systematic investigation.

A growing body of work shows that well-designed nanomaterials can temper microglia-driven ferroptosis and thereby limit secondary damage after SCI. Although each platform exploits a different carrier, all converge on three anti-Ferroptotic principles: (i) lower the free Fe²⁺ pool, (ii) quench iron-driven lipid peroxidation, and (iii) re-program microglia toward a reparative (M2-like) phenotype. Below, the studies are arranged from the most targeted genetic approach to multi-functional drug-delivery systems, highlighting the incremental design logic that runs through the field.

Macrophage-targeted phosphatidylserine-lipid nanoparticles (PS/LNPs) carrying Mms6 mRNA give injured spinal cords an "endogenous magnetite factory." Translated Mms6 protein nucleates magnetite cores that trap free iron, lowering the Fe²⁺ pool and lipid peroxidation. Selective PS-mediated uptake restricts expression to microglia/macrophages; when these cells are experimentally depleted, the motor and histological gains disappear—proving that ferroptosis control in this compartment is central to recovery (Fu et al., 2025).

An injectable carboxymethyl-cellulose/quaternised-chitosan hydrogel embedded with polydopamine nanoparticles (CQP hydrogel) chelates Fe²⁺ and scavenges ROS directly inside microglia. By re-activating the GPX4 lipid-repair axis, the gel attenuates ferroptosis and simultaneously tilts microglia from an M1 to M2 state, culminating in reduced lesion volume and improved locomotion (Shi et al., 2024).

Mesoporous selenium nanoparticles pre-loaded with resveratrol and cloaked in borneol-modified macrophage membranes cross the BSCB and dock onto resident microglia. Selenium-triggered GPX-like activity, together with resveratrol, raises GPX4/SOD, increases GSH, depresses malondialdehyde (MDA), and curbs ROS. The dual redox hit blunts ferroptosis, dampens inflammation and apoptosis, and translates into robust motor recovery (Liu X. et al., 2024).

Ferrostatin-1/selenium/ZIF-8 nanoparticles wrap selenium cores in zeolitic imidazolate framework-8 (ZIF-8) and ferrostatin-1. The hybrid particle scavenges ROS, safeguards mitochondria, and delivers a canonical ferroptosis inhibitor in one shot. As ferroptosis markers fall, microglia/macrophages polarize toward an anti-inflammatory M2 phenotype, inflammatory cytokines wane, and spared neurons support behavioral gains (Zhou et al., 2024).

A paclitaxel-idebenone prodrug self-assembles into nanomicelles that home to chondroitin-sulphate-rich lesion matrices. Idebenone restores the ferroptosis suppressor protein 1 (FSP1) ferroptosis-suppression pathway and curbs 4-HNE lipid adducts, while paclitaxel stabilizes microtubules to guide axon regrowth. The combined anti-Ferroptotic and pro-regenerative actions reduce microglial activation, preserve neurons, and enhance functional recovery (Xu Z. et al., 2024).

Collectively, these platforms prove that nanomedicine can hit multiple ferroptotic nodes simultaneously while steering microglial phenotype—something single-target small molecules rarely achieve.

Yet several limitations persist. Heterogeneous read-outs. Each study tracks a different slice of the ferroptosis pathway (GPX4, FSP1, Fe²⁺, 4-HNE, etc.), making head-to-head efficacy comparisons impossible. A consensus biomarker panel is urgently needed.

Cell-type specificity. Most reports infer microglial targeting from bulk tissue analyses or macrophage depletion, but single-cell RNA/proteomics would verify true selectivity and off-target risks. Selenium, magnetite, and ZIF-8 scaffolds can accumulate; only short-term biosafety is shown. Chronic retention and systemic redistribution remain open questions. Clinical translatability. Injectable hydrogels and PS/LNPs are already close to regulatory paradigms, whereas membrane-coated or metal-organic cages face sterility, scale-up, and reproducibility hurdles. The field is converging on multifunctional "all-in-one" particles, yet additive complexity may outpace manufacturing and regulatory capacity. Rational minimalism—identifying which ferroptotic lever truly dictates outcome—could streamline the path to the clinic.

In summary, nanoplatforms that couple iron detoxification with lipid-repair and immune re-balancing offer a compelling anti-ferroptotic strategy for SCI, but rigorous comparative studies and long-term safety profiling are essential before clinical translation.