Constitutive DAMPs in CNS injury: From preclinical insights to clinical perspectives

Damage-associated molecular patterns (DAMPs) are endogenous molecules released in tissues upon cellular damage and necrosis, acting to initiate sterile inflammation. Constitutive DAMPs (cDAMPs) have the particularity to be present within the intracellular compartments of healthy cells, where they exert diverse functions such as regulation of gene expression and cellular homeostasis. However, after injury to the central nervous system (CNS), cDAMPs are rapidly released by stressed, damaged or dying neuronal, glial and endothelial cells, and can trigger inflammation without undergoing structural modifications. Several cDAMPs have been described in the injured CNS, such as interleukin (IL)-1 α , IL-33, nucleotides (e.g. ATP), and high-mobility group box protein 1. Once in the extracellular milieu, these molecules are recognized by the remaining surviving cells through specific DAMP-sensing receptors, thereby inducing a cascade of molecular events leading to the production and release of proinflammatory cytokines and chemokines, as well as cell adhesion molecules. The ensuing immune response is necessary to eliminate cellular debris caused by the injury, allowing for damage containment. However, seeing as some molecules associated with the inflammatory response are toxic to surviving resident CNS cells, secondary damage occurs, aggravating injury and exacerbating neurological and behavioral deficits. Thus, a better understanding of these cDAMPs, as well as their receptors and downstream signaling pathways, could lead to identification of novel therapeutic targets for treating CNS injuries such as SCI, TBI, and stroke. In this review, we summarize the recent literature on cDAMPs, their specific functions, and the therapeutic potential of interfering with cDAMPs or their signaling pathways.


Early neuroinflammatory responses to CNS injury
The central nervous system (CNS) is separated from the periphery by at least four major blood-CNS barriers, which include the blood-brain barrier (BBB), the blood-spinal cord barrier (BSCB), the bloodmeningeal barrier (BMB), and the blood-cerebrospinal fluid barrier (BCSFB).These barriers are key sites of neuro-immune interactions, primarily responsible for immune surveillance of the CNS.They regulate the entry of peripheral immune cells and facilitate lymphatic drainage of cerebrospinal fluid (CSF) from the brain and spine to lymph nodes (for review, see (Rustenhoven and Kipnis, 2022)).Their integrity therefore plays a critical role in injury pathology as they limit inflammation and collateral damage in tissues whose regenerative capacity is hampered, thus minimizing homeostatic disturbances affecting vital functions.However, despite the existence of these borders, it is now acknowledged that various events may trigger disastrous sterile inflammatory responses in the injured CNS (Ransohoff, 2016;Salvador and Kipnis, 2022).Sterile inflammation is particularly relevant in cases of traumatic injuries affecting the CNS, such as spinal cord injury (SCI), traumatic brain injury (TBI), and stroke.In these instances, the primary lesion induces local inflammation, which may subsequently perpetuate secondary damage.SCI, TBI, and stroke are the most common injuries affecting the CNS.When damage occurs, a necrotic area rapidly forms at the center (lesion core) of the affected region (Fig. 1).Within minutes, the injury triggers an innate immune inflammatory response that rapidly leads to the recruitment of neutrophils and monocytes, whose primary goals are to sterilize and clear cell debris, two crucial steps towards tissue repair and healing.However, the considerable size of these CNS injuries is often associated with poor clearance of cell debris and a compromised resolution of inflammation at the site of injury.This may lead to exaggerated and prolonged inflammation, which is believed to promote further damage to the surrounding healthy tissue and the penumbra, a region where cells remain vulnerable but still salvageable (Salvador and Kipnis, 2022;Shi et al., 2019).From a molecular point of view, the initial phase of inflammation is triggered by necrotic cells releasing their cytoplasmic and nuclear contents into the surrounding parenchyma.Some of these released endogenous (self) molecules act as 'alarm' signals and, accordingly, are known as alarmins or damage-associated molecular patterns (DAMPs) (Oppenheim and Yang, 2005;Seong et al., 2021).While we will focus on the role of DAMPs in instances of inflammation, their expression in the healthy CNS is crucial for maintaining homeostatic functions.They stabilize gene expression, regulate cellular metabolism, and even act as transcription factors (Murao et al., 2021).However, when DAMPs are liberated from cells and flood the extracellular space, their primary function is to alert the immune system of the presence of tissue damage.
According to the new nomenclature recently proposed by Albert and colleagues (Yatim et al., 2017), DAMPs can be subclassified into at least two distinct categories: constitutive DAMPs (cDAMPs) and inducible DAMPS (iDAMPS).cDAMPS are constitutively expressed endogenous molecules released after cell damage, and do not require modification to induce their effects.iDAMPs on the other hand, are endogenous molecules that are actively produced, either by neo-transcription, neotranslation, or post-translational modification following tissue damage.The rapid release of cDAMPs at the lesion core by necrotic cells triggers the production of proinflammatory cytokines such as interleukin (IL)-1β and TNF by the surviving cells (Gadani et al., 2015a).Notably, the CNS endothelium expresses type 1 IL-1 and TNF receptors, IL-1R1 and TNFR1 respectively.The endothelium is thus strongly activated by this inflammatory environment, resulting in chemoattraction of innate immune cells to the site of injury and increased BBB/BSCB permeability (Bebo and Linthicum, 1995;Bretheau et al., 2022;Molino et al., 2016;Wong et al., 2019).IL-1β and TNF also induce non-apoptotic programmed cell death in parenchymal cells (necroptosis, pyroptosis) and leukocytes (pyroptosis, NETosis).This leads to the release of newly synthesized or modified DAMPs through active post-mortem transcriptional or post-translational programs, termed iDAMPs, thereby contributing to the overall immune response.The trauma also produces hemorrhages with the entry of undesirable blood-derived molecules and extravasation of erythrocytes and platelets, while also increasing water infiltration across the BBB/BSCB, resulting in edema (Miyanji et al., 2007;Shlosberg et al., 2010;Yang and Rosenberg, 2011).
This damaged environment will further promote the recruitment of inflammatory cells to the site of the injury.However, for unknown reasons, immune cells are unable to completely resolve inflammation Fig. 1.Necrotic CNS-resident cells release constitutive DAMPs that trigger inflammation and secondary damage after CNS injury.(A) Schematic of the acute phase of CNS injury, depicting blood vessels and circulating innate immune cells (top) and CNS parenchyma (bottom), separated by the blood-CNS barrier.The major CNSresident cells are shown: microglia (light green), astrocytes (dark green), oligodendrocytes (blue), and neurons (yellow).After CNS injury, ruptured blood vessels produce hemorrhages.In the lesion core (pink circle), necrotic cells release various molecules into the extracellular space, causing sterile inflammation, including several constitutive DAMPs (cDAMPs).(B) Representation of the consequences of the activation of the immune system and neuroinflammation in the subacute and chronic phases of CNS injury.The activated immune responses triggered by the cDAMPs promote activation of surrounding glial cells and disruption of the blood-CNS barriers, which is accompanied by the infiltration of innate immune cells such as monocyte-derived macrophages at the site of CNS injury.This proinflammatory environment causes an increase in glial and neuronal cell death and therefore an increase in lesion size, along with the creation of a glial scar mainly composed of microglia, pericytes, fibroblasts, and astrocytes (concentric circles surrounding the secondary lesion).
after clearance of debris, and in cases of SCI, they remain in the injured spinal cord for weeks to months, releasing additional waves of proinflammatory mediators (Schwab et al., 2014).In the case of TBI, this proinflammatory milieu can persist for months to years, thereby promoting the development of other neurodegenerative pathologies (Simon et al., 2017).Following primary brain damage induced by stroke, cytotoxic blood components and the production of reactive oxygen species further contribute inflammation and secondary neuronal damage in the penumbra region (Iadecola et al., 2020).In this review article, we explore the role of cDAMPs in SCI, TBI, and stroke.We examine the effects of specific cDAMPs in these CNS injury conditions (see Table 1), discuss future perspectives, and delve into potential therapeutic strategies targeting these molecules.

Constitutive DAMPs act as triggers of neuroinflammation after CNS injury
When trauma or damage occurs in the CNS, cDAMPs such as IL-1α, IL-33, nucleotides, high mobility group box-1 (HMGB1), heat shock proteins, and S100 proteins are likely released as 'S.O.S. messages' by damaged or dying cells (Gadani et al., 2015a).These cDAMPs bind to different pattern recognition receptors (PRRs), such as Toll-like receptors (TLRs) and nucleotide oligomerization domain (NOD)-like receptors (NLRs), along with non-PRR DAMP-sensing receptors, such as IL-1R1, P2 purinergic receptors as well as receptor for advanced glycation end products (RAGE), expressed by tissue-resident non-immune and immune cells (for review, see (Gong et al., 2020;Zindel and Kubes, 2020)).This in turn leads to the expression of multiple inflammatory genes that orchestrate the recruitment and sustained infiltration of innate immune cells at the site of injury, notably neutrophils and monocytes.The phagocytosis of cellular debris by myeloid cells is the first step towards tissue repair, and is typically followed by the resolution of inflammation.However, in the injured CNS, phagocytosis is slow, taking weeks to months in rodents, and sometimes years in humans (Griffin et al., 1992;Miklossy and Van der Loos, 1991), resulting in a chronic presence of cellular and myelin debris, thus preventing proper healing.In these contexts, axonal regeneration and tissue repair are compromised, and resolution of inflammation is unsuccessful, perpetuating inflammation and tissue destruction.The formation of a glial scar, which is concomitant with the release of proinflammatory mediators, also hampers axonal regeneration and neuroplasticity and can impose permanent disability on those suffering from CNS injury.In this review, we focus on cDAMPs, endogenous molecules that once released in the injured CNS quickly and effectively trigger inflammation without requiring structural modification to do so.We also discuss the receptors and signaling pathways by which these 'alarm' signals stimulate further production of chemokines and cytokines.
In the CNS, necrotic microglia and infiltrating platelets serve as potential sources of IL-1α in the context of SCI or stroke (Bastien et al., 2015;Luheshi et al., 2011;Thornton et al., 2010).Furthermore, neurons, glial cells, and CNS endothelial cells have been identified as expressors of IL-1R1 in different regions of the nervous system, including the brain (Bruttger et al., 2015;Cunningham et al., 1992;Ericsson et al., 1995), spinal cord (Holló et al., 2017;Lévesque et al., 2016;Wang et al., 2006), and dorsal root ganglia (Mailhot et al., 2020).This widespread presence of IL-1R1 suggests that IL-1α can elicit responses in multiple cell types, contributing to the complexity of neuroinflammatory signaling in the CNS (see Fig. 3).Together, these data make IL-1α a crucial cDAMP in the CNS response to damage.

IL-1α promotes neuroinflammation leading to secondary damage after CNS injury
The presence of IL-1α has been extensively demonstrated in SCI, TBI and stroke (Bastien et al., 2015;Boutin et al., 2001;Bretheau et al., 2022;Luheshi et al., 2011;Newell et al., 2018).In fact, its expression follows a similar pattern in all three pathological conditions in murine models.In the case of SCI, IL-1α can be found within the first 24 h after injury, with a peak expression at 4 h, which is directly associated with necrotic microglia at the injury site (Bastien et al., 2015;Bretheau et al., 2022).Similarly, 4 and 24 h after inducing stroke in mice, IL-1α expression colocalizes with Iba1, a marker of microglia, in the injury site (Luheshi et al., 2011).In TBI, IL-1α is also detected early on, between 6 and 24 h after injury, with expression localized in the parietal cortex, hippocampus, brainstem, and cerebellum (Newell et al., 2018).These findings substantiate the concept that IL-1α functions as a cDAMP in these pathological contexts, as its presence coincides with the initial phases of injury.
In order to study the role of this cDAMP in these pathologies, various authors have utilized a mouse model deficient in IL-1α (IL-1α-KO).In Bastien et al. 2015(Bastien et al., 2015), IL-1α-KO mice showed a significant decrease in neuroinflammation compared to WT mice, with a reduction in neutrophil and macrophage infiltration at the lesion epicenter and surrounding areas during the acute phase of SCI.Additionally, IL-1α-KO mice exhibited increased mobility during the first two weeks post-injury, which correlated with an increase in the total oligodendrocyte count at day 1 and a decrease in lesion volume at day 35 after SCI.Furthermore, we recently demonstrated that administration of IL-1α into the CSF induces an inflammatory response similar to that observed in SCI, promoting the release of proinflammatory cytokines and chemokines into the bloodstream (Bretheau et al., 2022).Blocking IL-1R1 led to an improvement in locomotion after injury, whether using mice lacking this receptor, or pharmacological inhibition via IL-1RA (also known as anakinra) (Bastien et al., 2015;Bretheau et al., 2022).Furthermore, we recently reported that the intra-cisterna magna injection of IL-1α alone in mice leads to neutrophil recruitment and a Fig. 2. Constitutive DAMPs activate pathways involved in inflammation.(A) IL-1α released after CNS damage will bind to IL-1R1, present on astrocytes and endothelial cells, causing the recruitment of the accessory protein IL-1RAcP which triggers a pathway that activates MAPK, JNK, and NF-κB, promoting the transcription of proinflammatory cytokines, chemokines, and cell adhesion molecules.(B) The receptor of IL-33, ST2, is present in microglia, and its activation triggers the recruitment of the accessory protein IL-1RAcP, promoting the activation of NF-κB.In CNS pathologies, this activation produces anti-inflammatory cytokines.In addition, the soluble form of ST2 (sST2) is produced, acting as negative feedback.GATA3 has been proposed as a key molecule in regulating the anti-inflammatory phenotype of macrophages; however, its effect in the context of CNS injury remains unknown.(C) HMGB1 activates RAGE and TLR receptors (only the RAGE receptor is depicted) providing the activation of MAPK and NF-κB, promoting the transcription of proinflammatory proteins after CNS injury.(D) ATP released during CNS damage will act as a cDAMP in the activation of different types of purinergic receptors, promoting the activation and proliferation of microglia, which favor the initiation of inflammatory responses.
decrease in mature oligodendrocytes in the spinal cord as a direct consequence of IL-1R1 activation in endothelial cells and astrocytes, respectively, suggesting their role in secondary damage in CNS injury.However, conditional knockout of this receptor in astrocytes or endothelial cells in the context of SCI did not produce improvement in mouse locomotion (Bretheau et al., 2022).This may suggest that the blockade of IL-1R1 in a single cell type is not sufficient in preventing the detrimental effects of IL-1R1 in SCI, as other cells expressing IL-1R1 may compensate by producing factors that can contribute to secondary damage.
IL-1-KO mice have also been utilized in models of ischemic brain injury to investigate the role of IL-1 cytokines.In Boutin et al. 2001(Boutin et al., 2001), IL-α/β double KO mice, but not mice lacking either IL-1α or IL-1β alone, exhibited a substantial (70 %) decrease in total infarct volume after transient middle cerebral artery occlusion (MCAO), a stroke model.As IL-1β may compensate for the effects of IL-1α and vice versa in the single cytokine KO models, intracerebroventricular administration of the IL-1 receptor antagonist, IL-1RA, was performed in IL-1α-KO and IL-1β-KO mice.IL-1α-KO mice treated with IL-1RA showed a reduction in lesion volume compared to saline-treated IL-1α-KO mice, suggesting that IL-1β may compensate for the lack of IL-1α in this model, binding to IL-1R1 to aggravate inflammation.Surprisingly, IL-1β-KO mice treated with IL-1RA showed no reduction in lesion volume compared to saline-treated IL-1β-KO mice, suggesting compensation might be independent of IL-1α in this mouse line or that IL-1α may exert proinflammatory functions independent of the traditional IL-1R1.Interestingly, in 2012, Ning Quan's laboratory discovered the presence of a truncated form of IL-1R1, hypothesizing based on its amino acid sequence and predicted structure, that this receptor lacks the extracellular portion prone to binding IL-1RA (Qian et al., 2012).Furthermore, we have determined that astrocytes express a smaller molecular weight IL-1R1 (Bretheau et al., 2022), which could correspond to the truncated version of the receptor described by Qian et al.Future studies should focus on investigating this truncated form of IL-1R1 to optimize the most effective treatment for inhibiting the IL-1/IL-1R1 axis in the context of CNS injury.
The use of IL-1RA in WT mice to block IL-1R1 has been reported as beneficial in SCI and TBI, with an improvement in locomotion and cognitive function, respectively (Bastien et al., 2015;Newell et al., 2018).Additionally, post-ischemic administration of an anti-IL-1α antibody or IL-1RA led to a decrease in the infarct size and reduction of neurological deficits after transient MCAO in mice (Bretheau et al., 2022;Liberale et al., 2020).These positive results in the inhibition of IL-1α or its receptor in different CNS injury contexts suggests that targeting the IL-1 pathway in CNS injuries has significant potential for shortwindow therapies.
Different studies have been conducted in humans for the blockade of IL-1R1 in CNS injury.Phase II studies in patients with stroke or TBI Fig. 3. CNS-resident cells serve as both the source and responders to cDAMPs.After CNS injury, necrotic CNS-resident cells release various cDAMPs that bind to DAMP-sensing receptors expressed by the surviving CNS-resident cells, triggering a sterile form of inflammation.Straight arrows represent the cells that release cDAMPs into the extracellular environment, while curved arrows represent those that respond to these molecules, following the color code: IL-1α (red), IL-33 (blue), HMGB1 (green), and ATP (black).IL-1α produced by necrotic microglia following injury promotes neuroinflammation and disruption of the blood-CNS barriers through activation of endothelial IL-1R1, while activation of astrocytic IL-1R1 induces oligodendrocyte death and secondary tissue damage.IL-33 produced by necrotic astrocytes and oligodendrocytes promotes an anti-inflammatory environment.Microglia and astrocytes both express the ST2 receptor; however, the implications of activating this receptor in these cell types in the context of CNS injury warrant further investigation.ATP is released by necrotic cells of all types and triggers inflammation and activation of microglia.HMGB1 released by necrotic cells of all types produces neuroinflammation and disruption of the blood-CNS barriers.Astrocytes and microglia are CNS-resident cells that could respond to HMGB1 after injury as they express the RAGE receptor.However, the specific role of these cells in the HMGB1 response remains uncertain.
determined that peripheral injections of IL-1RA (intravenously or subcutaneously) were safe (Emsley et al., 2005;Galea et al., 2018;Gueorguieva et al., 2008;Helmy et al., 2014;Singh et al., 2014).Furthermore, it was reported that IL-1RA injected systemically crossed into the CSF in stroke patients (Gueorguieva et al., 2008), as well as into cerebral microdialysate samples in TBI (Helmy et al., 2014).This suggests that despite the blood-CNS barriers blocking the passage of this molecule from the blood into the CNS in healthy conditions, injuries such as stroke and TBI increase BBB permeability, and allow IL-1RA to enter the CNS and carry out its function.Additionally, these studies reported a trend towards reducing inflammation compared to control patients, as evidenced by decreased neutrophil levels in blood (Emsley et al., 2005), decreased IL-6 levels in plasma and CSF in stroke patients (Singh et al., 2014), as well as a reduction in macrophage-derived chemoattractant (also termed as CCL22) in TBI patients (Helmy et al., 2014).However, determining the efficacy of this drug in achieving improved outcomes in the context of CNS injury is difficult due to the limited number of participants in these studies.
Overall, targeting IL-1α has shown promising results in the context of CNS injury.This cDAMP appears to promote inflammation by recruiting immune cells and favoring the expression of various other proinflammatory cytokines, producing an increase in lesion size and worsening injury outcomes.Additionally, the use of inhibitors targeting the IL-1α/IL-1R1 axis have been shown to be safe and to yield promising results in humans.However, larger-scale clinical studies are needed to truly assess the efficacy of IL-1α treatment in SCI, TBI, and stroke.

INTERLEUKIN-33
IL-33 is another member of the IL-1 family of cytokines, involved in regulation of the immune response.IL-33 is highly expressed in several tissues, including the lungs, spleen, lymph nodes, stomach, skin, brain, and spinal cord (Pichery et al., 2012;Schmitz et al., 2005).IL-33 is synthesized either as a precursor or in its full-length form, featuring the characteristic IL-1 fold in the C terminus (Lingel et al., 2009).The receptor for IL-33 is ST2, also known as the IL-1 receptor like 1 (IL-1RL1), which activates the NF-κB pathway via MyD88 (Martin and Martin, 2016) (Fig. 2).In homeostasis, IL-33 is crucial for maintaining regulatory T cells, as these cells can directly respond to this molecule (Molofsky et al., 2015).In necrotic processes, full-length IL-33 is released and can activate ST2 (Lüthi et al., 2009).Additionally, it has been reported that the full-length form of IL-33 released into the extracellular environment can be cleaved by neutrophil-derived proteases such as cathepsin G and elastase, thus generating an IL-33 cytokine with ~10-fold higher activity than full-length IL-33 in inflammatory conditions (Lefrançais et al., 2012).
IL-33 was first proposed as a DAMP in 2009 (Haraldsen et al., 2009), and various factors enable IL-33 to fulfill this function in the CNS.IL-33 has been detected in astrocytes and mature oligodendrocytes in different regions of CNS such as the corpus callosum, optic nerve, cortex, and spinal cord, among others (Gadani et al., 2015b).Moreover, IL-33 has been recently found in hippocampal neurons in adult mice (Nguyen et al., 2020).Regarding the expression of ST2, it is primarily found in microglia and astrocytes (Gadani et al., 2015b;Yang et al., 2017), allowing for the action of both IL-33 and its receptor in the event of CNS damage (Fig. 3).The recent discovery of ST2 expression by type 2 innate lymphoid cells (ILC2) present in the brain and spinal cord meninges, both in healthy and injured states, also highlights the importance of IL-33 outside of the CNS parenchyma (Gadani et al., 2017).

IL-33 exerts anti-inflammatory and neuroprotective effects in the injured CNS
As expected from a cDAMP, IL-33 is rapidly released into the extracellular environment of the CNS following an injury.In mouse models of stroke, an elevation in IL-33 levels is observed 1 day and 3 days after injury (Korhonen et al., 2015;Yang et al., 2017).Similarly, IL-33-cells expressing ST2, such as microglia and astrocytes (Fig. 3).Not to mention that cells located just outside of the injured CNS parenchyma, such as in the meninges, may also play a role in the protective effects of IL-33.Interestingly, Gadani et al. demonstrated that brain meningeal ILC2 cells are acutely activated in an IL-33-dependent manner after SCI, producing anti-inflammatory cytokines such as IL-5 and IL-13 (Gadani et al., 2017).Further, injections of lung-derived ILC2 cells into the cisterna magna of ST2-KO recipient mice 1 day before SCI resulted in smaller lesions and better locomotor recovery.
IL-33 has been administered exogenously in murine models of SCI, TBI, and stroke.Although the treatments differ in doses and timing, they all share the commonality of being administered shortly after inducing the injury (or 30 min prior in the case of Gao and colleagues (Gao et al., 2018)) and through peripheral administration (intraperitoneal or intranasal).Regardless of the type of injury, mice administered with IL-33 showed improvements in sensory and/or motor functions compared to vehicle administration, along with a reduction in lesion size (Gao et al., 2018;Pomeshchik et al., 2015;Xie et al., 2021;Xie et al., 2022).When exogenous IL-33 was administered to ischemic mice, no significant changes were observed between IL-33 and vehicle-treated groups when microglia were depleted (Xie et al., 2021).However, in mice retaining microglia, IL-33 exhibited a neuroprotective effect with a reduction in lesion volume and oligodendrocyte loss.These results highlight the crucial role of microglia as primary responders to IL-33, consistent with the strong expression of ST2 at the surface of Iba1 + cells in the ischemic brain (Xie et al., 2021;Xie et al., 2022;Yang et al., 2017).Considering these positive results in the context of CNS injury, it would be pertinent to conduct dose-response studies for IL-33 administration.Recent studies have shown that IL-33 might have an important role in synaptic modulation and plasticity in hippocampal neurons, potentially beneficial for learning and memory in mouse models (Nguyen et al., 2020;Wang et al., 2021).It is important that future studies exploring the effects of exogenous administration of IL-33 in the context of injury also analyze the possible effects this treatment could have on hippocampal neurons.Additionally, analyzing the optimal route of administration and time frame for administration is important for potential translational applications in humans.
Currently, no studies have been conducted on the exogenous administration of IL-33 in humans, and no data is available regarding its safety or efficacy in CNS or peripheral pathologies.In a clinical study conducted on two hundred and six patients, Qian et al. demonstrated the presence of IL-33 in the serum of patients who suffered an acute ischemic stroke (Qian et al., 2016).Notably, serum IL-33 levels were inversely correlated with the severity of the stroke, i.e. higher serum IL-33 levels were associated with smaller cerebral infarction volume and improved outcomes.Considering the beneficial effects of IL-33 in the injured CNS, investigating this cDAMP as a therapy to reduce neuroinflammation and tissue damage in stroke and other CNS-injury pathology seems promising.

Adenosine 5′ triphosphate
Nucleotides such as adenosine 5′-triphosphate (ATP) have long been recognized to play a fundamental role in cellular metabolism, cellular signaling, and neurotransmission (Abbracchio et al., 2009;Burnstock, 2020).ATP levels are primarily regulated by mitochondrial activity through the electron transport chain and oxidative phosphorylation (Szabo and Szewczyk, 2023).Although ATP acts as a key intracellular molecule, the idea that it may also serve as an extracellular messenger has drawn considerable attention since the early 2000′s.There is now a strong scientific consensus that extracellularly released ATP acting through P2 nucleotide receptors is involved in many pathophysiological responses (Franke et al., 2006;Khakh and North, 2006).Of particular interest here is the fact that signaling by P2 receptors is implicated in the initiation of neuroinflammation in response to CNS injury, establishing ATP as a potent cDAMP.
There are two types of purinergic receptors: P1 receptors, which have adenosine and AMP as their specific ligands, and P2 receptors, which exhibit greater specificity for nucleotides such as ATP and ADP (Burnstock, 2020).In the CNS, all types of P2 receptors are expressed in both neurons and glial cells, being widespread in most CNS regions (Köles et al., 2011;Surprenant and North, 2009).P2 receptors can be divided into two subfamilies: metabotropic receptors of the P2Y class (Gprotein coupled receptors) and ionotropic receptors of the P2X class (ligand-gated cationic channels) (Burnstock, 2007) (Fig. 2).Accordingly, their functions in homeostasis remain varied, and P2 receptors have wide applicability in the CNS as synapse mediators, mechanisms of glial communication, and regulators of synaptic plasticity (Köles et al., 2011).The role of ATP and purinergic receptors in inflammation and immunity is well-known and has been extensively reviewed elsewhere (Bours et al., 2006;Di Virgilio et al., 2017).Of relevance to this discussion is the demonstration that mitochondria-produced ATP released from necrotic cells has the ability to trigger a sterile inflammatory response, acting as a 'find-me' signal to recruit neutrophils (Iyer et al., 2009;McDonald et al., 2010).Moreover, scientific evidence generally supports the idea that after CNS injury such as trauma or ischemia, ATP may serve as an endogenous 'alarm' signal for microglia, a cell population known for their ability to quickly respond to damage (Nimmerjahn et al., 2005;Pineau and Lacroix, 2009).All of this highlights ATP as a promising molecule to study, especially for its role as a cDAMP in CNS injury.

ATP guides immune cells to sites of CNS injury and induces neuroinflammation
As expected from a cDAMP, ATP levels rapidly increase following CNS damage.In SCI models, ATP is detectable 2 h after the primary trauma, localized to the lesion site and immediately surrounding areas (Wang et al., 2004).In a mouse model of TBI, ATP levels were found to increase within the first 20 min post-injury (Faroqi et al., 2021;Moro et al., 2021).Specifically, this increase occurs after the primary impact, with a return to pre-injury levels after 20 min (Faroqi et al., 2021).Early detection of ATP was also reported using the MCAO model in rodents, with a continuous increase in extracellular ATP induced within the first minutes after the insult (Melani et al., 2005;Wilmes et al., 2022).
Different strategies have been employed to elucidate the role of ATP and the activation of P2 receptors following CNS damage.In 2005, Davalos et al. conducted an elegant study wherein they induced damage in the cerebral cortex using a two-photon laser, demonstrating a movement of microglial processes dependent on ATP, which was blocked after administration of ATPase apyrase or various inhibitors of P2Y receptors, but not with P2X inhibitors (Davalos et al., 2005).These findings, confirmed by Haynes and colleagues using P2Y12-KO mice (Haynes et al., 2006), provided direct evidence that ATP activation of P2Y12 is responsible for microglial mobility in the early acute phase of brain injury.Importantly, the role of the P2Y12 receptor in the early ATP-dependent chemotactic activity of microglia was also demonstrated in the context of laser-induced cerebrovascular damage, wherein microglial processes proved essential for the rapid closure of the BBB (Lou et al., 2016).Although 'resting' microglia respond within minutes to ATP released after laser-induced CNS injury by first extending their processes and then their cell bodies to rapidly shield the lesion site (Bastien and Lacroix, 2014), they subsequently retract their processes, adopt an amoeboid morphology and downregulate the P2Y12 receptor following activation (Bellver-Landete et al., 2019;Haynes et al., 2006).It should also be pointed out that the use of P2Y12 KO mice only delayed the migration of microglia by a few hours (Haynes et al., 2006).This suggests that P2Y12 is important, but not essential for the chemotactic response of microglia after CNS injury, and that other P2 receptors and/ or cDAMPs might be involved.The existence of such compensatory mechanisms emphasizes the importance of this microglial response, wherein they proliferate and migrate extensively toward the edges of the lesion during the first week post-SCI.Here, they aid in the formation of a glial scar that limits the spread of blood-derived innate immune cells and the expansion of tissue damage (Bellver-Landete et al., 2019).Future work will be needed to identify mechanisms allowing for microglia migration and whether they play a beneficial or detrimental role in the context of CNS injury.
The P2X receptor subfamily has been more extensively studied than P2Y receptors in the context of CNS injury.In a recent TBI study, pharmacological blockade of the P2X4 receptor was shown to reduce brain edema and inflammation, decrease apoptosis, and improve neurological function (He et al., 2022).In the context of SCI, mice lacking the P2X4 receptor present with impaired inflammasome signaling, resulting in reduced IL-1β levels and decreased infiltration of innate immune cells (de Rivero Vaccari et al., 2012).This translated into a reduced lesion volume and improved recovery of motor function, suggesting a proinflammatory role for the P2X4 receptor in SCI.Although the exact series of events leading to inflammasome activation and IL-1β release after SCI remains ambiguous, it has been suggested that stimulation of the P2X4 receptor could result in IL-1β maturation via a large multiprotein complex composed of pannexin-1, P2X7, a NODlike receptor pyrin domain-containing protein (e.g.NLRP1, NLRP2, NLRP3), and caspase-1 (de Rivero Vaccari et al., 2016).P2X4 global KO mice also exhibited a significant reduction in microglial activation and lesion volume at 3 days after ischemic stroke compared to WT mice (Verma et al., 2017).However, these differences diminished with time, and by 28 days after MCAO, both WT and KO mice showed similar tissue atrophy and neurological deficits.This suggests that in the absence of ATP-P2X4 signaling, other mechanisms produce reinstitution of inflammation, and consequently secondary tissue damage in the chronic phase of CNS injury.Interestingly, the same group reported a few years later that pharmacological blockade of the P2X4 receptor for 3 days starting 4 h after stroke protected against ischemic injury at both acute and chronic timepoints (Srivastava et al., 2020).This implies that the timing of the blockade of P2X4 receptors is crucial in optimizing recovery, and further research will be necessary to fully understand the molecular mechanisms underlying these effects and explore potential therapeutic interventions targeting P2X4 receptors.
Several studies have investigated the role of P2X7 receptors in CNS injury.In the initial hours following a CNS insult in rodents, an upregulation of P2X7 receptor expression has been witnessed in microglia/ macrophages and astrocytes (Kimbler et al., 2012;Liu et al., 2017;Zhao et al., 2016).In a rat model of SCI, pharmacological treatment with P2X7 receptor antagonists diminished neuronal and oligodendrocyte apoptosis in the peritraumatic area and improved long-term locomotor recovery (Wang et al., 2004).Further, the administration of P2X7 receptor inhibitors resulted in a significant reduction in IL-1β protein expression and astrogliosis after TBI (Kimbler et al., 2012;Liu et al., 2017), preserving the integrity of the BBB in the context of intracerebral hemorrhage (Zhao et al., 2016), suggesting a role for this purinergic receptor in inflammasome activation and neuroinflammation.This translated into a significant decrease in the number of apoptotic neurons in the cerebral cortex after TBI, and a greater number of surviving neurons at the site of trauma and in surrounding areas after CNS insult (Liu et al., 2017;Zhao et al., 2016).In addition, P2X7-KO mice showed a decrease in IL-1β production and a reduction in edema size compared to WT mice after TBI (Kimbler et al., 2012).Importantly, improvements in neurological outcomes were reported in rodents with TBI following either pharmacological or genetic inhibition of P2X7 receptors (Kimbler et al., 2012;Liu et al., 2017;Zhao et al., 2016).Transgenic mice overexpressing the P2X7 receptor exhibit a larger stroke size than WT mice after MCAO (Wilmes et al., 2022).However, after failing to observe any differences in cell death and infarct volume between P2X7-KO and WT mice in the MCAO model (Le Feuvre et al., 2002), Nancy Rothwell's group demonstrated that intracerebroventricular administration of P2X7 receptor agonists or antagonists both resulted in a significant increase in lesion volume 1 day after stroke in mice (Le Feuvre et al., 2003).
These contradictory results with the use of agonists and antagonists, which both produce a detrimental effect after MCAO, may be due to the presence of adaptive processes in response to pharmacological alteration of P2X7 receptors, or simply reflect the poor selectivity of these compounds.Recently, Koch-Nolte and colleagues designed nanobodies against P2X7 that are 1,000 times more potent in blocking the ion channel (i.e.effectively blocking ATP-induced release of IL-1β from monocytes) than other small-molecule P2X7 antagonists currently in clinical trials for the treatment of inflammatory diseases (Danquah et al., 2016).Intrathecal, but not intravenous, administration of these nanobodies resulted in a significant reduction in the stroke infarct volume (and % of parenchymal tissue loss) at 1 day after MCAO in mice (Wilmes et al., 2022).This suggests that the P2X7 receptor acts inside the CNS parenchyma by promoting inflammation and lesion formation.Furthermore, the more selective and efficient neutralization of P2X7 receptors using nanobodies compared to conventional antagonists could explain the conflicting results reported in previous studies.Along this line, future work should aim to compare the short-and long-term effects of nanobodies against P2 receptor subtypes, as it could demystify the role of these DAMP-sensing receptors in CNS injury.
Many clinical trials have been conducted for the inhibition of purinergic receptors, and a wide variety of substances have already been approved for use in humans (Han et al., 2022).Considering the positive results in inhibiting P2X7 in animal models, inhibitors of this receptor may be a good therapeutic target for clinical trials in patients with CNS injury.An interesting example of such therapeutics is JNJ-54175446, a P2X7 receptor antagonist that when given orally can passively cross the blood-CNS barriers and was proven to be safe in healthy participants enrolled in a randomised single-ascending dose study (Timmers et al., 2018).Notably, treatment of major depressive disorder patients with JNJ-54175446 was found to partially block the acute reduction of pleasure that is typically experienced after total sleep deprivation, applied here as a behavioral challenge to investigate the efficacy of the P2X7 inhibitor on the recovery of acute mood changes (Recourt et al., 2023).While inhibition of P2Y12 appears less promising for treatment of human CNS injury, this receptor may have an important role in the activation and migration of microglia, and studies have yet to determine whether this is beneficial or detrimental in the context of CNS injury.Additionally, P2Y12 is an important receptor in platelet activation (Dorsam and Kunapuli, 2004), and the use of P2Y12 inhibitors to limit platelet activity is included in the American Stroke Association guideline for stroke prevention (Kernan et al., 2014).However, an observational study conducted on TBI patients determined that individuals taking P2Y12 inhibitors had a higher mean in the Injury Severity Score (Yorkgitis et al., 2022), likely due to its antithrombotic activity.Thus, the use of these inhibitors in traumatic injuries such as SCI or TBI where platelet activity is important for controlling bleeding may not be recommended.
In conclusion, ATP elicits a rapid neuroinflammatory response to CNS injury, consistent with its role as a cDAMP.Due to the numerous receptors present in the CNS that can interact with ATP, the resulting effects can be diverse.Much work remains to be done in this field, as little is known about the underlying mechanisms triggering the described responses.Additionally, while most publications focus on the effects of ATP activation in microglia, the effects on other CNS cell types remain largely unknown.Future studies should also investigate the combination of blocking multiple P2 receptors, as initial findings indicate the existence of compensatory mechanisms when only a single receptor is inhibited.

High mobility group box 1
High mobility group box 1 (HMGB1) protein, also known as amphoterin, is a highly abundant and conserved non-histone DNA binding protein that belongs to the HMGB protein group (Agresti and Bianchi, 2003;Jantzen et al., 1990;Stros, 2010).Under physiological conditions, HMGB1 is predominantly located in the nucleus, acting as a molecular chaperone, stabilizing nucleosome structure and regulating gene transcription (Kang et al., 2014).HMGB1 plays a role in cell surveillance, motility, proliferation, maturation, repair of DNA damage and death, and inflammation (Lange and Vasquez, 2009;Lotze and Tracey, 2005;Yanai et al., 2009).A role for HMGB1 as a cDAMP and mediator of sterile inflammation was first suggested by the pioneering work of Marco Bianchi, who showed in the early 2000 s that necrotic or damaged mammalian cells passively release HMGB1 into the extracellular milieu (Scaffidi et al., 2002).Unlike necrotic cells, apoptotic cells firmly retain HMGB1 within their nucleus (associated with chromatin) and fail to release it even after undergoing secondary necrosis or partial autolysis.Additionally, they showed that cells lacking HMGB1 have a reduced capacity to induce inflammation.RAGE and TLR4 are the two main receptors to which HMGB1 binds, triggering a proinflammatory response that is dependent on NF-kB and MAPK signaling (Fig. 2).In addition to RAGE and TLR4, HMGB1 can also activate other receptors such as TLR2 and TLR9 (Andersson and Tracey, 2011;Wu et al., 2012).

HMGB1 contributes to neuroinflammation after CNS injury through the RAGE receptor
HMGB1 is a DNA-binding protein expressed in the nucleus of most neurons and glial cells (Faraco et al., 2007;Kigerl et al., 2018).Following exposure to different necrotic stimuli, but not apoptotic stimuli, HMGB1 was found to be released into the extracellular space, triggering glial cells to adopt a proinflammatory profile (Faraco et al., 2007).Aside from its role as a cDAMP, HMGB1 is synthesized via transcription and actively secreted by macrophages and microglia during the acute phase of traumatic SCI (Kigerl et al., 2018), thus conferring to HMGB1 the main characteristics of an iDAMP.Notably, HMGB1 protein levels reach their peak 1 day after SCI in the mouse spinal cord.Following microinjection into the intact mouse spinal cord, extracellular HMGB1 triggers a neuroinflammatory response characterized by the prevalence of proinflammatory microglia/macrophages as well as motor neuron death (Kigerl et al., 2018).In stroke, an increase in HMGB1 was reported in the CSF and plasma during the first day post-MCAO in rats (Kim et al., 2006;Zhang et al., 2011).HMGB1 plasma levels were also found to be upregulated during the first day post-injury in individuals who suffered from SCI, TBI, or stroke compared to healthy donors (Goldstein et al., 2006;Muhammad et al., 2008;Papatheodorou et al., 2017;Vogelgesang et al., 2010;Yang et al., 2022).These early detections of HMGB1 in the CNS and biological fluids such as CSF and blood in both animal models and humans after CNS injury underscore the potential dual role of HMGB1 as a DAMP and a biomarker for CNS damage.
To study the role of HMGB1 in CNS injury, various strategies to inhibit the activity of this protein or its main receptor, RAGE, have been employed.One of the most studied approaches has been neutralization using anti-HMGB1 monoclonal antibodies.In rodent models of TBI and stroke, beneficial effects have been reported following systemic administration of anti-HMGB1 antibodies compared to isotype controltreated mice, including alleviation of brain inflammation, reduced BBB disruption and edema, as well as a decrease in lesion size and improvements in motor function (Muhammad et al., 2008;Okuma et al., 2012;Zhang et al., 2011).In SCI, the work of Nakajo and colleagues focused on determining the suitable time window for therapies targeting HMGB1 in mice subjected to a 70-kdyn spinal cord contusion (Nakajo et al., 2019).Optimal results were obtained when the anti-HMGB1 neutralizing antibody was injected intraperitoneally at a concentration of 8 mg/kg of body weight within 6 h after SCI, as determined by longitudinal assessment of locomotor recovery using the Basso Mouse Scale (BMS).However, in Kigerl et al. 2018, a single daily intraperitoneal injection of the anti-HMGB1 neutralizing antibody (50 µg/injection/ day) during the first week post-contusion SCI (75 kdyn), starting 1 day prior to injury, failed to improve locomotor function or lesion pathology (Kigerl et al., 2018).Differences in the nature (e.g.specificity, affinity) of the neutralizing antibodies used, their concentration, and the timing of treatment may be the cause of these contradictions in treatment effectiveness.
Apart from therapies based on anti-HMGB1 antibodies, drugs interfering with HMGB1 synthesis and its action sites were also shown to be beneficial in SCI (Fan et al., 2020), TBI (Webster et al., 2019), and stroke (Gong et al., 2014;Xin et al., 2015).For example, interfering with HMGB1 transcription using short hairpin RNA (shRNA) reduced the expression of proinflammatory cytokines and hindered neuronal death in the MCAO model of stroke (Kim et al., 2006).Experiments with RAGE-KO mice were also conducted and revealed that mice lacking functional RAGE exhibit reduced BBB permeability in TBI and smaller lesions in stroke compared to WT littermates (Muhammad et al., 2008;Okuma et al., 2012).Seeing as HMGB1 can activate a broad range of receptors, these studies looked to confirm that the effects of HMGB1 were solely mediated by RAGE.Accordingly, a separate group of RAGE-KO mice were injected with an anti-HMGB1 neutralizing antibody, however this revealed no further improvement.The efficacy of glycyrrhizin (an inhibitor of HMGB1) or FPS-ZM1 (a RAGE receptor inhibitor), has also been tested in SCI mice (Fan et al., 2020).The administration of either one of these compounds was effective in reducing inflammation at the site of SCI, an effect correlated with decreased neuronal death and axonal and myelin pathology, reduced lesion size, and improved locomotor function.Together, these studies suggest that HMGB1 plays a detrimental role following CNS injury, functioning as a DAMP that stimulates neuroinflammation via the RAGE receptor.Given the close association between inflammation and secondary damage in these pathological contexts, HMGB1 stands out as a promising therapeutic target for various types of CNS injuries.However, further investigation is warranted to elucidate the cellular responders and mechanisms of action of HMGB1 in the injured CNS, thus enabling the development of targeted therapeutic interventions for its modulation.
To our knowledge, there are no clinical trials currently attempting to inhibit the HMGB1/RAGE axis in CNS injury.However, seeing as RAGE was found to interact with amyloid β protein and modulate its transport across the BBB, phase I and II studies have been conducted using antagonists of the RAGE receptor in Alzheimer's disease (Burstein et al., 2014;Galasko et al., 2014;Sabbagh et al., 2011).When administered at low doses, these drugs have been found to be safe and well-tolerated in patients with mild Alzheimer's disease.Although a potential decrease in cognitive decline was reported, the results were not entirely conclusive.Given the favorable outcomes of neutralizing the HMGB1/RAGE axis in animal models of CNS injury, it may be time to consider translating this knowledge to the clinic.

Conclusion
While the inflammatory response to traumatic events such as SCI, TBI, and stroke is both necessary and inevitable, the immune response to these events can be exaggerated and/or prolonged and become detrimental, leading to secondary tissue damage and exacerbation of neurological and physical deficits.After a CNS insult, necrosis occurs at the site of injury, wherein dying cells liberate their cytoplasmic and nuclear contents, ensuring a prompt inflammatory response.These released self-molecules serve as 'alarm' signals, attracting immune cells to the injury site for important tasks like phagocytosis.However, this immune response can also cause a significant and prolonged inflammatory surge that persists beyond the expected resolution time frame of days or weeks.Constitutive DAMPs, including IL-1α, ATP and HMGB1, exhibit rapid actions, saturating the CNS parenchyma post-injury and triggering an inflammatory cascade of events which promote secondary damage.The release of IL-1α after SCI induces permeabilization of the BBB/BSCB, the recruitment and infiltration of neutrophils and monocytes, and the death of mature oligodendrocytes, all contributing to deterioration of neurological and behavioral deficits.ATP's ability to bind to an extensive array of P2X and P2Y receptors firmly establishes it as a potent instigator and intensifier of inflammation.It also mediates the chemotactic activity of microglia when acting on the P2Y12 receptor and promotes neuroinflammation and secondary degeneration when activating P2X4 and P2X7 receptors.HMGB1 also plays a crucial role in promoting inflammation, as its presence can stimulate microglia and macrophages to adopt a proinflammatory phenotype and can exacerbate secondary damage by binding to the RAGE receptor.While most of the cDAMPs discussed in this review article demonstrate proinflammatory effects, IL-33 stands out by exerting neuroprotective effects and promoting the resolution of inflammation, as it also promotes the polarization of macrophages towards an anti-inflammatory phenotype.However, the specific mechanisms by which IL-33 promotes neuroprotection after CNS injury remain obscure.Although this review explored the impact of individual cDAMPs on the initiation of neuroinflammation, these molecules are most often released in combination in response to injury.However, studies have yet to examine these potential interactions and compensatory mechanisms between cDAMPs.This crosstalk is likely to be relevant for therapeutic targets and warrants further investigation.As research of DAMPs continues to develop, it will become possible to harness the neuroprotective potential offered by IL-33, while effectively mitigating the detrimental effects of excessive immune responses induced by inflammation-promoting DAMPs such as IL-1α, ATP, and HMGB1.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Table 1
The effects of cDAMPs on neuroinflammation and tissue damage after CNS injury.