Spinal cord injury (SCI) induces primary damage to neurons and glia due to mechanical trauma, followed by secondary injury characterized by a cascade of inflammatory processes that cause further cell death of oligodendrocytes, astrocytes, and especially neurons (1). A key feature of the inflammatory response in SCI is the infiltration of monocyte-derived macrophages (MDMs) into the lesion site and the activation of resident microglia in the central nervous system (CNS). Both cell types act as major phagocytes that remove tissue debris after injury, which is crucial for creating a regenerative environment. It is well established that tissue debris from peripheral injuries is cleared by MDMs within weeks after injury. However, after SCI, the persistent demyelination produces myelin debris and other lipid-rich histiocyte debris that induce the transformation of macrophages into foamy macrophages, resembling those found in atherosclerotic plaques. These foam cells persist in the lesion site and mediate chronic inflammation, which hampers nerve repair (2, 3). T The molecular mechanisms of foam cell formation and function are poorly understood, but their phagocytic and migratory abilities have been reported to be impaired (2–6). Thus, foam cells after SCI represent a potential target for novel SCI research and intervention.

In this review, we provide a brief overview of the lipid metabolism of macrophages in the inflammatory milieu, and how it influences their function and phenotype. This will help to elucidate the mechanisms and consequences of macrophage phenotypic alterations after SCI. Furthermore, we summarise and discuss the factors that induce foam cell formation after SCI and the potential targets for modulating them, aiming to identify the strategies for foam cell regulation after SCI.

Different types of stimuli induce different macrophage phenotypes and metabolic profiles, and immune activation can be divided into four main components: inducers (signals that trigger an inflammatory response), sensors (proteins that detect the inducers), mediators (molecules that transmit signals to activate an effector response), and effectors (downstream metabolic programmes that facilitate the maturation of the desired effector phenotype) (7, 8). Based on this general classification, the macrophage response to stimulation can be described more clearly. The phenotypes that emerge from macrophage activation are commonly categorised as the M1 phenotype and the M2 phenotype. The M1 phenotype is mainly associated with acute bacterial infections, and the typical inducers are pathogen-associated molecular patterns (PAMPs), which are sensed by cell surface pattern recognition receptors (PRRs), leading to the production of the intermediate mediator, HIF-1α, which connects glycolytic metabolism to the inflammatory and microbicidal programmes of the macrophage (9). The M2 phenotype is driven by IL-4 and IL-13 produced by innate and adaptive immune cells (e.g., mast cells, basophils, and T cells) during processes such as helminthic and parasitic infections (10, 11). After IL-4 and IL-13 bind to their respective receptors (IL-4Rα vs. IL-13Rα), the subsequent activation of the M2 cells largely depends on STAT6, which is transcriptionally activated to upregulate fatty acid β-oxidation, OXPHOS and metabolic genes essential for mitochondrial biogenesis (12). Additionally, STAT6 induces the expression of transcriptional regulatory proteins and co-activator proteins, such as PPARγ, PPARδ and PGC-1β, which cooperate with STAT6 to sustain the switch of the cellular metabolic profile to oxidative metabolism (12–15). Different metabolic profiles of the M1 and M2 phenotypes stem from this, with the M1 macrophage using Warburg metabolism and the M2 macrophage using OXPHOS, which has been extensively discussed by others in the literature (16–19). The metabolism of M1 and M2 macrophages is bi-directionally linked to their functional phenotypes. Glycolysis in M1 macrophages enhances PPP flux, which provides intermediates for the biosynthesis of ribose, fatty acids and non-essential amino acids to sustain vital metabolic pathways. In contrast, the activation programme in M2 macrophages requires high levels of both duration and intensity, making the TCA cycle more prominent than glycolysis. Fatty acid oxidation is especially important in these cells to support the TCA cycle. One source of fatty acids is triglycerides, which are internalised by M2 macrophages via CD36 and then hydrolysed by lysosomal acid lipase (20). Besides uptake from external sources, macrophages also have the ability for lipid synthesis (21, 22). Microbial stimulation can augment macrophage de novo lipogenesis (DNL). Microbial stimulation also increases glucose utilisation, redirects intracellular glucose metabolism, and supplies a reaction substrate for intracellular lipid regeneration (DNL) in macrophages (23, 24). On the other hand, it can also activate the signalling pathway triggered by the transcription factor NF-κB, which directly binds to the response element in the promoter of SREBP1a (an isoform expressed in macrophages), thereby enhancing the expression of SREBP1a, which in turn promotes the oxidation of fatty acids in the DNL, providing nutrients for the TCA cycle (25). The inhibition of lipolysis reduces M2 macrophage-mediated resistance to parasite infection (26). These metabolic changes are not irreversible. Unlike metabolic shifts in cancer (genetically driven aerobic glycolytic transitions) or slow exercise-induced myofibre type conversions in skeletal muscle, many immune cells can swiftly switch between glycolysis and OXPHOS in response to external signals.

It is noteworthy that some vitro studies have proposed that the key determinant of the two phenotypes is a difference in the metabolic pathways used by the cells, i.e. the use of metabolic substrates controls macrophage polarization. However, this conclusion is disputed, with Chang HR et al. demonstrating that cellular lipid metabolism does not correlate with macrophage phenotype, and that the tissue environment is the main factor in determining how subpopulations of macrophages respond to changes in lipid synthesis and metabolic capacity (27). Furthermore, although the M1 and M2 phenotypes of macrophage activation facilitate the understanding of macrophage responses to stimuli, the in vivo phenotype of macrophages is more complex, especially in spinal cord injury (28).

As discussed above, macrophages display various functional states that influence tissue repair, and understanding the basic macrophage metabolism in inflammation will help to elucidate the complex changes that occur in macrophages after SCI. It is well known that the surrounding microenvironment is a key determinant of macrophage function (29). The spinal cord microenvironment undergoes drastic changes shortly after SCI. The early microenvironmental changes in SCI are characterised by an inflammatory response that induces monocytes to migrate to the injury site and differentiate into macrophages, which further amplify the inflammation. Inflammation after SCI is complicated and involves multiple cell types and a variety of inflammatory cytokines, such as tumour necrosis factor α (TNFα), interleukin-1β (IL-1β) and interleukin-6 (IL-6). Despite some beneficial effects of inflammation after SCI, excessive infiltration of immune cells is the main cause of neurodegeneration. After the acute inflammation resolves, a large number of cells undergo apoptosis, generating large amounts of cellular debris, particularly myelin debris. The accumulation of myelin and cellular debris at the injury site, combined with the pre-existing CNS glial environment, forms an inhibitory environment in the centre of injury, which impairs nerve repair processes, such as oligodendrocyte precursor cells (OPCs) differentiation and synapse regeneration (30, 31). By removing myelin debris, lipid peroxidation of myelin-derived lipids in the injury centre is prevented and the inhibitory microenvironment is partially deregulated. Previous studies have demonstrated that the function of macrophages after SCI is significantly influenced by myelin debris, depending on the type and timing of stimulation, but this issue remains highly contentious (2, 32, 33).. Membrane lipids are classified into three major groups: cholesterol, phospholipids (e.g., plasminogen, lecithin, sphingomyelin), and glycolipids (e.g., galactose ceramide). Unlike most biological membranes (25%:65%:10%), myelin has a distinctive lipid composition, with a high proportion of cholesterol and glycolipids in a 40%:40%:20% ratio (cholesterol, phospholipids, and glycolipids) (34), and a total lipid dry weight of 70-75% (35, 36). This creates a unique lipid-rich environment. Dysfunctional lipid metabolism in macrophages, coupled with the lipid-rich environment, leads to the accumulation of intracellular lipids in macrophages, resulting in the formation of cell types resembling the specialised macrophages called foam cells in atherosclerotic plaques ( Figure 1 ), which are also referred to as foam cells after spinal cord injury. Foam cell is a term widely used in the study of diseases such as atherosclerosis and tuberculosis. Foam cell formation is induced by various factors, such as uncontrolled uptake of modified low-density lipoprotein (LDL), increased cholesterol esterification, and impaired cholesterol release mechanisms (37). Macrophages internalize modified lipoproteins via surface scavenger receptors (SRs) like CD36 and SR-A. These SRs are macrophage PRRs that recognize and bind oxidized LDL, thereby promoting foam cell formation by internalizing these lipoproteins. Coated-pit endocytosis, phagocytosis and pinocytosis also mediate lipoprotein internalization. After internalization, cleared lipoproteins are transported to endosomes or lysosomes for degradation, where cholesteryl esters (CE) are hydrolyzed by lysosomal acid lipase (LAL) to unesterified free cholesterol (FC). Free cholesterol is transported to the endoplasmic reticulum where it is re-esterified by acyl-CoA: cholesterol acyltransferase 1 (ACAT1) and stored as cytoplasmic lipid droplets. These lipid droplets look like bubbles within macrophages, hence the name foam cells. The foam cell formation process above has been detailed in articles (38, 39). It is noteworthy that while most studies have focused on macrophages as the main or only source of foam cells, the role of smooth muscle cell-derived foam cells in atherosclerosis has gained more attention (40). This article focuses on macrophage-derived foam cells, but the reader should be aware that foam cells are defined by their morphology and not by their precursor cells.

These foam cells are typically observed in the injury area 1 week after SCI in mice and persist for at least 4 weeks (2).. Early studies suggested that the uptake of myelin residues biased macrophage polarisation towards a beneficial anti-inflammatory phenotype (41), but more recent studies have shown that foamy macrophages in SCI, which lose their ability to phagocytose apoptotic cells and tissue debris, activate NF-κ B signalling, which promotes inflammation and may contribute to further neurodegeneration (2, 3). The direction and modes of intervention of foam cells will be discussed later.

Pathway analysis of genes involved in macrophage lipolysis metabolism revealed that lipolysis metabolic processes were a major function of macrophages seven days after spinal cord injury (compared to three days after spinal cord injury), with hubs including TNF, CD36, LPL, PPARγ, and ABCA1, among which the LXR/RXR pathways (CPs) were considered to be one of the most enriched pathways (42). Except for TNF expression, which was decreased, all gene-enriched pathways were increased, and the above receptors and pathways involved in phagocytosis of tissue debris and lipid efflux might be a key direction in dealing with foam cells ( Figure 2 ). To discuss the direction of foam cell intervention, we first need to clarify the functional differences between MDMs and microglia after spinal cord injury. Recent lineage tracing studies have indicated that almost all tissue-resident macrophages in adult mice originated from embryonic yolk sac progenitor cells and that their numbers were maintained by local proliferation throughout the organism’s life cycle (43–45). It is worth noting that there was confusion about the distinction between MDMs and microglia in previous studies due to the difficulty in distinguishing them from each other. The two types of cells showed similarities in morphology as well as gene and cell surface protein expression after activation, but the two types of cells differed in the timing of activation, location, and efficiency of processing of myelin debris after SCI. After SCI, as a result of primary mechanical injury and secondary death due to apoptosis, the number of microglia at and near the site of injury was significantly reduced (46). However, surviving microglia responded rapidly to ATP released by damaged cells. They also responded to danger and pathogen-associated molecular patterns (D/PAMPs) by releasing inflammatory molecules and were in turn affected by inflammatory molecules. Once activated, microglia could proliferate, migrate, and perform various effector functions including phagocytosis and proliferation of neuroinflammation by secreting additional chemokines and cytokines, and had better myelin debris processing efficiency compared to MDMs (46, 47). However, MDMs that subsequently infiltrated into the injury area inhibited microglia phagocytosis and gradually replaced microglia as the primary phagocytes after injury (48). Although the reason why microglia were less susceptible to foam cell formation remained to be explored, their potential as a target for intervention in the regulation of lipid metabolism after spinal cord injury had been well demonstrated. Since MDMs were the main phagocytes that formed foam cells, the potential interventions discussed subsequently would mainly target MDMs cells.

After spinal cord injury, the clearance of tissue debris is important for the inflammatory process and the nerve repair process. However, the clearance of tissue debris is hindered by the formation of foam cells. Therefore, the effective regulation of the phagocytosis and metabolic profiles of MDMs for the smooth clearance and metabolism of tissue debris is a key issue for promoting nerve repair after spinal cord injury. We propose two intervention strategies targeting foam cells: (I) reducing the phagocytosis of lipids by macrophages: the phagocytosis of lipid-rich tissue debris (including but not limited to myelin debris) leads to lipid accumulation in macrophages, and this can be improved by regulating or knocking down the relevant lipid uptake receptors (e.g., CD36, CD204, LOX-1); (II) regulating the RCT of macrophages: this process mainly involves ApoA-I, ApoE, and ABCA1 and ABCG1. Accelerating the lipid efflux from macrophages can reduce the excessive accumulation of intracellular lipids and thus prevent the transformation of macrophages into foam cells. It should be noted that the intervention strategies for foam cells are not limited to the ones proposed here; for example, since CE can only be transported to the extracellular compartment after being hydrolysed to FC, the hydrolysis of CE in macrophages limits the cholesterol efflux to some extent, and thus the intervention of related enzymes may also have a positive effect; moreover, since FC available for efflux is mainly transported to the autophagosome via CE in the autophagosome translocation to lysosomes to be supplied by hydrolysis by acid lipase rather than by hydrolysis by neutral cholesterol esterase, the induction of macrophage autophagy or the modulation of lysosomal function may also be effective intervention strategies. Traditionally, foam cell formation has been mainly linked to cholesterol uptake via natural and modified LDL, and the high cholesterol levels in the microenvironment due to the unique lipid composition of myelin sheaths after spinal cord injury have led to the article’s focus on cholesterol efflux. However, other lipids, such as nonesterified fatty acids and triacylglycerol (TAG)-rich lipoproteins (very low-density lipoproteins and chymotrypsin), can also interact with macrophages. Moreover, the intracellular composition of lipid droplets in foam cells after spinal cord injury and the proportion of lipid components phagocytosed and accumulated by macrophages are unclear, as there is a lack of analysis and studies on these aspects. Therefore, besides the foam cell treatment methods summarized in this article based on cholesterol efflux, we encourage researchers to explore more diverse approaches to foam cells after spinal cord injury. In conclusion, although many studies have obtained positive results, further studies are required to understand the differential effects of foam cell intervention and its potential as a pharmacological target for SCI repair.

XW: Conceptualization, Validation, Visualization, Writing – original draft, Writing – review & editing. ZL: Investigation, Software, Writing – review & editing. HD: Conceptualization, Formal Analysis, Software, Writing – review & editing. WL: Conceptualization, Visualization, Writing – review & editing. CZ: Formal Analysis, Project administration, Writing – review & editing. XX: Supervision, Visualization, Writing – review & editing. HK: Writing – review & editing, Resources. RP: Supervision, Writing – review & editing. D–GY: Supervision, Writing – review & editing. JL: Funding acquisition, Supervision, Writing – review & editing. FG: Supervision, Writing – review & editing.