Etiogenic factors present in the cerebrospinal fluid from amyotrophic lateral sclerosis patients induce predominantly pro-inflammatory responses in microglia

The exact pathomechanism of neurodegeneration in amyotrophic lateral sclerosis (ALS), esp. the sporadic forms of ALS, remains poorly understood, thus leading to the lack of an effective therapeutic intervention for the disease [1, 2]. Several studies have reported a non-cell autonomous glial involvement in the pathogenesis, and the relevance of innate as well as adaptive immunity in ALS has also been widely discussed [3]. However, the requirement of the dynamic modulatory signals from within the central nervous system (CNS) for the recruitment and regulation of the adaptive immunity across the blood-brain barrier (BBB)/brain-spinal cord barrier (BSCB) further highlights the importance of the resident immune cells of the CNS in the disease pathology [4]. Microglia, the specialized immune cells of the CNS, constantly survey and dynamically regulate the neuronal milieu in the healthy and diseased CNS [5]. Depending on the nature and extent of the insult, microglial cells have been proposed to adopt morphologically and functionally distinct reactive phenotypes. These distinct phenotypes may perform diversified functions ranging from facilitating the pro-inflammatory processes that promote neuroinflammation to inducing an anti-inflammatory process that is engaged in healing and wound repair [6]. A fine balance among these phenotypes is considered to be crucial for a competent surveillance system, and its disruption could lead to the self-propagating chronic neuroinflammation seen in several neurological disorders [7]. However, the typical classification based on microglial polarization into classically (M1) and alternatively (M2) activated microglia is controversial and has been parallelly challenged [8].

Activated microglia have previously been reported in the autopsy samples and animal models of familial ALS (FALS), as well as throughout the symptomatic stages as demonstrated by neuroimaging of the ALS patients [9‐11]. While many studies report microglial activation to be actively neurotoxic in ALS [12‐14], there are also studies that report the microglial involvement to ne either neuroprotective or having no significant role to play in the event of neurodegeneration in such ALS models [15, 16]. Further, the spatial microglia appeared to affect their activation status in these animal models [17]. Incidentally, majority of these studies were conducted with the transgenic models containing superoxide dismutase 1 (SOD1) mutations that are uncommon in sporadic ALS (SALS) (> 1%) [18]. In recent years, the focus has largely shifted to transgenic models with novel, gene mutations more commonly reported in ALS patients, thus narrowing the gap between animal models and actual disease etiopathogenesis. Of prominence are ubiquitinated cytoplasmic inclusions (~ 98%)/mutations (FALS 5%, SALS 1%) in 43-kDa TAR DNA-binding protein (TDP-43), fused in sarcoma (FUS) (FALS 4%, SALS < 1%) and, most prominently, the hexaneucleotide repeat expansion in C9orf72 (FALS 40%, SALS > 10%) [19‐21]. However, while these models recapitulate the etiopathogenesis of FALS more effectively, their relevance from the perspective of SALS pathogenesis, which constitutes 90% of the ALS etiology, is still dubious [22]. Although the emergence of these models has opened newer research avenues including aberrant RNA processing and protein degradation pathways, as well as perturbed nucleocytoplasmic transport [21], the neuroinflammatory pathways, esp. in SALS remain unraveled. Moreover, it remains unclear whether the microglial cell-mediated neurotoxicity or protection targeted the mutated protein or mimicked the overall disease pathology. The recent in vitro approaches targeting the induced pluripotent stem cells from SALS patients have been a significant step, but majority of the studies conducted till date have focused on the motor neuron pathology, precluding their relevance to microglial contribution [23, 24].

Therefore, we studied the effect of the disease-related factors circulated in the cerebrospinal fluid of ALS patients (ALS-CSF) on microglia to understand the direct nature of the insult in the SALS pathogenesis. Our hypothesis was based on the assumption that ALS-CSF may provide a mirror to the molecular effectors being produced and propagated in the diseased CNS, which may lead to progression of the disease. The neurodegenerative potential of ALS-CSF has already been documented in a plethora of molecular, electrophysiological, and behavioral studies, and the efficacy of the model in recapitulating the ALS pathology has been suggested [25‐32]. Our previous investigations that aimed at determining the putative toxic agent(s) by the proteomic analysis of ALS-CSF demonstrated multifold upregulation of chitotriosidase-1 (CHIT-1), an inflammatory protein secreted by activated macrophages [33, 34]. Therefore, we investigated the response of microglia towards the ALS-CSF, as well as towards CHIT-1 in doses comparable to those found in the CSF samples taken for the study [33]. Microglial cultures exposed to ALS-CSF were analyzed for the production of toxic factors including free radicals and nitric oxide release as well as for inflammatory cytokines interleukin-6 (IL-6), interleukin-10 (IL-10), interferon-γ (IFN-γ), and tumor necrosis factor-α (TNF-α), Cyclooxygenase-2 (COX-2), and prostaglandin E2 (PGE2). In order to determine the trophic modulation in microglia in response to ALS-CSF, the changes in the expression patterns of vascular endothelial growth factor (VEGF) and glial cell line-derived neurotrophic factor (GDNF) were studied.

Microglial activation is a dynamic process and may involve various temporally, physiologically, and/or spatially regulated events that govern the morphological and functional changes observed in the reactive microglia [6, 42, 43]. The pro-inflammatory microglial transformation observed in response to ALS-CSF indicates an inflammatory microenvironment of the CNS in SALS that may be further dynamically regulated by the actively transforming microglia, in a closed, amplifying feedback loop. The ability of the conditioned media from the microglial cultures exposed to ALS-CSF to exert toxicity to the NSC-34 cells corroborates the non-cell autonomous disease propagation in ALS [42]. Recent studies discuss the importance of MV/exosomes in the neuronal spread of TDP-43 proteinopathies in ALS [44, 45]. Studies conducted on the glial MVs released by the reactive astrocytes and glioma cells hint towards a possible engagement of MVs in the intercellular communication [46, 47]. Further, investigations on microglial microvesicular release in neuroinflammation also suggest its participation in disseminating neurotoxicity via proteins and/or microRNAs [48, 49]. The reports of pro-inflammatory nature of microglial MVs, especially at early time points (8 h) as opposed to the anti-inflammatory nature at later stages (> 24 h), support our observations of early pro-inflammatory secretion and concomitantly increased vesicular structures in the ALS-CSF-exposed microglial cultures [50, 51]. Interestingly, at 48 h, ALS-CSF also induced multinucleation and cytorrhexis in the microglial cultures, which hints at dysregulated pathways leading to failed cytokinesis and overactivation-induced self-moderation [52, 53]. Such a phenomenon has also been reported in another study involving the mutant SOD1 mice model, where it is explained as an accidental product of the high-intensity neuroinflammation [41]. Induction of NF-κB, the inflammatory regulator for glial activation, has been documented in the autopsy studies as well as TDP-43 models [54‐56]. Further, studies involving the C9orf72 pathology suggest towards a potential role of the microglial C9orf72 repeat expansions in exacerbating neuroinflammation, resulting in a loss-of-function endosomal-lysosomal aberration, thus favoring a pro-inflammatory microglial phenotype [57]. These observations highlight the role of neuroinflammation in the disease pathology irrespective of the nature of the etiogenesis.

Upregulation of microglial inflammatory mediators and retraction of the beneficial trophic support demonstrate the ability of ALS-CSF to shift the microglial functionality predominantly towards an activated pro-inflammatory physiology that is able to propagate inflammation by sending inflammatory signals in the circulation. Overexpression of iNOS, along with the downregulation of arginase, may suggest the induction of a pathological switch towards physiologically inflamed or the classically activated (M1) microglia in response to ALS-CSF [58]. This correlated with the increased expression and secretion of IL-6 and TNF-α, as well as a temporal decline in the secretion of the neuroprotective cytokines IL-10 and IFN-γ. However, the upregulation of IL-10 mRNA levels and the differential regulation of COX-2 mRNA in response to different ALS-CSF samples, contrary to their pro-inflammatory protein expression patterns, also depict a conflicting state of activation that may defy the typical M1/M2 polarization. Studies conducted in the mutant SOD1 mice model also suggested a disease-specific microglial profile, where a simultaneous dysregulation was reported in the expression of M1- and M2-specific markers in early stages, with either a slight bias [10] towards the M1 type or none [17] at the later stages. Since in our study the overall findings support the neurotoxic outcomes, we propose that these profiles may arise either as the result of a dynamic, reversible switch between a pro- and anti-inflammatory phenotype or a failed attempt of the repair processes at work.

l-Glutamate release from activated microglia downregulates astroglial glutamate transporter expression [59], a pathological phenomenon well documented in ALS [60]. Apart from this, the elevated microglial glutamate may also exacerbate neurodegeneration through dysregulated gliotransmission via purinergic receptors and Ca²⁺ wave propagation [61] or induce necrosis and necroptosis as seen in response to ALS-CSF [27, 38]. Alternatively, excess glutamate can also accentuate neuroinflammation by interacting with inflammatory modulators including ROS [62], PGE2 [63], and NO [64], which were found elevated in response to the ALS-CSF. The neurotoxic potential of the conditioned media from microglial cultures exposed to the ALS-CSF (ALS-MCM) corroborated the pro-inflammatory role of the factors predominantly upregulated in ALS-MCM, including IL-6, TNF-α, and NO. Further, the higher quantum of TNF-α levels, specifically in the initial phase, suggests its role as a potent, early initiator of neuroinflammation, much in accordance with the studies conducted in mSOD1 models [65, 66]. Recent literature elucidates the neuroprotective nature of microglial IFN-γ signaling in CNS pathology through the recruitment of immunoregulatory cells including macrophages, and their dysregulation in the neurodegenerative disorders [4, 67, 68]. The downregulation of microglial IFN-γ secretion may thus suggest a reduced anti-inflammatory leucocyte infiltration leading to the aggravation of the disease pathology.

Apart from a reduced trophic support, downregulation of microglial VEGF has been shown to correspond with a phagocytic microglial phenotype [69]. In a similar manner, a study by Matsushita and colleagues also reported a significant downregulation of GDNF in response to endotoxin LPS, suggesting that the reduction in GDNF expression could be a feature of inflammation [70]. These studies highlight the protective roles of both these trophic factors in disease pathogenesis through inhibition of inflammation and provide sufficient evidence to consider their relevance in clinical applications.

These observations clearly suggest that following exposure to ALS-CSF, microglial cells are activated and adopt a toxic phenotype. Additionally, we have provided compelling evidence inflammatory response of the spinal cord-derived microglia to ALS-CSF-mediated insult. Interestingly, in contrast to the early microglial activation, the maximal degenerative changes in the neurons exposed to ALS-CSF were observed at 48 h [27]. Similarly, the first known astroglial response to ALS-CSF in terms of gliosis or cytokine secretion was not observed before 24 h and became prominent at 48 h. The present observations, along with the previously reported, marginally delayed, inflammatory behavior of astrocytes towards ALS-CSF, might suggest a synergistic interplay between microglia and astrocytes in potentiating, sustaining, and aggravating the insult [36].

In addition, the present study attempts to explain the continued failure of anti-inflammatory therapeutic approaches in the clinical trials, despite obvious pathological contribution of inflammation in ALS [71, 72]. Since the alterations observed in the microglial physiology can at best be described as multifactorial with the interplay of components like pro-inflammatory and anti-inflammatory cytokines, PGE2, ROS, and glutamate, mitigating inflammation by blocking specific molecules/pathways would be inefficient. Moreover, the lacuna with employing anti-inflammatory drugs to target inflammation in general lies in the controversial observations of varied effects of these drugs, ranging from amelioration to exacerbation of the disease pathology, in different conditions, thus rendering them unsuitable for therapeutic interventions. For instance, minocycline was reported to be protective in the animal studies during the initial stages, as well as in the clinical trials at specific dosage, and stages of the disease, but had detrimental effects in the later stages or in higher doses [72‐74]. Hence, it is imperative to thoroughly investigate the mechanism of microglial inflammation in order to pin down the exact pathways affected in ALS, and design a combinatorial approach to target microglial activation and neuroinflammation.

CHIT-1, the protein primarily upregulated in the ALS-CSF, is considered as a marker of chronic activation of macrophages in the peripheral system [75]. Although CHIT-1 has also been implicated in various neurological diseases, its function in the CNS pathology remains least understood. While CHIT-1 is shown to have a pro-inflammatory potential in neurological disorders like stroke [76], its neuroprotective role in the inflammatory conditions like multiple sclerosis has also been suggested [77]. The microglial cell-specific upregulation of CHIT-1 expression in response to ALS-CSF suggests the possibility of neuroinflammatory process in ALS patients, led by the chronically activated microglia. The selective action of CHIT-1 on microglial cells suggests a vicious cycle of neurodegeneration primarily mediated by microglial cells and propagated to others via a cell-to-cell communication and/or circulating fluids. Since the reactive microglia in ALS are derived from the endogenous pool and not from the circulating monocytes [10], the endogenous microglial population may act as the predominant source of CHIT-1 in ALS-CSF, thus rendering them as potential clinical targets.

Based on our findings, we propose that microglial pathology in SALS is morphologically and functionally dynamic, leading to the neurodegeneration through myriad pathways (Fig. 8). It appears to operate by a push-forward mechanism, where microglia are the first to strike, possibly accompanied by astrocytes, thus setting up a neuroinflammatory cascade and eventually leading to accentuated neurodegeneration. In the light of the above, a combinatorial therapeutic approach is warranted, by suppressing the pro-inflammatory pathways and promoting the anti-inflammatory role of microglia.