The paradox of chronic neuroinflammation, systemic immune suppression and autoimmunity after traumatic chronic spinal cord injury
Abstract
During the transition from acute to chronic stages of recovery after spinal cord injury (SCI), there is an evolving state of immunologic dysfunction that exacerbates the problems associated with the more clinically obvious neurologic deficits. Since injury directly affects cells embedded within the “immune privileged/specialized” milieu of the spinal cord, maladaptive or inefficient responses are likely to occur. Collectively, these responses qualify as part of the continuum of “SCI disease” and are important therapeutic targets to improve neural repair and neurological outcome. Generic immune suppressive therapies have been largely unsuccessful, mostly because nflammation and immunity exert both beneficial (plasticity enhancing) and detrimental (e.g. glia- and neurodegenerative; secondary damage) effects and these functions change over time. Moreover, “compartmentalized” investigations, limited to only intraspinal inflammation and associated cellular or molecular changes in the spinal cord, neglect the reality that the structure and function of the CNS is influenced by systemic immune challenges and that the immune system is hardwired into the nervous system. Here, we consider this interplay during the progression from acute to chronic SCI. Specifically, we survey impaired/non-resolving intraspinal inflammation and the paradox of systemic inflammatory responses in the face of ongoing chronic immune suppression and autoimmunity. The concepts of systemic inflammatory response syndrome (SIRS), compensatory anti-inflammatory response syndrome (CARS) and ‘neurogenic’ spinal cord injury-induced immune depression syndrome (SCI-IDS) are discussed as determinants of impaired ‘host-defense’ and trauma-induced autoimmunity.
I. Introduction
Injury to the spinal cord elicits a robust intraspinal inflammatory response (Popovich et al., 1996; Schwab et al., 2001; Rice et al., 2005; Fleming et al., 2006 ). Consecutively, an immune modulatory response evolves in parallel including potent anti-inflammatory mechanisms both within and outside the spinal cord, presumably to regulate ongoing inflammation initiated by trauma (Lucin et al., 2007; 2009; Riegger et al., 2007; 2009, for review see Meisel et al., 2005; Irwin and Cole, 2011) (Fig. 1). Whether these are functionally effective inflammatory or anti-inflammatory reactions is questionable, especially since there is no evidence that either is resolved in affected individuals or animal models. Indeed, intraspinal inflammation persists indefinitely (Prüss et al., 2011; Popovich et al., 1997; Rosenberg et al., 2005) and depending on injury level, anti-inflammatory or even autoimmune pathologies develop (Popovich et al., 1997; Hayes et al., 2002; Ankeny 2006; Zajarias-Fainsod et al., 2012). Interactions between the central nervous and immune systems, i.e., the two main systems regulating homeostasis throughout the body, are not limited to aberrant immune cell activation/accumulation behind the blood-spinal cord barrier (BSB). Instead, SCI affects the entire immune system (Fig. 1).
The classical neuroimmunological perspective to research on SCI or other neurologic diseases has been to focus on leukocyte functions in brain and/or spinal cord, i.e., two immune-privileged sites. However, the modulatory effects of a CNS lesion on immune function were neglected (neurogenic immune ablation). For example, injury to the vegetative, sympathetic nervous system partly withdraws the control of the CNS on the immune and endocrine organs. Loss of this hardwiring has been designated as spinal cord injury-induced immune depression syndrome (SCI-IDS) and may increases the susceptibility to infection in a lesion height dependent manner and may also paradoxically promote autoimmunity. Non-neurogenic mechanisms of immune regulation including “systemic immune response syndrome” (SIRS) or “compensatory anti-inflammatory response syndrome” (CARS) are also elicited by injury and disease and likely increase incidence of infections. Infections are the major cause of death after SCI and have been identified as ‘disease modifying factor’ (DMF) characterized as independent risk factor for poor neurological recovery.
Dynamic interactions between the “immune-privileged/-specialized” spinal cord and systemic immune organs innervated by the CNS are critical considerations for the evolution of the “SCI-disease” state that is initiated by traumatic SCI (Popovich and McTigue, 2009). In fact, even though the immune response is vital for maintenance of tissue homeostasis, immune system function continues to change as the injury evolves from the acute to the chronic state. The loss or dysfunction of vegetative innervation to lymphatic and endocrine tissues leads to a defective immune response long after the initial trauma (Meisel et al., 2005; Zhang et al., 2013). Here, we highlight different mechanisms that help explain protracted immune dysregulation including the development of chronic immune suppression and systemic autoimmunity. By doing so, we will address the localized immune response at the spinal cord lesion site and the impact on immune system function.
II. Intraspinal inflammation caused by trauma persists indefinitely
The inflammatory response elicited by trauma has long been viewed as an acute response, distinct from chronic inflammation. Nevertheless, intraspinal inflammation after SCI includes a non self-limiting, smoldering inflammatory cascade (Prüss et al., 2011; Rosenberg et al., 2005), that is conserved in different species including humans (Popovich et al., 1997; Fleming et al., 2006; Dulin et al., 2013; Blight, 1991; Kigerl et al., 2006).
Historically, immune system activation has been implicated in the pathogenesis of post-traumatic secondary injury, a delayed and progressive form of neurodegeneration that exacerbates cell death beyond the site of primary mechanical trauma (Dusart and Schwab, 1994; Fleming et al., 2006; Dulin et al., 2013; Blight, 1991; Popovich, 2000; Schwab et al., 2006). The cellular and molecular inflammatory cascades induced by SCI include: activation and proliferation of resident microglia and astrocytes; infiltration of circulating innate immune cells including neutrophils, monocytes and lymphocytes; enhanced intraspinal synthesis and release of cytokines, chemokines and other vasoactive substances (e.g., histamine, complement proteins) by neuronal and non-neuronal cell types. The concomitant effects of triggering cell death with enhanced axonal sprouting or other indices of repair (e.g., progenitor cell proliferation/differentiation or revascularization), due to the simultaneous local release of neurotoxic and trophic factors by activated leukocytes, glia and neurons explains in part the paradoxical effects of intraspinal inflammation. The dichotomized perception of “good” and “bad” inflammation is overly simplistic and likely a reflection of our ignorance regarding the functional hierarchy and dynamics of neuro-immune interactions after SCI. Functional deciphering of distinct glia and leukocyte phenotypes and their relationship to changes in immunologic responsiveness throughout the body will offer a more precise understanding of the complexity of this multi-orchestrated response. More detail for each of the above components of intraspinal inflammation and the significance of these events are described throughout other reviews in this special issue and also in previous reviews (Ankeny et al., 2009; Donnelly and Popovich 2008; Popovich and Longbrake, 2008; Lucin and Wyss-Coray, 2009).
A key but poorly understood feature of intraspinal inflammation is the chronicity of this response (Prüss et al., 2011; Schwab et al., 2000). Chronic inflammation, also classified as “nonresolving inflammation”, implies that inflammation is normally controlled by active regulatory mechanisms (Nathan and Ding, 2010; Serhan and Savill, 2005). Failure to resolve inflammation results in aberrant tissue remodeling and organ dysfunction (Serhan and Savill, 2005; Serhan et al., 2007; Nathan and Ding, 2010). Consequently, effective repair and wound healing require coordinated inflammation with timely resolution (Serhan et al., 2007). Besides anti-inflammatory strategies, which aim to reduce leukocyte infiltration or survival at site of inflammation, naturally occurring resolution-promoting agonists, including resolvin E1 and protectin D1, exert similar effects (Schwab et al., 2007). Novel resolution phase promoting-agonists, such as the lipid-derived resolvins, protectins and maresins have been identified to exert potent stereoselective actions with human cells and animal disease models displaying features of incomplete resolution and are referred to as specialized pro-resolving mediators (SPM) and qualify as targets to treat neuropathology associated with a deficiency to resolve from inflammation (Serhan and Savill, 2005; Serhan et al., 2007).
There are no empirical data to show that inflammation resolves after experimental or human SCI. Conversely, an abundance of data indicate that inflammation persists indefinitely within the injury microenvironment. Although inflammatory cells persist at/nearby the injury site for months post-injury, none are as prominent as activated microglia and monocyte-derived macrophages, which persistently synthesize inflammatory enzymes that have been linked with chronic pathology and loss of function (Schwab at, 2000; Dulin et al., 2013). The mechanisms underlying sustained inflammation are not known but several variables must be considered. For example, the oxidative lesion microenvironment may damage leukocyte (and glial) mitochondria, which could impair normal cellular functions including the ability to sustain self-repair or maintenance functions (Lin et al., 2007; Ohri et al., 2011; Sullivan et al., 2005). Phagocytosis by microglia and macrophages is one example of a maintenance function gone awry that could exacerbate the chronicity of inflammation. Cellular and myelin debris are potent inflammatory signals that persist indefinitely at and nearby the injury site because of inefficient or impaired phagocytosis (Magnus et al., 2002; Vallieres et al., 2006; George and Griffin, 1994). A complementary explanation is that within an “immune privileged/-specialized” site like the spinal cord, the threshold of tissue damage caused by blunt trauma, compression or transection injury may exceed the capacity of normal resolution pathways to resolve inflammation (Prüss et al., 2011). This appears to be true even in patients with multiple sclerosis (MS), an inflammatory human CNS disease with comparatively less tissue damage (Prüss et al., 2013). A ‘neurogenic’ explanation for prolonged instraspinal inflammation is exemplified by the appearance of activated microglia, even in spinal regions remote from the injury site where intrinsic regulatory molecules expressed by neurons tonically regulate microglia. As a consequence of SCI and subsequent neurodegeneration, such control is lost and microglia become activated (Cardona et al., 2006; Meuth et al., 2008). Chronic remote activation of microglia has been implicated in the onset and progression of neuropathic pain (Schwab et al., 2005; Detloff et al., 2008; Hains and Waxman, 2006; Zhao et al., 2007). Lastly, metabolic changes could precipitate chronic inflammation. Just as obesity and associated metabolic syndromes contribute to atherosclerosis, another form of chronic inflammation (Nathan and Ding, 2010), a cardio-metabolic syndrome develops after SCI and could exacerbate neuroinflammation and neurologic dysfunction (Bigford et al., 2013; Gilbert et al., 2014).
Although protracted recruitment and activation of leukocytes and glia cannot be excluded as a mechanism to explain the “smoldering” chronic inflammatory response in traumatized spinal cord, a resolution deficit is also likely (Fig. 2). When objective, quantitative measures were applied to an experimental model of SCI, the persistence (“dwell time”) of different leukocyte subpopulations was revealed. Cell type-specific resolution intervals (time between maximum cell numbers and the point when they are reduced to 50%) vary from 1.2-1.5 days for neutrophils and T lymphocytes but are significantly increased to 15 or 55 days for B cells or microglia/macrophages, respectively. The absolute dwell time of leukocyte numbers exceeds the resolution interval as a resolution plateau may remain on supra-normal levels long after SCI. This holds in particular true for microglia/macropages and B-lymphocytes resulting in elevated resolution plateaus cell whereas T-lymphocyte and granulocytes get cleared in a much more effective manner from the lesion site (Prüss et al., 2011)(Fig. 2).
Resolution efficacy is measured as the number of cells being cleared from the lesion site. The cell-specific resolution interval Ri can be determined from the curve at the time when cell numbers decreased by 50% (data points represents the mean of N = 5). Several leukocyte subsets persist indefinitely implying that a ‘resolution-deficit’ exists after SCI. T-cell infiltrates plateau early with ~10% of maximal cell numbers remaining at chronic post-injury intervals. Macrophage and B-cell numbers decrease more slowly. Indeed, even several weeks post-injury, ~45% of maximal macrophage numbers persist (as can be predicted with reasonable certainty from nonlinear regression analysis) after SCI exemplified after a thoracic-level 8 SCI-Model in Lewis rats.
Residual activated microglia/macrophages are likely to undermine blood-spine-barrier closure after SCI thereby thwarting reconstitution of a critical feature of CNS immune privilege (Schwab at, 2000; Popovich et al., 1996b; Dulin et al., 2013). These distinct resolution intervals also suggest that a different repertoire of clearance receptors exists on these different leukocyte subpopulations. Thus, the aim of reducing dwell time of inflammatory cells at the CNS lesion or in peripheral tissues (see Anthony et al., this issue) is an alternative to the conventional anti-inflammatory strategies, which solely aim to reduce infiltrating leukocyte cell numbers at the lesion site (Prüss et al., 2011). Chronic inflammation is associated with impaired tissue remodeling (Serhan and Savill, 2005. Indeed, there is no context in which chronic inflammation is beneficial. Instead, it is associated with excessive parenchyma destruction and functio laesa Serhan et al., 2007).
Impaired resolution of inflammation leads to an extended presentation of self-epitopes, which are usually shielded from overt immune attacks by a blood brain barrier (BBB). The immunologically silent, non-phlogistic clearance of apoptotic cells is crucial for maintaining tolerance to self-antigens (Uederhardt et al., 2012; Savill et al., 2002). Uncleared apoptotic cells eventually lose membrane integrity and release immunogenic self-antigens. This can elicit autoimmunity. The autoimmune pathogenesis of systemic lupus erythematosus (SLE), in humans and mouse models, has been attributed in parts to failed clearance of apoptotic cells (Munoz et al., 2010; Nagata et al., 2010). Likewise, the inefficient phagocytosis of apoptotic cells by macrophages can impair or delay resolution pathways and the exacerbated inflammation can trigger autoimmunity (Nathan and Ding, 2010). After SCI, autoimmunity against CNS antigens develops (see also Jones et al., this issue) and represents a putative target for minimizing chronic neuronal degeneration and aberrant connectivity. Even though trauma and subsequent injury to the blood-spinal cord barrier (BSB) can be linked to neurologic relapses in Multiple Sclerosis (MS) patients and exacerbated pathology and loss of function in predisposed rodent models (Poser et al., 1994; Jones et al., 2002), this de novo autoimmunity elicited by SCI, i.e., trauma-induced autoimmunity, needs to be distinguished from the defined disease entity of MS, which is characterized by distinct diagnostic criteria with a different etiology. Nevertheless there are also overlapping disease mechanisms.
The non-resolving inflammation that occurs after SCI recapitulates relevant neuropathological aspects of the primary progressive form of MS (PPMS). In both SCI and MS, chronically activated microglia/macrophages are associated with ongoing neurodegeneration. Such a compartmentalized inflammatory reaction in the spinal cord may drive chronic progression of neurodegeneration and demyelination (Lassmann et al., 2007). In the progressive phase of MS, new demyelinating lesions are rare but diffuse atrophy of the gray and white matter and changes in the macroscopically “normal-appearing white matter” (NAWM) become prominent (Lassmann et al., 2007; Rise et al., 2013). Likewise, in human SCI, a greater magnitude of progressive atrophy in exofocal gray and white matter over a period of ~1 year correlates with significantly impaired neurological function (Freund et al., 2013).
III. ‘Crosstalk’/interactions between the nervous and immune systems, and neurogenic regulation of immune system
Functional cross-talk between the nervous and immune systems is achieved by the autonomic nervous system (ANS) and the hypothalamic-pituitary-adrenal (HPA) axis (Elenkov, et al., 2000). The ANS is comprised of the sympathetic nervous system (SNS) and the parasympathetic nervous system. These different arms of the ANS counter-regulate each other to control involuntary functions (e.g., heartbeat, digestion, secretion). A convincing body of evidence shows that primary and secondary lymphoid tissues are innervated by the SNS and cells of the immune system bear specific surface receptors for neurotransmitters, hormones and neuropeptides associated with the SNS and HPA axis (Meisel et al., 2005). Similarly, through the release of cytokines and/or chemokines during an active immune response, leukocytes communicate with and regulate CNS function. Cytokine and chemokine receptors are expressed on peripheral sensory nerves and can access cells in brain and spinal cord via specific transport mechanisms or the circumventricular organs (sites where blood-brain barrier is “leaky”). Cytokines are known to affect the survival and function of neurons and glia and can activate both the SNS and the HPA axis through a complex corticotropin-releasing hormone (CRH)-dependent pathway (long feedback loop between the immune system and the brain (hypothalamus) (Meisel et al., 2005).
SIRS and CARS
As a model for understanding the seemingly conflicting dynamics of acute and chronic intraspinal inflammation with delayed systemic immune suppression (see SCI-IDS below), it is useful to consider the pathophysiology of sepsis. Sepsis elicits a systemic inflammatory response syndrome (SIRS) and a compensatory anti-inflammatory response syndrome (CARS). In response to injury or infection, SIRS develops but in parallel CARS is triggered to terminate inflammation (Hotchkiss et al., 2013a,b). A balance in both the intensity and duration of these responses is needed to restore homeostasis within affected tissues throughout the body; an imbalance or overly exuberant response of either SIRS or CARS causes excess inflammation or chronic immune dysfunction, respectively. Neither result is ideal and alone or unregulated, each syndrome can have pathological effects. The dogma that sepsis is the result of an overzealous inflammatory reaction has caused most research and clinical therapies directed toward that disease to focus on limiting inflammation, although with the recent attention focused on the pathologic potential of CARS, therapies to enhance inflammation have started to emerge (Hotchkiss et al., 2013a,b).
Although distinct from sepsis, “sterile” inflammatory lesions in the CNS (e.g., SCI or brain injury) trigger a host response resembling SIRS (see Anthony and Couch, this issue). Autoregulatory mechanisms akin to CARS also develop; however, loss of vegetative input (i.e., SNS) to lymphoid and endocrine organs is unique to SCI and culminates in a condition referred to as spinal cord injury-induced immune depression syndrome (SCI-IDS) (Meisel et al., 2005: Riegger et al., 2007; 2009).
SCI-IDS and infection
SCI effectively ablates normal preganglionic innervation by the SNS to immune organs. This “neurogenic immune ablation” damages the normally well-balanced interplay between the immune system and the CNS culminating in the onset of SCI-IDS. SCI-IDS develops quickly, within 24h, affects cells of both the innate and adaptive immune systems, is independent of the iatrogenic application of high dose corticosteroids and is analogously represented in animal and human SCI justifying its translational relevance and extends into the chronic phase after SCI (Campagnolo et al., 1997; Furlan et al., 2006, Lucin et al., 2007; 2009; Riegger et al., 2007; 2009; Zhang et al., 2013). Similar complications of immunosuppression occur after traumatic brain injury and cerebral infarction (stroke) and are referred to as CNS injury-induced immune deficiency syndrome (CIDS) (Meisel, et al., 2005). Rapid-onset SCI-IDS may develop as a means to confine or prevent pathologic autoimmunity against self-antigens that are liberated or newly expressed by SCI. Conversely, autoimmunity might result from insufficient or weak induction of SCI-IDS („SCI-IDS-escape“), perhaps due to SCI at a lower spinal level or a less severe injury with spared neuro-immune communication (Figs. 1 & 4). SCI-IDS might also decrease the efficacy of protective immunizations (see below; also see Jones et al., this issue).
At the systemic level (upper row), initially after SCI the immune system is suppressed by ‘non-neurogenic’ (CARS/SIRS) and ‘neurogenic’ (SCI-IDS) mechanisms. With time after SCI, the imprint of SCI-IDS increases while the role of CARS/SIRS subsides in case of non-septic SCI-patients. The immune suppression facilitates the development of infections. Infections are the main cause of death in the acute and chronic phase after SCI and have been identified recently as an independent risk factor for poor neurological outcome in SCI patients.
(Lower row). In the injured spinal cord, sustained, non-resolving inflammation occurs and is associated with indices of neurodegeneration, demyelinisation and autoimmunity and neurologic changes associated with maladaptive plasticity (e.g.,pain).
(Middle row). Systemic and localized (intraspinal) immune dysfunction are interdependent. SCI-IDS could be protective by limiting excessive autoimmunity against CNS-epitopes shielded behind the intact blood-spinal cord barrier (BSB). The penetrance of SCI-IDS is sufficient to prevent classical CNS autoimmune disease (e.g., MS) but is not sufficient for eliminating activation of B- or T cell autoimmunity (See also Table 1). Infections generate excessive host and pathogen-specific RNA, DNA and cell membrane detritus – all are immunogenic and act as adjuvants that are able to boost inflammation including the non-resolved inflammatory cascades active in the injured spinal cord. Moreover, conserved microbial structural motifs, referred to as pathogen-associated molecular patterns (PAMPs), are able to bind to cognate receptors grouped as Toll-like receptors (TLRs) (Akira et al., 2006), which are expressed by neuron, astrocytes, microglia, and oligodendrocytes. Infection derived ligand recognition by TLR is a candidate mechanism to directly foster neuronal injury, neurodegneration, demyelinisation, autoimmunity and pain.
Previous reports from us and others indicate that post-SCI immune suppression during both the acute and chronic phases is level-dependent (T3 SCI but not T9 SCI) and is caused in part by dysregulation of the SNS and HPA axis, which results from the loss of descending modulation of spinal sympathetic pre-ganglionic neurons (SPNs) lying caudal to the lesion, as well as the exaggerated spinal circuitry/autonomic reflex triggered by a strong sensory input (Cruse, et al., 1996, Cruse, et al., 1993, Lucin, et al., 2007, Lucin, et al., 2009, Zhang, et al., 2013). Moreover, lesion-height dependent atrophy/involution of secondary lymphatic organs has been described (Lucin et al., 2007; Brommer et al., unpublished data).
SPNs are the source of all sympathetic outflow from the thoracic spinal cord (T3-T13) (Elenkov, et al., 2000). Activation of SPNs causes release of norepinephrine (NE) and GCs from post-ganglionic nerve terminals and the adrenal gland (Engeland and Arnhold, 2005). Visceral and somatic afferents activate SPNs, which send cholinergic projections to post-ganglionic neurons that innervate target organs including blood vessels, lymphoid tissues (e.g., spleen) and the adrenal medulla. SPNs also directly innervate the adrenal cortex providing a rapid and direct mechanism for enhancing GC release (Engeland and Arnhold, 2005). Physiological concentrations of GCs and NE ensure normal immune function but recurrent or prolonged activation of glucocorticoid receptors (GRs) or beta-2 adrenergic receptors (β2AR) is immunosuppressive (Nance and Sanders, 2007). Thus, sustained or uncontrolled activation of the SNS causes aberrant accumulation of NE and GCs in the circulation and immune organs and could promote immune suppression.
A decentralized SNS is also involved in the occurrence of autonomic dysreflexia (AD), which was recently identified as a cause/mechanism of secondary immune deficiency after SCI (Zhang, et al., 2013). Autonomic dysreflexia is triggered by diverse stimuli, including distension of bladder or colon, which cause reflex activation of SPNs and exaggerated discharge of sympathetic neurons. As a result, blood pressure rises to exceedingly high levels. In parallel, the increased firing of noradrenergic fibers innervating lymphoid tissues and the adrenal cortex trigger NE and GC release, respectively. Because the pressor response associated with autonomic dysreflexia can cause significant physical and psychological stress, the HPA axis is stimulated in parallel causing a vicious cycle leading to additional GC release.
Most immune cells express adrenergic and GC receptors. β2AR is the predominant type of adrenergic receptor expressed on B and T lymphocytes in the spleen. NE spike causes lethal overstimulation of β2AR on splenic lymphocytes, and circulating GCs synergize with NE to induce apoptotic signaling in these cells (Lucin, et al., 2007, Lucin, et al., 2009). Secondary lymphoid organs (e.g., spleen) are heavily innervated by sympathetic nerve fibers. Aberrant accumulation of CAs and GCs is associated with the global loss of B and T cells and leucopenia in these immune organs, sites where B cells interact with T cells and APCs to mount immune responses against pathogens and where antibodies are produced (Lucin, et al., 2007). Thus, the occurrence of episodic autonomic dysreflexia causes considerable systemic stress culminating in excess release of circulating GCs and intra-splenic NE, which aggravates lymphocyte depletion and corresponding impairments of immune response and antibody production, and finally chronic immune suppression (Zhang, et al., 2013).
Clinically relevant yet widely underappreciated; acute and chronic SCI-IDS also increases susceptibility to infection (Brommer et al., unpublished data). Infection is the leading cause of morbidity and mortality in patients with SCI (DeVivo et al., 1989; Soden et al., 2000). Infectious complications or diseases are prevalent (>50%) and often occur secondary to SCI, and increase in incidence in quadriplegics and high-level paraplegics leading to a mortality rate of ~4-17% (Failli, et al., 2012). Recent large neuroepidemiological studies with more than 1400 enrolled patients demonstrate that post-SCI infections are independent risk factors for significantly poorer neurological recovery (Failli, et al., 2012; Meisel, et al., 2005). Thus, infections qualify as independent neurobiological rehabilitation confounders that impair the endogenous recovery potential after SCI (Failli et al., 2012; Dietz and Fouad, 2013).
The underlying neurobiology of impaired neurological recovery is under investigation. Among others, exacerbated neurodegeneration (Cunningham et al., 2005) and axonal injury (Moreno et al., 2001) but also ‘sickness behavior’ (Teeling et al., 2010) were reported after systemic endotoxin challenge as a model of systemic infections (Perry, 2003). From a teleological view SCI-IDS could be understood as a conserved mechanism to prevent the development of autoimmunity against CNS specific epitopes, which become assessable by the innate and apdotive immune arm after SCI due to a defective BSB (Fig. 4). We hypothesize that in patients with a low SCI-IDS penetrance a higher riks for robust CNS-autoimmunity will be observed (Schwab et al., 2006). The pathoimmunological link between severity of SCI-IDS and developing autoimmunity will be investigated by the ‘SCIentinel Trail’ recruiting at present (Kopp et al., 2013).
IV. Autoimmune responses after SCI
Emerging clinical and experimental data indicate that SCI elicits systemic autoimmunity (Table 1). Autoimmunity develops when cells of the immune system recognize and mount an immune response against “self” antigens which include proteins, carbohydrates, lipids and nucleic acids that are normally found throughout the body. Autoimmunity can be viewed as a continuum of autoinflammatory (self-directed tissue inflammation dominated by innate immune system) and autoimmune reactions (loss of immunological tolerance with activation of self-directed adaptive immune responses) reactions (McGonagle and McDermott, 2006). The former involves the coordinated activation of inflammasomes and uncontrolled cytokine synthesis with robust activation of myeloid cells (e.g., neutrophils, monocytes/macrophages; also see Kigerl et al., this issue) and the latter activation of self-reactive lymphocytes and enhanced synthesis of autoantibodies (Hayes et al., 2002; Davies et al., 2007; also see Jones, this issue). Each has been implicated in progression of secondary injury after SCI (Ankeny and Popovich, 2010; Zhang and Popovich, 2011). The combination of primary and secondary injury after SCI causes extensive tissue damage to the spinal cord and blood-spinal cord barrier. Subsequently, CNS self-antigens (e.g., myelin, phospholipids, structural proteins) are released into the circulation and drain into secondary lymphoid organs where neuroantigen specific T and B lymphocytes become activated with a subpopulation of lymphocytes acquiring the ability to cross the intact blood-spinal cord barrier to form infiltrating clusters in Virchow-Robin spaces. These perivascular cell accumulations resemble those seen in the EAE model of MS (Popovich et al., 1996a) (Fig. 3a). Clinical signs of disease, such as flaccid tail paralysis, also were observed in naïve animals receiving lymphocytes from littermates suffering acute SCI (Popovich et al., 1996a). Despite the encephalitogenic potential of SCI-activated lymphocytes, protective T cells also are activated after SCI and a subset of endogenous autoreactive cells may exert ‘repairing’ effetcs the injured CNS (Schwartz and Kipnis, 2005; also see Jones, this issue).
A. T-lymphocytes are enceophalitogenic after acute SCI. Pioneering reports were able to unravel that adoptive T cell transfer after SCI are encephalitogenic and able to breach in between the intact endothelium and cross the blood spinal cord barrier (BSB) to form classical infiltration clusters accumulating in the perivascular virchow-robin spaces as entry routes for invading T-cells (Popovich et al., 1996a). Permission for reprint will be requested. B. B-lymphocytes form aberrant clusters and align along ectopic ‘follicle-alike’ structures in adjacent areas of the SCI lesion site and represent neuroimmunological hallmarks of a non-selflimiting inflammatory response after SCI. Remaining B-cell parenchymal clusters constitute a source of (auto-)antibody generation implied in degenerative complement mediated axonal damage but also autoimmune pathophysiology. Oligoclonal band synthesis has been reported after experimental SCI and in human CNS injury (A & B, 50 μm).
Table 1
Hallmarks and selected functionality/antigen specificity of natural autoimmunity and after spinal cord injury (SCI) ‘Naturally’ occurring autoimmunity and tissue maintenance
| Model | Serum/CSF | Functionality/proposed role | ||
|---|---|---|---|---|
| T cells | ||||
| Characteristics (antigen...) | ||||
| CD4, CD25, FoxP3 Tregs | Mouse | Neuroprotection and tissue homeostasis (for review see Schwartz and Kipnis, 2005; Beers et al., 2011) | ||
| CD8 > CD4 | Rat | Immune surveillance (Bradl et al., 2005), preferentially grey matter | ||
| B cells | ||||
| Characteristics (Ig-class, antigen...) | ||||
| IgM | ||||
| β3-tubulin-reactive? | Mouse | recomb | propagated neurite outgrowth/stabilization (Xu et al.,2011) | |
| Mouse | S & CSF | propagated myelin clearance and regeneration (after PNS-injury) (Vargas et al., 2010) | ||
| Ganglioside-reactive Myelin-reactive | Mouse/Human | Remyelinisation, Neuroprotection, neurite outgrowth (for review see Wright et al., 2009) | ||
| Neuronal membrane reactive | ||||
| Regulatory B cells, MOG reactive (B10cells) | Mouse | Tissue homeostasis (e.g. Matsushita et al., 2010) | ||
| IVIG | ||||
| IgG unspecific, derived from healthy donors | Rat | Human | Immunomodulatory effect in human autoimmune disease (Gold et al., 2007) Neuroprotective in xenogenous application after compression SCI in rats (Nguyen et al., 2012). | |
| Autoimmunity after SCI – Trauma-induced autoimmunity as a maladaptive, dysfunctional immune response and driver of neuropathology or suboptimal CNS repair | ||||
|---|---|---|---|---|
| T cells | ||||
| Model | Serum/CSF | Detection, Functionality/Role | ||
| Characteristic | ||||
| MBP81-99 and MBP151-169-reactive | human | S | n.d. | chronic SCI (Kil et al., 1999) |
| MBP-reactive | human | S | n.d. | chronic SCI, lesion severity dependent (Zajarias-Fainsod et al., 2012) |
| MBP-reactive | rat | S | n.d. | acute SCI – encephalitogenic (Popovich et al., 1996a) forming perivascular infiltrations (see also Fig. 3c) |
| MBP-reactive (MBP-TCR Tg) | mice | n.d. | increased secondary damage/demyelinisation impaired functional recovery (Jones et al., 2005) | |
| Functional assay | ||||
| EAE-like symptoms | rat | S | n.d. | Adoptive T cell transfer after acute SCI leads to EAE-like disease with impaired locomotion and elevated EAE disease score (Popovich et al., 1996a) |
| B cells | ||||
|---|---|---|---|---|
| Model | Serum/CSF | Detection | ||
| Characteristic (Class, antigen) | ||||
| IGM | Mice | S | n.d. | Acute SCI (form day 7 post SCI on) ELISA |
| IgG2a | Mice | S | n.d. | Subacute SCI (from day 14 post SCI on) ELISA |
| anti-DNA (ANA) | Mice | S | n.d. | ‘Chronic’ SCI (6 weeks) |
| anti-RNA | Mice | S | n.d. | ‘Chronic’ SCI (6weeks) |
| anti-CNS | Mice | Sf | n.d. | ‘Chronic’ SCI (6weeks) B-cells form ectopic parenchymal intraspinal lymphoid follicles juxtaposed to the lesion site (Ankeny et al., 2006, 2009, Kopp and Schwab, unpublished, see also Figure 3a). |
| anti-galactocerebroside (GalC) | Human | S | n.d. | Chronic SCI (Taranova et al., 1992) |
| anti-GM1 (IgG) | Human | S | n.d. | chronic SCI; (Hayes et al., 2002) associated with lesion severity (only in AIS A-C) |
| anti-GM1 (IgG) | Human | S | n.d. | associated with lesion severity, significantly enhanced in patients with infections such as UTI (Davies et al., 2007) |
| anti-MBP (IgG) | Human | S | n.d. | chronic SCI, lesion severity dependent (Zajarias-Fainsod et al., 2012) |
| Functional assay | ||||
| Reduced neurite outgrowth | Human | S | n.d. | Post-SCI sera on xenogen. neurons (Mizrachi et al., 1983) |
| Neurodegeneration | Mice | S | n.d. | Post-SCI sera injections in healthy brain (hippocampus) (Ankeny et al., 2009) |
| Impaired neurofunctional | ko-Mice | S | n.d. | Significantly improved hindlimb locomotion after SCI Neurological regeneration (Ankeny et al., 2009). |
| B and T cells | ||||
|---|---|---|---|---|
| Model | Detection, Functionality/Role | |||
| Lacking functional | Mice | RAG2−/− | Reduced demyelinisation (Wu et al., 2012). | |
| T and B lymphocytes | Significantly improved hindlimb locomotion after SCI (Wu et al., 2012). |
In classical CNS inflammatory diseases (e.g., MS), ectopic lymphoid-like follicles form adjacent to large intraparenchymal inflammatory lesions (Aloisi et al., 2007). Activated B and T cells also infiltrate and accumulate within the traumatically injured spinal cord where they form large cellular clusters that are reminiscent of germinal centers with follicles that are normally found in spleen and lymph nodes (Ankeny, et al., 2006; 2009) (Fig. 3b). The pathogenic significance of these clusters is not clear; however, if they are functionally analogous to those that form in MS, they may contribute to chronic gray and white matter pathology. Recent data show that in cases of MS, ectopic CNS B-cell follicles co-localized with a corresponding gradient of neuronal loss (Magliozzi et al., 2010).
Activated B and T cells secrete inflammatory cytokines and other co-stimulatory factors. Moreover, the formation of antibody-antigen complexes causes tissue damage through recruitment and activation of complement and Fc receptor-bearing cells (e.g. macrophages). In SCI mice, antibody and complement 1q (C1q) decorate neuronal somata in spared tissue and co-localize with regions of axonal injury and demyelination (Ankeny, et al., 2009). Also, axon loss, demyelination and inflammatory cell accumulation are reduced and functional recovery is improved in SCI mice that are deficient in complement C3 (Qiao, et al., 2006). When injected into intact spinal cord, antibodies purified from SCI mice cause pathology and neurological dysfunction; however, the destructive effects of SCI autoantibodies are attenuated when injected into mice deficient in complement C3 or Fc receptor γ chain (Ankeny, et al., 2009). Together, these data provide a basis for a deeper understanding why recovery from SCI is improved and neuropathology is reduced in mice lacking B cells (Ankeny, et al., 2009). SCI also triggers autoimmune T and B cell responses in humans but the pathogenic significance of these responses remains uncertain (Kil et al., 1999; Saltzman et al., 2013). Autoimmunity after SCI is considered a dynamic process, which is likely to be a function of the extent of SCI-IDS (Meisel et al., 2005; Kopp et al., 2013) and consecutive infections (Enouz et al. 2012) supported by clinical observations demonstrating elevated auto-antibody titers in patients with concomitant infections (Davis et al., 2007) (Fig. 4).
V. Possible reversal of immune dysregulation induced by SCI
Drugs exist that could be used to overcome immune dysfunction or trauma-induced autoimmunity after SCI. The onset of immunosuppression after SCI is primarily caused by unregulated activation of SPNs, which causes aberrant accumulation of NE and GCs, leading to prolonged activation of adrenergic receptors (e.g., β2ARs) and GRs expressed on immune cells. In mice, selective β2AR (butoxamine) and GR antagonists (RU486) can significantly reduce leukocyte killing and restore immune function after both acute and chronic high-level SCI (Lucin, et al., 2007, Zhang, et al., 2013). Other regimens of adrenoreceptor blockade have proven to be similarly effective in potentiating immune responses after stroke and could represent therapeutic targets for reversing CIDS (Ajmo, et al., 2009, Prass, et al., 2003). In addition, since autonomic dysreflexia was recently detected as an underlying mechanism of immune suppression after SCI, any strategy that can prevent AD should also alleviate SCI-IDS and consequently, reduce mortality and improve neurological outcome. For example, gabapentin, a GABA (Gamma-Aminobutyric acid) analogue used clinically to treat neuropathic pain, has also been shown in experimental models to significantly reduce tail spasticity and autonomic dysreflexia after severe SCI (Rabchevsky, et al., 2012).
Most experimental data indicate that SCI autoantibodies can cause tissue damage; therefore, strategies to remove or deplete autoantibodies or suppress B cell activation could be neuroprotective (Martin and Chan, 2006). Although some B cell depleting monoclonal antibodies (e.g., Rituxan®, Ocrelizumab®) fusion proteins or biologicals that regulate B-cell survival factors (e.g., BAFF and APRIL) have or are being tested as therapies for classical forms of autoimmune disease (Mackay, et al., 2003), similar strategies have not been tested in stroke, CNS injury or neurodegenerative disease.
The detection and identification of autoantibodies also could have diagnostic and therapeutic value for SCI. The presence of specific antibodies in patients offers an opportunity for early diagnosis of immune dysfunction after SCI. In the future, it may be possible to stratify the severity of injuries in SCI populations based on the magnitude or specificity of autoantibody profiles. Such an approach could be useful for predicting responders and non-responders for clinical trial design and perhaps for devising uniquely tailored immune-modulatory treatments. Indeed, a preponderance of autoantibodies of a specific isotype or antigen specificity may help predict changes in specific neurological functions and/or systemic pathologies.
Acknowledgment
This work was supported by the National Institutes of Health (Grant #R21 NS067260), the Ray W. Poppleton Endowment (to PGP) and the Else-Kröner-Fresenius Foundation, Wings for Life Spinal Cord Research Foundation (Grant Number WfL-DE-006/1) (to JMS). The Department of Clinical and Experimental Spinal Cord Injury Research, Department of Experimental Neurology, Charité, is an associated member of the European Multi-Center Study on Spinal Cord Injury (EMSCI, www.emsci.org).
Abbreviations
| AD | Autonomic dysreflexia |
| AIS | ASIA impairment scale |
| ANS | Autonomous nervous system |
| ASIA-score | American Spinal Injury Association (has been revised and referred to as INSCSCI = International Standards for neurological classification of spinal cord injury) |
| BSB | Blood spinal cord barrier |
| BDNF | Brain derived neurotrophic factor |
| CARS | counter regulatory anti-inflammatory syndrome |
| CIDS | CNS injury-induced immune depression syndrome |
| GCs | Glucocorticoids |
| HPA-axis | Hypothalamus-Pituitary-Adrenal axis |
| IVIG | Intravenous immunoglobulins |
| NE | Norepinephrine |
| NT-3 | Neurotrophin-3 |
| SCI-IDS | Spinal cord injury-induced immune depression syndrome |
| SIRS | Systemic inflammatory response syndrome |
| SNS | Sympathetic nervous system |
| SPN | Sympathetic pre-ganglionic neuron |
Footnotes
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.




