CXCL12 in brief

CXCL12 or stromal cell derived factor-1 (SDF-1) is a 68-amino-acid CXC chemokine with important roles in essential biological processes such as vascular and neuronal development and hematopoiesis. The response to CXCL12 occurs at a very early stage of embryonic development and appears to be widely operative whenever cell migration is required [1]. Indeed, mice lacking CXCL12 die prenatally and exhibit defects in vascularization, neuronal development, and hematopoiesis [2–5]. CXCL12 regulates the bone marrow homing and egress of stem and endothelial progenitor cells and their migration into peripheral tissues in both steady-state conditions and injury [6, 7]. Besides these physiological functions, CXCL12 seems to be involved in pathological processes such as neoplasia, tumor progression, and chronic inflammation [8–14]. In immune system, the principal role of this chemokine is to regulate the trafficking and localization of myeloid, lymphoid, and progenitor cells between central and peripheral compartments [15, 16].

CXCL12 is constitutively expressed in a broad range of tissues [17, 18], and the major sources of CXCL12 expression are bone marrow stromal elements and endothelial cells [19, 20]. It has been shown that the effects of CXCL12, such as mobilization of leukocytes from the bone marrow and transendothelial migration of inflammatory cells, are mainly dependent on the interaction of the chemokine with CXC chemokine receptor 4—CXCR4 [19–22]. CXCR4 has been considered as the unique receptor for CXCL12 and as the only mediator of its biological effects for many years. However, recent studies have found that CXCL12 binds not only to CXCR4 but also to CXCR7 (RDC1) [23]. While CXCL12 is exclusive ligand for CXCR4, CXCR7 binds both CXCL12 and CXCL11. This recently discovered receptor for CXCL12 is phylogenetically closely related to chemokine receptors but fails to induce typical chemokine receptor–mediated intracellular responses. Actually, CXCR7 functions as a decoy for CXCL12, thus regulating signaling of CXCL12 through CXCR4 [24, 25]. It seems that crucial activity of CXCR7 is to locally scavenge CXCL12 and thereby form a CXCL12 gradient for CXCR4-dependent migration [24]. However, there are other reports showing that CXCR7 actions are more than scavenging of CXCL12 and modulation of CXCR4 signaling. For instance, CXCR7 is critical for CXCL12-mediated survival of renal progenitor cells [5]. It is also essential for CXCL12-induced mitogen-induced protein kinases signaling and proliferation of astrocytes [26]. Finally, CXCR7-mediated signaling is involved in tumor cell growth, survival, and metastasis [27]. Similar to their ligand, CXCR4 and CXCR7 display a wide expression pattern in mammalian tissues. They are coexpressed on the surface of T- and B-cell subsets, monocytes, endothelial cells, dendritic cells, neurons, and in some tumor cells, primary human tumors, and tumor-associated endothelial cells [23, 28–32].

CXCL12 and inflammation

Much attention has been focused on the role of chemokines in inflammatory processes and especially, inflammatory autoimmune diseases. As chemoattractants of inflammatory leukocytes, chemokines are essential for leukocyte homing and the recruitment during inflammation, thereby directing the development of inflammatory responses, including inflammatory autoimmune diseases. CXCL12 is essential for attracting activated, CXCR4+ T cells to the sites of inflammation, and it also acts as a costimulator in T cell activation [33]. Moreover, it promotes the adhesion of T cells to ICAM-1 and VCAM-1 by upregulating the binding activity of LFA-1 and VLA-4 on T cells, respectively, and modulates the α4β7 integrin-mediated lymphocyte adhesion to mucosal addressin cell adhesion molecule-1 (MAdcam1) and fibronectin [34–36]. CXCL12 enhances the inflammatory infiltration of lymphocytes in diverse models and settings involving acute inflammation or fulminant infection [37]. In addition, upregulation of CXCR4 or CXCL12 has been reported in many inflammatory diseases such as rheumatoid arthritis (RA), multiple sclerosis (MS), inflammatory bowel disease (IBD), uveitis, nephritis, and lupus erythematosus [38–41]. Therefore, it has been suggested that CXCL12 plays a proinflammatory role in various autoimmune diseases and that it could be a valid target for neutralization in these diseases.

One of the basic agents in the research of CXCL12 functions is an antagonist of CXCR4, AMD3100. The rationale for using a chemokine receptor antagonist in inflammatory diseases is to selectively alter the recruitment of particular subsets of cells that initiate and maintain the inflammatory responses. AMD3100, a bicyclam molecule, has been identified as a specific inhibitor of CXCR4 [42, 43]. As CXCR4 serves as a coreceptor for human immunodeficiency virus (HIV) type 1 to infect T cells [44], it was shown that AMD3100 effectively and specifically blocked HIV entry into CXCR4-expressing cells [45]. In more recent studies, AMD3100 was shown to specifically inhibit CXCL12-mediated responses and to have beneficial effects in various animal models of inflammatory diseases, including asthma [46], RA [47], uveitis [48], and diabetes [49]. AMD3100 attenuated allergic lung inflammation and airway hyperreactivity in mice [46]. Treatment of allergic mice with AMD3100 significantly reduced number of pathological parameters related to asthmatic-type inflammation such as airway hyperreactivity, peribronchial eosinophilia, and the overall inflammatory responses. In addition, there was a shift from the pathogenic T helper (Th)-2 cytokine profile to the counteracting Th1 profile in the AMD3100-treated animals. Specifically, there was a significant reduction in interleukin (IL)-4 and IL-5 and a marked increase in IL-12 and interferon (IFN)-γ levels in the lungs of treated allergic mice [46]. CXCL12 expression has been demonstrated in rheumatoid synovium, and increased levels of this chemokine are detected in RA synovial fluid [50–52]. RA synovial tissues had higher levels of CXCL12 on high endothelial venules, where this chemokine was colocalized with heparan sulfate proteoglycans [53]. Importantly, interaction of CXCL12 with glycosaminoglycans is supposed to be important for leukocyte diapedesis and neovascularization during inflammation [53]. AMD3100 was also shown to be efficient in RA model, autoimmune collagen-induced arthritis (CIA) as it was demonstrated that exogenous CXCL12 injected in periarthritic tissue of mice elicited an inflammatory response that could be inhibited by AMD3100 [47]. Further, intraperitoneal injection of AMD3100 attenuated ovalbumine (OVA)-induced uveitis, and it significantly ameliorated ocular inflammation in vivo, thus suggesting that CXCL12 has pathogenic role in this inflammatory ocular disease characterized by the infiltration of T lymphocytes and other leukocytes into the eye [48]. Further, AMD3100 exerts therapeutic effects on an IBD model, experimental colitis. There, it inhibited colonic inflammation, decreased epithelial apoptosis and gut permeability, and consequently enhanced the epithelial barrier integrity [54]. In non-obese diabetic (NOD) mice, which are predisposed to develop type 1 diabetes (T1D) and serve as a model for studying diabetes in humans, AMD3100 mobilized T cells from the bone marrow to peripheral lymphoid tissues and simultaneously inhibited disease development [55]. It was suggested that elevated CXCL12 expression promotes T1D in NOD mice by altering T-cell and hematopoietic stem cell trafficking. On the other hand, there are findings that CXCL12 exerts beneficial regulatory functions in the NOD mouse model as it protects NOD mice from autoimmune diabetes. It was reported that when T splenocytes from NOD mice treated with AMD3100 were mixed with diabetogenic T cells during adoptive cell cotransfer experiments, prevalence of diabetes in the recipients significantly rose [49]. Also, AMD3100 reduced the number of CXCR4+ and CXCL12+ cells in the inflamed islets in mice, while CXCL12 attenuated diabetes and promoted pancreatic β-cell survival by the activation of the prosurvival kinase Akt [56]. Thus, AMD3100 application has helped us to understand that besides proinflammatory role in experimental models and inflammatory autoimmune diseases, CXCL12 has also anti-inflammatory properties. This observation challenges the concept of strict pathogenic role of CXCL12 in chronic inflammatory and autoimmune disorders. Such an assumption is further substantiated with data obtained in MS and its animal models, which are presented in the following paragraphs.

Experimental autoimmune encephalomyelitis (EAE) and multiple sclerosis (MS)

Experimental autoimmune encephalomyelitis (EAE) is an autoimmune disease of the central nervous system (CNS). EAE can be induced in susceptible experimental animals, including rodents and primates, by immunization with CNS-specific antigens, such as myelin basic protein (MBP), myelin oligodendrocyte glycoprotein (MOG), proteolipid protein (PLP), CNS tissue homogenate, etc. (active EAE). Alternatively, the disease can be evoked by the transfer of encephalitogenic CD4+ T cells, obtained from draining lymph nodes of animals immunized for active EAE induction, into syngeneic animals (transfer or passive EAE) [57]. Current concepts suggest two steps in the autoimmune pathogenesis of active EAE. First, dendritic cells (DC) located at the site of injection ingest the inoculated CNS antigens and migrate into the lymph nodes draining the sites of immunization. CNS antigens are also transported via the soluble routes into the lymph nodes where they are ingested by resident DCs. Within the lymph nodes, DCs present the antigenic peptides in the context of major histocompatibility complex (MHC) class II molecules to naive autoreactive CD4+ T cells. These cells then differentiate into IFN-γ- and/or IL-17-producing effector T cells, namely T helper (Th) cells, and even more specifically Th1 and Th17 cells, respectively [58]. Second, Th cells migrate from the secondary lymphatic organs into the CNS where they are reactivated after interacting with MHC class II+ CNS resident phagocytes [59, 60]. The stimulated effector T cells locally release proinflammatory cytokines and thus initiate an inflammatory reaction, which may finally result in demyelination and axonal degeneration [61, 62]. Inflammation, demyelination, and axonal loss result in various neurological deficits, usually an ascending progressive paralysis in a caudal to rostral direction starting with tail paralysis and eventually leading to forelimb paralysis and/or a moribund state and death in severe cases.

Since EAE shares many clinical, histological, immunologic, and genetic features with human demyelinating diseases, including MS, it has been widely used to gain important insight into MS pathogenesis and to validate new targets for MS therapy [63, 64]. MS is an inflammatory disease of the CNS. The most prominent feature of the disease is demyelination, which is the major cause for neurological symptoms expressed in the patients. Axonal loss and neurodegeneration are becoming increasingly appreciated as pathological hallmarks of MS as well [65]. There are at least four subtypes of MS regarding clinical expression: relapsing–remitting, primary progressive, secondary progressive, and relapsing progressive. Also, there are four defined pathological subtypes of the disease: macrophage-mediated, antibody-mediated, distal oligodendrogliopathy, and primary oligodendrocyte damage with secondary demyelination [66]. We can say that MS is not a single ailment, but rather a common name for a group of diseases that, although different in specific inductive and effector pathogenic mechanisms, converge at the point of inflammation and demyelination. Thus, it is relevant to have variants of the EAE model, which provide information of importance for understanding pathogenic specificities of MS forms. Importantly, depending on the species and strain used in experiments, on encephalitogen and adjuvants applied to these animals, as well as on the way the immunization is performed, there is a divergence in clinical manifestation of EAE, roughly corresponding to the major clinical subtypes of MS [63, 67]. Actually, among numerous strains of mice and rats in which EAE can be induced, every strain or even substrain has specificities regarding immunization protocol that leads to EAE induction. These specificities include myelin proteins and/or peptides, intensity of costimulation with adjuvants, number of injections, etc. Besides that, every strain has some pathogenic and clinical feature that makes it unique to other strains and species. Therefore, investigation of one particular EAE variant will give us useful information regarding some aspects of pathogenesis of one clinical subtype of MS. As EAE is mostly studied in inbred strains of animals, it is possible to consider one animal strain as of one human being. Unlike humans and natural populations of animals where every individual is genetically unique, with an exception of identical twins, all animals of an inbred strain are genetically identical to each other. Of paramount interest are those strains of mice and rats that show exceptional susceptibility to EAE induction, as well as those strains that are resistant to the disease induction. Having in mind correlation with humans, an extremely susceptible strain corresponds to a MS patient and a persistently resistant strain corresponds to a healthy person. Thus, through comparative investigation of such strains, it is possible to identify genes, molecules, mechanisms, etc. that define susceptibility and/or resistance to the autoimmune CNS disorder. In our laboratory, we are investigating such strains of rats, namely Dark Agouti (DA) rats and Albino Oxford (AO) rats.

Our way to CXCL12

AO and DA rats are at the opposite poles of susceptibility to EAE induction. We can claim this as we have shown that AO rats do not develop EAE in conditions under which other relatively resistant rat strains such as Fischer 344, Brown Norway (BN), or PVG rats do [68–71]. Also, we have demonstrated that DA rats develop the clinical disease after immunization with spinal cord homogenate (SCH) even in the absence of any adjuvant [72], which is a rare exception from the rule that EAE induction requires application of an adjuvant. Initial paper of Mostarica Stojkovic et al. [73] in which it was shown that rats of inbred AO strain are highly resistant to EAE induction was published some 30 years ago. This work was the beginning of the series of investigations aiming to elucidate mechanisms of AO rats’ resistance and DA rats’ susceptibility to EAE induction. During the course of the research, various cells and molecules relevant for EAE pathogenesis have been investigated in the peripheral blood, draining lymph node (DLN), and spinal cord (SC) lesions of AO and DA rats after the induction of EAE [74–76]. We have recently shown that upon immunization, the number of cells increased more vigorously in DLN of DA rats than in AO rats, suggesting a higher proliferation rate and/or a more efficient recruitment of immune cells, accompanied by a larger production of proinflammatory cytokines, including IFN-γ, IL-17, IL-6, IL-12, and IL-23 [71, 77]. Accordingly, number of cells infiltrating SC was markedly higher in DA rats than in AO rats at the time of peak of EAE in DA rats [78]. Thus, lower numbers of cells infiltrating the CNS as a consequence of a limited cell proliferation at the periphery could well explain the observed resistance of AO rats toward EAE. However, more than 20 years ago, Sedgwick et al. [79] showed that a small number of encephalitogenic cells were sufficient to induce severe EAE in Lewis rats. Therefore, the relatively low numbers of the cells infiltrating the SC of AO rats might not necessarily be the limiting factor for the induction of the disease. Moreover, histological lesions in the CNS tissue during the early phases of EAE (7–9 days after immunization) were similar in clinically diseased DA and symptom-free AO rats [74]. Further, SC infiltrations of similar degree and distribution were observed in AO and DA recipients of highly encephalitogenic (DA × AO) F1 T cells. Importantly, equal numbers of F1 cells did not induce any clinical symptoms in AO recipients while provoking a severe paralytic disease in F1 hybrids and DA rats [80]. All above mentioned suggests that mere number of cells infiltrating the CNS might not be the crucial factor for encephalitogenicity and that CNS intrinsic regulatory mechanism might contribute to the resistance of AO rats to the induction of EAE.

Therefore, we focused our research on the influence of the CNS on the encephalitogenic cells and CD4+ T cells in particular, as they are well known to represent the major culprits in the pathogenesis of EAE [58, 64]. We observed a substantial accumulation of CD4+ T cells in both rat strains at the peak of the disease in DA rats, but this increase was markedly pronounced in DA rats. Even in the early EAE, there were more CD4+ cells in the mononuclear cell infiltrates of DA rats [74]. Moreover, CD4+ T cells isolated from the SC of AO rats were less activated, and there were more naïve T cells. Also, there was a clear difference in the ratio of naïve and activated CD4+ T cells in the CNS versus peripheral blood between two strains [78]. Further, the expression of IFN-γ and IL-17 was also significantly lower in AO rats when determined in T cells immediately after isolation from the SC. However, if cells infiltrating the SC were removed from the CNS milieu and propagated in vitro, there was no difference in the percentage of cells expressing these cytokines, thus indicating that the CNS tissue of AO rats limits the expression of these cytokines in the infiltrating cells [78]. Finally, we observed that cells infiltrating SC of AO rats are more prone to apoptosis, which confirmed previous findings in passively transferred encephalitogenic CD4+ T cells in AO and DA rats [81]. These results further corroborated the assumption that factors in the CNS contribute to the limited autoimmune response and therefore to the resistance of AO rats to EAE induction.

CNS tissue may influence composition and activation state of the infiltrates by regulation of cell egression from circulation at blood–brain barrier (BBB) and by regulatory mechanisms working within the CNS parenchyma. Entrance of immune cells into the CNS is highly regulated at the level of BBB through various interactions of infiltrating cells and CNS-resident cells, including interactions between chemokines and their receptors [82]. While investigating gene expression of various neuroinflammation-related molecules in SC homogenates, we observed an interesting phenomenon regarding one of the chemokines, CXCL12. Its expression was remarkably downregulated in DA rats at the peak of EAE and strongly upregulated in AO rats analyzed at the same time point [78]. This result implied that CXCL12 might be of importance for the regulation of CNS inflammation or at least that this molecule’s abrogated expression is a marker of EAE pathogenesis.

CXCL12 in neuroinflammation

We have been aware of the potential importance of the interaction of CXCL12 and its receptor CXCR4 for neuroinflammation, as it has been previously shown that CXCL12 could be a key chemokine regulating the entrance of lymphocytes into the CNS [82]. Also, CXCL12 expression was detected in MS patients, both in cerebrospinal fluid and in MS plaques [83–86]. Therefore, we investigated CXCL12 expression in our system in details [78]. We found out that CXCL12 is expressed in similar amounts in the SC of non-immunized AO and DA rats. After immunization, it was increased in AO rats but decreased in DA at the time of peak of EAE in DA rats. As endothelial cells are among the major producers of CXCL12, we checked the expression of CXCL12 in these cells as well. Expression of CXCL12 in isolated SC microvessels of immunized animals was significantly higher than in total SC homogenates in AO, thus implying that the blood–brain barrier was the major source of CXCL12 expression. This notion was supported by the finding that by day 12–14 post-immunization, there was high CXCL12 expression on blood vessel cells in SC tissue sections of symptom-free AO rats and low, almost undetectable, expression of CXCL12 in the sections of paralyzed DA rats [78]. Importantly, when we examined expression of CXCR4, the principal receptor of CXCL12 in immune cells infiltrating SC, there was no significant difference in the expression of CXCR4 at the peak of EAE in DA rats and immunized AO rats killed at the same time. Thus, we concluded that it was not CXCR4 expression in immune cells that determined the difference in infiltration of CNS in AO and DA rats. Further, we have also detected downregulation of CXCL12 expression in SC of Lew rats at the peak of transfer EAE (unpublished observation), which showed that inhibition of CXCL12 was not limited to DA rats or active EAE, but that it was rather a general phenomenon linked to neuroinflammatory processes. In order to elucidate whether upregulated expression of CXCL12 has a protective role in EAE, we applied an antagonist of CXCL12, AMD3100, to AO rats. As a result, treatment with AMD3100 exacerbated otherwise mild EAE in AO rats [78], thus suggesting that CXCL12 could indeed be a factor that largely contributes to the limitation of encephalitogenicity.

Our results implying that CXCL12 has regulatory and not inflammatory role in neuroinflammation are not solitary. On the contrary, CXCL12 has been reportedly associated with down-regulation of EAE [82, 87, 88]. It was shown that CXCL12 prevented egression of infiltrating cells from perivascular space through the glia limitans into CNS parenchyma [87, 89]. Furthermore, CXCL12 was reported to redirect the polarization of antigen-specific effector Th1 cells into IL-10-producing T cells [88]. Also, it was found to induce CD4+ T-cell apoptosis via upregulation of the Fas–Fas ligand pathway [90]. Interestingly, while in rat EAE it was the level of expression that determined susceptibility to neuroinflammation, in a mouse EAE and in MS patients localization of CXCL12 was the most important parameter (Fig. 1). Namely, CXCL12 is normally expressed on basolateral side of endothelial cells, while in mice suffering from EAE and in active lesions of MS patients, its expression is shifted to luminal side [87, 89]. As suggested by the authors of the reports, it seems that basolateral expression of CXCL12 is necessary to keep infiltrating cells within perivascular cuffs and that the shift of CXCL12 expression to luminal side allows infiltrating cells to penetrate into CNS parenchyma. This change in localization of CXCL12 is induced by inflammatory cells, and it was shown that IL-1β had a major role in the shift [91]. Still, there is also a report which shows that the localization of CXCL12 expression in MS plaques blood vessels is strictly limited to the outer layer of blood vessels, consisted of pericytes and extracellular matrix, and not inner layer made of endothelial cells [84]. Also, there is a study that relates CXCL12 expression in astrocytes with activity of MS lesions [86]. Moreover, a truncated form of CXCL12 (P2G2) that antagonized CXCR4 also inhibited EAE in mice and reduced the accumulation of CD4+ T cells in the CNS [92]. This result seems to be in discrepancy with results of McCandless et al., as well as with our results. The discrepancy might be due to different EAE models used, that is, species and strains (SJL or C57BL/6 mice vs. DA rats) and/or encephalitogenic emulsions (PLP or MOG vs. SCH). This would imply that CXCL12 is not universally protective in EAE, but that its action is specific for some EAE variants and consequently for some variants of MS as well. While this possibility has to be seriously accounted in future research of the role of CXCL12 in neuroinflammation, especially regarding potential therapeutic approach aiming at neutralization of CXCL12, there are still more options for explanation of the diverse functions of CXCL12 in the reports. Namely, P2G2 inhibited the sensitization phase of the immune response and had no effect on EAE if applied to mice that received encephalitogenic cells [92]. Thus, we could say that the role of CXCL12 in propagating EAE is restricted to peripheral development of an immune response, while it does not have such effect within the CNS. The other important difference is the way in which CXCL12 effects were neutralized. While Kohler et al. used truncated form of CXCL12, AMD3100 was used by McCandless et al. and us. AMD3100 is an antagonist of CXCL12 signaling through CXCR4, but not through CXCR7, while P2G2 might affect both receptors. It has recently been reported that CXCR7 actually serves as a scavenger of CXCL12 in neuroinflammation and that it removes CXCL12 from basolateral side of blood vessel endothelial cells and thus allows for leukocyte to entry into the CNS parenchyma [93]. Therefore, if P2G2 affects both CXCR4 and CXCR7, its net effect on CXCL12-dependent route of leukocyte entry into the CNS might be neutral. This would then explain the lack of effect of P2G2 on transfer EAE course. Finally, the fact that CXCL12 level is important for EAE regulation in rats and CXCL12 localization is essential for the regulation of neuroinflammation in mice and humans implies that CXCL12 availability might be a conserved mechanism of inhibition of immune cell infiltration into the CNS, while ways in which this availability is regulated diverged among species in evolution.

Fig. 1
figure 1

Proposed roles of CXCL12 in neuroinflammation. a Basic elements of the CNS that interact with infiltrating immune cells at the level of a postcapillary venule: vessel lumen (VL), endothelial cells (EC), endothelial basal lamina (EBL), Virchow-Robin space (VRS), parenchymal basal lamina (PBL), astrocytic end feet (AEF), parenchyma (P). b CXCL12 (black dots) expressed at basolateral side of endothelial cells of the blood vessels arrests T cells (T) within VRS and prevents infiltrating cells to progress into the CNS parenchyma. Delocalization of CXCL12 from the basolateral side, induced by IL-1β and mediated by CXCR7, allows for the infiltration into the CNS parenchyma. Alternatively, high expression of CXCL12 in EC prevents infiltration of immune cells into the CNS. Reduction in CXCL12 expression under the influence of NO allows the infiltration of T cells. Also, CXCL12 provokes apoptosis in T cells and induces generation of regulatory T cells (Treg)

Full size image

Besides endothelial cells, astrocytes are important producers of CXCL12 in the CNS. Actually, they were identified as the most potent source of CXCL12 in active MS lesions [86, 94, 95]. CXCL12 expression in astrocytes was induced by neuroinflammation-related molecules IL-1β and MBP in vitro [86]. Astrocytes are considered to be major players in neuroinflammation, where their role might be both pro- and anti-inflammatory [96]. Therefore, their capability to express CXCL12 both at the area of glia limitans and within the CNS parenchyma is important for understanding EAE and MS pathogenesis and should be addressed in detail in future studies. Particularly interesting is the report that astrocytes attract immature DCs into the CNS through engagement of CXCL12/CXCR4 axis [94]. Immature DCs have been supposed to control inflammatory environment in MS if stimulated with IL-10 [97]. As stated previously, CXCL12 shifts Th1 cells toward IL-10-producing regulatory T cells [88]. Thus, we can speculate that high CXCL12 generation within the CNS will simultaneously induce IL-10 generation and immigration of immature DCs and that this will provide favorable conditions for limitation of neuroinflammation.

CXCL12 is also expressed in the healthy CNS where it serves as an important factor for survival and migration of neuronal and oligodendrocyte precursors [98, 99]. Thus, it is rational to assume that this chemokine might also contribute to the process of remyelination and neuroprotection. Indeed, CXCL12 expression on astrocytes was detected in cuprizone model of demyelination in mice, while CXCR4 expression on oligodendrocyte precursor cells (OPC) was shown essential for OPC migration and remyelination [100]. Similar results were obtained in virus-mediated demyelinating disease provoked by infection of the mouse CNS with the neurotropic JHM strain of mouse hepatitis virus (JHMV). Neural stem cells (NSCs) were introduced into spinal cords of JHMV-infected mice with established demyelination. This engraftment resulted in migration, proliferation, and differentiation of the cells into OPCs and mature oligodendrocytes and subsequent axonal remyelination [101]. However, if the NSC transplanted animals were treated with anti-CXCL12 blocking serum or with an antagonist of CXCR4, but not CXCR7, migration and proliferation of the engrafted stem cells were markedly impaired [101]. Finally, it was shown that CXCL12 is important for migration of human mesenchymal stem cells (MSC) in vitro [102, 103]. MSC have increasingly been appreciated as candidates for MS therapy, as they were shown to ameliorate clinical course and decrease demyelination, immune infiltrates, and axonal loss in EAE, and they were reported to be safe for application in humans [104].

It is clear that the observed differences in CXCL12 expression emerge during neuroinflammation, and it seems plausible to assume that immune factors, possibly extrinsic to the CNS, might influence CXCL12 expression. Cytokines previously related to the resistance/susceptibility of rat and mice strains to EAE induction, such as TNF, IL-4, TGF-β, and IL-10/IL-12 [105–108], must be taken into consideration. Precise dissection of the regulation of CXCL12 expression in AO and DA rats, as well as of the influence of CXCL12 on encephalitogenic cells is the focus of ongoing work in our laboratory. We are particularly interested in the influence of NO on CXCL12 expression. NO and its secondary metabolite peroxynitrite have been reportedly associated with inflammatory demyelination and neurodegeneration [109–111]. The overproduction of NO is related to the generation of peroxynitrite and consequently to extensive tissue damage in the CNS autoimmunity [109, 112, 113]. Although we showed that there was no disparity in NO generation in DLN of AO and DA rats [71] and that NO does not contribute to the resistance of AO rats to EAE induction, we also observed clear difference in inducible NO synthase (iNOS) gene expression within the CNS in AO and DA rats. Namely, opposite to CXCL12, iNOS gene expression was high in DA rats and low in AO rats SC during EAE course, and we observed negative correlation in the expression of iNOS gene and CXCL12 gene in SC of immunized rats (unpublished observation). Our preliminary data show that in vitro NO inhibits CXCL12 expression, both in astrocytes and in micro-blood vessels isolated from the CNS. Therefore, we are currently investigating influence of NO on CXCL12 expression and possibility that NO determines susceptibility of DA rats to EAE.

Concluding remarks

Taken together, the observed pattern of expression of CXCL12 in the SC of AO, DA, and Lewis rats along with data from the studies in mice and humans suggests that CXCL12 expression at the blood–brain barrier could be an important component in the net effect of the target tissue on encephalitogenic cells. Moreover, it appears that CXCL12 may largely contribute to the resistance of animals to EAE induction, and it is tempting to speculate that it might be protective in MS pathogenesis as well. Ongoing and future research should provide details about the regulation of CXCL12 expression during neuroinflammation.