MS progression and neurodegeneration highly depend on the severity of immune system reactions, rate of remyelination versus demyelination, and scar formation. As discussed earlier, cytokines and chemokines, such as TNF-α, IFN-γ, and TGF-β have an important role in the progression of these neurodegenerative processes. Nicoletti et al. indicated increased serum levels of TGF-β1 in both progressive and relapsing-remitting forms of MS patients, and those suffering from relapses even manifested elevated levels of serum TGF-β1 [49]. On the other hand, results from another study on patients with exacerbating MS indicated that compared to the control group, they had lower serum levels of TGF-β1. Nevertheless, TGF-β1 was increased in CSF samples of exacerbating patients [50]. Thus, the cytokine profile of native and infiltrated cells in MS patients’ CNS may explain more details about what is happening. According to the mentioned studies, we here tacle around the role of different isoforms of TGF-β in the immunopathogenesis of MS.
3.1. TGF-β and chronic inflammation in MS
As discussed earlier, chronic inflammation has a crucial role in MS pathogenesis. TGF-β is a cytokine well-known for its immunosuppressive effects. Among the three isoforms of TGF-β, the immunologic effects are predominantly orchestrated by the type 1 isoform, while TGF-β type 2 and 3 mostly drive differentiation and development of target tissue [51]. Considering it as a double-faced cytokine, All isoforms of TGF-β drive the differentiation of naïve T cells to T regs forkhead box P3 (FOXP3)+ T cells [52–54]. Also, all TGF-β1 and 3 has shown to inhibit proliferation, differentiation, and antibody production of human B cells [55]. Despite, It has been shown that in the existence of interleukin-6 (IL-6), IL-6 and TGF-1 synergistically proceed with the formation of pro-inflammatory Th17 cells [52]. Besides, all isoforms of TGF-β can also modulate Th17 organization through the enhancement of signal transducer and activator of transcription 3 (STAT3) activation [56, 57], which is essential for expressing retinoid-related orphan receptor gamma t (RORγt) and RORα. These two proteins are particular transcription factors for Th17 cells [58].
Moreover, TGF-β1 has been introduced as a mighty chemoattractant for inflammatory monocytes, neutrophils, and T lymphocytes by upregulating different chemokines and their receptors [59–61]. TGF-β1 can up-regulate chemokine receptors, such as C-C motif chemokine receptor 7 (CCR7) to stimulate the migration of naïve T cells toward secondary lymphatic organs and their maturation. Expression of CCR6 in Th17 cells and its ligand C-C motif chemokine ligand 20 (CCL20) in epithelial cells of choroid plexus, which are essential for Th17 and T-reg migration toward the inflammation site, are also upregulated by TGF-β1 [62]. Moreover, a study from Lou et al. indicated that inhibition of TGF-β1 attenuated the Th17 cells’ accumulation in the CNS of myelin oligodendrocyte glycoprotein (MOG) immunized mice model and reduced disease progression and severity [63]. However, a recent study demonstrated that transgenic cells differentiated with myelin antigen in the presence of IL-6 + TGF-β1 or TGF-β3 produced large amounts of Th17, but were not adequate to induce encephalomyelitis [64].
3.2. TGF-β and BBB integrity
BBB disruption is one of the hallmark pathological mechanisms of MS progression. Following this BBB impairment, reactive immune cells and their products can diffuse into the brain parenchyma, inducing an inflammatory environment that results in neurodegeneration. BBB mainly contains endothelial cells (ECs), pericytes, and astrocyte end feet in the presence of an extracellular matrix. Also, establishing endothelial tight junctions, including cytoplasmic adaptors like zonula occludens protein 1 (ZO-1), claudins and occludin, and create a barrier resistance to water and ions [65]. TGF-β1 plays a crucial role in BBB formation during fetal development by adjusting adhesion junctions. In the early stages of BBB establishment in the fetus, after organizing tight junctions, in response to CD146 expression on endothelial cells, pericytes are recruited and contribute to BBB maturation. Recruited pericytes secrete TGF-β1 that down-regulate CD146 expression on endothelial cells, diminishes the extra pericyte recruitment, down-regulate leukocyte infiltration, and contribute to BBB maturation and integrity [26]. Moreover, pericyte-derived TGF-β1 downregulate claudin-5 expression on endothelial cells by diminishing CD146 expression. Meanwhile, activating the TGFRII-ALK5-Smad2/3 pathway in pericytes by endothelial TGF-β inhibits their proliferation and facilitates their differentiation toward an ECM-productive form [66]. Likewise, TGF-β upregulates the specialized tight junction N-cadherin which mediates pericyte-endothelial attachment in collaboration with the NOTCH signal [26]. Nevertheless, it seems that dynamic expression and down-regulation of claudin-5 tight junction and pericyte organization affected by physiological levels of TGF-β1 may be an essential axis in maintaining CNS hemostasis.
Considering OPCs as an essential source of TGF-β1 secretion, OPC-specific TGF-β1 partial-knockout (KO) mice manifested cerebral hemorrhages and IgG leakage, as tight junction molecule ZO-1 was degraded [67]. In contrast, another study indicated that the TGF-β1 inhibitor could up-regulate claudin-5, occludin, and ZO-1 by blocking the TGF-β/ALK5 signaling pathway [68]. Moreover, the administration of exogenous TGF-β1 in a mouse model increased BBB permeability by down-regulating claudin-5 and up-regulating MMP9 [69].
Blood endothelial cell transition into the mesenchymal cell, a process termed EndMT, is highly induced by all three isoforms of TGF-β, especially by TGF-β2 [70]. This EndMT process which is present during several neurological inflammatory diseases like MS leads to the transition of endothelial cell adhesion molecule VE-cadherin to mese nchymal N-cadherin. Moreover, this reduces endothelial markers like CD31, induces mesenchymal phenotype, and increases their capacity for migration by disturbing the integrity of endothelial cells [71] and subsequent immune cells’ infiltration into CNS. Taken together, these evidence suggests that dysregulated expression of TGF-β isoforms may be a suspect of BBB impairment through disturbing BBB integrity and inducing the EndMT process.
3.3. Role of TGF-β in fibrogliosis and scar formation
Increased levels of intermediate filament proteins, such as glial fibrillary acidic protein (GFAP), vimentin, and nestin, and changes in actin cytoskeleton proteins are manifested by reactive astrocytes [72, 73]. In an in-vitro and in-vivo study by Romão et al., it was demonstrated that in astrocytes, the GFAP gene promoter is activated by TGF-β1, and this phenomenon resulted in astrogliosis [74]. Despite that, in another study, TGF-β1 did not alter GFAP protein levels but significantly decreased their organization toward a dense glial barrier and reduced laminin and fibronectin deposition in the lesion site [75]. In MS lesions, α-SMA and nestin-expressing reactive astrocytes were observed in close adjacency with TGF-β1 expressing microglia/macrophages. Following this observation, in-vitro immunohistochemistry showed increased organization and expression of α-SMA and nestin in TGF-β1 treated astrocytes [72]. In addition, astrocytes in the optic nerve head of mice presented reactivity in response to TGF-β2 [76]. Interestingly, TGF-β3 and TGF-β1 treatment in rat cortices elevated the levels of GFAP expression, which is known as the primary marker of astrocyte reactivity. However, the level of TGF-β3-induced reactivity was in less extent form compared to TGF-β1 treatment [77]. These results uncovered the effects of TGF-β1, TGF-β2, and TGF-β3 on astrocyte activation and subsequent ECM deposition.
Migration of reactive astrocytes to the injury site is also necessary for glial scar formation [78]. In a study of rat cerebellum cultured astrocytes, TGF-β1 proceeded migration of astrocytes through rearrangement of the actin cytoskeleton [79]. So far, it is known that reactive astrocytes increase aquaporin 4 (AQP4) expression in various CNS pathologies, a mechanism proposed to facilitate their migration toward the lesion [80]. Consistent with that, Nataf et al. demonstrated a TGF-β1-dependent upregulation of AQP4 in human astrocytes in progressive MS [81], which is also strong evidence indicating the indirect role of TGF-β1 in the case of astrocyte migration in MS lesions.
β-catenin expression is increased in reactive astrocytes, which makes it to be considered a functional marker. Also, it has been shown to enhance some gene expressions that activate MMP and facilitate astrocyte migration [82]. TGF-β1 protects β-catenin protein degradation and comforts its nuclear accumulation through SMAD3 and SMAD4 pathways [83], and therefore mediates astrocyte migration. One study investigated the interactions between TGF-β1 and 5-lipoxygenase (5-LOX) in astrocyte migration. 5-LOX metabolizes arachidonic acid and produces cysteinyl leukotrienes (CysLTs), which have been shown to increase after astrocytosis. CysLT can act as a ligand on CysLT1R in an autocrine manner in astrocytes. TGF-β1 up-regulates CysLT1R and activates 5-LOX mRNA. Moreover, the TGF-β1 effect was inhibited in the presence of 5-LOX inhibitor zileuton [84]. Therefore, it is not far from the expectation that TGF-β1 can drive astrocyte migration toward scar formation through its interaction with Aqp4, LOX/CysLT1R, and β-catenin pathways in MS.
As discussed above, normal-appearing astrocytes are not usually detected with GFAP staining, and the up-regulation of GFAP in reactive astrocytes may give the false impression of proliferation [85]. However, given that the term glial scar formation is under both morphological changes and proliferation of astrocytes, research revealed that TGF-β1 alone does not induce multiplication or proliferation in astrocytes [86] and that the proliferation of astrocytes may be under the control of other inflammatory cytokines or in a combination of TGF-β1 + LPS + IFNγ [87]. Thus, the TGF-β1 messaging pathway alone is not sufficient for inducing glial scar formation and activation of other pathways is also required.
The presence of fibroblast-like cells in the demyelinated lesions in MS and EAE has been widely studied [88]. The immunohistochemical studies of post-mortem MS patients’ biopsy demonstrated a lack of sub-endothelial mesenchymal pericyte cells (MPCs) and an increased level of perivascular MPCs in demyelinated lesions highly suspected to originate from sub-endothelial MPCs. these perivascular MPCs present a fibrotic phenotype that contributes fibrosis in MS scars [89]. As discussed above, TGF-β, irrespective of its isoform, can intercede the EndMT process, and this transition of BBB endothelial cells to mesenchymal cells can be considered a source of fibrotic cell formation in MS lesions, through pericyte and fibroblast formation [70]. Moreover, as reviewed elsewhere, TGF-β1 and 2 stimulate differentiation and proliferation of fibroblasts and induce them to produce ECM molecules, such as fibronectin and collagen, whereas TGF-β3 is suggested to have antifibrotic outcomes [90].
Basal membrane proteins, such as laminin, fibronectin, collagen type V, and heparan sulfate proteoglycans (HSPGs), are usually present around vascular endothelial cells. Nevertheless, they also become present in different plaques of MS and are highly associated with increased levels of TGF-β1 [91]. Chondroitin sulfate proteoglycan (CSPG), ordinarily present in CNS parenchyma, is also up-regulated in active and inactive MS lesions [92]. Astrocytes increase the production of CSPGs in response to TGF-β1 [93]. CSPGs protein core, in combination with the glycosaminoglycan carbohydrate chain, creates an inhibitory environment for tissue regeneration [94]. Subsequently, injection of TGF-β1 in a rat model of CNS injury demonstrated increased laminin, fibronectin, and collagen deposition. It also resulted in the formation of coalescent glial limitans bordering the lesion fibronectin site [75]. TGF-β1 also upregulates neurocan secretion in reactive astrocytes [106], a proteoglycan with neurite outgrowth inhibitory characteristics [95]. Results from animal models of spinal cord injury indicated that TGF-β1 mRNA was raised during the inflammatory phase, while during the scar formation phase, TGF-β2 mRNA was up-regulated around the lesion site. Consistent with the approved presence of TGF-β2 in the lesion site, this data might indicate that TGF-β2 collaborates more strongly with scar formation than TGF-β1 [96]. Injection of anti-TGF-β2 antibody in a model of CNS scar formation diminished laminin, fibronectin, and CSPG deposition [97]. This study highlights the role of TGF-β2 in inducing scar formation.
Previously, it was postulated that glial scar formation inhibits neural outgrowth. However, recently, pieces of evidence have shown the beneficial effects of astrocyte scar formation. It was hypothesized since reactive astrocytes secrete axonal growth permissive CSPGs, laminins, and other molecules like DSD-1‐PG, which have neurite‐outgrowth promoting properties [98, 99]. Therefore, the effects of TGF-β-induced astrogliosis in MS progression is still a matter of debate and further studies are still indicated.
3.4. TGF-β and remyelination
Remyelination is a process driven by CNS oligodendrocytes. The disturbance in remyelination process is one of the key pathologic features of MS progression. As mentioned before, oligodendrocyte progenitor cells' (OLC’s) proliferation, migration, and differentiation are necessary for adequate remyelination of the CNS neurons. TGF-β1 has been known as an inhibitory factor of OLC proliferation in many neural and glial cells [100]. A study investigated the TGF-β1 effect on remyelination. The findings of this study indicated that TGF-β1 induces OPC’s differentiation resulting in myelin-related protein (i,e. myelin basic protein (MBP) and myelin oligodendrocyte glycoprotein (MOG)) synthesis. Moreover, TGF-β1 activates cyclin-dependent kinase inhibitors and therefore suppresses OPC’s proliferation. At the same time, an increased number of proliferating OPCs and a reduced density of mature OLs in TGFRII-deficit transgenic mice were observed [101]. Besides, TGF-β2 and TGF-β3, secreted from PDGF-treated proliferating OPC culture, suppress OPCs proliferation, but at the same time induced its differentiation toward mature oligodendrocytes [102]. Moreover, in an EAE study, all 3 isoforms of TGF-β secreted by Th2/3 lymphocytes were able to induce migration of OPCs toward demyelinated lesions [103]. Taken together, all isoforms of TGF-β might be helpful in the differentiation and migration phase but not in the recruitment phase of oligodendrocytes. Nevertheless, it is not clear whether TGF-β contributes to or indirectly prevents the remyelination process.
Analyzing the genomic program of peri-plaques from Progressive Multiple Sclerosis Patients (PMSP) revealed epigenetic regulation of TGF-β1 in myelin-related genes [104]. N-myc downstream-regulated gene 1 (NDRG1) contributes to maintaining myelin sheaths after being co-expressed with myelin genes and ribosomal genes supporting mRNA translation. The binding of N-myc to the promotor region of NDRG1 has been shown to hypermethylate the promotor and suppress its gene transcription [104]. Consistent with hypermethylated NDRG1 promotors observed in NAWM and MS peri-plaques, the authors identified a relative protein network of N-myc and TGFRI which may suggest that this cytokine can silence NDRG1 through upregulating N-myc and deteriorating remyelination [105].
TGF-β promotes the differentiation of naïve macrophages into the M2 alternative phenotype, which is more phagocytic, less pro-inflammatory cytokine producer, and has lower expression of human leukocyte antigen DR (HLA-DR), resulting in less antigen presentation to lymphocytes [106]. In lesion samples of MS patients, M1 macrophages and microglia convert into M2 phenotype, induce differentiation of OPCs, and promote remyelination. These effects may be due to the Activin-A secretion by M2 macrophages [107] or due to the ability of M2 monocytes for myelin debris phagocytosis, as myelin debris is suggested to inhibit OPC’s differentiation, and therefore, neuron remyelination [108]. This myelin debris is highly rich in cholesterol and can not be broken down by phagocytes. Therefore, phagocytes need to transfer cholesterol to the extracellular matrix to deal with this intracellular lipid overload through a process mediated by activation of the liver X receptor (LXR)-regulated genes. A recent study has identified that TGF-β signaling suppresses the LXR pathway, which leads to lipid accumulation, and therefore suppresses remyelination [109]. To conclude, these data may indicate the adverse effect of TGF-β in the remyelination process and MS progression, even in the presence of M2 macrophages.