Role of CXCL13 in the formation of the meningeal tertiary lymphoid organ in multiple sclerosis

Immunomodulatory therapies available for the treatment of patients with multiple sclerosis (MS) accomplish control and neutralization of peripheral immune cells involved in the activity of the disease cascade but their spectrum of action in the intrathecal space and brain tissue is limited, taking into consideration the persistence of oligoclonal bands and the variation of clones of lymphoid cells throughout the disease span. In animal models of experimental autoimmune encephalomyelitis (EAE), the presence of CXCL13 has been associated with disease activity and the blockade of this chemokine could work as a potential complementary therapeutic strategy in patients with MS in order to postpone disease progression. The development of therapeutic alternatives with ability to modify the intrathecal inflammatory activity of the meningeal tertiary lymphoid organ to ameliorate neurodegeneration is mandatory.


Amendments from Version 2
We are pleased to conclude the writing of this manuscript after the total approval of its content by the referees. Their review was a strong source of motivation which led us to further investigate the existing scientific literature on the role that CXCL13 could play in the pathogenesis of MS and the therapeutic considerations that could arise should this chemokine indeed finds a pivotal role for itself in this condition. We emphasize the fact that the immune response in MS is still incompletely understood and profoundly complex. For this reason, extreme caution has to be exercised during attempts to extrapolate scientific findings from bench to bedside. We do hope that this final version leads to further discussions with colleagues and scientists dedicated to the study of the pathophysiology of multiple sclerosis.

Introduction
Although disease modifying therapy (DMT) agents in multiple sclerosis (MS) have contributed to reduction of neuroinflammation, they have not succeeded in the prevention of progression of disease. Inflammation is the appropriate tissue response to infection, autoimmunity, cancer, injury and allograft transplantation 1 . When inflammation does not resolve appropriately, a prolonged immune response persists leading to tissue destruction and loss of function 1 . Chronic infiltration by immune cells in the meninges is believed to form transitory lymphoid cell aggregates which simulate secondary lymphoid organs (SLO), and are known as meningeal tertiary lymphoid organs (mTLO) which play an important role in the pathogenesis of autoimmunity 1,2 . The mTLO seem to play a role in the intrathecal activity of immune system cells in MS 3 . The SLO, such as lymph nodes, show a cellular organization that includes germinal centers (GC) containing antibody secreting and proliferating B-cells with follicular dendritic cells (FDC), a T-cell zone that incorporates naïve cells from the blood stream, high endothelial venules for extravasation of lymphocytes, and a stromal cell network that provides chemokines and extracellular matrix for cell migration and structural integrity 1 . Chemokines are a family of proteins with the specific property of regulating leukocytes in the immune system and they may play a role in neurotransmission and neuromodulation 4 . Leukocyte trafficking is mediated by inflammatory chemokines in inflamed tissues and by homeostatic chemokines in lymphoid sites 5 (Figure 1). In this review, we focus on the role that CXCL13 (also known as B cell attracting chemokine [BCA-1], C-X-C motif ligand 13, or B lymphocyte chemoattractant [BLC]) plays in the organization of the mTLO in MS.

In normal conditions, the SLO acquire information and prepare for immune defense
The SLO have a genetically determined pattern of development and programming that allows trapping and concentration of foreign antigens to initiate an adaptive immune response 1 . Mucosal associated and non-encapsulated lymphoid tissue (including the Peyer's patches, adenoid tissue of the nasopharynx, tonsils, and the bronchial associated lymphoid tissue), together with lymphoid nodes and spleen, constitute the SLO 6,7 . The lymph node cortex contains clusters or primary follicles that include packaged B cells and FDC, whereas the node para-cortex has a lesser number of dendritic cells (DC) and T cells 6 . Generation of B cells with ability to produce auto antibodies usually occur in physiological conditions 8 . These auto antibodies are low affinity IgM, which exhibit a wide spectrum of reactivity and strong preference for soluble self-antigens on the cell surface 8 . Auto reactive low affinity B cells suffer apoptosis being unlikely they represent danger in normal conditions 8 .

Lymphoid cells are able to learn and exchange information at the GC
The GC present remarkable lymphocytic mitosis within SLO follicles 9 . Weyand et al. stated the GC are critical in the development of the B-cell normal immune response by drivingin cell division and maturation, B-cell selection with high affinity for immunoglobulin receptors and differentiation of B-cells and plasma cells (PC) 2 . Real time imaging technology has allowed visualization of the transit of the B cells from the dark zone to the light zone, and vice versa, during the maturation of the GC 9-11 . The GC light zone displays a predominance of FDC and follicular T-helper (Tfh) cells, whereas the dark zone contains closely packed lymphocytes and stromal cells 9,12 . The chemokine receptor CXCR4 is required for the positioning of the B cells in the dark zone where its ligand, CXCL12, is more abundant and is produced by stromal cells 13 . At the light zone, CXCL13 chemokine is concentrated in the FDC processes and, in conjunction with CXCR5, they contribute to the accumulation of B cells in this zone 12,13 . T-cells in the GC are essential to maintain signaling and represent approximately 5-20% of cell population 9 . Tfh cells are characterized by the expression of CXCR5 and ICOS, which has an inducible T cell co-stimulatory effect 9,14 . B cell growth and differentiation at the germinal center are supported by IL10 and ICOS 14 . Within the light zone, the three possible different outcomes for the centrocytes include death due to apoptosis; differentiation into memory B-cell or long lived plasma cells (LLPC); and re-entrance to the dark zone for a further round of cell mutation and selection 15 . The relevant function of the GC is, most likely, the primary production of memory B-cells and LLPC 15,16 ( Figure 2). Recent studies analyzing IgG heavy chain variable region genes in B cells from MS patients revealed that B cells are able to enter and exit the blood brain barrier (BBB) in order get exposed to somatic hypermutation (SMH) at the GC 17-21 .

Chemokines direct traffic of lymphocytes during the cell search for specific information
The induction of lymphoid chemokines, depends on lymphotoxin β (LT-β) and the tumor necrosis factor α (TNF-α) signaling on stromal cells and FDC 22 . Lymphotoxin α1β2 (LTα1β2) is expressed in the surface of B and T cells in the adult immune system and ligates to the lymphotoxin β receptor (LTβR) in reticular stromal cells thus inducing expression of lymphoid chemokines, such as CCL19, CCL21 and CXCL13 23 . These chemokines regulate the homeostatic traffic of lymphocytes in lymphoid organs and their distribution in the GC 23 . Homeostatic chemokines promote secretion of LTα1β2 by T and B cells, establishing a feedback loop that perpetuates the recruitment 1. CXCL13 increases its own production by stimulating the growth of FDC after regulating LTα1β2 on the membrane of B cells 5,26 .

CXCL13 is produced in the SLO by FDC and macrophages
and is an important chemoattractant to the CNS 27,28 .

Could the TLO become an operation center with ability to magnify an autoimmune response?
By maintaining antibody diversity, B cell differentiation, isotype switching, oligoclonal expansion, and local production of autoreactive PCs, the TLO perpetuate disease in response to environmental inputs 34 . Lymphoid organogenesis and formation of mTLO may be facilitated by expression of lymphotoxin α (LT-α) at the external layer of meningeal inflamed vessels leading to the compartmentalization of the immune response in MS 17,35 . It has been postulated that the perpetuation of neuroinflammation and disease progression results from mTLO induced differentiation and maturation of antigen specific effector lymphocytes 28 . The identification of independent centroblasts in the CSF of MS patients has suggested that there is an intrathecal B cell differentiation which is not dependent on the immune activity in the blood compartment 17 . The TLO, besides SLO, provide a thriving environment where PC differentiate from plasmablasts 6,28 . In the absence of recirculating immune cells from the periphery, the TLO exerts its remarkable ability to remain active for several weeks 36 . Therefore, the neutralization of TLO could play a significant role by blocking the re-emergency of auto reactive clones that could be able to drive relapses or resistance to therapy 36

Facts learned from the role of CXCL13 in other autoimmune diseases
Chronic and active inflammation in target organs such as thymus, thyroid, synovial tissue and salivary gland can be driven by formation of TLO in the corresponding target organs 2 . Expression of CXCL13 in pancreatic tissue has been associated with formation of ectopic lymphoid follicles and the induction of cascade events leading to diabetes in a transgenic mice model 46 .
In patients with Helicobacter pylori the blockade of CXCL13 has prevented the development of mucosa-associated lymphoid tissue (MALT) lymphoma and the propagation of inflammation by CXCL13 have been documented in non-lymphoid tissue 47 . Although formation of pulmonary lymphoid follicles has been seen in patients with complicated rheumatoid arthritis, idiopathic pulmonary hypertension and Sjogren's syndrome, the presence of lymphoid follicles has correlated with positive outcome in patients with lung cancer and infections of the respiratory tract 48 . Lung B cells are a major source of CXCL13 and it has a positive role in the lymphoid neogenesis in chronic obstructive pulmonary disease through a LT receptor and toll-like receptor signaling 48 .

Possible cellular sources of CXCL13 in the CNS
CXCL13 plays an important role in the formation of the GC in ectopic lymphoid follicles of several organs affected by inflammatory or autoimmune disease, or by infection 1 52 . The source of intrathecal production of CXCL13 during neuroinflammation has not been determined with certainty. However, stromal cells within the B cell follicles have been considered to be responsible for the chemokine production and the notion of simple passive transfer from the blood stream to the intrathecal compartment due to dysfunction of the BBB has been detracted 13,26,51,53 . According to Essen and collaborators, stromal cells in the meninges could produce CXCL13 in special circumstances and drive the focal accumulation and organization of lymphoid cells in specific sites 51 . In their study, they were able to demonstrate that microglia cells, in vivo and in vitro, are the main producers of CXCL13 in acute neuroinflammation induced by the Sindbis virus, which is not associated with demyelination 51 . They also found that type-1 interferon could suppress the production of CXCL13 by microglial cells 51 . In the rhesus macaque model of neuroborreliosis, Ramesh et al. and Narayan et al. found that infiltrating microglia and macrophage/ DC myeloid cells could be one source of CXCL13 in the CNS during inflammation 54,55 . In biopsy specimens from patients with primary CNS lymphoma, Smith and collaborators encountered the following: 1) FDC were not present in the analyzed specimens; 2) there was expression of CXCR5 and CXCL13 in malignant B cells with positive production of BCA-1 (CXCL13) mRNA; and 3) CXCL13 was present in endothelial vascular cells which had a negative production of BCA-1 (CXCL13) mRNA by in situ hybridization, a finding that could be attributed to transcytosis 56 . In different types of EAE, CXCL13 and BAFF mRNA transcripts were found to be significantly upregulated in the CNS of mice which developed the relapsing-remitting and the chronic-relapsing courses of disease opposite to those which developed a chronic progressive course. Besides, cells expressing CXCL13 were exclusively found in the brain stem meninges where infiltrating leukocyte proliferation was intense and vascular endothelial cells did not express CXCL13 57 . In specimens from patients with giant cell arteritis, arterial TLO with FDC precursors and lymphoid ducts were detected in the medial layer of the temporal arteries expressing CXCL13, BAFF, APRIL, IL 7, IL 17 and LTβ 58 .

Could CXCL13 be neutralized by direct action on itself, its receptor (CXCR5) or the lymphotoxin β (LT-β)?
A novel therapeutic monoclonal antibody against CXCL13 (Mab 5261 and Mab 5261-muIg) has been shown to induce functional in vitro inhibition of the chemokine in humans and mice 8 . An LTβ receptor blocking immunoglobulin inhibits CXCL13 interactions, suppresses the formation of mTLO in the CNS and ameliorates the symptoms of EAE in rodents 23

Would a complementary intrathecal therapy for deactivation of the mTLO be necessary to arrest disease progression?
A self-sustained intrathecal inflammation fostered by CSF chemokines involved in the traffic and survival of inflammatory cells occurs early in disease and is orchestrated by mTLO 3 . Studies have shown that lineage of B cells can travel through peripheral blood, cervical lymphoid nodes, and the intrathecal compartment where they can be exposed to SMH in the mTLO and return to peripheral blood 17 .

Conclusion
An early neutralization of CXCL13 could interfere with the organization and function of the mTLO thus modifying and reducing inflammation in the CNS of patients with MS. Studies in animal models where CXCL13 has been neutralized, or is not expressed (such as the CXCL13-/-mice), confirm its crucial role maintaining, rather than initiating, inflammation and its manipulation could lead to modification of disease in these models 39 . However, any therapeutic strategy unable to neutralize LLPCs or antibody secreting cells will not be successful in an attempt to impede the chronic progression of disease 73 . Neutralization of the CXCL13 should be carefully sought as complementary therapy to the DMT in MS.

Data availability
No data is associated with this article.

Grant information
The author(s) declared that no grants were involved in supporting this work. In this review article the authors describe the role of the chemokine CXCL13 in the formation of meningeal TLOs in MS and suggest it as a therapeutic target. The article is well written and the first half of the article provides a very detailed overview of the components and requirements for formation of secondary lymphoid organs. The authors then switch to tertiary lymphoid organs in the CNS assuming that all components and mechanisms of formation are identical to SLOs. While this may be the case for some of the most developed TLOs in some autoimmune diseases like Myasthenia gravis, it is not so clear which cell types and molecular players are required for formation of meningeal TLOs. In fact, to my knowledge it has not even been formerly proven that CXCL13 is required for formation of meningeal TLOs. Even though it is quite likely considering detection of CXCL13 in mTLOs and elevated CXCL13 levels in the CSF of MS patients, definitive proof even in the animal model is missing as also pointed out by reviewer 1, because a) CXCL13-deficient mice already have a defect in mounting proper immune responses in SLOs and b) active EAE induced by MOG-peptide/CFA immunization does not prominently feature mTLOs. The mouse models that do feature mTLOs such as the spontaneous 2D2xTh mouse have not been studied in the context of CXCL13 deficiency.
Furthermore, it would be very useful to discuss in this review cellular sources of CXCL13 in the CNS, as they may not be identical to SLOs. Thus, microglia (Ref 37) and meningeal stromal cells (Pikor et al., Immunity, 2015 ) have been suggested as sources for CXCL13 and should be discussed.
Another important point is that the authors state that the "mTLO maintain differentiation and maturation of antigen-specific lymphocytes which perpetuate inflammation and disease progression". This is not a fact but a hypothesis and should be stated as such. While it is clearly an attractive hypothesis there is no proof neither in mouse models, nor in MS. We agree with the authors that in MS occurence of TLOs has been with more severe disease course and cortical lesions, however, causality has not been associated demonstrated and even evidence for maturation of antigen-specific lymphocytes in mTLOs is very limited so far. Therefore, we believe that it is not justifiable to interfere with mTLO formation in MS patients, as 1 so far. Therefore, we believe that it is not justifiable to interfere with mTLO formation in MS patients, as long as their biological function and consequences are not much better understood.
As a side note, some sentences are a bit unclear, for example in the introduction "Inflammation is the appropriate immune response to...autoimmunity..." (pg 2) and "Tfh cells are characterized by expression of CXCR5 and ICOS, which is a subtype of Tfh cells" (pg 3).
Overall, the review has a very interesting and important topic, however, in my opinion as detailed above in several paragraphs the wording should be a bit more careful and precise in order not to be misleading.
interesting material to the discussion giving further support to the role of this chemokine in the pathogenesis of MS. Certainly, patient safety is a demanding priority and any application of the knowledge acquired from or EAE animal model studies to the treatment of patients with MS in vitro should be considered with extreme caution. In the review article the authors highlight the potential importance of tertiary lymph follicle like structures in the meninges of MS patients as driving forces for tissue damage and in particular cortical demyelination. They provide a very good summary of immunological mechanisms, which are involved in the organization and function of secondary lymph follicles in the peripheral lymphatic tissue and in particular highlight the importance of the interaction of CXCL13 with other cytokines and chemokines in these processes. They then describe in detail the evidence for the presence of structures with features of tertiary lymph follicles in the meninges of MS patients and their association with disease severity and cortical pathology. Finally they also review in detail the observations that CXCL13 is present in the CSF and may serve as a biomarker associated with poor prognosis of the disease. Based on the experimental observation that CXCL13 blockade or genetic ablation ameliorates EAE the authors propose that therapeutic blockade of CXCL13 in the CNS compartment of MS patients may be beneficial.
There is now good cumulative evidence that such follicle like inflammatory aggregates in the meninges are an important substrate of disease pathology in MS and that B-lymphocytes play an important, but so far not fully understood pathogenetic role in the disease. It is also clear that CXCL13 is an important chemokine, involved in B-cell recruitment into the central nervous system. However, it may be premature at the present time to propose intrathecal CXCR13 blockade as a therapy for MS patients. The EAE studies are only of limited value. It is not a surprise to ameliorate EAE with a therapy, which has major effects on the organization and function of peripheral lymphatic tissue. Although some EAE models show lymphocytic aggregates in the meninges, which share some features with those in MS, this is not the case in the majority of the models. Furthermore, in the respective mouse EAE models with lymph follicle like aggregates in the meninges there is no cortical demyelination. Thus lesion pathogenesis apparently is quite different between these models and MS. To what extent an intrathecal blockade of CXCL13 has an effect on CNS inflammation and what kind of effect will be achieved, is currently unknown. Whether this may induce dangerous side effects is also unclear. The suggestion to combine such a treatment with simultaneous intrathecal elimination of B-cells, T-cells and other immune cells is also far away from realization. Elimination with currently used antibodies requires complement of antibody dependent cellular cytotoxicity, and whether this is safe to induce within the CNS compartment in patients is also rather uncertain. Thus this review addresses a topic, which is interesting in a disease such as multiple sclerosis, but the suggestions for therapeutic translation are currently premature and potentially dangerous.

Are all factual statements correct and adequately supported by citations?
Partly