**6. Non‐IgG ITS**

#### **6.1. General background**

The IgG class accounts for the bulk of locally synthesized Igs in MS and most knowledge to date concerns this class. All Ig classes may give rise to ITS. IgG, IgM, and IgA are, respectively, present in 100, 20–36, and 6–9% of MS patients [69, 78]. The relative absence of IgE is commonly interpreted as a clue to type I immunity [147].

#### **6.2. IgM chains**

in a study including 1800 MS patients, the number of OCB was considered to correlate with the course of the disease [127]. The mean number of OCB in the group of benign patients was 2.9 ± 3.6 compared to 5.7 ± 4.9 in the severe group (*p* < 0.06) [127]. Therefore, the evidence points to a poorer clinical prognosis in MS patients in association with ITS. On the other hand, OCB negativity correlates with a better outcome in many studies [70, 127–131], even after demo‐

**5. Intrathecal Ig synthesis in animal models—technical limitations**

IgG synthesis level in CSF is highly correlated with CD19+CD138+ plasmablast levels [133], but although ITS is commonly thought to be associated with CSF floating cells, this minute cell number may account for less than 0.1% of the whole ITS, meaning that the bulk of IgG synthesis is provided by resident IgG‐secreting cells, residing either in the meninges or in the perivascular areas [134]. In experimental allergic encephalomyelitis (EAE) models, the parenchymal level of B‐cell infiltration increases from null to 134 B‐cell/cm<sup>2</sup> [135] and the whole

In MS, B‐cells remain a minor proportion of the infiltrating lymphocytes (<5%), are virtually absent from the parenchyma, and are mostly observed around the small veins and meninges

IT Ig synthesis in animal models has received little attention since the very small volume of CSF does not facilitate sampling. Although older experiments based on low‐sensitivity techniques failed to demonstrate any ITS during EAE, optimal techniques (IEF followed by anti‐IgG staining) have subsequently confirmed the local synthesis of OCB and/or an elevated IgG index [139–142]. In viral models of demyelination (i.e., measles or JHM strain of corona‐ virus infection), a mirror pattern of OCB is common but local OCBs are always revealed by immunoblot against viral antigens [21, 143]. We are not aware of any EAE experiments aimed at demonstrating the nonspecific pattern of ITS observed in MS, especially the animal coun‐ terpart of the human MRZ pattern using common animal viruses or vaccines. From a theoret‐ ical point of view, animals receiving intrathecal vaccination with foreign antigens such as albumin are able to mount a strong intrathecal response [144–146]. In a unique case challenged intrathecally with ovalbumin (OVA), an increase in AI‐herpes was also obtained apart from the expected increase in AI‐ovalbumin, suggesting that animal models are promising for future

The IgG class accounts for the bulk of locally synthesized Igs in MS and most knowledge to date concerns this class. All Ig classes may give rise to ITS. IgG, IgM, and IgA are, respectively,

are available regarding a possible role of meningeal B‐cell aggregates in ITS.

)[1, 138]. Plasma cells are occasionally found in demyelinated areas. No data

–105 cells per mouse or rat brain [136, 137].

graphic adjustment [132].

62 Trending Topics in Multiple Sclerosis

(4.6–6 cells/mm<sup>2</sup>

studies of nonspecific ITS [145].

**6. Non‐IgG ITS**

**6.1. General background**

number of infiltrating B‐cells is in the range of 10<sup>3</sup>

Intrathecal IgM synthesis is always higher in MS than in controls [110]. IgM OCB may occur in about 20–75% of MS patients [78, 113, 118] and in 74% of CIS [125]. CD19+CD5+ lymphocytes are the predominant population of B‐cells in CSF and are characterized by the T‐cell inde‐ pendent secretion of IgM natural antibodies directed against phylogenetically conserved structures that target non‐proteic epitopes. In a series of 53 CIS/relapsing‐remitting multiple sclerosis (RRMS), 46 of them had oligoclonal IgM‐recognizing lipid antigens, 30 targeted phosphatidylcholine (PC) alone, seven targeted PC and had additional reactivity against other myelin lipids (phosphatidylethanolamine, phosphatidylinositol, and sphingomyelin), and nine targeted myelin glycolipids, mostly sphingomyelin [113]. IgM OCB against lipids may be restricted to the RR and SP phases since they were not observed in PP patients [27]. Pentameric IgM strongly activates the complement cascade and is involved in the lesions of plaques (pattern II).

#### **6.3. IgA**

The IgA index is elevated in at least 9% of MS patients [54, 78, 148] and the mean intrathecal IgA index is modestly but significantly higher in patients than in controls [110]. Local synthesis of IgA may be the only clue to local Ig synthesis in some MS patients devoid of local synthesis of IgG [31]. About a quarter of Ig‐secreting cells are IgA in choroid plexuses [149]. IgA cells are also common in plaques (3–9%) and demonstrate a high clonality (up to 80%) [149]. IgAs are mainly mucosal secretory antibodies but could be a major component of CNS immune responses in certain viral infections [148, 150]. OCB IgAs were reported to have a low frequency (2/33) in the earlier CSF studies [71, 150]. In a post‐mortem analysis of plaques extracted from four MS brains, IgA+ plasma cells were recovered in 18/24 plaques and sometimes predomi‐ nated on Ig+ plasma cells [148]. Clonality assessment in the plaques of two patients demon‐ strated a low diversity of clones and a clonal restriction confined to plaques [148]. The somatic mutations were highly concentrated in the complementary determining region (CDR) and FR3, which are considered critical for antigenic specificities, suggesting an antigen‐driven matura‐ tion of affinity as observed for IgG. The locally synthesized IgA were demonstrated to target injured axons [148].

#### **6.4. Free light chains**

Light chains are normally synthesized more than heavy chains, with an FLCκ/*λ* ratio of 2:1 in normal blood. Serum circulating levels of FLC are about 1000‐fold lower than Ig levels owing to rapid renal clearance (serum half‐life of 2–6 h), whereas the rate of clearance from CSF is similar with Ig [151]. Therefore, the level of κFLC in MS is equal or higher in CSF than in serum. Although the absolute value of CSF FLC is variable among studies, the FLC level is always higher in MS than in controls, and the κFLC/protein ratio is even higher [152]. The κFLC index is more sensitive (Se 95%, Sp 91%) than the IgG index and OCB [153]. Since elevated CSF FLC is predictive of impairment [116], it has even been proposed as a therapeutic target [47].
