Though multiple sclerosis (MS) is considered the cornerstone of the central nervous system (CNS) demyelinating disorders, several clinical entities are important MS mimickers and need to be taken into consideration in the differential diagnosis (Brownlee et al., 2017; Thompson et al., 2018a; Thompson et al., 2018b; Solomon et al., 2019). Among those, systemic autoimmune diseases (SADs), neuromyelitis optica spectrum disease (NMOSD), and myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD) are the most common. Less frequently, neurosarcoidosis, neuro-Behçet disease, and chronic lymphocytic inflammation with pontine perivascular enhancement responsive to steroids (CLIPPERS) may present with MS reminiscent manifestations. Despite a distinct pathogenetic background, atherosclerotic small vessel disease or even cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL), Fabry disease, and Susac’s syndrome can also present with MS-resembling features (Thompson et al., 2018b).

SADs, mainly systemic lupus erythematosus (SLE), Sjögren’s syndrome (SS), and antiphospholipid syndrome (APS) have been previously shown to invariably affect the CNS, with transverse myelitis, optic neuritis, and white matter hyper-intensities (WMH) in CNS magnetic resonance imaging (MRI) being the main MS-like manifestations (Delalande et al., 2004; Ferreira et al., 2005; Pelidou et al., 2007; Theodoridou and Settas, 2008; Massara et al., 2010; Magro Checa et al., 2013; Piga et al., 2017; Hughes, 2018; Schwartz et al., 2019). In previous studies, higher rates of homozygous methylenetetrahydrofolate reductase (MTHFR) mutations have been found among patients with autoimmune features and MS-like or vasculitic CNS disease, and a link has been identified between anti-thyroid antibodies with nonspecific WMH and CNS involvement in the setting of APS (Mavragani et al., 2007; Mavragani et al., 2009).

While usually helpful, brain MRI (Stosic et al., 2010; Akasbi et al., 2012; Magro Checa et al., 2013; Magro-Checa et al., 2018) and cerebrospinal fluid (CSF) analysis (Govoni et al., 2016; Venhoff et al., 2019) cannot consistently differentiate between MS and CNS involvement in the setting of SADs. Though the detection of serum autoantibodies has a central role in the diagnosis of SAD, they are reported to precede disease onset by many years (Arbuckle et al., 2003). Additionally, low-titer antinuclear (ANA) and antiphospholipid antibodies are frequently found in the context of MS at variable frequencies ranging between 3%-–63% and 0.9%–8.5%, respectively (Spadaro et al., 1999; Szmyrka-Kaczmarek et al., 2012; Malyavantham et al., 2015; Merashli et al., 2017). Of interest, MRZ reaction [detection of CSF IgG antibody indices to measles (M), rubella (R), and varicella-zoster (Z)] was found to be prevalent in patients with MS and ANA positivity compared to those with rheumatological disease with CNS involvement, providing a potential differentiating tool between the two entities (Venhoff et al., 2019).

Despite the potential phenotypic similarities, distinct underlying pathogenetic mechanisms and therapeutic approaches characterize MS and SAD with CNS involvement. For instance, activation of type I interferon (IFN) pathway has been considered a central pathogenetic event in SADs in association with antibodies against ribonucleoproteins, disease activity indices, and renal involvement (Baechler et al., 2003; Kirou et al., 2005; Weckerle et al., 2011; Tsokos et al., 2016; Ronnblom and Leonard, 2019; Postal et al., 2020; Idborg and Oke, 2021), with type I IFN receptor blockade being recently approved for lupus patients (Morand et al., 2020; Deeks, 2021).

Compared to lupus, MS patients display 30-fold lower type I IFN serum activity (Feng et al., 2012), while type I IFN-induced gene expression has been shown to vary among relapsing–remitting MS (RRMS) patients (van Baarsen et al., 2006). Moreover, a subgroup of MS patients with high peripheral blood type I IFN signature, was found to be refractory to exogenous administration of recombinant IFN-β (Comabella et al., 2009), a well-established MS treatment (Author anonymous, 1993; Kieseier, 2011). Along the same lines, the latter has been shown to exacerbate lupus or other SADs (Borg and Isenberg, 2007; Airo et al., 2008; Bahri et al., 2012).

In view of these diagnostic and therapeutic challenges, in the current study, we sought to explore whether clinical, laboratory, and imaging characteristics as well as type I IFN inducible gene expression in peripheral blood could differentiate MS from systemic autoimmune diseases presenting with CNS involvement. Furthermore, we sought to investigate potential associations between activation of type I IFN pathway with distinct clinical, serological, and imaging features.

In the present study, we enrolled consecutive, undiagnosed, and treatment-naive patients with MS-like disease presentation, with the aim to explore whether distinct clinical, serological, CSF, and imaging features, as well as baseline peripheral blood type I IFN activity could aid in the formulation of a final diagnosis after 1 year of follow-up. Out of the 193 patients, included in our cohort, two-thirds were classified as MS spectrum disease and one-fourth as CNS autoimmune group. Among the latter, one-fourth met the criteria for SAD (predominantly SS and SLE), while the rest did not fulfill SAD criteria, constituting the DAF subgroup. In comparison to MS spectrum, CNS autoimmune patients displayed distinctive characteristics; they were older, more frequently females, more prone to suffer from hypertension and/or hyperlipidemia, and more likely to report a positive family history of autoimmunity. Cortical dysfunction as the first neurological presentation, atypical for MS MRI lesions, as well as less frequent T2 OCBs and IgG index positivity along with higher rates of ANA titers ≥1/320 and anti-cardiolipin IgM and anti-β2GPI IgG positivity also characterized the CNS autoimmune group. Moreover, peripheral blood type I IFN score at baseline was shown to be significantly elevated only in patients fulfilling classification criteria for SAD compared to all other groups. Of note, patients with high type I IFN scores were mainly females, and had increased rates of cortical/juxtacortical/deep white matter (non-periventricular) supratentorial lesions and ANA titers ≥1/320, together with lower neutrophil counts.

Our study suggests that underlying systemic autoimmunity is not infrequent in undiagnosed patients undergoing diagnostic evaluation for possible demyelinating CNS disease. In line with previous studies, stronger female predominance, frequent sparing of the corpus callosum, and preferential involvement of cortical/juxtacortical/deep white matter (non-periventricular) regions compared to the periventricular area have been well described in the context of neuropsychiatric SLE versus MS, with the exact opposite being true for the cerebellum, brainstem, and basal ganglia (Magro Checa et al., 2013). Moreover, hypertension and dyslipidemia were more prevalent in the CNS autoimmune group in accordance with previous comparative studies reporting a pronounced risk for cardiovascular events in lupus and rheumatoid arthritis compared to MS patients (Setyawan et al., 2021). Though lower ANA titers and antiphospholipid antibody positivity have been extensively reported in MS and MS-like populations (Spadaro et al., 1999; Szmyrka-Kaczmarek et al., 2012; Malyavantham et al., 2015; Merashli et al., 2017), in the current report, we emphasize the role of higher ANA titers (more than 1/320) pointing toward an underlying systemic autoimmune process, in support of previous findings (Magro Checa et al., 2013). Moreover, in agreement with previous works, a higher CSF IgG index and the presence of OCBs reinforce the diagnosis of MS without however ruling out a diagnosis of SAD (Magro Checa et al., 2013; Govoni et al., 2016; Venhoff et al., 2019). Of interest, MS spectrum patients displayed lower B12 vitamin levels in comparison to their CNS autoimmune counterparts, possibly due to malabsorption mediated by autoimmune reactivity toward gastrointestinal antigens commonly found in MS populations (Banati et al., 2013).

Of note, lupus and primary SS were the prevailing SAD diagnoses detected in our cohort, compatible with previous observations (Delalande et al., 2004; Bertsias and Boumpas, 2010; Massara et al., 2010; Unterman et al., 2011; Borowoy et al., 2012; Schwartz et al., 2019) and anti-Ro/SSA were the predominant reactivity among patients with SAD. This is in accord with previous studies on lupus and primary SS also highlighting the role of anti-Ro/SSA as a marker of neuropsychiatric involvement (Alexander et al., 1994; Mikdashi and Handwerger, 2004; Morreale et al., 2014; Fan et al., 2021; Tetsuka et al., 2021) and disease activity in the setting of NMOSD (Lin et al., 2022).

It is also intriguing that peripheral blood type I IFN activity, consistently found to be associated with anti-Ro/SSA antibodies in both lupus and SS patients (Kirou et al., 2005; Weckerle et al., 2011; Mathian et al., 2019), was a strong predictor for SAD diagnosis among patients with presentation suspicious of CNS demyelinating disease included in the current cohort. This is consistent with the work by Yusof et al. suggesting that an IFN score based on quantitation of type I and II IFN inducible gene expression in peripheral blood along with the family history of autoimmunity was able to predict the conversion to SAD in ANA-positive individuals (El-Sherbiny et al., 2018; Md Yusof et al., 2018).

Furthermore, the demyelinating disease with autoimmune features (DAF) group, displaying a low type I IFN score in peripheral blood, may represent a distinct clinical entity deserving close follow-up and tailored diagnostic and therapeutic approaches (Mavragani et al., 2007; Nikolopoulos et al., 2021). Interestingly, a recent study, described that almost half of the patients presenting with demyelinating lesions and coexisting clinical and/or serological evidence of autoimmunity, not fulfilling criteria for either MS or SAD at baseline, will also continue to lack a well-defined diagnosis after a median follow-up period of 3 years (Nikolopoulos et al., 2021). As previously reported, specific autoantibodies such as anti-Ro/SSA can be present years prior to disease diagnosis (Arbuckle et al., 2003), while the emergence of a well-defined autoimmune disease is in parallel with an elevation of type I IFN activity in peripheral blood (Munroe et al., 2016). Therefore, the possibility of DAF switching to SAD in the future is thought to be substantial.

While nonspecific brain WMH, as well as hypertension and dyslipidemia, were significantly higher in the CNS autoimmune group with respect to MS, none of these characteristics could differentiate patients with high versus low IFN activity. Although WMH is the most frequently observed MRI lesion in patients with systemic autoimmune diseases and suspected CNS involvement (Magro Checa et al., 2013; Magro-Checa et al., 2018), our CNS autoimmune group mainly includes the MS-like disease patients with autoimmune features (DAF group), in which type I IFN was not upregulated in contrast to individuals meeting classification criteria for SAD. Since IFN signature surfaces in parallel with clinically overt disease (Munroe et al., 2016), one could hypothesize that WMH represents an early sign of systemic autoimmunity prior to type I IFN pathway activation and full-blown disease manifestation.

Though activation of type I IFN pathway has been postulated as a major effector mechanism for several clinical manifestations in patients with lupus and SS (Mavragani, 2017; Postal et al., 2020), there is limited data supporting a role in the pathogenesis of neuropsychiatric manifestations in both serum (Fragoso-Loyo et al., 2012) and gene expression studies (Baechler et al., 2003). Of interest, variations of three prime repair exonuclease 1 (TREX1) genes–an exonuclease involved in the clearance of endogenous nucleic acids (Stetson et al., 2008) have been previously detected in European SLE cases with neurological involvement (Namjou et al., 2011). Mutations of the TREX1 gene are a hallmark of Aicardi–Goutières syndrome (AGS) (Crow et al., 2015), characterized by a lupus reminiscent picture with high IFN-α in peripheral blood and CSF, white matter lesions, autoantibody positivity and cytopenia (Ramantani et al., 2010; Crow and Manel, 2015; Cuadrado et al., 2015; Cattalini et al., 2016). Besides, previous studies in SS and lupus patients, as well as in TREX1-deficient human neural cells revealed a role of L1 (long interspersed nuclear element-1) retroelements as major triggers of type I IFN production in both target tissues (Mavragani et al., 2018) and astrocytes, respectively (Thomas et al., 2017). Moreover, ubiquitin-specific peptidase 18 (USP18) deletion in young adult mouse models [normally downregulating signal transducer and activator of transcription 1 (STAT1) signaling], led to increased activation of downstream type I IFN signaling in white matter microglial cells causing microgliopathy (Goldmann et al., 2015). Of interest, abrogation of type I IFN signaling in the Rnaset2^−/− mice characterized by AGS-resembling leukoencephalopathy syndrome, led to complete resolution of CNS features (Kettwig et al., 2021).

In line with previous reports, associations of type I IFN signature with female sex (Panchanathan et al., 2010; Gabriele et al., 2021), neutropenia (Hall et al., 2015; Mathian et al., 2019; Chasset et al., 2020) or high ANA titers (Wither et al., 2017) were detected. Moreover, the link between non-periventricular lesions and high IFN scores might reflect the presence of an immune-mediated vascular injury as the underlying potential mechanism, as formerly suggested (Skeoch and Bruce, 2015; Spinelli et al., 2018; Filippi et al., 2019; Wardlaw et al., 2019; Sereti et al., 2020). In support of these findings, it has been previously shown that type I IFN inhibition in SLE patients blocked NET formation and restored cholesterol efflux capacity impairment, already shown to contribute to lupus-related atherosclerotic disease (Giannelou and Mavragani, 2017; Casey et al., 2021; O'Neil et al., 2019; Carlucci et al., 2018). Of note, among MS patients, IFN-β therapy was associated with dysregulated metabolic profiles and higher blood pressure (Sternberg et al., 2014).

The main strength of our study was the large number and extensive evaluation of undiagnosed, treatment-naïve individuals, presenting with an episode described as potentially demyelinating. As for limitations, quantitation of type I IFN score was performed in a subset of recruited patients and the final number of SAD patients was rather small to draw definite conclusions. Moreover, our patients were not tested for lupus anticoagulant (LAC); however, the initial aPTT measurement did not reveal time prolongation in any case (Tripodi and Chantarangkul, 2017).

Taken together, the identification of a prominent type I IFN signature among SAD patients with evidence of CNS involvement as the cardinal manifestation might imply a potentially significant therapeutic role of type I IFN receptor blockade, already licensed for the treatment of lupus (Morand et al., 2020; Deeks, 2021). Conversely, treatment with IFN-β in patients with a systemic autoimmune background might lead to disease exacerbation and therefore alternative therapeutic strategies such as B cell depletion should be considered (Evangelopoulos et al., 2017; Sabatino et al., 2019).

In conclusion, our study indicates that underlying systemic autoimmunity is a considerable possibility in patients undergoing evaluation for possible demyelinating CNS disease and distinct clinical, imaging and laboratory features can reliably differentiate CNS autoimmune from MS spectrum individuals in an everyday clinical practice setting. Moreover, activation of type I IFN pathway seems to be a key mediator among SAD patients presenting with MS-like disease and therefore a potential therapeutic target. Finally, a close follow-up is recommended for CNS autoimmune patients not yet fulfilling well-defined SAD classification criteria, given the heightened risk for future development of full-blown systemic autoimmune disease. Thorough studies exploring the molecular pathways governing the generation of diverse aspects of multifaceted CNS demyelinating disease would allow the design and institution of tailored and most appropriate therapeutic strategies.