Chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) is a heterogeneous autoimmune disease affecting peripheral nerves. Diagnosis relies on clinical and electrophysiological criteria built to identify patients that may respond to immunomodulatory treatments [1]. The recent description of antibodies targeting proteins at node of Ranvier, such as contactin-1 (CNTN1) [2], neurofascin-155 (NF155) [3], contactin-associated protein 1 (CASPR1) [4], and nodal neurofascins [5], has revealed the existence of CIDP subsets with clinical features that associate specifically to each antibody. In the case of anti-NF155, patients show predominantly distal weakness, ataxia, and a low frequency tremor [3]. They respond to intravenous immunoglobulins (IVIg) less frequently than anti-NF155neg CIDP [6], and when resistant to conventional treatment, they may respond well to B cell depleting therapies [7]. Anti-NF155 antibodies are predominantly IgG4, an isotype unable to fix complement or activate inflammatory cells [6]. Interestingly, anti-NF155 CIDP patients show distinct pathological features that differ from those of typical CIDP patients, including lack of macrophage infiltrates and a selective loss of the transverse bands at the paranodal loops [8, 9]. Thus, the clinical, pathological, and electrophysiological diversity of CIDP disappears when patients are sub-classified according to specific autoantibodies.

There is no epidemiological, environmental, or genetic evidence that explains the appearance of anti-NF155 antibodies, and evidence for the existence of CIDP-specific risk factors is lacking, including genetic association studies. Human leukocyte antigen (HLA) loci are the group of genetic factors that has most frequently been associated with autoimmune diseases, including strong associations with other IgG4-mediated diseases [10–12]. Moreover, association of HLA genes with specific antigens can help discover the specific peptide sequence driving the immune response [10]. Considering the association of other IgG4-mediated diseases to HLA class II, we aimed our study to identify specific HLA class II alleles associated to anti-NF155 antibodies in CIDP.

Our study finds a strong association of DRB1*15 alleles with anti-NF155+ CIDP compared to anti-NF155neg CIDP and general population. Specifically, anti-NF155+ CIDP patients have a DRB1*15 allele 20 times more likely than anti-NF155neg CIDP patients. This finding represents the first strong association of a CIDP subset with a specific genetic background and opens a new field =of study to investigate the interactions of the antigen presenting B cells, the T cells, and the immunological environment leading to IgG4 anti-NF155 antibody production.

Association of specific HLA alleles with autoimmune diseases is one of the most frequent topics of research in translational immunology and is used in several diseases (Behçet disease, ankylosing spondylitis, and others) as a diagnostic biomarker. Recent studies have found strong associations of certain HLA alleles in diverse tissue-specific IgG4-mediated diseases, such as LGI1 encephalitis (DRB1*07:01–DQB1*02:02 haplotype) [10], anti-MusK myasthenia (DR14-DQ5 haplotype) [19], or anti-IgLON5 disease (DRB1*10:01-DQB1*05:01) [11]. Early studies in CIDP did not find any association with specific HLA alleles, including studies addressing specifically the DRB1 frequencies in chronic autoimmune neuropathies. Other studies describe a weak association of specific HLA class I alleles with CIDP in some populations [20]. Considering that the effector mechanism in anti-NF155+ CIDP is B cell mediated, we focused in HLA class II associations, instead of HLA class I. Also, although this study did not aim to study specifically the association of HLA class II with anti-NF155neg CIDP, we did not find any difference in class II allele frequencies compared to general population.

DRB1*15 was not present in three (out of 13) anti-NF155+ CIDP patients, suggesting that HLA DRB1*15 is a strong risk factor but probably not a necessary factor for the development of anti-NF155 antibodies. This agrees with results in other IgG4-mediated diseases in which the associated HLA alleles are not universally found in all patients but with odds ratios ranging 8 to 100 and frequencies ranging 50 to 90% of patients [10, 19]. The DRB1 alleles found in those three anti-NF155+ CIDP, but DRB1*15 negative patients were also present in controls at similar frequencies; it is unlikely that they influenced their anti-NF155 antibody development.

The two DRB1*15 alleles found in CIDP patients (DRB1*15:01 and DRB1*15:02) reach statistical significance for their association with anti-NF155 antibodies independently. However, it is likely that the association with two different DRB1*15 alleles is not influencing the underlying peptide presentation process. DRB1*15:01 and DRB1*15:02 only differ in one amino acid (codon 86). Moreover, in silico experiments predicted common peptides binding to both alleles, suggesting that probably the same (or a very similar) peptide of NF155 could be presented with similar efficiency in both DRB1*15 alleles.

A limitation of our study is that, due to the low frequency of the disease, patients from four different origins (Spain, France, Italy, and UK) were included, while the statistical analysis where performed with Spanish controls. HLA-DRB1*15 is found in 17% of Spanish population, 12% of North Italy population, 25% of Southern France population, and 20% of English population (www.allelefrequencies.net). Considering the country of origin of the anti-NF155+ CIDP (7 Spain, 3 Italy, 2 France, and 1 England), these differences would not change the fact that anti-NF155+ CIDP patients show very high frequencies of DRB1*15 alleles compared to control populations, regardless of the origin of the control population used as reference.

Interestingly, DRB1*15 is also associated with multifocal motor neuropathy [21], although anti-NF155 are absent in this disease [22]; DRB1*15:01 also associates with multiple sclerosis [23], in which some studies described antibodies against NF155 associating to specific MS subtypes [24]. In fact, anti-NF155 antibodies have been described in patients with combined central and peripheral demyelination, but these findings have not been replicated in independent cohorts [25, 26]. None of our patients presented with symptoms or signs in the neurological exam suggesting central demyelinating disease. Moreover, 9 out of 13 anti-NF155+ patients had, at least, a brain MRI available and none of them showed central demyelinating lesions. Although DRB1*15 alleles have also been associated to other autoimmune diseases, including the IgG4-mediated anti-PLA2R membranous nephropathy [27], the association of DRB1*15 with three different diseases (multiple sclerosis, multifocal motor neuropathy, and anti-NF155+ CIDP) in which myelinating cells (oligodendrocytes and Schwann cells) are targeted by immune system may suggest convergent pathophysiological mechanisms in which myelin antigen presentation with DRB1*15 HLA class II is central. Some reports suggest that Schwann cells express HLA-DR in their surface, particularly in inflammatory neuropathies [28], but whether DRB1*15 alleles are expressed more efficiently in myelinating cells in comparison with other DRB1 alleles has never been studied.

DRB1*15 alleles strongly associate to anti-NF155+ CIDP providing genetic evidence to support the idea that the clinical, electrophysiological, pathological, and genetic heterogeneity of CIDP disappears when the disease is segregated according to highly specific biomarkers like anti-NF155 antibodies.