Membrane Proteome-Wide Screening of Autoantibodies in CIDP Using Human Cell Microarray Technology

Abstract

Background and objectives: Autoantibody discovery in complex autoimmune diseases is challenging. Diverse successful antigen identification strategies are available, but, so far, have often been unsuccessful, especially in the discovery of protein antigens in which conformational and post-translational modification are critical. Our study assesses the utility of a human membrane and secreted protein microarray technology to detect autoantibodies in chronic inflammatory demyelinating polyradiculoneuropathy (CIDP).

Methods: A cell microarray consisting of human embryonic kidney-293 cells expressing >5,000 human proteins was used. First, a validation step was performed with 4 serum samples from patients with autoimmune nodopathy (AN) to assess the ability of this technology to detect circulating known autoantibodies. The ability of the cell microarray technology to discover novel IgG autoantibodies was assessed incubating the array with 8 CIDP serum samples. Identified autoantibodies were subsequently validated using cell-based assays (CBAs), ELISA, and/or tissue immunohistochemistry and analyzed in a cohort of CIDP and AN (n = 96) and control (n = 100) samples.

Results: Serum anti-contactin-1 and anti-neurofascin-155 were detected by the human cell microarray technology. Nine potentially relevant antigens were found in patients with CIDP without other detectable antibodies; confirmation was possible in six of them: ephrin type-A receptor 7 (EPHA7); potassium-transporting ATPase alpha chain 1 and subunit beta (ATP4A/4B); leukemia-inhibitory factor (LIF); and interferon lambda 1, 2, and 3 (IFNL1, IFNL2, IFNL3). Anti-ATP4A/4B and anti-EPHA7 antibodies were detected in patients and controls and considered unrelated to CIDP. Both anti-LIF and anti-IFNL antibodies were found in the same 2 patients and were not detected in any control. Both patients showed the same staining pattern against myelinating fibers of peripheral nerve tissue and of myelinating neuron-Schwann cell cocultures. Clinically relevant correlations could not be established for anti-LIF and anti-IFNL3 antibodies.

Discussion: Our work demonstrates the utility of human cell microarray technology to detect known and discover unknown autoantibodies in human serum samples. Despite potential CIDP-associated autoantibodies (anti-LIF and anti-IFNL3) being identified, their clinical and pathogenic relevance needs to be elucidated in bigger cohorts.

Figure 1. Prescreen of Autoimmune Nodopathy Samples With a Known Antibody

Prescreen of anti-NF155 (A) and anti-CNTN1 (B) CIDP serum samples against HEK293 cells overexpressing neurofascin (isoform 9 and 11), CNTN1, and Caspr1 and CD20 and EGFR. Screening of rituximab biosimilar (anti-CD20; C) was performed as a positive control and transfection control (D). CIDP = chronic inflammatory demyelinating polyradiculoneuropathy.

Figure 2. Immunocytochemistry of Anti-ATP4A/4B Antibodies

HEK293 cells were cotransfected with myc-DDK–tagged vectors ATP4A and ATP4B, double-stained with serum IgG (A, C) and with commercial antibody against c-myc (B, D). Patient 3 IgG binds to transfected cells (A), and colocalizes with c-myc (B); in contrast with the healthy control (D), that does not show any reactivity against ATP4A/4B-transfected cells.

Figure 3. Anti-ATP4/4B–Positive Patient Reactivity Against Stomach Tissue

Mouse stomach preparation stained with patient 3 serum in IgG (A) and healthy control (B). Patient 3 with anti-ATP4A/4B antibodies reacts against gastric parietal cells (A) while a healthy control negative for anti-ATP4A/4B does not (B).

Figure 4. Immunocytochemistry of Anti-LIF Antibodies

HEK293 LIF-transfected cells, double-stained with serum (A, C) and with commercial antibody against LIF (B, D). Patient 7 IgG binds to transfected cells (A), and colocalizes with LIF (B); in contrast with the healthy control (D), that does not show any reactivity against LIF-transfected cells. LIF = leukemia inhibitory factor.

Figure 5. Anti-LIF+ Patients' Reactivity Against Peripheral Nerve Tissue

Macaque peripheral nerve sections double stained in green with anti-LIF+ patient I serum in IgG (A), anti-LIF+ patient II (D), and healthy control (G) and in red with S100 (B, E, H). LIF = leukemia-inhibitory factor.

Figure 6. Anti-LIF+ Patients' Reactivity Against Myelinating Neuron-Schwann Cell Cocultures

Neuron-Schwann cell cocultures triple stained in green with anti-LIF+ patient I serum in IgG (A), anti-LIF+ patient II (D), and healthy control (G) and in red with anti-Caspr1 and in purple with anti–myelin basic protein (MBP) (B, E, H). LIF = leukemia-inhibitory factor.