Synapsin autoantibodies during pregnancy are associated with fetal abnormalities

Abstract

Anti-neuronal autoantibodies can be transplacentally transferred during pregnancy and may cause detrimental effects on fetal development. It is unclear whether autoantibodies against synapsin-I, one of the most abundant synaptic proteins, are associated with developmental abnormalities in humans. We recruited a cohort of 263 pregnant women and detected serum synapsin-I IgG autoantibodies in 13.3% using cell-based assays. Seropositivity was strongly associated with abnormalities of fetal development including structural defects, intrauterine growth retardation, amniotic fluid disorders and neuropsychiatric developmental diseases in previous children (odds ratios of 3-6.5). Autoantibodies reached the fetal circulation and were mainly of IgG1/IgG3 subclasses. They bound to conformational and linear synapsin-I epitopes, five distinct epitopes were identified using peptide microarrays. The findings indicate that synapsin-I autoantibodies may be clinically useful biomarkers or even directly participate in the disease process of neurodevelopmental disorders, thus being potentially amenable to antibody-targeting interventional strategies in the future.

Keywords: Antineuronal autoantibodies; Growth retardation; Maternofetal autoimmunity; Peptide microarray; Synapsin-I; Transplacental transfer.

Fig. 1

High prevalence of synapsin-I autoantibodies in pregnant women. (A) Representative examples of immunofluorescence stainings with human sera at 1:300 dilution (green) with (top row) or without binding (bottom row) to HEK293 cells overexpressing human synapsin-Ib. Protein expression is confirmed with a commercial synapsin-I/II antibody (red). Nuclei are stained with DAPI (blue). Scale bar = 20 μm. (B) Frequencies of synapsin-Ib autoantibodies in pregnant women and control cohorts, as determined by CBA. CBA scores: 0 = no binding, 1 = unspecific signal, 2 = intensive binding (positive); vertical dotted line represents cut-off for positivity. (C) Representative immunoblots of wild type (wt) and Syn1/2/3 triple knock out (TKO) mice cortex homogenates with a commercial synapsin-I/II antibody as positive control and sera from CBA-positive pregnant women (1:200 dilution). The strong band at ∼90 kDa corresponding to the molecular weight of synapsin-Ia/Ib was detected in wild type but not in TKO mouse tissue. Another band at ∼50 kDa likely represents the synapsin-IIb isoform or breakdown products of synapsin-I. Detection of GAPDH served as loading control. (D) Immunofluorescence staining on rat hippocampal neurons demonstrated strong IgG binding of human sera (green) in a characteristic synaptic pattern (insert shows high magnification, MAP2 [blue] staining for better visualization of neuronal processes). The punctate staining completely overlapped with a commercial synapsin antibody (red, insert for high magnification).

Fig. 2

Strong, titer-dependent association of synapsin-I autoantibodies with abnormalities of fetal development during pregnancy. (A-B) Several ultrasound parameters and status of having previous children with neuropsychiatric disorders were significantly associated with the presence of synapsin-I autoantibodies in serum of pregnant mothers (A), with odds ratios of 3–7 (B). Statistical analysis used Fisher's exact test when expected frequencies were <5 or Chi-square test with expected frequencies ≥5, * representing p < 0.05, **p < 0.01 and ***p < 0.001. (C) Synapsin autoantibodies reached the fetal circulation and were present with similar titers (±1 titer step, in one case 2 titer steps) in umbilical cord blood. (D) Synapsin autoantibodies were of IgG1 and/or IgG3 subclass in almost all pregnant women. (E) While the autoantibody titer alone did not discriminate between pathological and normal findings during fetal development (left), the combination of IgG1 subclass with high-level titers (≥1:30,000) was significantly associated with abnormal measurements (right).

Fig. 3

Detection of synapsin-I autoantibodies by ELISA, immunoblotting and microarray-based epitope mapping. (A) ELISA-based screening of all human sera using recombinant synapsin-I correlated with autoantibody binding in the CBA. (B) Immunoblots of synapsin-Ib-transfected HEK293 cells confirm that some human sera (1:200 dilution) strongly detect the synapsin-Ib band (middle), while in others binding to non-conformational epitopes is lost (right). Commercial synapsin-I/II antibody served as positive control (left, GAPDH is exemplarily shown; UT = untransfected cells). (C) An overlapping peptide library containing 20-mer peptides with an off-set of three amino acids was designed to fully cover synapsin-I primary sequence. Microarrays were probed with autoantibody-positive and -negative serum. The representative example shows detection of IgG epitope 2 within the synapsin-I slides and a representative negative control serum. Heat map visualization: IgG binding intensities are displayed in grayscale (0 = no binding, 1 = maximal binding). The core shared motif is highlighted in red. (D) Synapsin-I domain architecture together with a cartoon representation of the alpha-fold model of synapsin-I visualized in Pymol. Mapped synapsin-I autoantibody epitopes (1–5, marked in red) are located in domain A-B (epitope 1, peptide sequence: ²⁸PQPPPPPPGAH³⁸), domain C (epitope 2, ¹²⁴TDWAKYFKGKK¹³⁴) and domain D (epitope 3: ⁴⁷²PGPQRQGPPLQ⁴⁸²; 4: ⁵⁸³GGQQRQGPPQK⁵⁹²; 5: ⁶¹⁰VPRTGPPTTQQPRP⁶²³).