Fig 2: Representative image of neuronal pentraxins along the axons. Mouse primary cortical neurons were immunostained for MAP2, Syp, and either NP1 or NP2. On the representative images, synaptic NPs are clearly apparent, whereas synapses without NP-staining are also visible. Solid circles show synaptophysin signals colocalized with either NP1 or NP2, while dashed circles show synapses, which do not contain the neuronal pentraxins.

Fig 3: Co-occurrence of the C1q-tag with NP1/2 on synaptosomes examined by flow cytometry. According to the gating criteria, C1q-labeled synaptosomes were almost exclusively (~97%) positive for NP1 (A) and NP2 (B). In contrast to C1q labeling, only ~24% of NP-labeled synaptosomes were positive for C1q. Density plots show representative measurements where the percentages of synaptosomes that belong to each quadrant were also indicated. The secondary antibody controls went through the same procedure as the fully labeled samples. Statistically significant differences were determined with two-tailed Student’s t-test of paired samples (Means ± SEM; n = 4 mice). To test the viability of synaptosomes, calcein-AM labeling was applied separately ( Supplementary Datasheet 2 ).

Fig 4: Colocalization of synaptic C1q with NPs in mouse brain sections. (A, B) High-resolution confocal microscopy images of triple immunostained cerebral cortical sections were segmented to identify individual C1q, NP1/2, and Syp spots. White circles indicate several triple-colocalizing spots automatically identified according to the predefined criterion. Analyses demonstrated extensive colocalization of synaptic C1q with synaptic NPs, particularly with NP1. (T-test results: (NP1) P = 2,19007E-10, (NP2) P = 4,71338E-06 on normally distributed samples.) (C, D) Close proximity of synaptic C1q with NP1 (C) and NP2 (D) is not the mere consequence of their high abundance. Medians of the minimal center-to-center distance’s cumulative frequency distributions significantly differ between the observed and randomly shuffled samples (P = 0.00019, Wilcoxon signed-rank test). Means ± SEM are shown; n = 18 3D-images recorded from sections of three mice. Scale bar = 0.5 µm. (E) The lateral view of the colocalized proteins.

Fig 5: Interaction of neuronal pentraxins with C1q and C1 in vitro and activation of the complement classical pathway. (A) In an ELISA assay, NP1, NP2, and the positive control PTX3 were coated, and then, wells were incubated with purified C1q, and binding was detected with anti-C1q antibody. C1q showed direct interaction with both NP1 and NP2 in a dose-dependent manner, comparable to PTX3. Means ± SEM; n = 5. (B) Assay results with C1 complex immobilized. Recombinant NPs were detected with anti-His antibody and exhibited dose-dependent binding (representative experiment, Means ± SEM; n = 4.) Plots in (A, B) were fitted with hyperbolic function. (C) NP-dependent activation of the complement classical pathway was examined using microtiter plate binding assay. Neuronal pentraxins were able to activate the cascade in the presence of 2% serum, similarly to the positive controls (IgG and PTX3), as indicated by the deposition of C4 fragments, detected with anti-C4 antibody. In C1q-depleted serum, activation of the classical pathway by any of the investigated activators was dropped to baseline level. Supplementation of C1q-depleted 2% serum with purified C1q in the physiological concentration range (1.4 µg/ml) restored the activation. Columns represent the overall means ± SEM. Three biological replicates (each is a mean of two technical parallels) are also presented with hollow circles. The statistical analysis was two-sample t-test on normally distributed samples compared to the C1q-depleted respective samples (*p < 0.05, **p < 0.01, n = 3 biological replicates with two technical parallels).