Astrocytes modulate neuronal development by S100A6 signaling

Cinquina, Valentina; Tretiakov, Evgenii O.; Kalaba, Predrag; Alpár, Alán; Calvigioni, Daniela; Piscitelli, Fabiana; Keimpema, Erik; Di Marzo, Vincenzo; Verkhratsky, Alexej; Harkany, Tibor

doi:10.1038/s41467-025-64405-y

Download PDF

Article
Open access
Published: 13 October 2025

Astrocytes modulate neuronal development by S100A6 signaling

Nature Communications volume 16, Article number: 9049 (2025) Cite this article

Subjects

Abstract

Neuronal morphogenesis relies on intercellular signaling. Astrocytes release metabolites, trophic, and guidance factors to promote neuronal maturation. In contrast, the mechanisms by which astrocytes could limit and stabilize neuronal connectivity remain less explored. Here, we find cortical astrocytes to express and release S100A6, a Ca²⁺-binding protein (‘calcyclin’). Simultaneously, the majority of cortical neurons expressed calcyclin-binding protein (CaCyBp), a bona fide binding partner for S100A6. In neurons, CaCyBp maintains the unfolded protein response pathway, thereby controlling proteostasis. When released, S100A6 inhibits CaCyBp-mediated signaling, thus slowing protein turnover, and, consequently, neuritogenesis. In the cerebral cortex of male mice, S100A6-CaCyBp signaling during gestation is sensitive to the mother’s nutritional status, particularly eicosapentaenoic acid intake. Thus, a member of the S100 protein family acts as an astroglia-derived morphogen, whose action on neurons is modulated by environmental factors.

Introduction

Reciprocal communication with astrocytes underpins the development, survival, and excitability of neurons. Therein, astrocytes release gliotransmitters (such as canonical neurotransmitters, adenosine/ATP, and trophic factors), neurotransmitter precursors (glutamine, L-serine), and metabolites (chiefly lactate and pyruvate), and also clear excess neurotransmitters and waste products, including potential neurotoxins¹. During brain development, astrogliogenesis coincides with synaptogenesis¹. Astrocytes are broadly known to express and release differentiation and survival factors (e.g., brain-derived neurotrophic factor, glial cell line-derived neurotrophic factor, thrombospondin, hevin/Sparcl1) to promote the polarization, morphological differentiation, and synaptogenesis of neuroblasts¹. In contrast, astrocyte-derived factors that are inhibitory to neuritogenesis, thus acting to balance growth vs. synapse selection and stabilization, remain less explored.

Considering that major steps of cellular differentiation are regulated by Ca²⁺ signals (be this the activation of kinases, cytoskeletal instability, directional growth decisions, synaptogenesis)¹, Ca²⁺-binding and sensor proteins appear particularly important, considering that some can even undergo regulated release to support intercellular interactions². In astrocytes, S100β, a Ca²⁺-binding protein within the EF-hand superfamily, is best known as one of the 25 members of the S100 protein family² because it is constitutively expressed in most astrocytes across brain regions^3,4, and associated with inflammation in adults^3,5. However, if S100 proteins are expressed in astrocytes during brain development and could be assigned to regulate neuronal morphogenesis remains unknown.

Here, we explored the developmental expression patterns of S100 family proteins by processing open-label single-cell RNA-seq data during mouse fetal brain development, and identified S100A6/calcyclin, its binding partner (calcyclin-binding protein; CaCyBp), and the main downstream targets Siah1/Siah2, E3 ubiquitin ligases, with precisely timed expression coinciding with neuritogenesis and synaptogenesis in the late embryonic/neonatal period in mice. We then used in situ hybridization and immunohistochemistry to map S100A6 and CaCyBp to astrocytes and neurons, respectively, identified S100A6 as a secreted ligand by mass spectrometry, protein biochemistry, and neurochemistry in cellular models in vitro, and combined genetic and pharmacological tools to demonstrate that S100A6 could inhibit neuritogenesis, likely by slowing protein turnover. Notably, we found both S100A6 and CaCyBp expression to be sensitive in the developing fetal nervous system to nutritional imbalance, particularly to eicosatetraenoic acid (EPA) availability. Subsequently, we have associated the deregulation of S100A6-CaCyBp signaling between astrocytes and neurons with maternal exposure to hypercaloric/ω3-polyunsaturated fatty acid (PUFA)-enriched diets in male mice. Cumulatively, these data suggest that S100A6, once released, affects neuronal morphogenesis, at least in the mouse brain.

Results

S100 proteins in the developing mouse brain

First, we addressed the expression sites and developmental dynamics of S100 proteins in the mouse brain by processing open-label single-cell RNA-seq data on the fetal somatosensory cortex⁶. During corticogenesis (that is, from embryonic day (E)10 to postnatal day (P)4), cell identities were mapped in age-matched tissues by UMAP representation (Fig. 1a–d). Progenitors, as well as astrocytes, were subdivided into 16 clusters based on marker gene expression (see “Methods”; Fig. 1d). Fifteen of the S100 proteins were detected from E10.5 on (Fig. 1e). Four of the proteins (S100A1, S100A6, S100A13, and S100A16) gradually increased in expression in astrocytes (up to P4). Among these, S100a6 (also termed “calcyclin”) showed a sharp increase between E18.5 and P4, when astrogliogenesis peaks^1,6 (Fig. 1e). Detailed examination of the gene sets conferring cell identity in the single-cell RNA-seq data showed that S100A6 was chiefly expressed by astrocytes in neonatal and postnatal mouse cortices (Fig. 1f, g and Supplementary Fig. 1a). As such, UMAP representation revealed that S100a6 expression matched with astrocytes also enriched in e.g., S100b⁷, Apoe⁸, and Aldh1l1^9,10 (Supplementary Fig. 1b). qPCR and Western blotting validated the presence of S100a6 mRNA (n = 4–7 male mice/time point, p = 0.003 for E14.5 vs. P4, one-way ANOVA followed by Tukey’s multiple comparisons test) and protein expression (n = 3 male mice/time point; p = 0.017 for E18.5 vs. P4, Student’s t-test), respectively, during development (Supplementary Fig. 1c, d). These data suggest that S100A6 could control a specific event during brain development, noting that its expression can be maintained in mature astrocytes postnatally¹¹.

**Fig. 1: S100A6 and CaCyBp in the developing mouse brain.**

S100A6 and calcyclin-binding protein localization

S100A6 promotes cell proliferation and differentiation, and protects from noxious insults^12,13. S100A6 contains two EF-hand Ca²⁺-binding motifs, which are helix-loop-helix structural domains (Supplementary Fig. 1e). Upon Ca²⁺ binding, S100A6 undergoes a conformational change that exposes a hydrophobic region to allow for its interaction with target proteins. S100A6 binds to the C-terminus of calcyclin-binding protein (Cacybp or S100a6bp) to promote its interaction with Siah1a/Siah1b, E3 ubiquitin ligases, through its N-terminal domain (Supplementary Fig. 1f). Thus, S100A6, if released from, e.g., astrocytes, could regulate protein degradation/turnover in neighboring cells, including neurons, alike in other tissues^14,15, and malignant cells⁷.

Single-cell RNA-seq placed Cacybp in both apical progenitors and postmitotic neurons (Fig. 1e, h, i and Supplementary Fig. 1g). When testing the availability of Cacybp in neurons, we found it ubiquitously expressed at E10.5, with stable levels from E14.5 to P4 (Fig. 1e, h and Supplementary Fig. 1g). Siah1a and Siah1b were also detected by single-cell RNA-seq (Fig. 1e). At E18.5, when S100a6 expression begins to increase in the astrocyte lineage (Fig. 1e, g), UMAP assigned Cacybp expression to virtually all neuronal lineages in the mouse cortex (note its limited expression in astrocytes for cell-autonomous function; Fig. 1i). CaCyBp expression at both the mRNA and protein levels was confirmed by qPCR (Supplementary Fig. 1h; n = 5–7 male mice/time point, p = 0.009 for E14.5 vs. P4, one-way ANOVA followed by Tukey’s multiple comparisons test) and Western blotting (Supplementary Fig. 1i; n = 3 male mice/time point, p = 0.040 for E18.5 vs. P4, Student’s t-test). Both methods demonstrated stable levels of Cacybp mRNA and CaCyBp protein during development, followed by a significant increase postnatally.

Complementarity of S100A6 and calcyclin-binding protein expression

To map S100A6 and CaCyBp in brain tissue, we performed in situ hybridization experiments on coronal sections of E18.5 and P4 mouse brains (n = 3 male mice/time point) using probes specific for these markers (see “Methods”). In the somatosensory cortex (ssCtx; Fig. 2a), co-localization between S100a6 and S100b⁷, Apoe⁸, Aldh1l1^9,10, and Gfap mRNAs, pleiotropic markers of astrocytes, was found in the cortical plate at both E18.5 and P4 (Fig. 2a, d and Supplementary Fig. 2).

During the same developmental window, Cacybp expression was mostly confined to cells that co-expressed the neuronal marker Rbfox3 (which encodes NeuN; Fig. 2b, e and Supplementary Fig. 3a, b; n = 3 male mice/time point throughout). The neuronal identity of Cacybp⁺ cells was further confirmed by immunohistochemistry, showing CaCyBp expression in both excitatory and inhibitory neurons in the ssCtx of mice at both E18.5 and P4 (Fig. 2c, f and Supplementary Fig. 3c, d). At E18.5, 96.3% of the cells that populated the ssCtx co-expressed both CaCyBp and microtubule-associated protein (MAP2), a neuronal marker. Moreover, 17.0% of neurons were positive for both CaCyBp and calretinin (CR, Fig. 2cupper; Supplementary Fig. 3c; n = 3 male mice/time point). Similar results were found at P4, when 90.8% of cells co-labelled for CaCyBp⁺/MAP2⁺ and 19.1% for CaCyBp⁺/CR⁺ (Fig. 2cbottom; Supplementary Fig. 3d; n = 3 male mice/time point). Likewise, CaCyBp was observed in the somatodendritic axis of neurons in the cortical plate of human fetuses during the second trimester (Fig. 3a; gestational week (GW) 27; male subject). Similar results were obtained in the adult cerebral cortex of both mice (Supplementary Fig. 3e; P120; male) and humans (Supplementary Fig. 3f; age: 60 years, male subject).

**Fig. 3: CaCyBp in human fetal brain and its complementarity with S100A6 in vitro.**

We then used primary cell cultures enriched in either neurons or astrocytes to support the cell-type-specific enrichment of both S100a6 and Cacybp (Supplementary Fig. 3g). The S100a6:Cacybp mRNA (Fig. 3b) and protein (Fig. 3c) ratios were ∼15:1 in astrocytes, whereas ∼1:18 in neurons. Thus, we reveal the mutually exclusive expression of S100a6 and Cacybp in astrocytes vs. neurons during corticogenesis, and suggest that intercellular S100A6-to-CaCyBp signaling could constitute a mode of communication if S100A6 is indeed a released substance.

S100A6 is released from astrocytes

For S100A6 to be classified as a ligand released from astrocytes, among others cell types, its expression shall be regulated (e.g., by neurotransmitters and neuromodulators) and undergo activity-dependent release. We tested these properties in vitro, aided by IncuCyte-assisted live-cell imaging to control cell viability, by exposing primary astrocytes to glutamate (100 µM) for up to 24 h. Glutamate triggered S100A6 accumulation in the culture medium, with a ∼fourfold increase by 24 h (Fig. 4a; n = 3 biological replicates; mixed for sex; p = 0.044 [0 h] vs. [6 h]; p < 0.0001 [0 h] vs. [24 h], p = 0.001 [6 h] vs. [24 h]; one-way ANOVA followed by Tukey’s multiple comparisons). We propose this to be regulated S100A6 release given the lack of cell death (Fig. 4b and Supplementary Fig. 4a; see “Methods” for the 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay, and statistics), and the parallel (albeit disproportionate) increase in cytosolic S100A6 levels (within 6 h; Supplementary Fig. 4b; p = 0.021 [0 h] vs. [2 h]; p = 0.012 [0 h] vs. [6 h]; one-way ANOVA followed by Bonferroni’s multiple comparison; n = 6 biological replicates, mixed for sex).

**Fig. 4: Astrocyte-derived S100A6 limits neuritogenesis.**

Next, we asked if S100A6 could bind to CaCyBp by exposing Neuro-2a cells, which were invariably CaCyBp⁺ (Supplementary Fig. 4c), to His-tagged recombinant S100A6 (1 μg/μl; “pulse-chase experiment”). After 1 h, His-S100A6 was detected intracellularly in Neuro-2a cells by super-resolution microscopy (Fig. 4c). LC-MS/MS analysis supported these findings by detecting trypsin-cleaved peptide fragments specific to S100A6 in treated but not control Neuro-2a cultures (Fig. 4d and Supplementary Fig. 4d, e). Overall, these findings suggest a transcellular action analogous to those of trophic factors and morphogens¹⁶, wherein astrocytes could serve as a potential source of S100A6 regulating neuronal responses by CaCyBp-based ligand recognition¹⁷.

S100A6 limits neuronal morphogenesis

We tested this hypothesis by exposing cortical neurons to recombinant S100A6 (1 μg/μl) for 4 days, which significantly reduced neurite complexity (Fig. 4e; n = 3 biological replicates; mixed for sex; p = 0.003; Student’s t-test). This morphogenetic response coincided with reduced CaCyBp protein levels (Fig. 4f; n = 3 biological replicates; mixed for sex; p = 0.005; Student’s t-test), which we qualified as a desensitization response. We then knocked down Cacybp expression in primary cortical neurons by AAV-mediated shRNA transduction. Attenuated Cacybp expression (at 15 days in vitro; Supplementary Fig. 4f—h) reduced neurite outgrowth relative to mock-transfected controls (Fig. 4g; n = 23-27 cells; mixed for sex; p < 0.001 vs. mock-transfected controls; Student’s t-test).

We then combined Cacybp knock-down with exposure to recombinant S100A6 (1 μg/μl) 24 h later. qPCR analysis showed an additive effect on both Tubb3 (Fig. 4h; n = 5–8; mixed for sex; p = 0.045 [shCacybp] vs. [control]; p = 0.048 [His-S100A6] vs. [control]; p = 0.002 [shCacybp + his-S100A6] vs. [control]; one-way ANOVA followed by Tukey’s multiple comparison; ANOVA summary: F = 6.757; p = 0.002) and Actb (Fig. 4h; n = 9–7; mixed for sex; p = 0.027 [shCacybp] vs. [control]; p = 0.008 [His-S100A6] vs. [control]; p = 0.005 [shCacybp + his-S100A6] vs. [control]; one-way ANOVA followed by Tukey’s multiple comparison; ANOVA summary: F = 6.226; p = 0.002) following CaCyBp downregulation. Moreover, immunocytochemistry followed by single-cell reconstruction showed a reduction in neurite length (including the longest one designated as the putative “axon”) after treatment (Fig. 4j; “axon“: ANOVA summary: F = 11.88; p = 0.0004; p < 0.001 [shCacybp] vs. [control]; p = 0.003 [His-S100A6] vs. [control]; p = 0.002 [shCacybp + his-S100A6] vs. [control]; neurites: ANOVA summary: F = 7.16; p = 0.004; p = 0.010 [shCacybp] vs. [control]; p = 0.004 [His-S100A6] vs. [control]; p = 0.023 [shCacybp + his-S100A6] vs. [control]). The total number of neurites was not affected (Fig. 4i). These data identify a role for S100A6 in limiting neuritogenesis through CaCyBp^14,18 (Fig. 4k).

Neuritogenesis relies on the rapid expansion of the cytoskeleton and its enrichment in specific proteins (receptors, transporters) along the neurite axis, including the prospective synapse. Subcellular fractionation placed CaCyBp in the cytosol, in particular, associated to the endoplasmic reticulum (ER; Fig. 5a). Electron microscopy confirmed this localization with CaCyBp often found in the proximity of the cisternae of the ER of neurons (Fig. 5b and Supplementary Fig. 5a). This location suggested that CaCyBp could be poised to affect proteostasis¹⁹ (that is, protein synthesis and/or turnover at the ER).

CaCyBp modulates protein turnover in neurons

Cortical neurons after Cacybp knock-down had reduced protein content (vs. mock-transfected cells; Fig. 5c; n = 4 male mice; p = 0.005 vs. [mock-transfected control]; Student’s t-test). Therefore, we tested the expression of molecular constituents of the ER stress pathway, whose activation increases protein degradation²⁰ by serving as effectors of the unfolded protein response (UPR) machinery¹⁹: G-protein coupled receptor 78 (Grp78; p = 0.001), cyclic AMP-dependent transcription factor ATF-6 alpha (Atf6a; p = 0.008), unspliced X-box-binding protein 1 (Xbp1u; p < 0.001) and spliced X-box-binding protein 1 mRNAs (Xbp1s; p = 0.002) were all downregulated (Fig. 5d; n = 6 biological replicates; mixed for sex; Student’s t-test). This suggests the absence of a first phase (acute) cellular response to restore protein homeostasis. Besides cellular imaging (Fig. 4g—j), we used CCAAT-enhancer-binding protein homologous protein (CHOP)²¹ as a surrogate to show that attenuated Cacybp expression did not trigger pro-apoptotic signaling (Fig. 5d; p = 0.701).

Subsequently, we tested the expression of ubiquitin specific peptidase 9, X-linked (USP9X), and ubiquitin specific peptidase 7 (USP7), both acting as deubiquitinating enzymes to prevent protein degradation by the proteasome²², and thus being instrumental for cellular survival and function²³. USP9X levels were significantly decreased (Fig. 5e; p = 0.008 vs. [mock-transfected control], n = 3 biological replicates; mixed for sex; Student’s t-test). USP7 showed lesser reduction (Fig. 5e; p = 0.113 vs. [mock-transfected control]; n = 3 biological replicates). These results were compatible with increased tubulin acetylation (Fig. 5f; p < 0.001 vs. [mock-transfected control]; n = 3 biological replicates; mixed for sex; Student’s t-test), a post-translational mark of long-lived microtubules, suggesting slowed cytoskeletal dynamics^24,25. These changes were phenocopied by sustained exposure to recombinant S100A6, which similarly reduced Cacybp expression (Supplementary Fig. 5b—f).

Subsequently, we asked if treating neurons with another S100 family member, S100A4, which was also expressed by astrocytes during brain development, even if more sporadically (Fig. 1e and Supplementary Fig. 6a), could mimic the morphogenetic effects of S100A6. It is noteworthy that a binding partner for S100A4 is not known, and CaCyBp does not bind S100A4²⁶. S100A4 exposure did not affect neurite development in vitro (Supplementary Fig. 6b). Similarly, the expression of neither Cacybp nor of the molecular constituents of the ER stress pathway were altered (Supplementary Fig. 6c).

Lastly, we asked if S100A6 localization within Neuro-2a cells could be affected by shRNA-mediated Cacybp knock-down. To this end, Neuro-2a cells were transfected with siRNA targeting CaCyBp for 96 h, with subsequent exposure to His-tagged S100A6 for an additional 1 h. Super-resolution microscopy (Fig. 5g) revealed S100A6 intracellularly. This is not unexpected because CaCyBp does not represent a cell-surface binding site that would limit S100A6 internalization. Yet we found reduced S100A6 localization within the ER fraction where CaCyBp was shown to accumulate (Fig. 5h). Thus, we outline an intracellular signaling cascade centered on the UPS machinery for S100A6 to limit neuronal morphogenesis (see also Fig. 4k).

S100A6 signaling is sensitive to eicosapentaenoic acid

Next, we sought to address if environmental factors, particularly components of a maternal diet that is societally seen as “healthy” could affect S100A6-CaCyBp signaling in the developing brain. We hypothesized that eicosapentaenoic acid (EPA), an ω-3 polyunsaturated fatty acid (PUFA) that is particularly enriched in experimental diets (Supplementary Table 1), could be a candidate to affect S100A6 signaling because it modulates both neurogenesis^27,28 and astrocyte differentiation²⁹ by being incorporated in the membranes of glial cells^29,30,31. However, a biological substrate for EPA is as yet unknown.

First, we tested the dose-dependence of EPA action on both neurons and astrocytes. By using high-throughput live-cell imaging of neurons for 7 DIV, we found that 1 μM but not 5 μM concentrations of EPA induced neuritogenesis in cortical neurons (including branching; Fig. 6a; n = 7 biological replicates; mixed for sex; neurite length: p < 0.0001 for both [control] vs. [EPA, 5 μM] and [EPA, 1 μM] vs. [EPA, 5 μM]; neurite branches: p < 0.0001 for both [control] vs. [EPA, 5 μM] and [EPA, 1 μM] vs. [EPA, 5 μM]; Supplementary Fig. 7a). EPA did not impact astrocyte proliferation even at an excessive dose of 30 μM after 24 h (Fig. 6b and Supplementary Fig. 7b). No effect was seen on astrocyte viability either (Supplementary Fig. 7c). Yet, EPA increased the metabolic activity of astrocytes in the MTT assay (Supplementary Fig. 7c), which we interpreted as altered mitochondrial function likely upstream to ER stress³². Then we asked if EPA could be one of the effectors to modulate S100A6 secretion from astrocytes, particularly since increased cellular metabolism could underpin release events. Thus, we exposed astrocytes to EPA (30 μM) for 24 h and we found that it induced S100A6 release in vitro (Fig. 6c; n = 3 biological replicates; mixed for sex; p < 0.0001 [0 h] vs. [24 h], p = 0.0004 [6 h] vs. [24 h]). Both 5 μM and 10 μM EPA concentrations were also effective (Supplementary Fig. 7d). Thus, we characterized S100A6 as secreted ligand possibly released upon EPA exposure, at least in vitro.

**Fig. 6: S100A6 signaling is modulated by eicosapentaenoic acid.**

Next, we performed experiments in which conditioned media of astrocytes, collected 6 h after EPA exposure (30 μM), were transferred onto neurons (Supplementary Fig. 7e). After an incubation period of 24 h, neurons treated with conditioned media had reduced Cacybp mRNA levels (Fig. 6d; n = 3 biological replicates; mixed for sex; p = 0.021; Student’s t-test), as well as a lowered expression of the molecular constituents of the ER stress pathway (Fig. 6e; n = 3 biological replicates; mixed for sex) in vitro.

Thereafter, we sequentially knocked down S100A6 expression in astrocytes by a pool of siRNAs (1 μM) for 96 h, exposed them to EPA (30 μM) for 6 h, and then collected conditioned media for transfer onto neurons (for 24 h; 7a) to specify astrocyte-derived S100A6 as a mediator of EPA-induced growth effects on neurons. In astrocytes, both Western blotting (cytosol) and ELISA (medium) revealed attenuated S100A6 expression upon EPA treatment relative to either mock-transfected controls or EPA alone (Fig. 7b–d). In neurons, standard culture medium and astrocyte-conditioned medium (conditioning for 6 h only without treatment to the astrocytes; Supplementary Fig. 8a) led to a statistically significant increase in neuronal growth (Supplementary Fig. 8b). This effect is compatible with neuronal differentiation in response to growth factors released by astrocytes. From this point on, the morphological variables determined in neurons exposed to conditioned astrocyte medium served as baseline (“control”). In addition, we used direct stimulation by EPA (30 μM; Fig. 7a) to limit bias through its direct effect as “left-over contaminant” in conditioned media after 6 h. Once depleting conditioned media in S100A6 by RNAi, neurons underwent neuritogenesis (no mock-transfected controls at this point; Fig. 7e, f; p = 0.042 [control] vs. [sRNA alone]). Moreover, a differential, and statistically significant, neuritogenic response was observed when comparing EPA alone that reduced the length of the lead neurite vs. siRNA-mediated S100a6 knock-down in conjunction with EPA stimulation (p = 0.0006 [conditioned with EPA] vs. [siRNA + conditioned with EPA]). Likewise, the cumulative length of neurites and their branching were rescued when knocking down S100a6 before conditioning astrocytes with EPA (Fig. 7g; neurite length: p = 0.0005 [conditioned with EPA] vs. [siRNA + conditioned with EPA]; branching: p = 0.0001 [control] vs. [sRNA alone], and p = 0.025 [conditioned with EPA] vs. [siRNA + conditioned with EPA]). As before, increased S100A6 levels in the conditioned medium were coupled with a reduction of Cacybp mRNA expression in neurons (Supplementary Fig. 8c). Thus, we suggest that modulating S100A6 release impinges upon neuronal differentiation through CaCyBp engagement.

**Fig. 7: S100A6 availability modulates neuronal differentiation.**

Transgenerational modulation of S100A6 signaling

EPA is a key component of “ω-3 polyunsaturated fatty acid (PUFA)-enriched” diets in humans, which are preferred during pregnancy^33,34. Therefore, we exposed female dams for 8 weeks to a hypercaloric diet enriched in ω-3 PUFAs prior to conception and throughout pregnancy (Supplementary Fig. 9a and Supplementary Table 1)³⁵ (EPA enrichment: ∼fivefold vs. control diet).

We first determined cytoarchitectonic features of fetal cortices at E18.5 male fetuses after gestational exposure to the experimental diet. Wild-type C57Bl6/J dams were pulsed with 5-ethynyl-2′-deoxyuridine (EdU) on E14.5, the peak of cortically-oriented neurogenesis³⁶, with the distribution of EdU⁺ progeny in fetal cortices quantified and plotted on E18.5 (Fig. 8a). The total number of EdU⁺ cells did not differ between the groups (183 ± 29 cells/mm² [experimental diet] vs. 301 ± 85 cells/mm² [standard diet]; n = 3/group; Supplementary Fig. 9b). In contrast, the distribution of EdU⁺ cells across cortical laminae, which were separated by their different cellular packing densities^35,37, were dissimilar, with an increased density of EdU⁺ cells at the cortical plate-subplate boundary (layer 5: 15 ± 1 cells/mm² [experimental diet] vs. 7 ± 0 cells/mm² [standard diet]; p = 0.003), as well as the subventricular zone (layer 9: 18 ± 1 cells/mm² [experimental diet] vs. 12 ± 2 cells/mm² [standard diet]; p = 0.025; Fig. 8a). Conversely, the uppermost cortical plate/marginal zone that forms first in the outside-in hierarchy of cortical lamination ³⁸ had less EdU⁺ cells (layer 2: 4 ± 1 cells/mm² [experimental diet] vs. 12 ± 5 cells/mm² [standard diet]; p = 0.0008; Fig. 8a). These data suggest the layer-specific redistribution of postmitotic neuroblasts from equivalent pools of progenitors in the brains of male fetuses when dams received the experimental diet.

**Fig. 8: Transgenerational modulation of S100A6 signaling.**

Moreover, we observed an ∼45% decrease in the expression of growth-associated protein 43 (GAP43; n = 3 male mice/group; p = 0.018 vs. [standard diet]; Student’s t-test), and ∼65% reduction in microtubule-associated protein 2 (MAP2; n = 3 male mice/group; p < 0.0001 vs. [standard diet]) along the fetal somatosensory cortex (Supplementary Fig. 9c, d), suggesting a potential dysregulation of cortical neurogenesis during embryonic development. This decrease was also validated by iTRAQ methodology²⁴ (n = 3 mice/group; mixed for sex), allowing the unbiased testing if protein contents were affected by the experimental diet. Among the 266 proteins whose levels changed in the fetal cortices at E18.5 (129 vs. 137 proteins being up or downregulated, respectively; Supplementary Table 2), both GAP43 and MAP2 were found reduced in the brains of male fetuses in a gene ontology cluster identified as “structural activity/cytoskeleton” (Fig. 8b, c). Moreover, we found 18 proteins (7%) of the cluster “proteasome” and 48 proteins (18%) of “translation, ribosome assembly, endoplasmic reticulum” changed (Fig. 8b), providing faithful coverage of the proteostasis network³⁹. Within the “proteasome” cluster, CaCyBp was particularly reduced (Fig. 8c; male mice) and correlated with increased S100A6 levels. Western blotting corroborated the iTRAQ data (Fig. 8d; n = 3 male mice/group; p = 0.003 vs. [standard diet, CaCyBp]; p < 0.0001 vs. [standard diet; S100A6]; Student’s t-test). We have also detected reduced total protein content in fetal cortices (Fig. 8e; n = 3 male mice/group; p = 0.01 vs. [standard diet]), and increased tubulin acetylation (Fig. 8f; n = 3 male mice/group; p = 0.002 vs. [standard diet]) after gestational exposure to the experimental diet. Cumulatively, these data suggest that S100A6-CaCyBp signaling is sensitive to manipulation of dietary PUFA levels during pregnancy.

Finally, we have assessed if key parameters of cortical reorganization endure into the adulthood of male offspring prenatally exposed to an experimental diet. We particularly focused on anxiety-like behaviors because of the prevalence of this endophenotype upon manipulations with dietary PUFAs³⁵. In the elevated plus-maze paradigms, we detected a significant reduction of entries in the open arms (Fig. 8g; p = 0.023). These data emphasize the enduring impact of PUFA manipulation on corticogenesis in vivo²³, with a life-long manifestation of depression-like behaviors in adult male offspring.

Discussion

Astrocytes and neurons communicate through an array of fundamental mechanisms during brain development, each playing a crucial role in forming a functional nervous system⁴⁰. As a prototypical signalling axis, brain-derived neurotrophic factor (BDNF), alike other neurotrophins and growth factors, is known to be released by astrocytes to support neuronal survival, growth, and differentiation^1,41. Ca²⁺ signalling is another essential mechanism, where astrocytes use protracted Ca²⁺ rises to release, e.g., ATP/adenosine, modulating both neuronal activity and synaptic transmission⁴².

Among the numerous signalling molecules involved, the S100 protein family plays a significant role⁴³. These Ca²⁺-binding proteins can act as sensors (that is, to recruit partner proteins intracellularly to initiate signalling cascades) to modulate, e.g., cell proliferation, differentiation, and response to injury⁴⁴. Specifically, S100A6/calcyclin influences neurodevelopment by regulating cytoskeletal dynamics and cell survival^18,45,46. Through these actions, S100A6, when focally released by, among other cells, astrocytes in an activity-dependent fashion, might facilitate the proper alignment, synaptogenesis, and wiring of neurons, ensuring the accuracy of connectivity in the developing brain.

Here, we described a molecular signaling pathway priming astrocyte-to-neuron interactions. Stimulation of astrocytes can lead to S100A6 release, at least in vitro. S100A6 then could bind CaCyBp intracellularly. Our study is limited in addressing if and how free S100A6 could be internalized, which is likely given that CaCyBp knock-down did not curtail intracellular S100A6 accumulation. Yet, we note that CyCyBp is not a cell-surface receptor but a sensor protein adjoining the ER, and henceforth transport and targeting mechanisms ought to exist for S100A6 to pass through the plasma membrane and to become enriched in the cytosol, respectively. Notwithstanding, an ensuing shift in the activity of the ubiquitin-proteasome system (UPS), a pathway fundamental for protein degradation, likely alters cellular processes that shall operate at the peak of their activity during neuronal growth and development¹⁹. Particularly, slowed protein turnover could then reduce cytoskeletal instability, leading to, e.g., reduced neuritogenesis.

S100A6-CaCyBp signaling appears to be a molecular pathway sensitive to eicosapentaenoic acid (EPA), a key ω-3 fatty acid that has anti-inflammatory and neuroprotective properties^27,28,31,47. During pregnancy, ω-3 fatty acids play a crucial role in fetal brain development^48,49. In particular, EPA aids the formation of neuronal membranes, enhances membrane fluidity, and ensures the efficient transmission of neural signals⁴⁷. EPA also influences gene expression to facilitate the growth and differentiation of neurons, including synaptogenesis, and myelination^50,51. Thus, ensuring adequate ω-3 fatty acid intake during pregnancy, including EPA, is essential for physiologically adequate fetal brain development, being the foundation for cognitive and neurological health throughout life. We found that a hypercaloric experimental diet enriched in ω-3 fatty acids altered cortical architecture in male mice (that is, the positioning and molecular make-up of neurons). This phenotype was associated with an increase in S100A6, and a concomitant reduction in CaCyBp, corroborating in vitro results. These changes could also be linked to an altered status of the UPS, alike cells either treated with S100A6 or CaCyBp silenced by RNAi. It is of note that many cytoarchitectural changes seen at late-gestation became alleviated during postnatal life, yet an anxiety-like phenotype remained. This is not entirely unexpected since polyunsaturated fatty acid (PUFA) imbalance, whether due to the dominance of ω-6 or ω-3 PUFAs³⁵, can affect exploratory behaviors in stressful environments, at least in laboratory rodents. Thus, we suggest that EPA-enriched diets can influence fetal brain development through the modulation of modes of intercellular communication, potentially imposing long-lasting modifications that can compromise resilience to environmental challenges later in postnatal life.

In conclusion, our results describe a molecular axis, which limits neuronal morphogenesis through modulating the ER/UPS pathway in mouse brain. By using S100A6 as a released ligand, astrocytes, among other cells, could define endpoints of neuritogenesis through a signaling cascade centered on the regulation of proteostasis.

Methods

Animals and ethical approval of in vivo studies

Female mice (C57Bl6/J) were housed in groups (5 mice/group) in perplex cages on a 12 h/12 h light/dark cycle (lights on at 08:00 h) and in a temperature (22 ± 2 °C) and humidity (50 ± 10%)-controlled environment. Food and water were available ad libitum. Embryos and tissues were obtained from timed matings with the day of vaginal plug considered as embryonic day (E) 0.5. Group sizes were defined as per conventions in developmental biology^35,52. Experiments on live animals conformed to the 2010/63/EU European Communities Council Directive and were approved by the Austrian Ministry of Science and Research (66.009/0277-WF/V/3b/2017). Particular effort was directed towards minimizing the number of animals used and their suffering during experiments. Isoflurane (1–5%/L/min air flow, as appropriate) was used as anesthesia in all experiments.

Single-cell RNA-sequencing data analysis

We analyzed single-cell RNA-sequencing data from developing mouse cortex spanning embryonic day (E) 10 to postnatal day (P) 4. The dataset was obtained from Ref.⁶, and accessed through the Single Cell Portal (SCP1290)⁵³. Raw count data and metadata were downloaded and processed using Seurat (v5.1.0) in R (version 4.4.0 (24/04/2024))^54,55. We chose Seurat for its comprehensive toolset for quality control, analysis, and exploration of single-cell RNA-seq data, as well as its wide adoption in the field.

Data preprocessing and quality control

The raw count matrix was loaded using the “Read10X function” in Seurat. We performed the following pre-processing steps: (i) the log1p normalized matrix was converted back to raw counts by applying “expm1”; (ii) scaling factors were calculated based on the total UMI counts/cell. The count matrix was scaled by multiplying each cell’s counts by its scaling factor; (iii) a new Seurat object was created using the scaled count matrix; (iv) cells annotated as doublets, low quality, or red blood cells were removed using the “subset” function; (v) data were then normalized using the “NormalizeData” function, and 5000 highly variable features were identified using “FindVariableFeatures”.

Dimensionality reduction and clustering

We performed principal component analysis (PCA) on variable features using “RunPCA”. Based on the Elbow plot, which indicates the variability of each principal component (PC), we selected the first 30 out of 50 PCs for downstream analysis. This choice helped us to reduce noise in the data while ensuring biological reproducibility. Uniform Manifold Approximation and Projection (UMAP)⁵⁶ and t-distributed Stochastic Neighbor Embedding (t-SNE)^57,58 were used for dimensionality reduction, with embeddings stored in the Seurat object. Both techniques used the selected 30 PCs as input. Cells were clustered using the “FindNeighbors” and “FindClusters” functions. For community detection, we employed the Leiden algorithm (resolution = 0.7) instead of the commonly used Louvain algorithm or alternatives such as walktrap, multilevel, or infomap. The Leiden algorithm was chosen for its ability to find converged optimal solutions more efficiently, which is particularly beneficial for large-scale single-cell datasets⁵⁹.

Gene expression analysis

We analyzed the expression of S100 family genes and a curated list of genes of interest across different developmental stages and cell types. Feature plots, violin plots, and dot plots were generated using Seurat’s visualization functions (“FeaturePlot”, “VlnPlot”, “DotPlot”) and custom functions from the scCustomize package (v2.1.2).

Analysis of the astrocyte lineage

Cells annotated as astrocytes, apical progenitors, and cycling glial cells were subset for focused analysis of the astrocyte lineage. This subset was re-clustered using the same approach as above. We performed differential expression analysis between astrocyte clusters using both the “FindAllMarkers” function in Seurat (using a logistic regression test⁶⁰) and DESeq2 (v1.44.0)⁶¹ on pseudo-bulk data aggregated by cluster and developmental stage. The combination of these two approaches allowed us to leverage the strengths of both single-cell and bulk RNA-seq differential expression methods. Data preprocessing, quality control (including droplet selection, ambient RNA removal, doublet detection, and refined filtering), clustering, marker gene definitions, and the specific software packages used for data processing were made available at https://harkany-lab.github.io/Cinquina_2024/cortex_visualisation.html.

Human tissue, immunohistochemistry

Fetal brains (gestational week (GW) 27; n = 3 samples; mixed for sex) were collected at the Neurobiobank of the Institute of Neurology, Medical University of Vienna, Austria. Tissues were obtained and used in compliance with the Declaration of Helsinki and following institutional guidelines approved by the Ethical Committee of the Medical University of Vienna (IRB No. 396/2011). Fetal brain tissue was obtained from spontaneous or medically induced abortions with informed consent from the mothers. Fetuses whose cause of death was unrelated to genetic disorders, head injury, neurological diseases, or other known diseases (e.g., infections) were only included. Tissues were kept at 4 °C for 24–48 h due to local regulations before immersion fixation in formalin (10%) on average for 34–37 days before embedding in paraffin. Three-μm tissue sections of formalin fixed, paraffin-embedded tissue blocks were mounted on pre-coated glass slides (StarFrost). Following deparaffinization and rehydration, sections were incubated in a solution of H₂O₂ (3%) and MeOH (Millipore; #67_56_1) for 10 min to block endogenous peroxidase activity. After washing with distilled water, sections were pre-incubated in low pH EnVision™ FLEX antigen retrieval solution (PTLink; Dako) at 98 °C for 20 min, exposed to fetal calf serum (10%) for 10 min and subsequently stained with rabbit anti-CaCyBp antibody (Supplementary Table 3) at 4 °C for 72 h. The EnVision detection kit (K5007, ready-to-use, DAKO) was used to visualize the horseradish peroxidase/3,3′-diaminobenzidine (DAB) reaction product with H₂O₂ substrate (0.01%). Sections were counterstained with hematoxylin-eosin, dehydrated in an ascending series of ethanol, cleared with xylene, covered with Consil-Mount (Thermo Scientific), and viewed under a light microscope (Nikon Eclipse E400). We did not observe variations in the cellular distribution of CaCyBp across the samples. Therefore, a representative image was included.

Immunohistochemistry, light and electron microscopy in mice

Adult male mice were transcardially perfused with ice-cold 0.1 M PB (20 ml), followed by 4% PFA in PB (100 ml at 3 mL/min flow speed). P120 adult brains were post-fixed in 4% PFA overnight. After equilibrating in 30% sucrose for 48–72 h, coronal serial sections (1-in-6 series) were cryosectioned (Leica CM1850) at a thickness of 50 μm. Free-floating sections were processed as described⁶² for either fluorescence microscopy or chromogenic (using DAB as substrate) electron microscopy. E18.5 and P4 brain were immersion fixed in 4% PFA overnight, rinsed in PB, and cryoprotected in 30% sucrose (in PB) for 48 h. Serial coronal sections (20 μm) were cut, thaw-mounted onto fluorescence-free SuperFrost⁺ glass slides, and stored at −20 °C until processing.

For fluorescence histochemistry, E18.5, P4, and P120 sections from n = 4 male mice were extensively rinsed in PB, and blocked with PB containing 5% normal donkey serum (NDS; Jackson ImmunoResearch; #017000121), 1% bovine serum albumin (BSA; Sigma; #A2153), and 0.3% Triton ×-100 (Sigma) at 21–24 °C for 2 h to reduce non-specific labeling. Next, sections were incubated with a combination of primary antibodies (Supplementary Table 3) diluted in PB also containing 2% NDS, 0.1% BSA, and 0.3% Triton ×-100 at 4 °C for 72 h. DyLight Fluor 488, 560, or 633-tagged secondary antibodies (1:300; Jackson ImmunoResearch) were used for signal detection. Hoechst 33,342 (Sigma; #B2261) and 4′,6-diamin-2-phenylindol (DAPI, Sigma; #D8417) were routinely used as nuclear counterstains. After washing in PB and a final rinse in distilled water, free-floating sections were mounted onto fluorescence-free glass slides and coverslipped with Mowiol mounting medium (Sigma; #81381). Fluorescence images were acquired on a Zeiss 880 LSM confocal laser scanning microscope equipped for maximal signal separation and spectral scanning/unmixing.

Electron microscopy was performed as previously published²³. Briefly, free-floating sections from male adult mice were washed three times with PB for 5 min each and then treated with 1% H₂O₂ (Sigma; #216763) for 10 min to block endogenous peroxidase activity. After rinsing, sections were exposed to a blocking solution (see above) at 21–24 °C for 2 h. Sections were then incubated with a rabbit anti-CaCyBp primary antibody (Supplementary Table 3) for 72 h. After washing three times, sections were exposed to biotinylated anti-rabbit secondary antibodies (1:200; Jackson ImmunoResearch, 20–22 °C, 2 h), rinsed again in PB, and then exposed to preformed avidin-biotin-peroxidase complexes (1:200, ABC Elite; Vector Laboratories; #PK6100) at 4 °C overnight. Next, biotin-labeled cells were revealed using a peroxidase reaction with DAB alone (5 mg/ml, 0.03% H₂O₂ as substrate, brown reaction product). Sections were subsequently osmificated (1% OsO₄ in PB, 15 min), dehydrated in an ascending ethanol gradient followed by flat embedding in Durcupan (FLUKA, ACM). Ultrathin sections (60 nm) were cut on a Leica ultramicrotome, collected on Formvar-coated single slot grids, and analyzed with a Tecnai 10 electron microscope (FEI; primary magnifications: 6200×; 17500×).

In situ hybridization (HCR-FISH)

Fresh-frozen E18.5 and P4 mouse brains were cryosectioned at a thickness of 16 μm and collected on SuperFrost⁺ glass slides (ThermoFisher). Sections were immersed in 4% PFA for 15 min, and dehydrated in an ascending ethanol gradient (50%, 70%, and 100%; 5 min each). The HCR™ RNA-FISH protocol for either fresh-frozen or fixed-frozen tissue sections (Molecular Instruments) was used with pre-designed Cacybp, S100a6, Rbfox3, Gfap, S100b, Aldh1l1, and Apo4 probes. Samples were imaged on an LSM 880 confocal microscope (Zeiss) at 40× primary magnification.

Neuronal cultures, shRNA knockdown, treatment with recombinant S100A6

Cerebral tissues (cortex/hippocampus) of mouse embryos (mixed for sex) were isolated on either E18.5 or P0. Cells were mechanically dissociated into a single-cell suspension by papain digestion (20 U/ml, pH 7.4, 45 min; Worthington, #LK003176) and plated onto poly-D-lysine-coated (PDL; Sigma; #P728) 6-well (200,000 cells/well) or 24-well (50,000 cells/well) plates. Cells were maintained in Neurobasal medium (Fisher Scientific; #10888_022) supplemented with B27 (1%; Thermo Fisher; #17504-044), GlutaMAX^TM (1%; ThermoFisher; #35050_038), and penicillin/streptomycin (1%; Life Technologies; #15140_122). Media were replaced every other day.

For shRNA knock-down of Cacybp, P0 cortical neurons (200,000 cells/well) were seeded in 6-well plates, grown for 48 h, then transduced with either 1 μl of adeno-associated virus, serotype 2 (AVV-2) particles (6.09 × 10¹² genome copies/ml; GeneCopoeia^TM; #AA02-CS-MSE091282-AVE001-01-200) to carry shRNA or equal amounts of AVV-2s to deliver scrambled controls (GeneCopoeia^TM; #AA02-CS-CSECTR001-AVE001-01-200). Media were replaced with Neurobasal medium (as above) free of AAV-2 particles after 48 h, and changed every other day. Alternatively, P0 neurons in 6-well plates were exposed to either 1 μg of recombinant His-tagged S100A6 protein (Abcam; #ab108124) or 1 μg of recombinant His-tagged S100A4 protein (Abcam; #ab109341) for 24 h on days in vitro (DIV) 14. Analyses of neuronal morphology were performed on DIV15 because Cacybp mRNA and protein levels were found significantly reduced by then in shRNA-exposed samples vs. both technical controls and samples treated with scrambled shRNA. Cells were then lysed for Western blotting and qPCR after DIV15. Morphometric analysis of cultured neurons was aided by the ImageJ software and included the length of the primary neurite (μm).

In combined experiments, P0 neurons were plated in 6-well plates (200,000 cells/well), while E18.5 neurons in glass bottom dishes (25,000 cells/well; 35 mm diameter, 20 mm well size; 0.13–0.16 mm bottom; Cellvis, #D35-20-1-N). Cells were grown for 48 h, then transfected with AVV-2 particles as above. After 14 DIV, neurons were exposed to 1 μg of recombinant His-tagged S100A6 protein for 24 h. Neurons in 6-well plates were then processed for RNA extraction and qPCR. Neurons in glass bottom dishes were immersion-fixed in 4% PFA (in 0.1 M phosphate-buffered saline [PBS]) for 1 h, and then exposed to a blocking solution consisting of 10% NDS, 5% BSA, and 0.3% Triton ×-100 in PBS (pH 7.4) at 21–24 °C for 2 h. To visualize F-actin, cultured neurons were incubated with Alexa Flour 546-conjugated phalloidin (1:500; Invitrogen; Supplementary Table 3) for 2 h. Hoechst 33,342 (Sigma; #B2261) was routinely used as nuclear counterstain.

To test the effects of astrocytes (with/without genetic manipulation; see below), E18.5 neurons were plated first in 24-well plates (50,000 cells/well), 96-well plates (20,000 cells/well for morphometry), or 6-well plates (1 × 10⁶ cells/well for qPCR), and allowed to grow until 3 DIV. At this point, half of the neuronal media was replaced with either fresh neuronal media or astrocyte-conditioned media for another 24 h. Neuronal morphology and gene expression were then analyzed.

For S100A4 treatment, E18.5 neurons were plated at a density of 50,000 cells/well on glass bottom dishes (as above). After 3 DIV, neurons were exposed to 1 μg of recombinant His-tagged S100A4 protein (Abcam; #ab109341) for 24 h. Cells were then immersion-fixed in 4% PFA, and incubated in a blocking solution for 2 h (as above). Subsequently, samples were exposed to a guinea pig anti-MAP2 primary antibody (Supplementary Table 3) diluted in PBS also containing 2% NDS, 0.1% BSA and 0.3% Triton ×-100 at 4 °C for 24 h. DyLight Fluor 488-tagged secondary antibody were used to visualize MAP2 localization.

After rinsing the samples in PBS and finally in distilled water, coverslips were mounted onto fluorescence-free glass slides using Mowiol. Images were acquired on a Zeiss 880LSM confocal laser scanning microscope with maximal signal separation and spectral scanning/unmixing.

High-throughput neurite tracking

Mouse cerebral cortices were isolated on E18.5. Cells were mechanically dissociated and plated at a density of 20,000 cells/well in PDL-coated 96-well plates for on-line morphometry. After 24 h, E18.5 cultures were exposed to either recombinant S100A6 (1 μg/μl; Abcam #ab108124) for 4 DIV or EPA (either 1 μM or 5 μM; Cayman, #90165) for 7 DIV. Before use, EPA was flushed with N₂, then dissolved in DMSO. Aliquots were stored in glass amber vials at −20 °C until needed. Working concentrations of the EPA stock solution were diluted as appropriated to achieve the necessary final concentration in neuronal media containing BSA as antioxidant. Treatment was in quadruplicate and imaged live using an IncuCyte Zoom high-throughput device (Sartorius). Time-lapse images were acquired in loops with 2-h intervals. The growth rate of neurites in each well was obtained by measuring the surface area covered by neurites and expressed as mm/mm² surface area. Surface occupancy of the neurons after either 4 DIV or 7 DIV in culture was computed as the percentage surface area covered by live neurons (“body cluster area”), and taken as a measure of cell survival.

Primary cultures of astrocytes, confluence analysis, and S100A6 secretion

Cortices and hippocampi of P7 mice (n = 7) were isolated, dissociated into single cell suspension by papain (as above), and plated in T75 flasks. Cells were maintained in DMEM (high glucose, 1% GlutaMAX™; Thermo Fisher, #61965_026) supplemented with Na-pyruvate (1%; Sigma #11360_070), fetal bovine serum (FBS; 10%, Sigma, #A5256801), and 1% penicillin/streptomycin. After 7–9 days, T75 flasks were placed on a shaker (180 rpm, 37 °C, overnight) to remove cell debris, neurons, and oligodendrocytes.

For confluence analyses, astrocytes were plated at a density of 7000 cells/well into 96-well plates. After 24 h, astrocytes were placed in serum-free medium for 24 h, and then treated with glutamate (100 μM; Tocris, #56_86_0) or EPA (5, 10, and 30 μM; Cayman, #90110) for another 24 h. Before use, glutamate was dissolved in sodium hydroxide. EPA was flushed with N₂, dissolved in dimethyl sulfoxide, its aliquots stored at −20 °C, and used as above. Both glutamate and EPA treatments were performed in quadruplicate and imaged live using an IncuCyte Zoom high-throughput imaging device (Sartorius). Time-lapse images were acquired every 2 h. Confluence was determined as above.

To detect S100A6 secretion, astrocytes (1 × 10⁶) were trypsinated (0.25% trypsin-EDTA, Gibco, #25200_56), and seeded in 6-well plates. After 24 h, astrocytes were starved in serum-free medium at 37 °C for 24 h, and then treated with either glutamate (100 μM) or EPA (5, 10, and 30 μM) for an additional 24 h. Supernatants were collected and concentrated using either Amicon Ultra-4 centrifugal units (3 kDa cut-off; Sigma, #UFC800324) by centrifugation at 7500 × g for 40 min or Amicon Ultra-0.5 centrifugal units (3 kDa cut-off; Sigma #UFC500396) by centrifugation at 14,000 × g for 30 min. Concentrated supernatant was then used to detect S100A6 by Western blotting and/or an ultrasensitive mouse anti-S100A6 ELISA (antibodies-online.com; ABIN425090)⁶³ as per the manufacturer’s protocol. Astrocytes were pelleted, lysed, and used for Western blotting.

siRNA-mediated knockdown, benchmarking

To knock-down S100A6 expression, astrocytes were plated at a density of 300,000 cells/well in 12-well plates. After 24 h, astrocytes were placed in serum-free medium for 24 h. S100A6 expression was silenced by siRNA (1 μM; Dharmacon, #SO-3248729G). Scrambled siRNA (1 μM; Dharmacon, #D-001950-01-50) was used as control. Cells were lysed and collected for Western blotting after 96 h. To benchmark this manipulation, S100A6 protein was shown to be significantly reduced in siRNA-exposed samples vs. both technical controls and samples treated with scrambled siRNA (Fig. 7). After RNAi-mediated S100A6 silencing, astrocytes were treated with EPA (30 μM) for 6 h, when their pellets and supernatants were collected for Western blotting.

MTT assay

The 3-(4,5-dimethylthiazol-2-yl)−2,5-diphenyltetrazolium bromide colorimetric assay was carried out at the end of each IncuCyte run to determine drug effects on the viability of astrocytes. MTT (Abcam; #ab146345) was added to each well at a final concentration of 0.5 mg/ml, incubated for 3 h, with the medium then aspirated. One hundred μl of 100% DMSO (Sigma; #472301) was then added to each well. After the formazan crystals had dissolved, absorbance was determined spectrophotometrically at 570 nm. Data on astrocyte viability were analyzed using “Data analysis using Bootstrap-coupled estimation” (DABEST package v2023.9.12) in R for effect sizes and their confidence intervals⁶⁴ (see also: https://harkany-lab.github.io/Cinquina_2024/methods.html).

Neuro-2a cells, transfection, immunocytochemistry

Neuro-2a cells (CCL-131, ATCC) were kept in DMEM (high glucose, 1% GlutaMAX™, ThermoFisher, #61965_026) supplemented with 1% Na-pyruvate, 10% FBS, and 1% penicillin/streptomycin. Neuro-2a cells were trypsin dissociated every 4th day, and newly passaged cells were used in experiments 24–48 h after re-plating. Neuro-2a cells were directly lysed in either Laemmli buffer (for Western blotting; Bio-Rad, #1610747) or RNA extraction medium (for qPCR) or used for immunoprecipitation (see below).

Immunocytochemistry on Neuro-2a cells

Cells were plated at a density of 50,000 cells/well in 24-well plates. After 24 h, cells were exposed to recombinant S100A6 (1 μg/μl; Abcam, #ab108124;) for 1 h, immersion-fixed in 4% PFA (in 0.1 M PBS), and exposed to a blocking solution consisting of 10% NDS, 5% BSA, and 0.3% Triton ×-100 in PBS (pH 7.4) at 21–24 °C for 2 h. Next, samples were incubated with rabbit anti-S100A6 primary antibody (Supplementary Table 3) diluted in PBS also containing 2% NDS, 0.1% BSA, and 0.3% Triton ×-100 at 4 °C for 24 h. Alexa Flour 546-conjugated phalloidin (1:500; Invitrogen; Supplementary Table 3) and DyLight Fluor 488-tagged secondary antibody (1:300; Jackson ImmunoResearch) were used for F-actin and S100A6 signal detection, respectively. Hoechst 33,342 (Sigma, #B2261) was routinely used as nuclear counterstain. After washing in PBS and a final rinse in distilled water, coverslips were mounted onto fluorescence-free glass slides with Mowiol. Images revealing the subcellular localization of S100A6 were acquired on a Zeiss 880LSM confocal laser scanning microscope with maximal signal separation and spectral scanning/unmixing.

Subcellular distribution/localization of S100A6

Neuro-2a cells were plated at a density of either 50,000 cells/well in 24-well plates or 10 × 10⁶ cells in 10-cm Petri dishes for 24 h. Subsequently, they were transduced with 1 μl of AVV-2 particles for shRNA delivery (as above). Media were replaced with DMEM free of AAV-2 particles after 48 h. After an additional 24 h, cells were exposed to recombinant S100A6 (1 μg/μl) for 1 h and then processed for either immunocytochemistry or subcellular fractionation (see below). S100A6 were then detected in both the cytosolic and endoplasmic reticulum fractions by ELISA (antibodies-online.com; #ABIN425090) according to the manufacturer’s instructions⁶³.

Immunoprecipitation and Nano LC-MS/MS measurements

Neuro-2a cells were plated at a density of 10 × 10⁶ cells in 10-cm Petri dishes and exposed to recombinant S100A6 (1 μg/μl) for 1 h. Total protein was extracted, and immunoprecipitation performed using HisCube Ni-INDIGO His-tag MAgBeads MINI Kit according to the manufacturer’s instructions. Briefly, 20 μl of Ni-beads (Cube Biotech, #75201) were transferred to 1.5 mL low-binding tubes, magnetized in a magnetic rack, with the supernatant discarded. Beads were washed (3×) with 0.5 ml PB (each washing step with short vortexing, spin-down, magnetizing, and discarding the supernatant). Proteins were added to washed Ni-beads and incubated at 4 °C overnight. Subsequently, samples were magnetized and washed (3×). Beads were eluted (5×) by adding 30 ml imidazole (500 mM, pH 7.4), incubated for 10 min on a shaker at room temperature, and magnetized. Eluents were collected and added to ice-cold acetone (1:4) and left to precipitate at −20 °C overnight. Precipitated protein samples were centrifugated at 14,000 × g at 4 °C for 1 h, left to dry, and then solubilized in triethylammonium bicarbonate buffer (50 mM) on a shaker at room temperature for 2 h. Trypsin digestion was performed at 37 °C for 16 h. Formic acid was then added to the collection tubes, eluents were evaporated in a SpeedVac at 30 °C for 4 h, and processed for LC-MS.

Mass spectra were obtained using an Orbitrap Q Exactive mass spectrometer (Thermo Fisher Scientific) equipped with a nanospray ion source coupled to a HPLC system (UltiMate 3000, Dionex). Five µL of each sample was first loaded on a pre-column (trap column; Acclaim PepMap100, 100 µm × 2 cm, nanoViper C18, 5 µm, 100 Å, Thermo Scientific) for 10 min using a loading solvent (2% acetonitrile (ACN), 98% H₂O, 0.1% trifluoroacetic acid) at a flow rate of 10 μl/min. Separation was carried out on a C18 analytical column (Acclaim PepMap RSLC, 75 μm × 50 cm, nanoViper C18, 2 µm, 100 Å; Thermo Scientific) at a flow rate of 300 nL/min. Column oven temperature was kept at 40.0 °C. Mobile phase A composition was 2% ACN, 98% H₂O, 0.1% formic acid (FA), while mobile phase B was made up of 80% ACN, 20% H₂O, and 0.1% FA. Gradient conditions were within 135 min retention time at a flow rate of 0.3 μl/min, with mobile phase B set to 2%–90%. MS scans were acquired in positive ion mode over an m/z range of 400–1400 at a resolution of 70.000 (FWHM at m/z 200). Using a data-dependent acquisition mode, the 6 most intense precursor ions with charge states 2+ to 4+ were selected for fragmentation at 30% normalized collision energy and analysed in the Orbitrap at a resolution of 17.500 (at m/z 200). The dynamic exclusion for the selected ions was 30 s. The electrospray voltage was 2.2 kV and the ion transfer capillary temperature was 300 °C. Raw data were converted into mascot generic format (mgf) files using the msConvert tool from the ProteoWizard toolkit (ProteoWizard: Home (sourceforge.io)), and subsequently processed with ProteinProspector (version 6.5.0, University of California). Data were searched against N-terminally His-tagged mouse S100A6/PRA protein sequences (Accession No. P14069, MGSSHHHHHHSSGLVPRGSHMACPLDQAIGLLVAIFHKYSGKEGDKHLSKKELKELIQKELTIGSKLQDAEIARLMDDLDRNKDQEVNFQEYVAFLGALALIYNEALK), as well as the whole mouse proteome in the SwissProt database (version 2021), which yielded identical results. As further confirmation, pure recombinant S100A6/PRA protein was analyzed as standard.

RNA isolation and quantitative PCR

The RNeasy mini kit (Qiagen, #7326820) was used to extract RNA from mouse cerebral tissues and primary cultures. DNase I was used to eliminate traces of genomic DNA. Total RNA was then reverse transcribed to a cDNA library in a reaction mixture using a high-capacity cDNA reverse transcription kit (Applied Biosystems, #4368814). The cDNA library was then used for quantitative real-time PCR (CFX-connect, Bio-Rad). Pairs of PCR primers specific for Cacybp, S100a6, Grp78, Atf4, Atf6a, Chop, Xbp1u, Xbp1s, Map2, Tubb3, S100b, Aldh1l1, and Actb were designed with Primer Bank and the NCBI Primer Blast software (Supplementary Table 4). Quantitative analysis of gene expression was performed with SYBR Green Master Mix Kit (Life Technologies, #1725124). Expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (Gapdh), a housekeeping gene, for every sample in parallel assays (n ≥ 3 biological replicates/group).

Total protein determination and subcellular fractionation

Total protein was extracted from cerebral tissues and primary cultures by using a modified radioimmunoprecipitation assay buffer containing 5 mM NaF (Sigma, #7681_49_4), 5 mM Na₃VO₄ (Sigma, #13721_39_6), 0.1% N-octyl-β-d-glucopyranoside (Sigma, #08001), and a cocktail of protease inhibitors (Complete™, Roche, #11873580001). Samples were homogenized in a Dounce homogenizer, and protein lysates were pelleted at 10,000 rpm at 4 °C for 10 min. To fractionate endoplasmic reticulum (ER) and cytoplasm, samples were homogenized in 0.32 M Sucrose (Sigma, #57903); 1 mM NaHCO₃ (Sigma, #56297), 1 mM MgCl₂ (Sigma, #A0344733); 0.5 mM CaCl₂ (Sigma, #C7902) and a cocktail of protease inhibitors and centrifuged twice at 700 × g at 4 °C for 5 min. The supernatant was centrifuged several times to collect the ER fraction (6300 × g at 4 °C for 10 min, 20,000 × g for 30 min, and 100,000 × g for 1 h). The final pellet was then reconstituted as the ER fraction, whereas the supernatant was saved as the cytosol fraction. Protein concentrations were determined by the Pierce^TM BCA Protein Assay kit (Thermo Fisher, #23225). Cofilin and calnexin antibodies were used to control enrichment of the cytosol and ER fraction, respectively (Supplementary Table 3).

Quantitative Western blotting, including total protein normalization

Western blotting was performed with primary antibodies listed in Supplementary Table 3. Blots were scanned on a Bio-Rad XRS⁺ imaging system and then quantified in Image Lab 3.01 (Bio-Rad). GAPDH was used as loading control. For total protein normalization, labeling of the samples was initiated by carbocyanine (Cy)5 dye reagent (GE Healthcare) that had been pre-diluted (1:10) in ultrapure water. Samples were mixed and incubated at 21–24 °C for 5 min. The labeling reaction was terminated by adding Amersham WB loading buffer (GE Healthcare; 20 μl/sample) containing 40 μM dithiothreitol (Sigma, #10197777001). Equal amounts of the samples (20 μg/40 μl) were boiled at 95 °C for 3 min, and then loaded onto an Amersham WB gel card (13.5%). Electrophoresis (600 V, 42 min) and protein transfer onto PVDF membranes (100 V, 30 min) were at default settings in an integrated Amersham WB system (GE Healthcare). After blocking, membranes were incubated with primary antibodies overnight (Supplementary Table 3). Antibody binding was detected by species-specific Cy3-labeled secondary antibodies (1:1,000; GE Healthcare). Membranes were dried before scanning at 560 nm (Cy3) and 630 nm (Cy5) excitation. Automated image analysis was performed with the Amersham WB evaluation software with linear image optimization, if necessary.

Prenatal exposure to an experimental diet

Female C57Bl6/J animals at 6 weeks of age were randomly assigned to either a hypercaloric diet enriched ~70-fold in ω-3 PUFAs or a standard diet. Prospective dams consumed the experimental diet for 8 weeks prior to conception. Diets were from Special Diet Services (United Kingdom) and quality controlled by mass-spectrometry (Supplementary Table 1). The effect of maternal nutrition on fetal brain development was evaluated at E18.5³⁵. Embryos were collected from n ≥ 5 different pregnancies to keep the number of independent observations sufficient for statistical analyses, sexed if appropriate, and used for histochemistry, proteomics, and molecular biology. The use of specific or mixed genders for particular experiments was specified throughout. Embryos were collected by Cesarean sections from mothers terminally anesthetized by isoflurane (5%, 1 L/min flow rate), weighed, and decapitated with their heads either immersed in 4% PFA in 0.1 M PB (pH 7.4) overnight or snap-frozen in liquid N₂ and stored at −80 °C^52,65.

Click-iT EdU labeling

Dams were intraperitoneally injected (“pulsed”) with 5-ethynyl-2′-deoxyuridine (EdU, 33 mg/kg) at E14.5. At E18.5, n = 3 male embryos/group were selected by PCR genotyping (see ref.²⁴ for primer pairs and protocol). Their brains were immersion fixed in 4% PFA in PB overnight, rinsed in PB, and cryoprotected in 30% sucrose (in PB) for 48 h. Serial coronal sections (20 μm) were cut on a ThermoFisher NX70 cryostat, thaw-mounted onto fluorescence-free SuperFrost⁺ glass slides, and stored at −20 °C until processing. EdU was visualized with Alexa 488-azide using the Click-iT labeling method (Life Technologies)⁶⁶.

iTRAQ proteomics

Single cortical hemispheres (n = 6/group, mixed and balanced for sex from independent pregnancies) were obtained from E18.5 fetuses. Contralateral hemispheres were snap frozen for target validation by real-time PCR and/or Western blotting. Proteins were extracted by homogenization in lysis buffer (pH 10.0) of triethylammonium bicarbonate (25 mM, Sigma), Na₂CO₃ (20 mM, Sigma) in the presence of protease inhibitors (Complete™, EDTA-free Protease Inhibitor Cocktail; Sigma). Protein concentrations were determined by the bicinchoninic acid assay (Thermo Fisher). After quantification, proteins were precipitated in 6 volumes of ice-cold acetone. One hundred µg of proteins per sample were used for isobaric tagging for relative and absolute quantitation (iTRAQ) in an 8-plex layout per the manufacturer’s instructions (ABSciex). In brief, proteins were denatured, reduced, alkylated, trypsin digested (Promega), individually labeled with appropriate iTRAQ tags, pooled, concentrated, re-suspended in 1.4 ml loading buffer (10 mM KH2PO4, pH 3.0 in 25% acetonitrile) and sonicated.

Peptides were separated by cation exchange chromatography on a PolySulfoethyl A column (PolyLC) over 30 min with a KCl gradient increasing up to 0.5 M, with 0.5 ml fractions collected. Fifteen fractions across the elution profile of similar peptide concentration were generated and concentrated (SpeedVac). Fractions were re-suspended in 0.1% trifluoroacetic acid (TFA) and desalted on C18 spin columns (PepClean C18, Thermo Scientific). Half of each fraction was then injected on an Acclaim PepMap 100 C18 trap and an Acclaim PepMap RSLC C18 column (Thermo Scientific), using a nanoLC Ultra 2D plus loading pump and nanoLC as-2 autosampler (Eksigent). The peptides were loaded onto the trap in a mixture of 98% water, 2% ACN and 0.05% TFA, washed for 20 min to waste before switching in line with the column, and were then eluted with a gradient of increasing ACN, containing 0.1% formic acid (2–20% ACN in 90 min, 20–40% in a further 30 min, followed by 98% ACN to clean the column before re-equilibration to 2% ACN). The eluate was sprayed into a TripleTOF 5600 electrospray tandem mass spectrometer (ABSciex) and analyzed in Information Dependent Acquisition (IDA) mode, performing 120 ms of MS followed by 80 ms MSMS analyses on the 20 most intense peaks seen by MS, with the “Adjust collision energy when using iTRAQ reagent” box ticked. The data files were processed by ProteinPilot 4.5 (Sciex) using the Paragon algorithm, searching against the SwissProt database (March 2013 edition). The following settings were selected: sample type: iTRAQ 8-plex (peptide labelled), cysteine alkylation: MMTS, digestion: trypsin, instrument: TripleTOF 5600, species: mouse, ID focus: biological modifications and amino acid substitutions, and search effort: thorough. Results were statistical evaluated in SPSS v. 21.

Behavioral tests

Anxiety-like behavior in first-generation offspring was assessed in the elevated plus maze on P75 (n = 6 males/group from independent pregnancies) as described^67,68,69,70. Mice were placed in the center of a standard elevated plus maze apparatus (arms: 10 cm × 50 cm each) facing an open arm and allowed to explore the maze for 5 min. The distance travelled (cm) and the number of arm entries (n) were determined.

Statistical analysis and figure design

The number of independent samples is indicated in either the individual figure panels or their associated legends. Likewise, the number of animals is given where relevant. All values represent the means ± s.d. of independent experiments. Biological variation among the samples was similar throughout. Statistical significance was determined by either Student’s t-test (for independent groups, two-tailed) or one-way and two-way analysis of variance (ANOVA) followed by either Bonferroni’s or Tukey’s post-hoc correction (for multiple groups). Statistical analysis was performed in Prism 8.0 (GraphPad). Statistical significance was defined throughout unless stated in bulk as follows: *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. Multi-panel images were assembled in CorelDraw 2022 using linear color, brightness, and contrast enhancement to increase visual clarity, if necessary.

Data availability

All data relevant to judge and interpret this study were included in the paper and its accompanying Supplementary Information and Source Data Files. The mass spectrometry (proteomics) data have been deposited to ProteomeXchange (accession number: PDX066923), and are available at: https://www.ebi.ac.uk/pride/archive/projects/PXD066923 or for FTP download at: https://ftp.pride.ebi.ac.uk/pride/data/archive/2025/09/PXD066923. Single-cell RNA-seq datasets used in the current study are available in Gene Expression Omnibus (GEO SuperSeries GSE153164; https://www.ncbi.nlm.nih.gov/geo/query/acc.cgi?acc=GSE153164). Processed single-cell RNA-seq data are available at: https://harkany-lab.github.io/Cinquina_2024/cortex_visualisation.html. Source data are provided with this paper.

Code availability

The code developed and used in this report has been published at https://harkany-lab.github.io/Cinquina_2024. A whole system setup was also made available for download. Data analysis using Bootstrap-coupled estimation were also deposited (https://harkany-lab.github.io/Cinquina_2024/methods.html).

References

Verkhratsky, A. & Butt, A. Neuroglia: Function and Pathology (Elsevier, 2023).
Moore, B. W. A soluble protein characteristic of the nervous system. Biochem. Biophys. Res. Commun. 19, 739–744 (1965).
Article ADS CAS PubMed Google Scholar
Leclerc, E., Sturchler, E. & Vetter, S. W. The S100B/RAGE axis in Alzheimer’s disease. Cardiovasc Psychiatry Neurol. 2010, 539581 (2010).
Article PubMed PubMed Central Google Scholar
Savchenko, V. L., McKanna, J. A., Nikonenko, I. R. & Skibo, G. G. Microglia and astrocytes in the adult rat brain: comparative immunocytochemical analysis demonstrates the efficacy of lipocortin 1 immunoreactivity. Neuroscience 96, 195–203 (2000).
Article CAS PubMed Google Scholar
Xia, C., Braunstein, Z., Toomey, A. C., Zhong, J. & Rao, X. S100 Proteins as an important regulator of macrophage inflammation. Front. Immunol. 8, 1908 (2017).
Article PubMed Google Scholar
Di Bella, D. J. et al. Molecular logic of cellular diversification in the mouse cerebral cortex. Nature 595, 554–559 (2021).
Article ADS PubMed PubMed Central Google Scholar
Hernandez-Ortega, K. et al. S100B actions on glial and neuronal cells in the developing brain: an overview. Front. Neurosci. 18, 1425525 (2024).
Article PubMed PubMed Central Google Scholar
Xu, Q. et al. Profile and regulation of apolipoprotein E (ApoE) expression in the CNS in mice with targeting of green fluorescent protein gene to the ApoE locus. J. Neurosci. 26, 4985–4994 (2006).
Article CAS PubMed PubMed Central Google Scholar
Michalovicz, L. T. et al. Astrocyte-specific transcriptome analysis using the ALDH1L1 bacTRAP mouse reveals novel biomarkers of astrogliosis in response to neurotoxicity. J. Neurochem. 150, 420–440 (2019).
Article CAS PubMed PubMed Central Google Scholar
Dougherty, J. D., Schmidt, E. F., Nakajima, M. & Heintz, N. Analytical approaches to RNA profiling data for the identification of genes enriched in specific cells. Nucleic Acids Res. 38, 4218–4230 (2010).
Article CAS PubMed PubMed Central Google Scholar
Yamashita, N., Ilg, E. C., Schafer, B. W., Heizmann, C. W. & Kosaka, T. Distribution of a specific calcium-binding protein of the S100 protein family, S100A6 (calcyclin), in subpopulations of neurons and glial cells of the adult rat nervous system. J. Comp. Neurol. 404, 235–257 (1999).
Article CAS PubMed Google Scholar
Lesniak, W. & Filipek, A. S100A6 as a constituent and potential marker of adult and cancer stem cells. Stem Cell Rev. Rep. 18, 2699–2708 (2022).
Article CAS PubMed Google Scholar
Graczyk, A. & Lesniak, W. S100A6 expression in keratinocytes and its impact on epidermal differentiation. Int. J. Biochem. Cell Biol. 57, 135–141 (2014).
Article CAS PubMed Google Scholar
Schneider, G. & Filipek, A. S100A6 binding protein and Siah-1 interacting protein (CacyBP/SIP): spotlight on properties and cellular function. Amino Acids 41, 773–780 (2011).
Article CAS PubMed Google Scholar
Filipek, A., Puzianowska, M., Cieslak, B. & Kuznicki, J. Calcyclin—Ca(2+)-binding protein homologous to glial S-100 beta is present in neurones. Neuroreport 4, 383–386 (1993).
Article CAS PubMed Google Scholar
Allaman, I., Belanger, M. & Magistretti, P. J. Astrocyte-neuron metabolic relationships: for better and for worse. Trends Neurosci. 34, 76–87 (2011).
Article CAS PubMed Google Scholar
Wasik, U., Kadziolka, B., Kilanczyk, E. & Filipek, A. Influence of S100A6 on CacyBP/SIP phosphorylation and Elk-1 transcriptional activity in neuroblastoma NB2a cells. J. Cell Biochem. 117, 126–131 (2016).
Article CAS PubMed Google Scholar
Donato, R., Sorci, G. & Giambanco, I. S100A6 protein: functional roles. Cell Mol. Life Sci. 74, 2749–2760 (2017).
Article CAS PubMed PubMed Central Google Scholar
Karagoz, G. E., Acosta-Alvear, D. & Walter, P. The unfolded protein response: detecting and responding to fluctuations in the protein-folding capacity of the endoplasmic reticulum. Cold Spring Harb. Perspect. Biol. 11, a033886 (2019).
Article CAS PubMed PubMed Central Google Scholar
Almanza, A. et al. Endoplasmic reticulum stress signalling - from basic mechanisms to clinical applications. FEBS J. 286, 241–278 (2019).
Article CAS PubMed Google Scholar
Marciniak, S. J. et al. CHOP induces death by promoting protein synthesis and oxidation in the stressed endoplasmic reticulum. Genes Dev. 18, 3066–3077 (2004).
Article CAS PubMed PubMed Central Google Scholar
Trulsson, F. et al. Deubiquitinating enzymes and the proteasome regulate preferential sets of ubiquitin substrates. Nat. Commun. 13, 2736 (2022).
Article ADS CAS PubMed PubMed Central Google Scholar
Malenczyk, K. et al. A TRPV1-to-secretagogin regulatory axis controls pancreatic beta-cell survival by modulating protein turnover. EMBO J. 36, 2107–2125 (2017).
Article CAS PubMed PubMed Central Google Scholar
Tortoriello, G. et al. Miswiring the brain: delta9-tetrahydrocannabinol disrupts cortical development by inducing an SCG10/stathmin-2 degradation pathway. EMBO J. 33, 668–685 (2014).
Article CAS PubMed PubMed Central Google Scholar
Maruta, H., Greer, K. & Rosenbaum, J. L. The acetylation of alpha-tubulin and its relationship to the assembly and disassembly of microtubules. J. Cell Biol. 103, 571–579 (1986).
Article CAS PubMed Google Scholar
Filipek, A., Jastrzebska, B., Nowotny, M. & Kuznicki, J. CacyBP/SIP, a calcyclin and Siah-1-interacting protein, binds EF-hand proteins of the S100 family. J. Biol. Chem. 277, 28848–28852 (2002).
Article CAS PubMed Google Scholar
Borsini, A., Stangl, D., Jeffries, A. R., Pariante, C. M. & Thuret, S. The role of omega-3 fatty acids in preventing glucocorticoid-induced reduction in human hippocampal neurogenesis and increase in apoptosis. Transl. Psychiatry 10, 219 (2020).
Article CAS PubMed PubMed Central Google Scholar
Borsini, A. et al. Omega-3 polyunsaturated fatty acids protect against inflammation through production of LOX and CYP450 lipid mediators: relevance for major depression and for human hippocampal neurogenesis. Mol. Psychiatry 26, 6773–6788 (2021).
Article CAS PubMed PubMed Central Google Scholar
Yu, J. Z., Wang, J., Sheridan, S. D., Perlis, R. H. & Rasenick, M. M. N-3 polyunsaturated fatty acids promote astrocyte differentiation and neurotrophin production independent of cAMP in patient-derived neural stem cells. Mol. Psychiatry 26, 4605–4615 (2021).
Article CAS PubMed Google Scholar
Laye, S., Nadjar, A., Joffre, C. & Bazinet, R. P. Anti-inflammatory effects of omega-3 fatty acids in the brain: physiological mechanisms and relevance to pharmacology. Pharm. Rev. 70, 12–38 (2018).
Article CAS PubMed Google Scholar
Madore, C. et al. Essential omega-3 fatty acids tune microglial phagocytosis of synaptic elements in the mouse developing brain. Nat. Commun. 11, 6133 (2020).
Article ADS CAS PubMed PubMed Central Google Scholar
Chu, Q. et al. Regulation of the ER stress response by a mitochondrial microprotein. Nat. Commun. 10, 4883 (2019).
Article ADS PubMed PubMed Central Google Scholar
Haggarty, P. Effect of placental function on fatty acid requirements during pregnancy. Eur. J. Clin. Nutr. 58, 1559–1570 (2004).
Article CAS PubMed Google Scholar
Helland, I. B., Smith, L., Saarem, K., Saugstad, O. D. & Drevon, C. A. Maternal supplementation with very-long-chain n-3 fatty acids during pregnancy and lactation augments children’s IQ at 4 years of age. Pediatrics 111, e39–e44 (2003).
Article PubMed Google Scholar
Cinquina, V. et al. Life-long epigenetic programming of cortical architecture by maternal ‘Western’ diet during pregnancy. Mol. Psychiatry 25, 22–36 (2020).
Article CAS PubMed Google Scholar
Zeng, C. et al. Evaluation of 5-ethynyl-2’-deoxyuridine staining as a sensitive and reliable method for studying cell proliferation in the adult nervous system. Brain Res. 1319, 21–32 (2010).
Article ADS CAS PubMed Google Scholar
Mulder, J. et al. Endocannabinoid signaling controls pyramidal cell specification and long-range axon patterning. Proc. Natl. Acad. Sci. USA 105, 8760–8765 (2008).
Article ADS CAS PubMed PubMed Central Google Scholar
Kepecs, A. & Fishell, G. Interneuron cell types are fit to function. Nature 505, 318–326 (2014).
Article ADS CAS PubMed PubMed Central Google Scholar
Hipp, M. S., Kasturi, P. & Hartl, F. U. The proteostasis network and its decline in ageing. Nat. Rev. Mol. Cell Biol. 20, 421–435 (2019).
Article CAS PubMed Google Scholar
Allen, N. J. & Eroglu, C. Cell biology of astrocyte-synapse interactions. Neuron 96, 697–708 (2017).
Article CAS PubMed PubMed Central Google Scholar
Huang, E. J. & Reichardt, L. F. Neurotrophins: roles in neuronal development and function. Annu Rev. Neurosci. 24, 677–736 (2001).
Article CAS PubMed PubMed Central Google Scholar
Goenaga, J., Araque, A., Kofuji, P. & Herrera Moro Chao, D. Calcium signaling in astrocytes and gliotransmitter release. Front. Synaptic Neurosci. 15, 1138577 (2023).
Article CAS PubMed PubMed Central Google Scholar
Donato, R. et al. Functions of S100 proteins. Curr. Mol. Med. 13, 24–57 (2013).
Article CAS PubMed PubMed Central Google Scholar
Santamaria-Kisiel, L., Rintala-Dempsey, A. nneC. & Shaw, G. aryS. Calcium-dependent and -independent interactions of the S100 protein family. Biochem. J. 396, 201–214 (2006).
Article CAS PubMed PubMed Central Google Scholar
Lesniak, W. & Filipek, A. S100A6 protein-expression and function in norm and pathology. Int. J. Mol. Sci. 24, 1341 (2023).
Article CAS PubMed PubMed Central Google Scholar
Lesniak, W., Slomnicki, L. P. & Filipek, A. S100A6 - new facts and features. Biochem. Biophys. Res. Commun. 390, 1087–1092 (2009).
Article ADS CAS PubMed Google Scholar
Peng, Z. et al. EPA is more effective than DHA to improve depression-like behavior, glia cell dysfunction and hippcampal apoptosis signaling in a chronic stress-induced rat model of depression. Int. J. Mol. Sci. 21, 1769 (2020).
Article CAS PubMed PubMed Central Google Scholar
Schiavi, S. et al. Perinatal supplementation with omega-3 fatty acids corrects the aberrant social and cognitive traits observed in a genetic model of autism based on FMR1 deletion in rats. Nutr. Neurosci. 25, 898–911 (2022).
Article CAS PubMed Google Scholar
Cinquina, V., Keimpema, E., Pollak, D. D. & Harkany, T. Adverse effects of gestational omega-3 and omega-6 polyunsaturated fatty acid imbalance on the programming of fetal brain development. J. Neuroendocrinol. 35, e13320 (2023).
Article CAS PubMed PubMed Central Google Scholar
Bae, S., Bessho, Y., Hojo, M. & Kageyama, R. The bHLH gene Hes6, an inhibitor of Hes1, promotes neuronal differentiation. Development 127, 2933–2943 (2000).
Article CAS PubMed Google Scholar
Katakura, M. et al. Omega-3 polyunsaturated fatty acids enhance neuronal differentiation in cultured rat neural stem cells. Stem Cells Int. 2013, 490476 (2013).
Article PubMed PubMed Central Google Scholar
Calvigioni, D. et al. Functional differentiation of cholecystokinin-containing interneurons destined for the cerebral cortex. Cereb. Cortex 27, 2453–2468 (2017).
PubMed Google Scholar
Tarhan, L. et al. Single cell portal: an interactive home for single-cell genomics data. Preprint at bioRxiv https://doi.org/10.1101/2023.07.13.548886 (2023).
Stuart, T. & Satija, R. Integrative single-cell analysis. Nat. Rev. Genet. 20, 257–272 (2019).
Article CAS PubMed Google Scholar
Satija, R., Farrell, J. A., Gennert, D., Schier, A. F. & Regev, A. Spatial reconstruction of single-cell gene expression data. Nat. Biotechnol. 33, 495–502 (2015).
Article CAS PubMed PubMed Central Google Scholar
McInnes, L., Healy, J. & Melville, J. Umap: uniform manifold approximation and projection for dimension reduction. JOSS 3, 861 (2018).
Van der Maaten, L. & Hinton, G. Visualizing data using t-SNE. JMLR 99, 2579–2605 (2008).
MATH Google Scholar
Kobak, D. & Linderman, G. C. Initialization is critical for preserving global data structure in both t-SNE and UMAP. Nat. Biotechnol. 39, 156–157 (2021).
Article CAS PubMed Google Scholar
Traag, V. A., Waltman, L. & van Eck, N. J. From Louvain to Leiden: guaranteeing well-connected communities. Sci. Rep. 9, 5233 (2019).
Article ADS CAS PubMed PubMed Central Google Scholar
Ntranos, V., Yi, L., Melsted, P. & Pachter, L. A discriminative learning approach to differential expression analysis for single-cell RNA-seq. Nat. Methods 16, 163–166 (2019).
Article CAS PubMed Google Scholar
Love, M. I., Huber, W. & Anders, S. Moderated estimation of fold change and dispersion for RNA-seq data with DESeq2. Genome Biol. 15, 550 (2014).
Article PubMed PubMed Central Google Scholar
Berghuis, P. et al. Hardwiring the brain: endocannabinoids shape neuronal connectivity. Science 316, 1212–1216 (2007).
Article ADS CAS PubMed Google Scholar
Onsurathum, S. et al. Proteomics detection of S100A6 in tumor tissue interstitial fluid and evaluation of its potential as a biomarker of cholangiocarcinoma. Tumour Biol. 40, 1010428318767195 (2018).
Article PubMed Google Scholar
Ho, J., Tumkaya, T., Aryal, S., Choi, H. & Claridge-Chang, A. Moving beyond P values: data analysis with estimation graphics. Nat. Methods 16, 565–566 (2019).
Article CAS PubMed Google Scholar
Keimpema, E. et al. Differential subcellular recruitment of monoacylglycerol lipase generates spatial specificity of 2-arachidonoyl glycerol signaling during axonal pathfinding. J. Neurosci. 30, 13992–14007 (2010).
Article CAS PubMed PubMed Central Google Scholar
Ayoub, A. E. et al. Transcriptional programs in transient embryonic zones of the cerebral cortex defined by high-resolution mRNA sequencing. Proc. Natl. Acad. Sci. USA 108, 14950–14955 (2011).
Article ADS CAS PubMed PubMed Central Google Scholar
Archer, J. Tests for emotionality in rats and mice: a review. Anim. Behav. 21, 205–235 (1973).
Article CAS PubMed Google Scholar
Harkany, T. et al. N-Methyl-D-aspartate receptor antagonist MK-801 and radical scavengers protect cholinergic nucleus basalis neurons against beta-amyloid neurotoxicity. Neurobiol. Dis. 6, 109–121 (1999).
Article CAS PubMed Google Scholar
Harkany, T. et al. Beta-amyloid(Phe(SO3H)24)25-35 in rat nucleus basalis induces behavioral dysfunctions, impairs learning and memory and disrupts cortical cholinergic innervation. Behav. Brain Res. 90, 133–145 (1998).
Article CAS PubMed Google Scholar
Walf, A. A. & Frye, C. A. The use of the elevated plus maze as an assay of anxiety-related behavior in rodents. Nat. Protoc. 2, 322–328 (2007).
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank J. Mulder and M. Uhlén (Karolinska Institutet) for anti-S100A6 and anti-CaCyBp antibodies, G.G. Kovacs (Medical University of Vienna) for human tissues, and S.L. Shirran, C.H. Botting (University of St. Andrews, United Kingdom; supported by the Wellcome Trust [BB/T017686/1]), M.A. Fuszard (Martin-Luther University, Germany), P. Poullet (Institut Curie & INSERM U900, France) for iTRAQ proteomics and analysis; and A. Fabisikova (Mass Spectrometry Center at the University of Vienna) for LC-MS/MS analysis. This work was supported by the Austrian Science Fund (P 34121-B; E.K.; 10.55776/COE16; T.H.), the National Brain Research Program of Hungary (2017-1.2.1-NKP-2017-00002, NAP2022-I-1/2022; A.A.), the Excellence Program for Higher Education of Hungary (TKP-EGA-25; A.A.), the Swedish Research Council (2023-03058; T.H.), Novo Nordisk Foundation (NNF23OC0084476; T.H.); Hjärnfonden (FO2022-0300; T.H.), European Research Council (FOODFORLIFE, 2020-AdG-101021016; T.H.), and an intramural research grant of the Medical Neuroscience Cluster of the Medical University of Vienna (2021-1; T.H.).

Funding

Open access funding provided by Karolinska Institute.

Author information

Daniela Calvigioni
Present address: Department of Neuroscience, Biomedicum 4B, Karolinska Institutet, Solna, Sweden

Authors and Affiliations

Department of Molecular Neurosciences, Center for Brain Research, Medical University of Vienna, Vienna, Austria
Valentina Cinquina, Evgenii O. Tretiakov, Daniela Calvigioni, Erik Keimpema & Tibor Harkany
Institute of Biological Chemistry, Faculty of Chemistry, University of Vienna, Vienna, Austria
Predrag Kalaba
SE NAP Research Group of Experimental Neuroanatomy, Semmelweis University, Budapest, Hungary
Alán Alpár
Developmental Biology, Department of Anatomy, Histology, and Embryology, Semmelweis University, Budapest, Hungary
Alán Alpár
Endocannabinoid Research Group, Institute of Biomolecular Chemistry (ICB), National Research Council (CNR), Pozzuoli, Italy
Fabiana Piscitelli & Vincenzo Di Marzo
Canada Excellence Research Chair on the Microbiome-Endocannabinoidome Axis in Metabolic Health, Institut Universitaire de Cardiologie et de Pneumologie de Québec, Université Laval, Québec, QC, Canada
Vincenzo Di Marzo
Institut sur la Nutrition et les Aliments Fonctionnels, Centre Nutriss, Université Laval, Québec, QC, Canada
Vincenzo Di Marzo
Faculty of Biology, Medicine and Health, The University of Manchester, Manchester, UK
Alexej Verkhratsky
Achucarro Centre for Neuroscience, IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
Alexej Verkhratsky
Department of Neuroscience, Biomedicum 7D, Karolinska Institutet, Solna, Sweden
Tibor Harkany

Authors

Valentina Cinquina
View author publications
Search author on:PubMed Google Scholar
Evgenii O. Tretiakov
View author publications
Search author on:PubMed Google Scholar
Predrag Kalaba
View author publications
Search author on:PubMed Google Scholar
Alán Alpár
View author publications
Search author on:PubMed Google Scholar
Daniela Calvigioni
View author publications
Search author on:PubMed Google Scholar
Fabiana Piscitelli
View author publications
Search author on:PubMed Google Scholar
Erik Keimpema
View author publications
Search author on:PubMed Google Scholar
Vincenzo Di Marzo
View author publications
Search author on:PubMed Google Scholar
Alexej Verkhratsky
View author publications
Search author on:PubMed Google Scholar
Tibor Harkany
View author publications
Search author on:PubMed Google Scholar

Contributions

T.H. conceived the project. V.C., V.D.M., A.V., and T.H. designed experiments. A.A., E.K., and T.H. procured funding. V.C., D.C., F.P., and E.K. performed in vivo and in vitro experiments and analyzed data. P.K. performed mass-spectrometry-based ligand identification. E.O.T. performed and supervised single-cell data analysis. V.C. and T.H. wrote the manuscript with input from all co-authors. Correspondence and requests for materials should be addressed to T.H. (tibor.harkany@ki.se or tibor.harkany@meduniwien.ac.at).

Corresponding author

Correspondence to Tibor Harkany.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Blanca Díaz Castro and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Cinquina, V., Tretiakov, E.O., Kalaba, P. et al. Astrocytes modulate neuronal development by S100A6 signaling. Nat Commun 16, 9049 (2025). https://doi.org/10.1038/s41467-025-64405-y

Download citation

Received: 01 September 2023
Accepted: 12 September 2025
Published: 13 October 2025
DOI: https://doi.org/10.1038/s41467-025-64405-y