Fig 2: Peripheral M6a glycoprotein levels are affected by stress. (A,B) ELISA determination of M6a levels in serum samples of a control C57Bl/6 J mouse population. Both male (A) and female (B) samples show a normal distribution. Female mice n = 10; Shapiro-Wilk normality test P = 0.3; male mice n = 24; P = 0.6. (C-F) Female and male mice were subjected to a 3-week protocol of chronic restraint stress (CRS, black boxes). CRS consisted of 5 days of daily restraint stress followed by a 2-day interruption. Mice were sacrificed at the end of week 3. M6a levels in hippocampal homogenates (C, D) and in serum (E,F) were determined by ELISA. n = 9–10 for females and n = 6–8 for males. *P < 0.05. (G–J) To evaluate stress effect on M6a levels in individual male and female mice three blood extraction procedures were performed (arrowheads between boxes), the serum from the first and third extractions were assayed for M6a levels. (G, H) effect of blood extraction alone (white boxes) (I, J) blood extraction plus CRS (black boxes). Each line represents one individual. Gray line represents mean values. T-student test followed by Dunnet multiple comparison test *P < 0.05. n = 6 for males and n = 5 for females.

Fig 3: Phenotypic changes in COS-7 cells treated with M6a-containing EVs. EVs were isolated from COS-7 cells that were either non-transfected (NT), transfected with a plasmid coding GFP or with a construct coding the fusion M6a-GFP. (A) EV Western blot characterisation. Negative EV markers used were EEA1, calnexin and Rab5. HSP-70 was used as a positive EV marker. Similar amounts of protein (25 µg) were loaded in each lane. Western blot membrane was cut and incubated with the indicated antibodies. For comparison, cell lysates were included in the analysis. (B) The EV content assessed by Western blot with anti-GFP antibody to detect GFP and M6a-GFP. Bands representing M6a are indicated with stars while bands corresponding to GFP are indicated with hashtags. Full-length blots are shown in Figure S11. (C) Representative image showing EVs inside COS-7 cells. EVs were isolated from COS-7 cells transfected with M6a-RFP and subsequently stained with DiO lipophilic dye. Left panel shows the total population of EVs (DiO stain). Middle panel shows the EVs that carry the M6a-RFP protein. Right panel and magnifications show the overlay of left and middle images. Images show maximum projections of confocal Z-stacks. Scale bar 20 µm. (D) Phalloidin stained COS-7 cells were magnified to visualise filopodia. Recipient cells with phenotypic changes display multiple protrusions (arrows in right panel). (E) Quantification of the results reveals that M6a-positive EVs increased the percentage of COS-7 cells displaying filopodia compared to cells treated with EVs isolated from GFP-transfected COS-7 cells. n = 3 independent experiments, over 100 cells counted per each condition. *P < 0.05; t-student test. (F–I) Comparison of phenotypic changes between untreated COS-7 cells (F) or cells treated with neuronal EVs (G), M6a-GFP-containing EVs (H) and plasmid DNA coding M6a-GFP (I). Amplification of figures show no filopodia (F), long membrane extensions (arrowheads in G), long tiny filopodia in patches (arrows in H) or extensions throughout the cell surface (arrows in I). (J–L) Comparison of GFP expression between COS-7 cells transfected with a plasmid coding GFP (J), the fusion protein M6a-GFP (K) and treated with EVs carrying the fusion protein M6a-GFP (L). Cells were stained with phalloidin (red). Scale bar 20 µm.

Fig 4: M6a is released into neuronal extracellular vesicles (EVs) that induce membrane protrusions in recipient cells. (A) Representative TEM images of EVs isolated by ultracentrifugation from culture medium of hippocampal neurons. During the sample processing for TEM analysis, some vesicles may collapse while others may not. This explains the perfectly round shape of the isolated EVs. (B) Western blot analysis of neuronal lysate (positive control), ultracentrifugation supernatant and pellet (containing the EV fraction) with the positive marker flotillin 1 (upper panel) and with M6a (lower panel). (C) Western blot reacted with anti-M6a antibody. Scissors represent a cut in the blotting membrane. The left side was incubated with anti-M6a alone; the right side of the membrane was incubated with anti-M6a in the presence of blocked antibody (BA). Note that the blocking peptide greatly reduces M6a signal. Full-length blots are shown in Figure S9. (D) Western blot analysis of neuronal. Negative EV markers used were EEA1 and Rab5. HSP70 was used as a positive EV marker. Similar amounts of protein (25 µg) were loaded in each lane. Western blot membrane was cut and incubated with the indicated antibodies. MW: Molecular Weight Marker. Full-length blots are shown in Figure S10. (E) Representative image of COS-7 cells transfected with the PH-delta-RFP construct whose protein product locates in the plasma membrane (left, upper panel); stained with calnexin (an ER marker; left, middle panel) and treated with DiO labelled EVs (left, lower panel). On the right panel, the overlay of the three channels is shown. Panel 1 shows a projected (YZ) orthogonal section along the white line in the image derived from 11 z-slices 0.4 µm apart. Arrows and inset (panel 2) show direct contact of EVs with membrane protrusions. Scale bar 20 µm. (F) COS-7 cell phenotype change after addition of neuronal EVs. Scale bar 10 µm. (G) Quantification of the percentage of cells displaying phenotype change. COS-7 cells were treated with: EVs isolated from non-transfected COS-7 cells (COS-7), defined medium from cultured neurons, medium from hippocampal neurons maintained in vitro for 14 days (conditioned media) or EVs isolated and concentrated from neuronal culture medium. n = 3 independent experiments, more than 100 cells counted per each condition. **P < 0.0001. ANOVA followed by Dunnet post-test.