The immortalized mouse hippocampal cell line HT-22 (passages 15–20) was cultured with the following Thermo Fisher Scientific Inc., reagents: Dulbecco’s Modified Eagle Medium (Catalog Number: 11966025) supplemented with 4.5 g/L glucose, 10% Fetal Bovine Serum (FBS) (Catalog Number: 16000044), 2 mM glutamine (Catalog Number: 25030024), 100 I.U./ml penicillin-100 μg/ml Streptomycin (Catalog Number: 15140163), at 37°C under 5% CO₂. Once 60–70% confluency was reached, cells were washed twice with phosphate buffer saline (PBS) to remove any FBS. Then, cells were grown in the same medium but using 10% exosome depleted FBS (Thermo Fisher Scientific Inc., Catalog Number: A2720803). After 48 h, the conditioned medium (CM) of HT-22 cells was collected, and the cells were lysed with RIPA buffer containing Halt protease inhibitor cocktail (Thermo Fisher Scientific Inc., Catalog Number: 78445). The CM was subjected to three consecutive centrifugations: 2,000× g, 10 min; 20,000× g, 20 min; and 100,000× g, for 2 h at 4°C. The final pellet, called P100 or EV (Kowal et al., 2016), was resuspended in 0.2 μm filtered PBS, pelleted again, and stored at −80°C until further analysis or used immediately for downstream analysis.

P100 was re-suspended in PBS, re-pelleted and analyzed by transmission electron microscopy. To examine its fine structure, P100 was fixed in 2.5% glutaraldehyde/0.1 M sodium cacodylate buffer pH 7.4 at room temperature overnight (EMS, Catalog Number: 15960). Then it was washed with cacodylate buffer for 2 h and post-fixed with 1% aqueous osmium tetroxide for 2 h, rinsed with bidistilled water and stained with 1% uranyl acetate for 90 min. It was dehydrated twice stepwise with increasing acetone concentrations (50, 70, 95, and 100%) for 30 min each. Preinclusion was done with epon:acetone (1:1) overnight and then it was included in pure epon. The polymerization was carried out in an oven at 60°C for 48 h. Fine cuts of 80 nm thickness were obtained in a Leica Ultracut R Ultramicrotome, stained with 4% uranyl acetate in methanol for 2 min and with Reynolds’s lead citrate for 5 min. The slices were observed in a Philips Tecnai 12 BioTwin microscope (Eindhoven, Netherlands) at 80 kV. All post-fixation treatments of a sample were performed in the advanced Pontificia Universidad Católica de Chile Microscopy Unit.

For immunogold analysis, P100 pellets were fixed in 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer pH 7.0, 0.2% for at 4°C for 2 h, then rinsed for 30 min in phosphate buffer, dehydrated with 50, 70, 95, and 100% ethanol for 15 min each and then left overnight in ethanol/LR White 1:1. Then, P100 pellets were included in the resin in gelatin capsules and polymerized at 50°C for 8 h. A Leica Ultracut R Ultramicrotome was used to make fine 90 nm wide cuts, which were placed in nickel grids. A cut was blocked with PBS/0.1% BSA for 30 min, then incubated with a 1:20 dilution of a primary antibody (CD63, Santa Cruz Biotechnology, Inc., Catalog Number: sc-5275; Wnt3a, Thermo Fisher Scientific Inc., Catalog Number: PA5-44946; Wnt5a, Abcam, Catalog Number: ab174963; Santa Cruz Biotechnology, Inc., Wnt5a Catalog Number: sc-365370; Thermo Fisher Scientific Inc., Wnt7a, Catalog Number: PA5-80231; MyBioSource GPR177, Catalog Number: MBS769833, and washed three times with PBS/0.1% BSA/0.2% Triton X-100 for 20 min. Goat anti-Rabbit Gold 6 nm, (Abcam, Catalog Number: ab41498) or goat anti-mouse Gold 25 nm (EMS, Catalog Number: 25135) was used as secondary antibody and incubated for 1 h, then washed three times for 20 min with PBS/0.1% BSA/0.2% Triton X-100, and a rapid final wash in distilled water was done. A Philips Tecnai 12 BioTwin microscope (Eindhoven, Netherlands) at 80 kV was used to examine the sections stained with aqueous 1% uranyl acetate for 5 min.

A more precise size determination of the EV was obtained through performing NTA, which is a higher resolution technique than electron microscopy (Malloy, 2011). It is based on light scattering and Brownian motion behavior, which allows assessing nanoparticle size distribution and abundance of samples in liquid suspensions (NanoSight NS300, Malvern). The samples were diluted in sterile PBS allowing that the particle concentration is within the linear dynamic range of the equipment (10⁶–10⁹ particles/ml). The analysis was performed using the 532 nm laser, 565 nm long-pass filter, with a camera level set at 9 and a detection threshold of 3.

Western blots were used to analyze the protein composition of P100 pellets. We used primary antibodies: (dilution used for all primary antibodies 1/300) for, GPR177 (MyBioSource, Catalog Number: MBS769833; Abcam, GM130 Catalog Number: ab52649; Thermo Fisher Scientific Inc., Wnt3a Catalog Number: PA5-44946; Abcam, Wnt5a Catalog Number: ab174963; Thermo Fisher Scientific Inc., Wnt7a Catalog Number: PA5-80231. All the following antibodies were from Santa Cruz Biotechnology, Inc., Wnt5a, Catalog Number: sc-365370; Alix, Catalog Number: sc-53540; TSG101, Catalog Number: sc-7694; NCAM-L1 Catalog Number: sc-374046; Flotillin-1, Catalog Number: sc-133153; Tubulin, Catalog Number: sc-8035; CD63, Catalog Number: sc-5275. Lysate of HT-22 cells were homogenized in RIPA buffer (10 mM Tris–HCl, 1 mM EDTA, 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, 140 mM NaCl) supplemented with 1 mM PMSF, 7 μg/mL Pepstatin, 5–10 μg/mL Leupeptin, and 10 μg/mL Aprotinin. The protein content of the cell lysate and the P100 were determined by Pierce^TM BCA Protein Assay Kit. An aliquot containing 30 μg of protein in both the lysate and the P100 was mixed with standard sample loading buffer loaded and ran in a 10% acrylamide gel. Proteins were transferred on to a PVDF membrane overnight at 4°C.

HT-22 cells were cultured as described before to obtain the conditioned medium, but in parallel, the cells were incubated with 1 μM of the Wnt-C59 Porcupine inhibitor (Tocris, Catalog Number: 5148). The fresh drug was added after 24 h, and the CM was collected after 48 h to perform the enrichment procedure of EV as indicated above. As a stock of 20 mM of Wnt-C59 was prepared in DMSO, control cells were treated with 0.005% DMSO (vehicle) prepared in indicated cell media.

The sEVs were isolated by differential ultracentrifugation and resuspended in PBS. Samples were incubated in either PBS or 1.2 μg/ml Proteinase K (Promega, Catalog Number: V3021) in PBS, with or without 0.5% Triton X-100, in a final volume of 40 μl per sample for 5 min at 4°C. The assay was stopped by the addition of 8 μl 5X SDS-PAGE Loading Buffer and heating for 8 min at 97°C, and samples were analyzed by SDS-PAGE and immunoblotted.

The GraphPad Prism program was used for statistical analysis. Data derived from the western blot analysis of the Wnt ligands and other proteins were subjected to the Shapiro–Wilk test, which tested the normality of a data set since the sample size was less than 50. After confirming that all the data were normally distributed, a mismatched t-student test was performed. Error bars show SD. No sample size calculation was performed. No exclusion criteria were pre-determined.

To isolate extracellular vesicles, we used the mouse hippocampal cell line HT-22, a relevant model for studying glutamate and heavy metal neurotoxicity (Murphy et al., 1989; Fukui et al., 2009). The cell line HT-22 is accessible to culture and provides an in vitro model for biochemistry studies requiring large amounts of starting material. In fact, six to eight 150-mm culture plates were required to obtain 100 μg of protein in the P100 pellet. Transmission electron microscopy (TEM) was performed to morphologically characterize the vesicles in the P100 fraction (see section “Materials and Methods”). A representative image of the P100 pellet shows an enriched fraction of exosome-like vesicles containing a membrane with typical bilayer morphology and with an average diameter of 124.1 ± 6.8 nm (Figures 1A,B). Immunogold labeling allows us to visualize the CD63 exosomal signature marker (Figure 1B, right panel). The structure of the vesicles was partially lost due to a mild fixation used in immunogold studies to preserve protein epitopes. Considering that the fixation and dehydration of the sample for TEM analysis can cause deformation of the vesicles, a more precise size determination was obtained using NTA. Consistently, the NTA size distribution was similar to the reported size by TEM. The NTA profiles showed a peak of abundant vesicles of 127.4 ± 1.3 nm (Figure 1C; mode) and a population of vesicles with a size of <200 nm (defined as small extracellular vesicles, sEVs) representing 78% of the particles in the sample. The remaining 22% corresponds to medium or large extracellular vesicles (Figures 1C,D).

Then, we characterize the vesicles released by the HT-22 cells according to the International Society for Extracellular Vesicles (ISEV; Théry et al., 2018). Figure 1E shows a western blot analysis of the P100 protein content profile. The P100 pellet exosomal marker content was enriched relative to that of the lysate in the following rank order: Alix > TSG101 > CD63 > Flotillin-1 (Figure 1F). The neuronal origin of the vesicles released by HT-22 cells is corroborated by the presence of the adhesion molecule NCAM-L1, an accepted exosome biomarker of neuronal origin (Fauré et al., 2006; Figure 1F). Nevertheless, the cis-Golgi matrix protein, GM130, was absent from sEVs (Figure 1E), suggesting our preparation was not contaminated with intracellular membranes. Thus, our analysis revealed that vesicles released by HT-22 cells and collected in the P100 fraction effectively correspond to small extracellular vesicles like exosomes (size <200 nm) (Théry et al., 2018).

The enrichment and homogeneous population of exosome-like vesicles prompt us to continue with this cell model to study the presence of Wnt ligands on those vesicles. Next, we examined whether the canonical Wnt3a and Wnt7a ligands and the non-canonical Wnt5a (Korkut et al., 2009; Beckett et al., 2013) were secreted in association with sEVs released by HT-22 cells. Western blot analysis of the sEVs containing fraction (P100) showed enrichment of Wnt5a and the Wnt chaperone, GPR177/Evi, relative to the cell lysate (Figures 2A,B). On the other hand, although present, Wnt3a and Wnt7a were not enriched in this fraction (Figures 2A,B). To corroborate the presence of Wnts in sEVs, immunogold analysis was performed, revealing that Wnt3a, Wnt5a, Wnt7a, and GPR177/Evi were present in the sEVs (Figure 3B). Also, a double immunogold analysis was performed, revealing that Wnt5a (red arrows) and GPR177/Evi (black arrows) were jointly secreted together by the same sEVs (Figure 3D). On the contrary, the IgG respective controls did not show any electrodense mark (Figures 3A,C). Although these data provide evidence of the presence of Wnt ligands in HT-22 secreted vesicles with exosomes characteristics, it was not possible to detect double labeling of Wnt3a or Wnt7a with GPR177/Evi.

Porcupine O-acyltransferase is an O-acyltransferase whose activity affects Wnt ligand secretion in Wnt-producing cells (Kadowaki et al., 1996; Torres et al., 2019). It is unknown if inhibition of PORCN by Wnt-C59 affects only a single or all the secretory pathways that mediate Wnt release from cells (Proffitt et al., 2013; Bengoa-Vergniory et al., 2014). In the present work, we specifically assessed the effect of Wnt-C59 on Wnt ligand secretion associated with sEVs. Although this inhibitor should not affect the exosome formation and secretion pathways, we carried out control experiments. From now on, we will call control and Wnt-C59 the vesicles derived from the vehicle (DMSO) and Wnt-C59 treated cells, respectively. First, we evaluated whether treatment of HT-22 cells with Wnt-C59 alters the relative abundance of each of the vesicle subpopulations released by these cells. NTA was performed with the NanoSight NS300 system to analyze the microvesicle fraction content in the P100 pellet. Figure 4A provides a graphical description of the relationship between particle density (i.e., particles/ml) and their size distribution (nm). There were no significant differences between the control and Wnt-C59 extracellular vesicles (Figure 4B). For both the control and Wnt-C59, there is an enrichment of vesicles with a size of less than 200 nm, similar to the finding showed in Figure 1C. Even though there was a trend toward a decrease in the density of total particles, this decline was not significant (Figure 4C). Similarly, no significant differences were observed in the mean distribution size (Figure 4D) and the mode distribution size (Figure 4E) between the control and Wnt-C59 vesicles. The mean distribution sizes were 166.6 ± 2.5 nm and 164.8 ± 3.9 nm for the control and Wnt-C59 vesicles, respectively (Figure 4D). Their mode distribution size was 128.5 ± 0.8 nm and 128.1 ± 0.4 nm for the control and Wnt-C59 vesicles, respectively (Figure 4E). Therefore, Wnt-C59 did not affect the secretion of sEVs by HT-22 cells.

Western blot analysis was then used as a second control experiment to determine whether Wnt-C59 alters the content of exosomal markers in the sEVs (Figure 4F). As concentration standards have not been described for sEVs protein normalization (Koritzinsky et al., 2017), a statistical analysis of the sEVs protein content with three replicates were performed as shown by others (Imjeti et al., 2017) in which the staining intensity of each band derived from non-treated samples was assigned a value of 100%. The value obtained in the Wnt-C59 sEVs, derived for treated cells, was compared with its respective non-treated control. A t-student test was performed on each data pair of a protein. The results show that Alix, CD63, TSG101, and NCAM-L1 expression levels were unaffected by Wnt-C59 treatment (Figures 4F,G). On the other hand, Wnt-C59 treatment significantly decreased (p = 0.0011) the Flotillin-1 exosomal content relative to its content in the Wnt-C59 exosomes. The purity of the sEVs preparation was corroborated by the absence of the Golgi protein GM130. Hence, the porcupine inhibitor Wnt-C59 has no effect on the amount, size distribution, or content of exosomal markers but Flotillin-1 in the sEVs.

Next, we performed a western blot analysis to determine if Wnt-C59 altered the content of Wnt3a, Wnt5a, Wnt7a, and GPR177/Evi proteins in the P100 fraction obtained from control cells and Wnt-C59 treated cells (Figure 5A). Quantification was carried out in the same way as it was done for Figure 4. The results showed that Wnt ligands and GPR177/Evi levels decreased significantly (Figure 5B). Then, to confirm this finding, a second analysis was performed where the intensity values of the band were normalized dividing them by the number of particles contained in 30 μg of protein loaded per well, as previously reported by others (Dismuke et al., 2016; Shu et al., 2020). Accordingly, the Student t-test results show that the Wnt3a, Wnt5a, Wnt7a, and GPR177/Evi protein content decreased significantly (∼50%) compared to the control group (Figure 5C). Therefore, these declines suggest that palmitoylation is a necessary post-translational modification that allows Wnt ligand secretion by sEVs.

To establish whether Wnt ligands are localized within different regions of sEVs, we used proteinase K to perform a “shaving” procedure (Escrevente et al., 2011; Wang et al., 2017). In this procedure, proteinase K only degrades proteins on the surface of the vesicles, since it does not penetrate the membranous lipid bilayer. However, if proteinase K and Triton X-100 are applied together, the detergent lyses the outer limiting vesicles membrane rendering all sEVs proteins accessible to proteolysis by proteinase K. Hence, the effect of proteinase K on sEVs protein content in the presence and absence of Triton X-100 was compared (see section “Materials and Methods”). Western blot data shows that proteinase K treatment alone did not degrade Wnt3a but Wnt5a, and Wnt7a were degraded (Figure 6). On the other hand, co-incubation with proteinase K plus Triton X-100 completely digested all these proteinaceous Wnt ligands. As a control we used the exosomal biomarker CD63 which is a tetraspanin transmembrane protein (Théry et al., 2018) sensitive to proteinase K treatment as demonstrated by others (Vlassov et al., 2012; Phoonsawat et al., 2014). As luminal control we used Alix which was still present after proteinase K digestion but underwent a shift to a lower molecular weight protein. A similar result was found for Alix for exosomes derived from neurons treated with proteinase K (Wang et al., 2017). Moreover, the presence of two bands for Alix has been previously documented in several cell lines (Shtanko et al., 2011; Bongiovanni et al., 2012; Lopes-Rodrigues et al., 2019), with the 97 kDa band representing the full-length protein and the 75 kDa band would represent a cathepsin digestion product (Impens et al., 2010; Lopes-Rodrigues et al., 2019) that have been found in exosomes (Fiandaca et al., 2015; Kanninen et al., 2016).

Taken together, these results are unique since they suggest that Wnt ligands are segregated into different compartments of exosome-like vesicles secreted by HT-22 cells. Accordingly, it is likely that Wnt3a is restricted to the luminal interior, while Wnt5a and Wnt7a are delimited to the outer limiting membrane.