*3.2. Characterization of GLA-Null hESC Clones*

To confirm the pluripotency status of the established GLA-null hESC clones #19 and #27, several stemness markers were detected by RT-PCR, thus confirming that CRISPR/Cas9-mediated *GLA* knockout did not significantly influence the pluripotency gene expression in hESCs (Supplementary Figure S3A). Both GLA-null hESC clones displayed typical pluripotent morphology features, including small and tightly packed cells, high nucleus to cytoplasm ratio, and positive alkaline phosphatase activity (Supplementary Figure S3B). The expression of pluripotency markers OCT4, NANOG, TRA-1-60 and TRA-1-81 was verified by immunofluorescence staining (Figure 2A). Both null-GLA hESC clones were able to spontaneously differentiate in vitro into embryoid bodies expressing markers of three germ layers: HNF3β of endoderm, α-SMA of mesoderm, and nestin of ectoderm (Figure 2B). Both clones exhibited normal karyotypes (Figure 2C). To summarize, both established GLA-null hESC clones exhibited normal stem cell properties, and therefore, were suitable for differentiation into FD-specific cardiomyocytes to investigate the mechanisms of involvement of GLA deficiency in the pathology of FD cardiomyopathy.

**Figure 2.** Characterization of CRISPR/Cas9-edited GLA-null hESC clones. (**A**) Immunofluorescence staining demonstrating the protein expression of pluripotency markers OCT4, TRA-1-60, NANOG and TRA-1-81 in GLA-null hESC clones #19 and #27. Nuclei stained with DAPI. (**B**) Embryoid body formation assay showing differentiation of GLA-null hESC clones #19 and #27 into three germ layers: endoderm immunoreactive for HNF3β, mesoderm immunoreactive for α-SMA, and ectoderm immunoreactive for nestin. Nuclei stained with DAPI. (**C**) Representative karyograms GLA-null hESC clones #19 and #27. Visualization was performed using Agilent CytoGenomics software.

#### *3.3. Recapitulation of FD-Specific Cardiac Abnormalities in GLA-Null Cardiomyocytes*

Given the fact that CM hypertrophy is one of the symptoms of FD-associated cardiomyopathy, we examined whether GLA-null CMs could recapitulate such features. hESCs were differentiated into CMs according to the conventional CM differentiation protocol [21]. Both GLA-null CMs and CMs derived from the parental H9 hESCs (H9 CMs) exhibited spontaneous contractions as early as twelve days following initiation of cardiac differentiation. Noticeably, no significant differences in the efficiency of cardiac differentiation were observed between GLA-null CMs and H9 CMs, which suggested that the GLA deficiency did not significantly affect the differentiation of hESCs to CMs. The accumulation of Gb3 in CMs is the most prominent hallmark of FD-associated cardiomyopathy. Compared with H9 CMs, GLA-null CMs accumulated more Gb3, as was demonstrated by transmission electron microscopy (Figure 3A) and thin layer chromatography (Figure 3B). GLA protein was not expressed in #19 and #27 GLA-null CM clones at day 60 of differentiation (Figure 3C). Cardiac hypertrophy is often accompanied by reactivation of fetal genes which are active during fetal cardiac development and quiescent in adult hearts [22]. We measured the expression level of several of such genes by qRT-PCR and showed that atrial natriuretic peptide (ANP), B-type natriuretic peptide (BNP), but not troponin T (TnT) were upregulated in GLA-null CMs compared to H9 CMs. In addition, GLA-null CMs were characterized by shifted balance between α- and β-cardiac myosin heavy chain (MYH6/MYH7 ratio) expression, which is a common response to cardiac injury and a hallmark of cardiac hypertrophy [23] (Figure 3D). Immunofluorescent staining of cardiac troponin T (cTnT) revealed that GLA-null CMs were of significantly larger size (26% ± 9.7%, *p* < 0.001) as compared to the control H9 CMs (Figure 3E,F), which was consistent with the phenotype of FD-associated cardiomyopathy. In summary, we have shown that CRISPR/Cas9-edited hESC-derived GLA-null CMs recapitulated the typical phenotype features of FD-affected CMs and, therefore, could be a useful model to study the FD-associated cardiomyopathy.

#### *3.4. Proteomic Analysis of GLA-Null CMs*

To further investigate the potential mechanism of FD-related cardiomyopathy, we used proteomic analysis. We focused our analysis to the genes that where downregulated by *GLA* knockout, i.e., whose function was negatively affected. By applying LC-MS/MS, we identified 60 proteins downregulated in GLA-null CMs as compared to H9 CMs with a false discovery rate (FDR) below 0.01. By performing gene ontology (GO) analysis using FatiGO software [24], we found that these 60 proteins were significantly enriched in cellular component GO terms related to extracellular vesicle transportation, secretion, and exocytosis (Figure 4A,B). Among the downregulated proteins involved in regulated exosome release were Ras-Related Protein Rab-11 (RAB11), Rho GDP-dissociation inhibitor 2 (GDIR2, ARHGDIB), VPS36 and VTI1A (Supplementary Table S4). The levels of these proteins were validated and quantified by western blotting; they were shown to be downregulated in GLA-null CMs (Figure 4C,D).

**Figure 3.** Recapitulation of FD-specific cardiac abnormalities in GLA-null cardiomyocytes. (**A**) TEM images showing the ultrastructure of parental type (H9) and GLA-null CMs. The red arrowheads indicate the multilayered lysosomal structure. (**B**) Gb3 content in parental (H9) and GLA-null CMs (clones #19 and #27) analyzed by TLC. (**C**) Western blot showing lack of expression of GLA in clones #19 and #27 of GLA-null CMs as compared to the parental wild type CMs and hESCs (H9). Connexin 43 (Cx43) and NANOG served as cardiomyocyte and pluripotency markers, respectively. GAPDH used as a loading control. (**D**) qRT-PCR analysis of expression several fetal heart markers in CM clones #19 and #27. The results are expressed as fold change relative to H9 CMs. (**E**) Immunostaining of cTnT showing significantly enlarged size of GLA-null CMs compared to H9 CMs. (**F**) Quantification of area size of GLA-null CMs (green columns) and H9 CMs (blue columns). At least 200 cells were analyzed individually and statistical difference is *p* < 0.001.

**Figure 4.** Proteomic analysis of GLA-null CMs. (**A**) The most enriched cellular component GO terms in the list of 60 genes downregulated in GLA-null CMs. (**B**) Visualization of the most enriched cellular component GO terms using REVIGO software. (**C**) Western blot showing expression of the mediators of vesicular trafficking in GLA-null CMs #19 and #27 as compared to wild type control (H9). GAPDH served as a loading control. (**D**) Quantitation of expression of proteins annotated as the mediators of vesicular trafficking. The data are presented as mean ± standard deviation error bars from three independent experiments. (**D**) Flow cytometric analysis of CD63 expression in H9 and GLA-null CM-derived exosome isolated with magnetic beads directly from cell culture medium. (**E**) Quantification of exosome numbers in the supernatant from H9 and GLA-null CMs.
