In-silico study: Positive correlation of GPM6B with neural cell differentiation
GPM6B showed the highest level of expression in the human brain, compared to other primates 28 (Fig. 1a). The human GPM6B expression ratio in comparison to chimpanzee, Old-World monkeys, New-World monkeys, and lemurs was 6.1-, 3.6-, 5.4, and 4.3-fold, respectively. NT2 cells under RA treatment can differentiate into neural cells. To elucidate the expression status of GPM6B in neural cell differentiation, we performed an in-silico analysis on RNA sequencing (RNA-seq) datasets retrieved from the Gene Expression Omnibus (GEO) database (Table S1). To that end, we first analyzed the expression of GPM6B, NESTIN (NES) (a pluripotency marker) 29, Glial fibrillary acidic protein (GFAP) (a specific astrocytic marker) 30, neuronal-specific tubulin isoform 3 (TUBB3) (an early stage neural marker), and microtubule-associated protein 2 (MAP2) (a late stage neural marker) 31 on NT2 cells, using the RNA-seq datasets (Fig. 1b). We found that during differentiation, the expression of neural cell markers increased. GPM6B showed a positive correlation with the candidate markers. To confirm the accuracy of our in-silico analysis, we measured the expression of randomly selected candidate markers, TUBB3 and MAP2, at the RNA level in undifferentiated and differentiated NT2 cells, which were 21 days under RA treatments. The expression of TUBB3 and MAP2 were significantly increased in differentiated NT2 cells (Fig. 1c).
CRISPR/Cas9-mediated deletion of the GA-repeat in NT2 cells
CRISPR/Cas9 is a versatile and feasible gene editing technology targeting specific genome sites. To evaluate the potential effect of the GA-repeat on the GPM6B expression, we first designed a protocol based on using the CRISPR/Cas9 editing tool to delete the GA-repeat, by applying homology-direct repair (HDR) pathway (Fig. 2a). A donor vector, peGPM6B, comprised of a template without the GA-repeat at the target site. We mutated PAM sites in the donor vector to prevent re-targeting phenomena at the target site or unwanted targeting of peGPM6B in vivo. We designed gRNAs at specific sites, not disrupting TF binding sites at the regulatory region of the GPM6B gene (Fig. 2a). Moreover, we used eSpCas9(1.1) to express high specificity spCas9, which significantly decreases off-target effects 32. To evaluate the efficiency of designed gRNAs at the target site, the result of the T7E1 assay indicated that the editing efficiency in pool cells was 15.5% (Fig. 2b). The edited pool cells were diluted serially (1 cell/100 µL) and cultured in a 96-well plate to obtain a homogenous single clone. Subsequently, we used specific primer sets to determine the successfully edited cells, as follows: First, to exclude the clones, in which the donor vector was randomly integrated, we used the GPM6B F2 and R2 primer set outside of homology arms (Fig. 2a). Then, nested ARMS-PCR was applied to obtain PCR products, using specific primer sets (GPM6B-IF and GPM6B-R1; GPM6B-mutIF and GPM6B-R1). Clone (C) 1 and C4 were successfully amplified, using the F2 and R2 primer set (Figure S1a). The nested ARMS-PCR represented both clones being edited successfully (Figure S1b). Finally, to confirm the accuracy of donor template KI, Sanger sequencing was performed, and confirmed successful deletion of the GA-repeat at the target site (Fig. 2c). The same strategy was used to detect the successful KI in scrambled clones (SCs). Eventually, SC3 was selected, and the introduced mutations at the PAM sites were confirmed, using Sanger sequencing (Figure S1c).
Deletion of the GA-repeat significantly reduced the expression of the GPM6B gene
We used the CRISPR/Cas9 gene editing tool to precisely delete (GA)9 and evaluate its effect on the expression of the GPM6B gene. For this purpose, GPM6B was targeted in NT2 cells, which can be further differentiated into neural cells under RA treatment. The expression level of GPM6B was measured in the untreated and edited pool cells, using qRT-PCR, and the results showed a significant reduction in GPM6B expression in edited cells (Fig. 3a). Furthermore, we assessed the expression level of GPM6B in the single clones obtained, using qRT-PCR. The expression of GPM6B was not statistically different between the untreated and SC3 cells. However, the expression level of GPM6B in the C1 significantly decreased compared to the untreated and SC3 cells (p < 0.05) (Fig. 3b). To confirm the accuracy of the qRT-PCR results, the expression of GPM6B was measured at the protein level. Our results showed that the expression of GPM6B in the SC3 and C1 decreased by 14.8% and 49%, respectively, compared to the untreated cells. Moreover, the expression of GPM6B in C1 decreased by 34.1% compared to SC3 (Fig. 3c). Our observation of data retrieved from the ENCODE project at the UCSC database revealed that the GPM6B (GA)9 is the potential binding site for PRDM1, ZNF768, and Stat2 TFs (Fig. 3d). Moreover, this dinucleotide STR is close to the USF1 TF binding site, where the binding affinity was confirmed using chromatin immunoprecipitation-sequencing (ChIP-seq) (Fig. 3d).
Deletion of the GA-repeat in GPM6B disrupted optimal differentiation of NT2 cells into neural cells
Under RA treatment, NT2 cells are differentiated into neural cells. Previously, it was revealed that GPM6B has a crucial role in axon development 23. We monitored the SC3, C1, and DMSO-treated NT2 cells on Days 0, 8, and 21 after RA-treatment. According to our observations, the NT2 cells under DMSO treatment and the SC3 cells under RA treatment showed neural morphology on Days 8 and 21, while the differentiation of the C1 cells under RA treatment was disrupted to a considerable extent (Fig. 4a). Our results elucidated that DMSO and the introduced mutations at the PAM sites have no significant effect on the NT2 cell differentiation. At the same time, deleting the (GA)9 decreased the expression of the GPM6B and consequently reduced the differentiation of the modified NT2 cells.
To study a link between reduced amount of GPM6B and the deficient differentiation of the NT2 cells into neural cells, the expression of TUBB3 and MAP2 were measured, using qRT-PCR. The expression of both markers was decreased significantly in the C1, compared to the SC3 cells (p < 0.05) (Fig. 4b). Moreover, the expression of NES at the protein level was assessed before and after RA treatment at Day 0 and Day 21. The results showed that the C1 cells have higher expression levels, about 2.35-fold, than the SC3 cells (Fig. 4c). To that end, we first measured the ratio of Day 21/Day 0 for the SC3 and C1, then calculated the ratio of the C1 to the SC3. According to this, the expression of NES was increased 2.3-fold in the C1 compared to the SC3 cells. This observation confirmed that in the C1 cells, the number of undifferentiated cells is more than in the SC3 cells. To study whether the effect of decreased expression of GPM6B on differentiation of NT2 cells to neural cells (under RA treatment) is due to deletion of (GA)9, the expression levels of GFAP, TUBB3, and MAP2 were evaluated at the protein level. The results indicated that GFAP, TUBB3, and MAP2 expression decreased 0.77-, 0.57-, and 0.2-fold compared to the SC3 (Fig. 4b). Based on these findings, the reduced expression level of GPM6B, due to the specific deletion of (GA)9, consequently decreased the differentiation of NT2 cells under RA treatment.
To count differentiated cells in each experiment under RA treatment, specific cell markers such as NES, GFAP, TUBB3, and MAP2, were labeled, using specific antibodies and sorted by flow cytometry on Day 0 and Day 21. The number of SC3 and C1 cells expressing NES significantly increased compared to Day 0 (p < 0.05 and p < 0.001, respectively). Moreover, the number of cells expressing NES rose substantially in the C1 cells compared to the SC3 cells on Day 21 (p < 0.01). The number of undifferentiated cells in the C1 was more than in the SC3, and confirmed that a decreased level of GPM6B reduced the differentiation of NT2 cells to neural cells. The number of cells expressing GFAP on Day 21 significantly increased at the SC3 cells (p < 0.001) but not at C1, compared to their counterpart on Day 0. Furthermore, the number of the SC3 cells expressing GFAP was more significant than the ones in the C1 cells on Day 21 (p < 0.001). In line with these premises, it can be concluded that decreased levels of GPM6B inhibited differentiation of the NT2 cells to the neural cells expressing GFAP. TUBB3 is an immature protein that detects the differentiated neural cells at the early stage. Our findings indicated that the number of cells expressing TUBB3 on Day 21 was significantly increased compared to the cells at Day 0 in the SC3 (p < 0.001) and the C1 (p < 0.01). In addition, the number of differentiated cells expressing TUBB3 decreased significantly in the C1 compared to the SC3 (p < 0.05). MAP2 is a mature-specific cell marker to determine late-stage differentiated neural cells. Like the findings for TUBB3, the number of cells expressing MAP2 was significantly increased in the SC3 and the C1 cells on Day 21 compared to the cells on Day 0 (p < 0.0001 and p < 0.01, respectively). Moreover, the number of differentiated cells expressing MAP2 decreased significantly in the C1 compared to the SC3 cells (p < 0.01) (Fig. 5). Taken together, our findings elucidated that the reduced expression level of GPM6B due to the deletion of the GA-repeat significantly reduced NT2 cell differentiation to neural cells.