RPE Differentiation is a Selective Barrier Against Aneuploid hESC

doi:10.21203/rs.3.rs-5083824/v1

Download PDF

Article

RPE Differentiation is a Selective Barrier Against Aneuploid hESC

https://doi.org/10.21203/rs.3.rs-5083824/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

In this study, we show that undirected differentiation of human embryonic stem cells (hESC) to retinal pigment epithelium (RPE) acts as a selective barrier against aneuploid cells. Large-scale omics analysis reveals that 3–6% of cells of genetically normal hESC cultures are aneuploid, none of which progresses through RPE differentiation except for cells with a gain of 1q. We show that while all homogeneously aneuploid hESC lines carrying an array of different abnormalities have impaired RPE differentiation, co-culture with genetically normal cells specifically rescues the differentiation of cells with a gain of 1q. In turn, these aneuploid cells have an in vitro growth advantage, and progressively take over the differentiating culture. Time-course analysis shows that, only when co-cultured, cells with a gain of 1q follow the same differentiation path as genetically normal cells, which support the mutant cells by secreting extracellular matrix and ligands promoting differentiation.

Biological sciences/Stem cells/Embryonic stem cells

Biological sciences/Genetics/Genomic instability

Over 50 clinical trials are ongoing or have recently been completed involving the transplantation of cells derived from human pluripotent stem cells (hPSC), all of which are phase I/II trials with a focus on safety (www.clinicaltrials.gov)¹. Currently, the majority involve the use of retinal pigmented epithelium (RPE) cells derived from hPSC to restore or improve vision in patients suffering from retinal degenerative diseases including age-related macular degeneration and Stargardt’s macular dystrophy^1–3.

A remaining safety concern of these treatments is the susceptibility of hPSC to accumulate genomic abnormalities. The most common abnormalities involve segmental or full gains in chromosomes 1, 12, 17 and 20^4,5, but also include point mutations^6–8 and epigenetic changes⁹, often resembling mutations found in cancers^7,8,10. Despite their recurrence, insufficient is known about their downstream functional consequences, obscuring their impact on both the research and the clinical applications of hPSC^11–13. Only two of these recurrent abnormalities have been thoroughly characterised in the undifferentiated state. Gain of chromosome 12 results in increased hPSC proliferation rates and altered transcriptomic profiles, likely due to NANOG over-expression¹⁴. Higher levels of Bcl-xL in hPSCs with a gain of 20q11.21 result in decreased sensitivity to apoptosis¹⁵, which confers a survival advantage to the cells¹⁶. Regarding their impact on differentiation, hPSCs with a gain of 20q11.21 have impaired neuroectoderm commitment without affecting mesendoderm induction^17,18 and it has been shown that cells with an isochromosome 20q fail to survive RPE differentiation¹⁹. Even when chromosomally abnormal hPSC are capable of differentiating, they display altered gene-expression patterns suggestive of malignant transformation^{4,11,14,20,21}. Our current understanding of how these mutations impact the oncogenic potential of hPSC-derived cells remains limited. Research on this topic has predominantly focused on tumor formation by either residual undifferentiated cells or highly proliferative progenitor cells in the differentiated cell product¹⁰. Substantial effort has been devoted to developing methods for generating highly pure cell populations and avoiding the presence of any undifferentiated hPSC in the final product^11,22. Conversely, the abnormalities seen in hPSCs could be regarded as a first hit in cancerous transformation. Transplanted cells with genetic abnormalities could be more likely to undergo oncogenic transformation, requiring fewer genetic hits to initiate the process^13,22,23. In line with this, recent work in mice has shown that aneuploidy drives teratoma metastasis, with multiple organ dissemination²⁴.

Currently, all clinical centers subject their hPSC cultures and hPSC-based products to genetic screening prior to use in patients. This typically involves the use of G-banding and, more recently massively parallel sequencing (MPS), which enables the detection of copy number variation (CNV) as well as potentially harmful single nucleotide changes. Incidentally, the first clinical trial using hPSC-derived RPE was halted after potentially harmful mutations were identified in both the hPSC and the RPE cells derived from them²⁵. It is important to bear in mind that standard methods for genetic screening cannot detect low-grade mosaic abnormalities in hPSC cultures, which are common^26–29. Since specific abnormalities may lead to poor differentiation or increased malignant potential, their presence in a low-grade mosaic could result in the transplantation of a cell product that, at best, has decreased effectiveness, and at worst, has tumour-initiating properties.

In this study, we hypothesized that while some CNVs render hPSC refractory to correct differentiation, others may confer a growth advantage during cell specification, leading to the enrichment of specific mutant hPSC during differentiation. We focussed on investigating these events during RPE differentiation, to gain insight into how the genetic make-up of cells impacts their progression through differentiation in a clinically relevant cell type.

Genetically balanced hESC lines differentiate into RPE cell populations with unique gene-expression signatures containing astrocytes and pigmented ciliary body cells

In this work, we differentiated in-house genetically balanced hESC lines into RPE using an unguided differentiation protocol³⁰(Supplementary Table S1 and Supplementary Figure S1). After 4 to 8 weeks, pigmented areas appeared in the dish, indicating clusters of cells that differentiated spontaneously into RPE. These clusters were picked and passaged to yield a pure RPE cell culture (Fig. 1a and 1b). For this study, we were first interested in establishing if these clusters are of clonal origin, as this would generate a genetic bottleneck, the size of which we could modulate depending on the number of clones picked to establish the final RPE culture. For this, we differentiated an hESC line labelled with the RGB lineage tracing system³¹. Eight weeks into differentiation, we found that the pigmented clusters of cells were predominantly of clonal origin, as indicated by the expression of the same color within the cluster (Fig. 1c). We next differentiated five genetically balanced hESC lines to RPE, and purified the cultures through colony picking and subsequent passaging. Following two passages, the cultures consisted of homogeneous layers of RPE cells, characterized by their typical cobblestone morphology (Fig. 1d). The cultures were stained for markers of neuroectoderm (PAX6), pluripotency (NANOG) and RPE (PMEL, ZO-1 and BEST1 in Fig. 1e, all lines in Supplementary Figure S2). All RPE cultures were fully negative for NANOG while ZO-1 and PMEL were expressed by all cells. All cultures presented two cell populations, PAX6^high/ BEST1^low and PAX6^low/BEST1^high, reflecting different levels of maturation, in line with other studies^32,33.

We carried out single-cell RNA sequencing (scRNAseq) for the five hESC-derived RPE cultures, which yielded good quality results for 87168 individual cells, and utilized previously generated scRNAseq data of one hESC line as a reference for the undifferentiated state (N = 1996). The UMAP of all RPE cells shows that each line predominantly remained in its own cluster, with four small clusters originating from multiple cell lines (Fig. 1f). We first assessed the gene expression of RPE cells relative to that of undifferentiated hESC, to test the induction of RPE-related genes and residual pluripotency (Supplementary Figure S3a). We found that the five RPE cultures expressed pigment synthesis, visual cycle, phagocytosis, ion channel and tight junction-associated gene sets (Fig. 1g, Supplementary Table S2³⁴). We next integrated our dataset into that of Senabouth et al, that includes scRNAseq of RPE derived from 79 different iPSC lines (Fig. 1h³²). The results show that our cells express RPE markers at similar levels to those reported in this large study. In addition, and consistent with the PAX6/BEST1 immunostaining (Fig. 1e), the UMAPs show the presence of two distinct populations of RPE based on their maturation state in both datasets. Importantly, no RPE cells exhibited residual undifferentiated state gene expression (Fig. 1g and Fig. 1h). Taken together, cell morphology, immunostaining results and gene-expression profiling all indicate that our cells were efficiently differentiated into RPE cells.

In the next step of transcriptomic analysis, we analyzed the differences in gene-expression between the clusters of RPE cells originating from different hESC lines. We carried out differential gene expression analysis of each cluster against the other four and identified a set of 1206 genes that drive the differences between lines. The Venn diagram in Fig. 1i shows that the RPE derived from each hESC line expresses a specific set of genes, with few overlaps and each line having between 12 and 32% unique deregulated genes. The heatmap of differentially expressed genes in Fig. 1j illustrates that each cell line has its own transcriptomic signature, with some sets of genes being specifically higher or lower expressed per cell line. Gene set enrichment analysis for each of these gene sets revealed no enrichments for pathways related to RPE functionality or stem cell differentiation (Supplementary Table S3). Plotting the cells in a UMAP after excluding the genes driving the cell-line specific differences reduced the distance between the line-specific clusters, while the small clusters composed of cells from different cell lines remained apparent (Fig. 1k). The observed differences between the RPE are likely of minor significance and may be due to differences in the genetic and epigenetic landscape of individual cell lines^35,36.

Next, we focused on determining the identity of the four clusters composed of cells from different lines and that had transcriptomic signatures that differed from RPE cells (Fig. 1l). Cluster 2 represents 1.32% of the cells and contains RPE cells originating from all 5 hESC cell lines. This cluster shows a high expression of proliferation marker MKI67, and cell cycle state analysis showed that 47% of these cells were in G2/M and 53% in S phase, indicating that they are proliferating cells (Fig. 1m, Supplementary Figure S3b and S3c). Using UCell scoring of Seurat, we assigned to each cluster a score related to different neural and ocular cell types (Supplementary Table S4³⁷). The RPE clusters of VUB02, VUB04, VUB07, VUB14 and VUB32, as well as cluster 2, show the highest scores for the RPE gene set. Clusters 1 (1.40% of cells, originating from the 5 lines), 3 (0.63% of cells, originating from 3 lines) and 4 (0.29% of cells, originating from the 5 lines) also express RPE related genes, but at lower levels than the main RPE clusters. Conversely, clusters 1 and 3 show high expression of astrocyte markers, and cluster 3 also shows high scores for Schwann and Mueller glia cells. Finally, cluster 4 shows expression of genes marking pigmented ciliary body cells, a cell type related to the RPE lineage (Fig. 1n, Supplementary Figure S3d,).

RPE differentiation is a genetic bottleneck eliminating aneuploid cells except for cells with gains of 1q

We next investigated the genetic composition of the cell cultures at the single-cell level both before and following RPE differentiation. We performed single-cell DNA sequencing (scDNAseq) on three of the five hESC lines, at the start of the differentiation process. We were unable to perform scDNAseq on the two remaining lines and on the differentiated cells due to the discontinuation of the essential 10x Genomics products required for this specific analysis. Therefore, we employed the inferCNV algorithm on the scRNAseq data obtained from RPE cells derived from all five hESC lines to assess the genetic content at the endpoint of differentiation.

The analysis yielded high-quality scDNAseq results for a total of 1668 cells, including 954 cells from VUB02, 274 cells from VUB04, and 440 cells from VUB07. Their karyotypes are illustrated in Fig. 2a-c. Overall, single cells could be broadly categorized into different groups based on their genetic content. These groups included cells with genetically balanced content, cells with chaotic genetic content (defined as multiple abnormalities spanning more than 10% of the genome), and cells with gains or losses of other sizes. We further subdivided the abnormal non-chaotic cells based on whether the CNVs were of a size detectable by inferCNV on scRNAseq, with a detection threshold of 20 Mb. This subdivision was introduced to facilitate comparisons between the outcomes of these two distinct analyses. Figure 2d provides an illustrative representation of diverse cell types, including chaotic cells, cells with gains in chromosome 1q, cells with isochromosome 20, and cells with segmental gains in chromosomes 15q and 20q.

On average, 95% of cells were genetically balanced, with only minor variations observed between the three hESC lines (Fig. 2e). Interestingly, cells with chaotic genetic content were found in all three lines. These chaotic cells were similar to those found in human preimplantation embryos, where they originate from abnormal mitotic events with multipolar spindles^38–40. To investigate whether similar abnormal cell divisions occurred in hESC, we conducted live-cell imaging for 24 hours to visualize DNA and microtubules (Supplementary Video 1). Figure 2f shows two still images depicting the progression of a cell through tripolar spindle mitosis, with additional examples available in the supplementary video. This evidence confirmed that hESC undergo abnormal multipolar mitosis, resulting in daughter cells with chaotic genetic content as seen in scDNAseq.

The distribution and location of all gains and losses found in the cells (excluding chaotic cells) are depicted in Fig. 2g, and a comprehensive list of breakpoints is provided in Supplementary Table S5. While cells with a single loss were twice as common as cells with a single gain, overall, gains outnumbered losses (68 gains vs. 40 losses). Notably, all three hESC lines contained single cells with genetic abnormalities known to confer a growth advantage, potentially leading to culture dominance by cell competition. VUB02, for instance, included cells with a trisomy 12, a trisomy 20, a dup(1)(q32.1q44), and a del(18)(q21.2q22.1), whereas VUB04 had cells with a dup(1)(q21.3q44) and an isochromosome 20. VUB07 contained cells with a dup(20)(q11.21q13.2).

Remarkably, the inferCNV analysis at the endpoint of RPE differentiation revealed that, in contrast to undifferentiated hESC, the only genetic imbalance we identified in RPE cells was represented by different-sized gains of chromosome 1q, which were present in VUB04 and VUB07 and distributed across the cell clusters (Figs. 2h-l). VUB04 exhibited 42.8% of RPE cells with a 77.8 Mb gain of 1q21.3q44 and 2.95% of cells with a 20.4 Mb gain of 1q21.2q24.2. For VUB07, 2.66% of the cells displayed an 82.0 Mb gain of 1q21.3q44. Additionally, the clusters representing cells mis-specified as astrocyte and ciliary body-like cells contained 12.8% of cells with the 77.8 and 82.0Mb 1q gain and 28.7% of cells with the shorter 20.4 Mb 1q gain. Notably, cells with 1q gain showed no preference for allocation to the mis-specified cells, suggesting that this aneuploidy may not be the primary driver for their appearance (Fig. 2m).

Our results revealed a remarkable absence of any other aneuploidies in the RPE, which was widely different from undifferentiated hESC. Consequently, we conducted a probability calculation to determine the likelihood that this observation is the result of randomly picking only genetically balanced RPE colonies. To obtain pure RPE cultures, we pick approximately 60 individual colonies per line, which appear to be predominantly clonal in origin (Fig. 1c). Assuming that all aneuploid cells have equal probabilities of successfully differentiating to RPE and have no competitive advantage over genetically balanced cells, and based on the aneuploidies we found in the undifferentiated cells by scDNAseq, the chances of picking at least one aneuploid clone that would have shown in the inferCNV analysis are 47.8% for VUB02, 96% for VUB04 and 89.7% for VUB07. Considering these probabilities, the absence of aneuploidies other than the gain of 1q in RPE cells suggests that not only the differentiation to RPE represents a bottleneck with a strong selective pressure against most aneuploid cells, but that the gain of 1q may be conferring a culture advantage to the cells during this differentiation process.

Gain of 1q results in discreet transcriptional changes in RPE affecting the apoptosis pathway and eye disease-related genes

Since a large proportion of the RPE cells carried a large gain of 1q (RPE^1q), we aimed to investigate the potential differences in the transcriptome of individual RPE^1q in comparison to their euploid counterparts (RPE^wt). Cells with a smaller gain of 1q were excluded from this analysis due to their limited representation in the dataset.

We started by assessing the expression of a comprehensive set of genes typically associated with RPE cells. This initial analysis shows that RPE^1q consistently express all expected marker genes (Fig. 3a). The extent of their expression varied, and while all genes except MITF showed statistically significant differences between the two groups, some genes were expressed at higher levels while others at lower levels in RPE^1q cells. This suggests that these cells differentiated to bona fide RPE, similarly to the genetically balanced cells.

To gain further insights into the potential impact of the 1q gain on RPE cells, we conducted single cell Gene Set Variation Analysis (scGSVA) focused on the KEGG pathways. Four illustrative examples of pathway outcomes are presented in Fig. 3b (the remaining data is available in the Supplementary Datafile 1). It is noteworthy that although many of these pathways reached high statistical significance, the differences in normalized enrichment scores’ mean were minimal. We sought to explore whether the observed differences in gene expression were consistently linked to the presence of the 1q gain, or if they were influenced by inherent variations between cell lines. For this, we conducted scGSVA analyses on individual cell lines, as well as a comparative analysis of RPE^1q and RPE^wt cells in VUB04 and VUB07, as shown in Fig. 3c (the remaining data is available in the Supplementary Datafile 1).

In this more detailed comparison, we looked for consistent differences between the two groups, particularly in the context of KEGG pathways. However, it became evident that except for the apoptosis library their profiles did not consistently diverge. These differences were not consistently statistically significant, and the enrichment scores frequently shifted in opposite directions. The cumulative view encompassing all cell lines suggested that the observed variations were primarily attributable to line-to-line variation rather than being directly associated with the presence of the 1q gain.

The apoptosis pathway appeared as an exception, with consistent enrichment scores in the two isogenic pairs of VUB04 and VUB07. This consistency prompted us to look further into the roles of the differentially expressed genes within the apoptosis pathway (Fig. 3d). Our analysis revealed that only a subset of genes in the apoptosis pathway were expressed in RPE cells, and these genes did not all consistently differ between RPE^1q and RPE^wt cells. Interestingly, three genes involved in the execution of the apoptotic process, PARP1, MCL1 and LMNA, were significantly upregulated in RPE^1q cells. These three genes are all located within the region affected by the 1q gain, implying that their increased copy number likely contributed to their elevated expression and therefore to the deregulation of the broader apoptosis pathway. Given that these effectors often have opposing roles, predicting their overall biological impact remains challenging.

Lastly, we extended our analysis by exploring the disease ontology of our gene set in the context of human diseases relevant to RPE cells. Four sets of diseases were selected for analysis (Fig. 3e). The results indicated that RPE cells with a gain of 1q exhibited differential gene expression patterns related to eye diseases, retinal degeneration, eye degenerative diseases, and degeneration of the macula and posterior pole. Upon closer examination, we identified six genes that consistently and significantly differed between RPE^1q and RPE^wt (Fig. 3f and supplementary table 8). COL18A1 and CYBA exhibited lower expression in RPE^1q, while TMCO1, ATP6AP2, RPE65, and HTRA1 displayed higher expression. We explored the published roles of these genes in RPE function and found that their expression pattern did not suggest dysfunction or poor differentiation in RPE^1q cells. While downregulation of COL18A1 would be detrimental for RPE ⁴¹, downregulation of CYBA⁴² and upregulation of TMCO1 ⁴³, ATP6AP2 ⁴⁴, RPE65⁴⁵ and HTRA1⁴⁶ would be beneficial for the RPE function and survival.

Co-culture with genetically balanced cells cannot rescue the impaired RPE differentiation of aneuploid hESC except in cells with gain of 1q

In the subsequent experiments, we aimed to test the hypothesis that low-grade mosaic aneuploid hESC observed in genetically normal cultures do not contribute to the differentiated progeny due to their inability to undergo RPE differentiation. For this, we used three wt control lines and nine hESC lines with different aneuploidies recurrently present in hPSCs to undergo RPE differentiation. The differentiations were conducted in 1–7 independent differentiation experiments (the list of lines and replicates can be found in Supplementary Table S6).

After a 12-week differentiation period, we found that only the line with the smallest 1q gain (VUB03^1q32.21) had pigmented areas characteristic of RPE differentiation, while two hESC lines with a larger 1q gain (VUB03^1q21.1qter and VUB01^1q21.1qter) showed no visible pigmentation. Similarly, the lines with gain of 20q11.21, isochromosome 20q, gain of 17q, and derivative chromosomes involving the loss of 18q consistently failed to differentiate into RPE (Fig. 4a shows a representative example, additional images can be found in Supplementary figure S4). At the colony picking stage for RPE purification, we analyzed the mRNA expression levels of RPE markers RP65, BEST1, PAX6, and the undifferentiated state marker POU5F1. The results indicated that genetically balanced control lines induced the RPE gene-expression signature, whereas aneuploid cell lines either did not initiate the expression of these genes or did so significantly less than control cells (Fig. 4b). Collectively, these findings demonstrated that, except for the small focal gain of 1q32.21, the tested genetic abnormalities impeded the correct differentiation of hESC into RPE cells.

Given that in our first experiments we had identified correctly differentiated RPE harboring large gains of the 1q arm mixed into the genetically normal cells, we decided to carry out the differentiation of the 1q cells in co-culture with genetically balanced cells, which we hypothesized would promote the correct differentiation of the 1q cells. In the co-culture experiments, we introduced between 0.2 to 0.5% of two fluorescently labeled hESC lines with a gain of 1q alongside genetically balanced isogenic cells (N = 6 replicates). The differentiation progression was monitored by regular live fluorescent confocal imaging, which showed that 1q cells rapidly started outcompeting their genetically balanced counterparts in the dish (full dish images are shown in Fig. 4c). Pigmented colonies were picked at 12 weeks after initiation of differentiation, expanded, and passaged, and we used flow cytometry to quantify and to isolate the RPE^1q cells. The results show that at the endpoint of differentiation, the cell population with a gain of 1q had increased between 45 to 124-fold, accounting for 9%-24.9% of the differentiated cells. Moreover, these RPE cells expressed the RPE markers ZO-1, PMEL, PAX6 and BEST1 similarly to control cells (Fig. 4f), and were harvested for bulk RNA sequencing, along with the RPE obtained from VUB03^1q32.21 (shown in Figure S4).

Finally, we tested whether genetically balanced cells could rescue the differentiation of aneuploid hESC other than those with 1q gain. We mixed 1% of GFP-expressing hESC with a gain of 20q11.21 or with an isochromosome 20q with its genetically balanced and unlabelled counterpart and subjected them to RPE differentiation. We regularly imaged the cells and observed a steady increase in fluorescent signal until the pigmented colonies started appearing, after which the signal decreased (Fig. 4d). This suggested that the fraction of 20q11.21 and isochromosome 20q cells initially increased during the first part of spontaneous differentiation to then decrease upon RPE specification. We collected an entire dish, prior to RPE colony picking and tested for the presence of cells with a gain of 20q using a copy number assay. This revealed that in both cell competitions the mutant cells had taken over most of the dish, while silencing the fluorescent protein construct. We purified the RPE by colony picking and cultured until obtaining a homogeneous population. The cells were stained for ZO-1 and PMEL, and co-imaged for GFP to identify the 20q11.21 or isochromosome 20q cells (Fig. 4e). The results show that, respectively, 3 and 5% of the RPE cells consists of GFP⁺ aneuploid cells and that they do not express PMEL in the same manner as their genetically balanced counterparts, and that they represent a contamination of the pure RPE cell population derived from the control cells. Taken together, this suggests that both 20q11.21 and isochromosome 20q cells overgrow the genetically balanced cells during spontaneous differentiation, but do not specify to RPE but to an alternate cell fate.

RPE^1q show transcriptomic signatures of aneuploidy-related stress

We next studied by bulk RNA sequencing the transcriptome of the three RPE^1q cell lines we isolated in the previous experiments and compared it to that of RPE from control cells. We included triplicates for each RPE^1q line (N = 9 samples) and ten control samples (5 hESC lines, 2 replicates per line). The principal component analysis (PCA) (Fig. 5a) revealed a closer clustering pattern among RPE^1q samples, suggesting potential differences in their transcriptomic profiles. We first focused on examining the induction patterns of RPE markers in both groups compared to undifferentiated hESC. The results, shown in Fig. 5c, showed similar induction for the RPE differentiation genes, along with the simultaneous downregulation of pluripotent-state markers, in both control and 1q cells. This was in line with their similar RPE morphology and the stainings shown in Fig. 4d and confirmed that hESC^1q do have the ability to generate RPE cells similarly to their genetically balanced counterparts, though only when in co-culture with them.

We further explored potential differences between the two groups by carrying out differential gene expression. The results show that RPE^1q have higher expression of 372 genes and lower expression of 1183 genes, as compared to their RPE^wt counterparts (Fig. 5b). Gene set enrichment analysis (GSEA) showed that the RPE^1q cells had, in line with the findings in the scRNAseq, deregulation of the apoptosis pathway (Fig. 5d, Supplementary Datafile 2 and 3). Other pathways that were enriched were the unfolded protein response, DNA repair and p53 pathways. These pathways are known to associate to intracellular stress response due to aneuploidy^47–50, and may contributing to the deregulation of apoptosis-related genes we see here and in the scRNAseq data. It is interesting to note the negative enrichment score for the epithelial-to-mesenchymal transition (EMT) genes, an important process in RPE maturation^33,51. RPE^1q also showed an overall diminished expression of collagen-related genes, including genes not only the collagen genes themselves but also genes involved in their biosynthesis, trimerization and degradation (Fig. 5e, Supplementary Datafile 3).

Co-culture of 1q and wt cells rescues the differentiation trajectory within two days after growth factor removal

We hypothesized that the impact of the genetically balanced cells on differentiation of 1q cells when in co-culture would be detectable in the first days after growth factor withdrawal, as cell fate decisions are taken very early in the differentiation process. To test our hypothesis, we carried out scRNA sequencing of undifferentiated and spontaneously differentiating cells (day 1 and day 2), including a wt hESC line and its 1q-mutant counterpart, on their own and in a 1:9 1q:wt co-culture. We first controlled the karyotypes of the cells by inferCNV, and found that a small fraction of the wt cells carried a trisomy 20, and consequently the co-culture contained wt, 1q and trisomy 20 cells (Supplementary figure S5). Seurat analysis yielded 7 clusters (Fig. 6a). Differential gene expression analysis showed that this early after FGF2 and TGFBeta withdrawal there was only a very modest induction of differentiation-associated genes in two clusters, most of them of the ectodermal lineage (Fig. 6b). All cells still expressed high levels of pluripotency-associated markers, with none showing expression of canonical early differentiation markers such as TBXT, NES, or PAX6, while some cells did express ectodermal markers such as OTX2 and MAP2 (Fig. 6c). The location of the cells by day of differentiation and by karyotype are shown in Figs. 6d and e, and the composition of each cluster is plotted in Fig. 6f. Overall, one cluster was composed mostly of undifferentiated cells, both wt alone and the co-cultured cells, and another cluster of the undifferentiated trisomy 20 cells. The undifferentiated 1q cells resided in both clusters 1 and 2. The two clusters that induced predominantly ectoderm-associated genes were composed exclusively of day 1 and day 2 cells and were considered the standard route towards neuroectodermal differentiation. Two other clusters were composed of cells from the three days and had expression profiles that were predominantly of undifferentiated cells and with induction of genes that could not be associated to any specific lineage. The 1q cells, when cultured alone, remained in clusters 1 and 2 over the two days after growth factor withdrawal, and trisomy 20 cells transitioned from their undifferentiated cluster to cluster 1. This was in strong contrast to the trajectory of the wt cells, that transitioned from the undifferentiated cell cluster, and to day 1 and then day 2 of ectoderm induction. Remarkably, the 1q cells, when co-cultured, had one part of the cells following the same ectodermal trajectory as wt cells, and the other part transitioning to cluster 2 (Fig. 6g). These results prove that the change in differentiation trajectory induced by the co-culture of 1q cells with wt cells is occurring very early in differentiation, and further supports the findings that the differentiation of other mutant cells (in this case trisomy 20) are not rescued by the co-culture.

Finally, we analyzed ligand-receptor expression patterns in the two scRNAseq data sets. Figure 6h shows the cell-to-cell communication between wt cells and 1q and T20 cells, from the undifferentiated state to day 2, as well as between wt and 1q cells in the purified RPE state (Fig. 6i, full lists can be found in the Supplementary Datafile 4 and 5). We filtered for ligand-receptor pairs in which the ligands are provided exclusively by the wt cells, and the receptors are located on the 1q cells, as we hypothesized that wt cells secrete ligands that the 1q cells do not, and that are key to the differentiation process. From this analysis, two candidate networks emerge as the most interesting candidates: MFGE8 and collagens secreted by the wt cells, binding to the integrin receptors of the 1q cells, and the interaction between NLGN1, predominantly secreted by the wt, and NRXN1 and NRXN3 in the 1q cells. Figure 6j illustrates the progression of the expression of these genes in 1q cells differentiating alone or in co-culture, as well as of wt cells, and show how co-cultured 1q cells closely follow the expression pattern of wt cells, in contrast as when differentiating alone. Figure 6i shows two examples of the lower expression of collagen genes in RPE^1q as compared to RPE^wt. MFGE8 is known to regulate EMT in different cell types, which in turn is a key process in hPSC differentiation^52,53, and collagen is essential for the formation of an extracellular matrix promoting correct RPE differentiation^54,55. This is in line with the observation that both the EMT and the collagen biosynthesis appeared negatively enriched in the bulk RNA sequencing of the RPE (Fig. 5d, e). NLGN1, NRXN1 and NRXN3 are genes known to be involved in late neuroectodermal development ⁵⁶, but in our dataset also appear to be key regulators of the early spontaneous differentiation process. NRXN1 is for instance one of the top induced genes in the ectodermal clusters of day 1 and day 2 of differentiation (Fig. 6b). Together, this suggests these gene networks may be part of the mechanisms behind the impairment of 1q differentiation, as well as the rescue by the wt cells.

In this study we show that spontaneous RPE differentiation is a purifying bottleneck against aneuploid cells, with the notable exception of cells with a gain of 1q. These cells can progress through differentiation while being in co-culture with genetically balanced cells, which promote EMT and neural induction in the mutant cells and provide the correct extracellular matrix for their differentiation.

In the first part of our work, we carried out single-cell high-resolution karyotyping of a cohort of single hESC of unprecedented size, with results for 1674 individual cells originating from cultures that are genetically balanced based on shallow-whole genome sequencing. In comparison, in previous work we used single-cell array-based comparative genomic hybridization to study 60 and 59 individual cells^27,28 and single-cell DNA sequencing on a cohort of 56 cells²⁶. This extended dataset has allowed us to refine the map of aberrations carried by hESC cultures in the form of low-grade mosaicism. We find that our cultures carry between 3 and 6% of cells with chromosomal abnormalities, which included chaotic genetic contents, much reminiscent of those found in cleavage-stage embryos⁴⁰, as well as aneuploidies that are recurrently found to take over hPSC cultures^5,57. Remarkably, all three hESC lines screened carried cells with these aneuploidies, suggesting that all hESC cultures will sooner or later be taken over by any of these abnormalities, the speed of take-over depending on the selective pressure exerted on the cells due to potentially suboptimal culture conditions. Another interesting finding is that we identified aneuploidies in all chromosomes except for 19 and 22, which are also very rarely found in hPSC cultures worldwide⁵⁷, suggesting that they may be particularly deleterious to the cells.

Testing the effect of these chromosomal abnormalities on RPE differentiation, we find that all but one of the aneuploid hESC lines have impaired differentiation, in line with a previous report showing that cells with an isochromosome 20 are unable to differentiate to this cell type¹⁹. Remarkably, only cells with a gain of 1q can progress through differentiation but only when co-cultured with control cells and then even have a selective growth advantage. While these cells appear morphologically and transcriptionally undistinguishable from their genetically balanced counterparts, it is important to note that this does not guarantee that they are functionally fully equivalent. Further, when they are isolated and cultured alone, the cells show signatures of aneuploidy-related stress, such as p53 signaling, unfolded protein response and apoptosis^49,58. The time-course scRNAseq experiments prove that not only are the 1q cells able to follow the same trajectory of differentiation as wt cells when co-cultured with the latter, but also that cells with a trisomy 20 fail to follow the trajectory very early on. We found that the wt cells support the differentiation in different ways, promoting EMT and neural induction in the 1q as well as secreting the needed extracellular matrix, in line with the know key role of extracellular matrix and surrounding cells during RPE and neural differentiation^59,60.

In summary, our results suggest that while all hESC cultures carry small fractions of aneuploid cells, RPE differentiation acts as a purifying bottleneck, except for gains of 1q. This is reassuring as these small fractions of aneuploid cells are virtually undetectable with any of the existing technologies except for scDNAseq making it difficult to find an optimal approach to screening hPSC cultures prior to differentiation, especially in a clinical setting. On the other hand, gains of 1q take over the RPE differentiation and can therefore be detected by thorough genetic testing of the final cell product prior to transplantation. While RPE have a notoriously low tumor-formation rate in vivo⁶¹, and none of the completed clinical trials has reported any oncogenic process associated to hPSC-derived RPE transplantation, our results elicit caution. Even if they are transcriptionally indistinguishable from control cells, the cells with a gain of 1q still may be functionally different, or progress to an abnormal phenotype in the years after transplantation.

Resource availability

Further information and requests for resources should be directed to the corresponding author, Claudia Spits([email protected]).

Materials availability

All VUB stem cell lines in this study, including the genetically abnormal sublines and genetically modified lines, are available upon request and after signing a material transfer agreement.

Data availability

Raw sequencing data of human samples is considered personal data by the General Data Protection Regulation of the European Union (Regulation (EU) 2016/679), because SNPs can be extracted from the reads, and cannot be publicly shared. The data can be obtained from the corresponding author upon reasonable request and after signing a Data Use Agreement. The RNA sequencing count tables and all the data supporting the figures in this paper can be found at the Open Science Framework repository (https://osf.io/y8tzh/).

Ethics statement

For all parts of this study, the design and conduct complied with all relevant regulations regarding the use of human materials, and all were approved by the local ethical committee of the University Hospital UZ Brussel and the Vrije Universiteit Brussel (File number: B.U.N. 1432021000669).

hESCs lines, cell culture, genomic characterization and banking

All hESC lines in this study, were derived in-house in the past. The details on the derivation and results of the characterization, including tests for pluripotency, were reported previously ^62,63 and can be also found at the Open Science Framework repository (https://osf.io/esmz8/). The lines are registered in the EU hPSC registry (https://hpscreg.eu/), and available upon request. Some of the lines were genetically modified to stably express a fluorescent protein, so that they could traced in culture. This was achieved by lentiviral transduction of constructs encoding for mKate, GFP, mCherry, Pacific Blue or Venus. Briefly, Lentiviruses were produced in 293T cells by transfecting plasmids for VSV.G, gag-pol and the plasmid of interest in PEI (1/28) (Sigma-Aldrich) and Opti-MEM (Thermo Fisher Scientific) for 4h. The transfection cocktail was then replaced with complete medium, and the lentivirus-containing supernatant was harvested 48h-72h later and stored in aliquots at -80°C. One day before transduction, hPSC were seeded at a density of 50,000 cells per well of a 6well plate. Cells were then transduced in a transduction cocktail with 1:1000 protamine sulfate (LEO Pharma; 10mg/mL) and a 50:50 mix of Nutristem and lentivirus-containing medium. 800µL of the cocktail was added to each well of a 6-well plate and incubated for 4h. Cells were then washed with PBS 5x before adding fresh Nutristem medium. 24h later cells were again washed with PBS 5x and before being selected for by either puromycin, FACS or both.

Prior to the start of this study, cell working banks were created for each of the lines, which were karyotyped by shallow whole-genome sequencing and controlled for mycoplasma infection. Cells were drawn from the bank for the experiments and used for differentiation either immediately after thawing, or not beyond 6 passages after thawing (Supplementary Table S1 and Supplementary Figure S1).

The genetic content of the hESCs was assessed through shallow whole-genome sequencing by the BRIGHTcore of UZ Brussels, Belgium, as previously described ⁶⁴. Copy number variant analysis for the gain of 20q11.21 was done using quantitative real-time PCR (qRT-PCR). DNA was extracted with a DNeasy Blood and Tissue Kit (Qiagen) according to the manufacturers' protocol. qPCR was performed with the copy number assays: RNaseP (Thermo Fisher Scientific) as a reference and ID1 (Thermo Scientific) for the 20q regions. The reaction systems were prepared by mixing TaqMan 2× Mastermix Plus – Low ROX (Eurogentec), and the TaqMan assays together with the DNA samples. qPCR was performed on a ViiA7 thermocycler (Thermo Fisher Scientific), and Applied Biosystems Copy Caller v.2.1 was used to analyze the CNVs. The assays numbers are listed in the Supplementary table S7.

The hESCs were maintained in NutriStem hESC XF medium (NS medium; Biological Industries) with 100 U/mL penicillin/streptomycin (P/S) (Thermo Fisher Scientific) in a 37°C incubator with 5% CO₂, on Biolaminin 521 coated dishes (Biolamina®). The culture medium was changed daily. The cells were passaged as single cells using TrypLE Express (Thermo Fisher Scientific) and split when reaching 70–90% confluence. The medium was supplemented with 10 µM Rho kinase (ROCK) inhibitor Y-27632 (ROCKi, Tocris) for the first 24 h after passaging.

RPE differentiation

The cells were differentiated as described in ³⁰. HESC were cultured to confluence on Biolaminin-521®. Then, the cells were washed with PBS and medium was changed to NutriStem® hPSC XF GF-free with media changed every day. After 4 weeks, pigmented areas started to develop into the dishes. Following 8 weeks of differentiation, pigmented areas were mechanically cut out using a sharpened glass pipette. Then, cells were dissociated in accutase® for about 2h. Cells were seeded on Biolaminin-521® coated dish and fed twice a week with NutriStem® hPSC XF GF-free. When passaged, the cells were dissociated using accutase® for 1h30-2h and then through a 30 µm strainer and seeded at a density of 50’000-100’000 cells/cm².

Total RNA isolation, cDNA synthesis and quantitative real-time PCR (qRT-PCR) for gene expression analysis

Total RNA was isolated using RNeasy Mini and Micro kits (Qiagen) following the manufacturer’s guidelines, including on-column DNase I treatment. mRNA was reverse-transcribed into biotinylated cDNA using the First-Strand cDNA Synthesis Kit (Cytiva) with the NotI-d(T)18 primer. Quantitative real-time PCR (qRT-PCR) was carried out using TaqMan mRNA expression assays (Thermo Fisher Scientific, listed in Supplementary Table 5) and TaqMan 2× Mastermix Plus – Low ROX (Eurogentec) on a ViiA 7 thermocycler (Thermo Fisher Scientific) using the standard settings provided by the manufacturer. The relative expression was determined by the comparative threshold cycle (Ct) method and GUSB was used as the housekeeping gene. The assays numbers are listed in the Supplementary table S7.

Staining

RPE cells were fixed in 3.7% paraformaldehyde for 15 min, permeabilized in 0.1% Triton-X-100 for 10 min and blocked with 10% fetal bovine serum (FBS) for 1h at room temperature. Between each step, the cells were washed 3x with PBS for 5 min. The primary antibodies were diluted in 10% FBS and kept at 4°C overnight. The next day and after 3x washing with PBS for 5 min, secondary antibodies were diluted 1:200 in 10% FBS for 1h at room temperature. Nuclei were stained using a 1:1000 Hoechst dilution in PBS for 15 min. Finally, the cells were washed 3x in PBS for 5 min. Confocal images were acquired under an LSM800 (Carl Zeiss) confocal microscope at 20x magnification. Supplementary Table S8 lists the antibodies used in this study.

Live imaging

For live imaging of mitoses in hESCs, 75 000 cells per well were seeded on glass bottom µ-Plate Black 24-well plates (Ibidi) in Nutristem. NucBlue™ Live ReadyProbes™ Reagent (Invitrogen, Thermo Fisher Scientific, 1 drop/10mL) and SiR-Tubulin (Spirochrome, 100nM) were added to the culture the next day to visualize DNA and microtubules. After 1h incubation and without washing away the live imaging probes, the cells were imaged every 10–15 minutes for 15-18h using a laser-scanning LSM800 confocal microscope (Zeiss) equipped with a fitted on-stage incubator in 5% CO₂ at 37°C.

Single-cell DNA sequencing

hESC cultures were washed at least 3 times in PBS to remove all cell debris. The cells were dissociated using TrypLE™ for 15 min and strained through a 20µm cell strainer and centrifuged for 5min at 1000 rpm. Single cell libraries were generated using Chromium Single Cell CNV kit according to manufacturer’s instructions and sequenced on a Novaseq (Illumina) with a depth of 500k reads per cell. The fastq files were generated using cellranger-dna mkfastq (cellranger version 1.1.0) from the BCL file. The fastq were aligned to to the Genome Reference Consortium Build 38 (GRCh38) using cellranger-dna cnv. The CNV were detected using R package Aneufinder (version 3.14). Cell’s sequencings were kept only if Bhattacharyya distance ≥ 1 and a spikiness ≤ 0.20. From those, only CNV ≥ 10Mb were kept. The heatmaps were generated using copynumber R package (version 3.15).

Single-cell RNA Sequencing

The RPE cells were washed at least 3 times in PBS to remove cell debris and dissociated using Accutase for 1h30-2h. The cells were strained through a 20µm cell strainer and centrifuged for 5min at 1000 rpm. Cells were resuspended in culture medium. Single cell libraries were generated using the 10X Chromium Controller (10X Genomics) according to manufacturer’s instructions. Approximately 20k cells were sequenced per time point to minimize the occurrence of duplets. Sequencing was performed on a NovaSeq (Illumina) with 20k reads per cell (except one sample at 40k reads per cell). The fastq files alignment, filtering barcode counting and UMI counting were performed using CellRanger 3.1.0 (10X Genomics). The reads were aligned to the Genome Reference Consortium Build 38. The different 10X genomics runs were aggregated using the CellRanger aggr pipeline. The processed scRNAseq were analyzed using (version 4.3) with Seurat R package (version 4.3). Included cells in the analysis had nFeature count between 500 and 8000 and percent of mitochondrial genes < 20%. The cell-cell variation was regressed based on the nCount_RNA, percentage of mitochondrial gene content. The differential gene expression analysis was performed using the FindMarkers function from Seurat. The scGSVA package and scGSVA package.

InferCNV

First, we isolated a group of genetically balanced cells to use them as reference in InferCNV. We used an iterative process for each cell line. Cluster 0 was used as reference and the other clusters as test group. We isolated the subclusters that showed no sign of CNV and removed the ones showing CNV. Then, we did the same process for cluster 1, etc. Then, the genetically balanced subclusters set were compared to see if they were the same. When this was done for each cell line, we ran InferCNV with all the references to see if they were genetically balanced in a cell line specificity. Then, we used the references composed of a set of cells originating from different cell lines and ran inferCNV with it versus the other cells to identify their genomic content.

ACKNOWLEDGMENTS

Y.L. is a predoctoral fellow supported by the China Scholarship Council (CSC), and M.R., C.J., N.K. and E.C.D.D. are predoctoral fellows supported by the Fonds voor Wetenschappelijk Onderzoek Vlaanderen (FWO, grant numbers 1133622N, 11H9823N, 1169023N and 1S73521N respectively). This research was supported by the FWO (grant number G0713222N) and the Methusalem Grant to Karen Sermon and Claudia Spits (Vrije Universiteit Brussel).

AUTHOR CONTRIBUTIONS STATEMENT

ECDD carried out all bioinformatic analysis and all wet-lab experiments unless stated differently

YL carried out the RPE differentiation experiments of all aneuploid cell lines

NK generated the fluorescently labelled hESC lines and assisted with the cell culture work

DAD assisted with the cell culture work

AH assisted with the cell culture work and the quantitative real-time PCR

CJ carried out the live imaging of hESC

MR assisted with the confocal microscopy imaging

SV carried out the flow cytometry quantifications

LAVG advised during the design and progress of the study and provided access to the flow cytometer

OT assisted with the genetic analysis of the cell cultures

KM 10X Genomics Library preparation

KS provided funding for the study

CS designed, supervised and funded the study.

All authors contributed to the drafting of the manuscript.

COMPETING INTERESTS STATEMENT

The authors declare no competing interests.

Kobold, S. et al. A Manually Curated Database on Clinical Studies Involving Cell Products Derived from Human Pluripotent Stem Cells. Stem Cell Reports 15, 546–555 (2020).
Maeda, T., Sugita, S., Kurimoto, Y. & Takahashi, M. Trends of Stem Cell Therapies in Age-Related Macular Degeneration. Journal of Clinical Medicine 2021, Vol. 10, Page 1785 10, 1785 (2021).
Vitillo, L., Tovell, V. E. & Coffey, P. Treatment of Age-Related Macular Degeneration with Pluripotent Stem Cell-Derived Retinal Pigment Epithelium. https://doi.org/10.1080/02713683.2019.1691237 45, 361–371 (2019).
Nguyen, H. T., Geens, M. & Spits, C. Genetic and epigenetic instability in human pluripotent stem cells. Hum Reprod Update 19, 187–205 (2013).
Amps, K. et al. Screening ethnically diverse human embryonic stem cells identifies a chromosome 20 minimal amplicon conferring growth advantage. Nat Biotechnol 29, 1132–44 (2011).
Merkle, F. T. et al. Human pluripotent stem cells recurrently acquire and expand dominant negative P53 mutations. Nature 545, 1–11 (2017).
Avior, Y., Lezmi, E., Eggan, K. & Benvenisty, N. Cancer-Related Mutations Identified in Primed Human Pluripotent Stem Cells. 28, 10–11 (2021).
Merkle, F. T. et al. Whole-genome analysis of human embryonic stem cells enables rational line selection based on genetic variation. Cell Stem Cell 29, 472–486 (2022).
Bar, S. & Benvenisty, N. Epigenetic aberrations in human pluripotent stem cells. EMBO J e101033 (2019) doi:10.15252/embj.2018101033.
Ben-David, U. & Benvenisty, N. The tumorigenicity of human embryonic and induced pluripotent stem cells. Nat Rev Cancer 11, 268–277 (2011).
Keller, A. et al. Genetic and epigenetic factors which modulate differentiation propensity in human pluripotent stem cells. Hum Reprod Update 24, 162–175 (2018).
Halliwell, J., Barbaric, I. & Andrews, P. W. Acquired genetic changes in human pluripotent stem cells: origins and consequences. Nat Rev Mol Cell Biol (2020) doi:10.1038/s41580-020-00292-z.
Andrews, P. W. et al. The consequences of recurrent genetic and epigenetic variants in human pluripotent stem cells. Cell Stem Cell 29, 1624–1636 (2022).
Ben-David, U. et al. Aneuploidy induces profound changes in gene expression, proliferation and tumorigenicity of human pluripotent stem cells. Nat Commun 5, 4825 (2014).
Nguyen, H. T. et al. Gain of 20q11.21 in human embryonic stem cells improves cell survival by increased expression of Bcl-xL. Mol Hum Reprod 20, 168–77 (2014).
Avery, S. et al. BCL-XL Mediates the Strong Selective Advantage of a 20q11.21 Amplification Commonly Found in Human Embryonic Stem Cell Cultures. Stem Cell Reports 1, 1–8 (2013).
Markouli, C. et al. Gain of 20q11.21 in Human Pluripotent Stem Cells Impairs TGF-β-Dependent Neuroectodermal Commitment. Stem Cell Reports 13, 163–176 (2019).
Jo, H. Y. et al. Functional in vivo and in vitro effects of 20q11.21 genetic aberrations on hPSC differentiation. Sci Rep 10, 1–15 (2020).
Vitillo, L. et al. The isochromosome 20q abnormality of pluripotent cells interrupts germ layer differentiation. Stem Cell Reports (2023) doi:10.1016/J.STEMCR.2023.01.007.
Lund, R. J., Närvä, E. & Lahesmaa, R. Genetic and epigenetic stability of human pluripotent stem cells. Nat Rev Genet 13, 732–44 (2012).
Werbowetski-Ogilvie, T. E. et al. Characterization of human embryonic stem cells with features of neoplastic progression. Nat Biotechnol 27, 91–7 (2009).
Yamanaka, S. Pluripotent Stem Cell-Based Cell Therapy—Promise and Challenges. Cell Stem Cell 27, 523–531 (2020).
Keller, A. & Spits, C. The impact of acquired genetic abnormalities on the clinical translation of human pluripotent stem cells. Cells 10, (2021).
Xiao, R. et al. Aneuploid embryonic stem cells drive teratoma metastasis. Nat Commun 15, (2024).
Mandai, M. et al. Autologous Induced Stem-Cell–Derived Retinal Cells for Macular Degeneration. New England Journal of Medicine 376, 1038–1046 (2017).
Keller, A. et al. Uncovering low-level mosaicism in human embryonic stem cells using high throughput single cell shallow sequencing. Sci Rep 9, 14844 (2019).
Jacobs, K. et al. Low-grade chromosomal mosaicism in human somatic and embryonic stem cell populations. Nat Commun 5, 4227 (2014).
Jacobs, K. et al. Higher-Density Culture in Human Embryonic Stem Cells Results in DNA Damage and Genome Instability. Stem Cell Reports 6, 330–41 (2016).
Dekel-Naftali, M. et al. Screening of human pluripotent stem cells using CGH and FISH reveals low-grade mosaic aneuploidy and a recurrent amplification of chromosome 1q. Eur J Hum Genet 20, 1248–55 (2012).
Plaza Reyes, A. et al. Xeno-Free and Defined Human Embryonic Stem Cell-Derived Retinal Pigment Epithelial Cells Functionally Integrate in a Large-Eyed Preclinical Model. Stem Cell Reports 6, 9–17 (2015).
Weber, K. et al. RGB marking facilitates multicolor clonal cell tracking. Nat Med 17, 504–509 (2011).
Senabouth, A. et al. Transcriptomic and proteomic retinal pigment epithelium signatures of age-related macular degeneration. Nat Commun 13, 4233 (2022).
Petrus-Reurer, S. et al. Molecular profiling of stem cell-derived retinal pigment epithelial cell differentiation established for clinical translation. Stem Cell Reports 17, 1458–1475 (2022).
Zhang, Z. et al. A gene expression profile of the developing human retinal pigment epithelium. Mol Vis 18, 2961–75 (2012).
Kilpinen, H. et al. Common genetic variation drives molecular heterogeneity in human iPSCs. Nature 546, 370–375 (2017).
Abeyta, M. J. et al. Unique gene expression signatures of independently-derived human embryonic stem cell lines. Hum Mol Genet 13, 601–608 (2004).
Gautam, P. et al. Multi-species single-cell transcriptomic analysis of ocular compartment regulons. Nat Commun 12, 5675 (2021).
Chatzimeletiou, K., Morrison, E. E., Prapas, N., Prapas, Y. & Handyside, A. H. Spindle abnormalities in normally developing and arrested human preimplantation embryos in vitro identified by confocal laser scanning microscopy. Hum Reprod 20, 672–82 (2005).
Chatzimeletiou, K. et al. Cytoskeletal analysis of human blastocysts by confocal laser scanning microscopy following vitrification. Hum Reprod 27, 106–13 (2012).
Mertzanidou, A., Spits, C., Nguyen, H. T., Van de Velde, H. & Sermon, K. Evolution of aneuploidy up to Day 4 of human preimplantation development. Human reproduction 28, 1716–24 (2013).
Marneros, A. G. et al. Collagen XVIII/endostatin is essential for vision and retinal pigment epithelial function. EMBO Journal 23, 89–99 (2004).
Qiu, Y. et al. Downregulating p22phox ameliorates inflammatory response in Angiotensin II-induced oxidative stress by regulating MAPK and NF-κB pathways in ARPE-19 cells. Sci Rep 5, 14362 (2015).
McLaughlin, T. et al. Cellular stress signaling and the unfolded protein response in retinal degeneration: mechanisms and therapeutic implications. Mol Neurodegener 17, 25 (2022).
Kanda, A., Noda, K. & Ishida, S. ATP6AP2/(Pro)renin Receptor Contributes to Glucose Metabolism via Stabilizing the Pyruvate Dehydrogenase E1 β Subunit. Journal of Biological Chemistry 290, 9690–9700 (2015).
Cai, X., Conley, S. M. & Naash, M. I. RPE65: Role in the Visual Cycle, Human Retinal Disease, and Gene Therapy. Ophthalmic Genet 30, 57–62 (2009).
Zhou, J., Chen, F., Yan, A. & Xia, X. Overexpression of HTRA1 increases the proliferation and migration of retinal pigment epithelium. Advances in Clinical and Experimental Medicine 30, 859–864 (2021).
Santaguida, S. & Amon, A. Short- and long-term effects of chromosome mis-segregation and aneuploidy. Nat Rev Mol Cell Biol 16, 473–485 (2015).
Stingele, S. et al. Global analysis of genome, transcriptome and proteome reveals the response to aneuploidy in human cells. Mol Syst Biol 8, 608 (2012).
Donnelly, N. & Storchová, Z. Aneuploidy and proteotoxic stress in cancer. Mol Cell Oncol 2, e976491 (2015).
Krivega, M., Stiefel, C. M. & Storchova, Z. Consequences of chromosome gain: A new view on trisomy syndromes. Am J Hum Genet 109, 2126–2140 (2022).
Sripathi, S. R. et al. Transcriptome Landscape of Epithelial to Mesenchymal Transition of Human Stem Cell–Derived RPE. Investigative Opthalmology & Visual Science 62, 1 (2021).
Yu, L. et al. MFGE8 regulates TGF-β-induced epithelial mesenchymal transition in endometrial epithelial cells in vitro. Reproduction 225–233 (2016) doi:10.1530/REP-15-0585.
Jinushi, M. et al. Milk Fat Globule EGF-8 Promotes Melanoma Progression through Coordinated Akt and Twist Signaling in the Tumor Microenvironment. Cancer Res 68, 8889–8898 (2008).
Sorkio, A. et al. Structure and Barrier Properties of Human Embryonic Stem Cell–Derived Retinal Pigment Epithelial Cells Are Affected by Extracellular Matrix Protein Coating. Tissue Eng Part A 20, 622–634 (2014).
Wang, X. et al. Collagen vitrigels with low‐fibril density enhance human embryonic stem cell‐derived retinal pigment epithelial cell maturation. J Tissue Eng Regen Med 12, 821–829 (2018).
Harkin, L. F. et al. Neurexins 1-3 Each Have a Distinct Pattern of Expression in the Early Developing Human Cerebral Cortex. Cereb Cortex 27, 216–232 (2017).
Baker, D. et al. Detecting Genetic Mosaicism in Cultures of Human Pluripotent Stem Cells. Stem Cell Reports 7, 998–1012 (2016).
Li, R. & Zhu, J. Effects of aneuploidy on cell behaviour and function. Nat Rev Mol Cell Biol 23, 250–265 (2022).
Gong, J., Sagiv, O., Cai, H., Tsang, S. H. & Del Priore, L. V. Effects of extracellular matrix and neighboring cells on induction of human embryonic stem cells into retinal or retinal pigment epithelial progenitors. Exp Eye Res 86, 957–965 (2008).
Ma, W. et al. Cell-extracellular matrix interactions regulate neural differentiation of human embryonic stem cells. BMC Dev Biol 8, 1–13 (2008).
Shields, J. A. & Shields, C. L. Tumors and Related Lesions of the Pigmented Epithelium. Asia-Pacific Journal of Ophthalmology 6, 215–223 (2017).
Mateizel, I., Spits, C., De Rycke, M., Liebaers, I. & Sermon, K. Derivation, culture, and characterization of VUB hESC lines. In Vitro Cell Dev Biol Anim 46, 300–8 (2010).
Mateizel, I. et al. Derivation of human embryonic stem cell lines from embryos obtained after IVF and after PGD for monogenic disorders. Hum Reprod 21, 503–11 (2006).
Bayindir, B. et al. Noninvasive prenatal testing using a novel analysis pipeline to screen for all autosomal fetal aneuploidies improves pregnancy management. Eur J Hum Genet 23, 1286–93 (2015).

There is NO Competing Interest.

SupplementaryVideo1.mp4
Supplementary Video 1
Supplementarydata.docx

Download PDF

Version 1

posted

You are reading this latest preprint version

RPE Differentiation is a Selective Barrier Against Aneuploid hESC

Status:

Version 1

Abstract

Figures

INTRODUCTION

RESULTS

RPE^1q show transcriptomic signatures of aneuploidy-related stress

DISCUSSION

MATERIALS AND METHODS

Resource availability

Materials availability

Data availability

Ethics statement

hESCs lines, cell culture, genomic characterization and banking

RPE differentiation

Total RNA isolation, cDNA synthesis and quantitative real-time PCR (qRT-PCR) for gene expression analysis

Staining

Live imaging

Single-cell DNA sequencing

Single-cell RNA Sequencing

InferCNV

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

RPE Differentiation is a Selective Barrier Against Aneuploid hESC

Status:

Version 1

Abstract

Figures

INTRODUCTION

RESULTS

RPE1q show transcriptomic signatures of aneuploidy-related stress

DISCUSSION

MATERIALS AND METHODS

Resource availability

Materials availability

Data availability

Ethics statement

hESCs lines, cell culture, genomic characterization and banking

RPE differentiation

Total RNA isolation, cDNA synthesis and quantitative real-time PCR (qRT-PCR) for gene expression analysis

Staining

Live imaging

Single-cell DNA sequencing

Single-cell RNA Sequencing

InferCNV

Declarations

References

Additional Declarations

Supplementary Files

Status:

Version 1

RPE^1q show transcriptomic signatures of aneuploidy-related stress