Gain of 1q confers an MDM4-driven growth advantage to undifferentiated and differentiating hESC while altering their differentiation capacity

hESC^1q mis-specify to alternative cell fates during neuroectoderm differentiation and differentiate to more immature hepatoblasts and cardiac progenitors than hESC^wt

Figure 1A shows the overall setup of this study. All experiments were carried out on our in-house hESC lines VUB03 and VUB19 [29, 30]. In vitro culture led to three independent events of gain in chromosome 1 (Fig. 1B, C). VUB19 and one subline of VUB03 acquired a gain of the entire q arm of chromosome 1 (termed VUB19^1q21.1qter and VUB03^1q21.1qter), another subline of VUB03 gained a smaller region spanning 3.3 Mb in 1q32.1 (VUB03^1q32.1).

A Lentiviral transduction was used to induce stable expression of fluorescent proteins in hESC^1q. Working cell banks were established comprising mutant-labeled cell lines and their isogenic counterparts. Pure wild-type and 1q lines were utilized, along with mixes containing 10% of 1q cells, for differentiation to three lineages: neuroectoderm, hepatoblasts and cardiac progenitors. Differentiation capacity was assessed through immunostaining, qPCR, and RNA sequencing. Cell competition results were measured by analyzing the ratio of mutant and wild-type cells before and after differentiation using flow cytometry and immunostaining. To study the role of MDM4 in the growth advantage of hESC^1q, MDM4 was downregulated by siRNA prior to the competition assays. To study the mechanisms by which MDM4 provides the selective advantage, DNA damage was induced by Bleomycin, and apoptosis levels were evaluated using Annexin V staining and flow cytometry, while DNA damage levels were studied through immunostaining for gamma-H2AX. B Breakpoints for the minimal region of gain in chromosome 1 are 202,475,000 and 205,800,000. C Shallow DNA sequencing–based karyotyping of the hESC lines included in the study. Gain in 1q chromosome is visible for VUB03^1q32.1, VUB03^1q21.1qter and VUB19^1q21.1qter. D Representative images of immunostainings for POU5F1 (turquoise, 1st panel), PAX6 (magenta, 2nd panel), HNF4A (magenta, 3rd panel) and GATA4 (magenta 4th panel) of hESC^wt and hESC^1q before and after the completion of induction of differentiation to neuroectoderm (NE, day 8), hepatoblast (HEP, day 8), cardiac progenitors (CP, day 5) of VUB19^1q21.1qter. Figure S1 shows the images for cell lines VUB03^1q21.1qter and VUB03^1q32.1 included in the study. E Quantification of the percentage of PAX6, HNF4A and GATA4-positive cells in the immunostainings shown in panel D and in Fig. S1. F mRNA quantification by quantitative real-time PCR of PAX6 and SOX1 in NE^wt and NE^1q, HNF4A and GATA4 in HEP^wt and HEP^1q, and GATA4 and NKX2.5 in CP^wt and CP^1q. Results for POU5F1 and NANOG are shown in Fig. S1. Each differentiation was carried out in three technical replicates at the same time. Different hESC lines and independent experimental replicates (rep) are coded in different colors. *p < 0.05, **p < 0.01, ****p < 0.0001. ns = non-significant.

Full size image

We investigated the impact of the gain of 1q on trilineage differentiation by subjecting hESC^wt and hESC^1q to neuroectoderm (NE), hepatoblast (HEP) and cardiac progenitor (CP) differentiation (VUB03^1q21.1qter, VUB19^1q21.1qter, VUB03^1q32.1, VUB03^wt and VUB19^wt, all lines differentiated at least in triplicate). We measured the expression of six lineage-specific markers and of NANOG and POUF51 to evaluate differentiation efficiency and to test for residual undifferentiated cells (Figs. 1D–F, S1). All differentiated cells had almost undetectable levels of POU5F1 and NANOG mRNA (Fig. S1), and no POU5F1-positive cells appeared in the immunostainings (Fig. 1D), indicating that all cells exited the undifferentiated state (Figs. 1D, S1).

Immunostaining showed a lower percentage of PAX6-positive cells in differentiated hESC^1q than in their isogenic hESC^wt counterparts (Figs. 1D, 1E, NE^1q = 43.78%, NE^wt = 64.19%, p < 0.0001, unpaired t-test, N^wt = 26, N^1q = 36). While the number of SOX1-positive cells was similar between the two groups, we observed a significantly reduced intensity of SOX1 protein expression in hESC^1q (Fig. S2A, B, NE^1q = 6343, NE^wt = 9107, p < 0.0001, unpaired t-test, N^wt = 13, N^1q = 18). In line with this, NE^1q showed significantly lower mRNA levels of PAX6 and SOX1 compared to the levels in NE^wt (Fig. 1F p^PAX6 = 0.0017 and p^SOX1 = 0.0002, unpaired t-test, N = 15), indicating a decreased neuroectodermal differentiation efficiency. For the HEP differentiation, we found no significant difference between HEP^wt and HEP^1q in the mRNA expression levels of HNF4A while AFP had a significantly lower expression in HEP^1q (Fig. 1F, p^HNF4A = 0.2247, p^AFP = 0.0142, unpaired t-test, N = 9) and immunostaining for HNF4A showed similar percentages of HNF4A positive cells in both groups (Fig. 1E, HEP^1q = 26.27%, HEP^wt = 25.40%, p = 0.8585, unpaired t-test, N^wt = 25, N^1q = 32). Similarly, the mRNA levels of the CP marker GATA4 was not significantly different between CP^wt and CP^1q while NKX2-5 was significantly lower in CP^1q (Fig. 1F, p^GATA4 = 0.1537, p^NKX2.5 = 0.0398, unpaired t-test, N = 18), which was confirmed by the immunostaining for GATA4 (Fig. 1E, CP^1q = 33.35%, CP^wt = 24.74%, p = 0.1627, unpaired t-test, N^wt = 23, N^1q = 28).

Taken together, hESC^1q show a decreased differentiation efficiency into neuroectoderm, but do not remain undifferentiated, suggesting that part of the cells mis-specify to alternative cell fates. hESC^1q commit to mesendoderm as efficiently as their isogenic counterpart, showing similar early endoderm and mesoderm marker expression but reduced expression of markers for further cell-type commitment.

To further characterize the differentiated cells, we carried out bulk mRNA sequencing of 43 samples: 15 samples of neuroectoderm (NE^wt = 6, NE^1q = 9), 14 of cardiac progenitors (CP^wt = 6, CP^1q = 8) and 14 of hepatoblasts (HEP^wt = 6, HEP^1q = 8). Differential gene expression analysis showed that NE^1q differentially expressed 1603 genes as compared to NE^wt, while HEP^1q and CP^1q differentially expressed 189 and 241 genes, respectively, as compared to HEP^wt and CP^wt (Fig. S3). We surmised that this significant difference in the number of differentially expressed genes was caused by mis-specification of hESC^1q to neuroectoderm and were rather yielding a mixed cell population, which did not occur in the mesendoderm lineages.

To determine the alternate cell fate acquired by hESC^1q upon neuroectoderm differentiation, we carried out differential gene expression analysis of the NE^1q and the NE^wt relative to bulk RNA sequencing data of undifferentiated hESC (N = 38 samples, previously published data [26]). First, we tested the expression of neuroectoderm markers, which we found to be less induced in NE^1q than in NE^wt (Figs. 2A, S4A). We then queried the data for the expression of different sets of markers for embryonic and extra-embryonic lineages that appear in early human development. Table S1 shows the lists of gene sets we curated from published data [31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54]. We found that VUB19^1q21.1 showed high expression of markers of non-central nervous system (non-CNS) ectodermal lineages such as non-neural ectoderm and of cranial placode (Figs. 2B, S4B, S5). Staining for the non-CNS marker TFAP2A and placode marker SIX1 of independent differentiation experiments of VUB19^1q21.1 confirmed that the alternate cell fate of this line was indeed consistently a mix of neuroectoderm and non-CNS cells (Fig. 2C). The other two lines showed variable expression of genes of different lineages, including the neuronal, but we could not identify a gene expression pattern consistent with a specific cell type (Figs. 2B, S4B, S5).

A Lollipop diagrams representing the differential gene expression of neuroectoderm markers in NE^wt and NE^1q vs undifferentiated cells. All genes shown in this panel and in (D, E, G and H) are differentially expressed with a log₂fold-change over 0 and FDR < 0.05. B Heatmap representing the induction of expression of genes marking different ectodermal cell types and the undifferentiated hESC state in NE obtained from the five lines included in this study. C Representative images of the immunostaining of NE^wt and NE^1q for DNA (blue), PAX6 (white), TFAP2A (red, left panel) and SIX1 (red, right panel). D Lollipop diagrams representing the differential gene expression of hepatoblast markers in HEP^wt and HEP^1q vs undifferentiated cells. E Lollipop diagrams representing the differential gene expression of markers of different stages of hepatocyte differentiation in HEP^wt and HEP^1q vs undifferentiated cells. F Enrichment map showing the Top-15 deregulated pathways from gene ontology gene set enrichment analysis in HEP^wt vs HEP^1q. The size of nodes indicates the number of genes in each pathway and the color represents adjusted p-value. Pathways that cluster together have overlapping gene sets. G Lollipop diagrams representing the differential gene expression of genes marking different cardiac cell types in CP^wt and CP^1q vs undifferentiated cells. H Lollipop diagrams representing the differential gene expression of genes marking different stages of cardiomyocyte differentiation in CP^wt and CP^1q vs undifferentiated cells. I Enrichment map showing the top-15 deregulated pathways from gene ontology gene set enrichment analysis in CP^wt vs CP^1q. J Venn diagram showing the overlaps in differentially expressed genes between control and 1q-cells, for the three cell types with log₂fold-change over 0 and FDR < 0.05. MDM4 is part of the core of 53 commonly deregulated genes. K Reactome pathway gene set enrichment analysis of the C2 library canonical pathways related to p53 signaling in NE, HEP and CP from 1q and wildtype cells. L Graphical summary of the model in which MDM4 has a higher expression because of the gain of 1q, resulting in an inhibition of the p53 signaling pathway and in turn decreasing the sensitivity of the cells to DNA damage.

Full size image

Next, we used the same approach to study the cell types obtained from the HEP and CP samples. While hESC^1q equally induce some of the early HEP and CP differentiation markers as hESC^wt, 68.6% (24/35) and 63.2% (12/19) of HEP and CP lineage-specific markers have a lower expression in the mutant cells (Fig. 2D, G). Because the HNF4 and GATA4 staining (Figs. 1D, S1) and the RNA sequencing data (Fig. S6) did not suggest mis-specification for these lineages, we analyzed the gene expression data with the aim of establishing the degree of maturity of the HEP and CP cells.

Human ESC-derived HEP have a profile between hepatoblast stage 2 and fetal hepatocyte 1 (Fig. 2E), and cells with a gain of 1q show overall lower expression of 64% of markers (16/25). Gene-set enrichment analysis of HEP^wt versus HEP^1q showed that 145 of the 162 significantly enriched gene sets in the canonical pathways gene set had negative enrichment scores. They were frequently related to key processes of hepatocyte and liver function, including cholesterol metabolism, ferroptosis, transsulfuration, plasma lipoprotein remodeling, folate metabolism, selenium micronutrient network and metabolism of steroids (FDR < 0.05, Table S2). Gene ontology enrichment analysis showed that 1574 of the 1913 gene sets had negative enrichment scores, including amino acid metabolic process, fatty acid metabolic process and steroid metabolic process (Fig. 2F, Table S2).

In the case of CP, CP^1q and CP^wt equally express a profile between cardiac progenitors and cardiomyocytes, where CP^1q have lower expression of genes marking the later stages of differentiation (Fig. 2H). Gene set enrichment analysis of the CP samples showed that CP^1q have negative enrichment scores for 157 of the 178 significantly enriched canonical pathway gene sets, including sets related to dilated cardiomyopathy, folding of actin, striated muscle contraction and hypertrophic cardiomyopathy (Table S2). These genes are key to correct heart contraction functions. Gene ontology enrichment analysis identified negative enrichment scores for 1146 of the 1786 significantly enriched set, including terms such as l band, Z disc, sarcomere, myofibril, contractile fiber, muscle system process and muscle development (Fig. 2I, Table S2).

Human ESC^1q have an MDM4-driven competitive advantage that is retained during differentiation

Human ESC^1q have a well-established competitive advantage over their genetically balanced counterparts in the undifferentiated state [11, 13, 16]. We next aimed at determining whether the cells retain this selective advantage during differentiation, and which gene is driving their winning phenotype.

We first looked at the 53 commonly deregulated genes in our HEP, CP and NE samples (Fig. 2J), and found that MDM4, a regulator of p53 activity that is located in the common region of gain and which has been previously suggested as a key gene for the gain of 1q [13], is consistently upregulated in all samples (Figs. 2J, S2C). Gene set enrichment analysis of the differentially expressed genes in NE, HEP and CP from 1q and wildtype cells for the Reactome pathways related to p53 signaling indeed shows that the transcriptional regulation by p53, the regulation of p53 activity, the G1/S damage checkpoint and the p53 dependent responses to DNA damage in G1 and S are all significantly negatively enriched (Figs. 2K, S2D, Tables S2, S3). This led us to the hypothesis that the higher expression of MDM4 in cells with a gain of 1q results in the inhibition of p53-mediated transcriptional activity, leading to a decreased induction of apoptosis, thus providing a competitive advantage to the mutant cells (Fig. 2L).

To determine if cells with a gain of 1q retained their ability to take over the culture during differentiation, we carried out competition assays during NE, HEP and CP induction. For this, 10% of hESC^1q stably expressing a fluorescent protein were introduced into an unlabeled hESC^wt culture. Differentiation was initiated the next day and was controlled by immunostaining for PAX6, HNF4A and GATA4 (Fig. 3A). To measure culture takeover, the proportion of hESC^1q was determined by flow cytometry at the onset and at the end of differentiation.

A Representative images of immunostaining for PAX6 (magenta left panel), HNF4A (magenta, middle panel) and GATA4 (magenta, right panel) after induction of cell competition during the differentiation to neuroectoderm, hepatoblast and cardiac progenitors. Throughout all panels, 1q cells are shown in yellow, POU5F1 cells in turquoise and DNA in white. B Quantification of 1q-cells at the start and end of differentiation. C Western blot and intensity quantification of protein bands of MDM4 and p53 in hESC^wt, hESC^1q and siMDM4-treated hESC^1q. D Time-course of mRNA expression of MDM4 in siMDM4-treated and untreated hESC^1q. E Images of the competition assays in the undifferentiated state. hESC^1q (green) are mixed in at a 1:9 ratio with hESC^wt and traced by their fluorophore expression. In the siMDM4 condition, the hESC^1q have been treated with siRNA against MDM4 prior to mixing. F Images of the competition assays of hESC^20q11.21 (red) in the undifferentiated state. The experimental setup is as in panel (E). G Quantification of hESC^1q at the start of the cell competition (day 0) and at day 3 of in vitro culture. H Quantification of the numbers of hESC^20q11.21 at the start of the cell competition (day 0) and at day 3 of in vitro culture. I Quantification of 1q-cells at the start of the cell competition during neuroectoderm, hepatoblast and cardiac progenitor differentiation (day 0) and at day 3 or 4, with and without siRNA against MDM4. J Representative images of immunostaining for PAX6 (magenta left panel), SOX17 (magenta, middle panel) and GATA4 (magenta, right panel) after induction of cell competition during the differentiation to neuroectoderm, hepatoblast and cardiac progenitors, with and without treatment of the hESC^1q with siRNA against MDM4. Throughout all panels, 1q cells are shown in yellow, POU5F1 cells in turquoise and DNA in white. K Quantification of 1q-cells in embryoid bodies after 4 days of spontaneous differentiation and directed neuroectoderm differentiation. Four EBs were pooled for dissociation and counting of fluorescently labeled 1q cells. L Example of embryoid bodies for cell line VUB19 on day 1 and 4 of spontaneous and directed neuroectoderm differentiation. Cells with 1q gain are labeled with green fluorescence. *p < 0.05, **p < 0.01, ****p < 0.0001, ns = non-significant.

Full size image

We found that in all three differentiations, the cells with a gain of 1q outcompeted wildtype cells (Fig. 3B). During the NE induction, the proportion of 1q cells increased in average 33.9% ± 2.7% during the 8-day differentiation, from a mean of 10.9% at the onset to 44.9% at day 8 (p < 0.0001, unpaired t-test, N = 18–21). This was similar for the 8-day differentiation to HEP, where the 1q cells increased in average 49.3% ± 4.9% (mean at day 0 = 11.6%, mean at day 8 = 60.9%, p < 0.0001, unpaired t-test, N = 9). This increase was less pronounced during the 5-day CP differentiation, with an average 22.7% ± 3.2% increase (mean at day 0 = 10.9%, mean at day 5 = 33.5%, p < 0.0001, unpaired t-test, N = 18–21), which may be attributable to the shorter time span of the CP differentiation as compared to NE and HEP.

Next, we tested the role of MDM4 in the competitive advantage of cells with a 1q gain. We first quantified MDM4 and p53 protein levels in hESC^1q as compared to their isogenic counterparts and tested the effect of downregulating MDM4 by siRNA for 24 h (Fig. 3C, uncropped membranes shown in Fig. S7). Quantification of the western blot bands revealed a 1.5-fold increase in MDM4 in VUB03^1q32.1 and a 2.6-fold increase in VUB03^1q21.1qter compared to their isogenic controls. After siMDM4 treatment, MDM4 protein decreased by 0.3-fold in both cell lines compared to hESC^1q. Conversely, p53 levels were 0.2-fold and 0.1-fold lower in VUB03^1q32.1 and VUB03^1q21.1qter, respectively, compared to hESC^wt cells. After siMDM4 treatment, p53 levels increased 10.2-fold and 26.5-fold compared to untreated VUB03^1q32.1 and VUB03^1q21.1qter respectively.

We further controlled if the siRNA-mediated downregulation of MDM4 in hESC^1q restored its mRNA levels to those of hESC^wt and whether this was stable over the course of 4 days post transfection (Fig. 3D). MDM4 mRNA was 1.4-fold higher in hESC^1q than in hESC^wt (p = 0.0454, N^wt = 4, N^1q = 5). After 24 h of siRNA treatment, MDM4 mRNA was reduced to 0.5-fold of the hESC^wt levels, and by day 4 the expression was 0.9-fold of that of wildtype cells.

We then carried out competition assays in undifferentiated and differentiating cells. For the undifferentiated cells, we downregulated MDM4 for 24 h by siRNA in hESC^1q on the day prior to the start of the competition assay. The cells were imaged daily and counted at the start and end points (Fig. 3E). On average, hESC^1q increased from 14.0 to 33.7% in three days in the untreated condition, while no increase was observed after siRNA treatment (N = 3, p^1q = 0.0028, p^si1q = 0.7905, 2-way ANOVA, Fig. 3G). We carried out the same experiment on a hESC line carrying the recurrent gain of 20q11.21, which provides the cells with a Bcl-xL-mediated decreased sensitivity to apoptosis [18, 19]. Downregulation of MDM4 did not affect the growth advantage of the hESC^20q11.21 cells (N = 7, p^20q < 0.0001, p^si20q = 0.0033, 2-way ANOVA, Fig. 3F, H), showing that the suppression of the competitive advantage by modulating the MDM4 expression is specific to cells with a gain of 1q and not due to a general decrease in cellular fitness.

For the competition assays during differentiation, the hESC^1q were treated for 24 h with the siRNA and mixed at a 1:9 ratio, and differentiation was initiated the next day. Differentiation was shortened to 4 days, as this was the time we found that a single siRNA transfection could reliably sustain a gene’s downregulation (Fig. 3D). In the untreated conditions of the competition assays, the fraction of cells with a gain of 1q became significantly larger in all three differentiations (Fig. 3I, J). The increase was most pronounced after NE induction, with an average increase of 30.0% (p^1q = 0.0053, 2-way ANOVA). In HEP and CP, the mean increases were 11.3% (p^1q = 0.0131, 2-way ANOVA) and 11.67% (p^1q = 0.0149, 2-way ANOVA), respectively. Conversely, in the siRNA-treated competition assays, the fraction of 1q cells remained unchanged (Fig. 3I, J). Taken together, these results show that reducing the levels of MDM4 in the mutant cells abolishes their competitive advantage both in the undifferentiated state and during differentiation.

We next studied the competitive ability of 1q cells in a 3D differentiation using embryoid bodies (EBs, Figs. 3L, S8A, B). The EBs underwent both spontaneous and directed differentiation into neuroectoderm over a period of 4 days. For competition assays, mixtures were prepared with 10% fluorescently labeled hESC^1q cells or hESC^1q cells treated with siMDM4, as described previously. EBs were imaged daily, and the proportion of fluorescent cells was quantified at the end of the 4-day differentiation period (Fig. 3K). All EBs were characterized by RT-qPCR analysis of early differentiation markers (Fig. S8E–G). In the spontaneous differentiation conditions, at day four, EBs contained on average 52.83% 1q cells, while EBs with siMDM4-treated cells had 19.33% 1q cells (p < 0.0001, N = 6). The same experimental setup was applied for directed differentiation into neuroectoderm, where on day 4, EBs contained 56.83% 1q cells, whereas EBs with MDM4 downregulation exhibited 10.67% 1q cells (p = 0.0006, N = 6). Lastly, we carried out spontaneous differentiation by only withdrawing TGFβ and FGF2 from the cells, in monolayer cultures, which yielded similar results (Fig. S8C, D).

Higher MDM4 expression in hESC^1q results in a decreased sensitivity to DNA damage-induced apoptosis

Next, we sought to elucidate by which mechanisms higher expression of MDM4 confers the competitive advantage to the cells. We first tested the hypothesis that the higher expression of MDM4 by cells carrying a gain of 1q leads to a decreased p53-mediated apoptosis in response to DNA damage. For this, we induced DNA damage in hESC^1q and hESC^wt using Bleomycin and carried out a time-course measurement of apoptosis and cell death.

Figure 4A shows the percentages of live and apoptotic and dead cells for hESC^wt (N = 3), hESC^1q (N = 3), hESC^1q treated with siRNA against MDM4 (N = 3) and hESC^1q treated with Nutlin-3a, an inhibitor of MDM4 (N = 3), at the start of Bleomycin treatment and at the subsequent 2, 4 and 6-h time-points. hESC^wt start undergoing apoptosis 2 h after exposure to Bleomycin, followed by a rapid decrease in the numbers of live cells. In contrast, apoptotic cells start appearing in hESC^1q as from 4 h of exposure and reach 41.9% of apoptotic cells at 6 h, as compared to 78.6% in hESC^wt (unpaired t-test, p = 0.0169). Treating hESC^1q with siMDM4 significantly increases their sensitivity to DNA damage, with apoptosis initiating at 2 h, and reaching 49.4% at 4 h. siMDM4-treated hESC^1q cells do not reach same levels of dead cells at 6 h as in hESC^wt cells, although the differences are not statistically significant (54.9% vs 78.6%, unpaired t-test, p = 0.1530, Fig. 4A). This may be explained by an incomplete transfection of the cells, or to overall insufficiently stable downregulation of MDM4 to the levels of hESC^wt. Nutlin-3a treated hESC^1q show similar sensitivity to apoptosis as cells with downregulated MDM4. The apoptosis initiated at 2 h of Bleomycin treatment and reached 42.3% of apoptosis at 4 h.

A Percentages of apoptotic and live cells in hESC^wt, hESC^1q and treated hESC^1q with siMDM4 and Nutlin-3a upon 0 h, 2 h, 4 h and 6 h of exposure to Bleomycin. B Evolution of the numbers of cells from day 0 to day 8, for hESC^1q, hESC^wt and a 1:9 mix of hESC^1q and hESC^wt. C Percentage of hESC^1q over the course of 8 days in culture, starting from a 1:9 mix of hESC^1q and hESC^wt. The proportion of hESC^1q becomes significantly increased as from day 4. *p < 0.05. D Representative images illustrating the cell density at days 3 and 4 of the time-course experiment. Confluency is reached on day 4 of this experiment. 1q cells are shown in red and DNA in blue. E Examples of the staining for gammaH2AX (green) in hESC^1q and hESC^wt, upon exposure to Bleomycin and in low- and high-density cultures. DNA is shown in blue. F Cell densities for the low- and high-density culture conditions shown in (E and G). The dots represent the counts made in different images of dishes of different cell lines. The line represents the mean of the dots. G Counts of gammaH2AX foci per nucleus in hESC^1q and hESC^wt at low and high-density culture conditions. Each dot represents a single nucleus, hESC^wt-low N = 976, hESC^wt-high N = 2879, hESC^1q-low N = 2995, hESC^1q-high N = 7158. H Cell densities of hESC^wt, hESC^1q and mixes calculated from cell counts during a 5-day culture period.****p < 0.0001, one-way ANOVA and Šidák correction. ns = non-significant.

Full size image

This raised the question of how a decrease in sensitivity to DNA damage could provide a competitive advantage to hESC and differentiating cells. We observed during daily monitoring of the competition assays that hESC^1q started outcompeting their genetically balanced counterparts once the cultures became confluent. Previous work has indicated that hPSC are prone to replication stress and DNA damage [55, 56], which can be mitigated by addition of nucleosides to the medium [57] and exacerbated in higher cell density culture due to medium acidification [58, 59]. We therefore hypothesized that higher culture cell density generates the conditions for a strong competitive advantage of 1q gains, by increasing the levels of DNA damage.

We studied the cell proliferation dynamics of hESC^wt and hESC^1q over time in culture in pure form and mixed at a 1:9 ratio (N = 3 for each condition). Daily cell numbers count showed that the numbers of hESC^wt and hESC^1q increased similarly until day 7, suggesting that they have similar cell doubling times. As from day 8, hESC^1q continue proliferating even when they have reached a very high density (Fig. 4B). Daily analysis of the ratio between hESC^wt and hESC^1q showed no statistically significant changes until day 3, after which the proportion of hESC^1q started steadily increasing (Fig. 4C, N = 3). This coincided with culture dishes at day 3 still showing empty areas, whereas at day 4 the cells had reached confluence, suggesting that this is a flipping point for 1q gains to start providing a competitive advantage (Fig. 4D). To investigate the relationship between cell culture density and DNA damage, we measured DNA damage by γH2AX staining in hESC^1q and hESC^wt, in low and high cell density cultures (mean of 7547 cells/cm² and 200,050 cm/cm² respectively, representative images are shown in Fig. 4E and cell counts shown in Fig. 4F), and found that cells grown at low density had significantly lower numbers of γH2AX foci than when grown in high density (Fig. 4G, wt_low: 11.38, wt_high: 17.95, 1q_low: 13.71, 1q_high: 26.22, p < 0.0001, one-way ANOVA). While there were no differences in DNA damage between hESC^1q and hESC^wt grown at low density (Fig. 4G, 1.2-fold change, p = 0.2314, one-way ANOVA), at high density, hESC^1q showed higher γH2AX foci counts than hESC^wt (Fig. 4G, 1.5-fold change, p < 0.0001, one-way ANOVA). Lastly, we counted the numbers of cells in the cell cultures until day 5 and calculated the densities (Fig. 4H). The mean densities in days 1 and 2 are in the range of the ‘low density’ group in the DNA damage staining, with a mean of 33,711 cells/cm² at day 1 and 87,854 cells/cm² at day 2. From day 3, the densities are in the range and above of the ‘high density’ group, with 265,224 cells/cm² at day 3, 331,680 cells/cm² at day 4 and 351,475 cells/cm² at day 5. This indicates that at this point, the cells start undergoing DNA damage more frequently, reaching the condition when a decreased p53-mediated induction of apoptosis starts providing a selective advantage, so cells with a gain of 1q will start outcompeting their genetically balanced counterparts.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41419-024-07236-x