Origin of Chromosome 12 Trisomy Surge in Human Induced Pluripotent Stem Cells (iPSCs)

Abstract

Trisomy 12 is the most common whole-chromosome abnormality in human pluripotent stem cells. Conventionally, this acquired aneuploidy is ascribed to a rare single-cell event followed by selective growth advantage. Instead, we show that trisomy 12 emerges simultaneously in a very high percentage of cells in critical transition passages. Mis-segregation and incorporation of chromosome 12 into micronuclei occur through bridging of the short p arms of chromosome 12. Subsequently, single, unreplicated chromosome 12 chromatids are observed in mitotic cells. Erosion of the subtelomeric regions of the 12p arms is found during the passages when chromosome 12 bridges become frequent and trisomy 12 increases. Trisomy 12 cells persist due to a slight growth advantage. Among the shortest telomeres in humans are those on the 12p arms, making them particularly vulnerable to damage and bridging during mitosis. These findings reveal a novel mechanism of whole-chromosome instability in human stem cells, with broad implications for understanding the genesis of aneuploidy across diverse biological systems.

Summary:

Narozna et al. reveal that trisomy 12 arises en masse during crucial passages of iPSCs. The short p-arms whose ends are eroded by replication stress, trigger p-arm chromosome bridges at anaphase leading to chromosome mis-segregation, formation of micronuclei, and trisomy. A slight growth advantage allows trisomic cells to persist and eventually dominate.

Graphical Abstract

1. Introduction

Aneuploidy, defined by the gain or loss of whole chromosomes and chromosome segments, is a hallmark of human tumors (Ben-David and Amon, 2019; Girish et al., 2023) and a major challenge in studying the cell biology of normal diploid cells in culture (Duesberg et al., 2004). Primary cell cultures have limited proliferation potential due to replicative senescence caused by telomere erosion (Deng et al., 2008; Beauséjour et al., 2003). To overcome this, primary cells are immortalized through stable transfection with telomerase (Bodnar et al., 1998) or with oncogenes (de Bardet et al., 2023), but the resulting somatic cell lines invariably contain significant chromosome abnormalities (Ippolito et al., 2021; Cohen-Sharir et al., 2021; Zerbib et al., 2024; Voloshin et al., 2023; Ouellette et al., 2000; Zongaro et al., 2005). Consequently, previous studies have often relied on transformed or tumor-derived cell lines (Taylor et al., 2018; Cohen-Sharir et al., 2021), making it challenging to find diploid and non-transformed human cell lines that acquire aneuploidy through natural chromosome mis-segregation to study specific aneuploidies in its native context (Zerbib et al., 2024; Ben-David et al., 2014).

Human pluripotent stem cells, because they innately express telomerase (Boyle et al., 2020; Rivera et al., 2016; Hiyama and Hiyama, 2007), are perhaps the only true diploid cells that can be propagated indefinitely in an undifferentiated, pluripotent state in culture, offering a unique opportunity to study the rise of aneuploidy from diploid cells (Moradi et al., 2019; Odorico et al., 2001). Cultured pluripotent stem cells arise from two sources (Romito and Cobellis, 2016): embryonic stem cell lines are generated from embryos (Moon et al., 2006) and induced pluripotent stem cells (iPSCs) are reprogrammed from primary cell cultures via transient expression of specific transcription regulators (Takahashi et al., 2007). However, up to one-third of all cultured human pluripotent stem cell lines, although initially diploid (Mayshar et al., 2010), acquire spontaneous chromosome abnormalities over time (Baker et al., 2016; Halliwell et al., 2020; Ben-David et al., 2014; Lamm et al., 2016; Na et al., 2014). These chromosomal abnormalities are not random but predominantly involve duplications of whole or parts of chromosomes 1, 12, 17, 20, and X (The International Stem Cell Initiative., 2011; Baker et al., 2016; Halliwell et al., 2020; Vaz et al., 2021; Mayshar et al., 2010; Ben-David et al., 2014; Lamm et al., 2016; Na et al., 2014). Chromosome 12 abnormality most often manifests as a whole-chromosome trisomy, whereas gains involving other chromosomes tend to occur as duplications of single chromosome arms or arm segments (Andrews et al., 2022; Draper et al., 2004; Mayshar et al., 2010; Ben-David et al., 2011, 2014). Trisomy 12 is one of the most frequent and significant changes (Reid et al., 2021; Baker et al., 2016; Halliwell et al., 2020; Ben-David et al., 2014; Lamm et al., 2016). It poses serious problems for use of iPSCs in therapy and research (Ben-David et al., 2014, 2011; Yamanaka, 2020; Kim et al., 2022) given that it is also characteristic of various malignancies, including chronic lymphocytic leukemia (Abruzzo et al., 2018; Reid et al., 2021) and germ cell tumors (Rosenberg et al., 2000; Freitag et al., 2021). The conventional explanation for the appearance and dominance of chromosome aberrations in stem cell culture is random replication and chromosome segregation defects followed by selection of chromosome gains that enhance proliferation (Ben-David et al., 2014; Price et al., 2021; Mayshar et al., 2010; Draper et al., 2004; Keller and Spits, 2021; Vaz et al., 2021; Andrews et al., 2022). Here we provide evidence that selection, on its own, does not explain the origin and eventual dominance of chromosome 12 trisomy in cultured human iPSCs.

In our previous study, we leveraged a single human iPSC line that acquires trisomy 12 during passaging without other chromosomal aberrations (Dubose et al., 2022). Our initial analysis showed that the proliferative advantage of chromosome 12 trisomic cells was insufficient to explain the rapid dominance of the trisomic cells and implicated additional contributors. Our present work aimed to address key questions of how and why does chromosome 12 trisomy arise and rapidly dominate human iPSC cultures. We conclude that the takeover of chromosome 12 trisomy in human iPS cells reflects the gain of an extra copy of chromosome 12 simultaneously in a very high percentage of diploid cells per cell division. Specifically, we provide a new mechanism for the simultaneous conversion of at least 1% of diploid cells to trisomic per cell division due to a specific chromosome 12-mis-segregation error. Because trisomy 12 cells proliferate faster than their diploid precursors, they are maintained in the population. Eventually, the combination of the generation of new trisomy 12 cells from diploid parents and their faster proliferation rate drive the trisomy 12 cells to rapidly and completely dominate the population. Altogether, our findings uncover a novel source of aneuploidy in human iPSCs with significant implications for the use of stem cells in research and regenerative medicine but may also provide potential intervention strategies to reduce the incidence of spontaneous aneuploidy in stem cell culture.

2. Results and Discussion

2.1. Chromosome 12 trisomic cells appear spontaneously in iPS cell culture and rise to almost 100% over 12 – 14 passages.

In our previous study, trisomy 12 was first detected at passage 21 in 12% of the cell population and rapidly increased to 80% by passage 26 (Dubose et al., 2022). However, the chromosome spreads approach we used previously was limited to small sample sizes. To precisely map the timing of the initial occurrence and eventual dominance of trisomy 12, we used dual-color fluorescence in situ hybridization (FISH) to label centromeric regions of chromosome 12 and, as a control, chromosome 10 (Fig. 1A-C). Karyotypic alterations involving gains or losses of the entire chromosome 10 have not been reported in human iPSCs (Akutsu et al., 2022; Henry et al., 2019; Mitalipova et al., 2005; Na et al., 2014; Rajamani et al., 2014; Assou et al., 2020; Vaz et al., 2021; Yamamoto et al., 2022), with only rare instances of 10p loss observed (Baker et al., 2016). We cultured and mapped the cells using FISH over 170 days of continuous culture (57 passages). By analyzing hundreds of cells at each passage, we pinpointed the critical transition passages where chromosome 12 trisomic cells arose. Chromosome 12 trisomic cells appeared in a small percentage of cells (>5%) and gained almost 100% dominance within ~13 passages. Trisomic cells remained dominant (>95%) for at least 28 more post-transition passages (Fig. 1C, S1A).

Fig. 1. Trisomy 12 dominance in human iPS cells arises spontaneously through multicellular origin and rapidly dominates the cell culture within ~13 passages.

(A) Chromosome 12 diploid and trisomic cells after dual-color FISH performed on human iPS cells. DNA labeling is shown in grey. The overlays merge DNA (blue), chromosome 12 centromeres (magenta), and chromosome 10 centromeres (yellow). Maximum intensity projections are shown. Scale bars = 10 µm. (B) Microscope field after FISH performed on human iPS cells from passage 21 with 36% trisomy 12. Arrows indicate chromosome 12 trisomic cells. DNA labeling is shown in grey. The overlays merge DNA (blue), chromosome 12 centromeres (magenta), and chromosome 10 centromeres (yellow). Maximum intensity projections are shown. Scale bar = 10 µm. (C) The observed trisomy 12 takeover during transition passages 17 to 29. Transition passages are defined as those where trisomy 12 frequency rises from >5% to >95%. Each dot on the graph represents a microscope field of analyzed cells, with the numbers above indicating the total number of analyzed cells. At passage 12, all analyzed cells are still diploid for chromosome 12 (depicted as two magenta dots labeling two chromosome 12 centromeres) and the control chromosome 10 (two yellow dots labeling two chromosome 10 centromeres). This results in an average chromosome 12 to chromosome 10 ratio of approximately 1.0. By passage 29 until at least passage 57, nearly all analyzed cells are trisomic for chromosome 12 (possess 3 copies of chromosome 12). This results in a ratio of chromosome 12 to chromosome 10 of 1.5. Transition passages 17–29 show a rapid takeover of the chromosome 12 trisomy. Trisomy 12 persists in the cell culture for at least 28 more post-transition passages since it gained almost 100% dominance at passage 29. Passage numbering reflects the number of passages performed during this study starting from passage 2 of human iPS cell line AICS-0012. Error bars represent SEM. (D) Calculated doubling times of diploid and trisomic cell populations: 19.25 ± 1.29 h and 17.68 ± 1.41 h, respectively (n = 36 per group). Each data point represents a technical replicate, pooled from three diploid and three trisomic passages. Data points are shape-coded by biological replicate. Data presented as mean ± SD; p values were determined by unpaired, two-sided Mann-Whitney U test, **** p<0.0001. Normality was assessed using the Shapiro-Wilk test. (E) Theoretical outgrowth of trisomic cells over diploid during 13 passages (39 days in cell culture) based solely on the doubling times difference. The variable representing the percentage of diploid cells converting to trisomy 12 per cell division is set at 0%, relying only on differences in doubling times. (F) The average dominance of the chromosome 12 trisomic cell population overtime observed in three FISH replicates where trisomy 12 took over the cell culture. Despite expectations of observing 38% trisomy 12 after 13 passages (39 days in cell culture), FISH experiments consistently showed almost 100% trisomic cell population during that time. Error bars represent SEM. (G) Trisomy 12 outgrowth modeled by incorporating the doubling time difference and an additional variable representing the percentage of diploid cells converting to trisomy 12 per cell division. Applying this variable, the mathematical model showed that to reach 96% of trisomy 12 in 13 passages (39 days in cell culture), 3% of diploid cells must convert to trisomic during each cell division. (H) Early transition passage (~6% trisomy 12) showing a predominantly diploid cell colony with clustered trisomic cells, likely arising from a trisomic cell that was seeded among diploids and underwent approximately four doublings over 72 hours of growth. Arrows indicate chromosome 12 trisomic cells. (I) Single chromosome 12 trisomic cells observed in various colonies in diploid passage on the same coverslip. In each field, only single trisomic cells were present after three days of growth, with no neighboring trisomic cells detected within at least 90 µm. In (H) and (I), DNA labeling is shown in grey, with overlays merging DNA (blue) and chromosome 12 centromeres (red). Maximum intensity projections are shown. Scale bars = 20 µm. Scale bars in the enlarged regions = 10 µm.

To determine if trisomy 12 detected by centromeric probes reflected the gain of a whole chromosome 12 or a fragment, we used whole chromosome paint for chromosome 12 and applied it to chromosome spreads. This revealed the presence of two complete chromosome 12’s in diploid cells and three in trisomic cells (Fig. S1B–C). To investigate the robustness and consistency of the trisomy 12 takeover, we conducted independent biological replicates of the previous dual-color FISH mapping, initiated from cryopreserved cells from passage 3 of the AICS-0012 iPS cell line (Fig. S2). The combined findings show that chromosome 12 trisomic cells spontaneously appear in the culture, and once they constitute >5% of the population, they rise to nearly 100% within ~13 passages.

Overall, we have cultured the same AICS-0012 iPS cell line through transition passages a total of eight times, once in the previously published study (Dubose et al., 2022) and seven times more recently. In seven out of eight experiments, trisomy 12 became dominant to over 95% of the population, demonstrating that this is the most frequent outcome. In a single replicate, we did not observe significant trisomy 12 (Fig. S3A–B), and we considered the possibility that the cell population acquired other major chromosomal aberrations during culturing. However, KaryoStat+ analysis revealed no additional chromosomal gains or losses, indicating that in rare cases, iPSC cultures do not acquire significant chromosomal aberrations (Fig. S3C).

2.2. The difference in doubling times between chromosome 12 diploid and trisomic cells of the same origin is insufficient to account for the rapid dominance of the trisomic population.

In our previous study, the change in growth rate due to amplification of chromosome 12 was insufficient to explain the dominance of trisomic cells in the cell culture. However, the growth rates of chromosome 12 diploid and trisomic cells calculated in that study were based only on cell counting during passaging (Dubose et al., 2022). To obtain high-resolution proliferation measurements, we performed microwell growth assays (Fig. S3D–F) and determined that diploid cells showed a doubling time of 19.25 ± 1.29 hours (mean ± S.D.), while trisomic cells 17.68 ± 1.41 hours (mean ± S.D.) (Fig. 1D). These results are consistent with the literature data indicating human iPS cell division time of 18-20 hours (Chatterjee et al., 2016; Dubose et al., 2022).

We used doubling times to model the theoretical outgrowth of the trisomic population over diploid, starting from an initial average of 6% trisomic cells in the first transition passage. Based on the doubling times, we would have expected only 38% of trisomic cells after 13 passages (39 days in the cell culture) (Fig. 1E). However, our experiments consistently showed almost 100% trisomy 12 at that time (Fig. 1F), confirming that differences in doubling times alone cannot explain the rapid takeover by trisomic cells. We then introduced to the model a variable representing the percentage of diploid cells converting to trisomic per cell division. Our modeling showed that to reach 96% of trisomy 12 in 13 passages (as observed in FISH experiments), 3% of diploid cells must convert to trisomic during each population doubling (Fig. 1G).

To verify whether trisomy 12 cells arise simultaneously in multiple cells, we examined diploid and early transition passages. In early transition passages (~6% trisomy 12), we found primarily diploid cell colonies containing clusters of trisomic cells (Fig. 1H). This clustering suggests that trisomics were seeded and grew for 3 days, dividing approximately four times to form colonies with grouped daughter trisomic cells. However, after 3 days of growth, we also found multiple, isolated, single trisomic cells (Fig. 1I). These trisomic cells must have arisen less than one doubling time before fixation. Altogether, our data indicate that trisomy 12 arises simultaneously and independently in multiple cells.

2.3. Chromosome 12 shows a high propensity for bridging during anaphase in transition passage iPSCs.

To understand how chromosome 12 trisomy arises in transition passages we examined the segregation of chromosome 12 and control chromosome 10 in mitosis. Exclusively during the transition passages, 55% of detected anaphase bridges involved chromosome 12, a frequency 13-fold higher than expected by random chance (approximately 4.3% for any of the 23 human chromosomes). This level of bridging was never observed for chromosome 10. However, once trisomy 12 became dominant (in ~100% trisomic passages), the tendency for chromosome 12 to form bridges significantly decreased. In very early fully diploid passages, chromosome 12 bridging was not elevated (Fig. 2A-B).

Fig. 2. During transition passages only, chromosome 12 shows a high propensity for bridging during anaphase and is significantly overrepresented in micronuclei.

(A) The percentage of bridged chromosome 12 in pre-transition (~100% diploid), transition, and post-transition (~100% trisomic) passages. In transition passages, bridging chromosome 12 exceeds to the highest extent the theoretically assumed percentage of each chromosome bridging in anaphase (1 out of human 23 chromosomes randomly bridging would give approximately a 4.3% chance). The n values represent the total number of analyzed bridges from three biological replicates of trisomy 12 takeover. The dotted line indicates the random expected frequency for any single chromosome bridging in anaphase. (B) Anaphase cells with one copy of chromosome 12 bridging found in transition passages 15, 17, 18, and 21. DNA labeling is shown in grey, the overlays merge DNA (blue), chromosome 10 centromeres (yellow), and chromosome 12 centromeres (magenta). Overlays show maximum intensity projections, while DNA images show single z-slices centered on the anaphase chromosome bridges. Scale bars = 10 μm. Single z-slices focused on the bridged chromosome 12 chromatids are shown in the enlarged DNA regions, while enlarged overlays show maximum intensity projections. Scale bars in the enlarged regions = 5 μm. (C) Delayed separation of chromosome 12 chromatids observed in transition passages 21, 23, and 24 anaphase cells, possibly due to prior bridging. The overlays merge a single plane of DNA labeling (gray) with chromosome 12 centromeres (red). Red-boxed areas indicate the bridged and lagging chromatids. Maximum intensity projections are shown. Scale bars = 10 µm. (D) The percentage content of chromosome 10 and 12 in micronuclei found in diploid cells (pre-transition passages), transition passage cells, and trisomic cells (post-transition passages) in three biological replicates of trisomy 12 takeover. The dotted line indicates the random expected frequency for any single chromosome being mis-segregated into the micronuclei (1 chromosome out of 23 = 4.3%). The percentage of chromosome 12 in micronuclei of transition passages exceeds the randomly expected percentage of 4.3% to the greatest extent. The n values represent the total number of analyzed micronuclei in three biological replicates of trisomy 12 takeover. Data points are shape-coded by biological replicate and presented as mean ± SD; p values were determined by unpaired two-tailed Student’s t-test, * p=0.0318, ns = not significant. Data distribution was assumed to be normal but was not formally tested. (E) Representative micronuclei observed in transition passages 21 and 22 containing chromosome 12 chromatids. In FISH, identifying cells in S or G2 phases - where chromosome 10 and/or 12 are replicated and exhibit distinct signals is rare. However, passage 22 depicts chromosome 12 diploid cell with already replicated chromosome 12 (4 centromeric signals). Remarkably, the fifth chromosome 12 signal of comparable fluorescence intensity to those in the main nucleus originates from the micronucleus, suggesting the presence of a fifth chromosome 12 chromatid in the cell. DNA labeling is shown in grey, the overlays merge DNA (blue), chromosome 10 centromeres (yellow), and chromosome 12 centromeres (magenta). Maximum intensity projections are shown. Scale bars = 10 µm. Single z-slices centered on the micronuclei are shown in DNA enlarged regions, while enlarged overlays show maximum intensity projections. Scale bars in the enlarged regions = 5 μm. (F) Counts of micronucleated cells as a fraction of the total number of micronuclei and the total number of analyzed cells in three biological replicates of trisomy 12 takeover. The values for pre-transition passages (18520 cells), transition passages (24208 cells), and post-transition passages (8854 cells) represent the cumulative results of three independent counts from three experiments. (G) Fluorescence intensity (F.I.) of chromosome 12 centromeres in metaphase cells was normalized to 2. Each dot in the metaphase group (n = 112) represents the F.I. of two adjoined chromosome 12 centromeres. Chromosome 12 F.I. in anaphase cells (n = 130), as well as in interphases (n = 398) and representative micronuclei (n = 18) was normalized to metaphase cells. Data present F.I. measurements obtained from representative micronuclei found in three biological replicates of trisomy 12 takeover, with mitotic and interphase cells analyzed from the same coverslips as the analyzed micronuclei. Each dot represents the F.I. of a single signal. Data points are shape-coded by biological replicate. Data presented as mean ± SD; p values were determined by unpaired two-sided Mann-Whitney U test **** p<0.0001. Normality was assessed using the Shapiro-Wilk test.

An anaphase chromosome bridge may occur when cells enter mitosis with fused sister chromatids resulting in lagging of the bridged chromosomes in anaphase. This lag can position bridged chromosomes near the spindle equator, leading to mis-segregation of both chromatids into the same daughter nucleus or into a micronucleus (Pampalona et al., 2016). In late anaphase cells, we detected separated but lagging chromosome 12 chromatids consistent with the idea that the bridged chromatids eventually rupture their connection and separate (Fig. 2C) (Anglada et al., 2025). Moreover, although chromosome 12 is average in size and gene content (Chromosome Map - Genes and Disease - NCBI Bookshelf), in our observations it forms bridges that localize at the spindle periphery of anaphase cells, in contrast to the general tendency of primarily the largest chromosomes to segregate toward the cell’s exterior during mitosis (Booth et al., 2016; Mosgöller et al., 1991; Klaasen et al., 2022).

2.4. Chromosome 12 is entrapped in micronuclei at high frequency.

We reasoned that bridging chromosome 12 chromatids may become entrapped in micronuclei of transition passage cells. The high proliferation rates of human pluripotent stem cells make them particularly susceptible to micronuclei formation (Saxena and Zou, 2022; Henry et al., 2019). We compared the rates of mis-segregation of chromosomes 10 and 12 into micronuclei during continuous culture (Fig. 2D-E). We found that chromosome 12 was enriched in micronuclei only in transition passages, where its mis-segregation rate (12.9%) was significantly higher (p=0.0318) than that of chromosome 10 (4.4%). In contrast, in pre-transition and post-transition passages, the mis-segregation rates for chromosomes 10 and 12 were not significantly different (Fig. 2D). Overall, the frequency of micronucleated cells remained stable across pre-transition, transition, and post-transition cells, with no significant change in their frequency (1.9%, 2.4%, and 2.3%, respectively) (Fig. 2F).

To investigate whether chromosome 12 signals in the micronuclei reflected single chromatids or whole chromosomes we used quantitative fluorescence. We generated standards by measuring the fluorescence signals exhibited on metaphase chromosomes that contain two centromeres and signals from anaphase chromosomes that contain single centromeres. Our analysis showed that the fluorescence of chromosome 12 centromere in micronuclei was significantly lower (p<0.0001) than the intensity of metaphase chromosomes and comparable to the intensity of anaphase chromatids (Fig. 2G), indicating that these signals originate from single chromatids.

2.5. The presence of intact lamin B1, γ-H2AX expression, and the percentage of large micronuclei indicate that over 50% of micronuclei in human iPSCs remain stable.

Micronuclei are a hallmark of chromosome instability in cancer cells where they often exhibit defects in nuclear lamina, leading to rupture of their envelopes, followed by enzymatic shattering of the enclosed chromosomes, a process termed chromothripsis (Zhang et al., 2015; Hatch et al., 2013; Terradas et al., 2018; Hatch and Hetzer, 2015; MacIejowski and Hatch, 2020; Liu et al., 2018; Lusk and King, 2018). We examined the distribution of lamin B1 in diploid and early transition passages of the AICS-0013 cell line, which carries an mEGFP tag on the N-terminus of the lamin B1 gene LMNB1. Lamin B1 was present in all primary nuclei and approximately 60% of micronuclei (Fig. 3A-C). Any micronuclei exhibiting gaps in the nuclear lamina meshwork, which create weak points susceptible to membrane rupture (Mammel et al., 2022; Hatch et al., 2013), were scored as lamin B1-negative.

Fig. 3. Transition passage cells enter mitosis with one unpaired chromosome 12 chromatid originating from prior mis-segregation into micronuclei.

(A) and (B) Representative iPS cell micronuclei with and without lamin B1 expression in the micronuclear membrane. 3D projections were used to assess lamina integrity. DNA is shown in grey, and the overlay represents DNA (blue) and lamin B1 (green). Maximum intensity projections are shown. Scale bars = 10 µm. Scale bars in the enlarged regions = 4 µm. (C) Percentage of lamin B1 presence in micronuclei of diploid and early transition passages. GFP visualization of the LMNB1-GFP-tagged iPS cell line showed that lamin B1 was present in all primary nuclei and ~60% of micronuclei. The n values represent the total number of analyzed micronuclei. (D) Representative images of γ-H2AX immunolabeling of transition passage iPS cells. DNA labeling is shown in grey, and the overlay represents DNA (blue) and γ-H2AX (yellow). Maximum intensity projections are shown. Scale bars = 10 µm. Scale bars in the enlarged regions = 4 µm. (E) iPS cells with micronuclei in transition passages analyzed for γ-H2AX presence in both primary nuclei and micronuclei. Of over 130 cells with micronuclei, ~50% had γ-H2AX-negative micronuclei, the majority of which were also negative for the presence of double-strand breaks in the main nuclei. (F) Area of micronuclei measured across five groups, with the first group representing micronuclei containing missegregated chromosome 12 chromatids. Micronuclei with chromosome 12 (12.45 ± 9.58 μm², mean ± S.D.) were compared to lamin B1-positive (12.34 ± 6.39 μm²), lamin B1-negative (3.12 ± 2.81 μm²), γ- H2AX-positive (3.46 ± 2.63 μm²), and γ-H2AX-negative (10.41 ± 5.90 μm²) micronuclei. Micronuclei containing chromosome 12 were significantly larger than lamin B1-negative and γ-H2AX-positive micronuclei but were not significantly different from lamin B1-positive or γ-H2AX-negative micronuclei. Data presented as mean ± SD (n = 30 for each group); p values were determined by unpaired two-sided Mann-Whitney U test and unpaired two-tailed Student’s t-test, **** p<0.0001. Normality was assessed using the Shapiro-Wilk test. (G) Model of chromosome mis-segregation and lamin B1 recruitment in the subsequent interphase, adapted from Liu et al. (Liu et al., 2018). Micronuclei formed outside the spindle are more stable, as they recruit non-core proteins, such as lamin B1. Lagging chromosome 12 was never observed within the midspindle. However, when chromosome 12 bridges or mis-segregation events were observed, they predominantly occurred as peripheral bridges or mis-segregation outside the midspindle, as illustrated in the example anaphase cell. DNA labeling is shown in grey, and the overlay merges DNA (blue), chromosome 12 centromeres (magenta), and chromosome 10 centromeres (yellow). Maximum intensity projections are shown. Scale bar = 10 µm. (H) Human iPS cells with micronuclei in transition passages with incorporation of EdU showing DNA replication in nuclei and micronuclei within the same cells. The most frequent outcome was EdU incorporation into primary nuclei but not micronuclei within the same cells. (I) The four scored patterns of EdU incorporation in iPSCs with micronuclei after pulse labeling with 20 μM EdU for 20 min. DNA labeling is shown in grey, the overlay merges DNA (blue) and EdU labeling (red). Maximum intensity projections are shown. Scale bars = 5 µm. (J) and (K) show schematics and the corresponding FISH labeling on the regular metaphase and anaphase cells of chromosome 12 diploid and trisomic cells. (L) Metaphase cells observed in transition passages with signals from one potentially unpaired chromosome 12 chromatid apart from the metaphase plates. In metaphase cells, two copies of chromosome 12 are found in the aligned metaphase plate, while the third copy lies at some distance off the metaphase plate. In passage 26 (bottom), chromosome 10 signals appear as a pair of centromeres, representing sister chromatids still aligning at the metaphase plate. In contrast, the off-plate chromosome 12 signal appears as a single centromere. In a separate experiment using whole chromosome 12 probes, again, two copies of chromosome 12 are found in the aligned metaphase plate (one is weakly visible due to being positioned deeper within the metaphase plate), while the third copy lies off the metaphase plate. (M) Anaphase cells from multiple transition passages with the normal 2 and 2 distribution of control chromosome 10 chromatids, but anomalous organization of 3 chromatids of chromosome 12 oriented to one pole and 2 chromatids oriented to the other pole (3:2 anaphase cells). In (J) - (M), DNA labeling is shown in grey, and the overlays merge DNA (blue) and centromeric probes for chromosome 10 (yellow) and chromosome 12 (magenta). The right overlay in panel (L) merges DNA (blue) and whole chromosome paint for chromosome 12 (red). Maximum intensity projections are shown. Scale bars = 10 µm. (N) Fluorescence intensity (F.I.) of chromosome 12 centromeres in diploid and trisomic metaphase cells was normalized to 2. Each dot in the metaphase group (n = 202) represents the F.I. of two adjoined chromosome 12 centromeres. Chromosome 12 F.I. in diploid and trisomic anaphase cells (n = 393), as well as in diploid poles of 3:2 anaphase cells (n = 26) and outside of the metaphase plates (n = 12) was normalized to metaphase cells. Each dot in these groups represents the F.I. of an analyzed signal. Data present F.I. measurements from three biological replicates of trisomy 12 takeover, with normal mitotic cells F.I. analyzed from the same coverslips in which the abnormal mitotic cells were identified. Data points are shape-coded by biological replicate and presented as mean ± SD; p values were determined by unpaired two-sided Mann-Whitney U test. P-values are as follows: ** p=0.0017, **** p<0.0001. Normality was assessed using the Shapiro-Wilk test.

Smaller micronuclei have been reported to inefficiently form intact lamina, making them prone to rupture and accumulation of DNA damage (Kneissig et al., 2019). We found that the presence of an intact lamin B1 shell correlated with larger micronuclei size. Micronuclei containing chromosome 12 were not significantly different in size from lamin B1-positive micronuclei and were significantly larger (p<0.0001) than lamin B1-negative micronuclei (Fig. 3F). Thus, micronuclei that contain chromosome 12 are expected to contain intact nuclear lamina and exhibit a reduced frequency of rupture.

In cancer cells, micronuclei with absent, faint, or irregular lamin B1 exhibit significantly higher levels of DNA damage (Vázquez-Diez et al., 2016). To assess DNA damage in iPSC micronuclei, we used γ-H2AX immunolabeling (Fig. 3D). Approximately 50% of micronuclei analyzed were γ-H2AX-positive (Fig. 3E), consistent with our observation that approximately 40% of micronuclei lacked intact lamin B1. Micronuclei containing chromosome 12 were similar in size to γ-H2AX-negative micronuclei and were significantly larger (p<0.0001) than γ-H2AX-positive micronuclei (Fig. 3F). Thus, chromosome 12 chromatids were mis-segregated into larger micronuclei containing intact lamina, shown to be protected from DNA damage.

Central spindle microtubules impede the assembly of nuclear envelope proteins upon micronuclei formation on mis-segregating chromosomes, leading to defects in envelope formation, while chromosomes mis-segregating on the spindle periphery or at poles generate micronuclei with intact lamina (Liu et al., 2018). Consistently, we found peripheral positioning of chromosome 12 bridges correlating with the formation of larger micronuclei that exhibit a lower propensity for rupture (Fig. 3G).

2.6. Micronuclei in human iPSCs fail to replicate DNA in S phase.

We examined DNA replication in micronuclei by incubating cells with the thymidine analog, 5-ethynyl-2’-deoxyuridine (EdU). Following a short EdU pulse, cells were fixed and labeled. We found that nearly all micronuclei (~95%) failed to incorporate EdU when their primary nuclei did. We did not observe evidence of delayed replication in micronuclei at later stages of the cell cycle (Fig. 3H-I). We concluded that DNA in iPSC micronuclei rarely undergoes replication, leading to chromosomes in micronuclei entering mitosis without being replicated.

2.7. Transition passage iPS cells enter mitosis with a single unpaired chromosome 12 chromatid.

In diploid cells, 2 signals for both chromosomes 10 and 12 appear at metaphase and 4 signals at anaphase (Fig. 3J). In trisomic cells, 2 chromosome 10 signals and 3 chromosome 12 signals appear in metaphase. These resolve into 4 chromosome 10 signals and 6 chromosome 12 signals in anaphase (Fig. 3K). However, in 1.5% of 798 analyzed metaphase cells during transition passages, we observed a chromosome 12 signal positioned away from the metaphase plate (Fig. 3L). Failure to fully align at metaphase would be expected if these signals reflected an unpaired chromosome 12 chromatid. This was never seen for chromosome 10. Similarly, in 2% of 655 analyzed anaphase cells in transition passages, chromosome 12 showed 2 signals oriented to one pole and 3 to the other (Fig. 3M). This 3:2 anaphase segregation was never detected for chromosome 10.

Given the unprecedented nature of the existence of a single, unpaired, and unreplicated chromosome 12 chromatid, we devised a method for quantitative fluorescence measurements of centromere FISH probes to distinguish between single chromatids and whole chromosomes. This approach allowed us to address two important questions. First, do the chromosome 12 signals lying apart from the metaphase plate reflect whole chromosomes or single chromatids? Second, might the 3:2 anaphase cells be explained by trisomic anaphase cells where two chromosome 12 centromeres on one side overlapped, thus generating one anaphase signal that contained two chromatids? Our data revealed that the chromosome 12 fluorescence signals lying apart from the metaphase plate were consistent with most originating from a single chromatid. Similarly, the fluorescence signals in each of the centromeres in the diploid pole of 3:2 anaphase cells were primarily due to single chromatids rather than two superimposed centromeres (Fig. 3N).

Although chromosome mis-segregation has been extensively characterized in human stem and cancer cells, our dual-color FISH analysis identified a previously undocumented 3:2 distribution of chromosome 12 chromatids during anaphase. Our data show that at least 1% of transition passage cells enter mitosis with an abnormal chromosome 12 configuration, two intact chromosomes, and one unpaired chromatid, originating from the previous mis-segregation into the micronuclei.

2.8. Why chromosome 12? Short telomeres and faster proliferation.

2.8.1. Chromosome 12p arms create bridges in anaphase during trisomy 12 takeover.

Our next focus was to address the question: why is chromosome 12 prone to this mis-segregation? Intriguingly, the shortest telomeres in humans are located on the 17p, 20q, and 12p arms (Karimian et al., 2024), and the same chromosome arms are most frequently involved in chromosomal aberrations observed in cultured human embryonic and pluripotent stem cell lines (Seol et al., 2008; Draper et al., 2004; Na et al., 2014; Ben-David et al., 2014; Andrews et al., 2022; Halliwell et al., 2020; Stavish et al., 2024). Chromosome 12 exhibited a high propensity for bridging and these bridges occurred at the ends of the arms (Fig. 2A-C). To test whether bridging involved p or q arms, we compared their lengths in diploid and trisomic metaphase spreads (Fig. 4A-B). 12p arms were significantly shorter than 12q arms (p<0.0001), with no significant difference in 12p length between diploid and trisomic cells, and a slight increase in 12q length in trisomic spreads (p=0.0313). We next examined the length of bridging arms and found they were significantly shorter than 12q (p=0.0097), indicating they most likely originated from 12p arms (Fig. 4C, E). A stretch ratio analysis further supported this conclusion (Fig. 4D). Together, our data indicate that the bridges originate from the shorter 12p arms, contributing to the limited understanding of how specific chromosome arms are involved in mitotic mis-segregation (Lee et al., 2023; Saunders et al., 2000; McClintock, 1950).

Fig. 4. Chromosome 12p arms show a high propensity for bridging in anaphase cells and show disrupted subtelomeres, providing a pathway for the generation of trisomy 12 cells.

(A) Chromosome 12 in diploid and trisomic metaphase spreads after labeling with whole chromosome 12 probes used to analyze the length of 12p and 12q arms. The overlays merge DNA (grey) and chromosome 12 (red). Maximum intensity projections are shown. Scale bars = 2 µm. (B) Quantification of 12p and 12q arm lengths in diploid and trisomic metaphase spreads labeled with whole chromosome 12 probes. The 12p arm measured 1.16 ± 0.21 µm (mean ± S.D.) and the 12q arm measured 2.85 ± 0.60 µm (mean ± S.D.). Each dot on the graph represents the length of a single arm. The n values represent the number of measured arms. Data presented as mean ± SD; p values were determined by unpaired two-tailed Student’s t-test and unpaired two-sided Mann-Whitney U test. P-values are as follows: * p=0.0313, **** p<0.0001, ns = not significant. Normality was assessed using the Shapiro-Wilk test. (C) The total length of both bridging chromosome 12 arms (4.91 ± 0.63 µm; mean ± S.D.) compared to the doubled 12q arm length in the metaphase spreads. Bridging arms were significantly shorter than the doubled 12q arms. Each dot on the graph represents the length of two arms. Data presented as mean ± SD; p value was determined by unpaired two-sided Mann-Whitney U test, ** p=0.0097. Normality was assessed using the Shapiro-Wilk test. (D) Comparison of measured bridging 12p arm length to the predicted bridging 12q length after applying the average p-arm stretch ratio. If bridges originated from 12q, their expected length would be ~5.9 µm. Each dot on the graph represents the length of a single arm. Data presented as mean ± SD; p value was determined by unpaired two-sided Mann-Whitney U test, **** p<0.0001. Normality was assessed using the Shapiro-Wilk test. (E) Examples of chromosome 12 bridges in transition passage anaphase cells. The overlays merge DNA (grey) and chromosome 12 centromere (red). Single z-slices focused on the bridged chromosome 12 chromatids are shown. Scale bars = 10 µm. (F) Percentages of retained chromosome 12 subtelomeric regions in diploid, parental WTC-11 cells, and AICS-0012 cells before, during and after trisomy 12 takeover. The n values represent the number of analyzed 12p and 12q subtelomeres. (G) Passaging history of the diploid, parental WTC-11 cell line during the development of the AICS-0012 cell line and further passaging of the AICS-0012 cell line through the initial appearance and takeover of trisomy 12. (H) and (I) Representative images of metaphase chromosomes labeled with 12p and 12q subtelomeric probes in the diploid WTC-11 cells, and diploid and transition passages of the AICS-0012 cell line. Red arrows indicate 12p arms lacking subtelomeric signals and white boxes show enlargements of chromosomes missing 12p subtelomeric repeats. DNA is labeled in grey and the overlays merge DNA (blue) and chromosome 12 subtelomeres (red). Maximum intensity projections are shown. Scale bars = 10 µm. Scale bars in the enlarged regions = 2 µm. (J) Sequence of cell cycle events generating new chromosome 12 trisomic cells. When trisomy 12 takes over the culture, bridging chromosome 12p arms create one chromosome 12 monosomic cell that will not survive, and another daughter cell, where, in some cases, a chromosome 12 chromatid generates a micronucleus. This chromosome 12 trapped in the micronucleus fails to replicate properly in the next S phase. The resulting G2 cell has 2.5 copies (5 chromatids) of chromosome 12. This would generate a metaphase cell with one unpaired chromatid, and a 3:2 anaphase, both of which we observed in the transition passages. This 3:2 anaphase would generate a new chromosome 12 trisomic cell. (K) Trisomy 12 outgrowth modeled by incorporating the doubling time differences and an additional variable of 1% of diploid cells converting to trisomic at each cell division. After applying this variable, the mathematical model showed 80% of trisomy 12, closer to the observed trisomy 12 dominance.

2.8.2. During trisomy 12 takeover almost 20% of chromosome 12p arms appear with lost subtelomeric repeats.

Subtelomeric sequences, located near the ends of chromosome arms and adjacent to telomeres, play a crucial role in protecting telomeres from degradation (Karimian et al., 2024; Meena et al., 2015; Günes and Rudolph, 2013; Kwapisz and Morillon, 2020; Henry et al., 2019; Li et al., 2020). The inherent shortness of 12p telomeres (Karimian et al., 2024) may enhance their vulnerability to DNA replication stress-induced erosion, thereby increasing the likelihood of bridge formation (Rivosecchi et al., 2024; Beishline et al., 2017; Almeida et al., 2018; Romanov et al., 2001; O’Sullivan et al., 2002; Stewénius et al., 2005; Pampalona et al., 2010). This led us to hypothesize that the bridging of chromosome 12 p arms might be due to damage at the telomere and subtelomere regions of the 12p arms.

To specifically assess damage to chromosome 12, we used probes that specifically label subtelomeres of both p and q arms of chromosome 12 in metaphase spreads. We labeled diploid parental cells (WTC-11), as well as diploid, transition, and trisomic passages from the AICS-0012 cell line undergoing trisomy 12 takeover (Fig. 4F-I). In the parental WTC-11 population, 94% of 12p and 99% of 12q subtelomeric regions were labeled. After ~30 passages, in early but still diploid passages of the AICS-0012 cell line (prior to trisomy 12 emergence), 12p subtelomere labeling had decreased by 7%, while 12q showed only a 1% reduction. Subsequently, during the emergence of trisomy 12 (0% to 70% over 18 passages), 12p subtelomeres declined by an additional 9%. In total, 12p subtelomere loss reached 16%, compared to only 5% for 12q. Subtelomere levels remained stable from passage 25 onward, as trisomy 12 became fixed in the population (Fig. 4F).

Our findings show that subtelomeric repeats loss or damage on the 12p arms during the critical, transition passages can play a role in the onset and dominance of trisomy 12. This loss can disrupt the proper separation of 12p arms during anaphase, resulting in the observed bridging. The elevated propensity for subtelomere loss in the 12p arms, along with the high incidence of bridging in these arms, supports a link between subtelomeric instability and the increased rates of chromosome 12 mis-segregation observed during transition passages.

We previously observed that during transition passages, 55% of the detected anaphase bridges involved chromosome 12. However, once trisomy 12 became dominant, the tendency for chromosome 12 to form bridges significantly decreased (Fig. 2A-B). This decline in bridging coincided with the stabilization of 12p subtelomeric regions, as evidenced by the absence of further loss of subtelomeric repeats after cells reached full trisomy 12. At 70% trisomy 12, nearly 20% of 12p subtelomeric repeats were lost, suggesting that rapid aneuploidy progression might have temporarily overwhelmed telomeric maintenance mechanisms. However, the stabilization observed in trisomic passages suggests that 12p ends are somehow stabilized. We hypothesize that the decline in 12p bridging reflects successful restoration of shortened telomeres and/or chromosome ends, enabling stabilization of the trisomic state. Future work will focus on directly analyzing telomere dynamics across iPSC passages.

2.9. Mechanisms driving trisomy 12 takeover in human iPSCs.

Based on our findings, we propose a model for how chromosome 12 trisomic cells arise in human iPSCs (Fig. 4J). From a diploid parent, bridging chromosome 12p arms create a monosomic daughter cell that rapidly dies, and a diploid cell with an extra chromatid 12 trapped in a micronucleus. This micronucleus fails to replicate DNA during S phase and creates a G2 cell with 5 chromosome 12 chromatids. Therefore, in the next mitosis, this G2 cell generates a 3:2 anaphase. The cell receiving three copies of chromosome 12 becomes a new trisomic cell.

To test the model, we introduced a variable representing the percentage of diploid cells converting to trisomic per cell division observed in our FISH mitosis studies. Our data showed that 2% of anaphase cells in transition passages exhibited 3:2 segregation, generating one diploid and one trisomic cell (Fig. 3M). Using a conservative rate of 1% resulted in the trisomy 12 reaching a predicted 80% after 13 passages (Fig. 4K). This prediction more closely matches the observed outgrowth of trisomic cells in FISH experiments (Fig. 1F). To independently verify the conversion rate of diploid cells to trisomic, we performed single colony analyses of the transition passage with ~6% trisomy 12 (Fig. S4). We estimated an average diploid-to-trisomic conversion rate of 1.46% per cell division, supporting the high missegregation frequency observed in our mitotic cells FISH analyses (Fig. 3L-M).

To highlight the importance of both mechanisms, growth rate differences and the high magnitude of mis-segregation events through mathematical modeling, we modeled scenarios where the doubling times of diploid and trisomic cells are identical, and high diploid-to-trisomic conversion rates (1% and 2%) are applied (Fig. S5). These simulations show that even with a 2% conversion rate, trisomic cells do not reach 70% of the population within 13 passages, emphasizing the critical role of faster growth rates in driving the takeover (Fig. S5A). We also modeled the effect of the doubling time differences between diploid and trisomic cells but with lower mis-segregation rates (0.01% and 0.1%). Under these conditions, trisomic cells fail to reach 50% of the population within 13 passages, demonstrating that the frequency of chromosome 12 mis-segregation events must be remarkably high to account for that rapid takeover (Fig. S5B). Finally, we present comprehensive models incorporating both mechanisms: the doubling time advantage of trisomic cells and high conversion rates (1%, 2%, and 3%). Only when both factors are combined does the model replicate the observed rapid dominance of trisomic cells in culture (Fig. S5C). Together, these findings underscore that both the difference in growth rates and the exceptionally high mis-segregation frequency are crucial for driving the rapid emergence and dominance of trisomy 12 in human iPSCs.

While our study identifies chromosome 12-specific mis-segregation as one of the drivers of trisomy 12, gene dosage effects in trisomic cells may potentially contribute to both their selective advantage and chromosomal instability. Chromosome 12 harbors several genes associated with pluripotency, including NANOG and GDF3, as well as genes like KRAS, MDM2, and CDK4, which could enhance cell survival and growth while overexpressed (Ben-David et al., 2014; Draper et al., 2004; Baker et al., 2007; Mayshar et al., 2010; Goel et al., 2022; Yao et al., 2024). The upregulation of these genes may contribute to the growth advantage we observe in trisomic cells. Transcriptomic studies will be essential to clarify whether specific genes influence chromosome 12 instability and cell proliferation to drive trisomy 12 dominance. Moreover, our current approach could not detect potential, small chromosomal changes. Future sequencing-based studies, targeting particularly the subtelomeric and telomeric regions of chromosome 12, will be important to confirm the causes and consequences of chromosome 12 mis-segregation.

2.10. Chronic low-dose hydroxyurea treatment advances chromosome 12 mis-segregation and accelerates trisomy 12 onset.

We observed almost 20% loss of subtelomeric regions on the 12p arms during the trisomy 12 takeover, which we reasoned is driven by DNA replication stress (Karimian et al., 2024; Meena et al., 2015; Günes and Rudolph, 2013). To test this, we induced chronic replication stress using low-dose hydroxyurea (HU), a ribonucleotide reductase inhibitor that perturbs dNTP synthesis and DNA replication fork progression (Arlt et al., 2011; Shaw et al., 2024). We treated early-passage diploid iPS cells with 50 μM HU for 36 days (12 passages) (Fig. 5A). By the end of treatment, our analysis showed that HU-treated cells exhibited significantly higher (p=0.0068) levels of trisomy 12 (8.2%) compared to controls (4.6%) (Fig. 5B). HU-treated cells also showed a greater number of detected chromosome 12 anaphase bridges and an increased percentage of micronuclei containing chromosome 12 (19.6%), representing a two-fold increase over controls (Fig. 5C-D, F). Overall, micronucleation frequency was also elevated in the HU-treated group (Fig. 5E). Importantly, at this concentration, HU did not affect the growth rate or colony morphology compared to control conditions.

Fig. 5. Hydroxyurea treatment drives earlier trisomy 12 acquisition.

(A) Schematic of the experiment used to test whether DNA replication stress accelerates the onset of trisomy 12. Fully diploid iPSCs were cultured in parallel under control conditions or treated daily with 50 μΜ hydroxyurea (HU) from passage 8 to passage 20 (36 days in the cell culture). Each condition was carried out in three independent replicates, starting from the same diploid population. (B) Percentage of trisomy 12 measured by FISH in three replicates at the start of treatment (passage 8), and after 12 passages under control and HU-treated conditions. While trisomy 12 was undetected at passage 8, by passage 20, HU-treated cells exhibited a significantly higher percentage of trisomy 12 compared to controls. The n values represent the total number of analyzed cells. Data points are shape-coded by biological replicate and presented as mean ± SD; p value was determined by unpaired two-tailed Student’s t-test, ** p=0.0068. Data distribution was assumed to be normal but was not formally tested. (C) The number of chromosome bridges observed in anaphase cells at passage 20 in control and HU-treated conditions. HU-treated cells showed a higher number of chromosome 12 bridges compared to controls. (D) Percentage of micronuclei containing chromosome 12 at passage 20 in three replicates of control and HU-treated iPSCs. HU treatment resulted in a two-fold increase in chromosome 12 content in micronuclei compared to controls. Both control and HU-treated samples exceeded the frequency expected from random segregation (indicated by the dotted line at 4.3%, corresponding to 1 out of 23 chromosomes). Chromosome 12 mis-segregation into micronuclei under HU treatment occurred at almost five times the random expectation. The n values represent the total number of micronuclei analyzed. Data points are shape-coded by biological replicate and presented as mean ± SD; p value was determined by unpaired two-tailed Student’s t-test, * p=0.0498. Data distribution was assumed to be normal but was not formally tested. (E) Counts of micronucleated cells as a fraction of the total number of micronuclei and the total number of analyzed cells in control (2754 cells) and HU-treated (3428 cells) passage 20 represent the cumulative results of three replicates. (F) Representative images of passage 20 under control and HU-treated conditions. DNA labeling is shown in grey, and the overlays merge DNA (blue) and chromosome 12 centromeres (red). White arrows indicate chromosome 12 trisomic cells. Enlarged regions highlight chromosome 12 mis-segregation. Maximum intensity projections are shown. Scale bars = 20 µm.

These results support a model in which increased replication stress can promote chromosome 12-specific mis-segregation events, thereby accelerating the emergence of trisomy 12 in human iPS cells. While additional experiments will be needed to determine whether endogenous replication stress fluctuates during the transition phase, our findings suggest that replication stress may play a mechanistic role in driving the selective advantage of trisomy 12.

Here we show that trisomy 12 appearance and dominance in iPSC culture does not arise from rare, aberrant cells, but stems from massive mitotic mis-segregation specific to chromosome 12. The newly formed trisomic cells subsequently persist and outgrow the diploid population. By uncovering de novo emergence of trisomic cells, bridging exclusively through 12p arms, unpaired chromosome 12 chromatids in mitosis, and subtelomeric instability during transition passages, we provide new insights into trisomy 12 dominance in iPSCs. Importantly, the identification of DNA replication stress as a trigger for this process offers a valuable target for future intervention strategies. Our discovery represents a novel paradigm for whole chromosome instability, with significant implications for research and regenerative therapies involving stem cells, as well as for understanding the origins of aneuploidy.

3. Materials and Methods

3.1. Human induced pluripotent stem cell lines

This research was made possible through the Allen Cell Collection, available from Coriell Institute for Medical Research. Human iPS cell lines used in this study, identifiers AICS-0012 cl. 105 (TUBA1B-mEGFP, Coriell Cat# AICS-0012, RRID:CVCL_IR34) and AICS-0013 cl. 210 (LMNB1-mEGFP, Coriell Cat# AICS-0013, RRID:CVCL_IR32) were derived from the WTC-11 cell line (Coriell Cat# GM25256, RRID:CVCL_Y803) at the Allen Institute for Cell Science (allencell.org/cell-catalog). Cell lines were obtained from the Allen Cell Collection via the Coriell Institute for Medical Research. Briefly, the WTC-11 cell line has been derived from healthy, male donor fibroblast cells using episomal vectors. The identity of the unedited WTC-11 line was confirmed with short tandem repeat profiling. Whole genome sequencing and population-level RNA-seq data for the WTC-11 iPSC line are publicly available from the Allen Cell Institute/UCSC Genome Browser (Kreitzer et al., 2013). Each gene-edited cell line harbors a fluorescently tagged protein localizing to a cellular structure. The CRISPR/Cas9-mediated genome editing methodology used to generate these cell lines was described in (Roberts et al., 2017). The donor plasmid used to generate the AICS-0012 cell line is AICSDP-4 TUBA1B-mEGFP (Addgene, Cat. #87421), and to generate the AICS-0013 cell line is AICSDP-10: LMNB1-mEGFP (Addgene, Cat. #87422) and can be obtained through Addgene (https://www.addgene.org/The_Allen_Institute_for_Cell_Science/). Specifically, to obtain the AICS-0012 cell line, using CRISPR/Cas9, WTC-11 iPSCs at passage 33 were endogenously tagged with mEGFP in a single allele of the TUBA1B gene and then cloned (clone 105) at the Allen Cell Institute. AICS-0012 cells were obtained from the Coriell Institute at passage 32 from the Allen Cell Collection and then cultured in this study between passages 1 to 61. To obtain the AICS-0013 cell line, using CRISPR/Cas9, WTC-11 iPSCs at passage 33 were endogenously tagged with mEGFP in a single allele of the LMNB1 gene and then cloned (clone 210) at the Allen Cell Institute. AICS-0013 cells were obtained from the Coriell Institute at passage 30 from the Allen Cell Collection and then cultured in this study between passages 1 to 16. Edited AICS-0012 and AICS-0013 cell lines were karyotypically normal based on at least 20 metaphase cells analyzed by G-banding, did not possess any mutations at off-target sites, and had no acquired mutations 8 passages beyond the original line WTC-11 that affect genes in Cosmic Cancer Gene Census.

3.2. Human induced pluripotent stem cell culture

Human iPS cell lines were cultured in 25 cm² and 12.5 cm² flasks during standard passaging, in 96-well plates for proliferation assays, 6-cm cell culture dishes for chromosome spreads, and in 6-well plates on square cover glass coverslips for FISH and immunofluorescence assays. All growth surfaces were coated in Growth Factor Reduced (GFR) Matrigel (Cat. #354230, Corning) diluted in phenol red-free DMEM/F12 (Cat. #21041025, Gibco, Thermo Fisher Scientific) at a ~1:30 ratio (Matrigel final protein concentration = 0.337mg/mL). Matrigel plates and flasks were stored at 4°C and used within 1 week from coating.

Cells were maintained with mTeSR Plus media (Cat. #1000276 and #1000274, STEMCELL Technologies) supplemented with 1% penicillin-streptomycin (Cat. #SV30010, Cytiva). Cells were passaged every 3 days (~72 hours) as single cells. At each passage, iPSCs were split between a 1:4 to 1:8 ratio depending on density, the following experiment design, or when colonies began to merge. Cells were transferred to a new flask following detachment via Accutase (Cat. #A1110501, Gibco, Thermo Fisher Scientific) and blocking via 1× Dulbecco’s Phosphate-Buffered Saline (DPBS) (Cat. #14190144, Gibco, Thermo Fisher Scientific) according to the manufacturer’s instructions. Culturing media was supplemented with 10 μM Y-27632 Rho-associated protein kinase (ROCK) inhibitor (Cat. #HY-10583, MedChem Express) for approximately 24 hours following passaging, then changed to media without ROCK inhibitor (Rivera et al., 2020). Cells were fed daily with fresh room temperature (RT) complete mTeSR Plus media and maintained at 37°C under a humidified atmosphere of 5% CO₂ in a water-jacketed incubator. Cells were passaged and maintained continuously for various passaging times but a maximum of approximately 180 days (61 passages). Human iPSC colonies maintained normal morphology and had an average death rate of 2% Trypan-blue stain (Cat. #T10282, Invitrogen, Thermo Fisher Scientific) was used for cell counts and to assess the percentage of dead cells at each passage using the Countess Automated Cell Counter. A detailed protocol is available at the Allen Cell Explorer (https://www.allencell.org/cell-catalog.html).

3.3. Fluorescent in situ Hybridization (FISH)

3.3.1. Preparation of metaphase chromosomes from human iPSCs

Human iPSCs were grown in a 6-cm cell culture dish until reached 80-90% confluence and were treated with 0.1 μg/mL Colcemid (Cat. #15212012, KaryoMAX Colcemid, Gibco, Thermo Fisher Scientific) and 0.5 μM Wee1 inhibitor (PD0166285, Cat. #HY-13925, MedChemExpress) for 2.5 hours. The medium was transferred to the 15 mL conical; cells were washed with 5-10 mL sterile 1× PBS (without Ca and Mg) and added to the conical. Cells were treated with Accutase, transferred to the same conical, and centrifuged for 5 min at 190 rcf. Leaving approximately 0.5 mL of medium, the cell pellet was gently resuspended and treated with approximately 5 mL of prewarmed (37°C) 0.075M KCl swelling buffer (volume of hypotonic solution was dependent upon the size of the cell pellet), drop-by-drop, while agitating gently. Cells were incubated in a 37°C water bath for 25 min. 4-5 drops of freshly prepared Carnoy’s fixative (methanol: glacial acetic acid; 3:1 per volume basis (v/v)) were added to stop the reaction. Cells were centrifuged for 5 min at 275 rcf, and 5 mL of freshly prepared fixative was added gradually, drop-by-drop, while gently agitating the tube and mixing well by flicking the tube so no clumps of cells remain. Centrifugation and resuspension in fresh fixative were repeated twice. Microscope slides and coverslips were washed in soapy water, rinsed 3 times with Milli-Q water before use, kept in Milli-Q water at 4°C for 1 hour, and dried well right before use. Slides and coverslips were humidified at 42°C in the oven on wet paper towels for 2-3 min right before spreads preparation. Metaphase chromosomes were prepared by placing 50-150 μL (depending upon the size of the initial cell pellet) of preparation dropped onto clean glass from a height of 60 cm. Coverslips were dried overnight on parafilm in a humidified chamber (150 mm dish on top of wet filter paper). The next day, metaphase preparations were checked for the quality of chromosome spreads using 4′, 6-diamidino-2-phenylindole (DAPI) staining (1000 ng/mL). Imaging was performed at RT with MetaMorph software (RRID:SCR_002368) using a Zeiss Axioplan II with oil immersion 100× (NA 1.4) objective and a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophore was excited and detected using a DAPI filter set (387/440). Images were analyzed with the FIJI ImageJ software (RRID:SCR_002285). Prior to FISH, coverslips were incubated at 37°C overnight.

3.3.2. Dual-color FISH labeling of chromosome 10 and chromosome 12 centromeres

6-well plates with 18 mm or 22 mm square glass coverslips were coated with 1.5 mL of Matrigel and used within one week from coating. FISH experiments for trisomy 12 takeover calculations were performed on coverslips made every 3 days alongside regular cell passaging using cells coming from the corresponding flasks. Human iPSCs were plated as a single cell suspension to assure monolayer cell growth and the confluence matching the corresponding flasks after 72 hours of growth.

Oligonucleotide DNA FISH probes designed to label alpha satellite repeats on centromeres of chromosome 10 (locus D10Z1, region 10p11.1 - q11.1) in green (Cat. #LPE 010G) and chromosome 12 (locus D12Z3, region 12p11.1-q11.1) in red (Cat. #LPE 012R), as well as hybridization buffer were purchased from the Cytocell (OGT Sysmex Group Company). Probes were delivered as 15 μL concentrated solutions. FISH protocol was adjusted based on the official Cytocell protocol available online (https://www.ogt.com/media/za0brc0w/lpe-r-g-v012-00.pdf).

Day 1:

Cells were grown as colonies and before fixation were rinsed with 1× DPBS (Cat. #14190144, Gibco, Thermo Fisher Scientific). Coverslips were fixed for 15 min in Carnoy’s solution on an orbital shaker, fixative was aspirated, and coverslips were air-dried. Coverslips were rehydrated in 2× concentrated saline sodium citrate (SSC) for 2 min, dehydrated in a series of ethanol dilutions (70%, 80%, and 100%, 2 min each), and left to air-dry. Dual-color probe mix was prepared by adding 0.80 μL of each concentrated probe to 5.90 μL of hybridization buffer (7.5 μL total per 18mm coverslip). Microscope slides, coverslips, and probes solution were pre-heated on a 37°C hotplate for 5 min. Cells were then sealed with rubber glue cement with premixed probes solution in hybridization buffer, denatured on a 75°C hotplate for 2 min, and placed in a humidified lightproof container at 37°C for approximately 20 h.

Day 2:

Coverslips were washed in 0.4X SSC (pH 7.0) by immersion for 2 min at 0.4× SSC wash (pH=7.0) at 72°C and 30 s in 2× SSC/0.05% Tween 20 at RT. Coverslips were air-dried, stained with DAPI (200 ng/mL) for 5-10 min, mounted using Vectashield mounting media (Vector Laboratories, Cat. #H-1000-10 or #H-1700) and sealed with clear nail polish. Slides were stored in the dark at −20°C until image capture.

Images were taken with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II microscope with 63× (NA 1.4) and 100× (NA 1.4) oil immersion objectives, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440), FITC for chromosome 10 centromeres (485/525), and Cy3 for chromosome 12 centromeres (560/607). Additional images were acquired on a Nikon TiE microscope with a 40× (NA 0.6) air objective, using a DS-Qi2 camera (Nikon Instruments). Fluorophores were excited and detected using filter sets: DAPI (392/447), FITC/GFP (466/525), and Cy3 (554/609). Representative images shown in Fig. 1A-B were taken with ZEN software (RRID:SCR_013672) on a Zeiss LSM 880 confocal microscope with a 63× (NA 1.4) oil immersion objective, and laser lines 405 nm (DAPI), 488 nm (FITC), and 561 nm (Cy3/Texas Red). All imaging was performed at RT. Images were processed using FIJI ImageJ software (RRID:SCR_002285). Dozens of positions were chosen at random to ensure hundreds of nuclei for trisomy 12 ratio calculations. Imaging assured the height of the z-stack capturing the entire nuclei of interphase and mitotic cells. The imaged loci were D10Z1 (chromosome 10 centromeres) labeled with fluorescein isothiocyanate and D12Z3 (chromosome 12 centromeres) labeled with Texas Red. Z-stacks were taken every 200-400 nm. Before the probes were used on fixed cell colonies, the localization of FISH probes to the centromeres was confirmed on metaphase preparations.

Trisomy 12 calculations:

Trisomy 12 ratio in iPSCs was calculated based on FISH performed on fixed cell colonies from multiple passages. Chromosome 10 and chromosome 12 signals were counted in hundreds of nuclei per passage. Centromeric signals for each chromosome to define trisomy 12 percentage were identified from maximal z projections of the green (chromosome 10) and red (chromosome 12) channels, followed by ImageJ’s find maxima selection. Choosing points as an output type and the level of prominence, points’ selection represents point counts of chromosome 10 and chromosome 12 in each microscope field. ~100% cell population diploid for chromosome 12 resulted in an average chromosome 12 to chromosome 10 ratio of ~1.0. ~100% trisomic cell population for chromosome 12 resulted in a ratio of chromosome 12 to chromosome 10 of ~1.5. Counts for red and green signals within a microscope field were calculated on an unbiased spot-basis manner before each image was merged with its corresponding DNA stain. Chromosome 10 was used as a control due to its comparable size and gene density with chromosome 12 (Chromosome Map - Genes and Disease - NCBI Bookshelf). Furthermore, there is no mention of whole chromosomal gains or losses of chromosome 10 in human iPS or ES cells (Akutsu et al., 2022; Henry et al., 2019; Mitalipova et al., 2005; Na et al., 2014; Rajamani et al., 2014; Assou et al., 2020; Vaz et al., 2021; Yamamoto et al., 2022), with only rare instances of 10p loss observed (Baker et al., 2016). Consequently, the number of cells in each microscope field was determined by dividing the total count of chromosome 10 centromeres by two. In all FISH experiments, chromosome 10 labeling consistently indicated a diploid cell population. We acknowledge that the ratio calculations might be subject to slight errors due to imperfect labeling of the centromeric probes. Variations in the intensities of chromosome 10 or chromosome 12 centromeres labeling could potentially shift the ratio. To minimize these possible errors, we analyzed hundreds of cells from different microscope fields per passage.

Fluorescence intensity:

The integrated fluorescence intensity of chromosome 12 signals in micronuclei, mitotic, and interphase cells was measured from z-stack images and calculated relative to the background fluorescence. Images were first converted to maximum projections and then to sums with MetaMorph software (RRID:SCR_002368). F.I. was quantified by selecting regions of interest (ROIs) centered on centromeric signals (diameter ~30 pixels, corresponding to ~7.4 μm²), and background regions of larger size (diameter ~40 pixels, ~13.1 μm²) in adjacent areas. Corrected F.I. was calculated using the formula: F.I._corrected=I_signal−(A_signal×((I_background−I_signal)/(A_background−A_signal))), where: I_signal: integrated F.I. of the centromeric signal ROI; A_signal: area of the centromeric ROI; I_background: integrated intensity of the background ROI, and A_background: area of the background ROI. The F.I. of two chromosome 12 centromeres adjoined in metaphase cells was normalized to a value of 2. Chromosome 12 F.I. in micronuclei, other mitotic and interphase cells was normalized to the F.I. of chromosome 12 centromeres in metaphase cells from the same sample. F.I. measurements and comparison were assessed only for images with the same z-stack distance and exposure time, imaged during the same imaging session to maintain uniformity of fluorescence intensity measurements.

Measurement of bridging arms in anaphase cells:

To measure the length of bridging chromosome 12 arms, single z-slice images were analyzed after FISH labeling with centromeric probes and DNA staining. For each chromosome 12 bridge, the distance from the mid-centromeric region to the most distal visible end of the chromatin arm was measured using the straight-line tool in FIJI ImageJ software (RRID:SCR_002285). Even though chromosomes are expected to stretch during bridging, to conservatively evaluate the origin of bridging arms, the total length of bridged arms was compared to twice the measured 12q arm length from metaphase spreads. Only bridges with clearly identifiable centromeric signals, DNA labeling, and straight chromosome arms were included in the analysis.

3.3.3. Whole chromosome 12 labeling

FISH labeling of the whole chromosome 12 was performed on metaphase spreads and on cells grown in colonies prepared as described in 3.3.1. and 3.3.2. Probes designed to label the whole chromosome 12 with Rhodamine were purchased from ASI (Cat. # FPRPR0167, Applied Spectral Imaging Inc.) and delivered as a ready-to-use complete paintbox set. FISH protocol was adjusted based on the delivered ASI protocol.

Day 1:

Coverslips were rehydrated in 2X SSC at RT for 2 min and then dehydrated in ethanol series: 70%, 80% and 100%, 2 min, and air-dried. Then they were placed in the denaturation solution (70% formamide/2× SSC, pH = 7) at 72°C for 1.5 min. Coverslips were immediately placed in ice-cold 70%, RT 80%, and RT 100% ethanol, 2 min each, and air-dried. 7.5 μL of the probe mixture per 18 mm coverslip was denatured by incubation at 80°C in a water bath for 7 minutes and then at 37°C for 10 min. Chromosome preparations were sealed with rubber glue cement with 7.5 μL of probe solution, denatured on a 74°C hotplate for 4 min, and placed in a humidified lightproof container at 37°C for approximately 20 h.

Day 2:

Excess probe was removed by washing for 4 min in 0.4× SSC stringent wash (pH=7.0) at 74°C and for 2 min in 4× SSC/0.1% Tween 20 at RT. Coverslips were stained with DAPI (1000 ng/mL) for 5-10 min, and mounted using Vectashield mounting media (Vector Laboratories, Cat. #H-1000-10 or #H-1700) and then sealed with clear nail polish. Slides were stored in the dark at −20°C until image capture.

Imaging on chromosome 12 was performed with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II microscope with a 100× (NA 1.4) oil immersion objective at RT, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Images were processed using FIJI ImageJ software (RRID:SCR_002285). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440) and Cy3 for chromosome 12 (560/607). Z-stacks were taken every 200-400 nm.

Measurement of chromosome 12 arm lengths in metaphase spreads:

Chromosome 12p and 12q arms were measured from diploid and trisomic metaphase spreads after whole chromosome 12 FISH labeling. Measurements were performed on single-plane images merged with DAPI staining by tracing a straight line from the mid-centromeric region to the distal end of each arm using FIJI ImageJ software (RRID:SCR_002285). Although the whole chromosome was labeled, the 12p and 12q arms could be readily distinguished by their relative sizes, with 12p being noticeably shorter than 12q. To minimize hypercondensation across metaphase chromosomes caused by colcemid treatment, we arrested the cells for only 2.5 hours and analyzed large numbers of each chromosome 12 arms (>60 arms per group of diploid and trisomic metaphase spreads). These measurements served as a reference for comparisons with the bridging arms observed during anaphase.

3.3.4. Labeling subtelomeres on p and q arms of chromosomes 12

Oligonucleotide DNA FISH probes designed to label in red (Cy3 spectra) subtelomeres on p (Cat. #LPT12PR) and q (Cat. #LPT12QR) arms of chromosome 12, as well as hybridization buffer, were purchased from Cytocell (OGT Sysmex Group Company). Cytocell’s subtelomere probes label the most distal region of chromosome 12-specific DNA. Beyond this unique sequence is only the region of telomere-associated repeat, followed by the cap of tandemly repeated (TTAGGG)_n sequence. FISH was performed on metaphase chromosome spreads prepared as described in 3.3.1. The FISH protocol was performed as the official Cytocell’s protocol available online: https://www.ogt.com/media/zc5bqjcb/lpt-r-g-v012-00.pdf. Imaging was performed as described in 3.3.3.

Assessment of subtelomeric repeat loss:

To evaluate the presence or loss of subtelomeric repeats, the integrated fluorescence intensity (F.I.) of labeled chromosome 12p and 12q subtelomeric regions was analyzed in metaphase chromosome spreads from z-stack images with MetaMorph software (RRID:SCR_002368). Only separated 12p and 12q arms with distinguishable subtelomeric signals were used for F.I. quantification. For each chromosome arm, the subtelomeric F.I. was calculated relative to the background. Z-stacks were first converted to maximum projections and then to sums. F.I. was quantified by selecting regions of interest (ROIs) centered on subtelomeric signals (diameter ~20 pixels, corresponding to ~1.3 μm²), and background regions of larger size (diameter ~30 pixels, ~2.9 μm²) in adjacent areas. Corrected F.I. was calculated using the following formula: F.I._corrected=I_signal−(A_signal×((I_background−I_signal)/(A_background−A_signal))), where: I_signal: integrated F.I. of the centromeric signal ROI; A_signal: area of the centromeric ROI; I_background: integrated intensity of the background ROI, and A_background: area of the background ROI. Fluorescence intensities of 12p or 12q subtelomeric signals were normalized within each metaphase preparation to 1. Subtelomeric regions were classified as present if their normalized F.I. was ≥ 0.2 and as missing if F.I. was < 0.2. All imaging was performed under identical conditions, with the same z-stack intervals and exposure settings across metaphase preparations to ensure consistency in F.I. measurement and classification.

3.4. EdU pulse labeling

Human iPSCs from early transition passages 13 and 14 (20-30% of trisomy 12) were seeded on 22 mm coverslips in 6-well plates. Cells were initially cultured for 24 hours in media supplemented with ROCK inhibitor. The following day, the media was replaced with regular mTeSR Plus, and cells were allowed to grow until day 3. On day 3, cells were pulsed with 20 μM EdU (5-ethynyl-2’-deoxyuridine, Life Technologies, Cat. #A10044), a nucleoside analog of thymidine, for varying time intervals (10, 15, 20, and 30 min). After EdU incorporation, the media was removed, and cells were washed with 1× PBS. Cells were fixed in 1.5% paraformaldehyde (PFA)/1× PHEM buffer (60 mM PIPES, 25 mM HEPES, 10 mM EGTA, and 4 mM MgCl₂) for 15 minutes, rinsed with 1× PHEM buffer, and permeabilized with 1× PHEM buffer containing 0.5% Triton X-100 for 8 min. Subsequently, cells were washed with 1× PBS. EdU was detected using click chemistry. Coverslips were incubated with the reaction mixture containing 2 mM CuSO₄, 50 mM ascorbic acid, and 2 mM Alexa Fluor 555 Azide (Life Technologies, Cat. #A20012) for 30 min at RT in the dark. The incubation step was repeated twice. Following the staining, coverslips were washed twice with 1× PBS for 10 min each. Coverslips were stained with DAPI (500 ng/mL) for 10 min, mounted using Vectashield mounting medium (Vector Laboratories, Cat. #H-1000-10 or #H-1700), and sealed with clear nail polish. Prepared slides were stored in the dark at −20°C until imaging. Images were acquired with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II microscope with a 100× (NA 1.4) oil immersion objective at RT, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440) and Cy3 for EdU (560/607). Z-stacks were taken every 200-400 nm. Image processing and analysis were performed using FIJI ImageJ software (RRID:SCR_002285). EdU foci presence in primary nuclei and their micronuclei were assessed. All mitotic cells were EdU negative, and at each time point of EdU treatment, approximately 60% of labeled cells were in the S phase, aligning with literature data regarding cell cycle timing in human iPSCs (Weissbein et al., 2014; Parrotta et al., 2017; Liu et al., 2019; Kapinas et al., 2013).

3.5. GFP visualization of lamin B1 in fixed LMNB1-GFP cells

Human iPS cell line AICS-0013, which has an mEGFP tag on the N-terminus of the lamin B1 gene, was cultured from passage 1 to 16. Cells from passages 12-14 (corresponding to the early transition passages in the AICS-0012 cell line) were used to assess lamin B1 expression in micronuclei of human iPSCs. Cells were grown on 22 mm coverslips until they reached approximately 80% confluence. Fixation was performed using 1.5% paraformaldehyde (PFA) in 1X PHEM buffer, followed by extraction using 1X PHEM, 0.5% Triton X-100, and 1:1000 protease inhibitors. After fixation for 15 min, cells were incubated in the extraction buffer for 8 min. Post-extraction, coverslips were rinsed in 1 × PHEM, stained with DAPI (500 ng/mL) for 10 min, mounted using Vectashield mounting medium (Vector Laboratories, Cat. #H-1000-10 or #H-1700), and sealed with clear nail polish. Prepared slides were stored at −20°C in the dark until imaging. Imaging was performed with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II microscope with 63× (NA 1.4) and 100× (NA 1.4) oil immersion objectives at RT, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440) and FITC for LMNB1 (485/525). Z-stacks were taken every 200-400 nm. Image processing and analysis were performed using FIJI ImageJ software (RRID:SCR_002285). To assess lamina integrity, 8-bit stacks of the DAPI and FITC channels were processed using 2D deconvolution with the nearest neighbors algorithm in MetaMorph software (RRID:SCR_002368). Deconvolved images were combined to generate 3D projections using the 4D Viewer MetaMorph plugin, enabling qualitative assessment of lamin B1 enclosure around chromatin. 3D projections were examined to evaluate the presence or absence of a continuous lamin B1 signal encasing micronuclear DNA.

3.6. γ-H2AX immunolabeling

Human iPSCs from late transition passages were grown on 18 mm coverslips. Cells were pre-permeabilized in ice-cold 1X PHEM/0.5% Triton X-100/1:1000 PI for 5 min and washed twice with icecold 1X PBS for 5 min each. The cells were then fixed in freshly dissolved 1.5% paraformaldehyde (PFA) in 1X PHEM for 15 min at RT. Following fixation, the cells were permeabilized in 0.2% Triton X-100 in ice-cold 1× PBS for 10 min. Blocking was performed using 20% BNGS (boiled normal goat serum) in 1× MBST (MOPS-buffered saline with 0.05% Tween 20) for 30 min. Coverslips were incubated with primary antibody (monoclonal mouse anti-phospho-histone H2A.X (Ser139)) (Millipore, Cat# 05-636-I, RRID:AB_2755003) diluted 1:1000 in 5% BNGS/1× MBST, for 2.5 hours. After primary antibody incubation, coverslips were washed twice with 1× MBST for 5 min each and then incubated with secondary Cy3 goat anti-mouse antibody (Jackson ImmunoResearch, Cat. #115-165-146, RRID: AB_2338690) diluted 1:800 in 5% BNGS/1× MBST for 1 hour in a dark, humidified chamber. Following secondary antibody incubation, coverslips were washed twice with 1× MBST for 5 min each. The DNA was labeled with DAPI (500 ng/mL) for 10 min, mounted using Vectashield mounting medium (Vector Laboratories, Cat. #H-1000-10 or #H-1700), and sealed with clear nail polish. A negative control was processed identically to the test samples, omitting primary antibodies. γ-H2AX presence was analyzed in both primary nuclei and micronuclei relative to the negative control, which was used to standardize image scaling. Prepared slides were stored at −20°C in the dark until imaging. Imaging was performed with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II microscope with 63× (NA 1.4) and 100× (NA 1.4) oil immersion objectives at RT, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440) and Cy3 for γ-H2AX (560/607). Z-stacks were taken every 200-400 nm. Image processing and analysis were performed using FIJI ImageJ software (RRID:SCR_002285).

3.7. Micronuclei area measurements

To determine the sizes of micronuclei, z-stack images were converted to the maximum projected areas. Micronuclei displayed an elliptical shape, and their areas were calculated using the formula for an ellipse: area [μm²] = π × (length/2) × (width/2). Micronuclei sizes were evaluated in human iPSCs across three experimental protocols, as detailed in sections 3.3.2, 3.5, and 3.6. Thus, the primary nuclei sizes were measured first, with 100 nuclei analyzed per protocol, across at least three independent samples for each condition. Nuclei sizes obtained following Carnoy’s fixation (methanol: glacial acetic acid; 3:1 (v/v)) in the dual-color FISH protocol (section 3.3.2) were found to be larger than those observed following γ-H2AX immunolabeling protocol (section 3.6) and LMNB1-GFP fixation (section 3.5). To account for this variation, nuclei sizes measured from the dual-color FISH protocol were normalized to a value of 1. Correction factors were subsequently applied to the nuclei sizes obtained from γ-H2AX immunolabeling and GFP-LMNB1 visualization to align with the FISH-derived measurements. These correction factors were then applied to the micronuclei area measurements for both γ-H2AX-labeled and LMNB1-GFP expressing cells.

3.8. Microwell growth assay

The microwell growth assay was used to determine the doubling times of chromosome 12 diploid and trisomic cell populations. SYBR Gold nucleic acid stain (10,000× in DMSO, Cat. #S11494, Invitrogen) was used to measure fluorescence intensity based on the DNA content (Prole et al., 2020; Schreier et al., 2019).

Cell seeding and preparation:

Cells were cultured in 96-well black-walled plates with a clear, flat bottom (Cat. #655090, Greiner Bio-One). The wells were pre-treated with Matrigel (100 μL/well) the day before seeding. Matrigel was then removed, and cells were seeded in a 100 μL cell suspension per well with media containing ROCK inhibitor for the first 24 hours. Negative control wells pre-treated with Matrigel and containing only media (with and without ROCK inhibitor) were included on each plate in eight technical replicates to account for background fluorescence.

Standard growth curves:

To establish a correlation between cell counts and fluorescence intensity, cells from diploid and trisomic populations were incubated for 3-4 hours to allow attachment. The media was removed, and cells were fixed and stained with 1% PFA, 1× PHEM, 1:1000 SYBR Gold, and 0.5% Triton X-100 for 30 min at RT. Fluorescence intensity was measured using a microplate reader (Tecan Phenix GENios) with consistent gain settings to ensure comparability across experiments. Measurements were taken from the bottom of the wells, using excitation and emission wavelengths of 485 nm and 535 nm, respectively. Diploid and trisomic cell populations were examined in three biological replicates, with varying cell counts (800, 1600, 3200, 6400, 12800, 26500, 51200, and 102400 cells) seeded in eight technical replicates per count. Fluorescence intensities were then used to generate linear equations from power trendlines for each population.

Doubling times calculation:

The linear equations for both cell populations were used to calculate the end-point cell counts during the growth assays. Two types of microwell growth assays were conducted on diploid and trisomic passages. Time-course growth assay: Five different cell counts (ranging between 400 to 2000 cells) were seeded in four technical replicates per count, and end-point fluorescence was measured after 12, 24, 49, 74, and 96 hours of growth. Single time-point growth assay: Varying cell counts (ranging between 500 to 4000 cells) were seeded per microwell in four technical replicates, and end-point fluorescence was measured after 70 hours. End-point cell counts were calculated using the linear equations from standard growth curves. Doubling times were determined using the known seeded cell counts, calculated end-point cell counts, and growth duration. All doubling time data were analyzed using the ROUT method (Q = 5%) to identify and exclude outliers. Final doubling times were calculated from the cleaned dataset after removing identified outliers.

3.9. Predictive mathematical modeling

A mathematical model created using Microsoft Excel predicted fractions of cells within a culture containing a mixed population of diploid and trisomic cells, exhibiting unique characteristics. The model allows for manipulation of key parameters, including cell doubling times, death rates, and the rate of conversion from diploid to trisomic cells. By adjusting these parameters, the model can predict the theoretical fractions of diploid and trisomic cells within the culture at any given time, depending on the input characteristics. In the model incorporating differences in doubling times, the division rate of diploid cells is normalized to 1, while the division rate of trisomic cells is 1.09, reflecting their faster growth calculated as the ratio between diploid to trisomic cell doubling times. Additionally, the initial percentage of diploid cells in the culture is an adjustable input. The model includes a variable for the conversion rate of diploid cells to trisomic cells. In the initial theoretical scenario where trisomic cells overtake diploid cells purely by growth rate difference, this conversion rate was set to 0%. However, in a revised model that accounts for the spontaneous conversion of diploid cells to trisomic cells, this variable was set to 1%, meaning that with every cell division, 1% of diploid cells convert to trisomic cells due to chromosome 12-specific mis-segregation error. Microsoft Excel was used to create stacked area graphs to visually represent the predicted growth dynamics of both cell populations over time, illustrating how varying the input parameters affects the outcome of the culture’s cellular composition.

Loss of trisomy 12 or restoration of diploidy through reverse mis-segregation events was not incorporated into the current model. This decision was based on experimental data from fully trisomic passages, in which chromosome 12 mis-segregation events such as 4:2 anaphase configurations, expected to yield one diploid and one tetrasomic daughter cell, were never observed. However, we acknowledge that the fully trisomic population does not reach 100%, and a small fraction of diploid cells (<4%) remains even in the late passages. This raises the possibility of rare reversion events via non-mitotic mechanisms. We plan to explore this hypothesis in future experiments and, if supported, revise the model accordingly.

3.10. Quantitative analysis of trisomy 12 conversion events using single colonies

To verify the rate of diploid-to-trisomic conversion per cell division, we analyzed single colonies in transition passage 17 (~6% trisomy 12), following FISH labeling (as described in section 3.3.2). To focus on de novo conversion events, colonies containing >8 trisomic cells were excluded, as they most likely originated from trisomic cells present in the seeded cell suspension. Colonies with 1–8 trisomic cells were used to estimate how many cell divisions (doubling times) had elapsed since the first trisomic cell emerged. Because the cells were cultured for ~72 hours after plating, corresponding to 4 doublings, colony composition could be retrospectively linked to the timing of simultaneous conversion events generating the first trisomic cells in those colonies. Specifically, colonies containing 5–8 trisomic cells were assumed to have originated from a single trisomic cell that appeared 3 doublings ago. Colonies with 3–4 trisomic cells were assumed to originate from a trisomic cell that appeared 2 doublings ago, those with 2 trisomic cells from a cell that appeared 1 doubling ago, and colonies with a single trisomic cell from a parent that appeared within the most recent division. To account for the total diploid population at each time point, we included diploid cell counts from fully diploid colonies, under the assumption that only diploid cells can undergo conversion to trisomy 12. This allowed us to calculate the frequency of de novo trisomic events against the total number of diploid cells. Based on this analysis, we estimated an average diploid-to-trisomic conversion rate of 1.46%. This approach could only be applied in early transition passage, where the low prevalence of trisomic cells enabled tracking of single-cell conversion events. In later passages, widespread trisomy 12 made it difficult to distinguish between de novo and pre-existing trisomic cells. Imaging was performed with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II with a 40× (NA 0.6) air objective at RT, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440) and Cy3 for chromosome 12 (560/607). Z-stacks were taken every 400 nm. Image processing and analysis were performed using FIJI ImageJ software (RRID:SCR_002285).

3.11. Induction of DNA replication stress with hydroxyurea

Hydroxyurea (HU) (Sigma Millipore, Cat. #80056–742) dose optimization was performed to determine the concentration that does not impair cell growth or colony morphology. Human iPSCs were treated daily with HU at concentrations ranging from 25 μΜ to 2 mM, and cell counts were recorded at each passage across 3 consecutive passages for all conditions where cells remained viable. Concentrations exceeding 50 μΜ resulted in reduced cell counts and morphological abnormalities. To test the long-term effects of HU treatment, three biological replicates of AICS-0012 cells from passage 7 (confirmed diploid for chromosomes 12 and 10) were divided into two groups (controls and HU-treated). Three were cultured under standard conditions, and three received daily treatment with 50 μΜ HU for 12 passages (36 days in the cell culture), from passage 8 to 20. Cells were maintained in mTeSR Plus medium and passaged every 3 days, as described in section 3.2. Trisomy 12 levels were assessed at passages 8 and 20 in all replicates using FISH, following the protocol in section 3.3.2. Chromosome 10 was not labeled in this assay, as prior validation in eight independent replicates confirmed consistent diploidy of chromosome 10 throughout trisomy 12 emergence. Chromosome 12 centromere signals were analyzed in conjunction with DAPI- labeled nuclei, and total nuclei counts were used to calculate the trisomy 12 percentages. Chromosome 12 signal quantification was performed using FIJI ImageJ’s (RRID:SCR_002285) ‘find maxima’ selection as described in section 3.3.2. Images were taken with MetaMorph software (RRID:SCR_002368) on a Zeiss Axioplan II with a 40× (NA 0.6) air objective at RT, using a Hamamatsu ORCA II camera (Hamamatsu Photonics). Fluorophores were excited and detected using filter sets: DAPI for DNA (387/440) and Cy3 for chromosome 12 (560/607). Z-stacks were taken every 500 nm.

3.12. Statistical Analysis

Statistical analysis (Mann-Whitney, t tests, tests for normality as indicated in figure legends), and outliers’ identification were performed using GraphPad Prism software version 10.2.3 (RRID:SCR_002798). Star values represent the level of statistical significance, with p-values as follows: * = p < 0.05, ** = p < 0.01, *** = p < 0.001, and **** = p < 0.0001.

Data availability:

All data are available in the main text or the supplementary materials. Other materials generated in this manuscript are available from the corresponding authors upon request.