Common fragile sites are chromosomal regions of hindered replication which can remain unreplicated by the mitotic entry. These under-replicated chromosomal segments are shared by two chromatids and, if unresolved, can lead to chromosome mis-segregation or DNA breaks converting into focal CNVs or terminal losses (Bjerregaard et al., 2018; Debatisse and Rosselli, 2019; Shaikh et al., 2022). CFSs are involved in formation of cancer-specific and hereditary genomic aberrations (Mitsui et al., 2010; Le Tallec et al., 2013; Y. Li et al., 2020). The repertoire of cFSs is tissue-specific and has not been previously characterized in iPSCs. Given the high level of replication stress in PSCs and frequently observed karyotype instability, the identification of the genomic regions most vulnerable to replication stress is an actual task for developmental biology, genotoxicology and medical applications of iPSCs (Vallabhaneni and Hursh, 2022). Our study provides an insight into conditions of cFS induction in iPSCs and presents the molecular mapping of FRA12L within the ANKS1B gene on Chromosome 12.
Moderate replication stress severely delays mitotic entry in iPSCs
Mild Aphidicolin treatment (up to 6 µM concentration) of lymphocytes and fibroblasts is known to provoke fork speed reduction and moderate replication stress, which tolerates further cell cycling despite the presence of under-replicated DNA and elicits mitotic chromosome breakage (Koundrioukoff et al., 2013). At the same time, mitotic delay occurs through a yet unknown classical checkpoint-independent mechanism, while ATR plays a crucial role in fork stabilization and modest role in control of mitotic entry (Koundrioukoff et al., 2013). Effects of exogenous replication stress on iPSC genomes have been previously studied under 0.1–0.8 µM concentration of Aphidicolin and relatively short incubation time (16–24 hours) (Ahuja et al., 2016; Lamm et al., 2016; Paniza et al., 2020). Interphase markers of replication stress (FANCD2 accumulation, reduced replication fork speed) have been reported, as well as mitotic defects (micronuclei, chromosome decondensation, segregation errors), while no direct evidence of chromosome fragility provided. Our analysis of dynamics of mitosis entry under replication stress suggested that in iPSCs treated with Aphidicolin during the substantial part of the S-phase, mitosis was delayed by the 48th hour. At the 21th and 27th hours of the experiment, metaphases were mostly attributed to the cells treated at the very end of S-phase and at G2-phase, thus no exogenous replication stress could be suggested. Taking into account short cell cycle of iPSCs under unperturbed conditions, we can hypothesize that minor elevation of mitotic markers of genotoxicity observed by us and others can be explained by the prevalence of mitotic cells experiencing endogenous mitotic delay, and hindered mitotic entry by other cell populations at early experimental time points. Under both conditions of replication stress induction used in our study, the level of mitotic chromosome damage in iPSCs greatly exceeded that in lymphocytes when the duration of treatment was prolonged in iPSCs according to cell cycle delay. In contrast, similar or even lower rates of mutation accumulation by cultured iPSCs, have been revealed in multiple reports under normal conditions and genotoxic stress compared to differentiated cells, as well as better resistance to DNA damage (Dannenmann et al., 2016; Rouhani et al., 2016; Kwon et al., 2017). Thus our results provide a clue to interpretation of the Aphidicolin effects reported by other groups and facilitate the design of further studies on replication stress mutagenesis in iPSCs.
Localization of focal EdU signals on metaphase chromosomes indicated that the latest replicating regions in iPSCs were centromeres and MiDAS sites. Recent profiling of replication timing of heterochromatic regions based on the T2T-CHM13 genome assembly has shown that it varies consistently in five differentiated cell lines studied, with centromeric regions being replicated in mid-to-late S-phase, but not unusually late-replicated relative to the rest of the genome (Massey and Koren, 2022). One of the distinctive features of human embryonic cells is their internal nuclear centromere localization (Wiblin et al., 2005). Although mechanisms governing cell type-specific pan-centromeric replication timing are unknown, they can be associated with spatial organization of nuclei (Massey and Koren, 2022). Analysis of centromere composition in pluripotent cells revealed reduced centromeric chromatin size due to weaker recruitment of the centromere-specific histone CENP-A during the shortened G1 phase (Milagre et al., 2020). CENP-A depletion in hTERT RPE-1 cells provoke replication stress and subsequent MiDAS foci at centromeres (Giunta et al., 2021). Thus an intriguing question on the interplay of nuclear location, replication timing and functioning of centromeres in iPSCs emerges, as well as their impact on pluripotent genome instability. In our study, relatively prolonged timings of incubation with EdU were used, therefore, centromeric EdU foci were observed only in cells treated with Aphidicolin. Thus it remains to be defined whether centromeres are the latest replicated regions in untreated iPSCs.
The ANKS1B locus can impact to genome instability in iPSCs
Trisomy 12 accounts for approximately 46% and 30% of the aneuploidy cases in human ESCs and iPSCs, respectively, and impose a selective growth advantage in culture and tumorigenic potential (Ben-David et al., 2014; Poetsch, Strano and Guan, 2022). The recurrently gained 12p arm harbors cell cycle-related genes, specifically, the NANOG gene, which is essential for pluripotency maintenance, DPPA3, GDF3, and KRAS. In contrast, the most active fragile site on Chromosome 12 in our study, FRA12L, was located at the 12q arm. Mechanistically, breaks and ultrafine anaphase bridges provoked by under-replication at FRA12L can promote Chromosome 12 mis-segregation and formation of truncated chromosomes with terminal 12q deletions. The FRA12L breakage frequency of 2.6% of all chromosomal breaks is relatively low compared to other cFSs in iPSCs, ranking approximately 12th most active site in their genome (our unpublished data), however, subsequent rearrangements could be further bolstered by clonal selection. In addition, cFSs are subject to submicroscopic genomic alterations that frequently go beyond karyotype analysis during standard stem cell line characterization. Notably, FRA12B is sensitive to endogenous oncogene-induced replication stress in fibroblast, and potentially can be activated during cell reprogramming. However, our survey of CNVs in iPSCs and ESC published by different groups (Hussein et al., 2011; The International Stem Cell Initiative, 2011; Ruiz et al., 2015; Huang et al., 2020; Yamamoto et al., 2022), did not reveal any CNV located at FRA12L, while other loci of the chromosome were affected. This observation may be explained by low clonality of focal ANKS1B alterations which are not supported by further selection.
The data on somatic and germline genetic alterations at ANKS1B are limited in the literature. The ANKS1B gene is frequently downregulated in a smoking-related cases of clear cell renal cell carcinoma (Eckel-Passow et al., 2014), and disrupted by hepatitis B virus integration in non-Hodgkin lymphoma (M. Li et al., 2020), which are typical for many known cFSs. Non-recurrent CNVs have been reported as secondary events in multiple enchondromas of patients with Ollier disease (Pansuriya et al., 2011). Our analysis of ANKS1B translocations recorded in the DepMap database revealed high prevalence of focal deletions in tumor cell lines. At the same time, ANKS1B is not among known tumor suppressor genes, and according to the evaluation of the homozygous to hemizygous deletion ratio conducted by Bignell et al., this chromosomal region exhibits high mutational rate with low selection pressure in cancer cell lines, which is a distinctive feature of cFSs contrasting them from recessive cancer genes (Bignell et al., 2010). Thus focal ANKS1B rearrangements might be neutral for cell survival in culture and present a passenger marker of experienced replication stress in parental fibroblasts or iPSCs.
Adjacent to the centromeric flank of ANKS1B, there is a pro-apoptotic gene, APAF1 (Apoptotic protease activating factor 1), whose stability can be threatened by the neighboring fragile site. Indeed, two ANKS1B “translocations” listed in the DepMap database disrupt the APAF1 sequence. Loss of heterozygosity and deletions of APAF1 have been reported in glioblastoma, melanoma and testicular germ cell tumors, and have been associated with poor outcome (Bala et al., 2000; Fujimoto et al., 2004; Watanabe et al., 2006; Shakeri, Kheirollahi and Davoodi, 2021). Since downregulation of APAF1 is a marker of more aggressive tumor phenotypes (Shakeri, Kheirollahi and Davoodi, 2021), its loss in iPSCs might promote cell adaptation to culture conditions.
Typically for most cFS genes, ANKS1B is an extremely large (approximately 1.3 Mb) gene encoding for the ANS1B protein (also known as AIDA-1, EB-1, Kakhalin-2 or ANKS2) highly expressed in the developing and adult brain (Younis et al., 2019), which interacts with amyloid-β precursor protein and regulates synaptic plasticity. Rare and de-novo gene variants may lead to an increased risk of psychiatric disorders, late-onset Alzheimer’s disease and age-related cognitive decline (Younis et al., 2019). Patients with copy number alterations exhibit the ANKS1B haploinsufficiency syndrome, a spectrum of neurodevelopmental phenotypes, including autism, attention-deficit hyperactivity disorder, and speech and motor deficits (Carbonell et al., 2019). According to the Database of Genomic Variants, there is a cluster of CNV (losses) within the ANKS1B gene in healthy individuals (MacDonald et al., 2014). The DECIPHER database contains 24 copy number variants of ANKS1B, including 20 (83%) losses of variable inheritance and pathogenicity status, with the prevalence of neuro-psychiatric phenotypes, 17 (85%) of which are focal aberrations (less than 1Mb in size). Thus, mosaic ANKS1B alterations caused by replication stress can potentially influence tissue-specific functions in iPSC-based cell products, as well as contribute to phenotypically relevant somatic mosaicism, which could be most pronounced in neuronal lineages.