Mitotic antipairing of homologous chromosomes

Abstract

Chromosome organization is highly dynamic and plays an essential role during cell function. It was recently found that pairs of the homologous chromosomes are continuously separated at mitosis and display a haploid (1n) chromosome set, or “antipairing,” organization in human cells. Here, we provide an introduction to the current knowledge of homologous antipairing in humans and its implications in human disease.

INTRODUCTION:

Advantages of a diploid (2n) organism derived from maternal and paternal genomes and chromosome organization

The human genome consists of a diploid (2n) set of homologous chromosomes, each set of maternal and paternal origin (Levy et al, 2007). The idea of genetic order via conserved chromosomal domains emerged when cell biologists/geneticists first began to study the mechanisms of inheritance (Misteli, 2020; Bickmore, 2013; Cremer and Cremer, 2010; Strickfaden et al, 2010; Haaf and Schmid, 1991). At fertilization, there is a contribution from the maternal genome of the egg and paternal genome of the sperm to form the single-cell zygote (Barresi and Gilbert, 2020). During development, cells will undergo cellular division to increase their numbers and generate diverse cell types, which will faithfully inherit both maternal, and paternal genomes by all daughter cells (Barresi and Gilbert, 2020). This process of cell division, or mitosis, continues in the developing embryo and through adulthood and will repeat itself through many generations (Barresi and Gilbert, 2020).

The molecular mechanisms underlying mitosis must be precise to successfully produce daughter cells with identical genetic information as the parental cell. Indeed, cells perform multiple rounds of mitosis throughout their lifespan, yet, how is a complete and correct delivery of maternal and paternal genetic information ensured to maintain its genetic stability and avoid chromosomal errors that could lead to genetic aberration?

Cell development and differentiation may serve as key determinants of whether chromosomal order is justified (Lomiento et al, 2018; Kuroda et al 2004; Dixon et al, 2015; Cremer et al, 2003; Sehgal et al, 2016). Cell types that tend to divide more often, such as embryonic/neonatal epithelial cells, are statistically more likely to develop genetic aberrations due to mitosis than those that rarely divide. This risk may be mitigated in part by the genetic organization that reduces the chance of these abnormalities and acts as a safeguard to ensure cell divisions proceed with fidelity in ideal circumstances.

Parental genome segregation in the zygote

The single cell zygote to the 8-cell embryo has shown clear genomic domains between parental genomes in mammalian cells (Fig. 1A) (Reichmann et al, 2018, Gondos et al, 1970; Odartchenko and Keneklis, 1973; Brandriff et al, 1991; Brandriff et al, 1992; Mayer et al, 2000; De La Fuente et al, 2015; Payne et al, 2021; Schneider et al, 2021). Mouse oocytes fertilized with 5-bromodeoxyuridine (BrdU), or tritiated thymidine (³H-Tdr) labeled sperm showed topological separation of BrdU, or ³H-Tdr, positive regions representative of the paternal genome, apart from the negative regions in the zygote and up to the 4-cell embryo (Mayer et al, 2000, Odartchenko and Keneklis, 1973). Classic studies using the human/hamster hybrid zygote, and 2-cell embryo have also reported this separation of parental genomes by using human specific DNA fluorescence in situ hybridization (FISH) probes to visualize one complete parental genome set (Brandriff et al, 1991; Brandriff et al, 1992).

Figure 1: Current models for nonrandom chromosome organization.

Different models of global chromosome organization have been proposed for the human interphase cell. Studies have argued against a random model as in A) (Cornforth et al, 2002; Lesko et all, 1995). B) In the parental origin model, maternal and paternal chromosomes are spatially segregated (Reichmann et al, 2018, Gondos et al, 1970; Odartchenko and Keneklis, 1973; Brandriff et al, 1991; Brandriff et al, 1992; Mayer et al, 2000; De La Fuente et al, 2015; Payne et al, 2021; Schneider et al, 2021). C) Size-dependent organization proposes larger chromosomes are positioned along the nuclear periphery, close to the nuclear envelope, and smaller ones in the nuclear interior. (Cremer et al, 2001; Sun et al, 2000; Bolzer et al, 2005; Emmerich et al, 1989; Popp et al, 1990; Fritz et al, 2014) D) Although similar in size, chromosomes 18 and 19 have been shown to distribute based on gene-density. (Boyle et al, 2001; Bridger et al, 2000) E) Chromosome positions have been proposed to change under various physiological conditions that include differentiation and cell type identity. (Bolzer et al, 2005; Cremer et al, 2001; Boyle et al, 2001; Parada et al, 2004; Marella et al, 2009; Mayer et al, 2005; Zeitz et al, 2009; Foster et al, 2005; Alcobia et al, 2000). F) It has been shown that chromosomes positions are more similar in cells with a common progenitor suggesting differentiation may be important for nuclear organization (Cremer et al, 2003).

Maternal and paternal centromeres can also be distinguished among mouse interspecific Mus musculus/Mus spretus hybrid cells and have been used to evaluate nuclear organization of the parental genomes (Wong et al, 1990; Narayanswami et al, 1992; Matsuda et al, 1993). Pericentromeric repeat sequences found in differential quantities between M. spretus and M. musculus chromosomes allows for visualization of each parental set (Mayer et al, 2000; Miyanari et al, 2013). Using DNA FISH, and transcription activator-like effector-mediated genome visualization (TGV) against the differential pericentromeric repeat major and minor satellite sequences, it was found that Mus musculus/Mus spretus zygotes and up to the 8-cell embryo displayed parental genome separation (Mayer et al, 2000, Reichmann et al, 2018). Asymmetric distribution for DNA demethylation, and kinetochore-associated proteins of the parental genomes in the zygote have previously hinted towards a spatial compartmentalization of the individual parental genomes (Mayer et al, 2000; Oswald et al, 2000; Santos et al, 2002; De La Fuente et al, 2015). This topological compartmentalization has been attributed to the presence of dual mitotic spindles surrounding each genome that persists to the 8-cell stage (Reichmann et al, 2018). A dual spindle apparatus has also been reported in bovine embryos by the same group (Schneider et al, 2021). However, it is less clear whether this parental genome separation remains in later developmental stages.

It has been proposed that the parental genome compartmentalization may persist to late development and adulthood (Hua and Mikawa, 2018a; Chaudhuri and Walther, 2003; Chaudhuri and Walther 2013; Weise et al, 2016; Nagele et al, 1995; Koss 1998; Nagele et al, 1998; Nagele et al, 1999). Past cytogenetic assays using metaphase spreads have been used to propose a parental chromosome segregation (Chaudhuri and Walther, 2003; Chaudhuri and Walther, 2013; Weise et al, 2016). However, metaphase spreads lack a precise and systematic reference point to be used to identify the two discrete groupings for each parental chromosome set, and require rupturing of the cellular membranes. In addition, mouse embryonic stem (ES) cells grown in cell culture also lack a defined polarity that can be used to systematically map the parental chromosome sets within a coordinate system (Miyanari et al, 2013; Reichmann et al, 2018).

We have recently found that sex and autosomes segregate into haploid (1n) chromosome sets along a subcellular axis in dividing human neonatal endothelial and fibroblast cells (Fig. 2B) (Hua and Mikawa, 2018a). Individual pairs of homologous chromosomes were found to be continuously separated along the centrosome, or nuclear division, axis during mitosis, thus forming two haploid (1n) chromosome sets (Hua and Mikawa, 2018a). Our experimental tracking of a maternal supernumerary chromosome in a translocation mouse model revealed a tendency of the maternal-derived translocation chromosome to segregate by maternal origin, within the X-containing haploid set, in male embryos along the defined centrosome axis (Hua and Mikawa, 2018a). Our data is suggestive that each of the haploid chromosome sets may be derived by parental origin, although the underlying mechanisms are still unknown. Limitations in precise identification of individual maternal or paternal chromosomes using a defined axial reference system, or a lack of intact spatial chromosome topology from metaphase spreads makes the parental genome separation model still unclear, and warrants further investigation in its maintenance throughout development and adulthood.

Figure 2: Model of the haploid (1n) chromosome set organization in human endothelial cells at mitosis.

A) We have recently shown that a pair of homologous and sex chromosomes segregate to individual nuclear hemispheres forming a haploid (1n) chromosome set (Hua & Mikawa, 2018). A 3D coordinate axis was defined as following: the x-axis is a line crossing both centrosomes (line of nuclear division); the z-axis is along the laser line of the microscope; and the y-axis is perpendicular to both x- and z-axes. B) Top view of a tetraploid mouse C2C12 mouse myoblasts at prometaphase stained for Chromosomes 6 (red), Chromosome 1 (green), Chromosome 3 (magenta), and TO-PRO3 for DNA (blue) from confocal optical sections. Individual homologous chromosomes are spatially segregated from one another.

Radial chromosome organization

Besides parental genome organization present at early developmental stages, numerous studies have reported a nonrandom pattern of radial chromosome organization (Crosetto and Bienko, 2020). Earlier studies using individual centromere or constitutive heterochromatin based FISH found variations in separation distances between homologous and heterologous centromeres that suggested a spatial randomness for individual chromosomes within the interphase nucleus (Fig. 1 B) (Emmerich et al, 1989; Lesko et al, 1995). Experiments using ionizing radiation were also used to measure whether there were specific chromosome interchanges following DNA double strand break/rejoining in primary blood lymphocytes (Cornforth et al, 2002). These experiments did not find any patterns for specific chromosome-chromosome exchanges to occur between all possible combinations of heterologous chromosomes (Cornforth et al, 2002). Thus, information regarding large scale whole chromosome organizational patterns are still unclear.

Radial organization describes the positioning of individual chromosomes from the periphery to the center of the nucleus, and has been typically characterized in interphase cells. The radial organization has been proposed to be based on chromosome base pair size, gene density, or a ratio of their gene density to chromosome size (Fig. 1 C, D) (Heride et al, 2010; Kupper et al, 2007; Cremer et al, 2001; Leitch et al, 1994; Cremer et al, 1982; Bolzer et al, 2005; Foster et al, 2012; Mayer et al, 2005; Kupper et al, 2007; Croft et al, 1999; Hoo and Cramer, 1971; Sun et al, 2000; Boyle et al, 2001; Sajid et al, 2021).

A chromosome size dependent organization is proposed where larger chromosomes are preferentially located along the nuclear periphery in comparison to smaller chromosomes that are found to be located in the nuclear interior (Fig. 1 C) (Cremer et al, 2001; Sun et al, 2000; Bolzer et al, 2005; Emmerich et al, 1989; Popp et al, 1990, Fritz et al, 2014). A size dependent radial positioning of human chromosomes was first proposed for human fibroblast cells at interphase using q-arm telomere FISH probes (Sun et al, 2000). Although, the analysis had limited resolution of the entire chromosome, and lacked information for the remaining 14 chromosomes in the human karyotype, it was corroborated in subsequent studies in fibroblasts that demonstrated a chromosome size dependent organization (Cremer et al, 2001; Bolzer et al, 2005). A specialized Multiplex DNA FISH approach with two rounds of combinatorial labeled hybridization probes using seven differentially labeled nucleotide probes was later developed to allow for simultaneous visualization of all human chromosomes in the fibroblast cell (Bolzer et al, 2005). A comprehensive chromosome map at interphase describing these “chromosome territories” was then generated and is now widely used by many researchers (Bolzer et al, 2005). The study confirmed that in quiescent, non-dividing human fibroblasts with ellipsoid nuclei, larger chromosomes were preferentially positioned along the nuclear periphery, while smaller chromosomes were positioned towards the nuclear center (Bolzer et al, 2005). This result was consistent in earlier studies from the same research group and others that have used centromeric based FISH analysis with limited spatial resolution (Cremer et al, 2001; Emmerich et al, 1989; Popp et al, 1990, Hofers et al 1993).

Alternatively, a gene-density organization was initially proposed based on the spatial location of whole/partial arm chromosome painting of chromosomes 18 and 19 in human lymphoblastic derived cells (Fig. 1 D) (Bridger et al, 2000, Boyle et al, 2001). Chromosomes 18 and 19 have been historically used as a paradigm to understand chromosome organization since they are similar in size but exhibit differences in gene density (Bridger et al, 2000; Boyle et al, 2001). Chromosome 18 is gene poor with approximately 4.4 genes per Mbp, and 76 Mbp in length, while chromosome 19 is gene rich with approximately 26 genes per Mbp, and 63.8 Mbp in length (Nusbaum et al, 2005; Grimwood et al, 2004). However, the quantitative gene density per chromosome may be inaccurate based on the annotated human genome (Kapranov et al, 2002; Pederson, 2004). In these studies, chromosome 18 was found to be located along the nuclear periphery, while chromosome 19 was located in the nuclear center in human lymphoblast and lymphocyte cells that possess spherical nuclei (Boyle et al, 2001; Croft et al, 1999). The study was then extended to include all the human chromosomes, and found that they preferentially followed an organization based on gene density rather than size (Boyle et al, 2001). In addition, the gene-density based organization was also observed in human ES, and hematopoietic stem cells (HSC) (Wiblin et al, 2005; Lomiento et al, 2018). The central location of five gene-rich chromosomes including chromosome 19 have also been confirmed in Hi-C data sets at interphase (Kalhor et al, 2011).

A possible interpretation is that chromosomes 18 and 19 may be exceptions to size dependent organization, as this organization was later found to be conserved in both normal and cancerous cells with differing nuclear morphologies (Cremer et al, 2003). Tumor cells also showed an inverted position for chromosomes 18 and 19 as compared to non-diseased cells (Cremer et al, 2003), although, the Bickmore research group has provided conflicting results for human fibroblast cells that organize based on gene density rather than by chromosome size as previously reported (Boyle et al, 2001). This discrepancy has been attributed to differential experimental fixation and processing steps (Bolzer et al, 2005).

In addition, dividing human fibroblasts have also reported the organization of larger chromosomes to be peripheral, and smaller chromosomes to be within the center/middle of the metaphase plate, with confirmation that gene dense chromosomes were located within the center of the metaphase plate as well (Mosgoller et al, 1991; Hens et al, 1982). Another report noted that chromosome 18 was located in the nuclear interior of fibroblasts at G₀, but returned back to its peripheral location when it re-entered the cell cycle, suggesting that the cell cycle stage may also influence higher order chromosome organization (Bridger et al, 2000). Chromosome re-positioning in human fibroblasts has been proposed to occur in less than 15 minutes (Mehta et al, 2010). These data taken together suggest the non-random radial chromosome organization may be transmitted throughout the cell cycle.

Chromosome polarity based on early/late replicating regions has also been proposed to support that gene poor, G/C-enriched bands localize mostly to the nuclear periphery, while gene rich, R bands localize in the interior (Sadoni et al, 1999). G band DNA has interpreted to represent mid-replicating chromosomes, while R band DNA represents early-replicating chromosomes (Sadoni et al, 1999). Other reports have not found that G-bands are correlated with peripheral positioning, however (Kupper et al, 2007). Regional gene density was also found to influence radial nuclear positioning within chromosomes, as there is an asymmetrical distribution of gene-dense and gene-poor chromatin, adding to the complexity of whole chromosome organization (Kupper et al, 2007).

In the aforementioned studies, the precise position of individual chromosomes were defined differently due to the technical limitations of visualizing decondensed interphase chromatin using chromosome paint probes, thus contributing to the confusion. Staining of the decondensed chromosomes can be diffuse and the signal intensity may be inaccurate to determine the whole chromosome position. There seems to be both a size and gene-density related radial chromosome organization that can vary between non-dividing and dividing cells, and cell types. Cell type differences were similarly shown in non-human primate cells where lymphoblastic cell lines displayed a gene-density based radial organization, while primary fibroblast nuclei showed a size-dependent organization (Neusser et al, 2007). Overall, gene-density based radial organization has been consistently observed in cells with spherical nuclei, such as lymphocytes, and other lymphoblastoid cell lines, while size-based radial organization has been observed in cells with elliptical nuclei, such as fibroblasts and amniotic fluid cells, suggesting nuclear morphology may be another parameter that regulates overall chromosome organization.

It is not surprising that the higher-order chromosome organization can vary among different cell types with varying nuclear shapes (Fig. 1 E). By using 2D/3D DNA FISH analysis for six chromosomes, mouse lymphocyte nuclei were found to display a gene-density radial organization, while mouse ES derived macrophages displayed a size-dependent radial organization (Mayer et al, 2005). Primary mouse cell types, and embryonic stem (ES) cell line derivatives also showed variations in individual chromosomes localization within the nucleus (Parada et al, 2004, Mayer et al, 2005). For example, Chromosome 5 was localized to the interior in liver hepatocyte cells, mid-concentric subnuclear region in lymphocytes, and peripheral regions in lung cells, suggesting that chromosome positions can vary between different cell/tissue types (Parada et al, 2004). Centromeric spatial organization has been shown to be non-random and cell type specific for hematopoietic cells and fibroblasts as well (Alcobia et al, 2000; Kim et al, 2004). Cell type specific chromosome positions were also found for a normal epithelial cell line, as compared to its equivalent metastatic associated breast cancer cell line when employing shortest edge-to-edge distance measurements between six different chromosome pairs (Marella et al, 2009). Other cancers, such as thyroid carcinoma cells and pancreatic ductal adenocarcinoma, have shown alterations of chromosome position including in in vivo tissue sections (Murata et al, 2007; Timme et al. 2011). However, the radial chromosome positions for selected chromosomes was found to be preserved in multiple cell types in porcine ex vivo tissue sections (Foster et al, 2012). Mouse mutant lymphoma cell lines have shown conserved positions for chromosomes involved in translocations that are also present in normal splenocyte cells (Parada et al, 2002).

Chromosome arrangements may be more similar among cell types that share differentiation pathways, suggesting cellular differentiation may play a role in facilitating these positional changes (Fig. 1 F). Human hematopoietic and derived myeloid CD14-precursors did not show changes in chromosomal preferential locations following differentiation (Lomiento et al, 2018). In comparing the positions of Chromosome 18 and 19, lymphocytes and granulocyte-macrophage colony-forming cells demonstrated more similar positions of the two chromosomes as compared to other cell types from non-hematopoietic lineages (Cremer et al, 2003). However, a subsequent report showed that chromosome positions for lymphocytes and macrophages, both of which are hematopoietic derivatives, were varied (Mayer et al, 2005). It was also found that preadipocyte cells had a different distribution than the differentiated adipocyte cells by examining the positions of chromosome 12 and 16 (Kuroda et al, 2004). Analyzing a subset of seven chromosomes also found variations of chromosome positions during keratinocyte differentiation (Sehgal et al, 2016).

It is important to note that studying nuclear architecture in primary, immortalized, and transformed mammalian cells may also produce complexities in the interpretation. Although chromosome size and gene density remains somewhat constant, other variables such as cell cycle stage, differentiation status, cell type, and nuclear shape can be factors that regulate chromosome positioning. Collectively, these studies demonstrate the probabilistic nature of individual chromosome positions among multiple cell types that have thus far been elusive, and lack a defined set of rules to precisely regulate nuclear architecture (Sehgal et al, 2016; Fritz et al, 2014, Zeitz et al, 2009, Parada et al 2003; Leitch et al, 1994; Wollenberg et al, 1982; Mosgoller et al, 1991; Manuelidis 1984; Manuelidis and Borden 1988; Arnoldus et al, 1989; Arnoldus et al, 1991; Marella et al, 2009; Nagele et al, 1999; Keenan et al, 2021).

Significance of homologous chromosomes position/distribution

Previous studies have consistently shown an absence of association between homologous chromosomes in the mammalian cell (Fig. 2 A) (Heride, et al, 2010; Pliss, et al, 2015; Sun et al, 1999; Popp et al, 1990; Emmerich et al, 1989; Vourc’h et al, 1993; Hepperger et al, 2009; Zeitz et al, 2009; Leitch et al, 1994, Cremer et al, 2001, Lesko et al, 1995, Chandley et al 1996, Khalil et al, 2007; Ferguson, 1992; Alcobia et al, 2000; Marella et al, 2009; Nagele et al, 1995; Hua and Mikawa, 2018a), including a low frequency of interactions using multi-omic data sets including Hi-C (Selvaraj et al, 2013). Measurements of the inter-homologous chromosome distances were frequently greater than inter-heterologous chromosome distances (Heride et al, 2010, Ferguson, 1992; Alcobia et al, 2000; Hua and Mikawa, 2018a; Caddle et al, 2007; Mora et al 2006). Homologous chromosome rearrangements have also been shown to be minimal after irradiation in human and mouse cells, supporting their spatial segregation in the nucleus (Boei et al, 2006; Caddle et al, 2007). However, an exception of Sertoli cells have been reported although limited in its six autosomes analysis (Chandley et al, 1996). This absence of interactions between homologs have been reported to occur not only in interphase but during mitosis as well, suggesting an important function for chromosome organization and regulation throughout the complete cell cycle.

Haploid set based segregation of homologous chromosomes

We have recently found a nonrandom haploid (1n) chromosome set organization pattern during mitosis that may persist throughout the cell cycle. The two compartmentalized haploid chromosome sets were proposed to disrupt homologous chromosome pairing during mitosis by positioning individual homologs on either side of the nuclear division axis defined by the centrosomes (Fig. 2 B) (Hua & Mikawa., 2018a). The haploid set, or antipairing, organization is proposed to serve as a structural pattern to mitigate abnormal chromosome pairing, and exchange during cell division through minimizing interactions between homologs to prevent allelic mis-regulation and genetic recombination (Hua & Mikawa, 2018a).

Although necessary during meiosis for chiasma formation, crossing over and recombination, homologous chromosome pairing may be deleterious during mitosis. It can cause structural changes in chromosomes that have been associated with human pathologies (Potapova and Gorbsky, 2017). For example, chiasmata formation between paired homologous chromosomes in mitotic cells has been postulated to be a potential precursor for cancer. It can result in newly homozygous cancer alleles within daughter cells via recombination induced loss of heterozygosity (LOH) (Fig. 3) (Groden, 1990, Atkin, 1992, Faruqi, 1994; Cavanee, et al, 1983, et al, 1999). A common cause of genetic disorder and cancers arise from chromosome misregulation during mitosis, whereby portions of chromosomes or whole chromosomes may not segregate toward the cellular poles (Potapova and Gorbsky, 2017; Levine and Holland 2018). The mechanisms by which cells regulate this segregation have not been fully elucidated.

Figure 3: Schematic representation of a loss of heterozygosity (LOH) because of mitotic recombination that contributed to abnormal somatic pairing.

Heterozygous parent cells carry a wild type (wt, yellow), and mutant (mt, red) allele on two different homologous chromosomes. During mitosis, if there is a loss of antipairing and an abnormal pairing event, mitotic recombination can occur between the homologous chromosomes (Inset). During prophase and up to metaphase, there could be a regional exchange between homologous chromosomes, and resulting daughter cells could have different allelic composition. Mitotic recombination generates genetically different daughter cells than the parental cell. Both daughter cells would be homozygous for the alleles. One daughter cell would have two wt alleles and the other daughter cell would have two mutant alleles which demonstrate a loss of heterozygosity (LOH) event. In retinoblastoma, there is a loss of both alleles that promote tumorigenesis.

Intranuclear location of individual chromosomes vary

Currently, we lack a unifying theory regarding chromosome order within the nucleus in any single organism. Nagele et al. (1995) argued for the presence of an “antiparallel” organization of homologous chromosomes, where the individual homologous chromosomes are not only spatially separated but also positioned opposite to each other with a near 180° angle of separation between them at prometaphase that persists to interphase (Nagele et al, 1995; Nagele et al, 1999). However, it is important to note that the researchers used stringent selection criterion such as a well-defined circular chromosome rosette morphology and correct fluorescent signal numbers that allowed for only a small portion, ~20%, of all of prometaphase rosettes to be analyzed (Nagele et al, 1995). As such, it may not be a clear representation of all prometaphase cells thus reconciling other reports of highly variable proximity patterns between homologs at prometaphase (Bolzer et al, 2005; Allison and Nestor, 1999). This data may have implications in inter-homologous chromosomes positioning, however our current understanding of spatial organization of chromosomes is limited by resolution, and other technical challenges that are difficult to properly analyze and interpret the data.

We found no conserved position of homologous chromosomes within the cell at mitosis, nor that they are antiparallel, but rather they organize into discrete haploid chromosome sets along the centrosome, or nuclear division, axis (Hua and Mikawa, 2018a). Our finding of the variable positions of individual homologous chromosome within the haploid chromosome set is consistent with previous reports of a lack of fixed position/address for individual chromosomes (Cremer et al, 2001; Mayer et al, 2005; Allison et al, 1999).

Our findings do not align with the antiparallel model, although mitotic chromosomes clearly exhibit a pattern of spatial separation, or antipairing, as there was no evidence for fixed relative positions of the homologs or of any particular angular relationship between them aside from their segregation into nuclear hemispheres along the centrosome axis (Hua and Mikawa, 2018a).

Visualization of individual homologous chromosomes

Chromosomal organization and homologous antipairing have been studied in mammalian models using a variety of methods. Many of these classical studies have been performed using standard cytogenetic tools such as conventional banding techniques on metaphase spreads using phase-contrast microscopy and chromosome morphology assays through electron micrographs (Mosgoller et al, 1991; Leitch et al, 1993; Liehr, 2021). These methods have identified abnormal homologous pairing in human pathologies (Lewis et al., 1993a; Lewis et al., 1993b; Atkin & Jackson, 1996; Brown et al., 1994; Atkin, N., 1992). However, a limitation of this approach is loss of the essential 3D spatial information of chromosomes in an intact cell as metaphase spreads are generated by physical bursting of cellular membranes.

Molecular cytogenetics of nucleotide detection using autoradiography, immunofluorescence, fluorescent in situ hybridization (DNA FISH), and whole/partial chromosome painting have provided a higher resolution for understanding the 3D spatial positions of chromosomes within cells (Liehr, 2021; Mayer et al, 2000; Boyle et al, 2011; Lichter et al, 1988). Past studies used centromeric/telomeric, or gene/locus-specific DNA FISH probes, and centromere morphology to visualize homologous chromosome positions in vivo (Sun et al, 1999; Koeman et al., 2008; Lewis et al., 1993a; Lewis et al., 1993b; Atkin & Jackson, 1996; Brown et al., 1994; Atkin, N., 1992; Hofers et al, 1993). These studies interpreted two separate FISH signals to represent discrete homologous positioning or unpairing, and a single FISH signal to represent homologous pairing in a diploid interphase cell (Sun et al, 1999; Koeman et al., 2008; Lewis et al., 1993a; Lewis et al., 1993b; Atkin & Jackson, 1996; Brown et al., 1994; Atkin, N., 1992, Mosgoller et al, 1991; Leitch et al, 1993). High throughput assays using DNA cross-linking and ligation to interpret the spatial relationships between genomic loci have also been used to study the stochastic global chromosome interactions (Kalhor et al, 2011; de Wit E et al, 2012). However, visualization of centromere/telomere positions or other genomic regions, and cross-linked DNA sequencing may not precisely represent the position and orientation of the entire chromosome within the cell. Therefore, to gain a more precise global view of whole chromosome positions in its native in vivo state, we utilized molecular assays using whole chromosome paints to directly visualize individual homologous chromosomes (Hua and Mikawa, 2018b).

Whole chromosome paint probes are sequence-specific DNA oligonucleotides that can be directly applied to cells, and allow for subsequent visualization of the entire chromosome using an optical microscope (Hua and Mikawa, 2018b; Nagele et al, 1995; Bolzer et al, 2005; Heride et al, 2010; Beliveau et al, 2015; Lichter et al, 1988). Mitotic chromosomes generally provide higher intensity of chromosome paint signals due to the condensation of DNA as compared to interphase chromosomes (Hua and Mikawa, 2018b).

Recent advances in high-resolution imaging and 3D data analysis, including live imaging technologies, have allowed for a better understanding of the chromosomal organization. For example, researchers have been able to use the CRISPR/Cas9 system to label and visualize individual chromosomes, thus painting whole chromosomes in a living cell (Zhou et al, 2017). Previously, fluorescent real-time imaging of individual chromosomes using CENP-A/centrin-1 GFP cell lines allowed for an understanding of global chromosome dynamics during the cell cycle but lacked identification of individual homologs (Hua and Mikawa, 2018a; Magidson et al, 2011). Individual homologous chromosomes were recognized using a complementary approach of fixed cell analysis using chromosome painting and subsequent 3D data reconstruction/analysis (Hua and Mikawa, 2018a). Other live imaging methods using fluorescent repressive-operator lac arrays (Straight et al, 1997; Aragón-Alcaide and Strunnikov, 2000; Jain et al, 2012), photobleaching/photoactivation of fluorescently labeled histones/centromeres (Gerlich et al, 2003; Strickfaden et al, 2010; Essers et al, 2005), or transcription activator-like effector (TALEs) technology for hybrid mouse genome identification (Miyanari et al, 2013; Reichmann et al, 2018) also could not visualize specific chromosomes within a cell. In addition, high-throughput chromosome conformation capture (Hi-C), and other DNA sequence-based genomic architecture mapping lack the 3D spatial position information provided by whole chromosome paints in vivo (Han et al, 2020; Tan et al, 2018; Nagano et al, 2017). The new method utilizing CRISPR/Cas9 would allow the study of individual homologous chromosome dynamics throughout the cell cycle. In particular, the spatial arrangement of homologous chromosomes throughout the cell cycle and its functional consequence could be further investigated. Together, these advances in live imaging and fixed cell molecular cytogenetics have allowed for a better understanding of nuclear organization present in the human cell.

Definition of homologous chromosomes pairing/antipairing

Homologous pairing is defined as the end-to-end spatial alignment of homologous chromosomes along their longitudinal axes, where homologous chromosomes are held in close proximity, or in register (Joyce et al, 2016; Apte and Meller, 2011; McKee, 2003). Pairing of homologous chromosomes occurs when both the maternal and paternal homolog are aligned end-to-end, demonstrating inter-homolog chromosome associations (Joyce et al, 2016; Apte and Meller, 2011; McKee, 2003). The specific distance would be expected to be minimal between the homologous chromosomes if they are paired (Hua and Mikawa, 2018a).

In our previous study, we defined homologous chromosome pairing when the minimal distance between the homologs was less than 0.3 μM, the axial resolution limit of our confocal microscope, and the homologous chromosome axes displayed an angular orientation of 0° relative to each (Hua and Mikawa, 2018a). We determined inter-homologous distances by using fluorescent edge-to-edge measurements between the homologous chromosomes (Hua and Mikawa, 2018a). The minimum edge-to-edge, or shortest 3D distance, was calculated between the boundaries of a pair of homologs determined by the two closest chromosome paint signals (Hua and Mikawa, 2018a). Inter-homologous angles were determined by finding the angle between two vectors along the individual chromosomes’ longitudinal axes, or the longest axis of an ellipsoid contained within each homologous chromosome (Hua and Mikawa, 2018a). Therefore, a pairing of homologous chromosomes was present if the distance was ≤0.36 μm, and angular orientation = 0° (Hua and Mikawa, 2018a).

Whereas “pairing” denotes a state of proximity and spatial alignment of homologous chromosomes, “antipairing” denotes a non-random organization of homologous chromosomes with segregation of one from each pair along a defined subcellular, centrosome, axis (Hua and Mikawa, 2018). We have found that homologous chromosomes are segregated into one homologous chromosome per nuclear hemisphere, as defined by the centrosome, or nuclear division, axis, and are not ‘paired” as previously defined (Hua and Mikawa, 2018). Antipairing is the non-random spatial separation of homologous chromosomes within a cell to minimize/prevent homology-induced pairing (Hua and Mikawa, 2018). Although there have been extensive studies for homologous pairing mechanisms within the cell, homologous antipairing has not been well studied or described (Joyce et al, 2016; Apte and Meller, 2011; McKee, 2003).

Here, we discuss antipairing, which we define as the spatial segregation of homologous chromosomes into discrete hemispheres, and absence of spatial juxtaposition of an entire homologous chromosome along with its allelic locations. To note, we will not discuss other types of anti/pairing that may occur in centromere, telomere, and nucleolar organizational regions (NOR).

Past reviews offer different perspectives for how antipairing is coordinately balanced with pairing mechanisms and chromosome dynamics, which have been best demonstrated in Drosophila, and its significance in meiosis and mitosis (Joyce et al, 2016; Apte and Meller, 2011; McKee, 2003). Our review is primarily focused on discussing the antipairing mechanism in the mammalian model, with emphasis for the past evidence which are relevant to human disease.

Mechanisms underlying haploid set based segregation

The mechanisms underlying the haploid set organization have yet to be elucidated. Recent studies have implicated the cytoskeleton as a potential regulator for the haploid set compartmentalization. In the mouse and bovine zygote, two physically discrete mitotic spindles have been visualized and proposed to separate the individual maternal and paternal pronuclei at fertilization, and persists until the 8 cell stage in the mouse (Reichmann et al, 2018; Schneider et al, 2021). Although we did not identify the presence of dual spindles in our neonatal endothelial cells (Hua and Mikawa, 2018), microtubules may regulate the haploid set organization and warrants further investigation. Surprisingly, the biased pronuclear dependent microtubule based spindle assembly was not based on the sperm-derived centrosomes (Schneider et al, 2021).

The centrosomes may also be implicated in the haploid set sequestration. Centrosome contribution is asymmetric in the human zygote, and provided by the spermatozoan (Avidor-Reiss et al, 2019; Meaders and Burgess, 2020). There is an association between the paternal pronuclei and sperm centrosome, which nucleates with microtubules to form the sperm aster in the zygote (Reinsch and Gonczy, 1998). In contrast, the maternal pronuclei lack centrosome associations and microtubule-nucleating activity (Reinsch and Gonczy, 1998). It remains unknown how the centrosomes remain closely associated to the paternal pronuclei during pronuclear migration. Embryos also demonstrated the ability to differentiate between maternal versus paternal centrosomes suggesting there may be differential molecular markers (Wu and Palazzo, 1999). Taken together, this brings up the possibility that the paternal chromosomes may possess a biased association through the sperm-derived centrosomes that persist throughout development to maintain the haploid set spatial segregation.

Past studies have explored and elucidated molecular factors that are involved in antipairing of homologous chromosomes (Joyce et al, 2012; Thatcher 2005, Williams 2007, Gandhi 2012). There is an argument for existence of a complex and intricate balance that the cell maintains between the opposing forces of pairing and antipairing to mitigate and manage the molecular processes of recombination induced LOH, and other chromosomal abnormalities, in addition to homology dependent gene regulation (Faruqi 1994, Wu 1999, Joyce 2012, Joyce 2016). In Drosophila, a high-throughput fluorescent FISH and genome wide RNAi screen provided a candidate list for pairing antagonists, or antipairing factors, that inhibit homology-dependent chromosomal interactions (Joyce et al, 2012). In particular, several molecules were implicated as antipairing candidates including condensin II, HP1a, and the human ORC complex that localized to heterochromatin (Joyce et al, 2012). Condensin II was also shown to sequester repetitive heterochromatic sequences, and participates in higher order establishment of chromosome territories at interphase (Rosin et al, 2018). This is consistent with previous reports that it functions to spatially separate homologous chromosomes and prevent pairing interactions (Joyce et al., 2007, 2012; 2013, Williams et al., 2007; Hartl et al., 2008). This was demonstrated in Drosophila, with the overexpression of Cap-H2, a subunit of condensin II, that lead to dramatic separation of aligned polytene chromosome structures of the salivary gland to promote chromosome disassembly (Hartl et al., 2008; Bauer et 2012,). A model for how this is accomplished is through intrachromosomal folding in the heterochromatic regions that minimizes inter-homologous chromosome interactions (Rosin et al, 2018, Joyce et al, 2012; Bauer et al, 2012). This heterochromatic intrachromosomal folding is proposed to commence once the cell exits mitosis, at the beginning of G1 interphase (Rosin et al, 2018). However, it remains unclear how condensin II, and the other candidates would regulate this process during mitosis when the chromosomes are condensed. Condensin II, HP1a, and the human ORC complex are required for proper compaction of satellite repeat sequences that suggest chromosome compaction may also be involved in antipairing organization (Joyce et al, 2012; Vermaak et al, 2009; Kellum and Alberts, 1995; Prasanth et al, 2010). Taken together, this suggests that the organization of the heterochromatin may play a higher order chromosomal structure role to facilitate antipairing and warrants further investigation.

Tetraploidy

It is also interesting that spontaneously derived tetraploid (4n) human endothelial cells also displayed haploid set organization following endoduplication (Hua and Mikawa, 2018a). Surprisingly, the tetraploid cells segregated their haploid sets into “quadrispheres”, along the centrosome axis (Hua and Mikawa, 2018a). Our preliminary investigations of other tetraploid cells, such as the C2C12 mouse myoblasts cell line (Fig. 2 C), have also indicated the presence of haploid set organization, suggesting that parental origin may not be the only mechanism involved in the haploid genomic organization. This is consistent with past reports proposing segregation of haploid sets in polyploid individuals (Nagele, et al 1998; Weise et al, 2016). Thus, suggesting that the underlying mechanism for the haploid organization is maintained in cell lines, and that a fertilization step may be unnecessary.

Loss of antipairing in tumorigenesis/cancer

Previous reports support that the loss of antipairing mechanisms, when homologous chromosomes pair in somatic cells, are correlated with gene misregulation and are characteristic of cancer (Koeman et al., 2008; Lewis et al., 1993a; Lewis et al., 1993b; Atkin & Jackson, 1996; Brown et al., 1994; Atkin, N., 1992). Renal carcinoma cells, Caki1, have been reported to have an abnormal homologous pairing of chromosome 19 throughout the cell cycle (Hua & Mikawa., 2018a; Koeman et al., 2008). The regions of chromosome 19 that were abnormally paired at interphase, as compared to regions of chromosome 19 that were not paired, showed specific changes of gene expression and misregulation in the Caki1 cells (Koeman et al., 2008). These data suggested that abnormal pairing of homologous chromosome arms deregulated gene expression at these particular paired chromosome 19 regions, with many genes implicated in tumorigenesis and tumor development.

One possible consequence of the loss of antipairing is an increased frequency of abnormal homologous pairing for chromosome 19. The abnormal pairing may perturb higher-order chromatin structure, causing the consequential positioning of cis/trans gene regulatory regions that could impact the transcriptional regulation of multiple genes at interphase. This was demonstrated for the EGLN2 gene located on the homologous paired region of chromosome 19, that was significantly deregulated in Caki1 cells, which led to direct deficits in the oxygen-sensing network of renal oncocytoma (Koeman et al., 2008). Interestingly, the abnormal pairing of chromosome 19 at interphase was also observed during mitosis (Hua & Mikawa., 2018a). These data taken together suggest the antipairing mechanisms are present throughout the entire cell cycle.

Other cancer studies of lymphomas and leukemias have also reported the presence of abnormal somatic pairing. For example, chromosome 15 was found to have a high frequency of pairing in hairy cell leukemia cells, and in benign/neoplastic cells in the lymphopoietic system at interphase, in particular in the centromeric region and the p-arm (Lewis et al.,1993a; Lewis et al.,1993b). It is also important to note that somatic pairing of the acrocentric/NOR chromosomes (Chr. 13, 14, 15, 21, and 22) may be more prevalent in diseased models compared to other chromosomes due to recent evidence of a ribosomal DNA (rDNA) linkage on the NOR chromosomes present for both homologous and heterologous associations (Patopova et al, 2019). These interchromosomal rDNA linkages were found to connect small and large heterologous acrocentric/NOR chromosomes together (Patopova et al, 2019).

Studies of benign and neoplastic prostate tissue also showed pairing of homologous chromosome 17, although the researchers used centromeric specific probes with lower whole chromosome resolution (Brown et al., 1994). Previous research using centromeric probes and C banding techniques on metaphase spreads were not uncommon and have also argued for the presence of abnormal homologous pairing at interphase and metaphase for chromosomes 1, 7, and 10 in follicular and non-Hodgkin’s lymphoma (Atkin and Jackson, 1996; Atkin, 1992). GC-banding performed on metaphase spreads derived from lymphocytes and tumor cells from patients with bladder carcinoma also showed pairing of chromosome 1 (Atkin, 1992). However, using centromere staining to interpret whether whole homologous chromosomes are paired may be limited as the chromosome arms may not necessarily be paired in the cell. It is possible that although centromeres for the individual homologous chromosomes are paired, the chromosome arms are oriented away from each other. In this scenario, the homologous chromosomes would not be in close proximity to pair and undergo potential recombination events. These studies support that homologous pairing in somatic cells may be a recurring characteristic found in certain cancers. These reports also suggest the underlying mechanisms of homologous chromosome antipairing may be preserved throughout the cell cycle.

Mitotic chiasmata/quadriradial configuration and recombination

Not only is a loss of mitotic antipairing, and somatic pairing associated with genomic instability (Hua and Mikawa, 2018a), but also it can lead to unregulated mitotic recombination, which under certain circumstances, can be detrimental to the cell. Although homologous pairing can be independent of recombination (Dernburg et al, 1998; McKim et al, 1998; Wolf et al, 1994), it can lead to rare mitotic recombination events through the initiation of DNA double-strand breaks (DSBs) (Lee et al, 2009).

In addition to cancer, abnormal pairing of homologous chromosomes has been detected in other human diseases. For example, Immunodeficiency-centromeric instability-facial anomalies syndrome (ICF) patients have also shown the presence of somatic pairing in lymphocytes and fibroblasts using centromeric staining (Marschio et al., 1992). The pairing of chromosome 1 and 16 caused chromosome breaks and/or recombination within the chromosomes that led to chromosome instability, with downstream facial dysmorphism (Marschio et al., 1992). Another example is a boy with a common variable immunodeficiency that displayed abnormal pairing of chromosome 1 homologs at confined segments (Hulten, 1978). These confined, paired segments on chromosome 1 (1q12) contributed to increased “branched” chromosome fragments and thus impacted chromosome structure and fragility (Hulten, 1978). The study proposed a correlation between somatic pairing and chromosome structure/recombination events (Hulten, 1978).

Metaphase spreads have also allowed for identification of a cytogenetic abnormality of mitotic chiasmata, or quadriradial chromosome configuration (Therman et al, 1981; Kuhn and Therman, 1986; Faruqi et al, 1994; Chaganti et al, 1974; Schroeder, 1974). Quadriradial chromosome configurations are rarely observed in normal human cells, and can result from complete/partial endoreduplication, when a chromatid or its fragment remains associated with its sister chromatid in anaphase or when there is an insertion of a broken chromosome into a gap formed by a break in the chromatid (Kuhn and Therman, 1986). Interestingly, a case study of a woman with B-cell prolymphocytic leukemia reported abnormal chromosome 16 pairing and the presence of a quadriradial chromosome configuration in metaphase spreads (Ittel et al, 2017). The formation of quadriradials and abnormal pairing of chromosome X were also reported in a patient with Klinefelter’s syndrome (Welter, et al, 1967). The presence of abnormal somatic pairing of homologous chromosomes has been reported in multiple human pathologies, suggesting that when antipairing mechanisms are compromised and abnormal pairing ensues, it could lead to mitotic recombination events. These observations support an instrumental role for mitotic antipairing organization for proper chromosome organization and function, minimizing the occurrence of mitotic recombination events, in addition to gene regulation and expression.

Mitotic chiasmata and crossing over

The previous studies showed correlative information for the loss of antipairing mechanisms in various diseases, however, there is a lack of direct evidence for the consequential mitotic exchange/crossing over that can occur resulting in an abnormal pairing of homologous chromosomes.

Bloom’s Syndrome (BS) is a rare and unique chromosome breakage syndrome that predisposes the individual to cancer. It is characterized by genomic instability with an increased exchange between homologous chromosomes through mitotic crossing over events (Kuhn, 1986). It is caused by a mutation in the BLM gene, which encodes a ReqQ family DNA helicase (Schawalder et al, 2003; Ellis et al, 1995; Ellis et al, 1996; Luo et al, 2000). In patients with BS, an autosomal recessive disorder, one distinctive feature found in lymphocytes are the presence of quadriradials, or mitotic chiasmata, where a pair of homologous chromosomes is paired, and has undergone recombination (German, 1993; German, 1964; Kuhn, 1986; Schroeder, 1974; Chaganti et al, 1974). Quadriradial configurations have also been described in cancer translocation patients and in vitro cell lines following chemical perturbations, irradiation and reflects compromised genomic integrity, and instability (German, J. 1964; Newell, et al, 2008; Owen et al, 2015).

The quadriradial, or mitotic chiasmata, formation in metaphase spreads of BS cells showed abnormal somatic pairing between homologous and heterologous chromosomes (Therman et al, 1981; Kuhn and Therman, 1986; Faruqi et al, 1994; Chaganti et al, 1974; Schroeder, 1974; Owen et al, 2014). The quadriradial, mitotic chiasmata underwent recombination events and was not limited to only homologous chromosome exchange, as there were heteromorphic chromosome bivalents also found by cytogenetic banding techniques where the chromosomes were structurally different and partly homologous to each other (Kuhn, 1985). It is also interesting to note that the quadriradial formation seemed to be non-random among the chromosomes, and predominantly found to be between non-homologous chromosomes (Owen et al, 2014). This result is consistent with homologous chromosomes positioning further away from each other as compared to a heterologous chromosome (Heride et al, 2010; Hua and Mikawa, 2018).

In addition, two clonal lymphoblastic BS cells lines grown in vitro also demonstrated evidence of somatic recombination via homolog chromatid exchange (Groden, 1990), suggesting the quadriradial configuration could undergo mitotic recombination and homologous chromatid exchange. The combinatorial events were spontaneous for the quadriradials, and the loss of heterozygosity (LOH) was identified in BS cells using Giemsa staining techniques (Groden, 1990). These observations are significant because they demonstrate that quadriradials/mitotic chiasmata and subsequent crossing over can result from a loss of antipairing of homologous chromosomes in mitosis, which allows homologous recombination to occur.

BS individuals also show a high incidence of cancer caused by a high rate of homozygosity, or loss of heterozygosity (LOH), resulting from mitotic crossing-over events, which allow the expression of recessive cancer genes (Kuhn 1986; Larocque et al, 2011; 1974 Chaganti; Kuhn et al, 1985). In addition, unequal crossing-over could amplify recessive cancerous genes, and abnormal chromosome structural changes may transfer oncogenes to new positions that lead to abnormal activation (Kuhn 1986; Schawalder et al, 2003). This could result in immunodeficiency to promote malignant growth and tumorigenesis. Metaphase spreads with lymphocytes stained with BrDU, and Giemsa staining showed exchanges between the homologous chromosomes and not between sister chromatids (Chaganti, 1974; Schroeder, 1974; Owen et al, 2014).

Mitotic recombination and loss of heterozygosity (LOH)

Homozygosity, or loss of heterozygosity (LOH), can be achieved by mitotic recombination, resulting in the loss of the homologous wild type chromosome or allele (Fig. 3) (Hagstrom, 1999; Lasko and Cavenee, 1991; Couto, 2011; Morley et al, 1990; Gupta et al, 1997; de Nooij-Van Dalen et al, 2001;Van Sloun et al, 1998; Shao et al; 1999). One particular disease that LOH has been well characterized is Retinoblastoma (Rb) in its sporadic occurrence through mitotic recombination following the inheritance of a heterozygous mutant Rb-1 locus in human tumor tissue (Cavenee et al, 1983; Couto, 2011; Lasko and Cavenee, 1991). Rb is a heritable childhood cancer that can arise from a single somatic event with a cell that carries an inherited germline mutation at the Rb-1 locus (Cavenee et al, 1983; Lasko and Cavenee, 1999; Couto, 2011). Hereditary Rb is transmitted through the germline, and can undergo subsequent sporadic events to become Sporadic Rb, leading to homozygosity at the Rb-1 locus (Couto, 2011). For example, germline inheritance of a heterozygous retinal cell with a mutant Rb-1 locus can undergo a sporadic mitotic recombination event that would result in homozygosity for the mutant allele (LOH) and result in tumorigenesis, where both Rb-1 loci on chromosome 13 is mutant (Lasko and Cavenee, 1999; Couto, 2011). This occurs through loss of the homologous chromosome carrying the wild type Rb-1 allele, followed by reduplication of the mutant chromosome (Fig. 3). Mitotic recombination between the two different homologs generates daughter cells that are homozygous for the mutant Rb-1 allele at both Rb-1 loci (Cavanee, 1985; Lasko and Cavenee, 1999; Couto, 2011).

In addition, these shared LOH mechanisms observed in retinoblastoma have also been observed in other cancers such as osteosarcoma (Hansen 1985), astrocytoma (James 1989), and Wilm’s tumour (Orkin, 1984) suggesting a common mutagenic chromosomal mechanism that is involved in tumorigenesis (Hansen 1985, Dracopoli 1985). Survivors of heritable Rb, with a heterozygous mutant Rb-1 locus, can subsequently develop second primary osteocarcinomas (Hansen et al, 1985; Dryja et al, 1986). This demonstrates a shared pathogenetic mechanism that can unmask and reveal recessive mutations at a locus exerting more than one phenotypic tissue traits/effects. In addition to cancer, somatic recombination and its acquired LOH at various loci, has also been proposed to be a mechanism for the development of Prader Willi syndrome and other mutagenic pathways involving human lymphocytes (Gregory 1991; Morley et al, 1990; Gupta et al, 1997; de Nooij-Van Dlaen et al, 200).

Mitotic recombination has been observed in LOH in mouse somatic cells (Van Sloun et al, 1998; Shao et al; 1999; Potter et al, 1987; Panthier et al, 1990; Sepulveda, 1995). Gene targeting experiments in transgenic mice, where the establishment of modification of host genes in somatic cells is possible through recombination between homologous endogenous and the engineered DNA (Van Sloun et al, 1998; Shao et al; 1999; Stambrook et al, 1996), already demonstrate the occurrence of mitotic recombination. In earlier reports, Potter et al (1987) showed the occurrence of mitotic recombination between homologous chromosome 17 in vitro by quantitative Southern blot analysis in mouse cells. When growing somatic cell lines in vitro, they found spontaneous generation of clones, which were originally heterogeneous in particular loci for the parental cell line, to be homozygous after subsequent growth in culture (Potter et al, 1987). This was also demonstrated in human lymphoblastoid cell lines (de Nooij-van Dalen et a, 1997; de Nooij-van Dalen et a, 1998; Pongsaensook et al, 1997; Yandell et al, 1986), essentially showing the emergence of homozygous cells from wild type heterozygous cells. Mitotic recombinant cells that are mutant homozygous for deleterious alleles can be generated by a heterozygous individual through loss of heterozygosity (LOH). Mitotic recombination and LOH were also demonstrated in murine lymphoid cell lines, where the somatic recombination rate was determined to be approximately 10⁻⁵ per cell/generation (Nelson et al, 1988). Thus, somatic recombination may occur spontaneously and may also have advantages in certain situations and not always be associated with neoplasia.

Mitotic recombination/LOH and impact of disease (Molecular self-correction)

Fanconi anemia (FA) is a genetically heterogeneous disease that provides individuals with an elevated risk for leukemias and other malignancies like acute myelogenous leukemia or other solid tumors (Gross et al, 2002). In FA individuals, there have been reports of elevated levels of homologous chromosome recombination activity that contribute to their genomic instability and cancer susceptibility (Thyagarajan 1997). These studies analysed protein extracts from the FA derived fibroblasts (Thyagarajan 1997). However, it is important to note that in FA individuals, there are also reports of utilizing mitotic recombination advantageously, as a spontaneous self-molecular correction mechanism that generates genotype/revertant mosaicism (Gregory et al, 2001; Gross et al 2002). Although in these studies, the FA patients have a mixture of defective and self-corrected peripheral blood cells (Gross et al, 2002). In some FA patients, and heterozygous BS/FA patients, there have been reports of a restoration mechanism to the wild-type allele using a back mutation, or gene conversion through phenotypic reversion (Gross et al, 2002; Gregory et al, 2001). Genetic instability appears to increase the opportunity for self-cellular correction of the genetic errors in a cell population by genotype reversion, and/or compensation for the original mutation (Gross et al, 2002; Gregory et al, 2001). Somatic cell mosaicism can have an advantageous proliferative capability following self-correction. For example, the daughter cells of the self-corrected precursor cells can gradually replace the defective parent/progenitor cell population (Gross et al, 2002; Gregory et al, 2001). For these FA/BS patients, the corrected parent/progenitor cell population can undergo clonal expansion that may lead to a reversal of the disease phenotype. This new type of understanding needs to be further investigated as genomic instability may not always be detrimental but potentially beneficial in certain situations.

Nevertheless, the question remains for the underlying mechanisms governing the antipairing of homologous chromosomes and how these mechanisms are altered in human pathology. Uncovering the next steps of mitotic antipairing will be useful for understanding the fundamental mechanisms of nuclear organization and its significance in human disease.