Complex chromosomal rearrangements have been found since introduction of cytogenetic techniques. At present, due to development of new molecular-cytogenetic and molecular methods, the nature of CCRs became apparent making possible their classification.

The first documented CCR case was a translocation affecting three chromosomes in a child with mental retardation and associated dysmorphic features (Nuzzo et al., 1968). In 1970, a team of Lund University researchers discovered another translocation involving three, or possibly, four chromosomes and characterised it as a “complex translocation” and “complex rearrangement” (Fredga and Hall, 1970). Subsequently, complex chromosomal translocations were given the definition that currently extends to the term “CCRs”: complex chromosomal translocations involve more than a reciprocal exchange of segments between two chromosomes resulting in multiple derivative chromosomes (Pai et al., 1980). As molecular genetic techniques gained popularity, our understanding of the nature and origins of structural chromosomal abnormalities increased. As a result, the initial definition of CCRs is frequently updated in terms of the number of breakpoints and number of involved chromosomes. At present, CCRs are understood to be structural chromosomal abnormalities that arise as a result of three or more breakpoints in one or more chromosomes, with the exception of inter- and intrachromosomal insertions (Madan, 2013; McGowan-Jordan et al., 2016).

In January 2011, Stephens et al. (2011) published a paper on CCRs in chronic lymphocytic leukaemia. Using paired-end DNA sequencing, they revealed 42 rearrangements affecting chromosome 4 and several rearrangements affecting chromosomes 1, 12, and 15 in the tumour cells of one patient. The detected rearrangements were characterised not only by numerous breakpoints in a relatively short genome region but also by multiple deletions in the almost complete absence of duplications. Subsequently, when studying similar genome alterations in a small-cell lung cancer cell line (SCLC-21H), the authors observed the formation of double minutes from fragments of derivative chromosome 8 (Stephens et al., 2011). Notably, the rearranged chromosomes and the double minutes comprised material from only one of the homologous chromosomes, the other remaining intact. The authors suggested the term “chromothripsis” to describe this phenomenon (from the Greek “chromos” – “chromosome” – and “thripsis” – “shattering” into small fragments) (Stephens et al., 2011).

Importantly, apart from chromothripsis, over the last 7 years two more CCR types have been described: chromoanasynthesis and chromoplexy. The three types of aberrations are covered by the umbrella term “chromoanagenesis” (from the Greek “anagenesis” – “rebirth”), which indicates a structural chromosome reorganisation (Holland and Cleveland, 2012). It is believed, however, that chromothripsis differs from other chromoanagenesis phenomena by the mechanisms of its occurrence and the nature of genetic alterations (Poot, 2018).

The results of whole genome sequencing, followed by mapping reads against a reference genome, lead us to believe that chromothripsis is based on the process of chromosome shattering triggered by double-strand DNA breaks (Stephens et al., 2011). The repair of double-strand breaks in a cell may occur through either a homologous recombination or NHEJ (reviewed in Ceccaldi et al., 2016). NHEJ is believed to be the primary repair mechanism in chromothripsis cases (Stephens et al., 2011). Once the DNA has been repaired through NHEJ, the reassembled chromosome may have errors in the order and orientation of segments. Fragments that do not ligate together with a centromere may be lost during subsequent cell divisions resulting in deletions (Figure 1; MacKinnon and Campbell, 2013). When double-strand breaks occur in two or more chromosomes, chromosome fragments may fuse, forming derivative chromosomes.

In theory, such CCRs may result from either chromosome pulverisation or sequential, independent rearrangements. The Monte-Carlo simulation method, which includes repeated random sampling and is traditionally used in stochastic process research, has established that the chromosome pulverisation model, which implies an absence of duplications, more accurately matches the genome alterations observed in chromothripsis. These data have given rise to an assumption that chromothripsis is the result of a single catastrophic event (Stephens et al., 2011).

The discovery of chromothripsis in the tumour cells of patients with chronic lymphocytic leukaemia was followed by a description of constitutional chromosomal rearrangements that are comparable with chromothripsis by number of breakpoints and breakpoint clustering but have different copy-number profiles. Microarray results have revealed that the karyotype of 17 patients with various developmental problems featured not only deletions but also multiple duplications and triplications, which could not have arisen as a result of NHEJ (Liu et al., 2011). This enabled the authors to hypothesise that such copy number alterations may result from replication and repair errors caused by DNA microhomology (MMBIR, microhomology-mediated break-induced replication; MMIR, microhomology/microsatellite-induced replication) (Payen et al., 2008; Hastings et al., 2009). Since chromothripsis does not fully reflect the characteristics of the observed genome alterations, the authors suggested replacing the term “chromothripsis” with “chromoanasynthesis,” which stands for chromosome reconstitution or chromosome reassortment (Liu et al., 2011).

In their review article on the hypothetical mechanisms and consequences of chromoanagenesis, Holland and Cleveland (2012) contrast the terms “chromoanasynthesis” and “chromothripsis.” According to the authors, chromoanasynthesis and chromothripsis are two independent phenomena with different underlying mechanisms. However, multiple chromosomal aberrations, which are observed in both, are most likely the result of a single catastrophic event, and not a successive series of rearrangements.

By contrast, chromoplexy, the third example of chromoanagenesis, is the result of an accumulation of chromosome rearrangements. The term “chromoplexy” (from the Greek “pleko” – “to weave”) was introduced in 2013 to indicate complex rearrangements of prostate cancer genomes (Baca et al., 2013). To analyse the results of whole-genome sequencing and microarray-based comparative genomic hybridisation (aCGH), the researchers developed the ChainFinder algorithm, which identifies the chained rearrangements that resulted in the CCRs. They demonstrated that, in the majority of samples (50 out of 57), multiple deletions and translocations occurred successively, which is uncharacteristic of either chromothripsis or chromoanasynthesis. Chromoplexy is also characterised by fewer breakpoints and a larger number of rearranged chromosomes (up to eight) compared to chromothripsis (Baca et al., 2013). Importantly, the breakpoints are presumably localised in open chromatin regions (Berger et al., 2011). Therefore, the high transcription level of certain loci may serve as a chromoplexy trigger.

Recent study showed a novel potential mechanism of chromoanagenesis: DNA polymerase 𝜃-dependent alternative homologous end joining (Masset et al., 2016). Thus, chromoanagenesis may be induced by a variety of mechanisms that lead to CCRs. In contrast to chromoplexy and chromoanasynthesis, chromothripsis is characterised by a larger number of breakpoints and a random order and orientation of chromosome segments after reassembly. Chromothripsis features a high frequency of deletions in the almost complete absence of duplications in localised genome regions. However, chromothripsis identification among the multitude of CCRs is challenged by a lack of distinct limitations on the number of breakpoints and other features. The authors suggest six criteria to distinguish chromothripsis from other CCRs (Korbel and Campbell, 2013):

The authors used statistical algorithms to justify some of the criteria, but they did not report the minimal number of breakpoints, and admitted the possibility of partial tri- and tetrasomies (Korbel and Campbell, 2013).

Initially, CCRs with over 50 breakpoints were classified as chromothripsis (Stephens et al., 2011). However, this criterion was not always fulfiled in subsequent works. In a number of cases, rearrangements with 20 (Molenaar et al., 2012), 10 (Northcott et al., 2012; Rausch et al., 2012), or fewer (Chiang et al., 2012) breakpoints were treated as chromothripsis. Kinsella et al. (2014) drew attention to this issue in 2014. Using statistical simulation, they demonstrated that chromothripsis-like rearrangements may result from sequential rearrangement. Importantly, these results do not debunk the traditional hypothesis of the origins of chromothripsis but only emphasise the need for further research.

It has been possible to describe the features of chromothripsis due to mate-pair sequencing and paired-end sequencing. These methods work for structural variant detection and CCRs, as well as genome assembly and de novo sequencing (Miller et al., 2010). In the case of a CCR, mate-pair and paired-end sequencing with subsequent verification by Sanger sequencing not only determine the precise localisation of breakpoints, but also gains data on nucleic acid sequences at breakpoint junctions (Gao and Smith, 2017). In spite of their high cost and challenging methodology, mate-pair sequencing and paired-end sequencing are widely used in chromothripsis studies.

Another efficient method of detecting and studying chromothripsis is microarray-based comparative genomic hybridisation (array CGH, aCGH), which is frequently referred to as “virtual karyotyping” or “chromosomal microarray analysis.” Copy number analysis allows detection of deletions, duplications, and other aberrations as well as identification of their precise genome localisation and size. The resolution of this method is sufficient to detect submicroscopic aberrations. For higher resolution and information capacity, aCGH is combined with a single nucleotide polymorphism (SNP) array (Keren, 2014). As a method, aCGH is not without considerable limitations: it cannot detect balanced structural chromosomal aberrations or determine the order and orientation of derivative chromosome segments (Balajee and Hande, 2018).

For detection and localisation of a specific DNA or RNA sequence on a chromosome or in a cell, fluorescence in situ hybridisation (FISH) is frequently used. In chromothripsis studies, various FISH techniques are used, each of them addressing specific aspects in the identification of the derivative chromosome structure. SKY and multicolour FISH (M-FISH), with the use of whole chromosome probes conjugated with different fluorochromes, enables identification of chromosomes involved in a rearrangement. The multicolour-banding FISH technique (MCB-FISH) is a segment-specific variant of chromosome banding that allows one to determine the structure of an aberrant chromosome (Balajee and Hande, 2018). To map breakpoints on the chromosomes, locus-specific probes with known cytogenetic localisation may be used for FISH. A combination of SKY and in situ hybridisation with fluorescent locus-specific probes is used to determine the precise structure not only of derivative chromosomes but also of double minutes (Stephens et al., 2011).

In patients with hereditary diseases, chromothripsis may be detected by a conventional karyotyping of metaphases from peripheral lymphocytes. This technique allows identification of numerical and structural chromosomal abnormalities including translocations and inversions, which are frequently observed in CCR cases. However, the complex nature of CCRs makes their interpretation by conventional karyotyping alone difficult. Therefore, to precisely determine the structure of rearrangements in chromothripsis, it is necessary to use a complex approach that includes classical chromosome banding, visualisation of the aberrations on metaphase chromosomes by FISH and molecular genetic techniques.

The first assumptions regarding the mechanisms of chromothripsis were made by Stephens et al. (2011). The authors argue that DNA junction sequences and their localisation in the genome attests to chromosome pulverisation during mitosis at the stage of their highest condensation, not at the interphase stage. Today, several presumed causes of chromothripsis are listed (Meyerson and Pellman, 2011; Forment et al., 2012; Jones and Jallepalli, 2012; Maher and Wilson, 2012).

In 2015, ChromothripsisDB¹ database was created (Yang et al., 2016) to categorise cases of chromothripsis in human and model organisms by disease, research method, and criteria that enabled the authors to classify the observed chromosomal abnormalities as chromothripsis. As of March 2018, the database counted 500 chromothripsis cases across 46 cancers. The authors of ChromothripsisDB update it on a regular basis and standardise the information on all the rearrangements that are treated as chromothripsis (Cai, 2018). At present, ChromothripsisDB is the most informative source of information for accessing and comparing the results of chromothripsis studies.

Chromothripsis may also be a constitutional karyotype abnormality caused by chromosome damage in germline cells or preimplantation embryos. Cases of CC are extremely rare and usually coincide with congenital malformations or reproductive failure in the patient (Table 2; Kloosterman et al., 2011;de Pagter et al., 2015). In the virtually complete absence of any genetic imbalance, CC may co-occur with breakage of multiple genes or changes in their expression (Table 2; van Heesch et al., 2014; de Pagter et al., 2015; Bertelsen et al., 2016; Middelkamp et al., 2017). CC may include structural chromosomal abnormalities associated with genetic disorders (Table 2; Fontana et al., 2014; Genesio et al., 2015; Kurtas et al., 2018). In this case, the patient displays symptoms of an inherited disease. However, certain non-specific phenotypical features complicate the diagnosis and prognosis of the clinical outcome of the CC (Table 2).

Constitutional chromothripsis carriers may transmit the rearrangement to the offspring either stably or with de novo events (Gu et al., 2013; Weckselblatt et al., 2015; Bertelsen et al., 2016; Nazaryan-Petersen et al., 2016; Kurtas et al., 2019). Whereas the majority of de novo CC cases result from chromosomal aberrations arising from male gametogenesis (Pellestor et al., 2014), chromothripsis is inherited primarily from the mother (Table 3). To all appearances, it is determined by differences in DNA repair capacity and specific features of spermatogenesis and oogenesis.

Chromothripsis may arise during mitotic and meiotic divisions of spermatogenic cells as well as during spermiogenesis (round spermatid differentiation in spermatogonia) (Pellestor and Gatinois, 2018). Considering that spermatogonia undergo a succession of mitotic divisions, the replication stress may lead to errors during mitosis. Meiotic recombination may also feature double-strand break repair errors (Pellestor and Gatinois, 2018). The DNA breaks in spermatids that occur at the stage of histone-to-protamine transition during spermiogenesis can only be repaired through NHEJ because of the haploid chromosome number in cells at this stage (Gunes et al., 2015). In rodent male germ cells, scaffold/matrix-attached and differentially packaged chromatin regions are highly sensitive to endogenous nucleases, and, thus, to damage (Arpanahi et al., 2009; Grégoire et al., 2013). Accumulation of DNA strand breaks may be also caused by the epigenetically mediated bystander effect in non-irradiated whole testis tissue (Tamminga et al., 2008). This phenomenon is also involved in the production of delayed DNA damage in mouse elongated spermatids due to upregulation of proapoptotic genes 21–33 days later after spermatogonia exposure to X-rays (Cordelli et al., 2012). However, the apoptotic elimination of spermatogenic cells with DNA damage may be aborted before completion (the so-called abortive apoptosis or anoikis), allowing such cells to continue to differentiate and participate in fertilisation (Tubio and Estivill, 2011; Tang et al., 2012). In addition, there are some evidence of aberrant DNA methylation and tissue-specific accumulation of chromosome aberrations in unexposed progeny of cranially irradiated rodents (Koturbash et al., 2006; Tamminga et al., 2008). These data indicate an epigenetic link between DNA damaging agents and occurrence of chromosome aberrations both in unexposed parental germline and offspring’s somatic cells.

In contrast to male germ cells, oocytes may repair breaks through both homologous recombination and NHEJ (Marchetti et al., 2007). Consequently, chromothripsis during oogenesis appears to be less likely than during spermatogenesis. Aberrations in chromosome segregation and premature chromatid separation may cause chromosomal rearrangements during female gametogenesis (Pellestor and Gatinois, 2018). In addition, the DNA repair capacity of an oocyte is the crucial factor of zygote viability, because the repair of maternal and paternal chromosome damage after fertilisation and prior to embryo genome activation occurs through DNA repair factors accumulated in the oocyte cytoplasm.

De novo CC may also be induced by DNA damage during early embryogenesis. Preimplantation embryos typically demonstrate micronuclei formation, blastomere fragmentation, and abnormal mitosis at the cleavage stage (Chavez et al., 2012). This could be a consequence of imperfect repair in germ cells or DNA damage in embryo. In addition, asynchronous pronuclear development and resulting under-replication of the paternal DNA may induce chromosome pulverisation in a zygote (Eichenlaub-Ritter et al., 1995).

Importantly, CCRs are hardly ever detected during conventional karyotyping of chorion cells in a miscarriage, which is conducted starting from 4 to 5 weeks of gestation (i.e., after embryo implantation) (Pendina et al., 2014; Massalska et al., 2017; Soler et al., 2017; Pylyp et al., 2018). As of today, the literature describes only one case of CC in an embryo with multiple malformations (Macera et al., 2015). Apparently, most embryos with CCRs, including chromothripsis, are eliminated at the implantation stage. Despite the wide use of preimplantation genetic testing, the actual frequency and the specific mechanisms of chromothripsis occurrence in gametes and embryos at early stages of development are yet to be determined.

Constitutional chromothripsis is generally characterised by fewer chromosome breaks and almost complete absence of deletions in comparison with malignant tumours (Figure 1; Kloosterman and Cuppen, 2013). A number of studies treat CCR cases with duplications of chromosome regions as chromothripsis (Table 2; Gamba et al., 2015; Wang et al., 2015; Del Rey et al., 2016; Kurtas et al., 2018). It is yet to be established, however, whether such genetic abnormalities in patients are cases of true chromothripsis or variations of other CCRs.

As is the case with any recently discovered phenomenon, the concept of chromothripsis is ambiguous. In our opinion, the most comprehensive definition of chromothripsis has been suggested by Ly and Cleveland: “Chromothripsis is a catastrophic event in which one or a few chromosomes are shattered and stitched back together in random order, producing a derivative chromosome with complex rearrangements within a few cell cycles” (Ly and Cleveland, 2017). Considering that chromothripsis is a highly complex genomic aberration, its reliable detection necessitates the use of a comprehensive approach, combining molecular genetic, molecular cytogenetic, and cytogenetic methods.

Chromothripsis was first detected in chronic lymphocytic leukaemia. To date, it is most frequently found in cancers, even though there are registered cases of chromothripsis both in benign tumours and as constitutional chromosomal abnormality. Both somatic and CC feature multiple rearrangements of one or more chromosomes with a random order and orientation of reassembled fragments, as well as alteration of regions with loss and retention of heterozygosity. However, these aberrations are less pronounced in CC, which normally has fewer breaks and shorter chromosome regions with copy number alterations or a complete absence of such.

The causes and mechanisms underlying chromothripsis remain a subject for discussion. The most probable are telomere damage, exposure to ionising radiation, and viral infections. Along with these already known causes and mechanisms, we suggest aberrant epigenetic regulation as a possible pathway to chromothripsis. The above-mentioned factors may directly destruct chromosomes or activate cell mechanisms associated with chromothripsis. To clearly understand chromothripsis mechanisms, it is necessary to develop models of chromosome pulverisation in micronuclei, reversible apoptosis, and dicentric chromosome breaks.

As of today, it is not clear whether somatic chromothripsis is the cause of tumours or a consequence of pathological processes in tumour cells. Considering that cases of chromothripsis are observed in both malignant and benign tumours, as well as in the karyotype of healthy individuals, it cannot be unambiguously associated with poor clinical outcomes. Apparently, what matters most for neoplasia pathogenesis and a chromothripsis carrier’s phenotype are the genome regions involved in the rearrangement, their localisation, and the size of deleted or amplified fragments – not the presence of chromothripsis itself.

Regardless of the fact that chromothripsis was discovered over 7 years ago, we are still facing challenges in its differentiation from other multiple chromosomal rearrangements and in the understanding of its causes, mechanisms, and consequences – all of which requires further in-depth research.

AK, AP, OE, OC, TK, and VB contributed to the conception, writing, and checking of the manuscript for important intellectual content.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.