Repair and misrepair of telomeric DNA in dynamic interactions with PML nuclear bodies and lamin B1. A gear-wheel nuclear traffic

1. Introduction

Accelerated cell senescence (ACS) imposed by oncogenes, oxidative stress and drug treatments on cancer cells [1] is characterised by interruption of proliferation, telomere attrition [2,3], degradation of lamin B1 [4,5,6], activation of transposable elements [7] and the transition by mitotic slippage (MS) to reversible polyploidy, as well [8,9,10].

Paradoxically, ACS also serves and even is needed for reprogramming [11], cancer cell stemness [12], and drug-resistant cell survival [13,14,15]. Usually, telomeres are maintained in epithelial tumours by telomerase (reverse transcription mechanism), however 10-15% of tumours, mostly mesenchymal, support telomere length by alternative mechanism based on telomere recombination aided by promyelocyte leukaemia PML bodies, both mechanisms can coexist [16]. In our previous article [10], on triple-negative breast cancer MDA MB 231-doxorubicin-treated (MDA-DOX) cells underwent mitotic slippage (MS) revealing senescence and DNA damage in most cells lasting for three weeks (Figure 1 A-D). During MS resulting in polyploidisation, the cells cast off telomere ends, shelterin TRF2 (Figure 1A,B), telomerase and were transiently undergoing the alternative telomere lengthening (ALT) in PML bodies (APB) [10]. Nevertheless, three weeks later, the telomerase activity was restituted in the clones of the recovered depolyploidised survivors, resuming mitotic divisions (Figure 1C). Interestingly, in this interim ALT-period, we found by qPCR upregulation of several proteins of meiotic prophase, particularly high of the DNA recombination-related nuclease SPO11 (Figure 1.E). As well, induction of recombinase DMC1 and Mos-kinase were revealed [10]. The strong dependence of ALT on the other two genes used for meiotic homology search Hop2-Mnd1 along with the directional telomere movement in all examined ALT cells was also earlier reported by Cho et al [17] hinting at the mutual mechanisms of ALT with meiotic prophase [18] for search and alignment of homologous chromosomes [19]. Assuming all this data and because the meiosis started by the recognition of the chromosome-specific subtelomere homologue sequences [20] particularly favoured by polyploidy [21], we suggested that meiotic recombination proteins and inverted meiosis, based on the pairing of subtelomeric sequences [20] may be involved in the repair of the damaged DNA in cancer ALT as schematized on Figure 1F. [10].

With these assumptions, we added to our immunofluorescence set, the antibodies for SPO11 and DMC1 (see Table in Methods). To generalise our findings, we performed experiments on this triple-negative breast cancer (MDA-DOX) and BRAF-CDk4-mutant melanoma cell line SkMel28 [22] (SkMel-DOX). Both used cell lines are also TP53 mutant, metastatic and aneuploid (para-triploid and para-tetraploid, correspondingly). We applied mild and sublethal concentrations of DOX-24-h treatment. As shown below in the Results, we could confirm the association of SPO11, as well as of RAD51 and DMC1 with APB after DOX-treatment of both cell lines, at the mild DOX100nM concentration. At sublethal DOX concentrations, we also found a formation of thread PML structures (apparently, PMLII isoform known for senescence [23,24] with affinity to the nuclear envelope [25] and characteristic of laminopathic progeria [26]). PML threads were particularly often seen in SkMel-DOX. We investigated by epifluorescent and confocal microscopy the topological relationship between telomere-enclosing thready PML structures in their dynamic interaction with lamin B1 in the progressively senescing cells and the capacity to repair DNA. The results, analysed together with the literature and own previous data on the lamin B1-associated the most peripheral chromatin layer (epichromatin) and its mobile derivatives – nuclear envelope limited chromatin sheets (ELCS) enriched in telomere sequences, ALU retrotransposons and PML, brought us to the hypothesis of the novel mechanism safeguarding the genome functionality by the concerted activity of PML and ELCS traffic with participation of meiotic proteins.

2. Materials and Methods

2.1. Cell Lines and Treatment

The triple-negative breast adenocarcinoma MDA-MB-231 cell line was obtained from the ECACC (European Collection of Authentic Cell Cultures, Wiltshire, UK). The human melanoma SK-MEL-28 cell line (mtB-RAF V600E, mtTP53) was obtained from the ATCC (The American Type Culture Collection, Manassas, VA, USA). Cells were cultured in flasks in Dulbecco’s modified Eagle’s media (DMEM) supplemented with 10% foetal bovine serum (FBS; Sigma-Aldrich, St. Louis, MO, USA) at 37 °C in a 5% CO2 humidified incubator without antibiotics. For experimental studies, cells were maintained in the log phase of growth and treated at 60-80% confluence with DOX (doxorubicin, D1515, Sigma-Aldrich, St. Louis, MO, USA) for 24 h at several concentrations (100, 250, 500, 1000 nM) for SK-MEL-28 cell line and 100nM for MDA-MB-231 cell line. After drug removal, cells were maintained by replenishing the culture medium every 2–3 days and sampled over 3 weeks post-treatment until the appearance of escape clones. In some experiments, cells were grown on chamber slides.

To determine growth kinetics, cells were seeded and treated with DOX at a density of 100,000 cells per well in a 6-well plate and counts were performed using a Neubauer camera (Heinz Herenz Medizinalbedarf GmbH Hamburg, Germany) and Trypan blue dye (0.4%) exclusion.

2.2. Immunofluorescence

Immunofluorescence staining was performed as described earlier [10]. Primary antibodies and their sources are listed in Table 1.

For microscopic observations, a fluorescence light microscope (Leitz Ergolux L03-10, Leica, Wetzlar, Germany) equipped with a colour video camera (Sony DXC 390P, Sony, Tokyo, Japan) and laser scanning confocal microscope (LEICA TCS SP8, Wetzlar, Germany) were used. To capture epifluorescence images, in addition to separate optical filters, a three-band BRG (blue, red, green) optical filter (Leica, Wetzlar, Germany) was applied.

2.3. Fluorescence in Situ Hybridization (FISH)

Telomere FISH for Telo PNA Cy3/Cen#2 FITC was performed with a peptide nucleic acid (PNA) telomere probe (Dako Inc., Glostrup, Denmark) in conjunction with a differentially coloured centromere 2 PNA probe (a gift from Dako, Inc., Glostrup, Denmark) as an internal reference point. FISH was applied following the procedure previously described [10,27].

2.4. [3H]-thymidine autoradiography

DNA replicative activity in S-phase was determined in pulse experiments on irradiated (10 Gy) Namalwa lymphoma cell line by adding fresh [3H]-thymidine (Amersham) to a final concentration of 5 Ci/ml to the culture medium for 60 min. For more detail, see [28].

3. Results

3.1. The topological relationship of the components of ALT and lamin B1 in the time course post-DOX action. DNA repair and misrepair

Non-treated breast cancer MDA MB 231 and melanoma SkMel28 cell lines in immunofluorescent staining of the used antibodies express some amount of meiotic recombination nuclease SPO11 and meiotic recombinase DMC1 particles, with affinity of SPO11 to mitotic telomeres; the cells are also Mos-kinase positive, particularly polyploid as earlier reported for these cell lines [10,29].

The spatial interactions between the marker of telomere end shelterin TRF2, PML body protein, γH2AX marker of DSBs, SPO11, RAD51, DMC1, and lamin B1 structures were examined by confocal microscopy in non-treated (NT) cells and after 24h-DOX treatment on days 4-9. In general, the topology and interaction of these nuclear components in response to DOX can be divided into five phases.

Phase I-II (juxta-position of ALT components).

Phase I: Co-parallel orientation of the long sparse arrays of the small PML-prebodies (0.2<1 µm), tiny TRF2 loci, and very small multiple SPO11 foci, with occasional colocalization, were seen in ~30-40 % of non-treated melanoma and breast cancer cells and in about 60-65% of DOX250 nM-treated cells on day 4-5 after 24h treatment; this pattern was more typical for multilobular polyploid nuclei (Figure 2A).

Phase II (Figure 2B insert and Figure 3): short arrays of more approached and increasingly colocalising components of maturing ALT-PML (APB) nuclear bodies (NB) were observed: PML, TRF2, γH2AX, SPO11 with RAD51 and also DMC1 recombinase (Figure 2B, boxed and Figure 3A-C). The arrays and clusters of telomere DNA sequence indirectly labelled by TRF2 were also confirmed by Telo-FISH (Figure 3D; for control, see insertion on Figure 1B).

3.1.2. Phase III (maturation of recombinogenic APBs)

Next, these short arrays formed mature ~ 1-2 µm round APB bodies including TRF2 foci, SPO11, γH2AX, RAD51 as found in 11-12% of cells (DOX 250 nM), mostly ~ 8C, on days 4-9 post-DOX (Figure 2B-D and Figure 3E-H). It was noted that the topological patterning of SPO11 and TRF2 foci with PML bodies, either in short arrays or colocalising in mature APB, were very similar, where PML bodies interacting with either of them was often seen in doublets, as schematised in Figure 3, I.

3.1.3. Phase IV: the formation of PML II isoform fibrillar structures

PML protein has seven isoforms, of them only one PMLII with a highly disordered Exon 7b forms the fibrillar structures as characteristic for senescing cells [25] and laminopathy progeria [26]. Here, PML fibrils constituting long ribbons were observed after sublethal doses 250 and 500nM DOX, more often in melanoma than MDA-DOX cells. It appears that the long PML tracks were formed by dimeric rods of PML NBs joined by TRF2-positive foci (Figure 4A). Concurrently, PML dimeric segments were dashed by yH2AX links (Figure 4B, small arrows), which could indicate to the non-homologous end-joining of DNA double-strand break-containing telomere repeats sticking in the PML tracks. These long strands of PML tended to form tangles converging to one knot (Figure 4A) and also rosettes around large clusters of yH2AX (Figure 4B, big arrows) or were seen as one or two concentric rings with small loops circumventing the nuclear periphery, supposing rotation around it (Figure 5A, B). Peripheral tracks of PML colocalized with the peripheral heterochromatin (Figure 5C, arrows); in addition, an inner network of PML tracks could also be seen in the continuity with peripheral threads. Their small loops containing fragments of lamin B1 (asterisk), sometimes were twisted in doublets (labelled by double asterisks). In these cells, lamin B1 was generally diffusely under-stained supposing its partial degradation (Figure 5C). Similarly, small loops in the concentric rings of Lamin B1 and of heterochromatin were revealed by us earlier in irradiated HeLa cells as described [10,30] and example in Figure 5D. The staining for a-tubulin and highly overexpressed MOS-kinase in the DOX-treated cells occasionally revealed concentric whorls of microtubules around nuclear envelope, interlaced with MOS, suggesting nuclear spinning by Mos-kinase phosphorylating the tubulin (Figure 5E). Mos is capable of organising a monopolar spindle for the meiotic bouquet rotation [10,31]. Moreover, to our surprise, we occasionally encountered in SkMel-DOX500 nM preparations, the tangled figure of PML threads interspersed by TRF2 foci reminding a telomere „bouquet” in meiotic zygotene (Figure 5F). The proteins of synaptonemal complexes (SCP1, SCP2, SCP3) were not studied here but are known to be expressed in cancers and melanoma; SCP3 was reported and previously seen in cancer cells at the nuclear envelope and cytoplasm [29,32].

3.1.4. Thready PML and DNA recombination repair in DOX-treated cells

We wondered if and how thready PML were involved in the recombinative repair of the DNA damage. The counts of cells containing striped structures of PML for the presence of RAD51 foci after using DOX for both cell lines in comparison with cells displaying the APB foci are presented on Figure 6A. They showed that proportion of such Rad51-positive cells with striped PML structures was much lower than the proportion of RAD51-positive cells with typical small and large APB-bodies, still melanoma cells were relatively proficient in that and at 250-500 nM DOX concentrations still survived the damage (Figure 6B). As to the topology of the RAD51 foci in relation to the PML striped in these rare cells, Rad 51 is seen juxta-posed-colocalized with these structures but at the same time the pattern is chaotic (Figure 6C). In some cases, we could find the intermediate states between short tandem arrays of RAD51 foci, rarely colocalized with PML pre-bodies in continuation with short thread-like PML casually colocalized with small foci of RAD51 (Figure 6D), such indefinite patterns were occasionally seen in MDA-DOX. This observation may indicate that the state of short chains in preparation for telomere repair in APB is unstable and may as well not reach its goal - to form recombinative mature APB but turn to the formation of thready PML structures employing NHEJ which might acquire the affinity to lamin B1 starting rotation around nuclear periphery. The factors limiting recombinative maturation of APB in them may be due to the lack of SP-100 and SUMO reported for PML II thready isoform [25]. The rotation of PML/telomere threads around cell nuclei, indirectly observed as presented in Figure 5, can end in stage V, at the brink of cell death.

3.1.5. Phase V, at the brink of catastrophe

In the transition from phase IV to V, with the low diffuse nuclear staining for lamin B1, we found both nucleus-circumventing PML tracks along with their segments dashed in large loops of lamin B1 suprastructures collapsing inside the nucleus (Fig.A). In terminal cells, the lamin B1 thready suprastructures were formed tightly contorting within one or a few very large PML bodies, near or around nucleoli (Figure 7B). Less deeply senescent cells contain large yH2AX patches of unrepaired γH2AX /DNA surrounded or colocalized with the PML rosettes knotted near the intranucleolar invaginations of lamin B1 into the perinucleolar heterochromatin (Figures 4B and 7C, D). The lamin B1 folds invaginate towards multifacet PML structures around clusters of γH2AX formed within the perinucleolar (pericentric) chromatin and are also connected with PML tracks close to the nuclear envelope; these large PML bodies accumulate p62, a reporter of autophagic activity in this material as previously described [10].

In general, our observations correspond to those published by Condemine et al [25,33] on the transfected isoforms of PML: “PML NBs assemble in “rosettes” surrounding DNA centromeres or are distributed in tracks bridging two centromeres”. The APB PML enclose two telomere ends, while the giant PML NBs seen near nucleoli enclose satellite centromeric DNA [34]. Condemine [25] also mentions that PML “tracks can cross the nucleus and were frequently observed filing lamina gaps at the nuclear envelope”. As well, our results are consistent with the recent paper applying superresolution microscopy showing that lamin B1 has a dashed character with γH2AX-clusters filling its gaps [35], coinciding with the earlier study by Olins showing the “partial colocalization of Lamin B1 and lamin B receptor (LBR) with the tendency of intermittent dashed pattern” [36,37].

The schematic, summarising our results on the topological relationship between PML- structures, ALT and nuclear lamin B1 is presented in Figure 8

The observed direct interaction of PML with lamin B1 and peripheral chromatin motivated us further to analyse their cooperation operating with the bulk literature related to the most peripheral heterochromatin and nuclear envelope-limited chromatin sheets (ELCS).

3.2.1. The epichromatin and ELCS rotating around cell nuclei

So, we found the topological relationship of PML thready form associated with the damaged telomeres presumably coupled to the spinning mobility of the degrading nuclear lamin B1. Degradation of lamin B1 is occurring along with telomere attrition [3,38]. Lamin B1 and LBR embedded within the nuclear envelope, with the N-domain of the amphiphilic LBR interacting with heterochromatin, were shown as underlining the most peripheral and densely-packed 30-nm granules of “epichromatin” [36,37]. These DNAse-high salt-resistant chromatin granules were also coined “anchorosomes” by Fais et al [39]. Epichromatin forms flat folds - ELCS, lined by the inner and outer nuclear envelope [40]. They were first discovered by Davies [41] and seen in many animals and tissues. ELCSs are considered pathognomonic for leukaemia and many cancers as such [42]. Epichromatin and ELCS were extensively studied by Donald and Ada Olins in myeloid HL-60 cells differentiated by retinoic acids into multilobular neutrophils and also reviewed [37]. They revealed with collaborators more than 6,000 epichromatin fragments, ~1Kb average length, dispersed along all chromosomes, more densely at the telomeric ends [43], retaining canonical 30-nm chromatin structure, different from all other heterochromatin in cell nuclei [44].

In our EM and IF studies of the heavily γ-irradiated TP53 mutant lymphoma cell lines [45], we described the formation of ELCS correlating with induced polyploidy, nuclear segmentation and micronucleation enriched in the Ku70. The latter was shown by Celli et al [46] as promoting the non-homologous end-joining (NHEJ) of DNA breaks and the fusion of dysfunctional telomeres. In our study of irradiated p53 mutant lymphoma, we described the protrusions of heterochromatin into narrow folds of the nuclear envelope fusing two strands, from each nuclear side, with further separation and reunion of each strand with heterochromatin in another place of cell nucleus (as exampled on Figure 9A) [45]. In similar experiments on HeLa, we found the colocalization of lamin B1, peripheral heterochromatin and α-tubulin along the edges of the lobulating giant cell nuclei suggesting moving of these edges by microtubules (Figure 9 B-D), while in some experiments carried out on irradiated WI-L2-NS lymphoblastoma, the lobulating nuclear contours were found outlined by DMC1-recombinase (Figure 9E). The concentric rings of ELCS in the perinuclear space around nuclear lobes were beautifully demonstrated by A&D Olins [37] and reproduced with permission in Figure 9F, also undoubtedly indicating the nuclear lobulating mobility of ELCS. These images of fixed cells confirm the study of nuclear rotation on living cells by Paddock [47] showing that the interface of nuclear rotation is located either between the two nuclear membranes or in the adjacent cytoplasm.

Furthermore, in another study on intact MCF-7 breast cancer cells, we showed in situ by Acridine orange DNA structural test that differently from A-T-rich large lamin-associated domains (LADs) melting at 82°C, the DNA of epichromatin is melting only at 94°C suggesting that it is assuming a non-canonical hydrophobic DNA A-form revealed by R. Franklin and Gosling [48] along with the more hydrated canonical B-form. We suggested that due to this particular secondary DNA structure, the epichromatin can densely pack and favour the hydrophobic stacking of its supra nucleosome ~30 nm beads in the immediate hydrophobic microenvironment of the nuclear envelope [49]; a scheme of the epichromatin fragments assembly and relationship with LADs is reproduced on Figure 10.

Based on all these observations and the physical properties of the hard-packed epichromatin granules, we suggested that ELCS may be involved in the gear-wheel-like nuclear traffic for the chromosome homology search and rearrangements, enabling DNA repair or autophagic sorting, at the brink of survival [30,50,51]. Recently, Olins and colleagues, using cryo-EM demonstrated a criss-cross topology of the two epichromatin granular strands of ELCS [44] which well fits the supposed gear-wheel mechanism of the nuclear traffic of the two epichromatin globular strands in ELCSs.

D&A Olins worked out a monoclonal antibody HL2-6 for the epichromatin-specific histone-DNA epitope (H2A/H2B/DNA) which was also used by their methodology (double fixation) [52,53] in our lab confirming epichromatin as outlining telomeres in mitosis and underlining nuclear envelope as a thin rim in interphase described by the authors [49]. This antibody was used by Olins, Teif and colleagues for characterisation of epichromatin in chip-Seq studies.

3.2.2. Chip-Seq studies: Epichromatin meets PML and ALU

In chip-Seq studies of epichromatin domains in HL-60 cells, the important characteristics were described: the dominance of G-C over A-T sequences, GpC methylation, super-enrichment with ALU (S- and Y human) species - epichromatin encloses 30% of all nuclear ALU, while itself makes only 4-6% of the nuclear DNA [43,54] , and high enrichment with PML proteins (18-fold) {53] shown in Figure 11A, B.

3.2.3. ELCS are directly composed of telomere repeats

So, by Chip-Seq the epichromatin met with PML and ALU. Moreover, the study of ELCS in mouse brain stem B-cells with typical EM morphology described above, using FISH and IF revealed that the ELCS enclosed heterochromatin is consisting of telomere sequences (proved by Telo-FISH), tethered to lamin B1 by TRF2 [55]. This data do not contradict but, on the contrary, well fit with the affinity of the epichromatin conformation antibody to telomere ends clearly seen by IF in mitosis and in chip-Seq studies [43].

Summarising all this data, we have PML/ALT bodies in continuation with telomere-dashed fibrillar PML II form which interacts with lamin B1 and rotates in cell nuclei like mobile ELCS. This begins to look as a united, dynamic system. It remains to understand how this gear-wheel might be acting.

3.2.4. The traffic of ELCS structures from the analysis of EM pictures

A typical fragment of the thin section of irradiated multilobular Burkitt’s lymphoma nucleus is reproduced in Figure 11C. This fragment showing the nuclear area enclosing the margins of three nuclear lobes (Nu) of a polyploid cell resembles a railway station junction (or a gear-wheel, see for the mechanical similarity Figure 11D) in a quite busy cell. At this site, the multiple ELCS on the upper nuclear lobe, powered by a large mitochondrion, protrude and split their double-strands, which pair with the single strand from another split ELCS or fuse with the heterochromatin from another site (shown with twig and double arrows) acting by the gear-wheel principle. In the bottom nuclear lobe, one strand of a splitted ELCS is forming a nuclear pocket (NP) with cytoplasmic content, such as earlier shown being involved in the local autophagy [45,51]. But in some places ELCSs typically form ‘eyes’, sized 0.3-0.5 ϻm, shortly disconnecting the two tightly apposed rows of epichromatin strands, this should certainly hamper this motion. Their function is unclear. However, we also see here the “double-eyes’ (dash-boxed), which seems topologically allowing a crossover between four strands. Notably, in the immunoelectron study of the PMLII transfected form by [25], besides the lining nuclear envelope, the labelled bubbles of a similar size (~ 0.4 µm) found in the vicinity of the nuclear envelope were described. This similarity in size and location may suggest that PMLII is participating in the formation of these “eye” structures. Moreover, the observation presented in Figure 5A-C shows in our current SkMel-DOX250 experiment similar small PML-positive loops along the PML threads, sometimes as twisted doublets containing the fragments of lamin B1 (Figure 5C). Collectively, these observations suggest that ELCS and their derivatives composed of the epichromatin telomere sequences employ the PMLII isoform interacting with fragments of lamin B1 for the gear-wheel traffic. It is moved by the microtubules as clearly seen in Figure 9B-D, which should be driven from a centrosome (designed Cs on Figure 11D). Therefore, paradoxically, we found occasionally a typical gear-like heterochromatic structure of the densely pulse-labelled (60 min) DNA along with the dashed label of elongated chromatin threads (Figure 11E).

4. Discussion

4.1. The telomere repair and misrepair in accelerated cellular senescence (ACS)

Telomeres have a central role in chromosomal and nuclear stability events. On the other hand, free chromosome ends, dysfunctional telomeres, and interstitial telomeric sequences (ITS) can trigger chromosome rearrangements and are involved in evolutionary adaptations [56]. Accelerated cellular senescence (ACS) induced in TP53 mutant cancer cells by stress and anticancer drugs response also unites two opposite processes: it stops mitotic divisions by MS for one to two weeks but this interval accumulating stressed polyploid cells seems needed for proliferation to be further resumed in cell minority which rescue from DNA damage and provide resistant tumour growth. Besides telomere attrition and decrease of lamin B1, the ACS depth is characterized by progressive DNA demethylation, and activation of retrotransposons [57].

With interruption of mitotic division by MS, the halving of the semi-conservatively replicated chromosomes for the linear transfer of genetic material is also interrupted. We can consider as a hypothesis that this interruption is needed for DNA repair and cell survival by the mechanism, which is not associated with the linear transfer of genetic information. Then, what?

4.2. Telomeres, ALU flipons, and non-linear information transfer

For the alternative solution, we have four hints : (1) It is abundance of SPO11 meiotic nuclease in these NT and much more, in DOX-treated polyploid cancer cells (Figure 1E and Figure 3A, H); (2) the observation of SPO11 identical topology with telomeric shelterin TRF2 in relation to PML pre-bodies and mature NBs; (3) yet unexplained super-enrichment of the epichromatin composed of telomere sequences [55] with ALU-retrotransposons [43,54]; (4) our previous observations on the unscheduled DNA synthesis moving the 90-min selectively BrdU-labelled material from perinucleolar rim to ELCS loops together with fibrillarin, in the embryonal carcinoma induced ASCs which can be interpreted as the involvement of activated retrotransposons in the nuclear traffic [30].

The meiotic recombinative nuclease SPO11 is rather peculiar. Boateng et al [58] showed in mice that significant homolog pairing needs SPO11 (and SUN1 of the nuclear envelope) and is occurring still before the programmed meiotic DNA cleavage by SPO11. Moreover, SPO11 initiates DNA DSB not in random sites and maybe, not only in conventional meiosis. Recent studies showed that SPO11 prefers to cut sequences with similarity to a DNA-bending motif. Double DSB signals induced by SPO11 overlap and correlate with topoisomerase II-binding sites, which points to a role for topological stress and DNA crossings in break formation and suggests a model for the formation of DSBs and double DSBs in which Spo11 traps two DNA strands. Double DSB gaps, formed by SPO11 which make up an estimated 20% of all initiation events, can account for full non-Mendelian gene conversion events [59] which is favoured by polyploidy [60]. The core meiotic genes including those revealed in our study in cancer ALT are also active in the asexual Entamoeba (Spo11, Hop1, Hop2, Mnd1, Mlh1, Mlh2, Pms1, Dmc1, Msh2, Msh4, Msh5, Msh6, Rad50, Rad51 and Rad52) suggesting the possible ancestral traits in human genome [61].

Compared to other mammals, the genomes of primates and particularly humans are enriched with large, interspersed segmental duplications (SDs), repeated in two or more genomic locations, with high levels of sequence identity. These duplications show clustering and up to 10-fold enrichment within pericentromeric and subtelomeric regions [59,62].

Detailed analyses of the sequences of pairwise SD alignments have revealed that SINE/AlU, the most abundant RT class of mobile elements, is significantly enriched at the boundaries of SD pairs and restricted to younger subfamilies (AluY and AluS). The pairwise SD boundaries were shown to be fragile and the preferential sites of double-strand breakage. The fragile human genome sites assume a left-handed zigzag-like Z-DNA form of high energy tension and represent the loci of the high mutation and deletion rates [63]. In particular, interstitial inverted telomere duplications in the mammalian genome are flanked by the fragile ALU sequences [64] typically assuming the DNA structure of so-called Z-flipons [65].

So, SPO11 is affine to and eager of cutting through ALU sequences flanking telomere repeats. Shelterin TRF1 controls common fragile sites containing interstitial telomere repeats [66] and attracts TRF2 and also, potentially PML as schematized in (Figure 12).

These potential actors of ALT juxta-pose along the chromosomes to epichromatin telomere domains and can become mobile by SPO11 cuts through ALU-Z-DNA (occurring, possibly, in replicative stress of polyploidisation) as molecular ensambles, find a pair, attract RAD51, and repair the telomere damage by recombination in mature APBs. The mechanism may be similar to inverted meiosis recombining on subtelomeres [10]. The direct participation of DMC1 in ALT remains unclear. Alternatively, the unrepaired by HR telomere pairs stack into PMLII tracts by NHEJ. This process occurring with active involvements of the potentially mobile ALU-Z flipons is orthogonal to linear transfer of genetic information in conventional cell cycle, which is interrupted (with purpose?) by ACS. By the model of Alan Herbert [65], the flipons (the sites of non-conventional DNA conformations) serve as dissipative structures for binary switches. They expand the repertoire of RNAs compiled from an involved gene. Their action greatly increases the informational capacity of linearly encoded genomes. Likely, just this process is involved in the described events associated with ASC involving telomeres, PML in NBs and thread isoforms, ALU and the lamin B1-lined two-stranded mobile ELCS enclosing them. It is not excluded that the ELCS traffic using the resource from autophagy of sorted micronuclei and ELCS nuclear pockets may return cancers from the brink of cell death [30,51,67].

4.3. ELCS nuclear traffic and chromothripsis

Several points indicate that the discussed issue associating telomere surveillance with the proposed mechanism of the gear-wheel ELCS traffic interacting with a PML isoform system may be related to the karyotype restructuring by chromothripsis (chromoanagenesis). Below, to be short, we just compile the literature data and, highlight in bold the key words which are consistent with this idea.

Chromothripsis events characterised by massive, clustered genomic rearrangements pervasive across 46% of patient cancers have a link with telomere damage and cancer hyperploidy [68,69], and it may be reversible [70]. This punctuated karyotype evolution represents a powerful survival strategy for somatic cells under high levels of stress/selection [71]. Chromothripsis uses non-homologous end-joining, shaping the genomic landscape [70,72]. It is an adaptive survival tool of macroevolution [73].

5. Conclusions

The XXI century started the era of New Biology, dealing with the complexity and the latest evolution of the mammalians and humans [74]. The attempts of most researchers in the XX century to understand cancer with the rules and means of basic molecular biology largely failed [74,75]. However, the tremendous efforts to win against cancer began to paradoxically reveal the Biology of Complexity itself as dealing with dissipative adaptive Systems [76]. While the Weisman’s and Mendelian inheritance laws keep us to the conventional mitosis, meiosis and recombination between homologues, the presence of jumping transposable elements discovered by [77] comprising nearly half of human DNA, segmental duplications, including interstitial telomeres, and the disordered proteins indicate the existence of non-linear regulation and transfer of genetic information in addition to the linear one, allowing survival of complex systems. The story of the two telomere repair pathways in cancer, occurring with shifting between PML isoforms showing the adaptive switch between two pathways – the error-free HR in APBs/PML I and the error-prone NHEJ aided by PML II and lamin B1, performed during accelerated cell senescence and still compatible with survival, just reflects this duality in regulation of complex systems. It appears that meiotic nuclease SPO11 is starting both pathways. We hope that our observations and the hypothesis of the gear-wheel nuclear traffic will stimulate further research.