3.1. Role of NER in Simple and Complex Chromosomal Rearrangements
The loss of proper nuclear compartmentalization can have drastic consequences in terms of genome evolution. During NER, the unprotected DNA in contact with the cytoplasm can result in DNA breaks
[14][15][16][17] (
Figure 2). Indeed, the use of DNA damage markers such as γH2AX and 53BP1 foci revealed important double strand DNA (ds-DNA) breaks in the constricted nuclei during migration through narrow areas
[14][15][16][17][18]. Importantly, nuclear compression and deformation, even in the absence of NER, leads to increased replication stress in the S/G2 phase and provokes DNA breaks
[17]. Moreover, in an elegant study, Nader and collaborators demonstrated that during NER, TREX1, a cytoplasmic DNA exonuclease that clears normal endogenous cytosolic DNA, can enter and become trapped inside the nucleus after NER repair causing massive DNA breaks
[16]. Thus, the DNA damage observed during nuclear squeezing and NER can have profound consequences on tumor evolution.
Figure 2. Genomic instability associated with nuclear squeezing and nuclear envelope rupture leads to the generation of genomic diversity. Nuclear deformation and nuclear envelope collapse, displaying a transient loss of compartmentalization, lead to replication stress and DNA damage. This DNA damage can initiate a cascade of genomic alterations in the next mitosis with the generation of micronuclei and telomere fusion, provoking a chromatin bridge. Such events are known to initiate a myriad of diverse genomic events such as chromothripsis, extra chromosomal DNA (ecDNA), the hypermutated pattern Kataegis, events of insertions, deletion, and translocation, as well as the mutational signature APOBEC (Adapted from Gauthier BR, et al.
[19]).
Taking advantage of advances in high throughput sequencing and in single cell analysis technologies, a plethora of genomic anomalies have been discovered in cancer cells. Besides the classical and simple genomic aberrations such as mutations, deletions, translocations, and insertions, cancer cells can present complex chromosomal rearrangements such as chromothripsis and kataegis (
Figure 2). Chromothripsis is a chromosomal instability phenomenon where hundreds of chromosomal rearrangements occur during one single event in a localized region of one or few chromosomes. This type of chromosomal rearrangement is highly frequent in cancer with a prevalence of 49% to up to 80%
[20]. Chromothripsis is associated with the formation of circular extrachromosomal DNA (ecDNA)
[21][22] as well as with segmental deletion. Kataegis is a pattern of localized hypermutations occurring in a small region of DNA.
Importantly, all these complex and simple chromosomic events are known to originate from the NER of micronuclei
[22][23][24][25] or during the NER of cells presenting a chromatin bridge during a telomere crisis
[26][27][28] (
Figure 2). Micronuclei are small nuclei found next to the main nucleus in cancer cells that contain a full chromosome or a chromosome fragment, and are the result of aberrant mitosis
[19]. It is noteworthy that the NE of these micronuclei is fragile and tends to disrupt without the possibility of proper repair
[29][30]. On the other hand, the chromatin bridge appears when cells experiencing telomere fusion connect two daughter cells. This implies the generation of additional tension forces affecting the NE during movement, leading to NER that can last up to two minutes
[27].
The mechanisms responsible for the genomic aberrations linked with NER are still under debate. Some authors have demonstrated that in the cytoplasm the unprotected DNA can be attacked by DNAses such as TREX1 or by the immune DNA mutator APOBEC upon NER
[28]. The recent discovery of the nuclear internalization of TREX1 after NER repair reinforces its role in generating DNA damage
[16]. APOBEC that plays a role against retrovirus attack can lead to the Kataegis pattern
[28] or to the mutational signature APOBEC, characterized by an increase in mutations with the substitutions C-to-G and C-to-T
[31]. Additionally, it has been demonstrated that during the chromatin bridge, the mechanical forces generated can trigger the breakage of the chromosome bridge, leading to extensive DNA breaks
[25]. Furthermore, the loss of compartmentalization can affect the replication inside the micronuclei, provoking a desynchronization with the main nucleus. Thus, the main nucleus may start the mitosis too early for the DNA trapped inside the micronuclei, which is not folded nor protected and can lead to its pulverization within the cytoplasm, resulting in chromothripsis
[23][25].
Altogether, these data reveal the profound and hereditary consequences that NER can have on the genome and the creation of genomic diversity. It is not yet known if the NER observed during migration
[32] and passage through tight spaces
[14][15] can also generate drastic complex genomic reorganization. Nevertheless, a recent study using CRISPR-Cas9 gene editing has shown that a single ds-DNA break can lead to a cascade of events resulting in the formation of micronuclei and chromosome bridges
[33]. As such the DNA damage caused by nuclear deformation and NER can be amplified into far more extensive genomic alterations during subsequent mitosis (
Figure 2), leading to a myriad of genomic diversity. Thus, the biophysics behind nuclear squeezing may be a major player in the increase in genomic heterogeneity during metastasis.
3.2. Metastasis and Genomic Evolution
In the past decade, extensive sequencing of numerous patient tumors revealed that cancer genomes are highly complex and heterogeneous. Tumors are composed of several clones with different genomic alterations that compete with each other. Upon specific conditions such as therapy treatment, some of these clones will be favored, thus becoming more prevalent than others (
Figure 3A). In order to identify the major driver events for tumor initiation, the genome of tumor samples of the same patient but from different body locations and at different time points were sequenced to generate a phylogenetic tree that highlights the tumor history. The analysis of these phylogenetic trees revealed that complex structural events, including chromothripsis, are major drivers during the early phase of tumorigenesis. However, these events can also occur in the later phases of the disease
[20][34]. Indeed, the comparative analysis of cancer cells from primary sites and from metastases revealed an enrichment in chromosomal instability in the metastases of several cancer types
[35][36].
Figure 3. Modes of metastatic dissemination from the primary tumor. (A) The primary tumor is composed of a multiclonal population with cells competing between each other. Metastatic tumors are formed by clones found in the primary site, as well as by new independent subclones. (B) The phylogenetic tree shows the history of the tumor evolution. Genomic diversity can arise at all steps of tumor progression. Importantly, the metastatic site can be composed of clonal populations found in the primary tumor site or independent subclones.
These analyses of whole genome data have also demonstrated that metastatic dissemination can be monoclonal or polyclonal. Furthermore, some metastatic clones can have their own subclone evolution through the increase in genome complexity of a metastatic precursor (
Figure 3A,B)
[37]. The comparative analyses of a metastatic tumor versus a primary tumor have indicated the absence of universal metastatic-specific driver alterations exclusive to metastatic disease. It rather shows a continuous evolution of the tumor with increased genome heterogeneity and complexity
[38]. Nevertheless, recent data suggest that certain alterations, also found in primary tumors, are enriched at metastatic sites, revealing the possible existence of drivers specific to metastatic clones (
Figure 3A)
[38][39]. Moreover, distinct patterns of copy number alterations have been observed in metastases from different tumor types, highlighting that a specific gain/loss of chromosomes confers advantages in certain tumor types
[36]. It is still unknown whether these specific drivers confer a better resistance to drugs and/or a better ability to succeed in the metastasis process. Interestingly, the analysis of specific metastatic tropism suggests that some genomic characteristics may be linked with the potential to seed in a specific organ. For example, a gain in semaphorin 4D was shown to support tumor cell transmigration through the blood–brain barrier and MYC has been suggested as a key factor for the adaptation of disseminated tumor cells to the activated brain microenvironment
[40].
4. Nuclear Squeezing and Its Role in Activating the Innate Immune Response cGAS
Besides genomic alterations, NER-derived leakage of DNA into the cytoplasm can also trigger an immune response. The presence of DNA in the cytoplasm can be interpreted as a viral or bacterial attack, and mammalian cells have several mechanisms to detect intrusion and trigger an anti-viral immune response. One of these responses is the activation of the cyclic GMP-AMP synthase (cGAS), a cytosolic DNA sensor that binds to cytosolic ds-DNA and catalyzes the synthesis of the second messenger 2′3′-cyclic-GMP-AMP (cGAMP), which in turn activates STING, eventually leading to the production of several inflammatory factors such as type I interferons, interleukins, and the tumor necrosis factor (
Figure 4A)
[41]. Importantly this pathway is also activated through the release of ds-DNA from replication stress or from mitochondria DNA damage
[41][42].
Figure 4. The activation of cGAS can support metastasis survival. (A) Activation of inflammatory genes through the detection of double strand DNA (ds-DNA) by the enzyme cGAS. Double strand-DNA bound cGAS induces the production of the second messenger cGAMP that in turn activates STING, leading to the transcription of several inflammatory response genes. cGAMP can also be a paracrine signal by being released in the extracellular compartment or transferred to other cells. cGAS pathway is involved in several processes such as alerting the immune cells but is also involved in senescence, autophagy, and surprisingly in favoring metastasis survival. (B) cGAS activation in metastatic cells. (1) cGAMP supports cell own growth as an autocrine signal by the induction of inflammatory genes. (2) To avoid extracellular cGAMP release and activation of immune cell attack, cancer cells express ENPP1 that selectively hydrolyze the extracellular pool of cGAMP. (3) In the context of brain metastasis, cGAMP can transfer to neighbor astrocyte cells by carcinoma–astrocyte gap junctions. This paracrine signal supports the growth of metastatic cells by the astrocytes.
It is noteworthy that the cGAS/STING pathway has important anti-tumorigenic functions, helping in the clearance of genetically unstable cells by alerting the immune cells (
Figure 4A). The secretion of type I interferon favors the establishment of an immune infiltration of T cells
[43] that participate in the clearance of defective cells. The secretion of cGAMP into the extracellular space is also an important signal for the activation of dendritic cells and enhanced cross-presentation of tumor-associated antigens to CD8 T cells
[44] (
Figure 4A). Additionally, the cGAS/STING pathway is also involved in two other barriers against oncogenic transformation by the elimination of pre-cancerous cells through autophagy of cells under crisis
[45], and in favoring cell senescence, a permanent arrest of the cell cycle
[46] (
Figure 5A).
However, recent studies have shown that the cGAS/STING pathway can also be kidnaped by tumor cells to favor tumor progression in metastatic sites. The cGAS/STING pathway can have an autocrine effect by inducing a local inflammation that supports metastatic tumor cell growth
[35], opposite to its anti-tumorigenic action at tumor primary site (
Figure 5B). Using a mouse model, Bakhoum and coworkers showed that highly genetically unstable cancer cells with high chromosomal instability and an activated cGAS/STING pathway are more prone to form metastases than cancer cells with a more stable genome that do not activate the cGAS/STING pathway
[35]. Interestingly, the metastases harboring cancer cells with unstable genomes engage a STING-dependent noncanonical activation of NF-κB and inflammatory responses that favor invasion and metastasis
[35][47] (
Figure 4B).
Thus, it is intriguing to observe that metastatic tumor cells adopt inflammatory signaling and the induction of chronic inflammation while evading the immune attack in the newly seeded site. A recent study combining data from patients and mouse models has demonstrated that the expression of ENPP1 in metastases is a key factor for this outcome
[48]. ENPP1 is an enzyme that can hydrolyze the extracellular cGAMP, preventing its transfer from cancer cells to the microenvironment, thus avoiding its transfer to immune cells
[48]. ENPP1 activity leads to a reduction in immune cell infiltration at the metastatic site. In clinic,
Enpp1 expression has been associated with reduced lymphocytic infiltration in human cancers in accordance with the role of ENPP1 in escaping the immune system
[48] (
Figure 4B).
In addition, at the specific metastatic brain niche, cGAMP can act as a paracrine signal between disseminated cancer cells and their environment. In brain metastases, invasive breast and lung cancer cells establish gap junctions with astrocytes allowing cGAMP transfer. In return, astrocytes activate the innate immune response leading to the secretion of factors that support metastatic growth and chemoresistance
[49] (
Figure 4B). In this particular study, the origin of the cytoplasmic ds-DNA that leads to cGAMP production was not identified, but it is tempting to speculate that NER can be one of the sources. Then, NER associated genomic instability can initiate a paracrine crosstalk, that is often underestimated in the study of metastasis, providing a pro-survival signaling pathway necessary for its growth.
Moreover, the cGAS/STING pathway can also support metastasis by promoting a welcoming tumor microenvironment. cGAS is indispensable for senescence
[50] and initiates the secretion of senescence-associated secretory phenotype (SASP)
[51]. SASP paracrine signaling from cells that failed to form metastasis can mediate several pro-tumorigenic effects, such as promoting the induction of tumor-associated angiogenesis
[46]. Then by inducing senescence in cells failing successful metastasis, the cGAS/STING pathway influences and primes the tumor microenvironment.
To conclude, cGAS/STING pathway activation can have opposite outcomes depending on its location. In primary tumor sites, the cGAS/STING pathway has an anti-tumorigenic action, being a major driver of cancer immunity, while at metastatic sites, this pathway has a pro-survival activity.