Genetic Mosaicism and Cancer: Cause and Effect

Just like nothing has contributed to the flourishing of physics more than war, nothing has stimulated the development of biology more than cancer. The unprecedented intellectual and material efforts invested into combating the ongoing cancer pandemic have greatly enriched our understanding of fundamental processes of life and the organization of a living cell.

With regard to oncology, it has been established that “cancer is a disease of genes,” and that genetic instability is the driving force of carcinogenesis and a key feature of tumor cells (1–3), which is based on the assumption that the genome of a normal cell is generally stable. However, recent evidence contradicts this notion, as it appears that the human body represents a mosaic composed of trillions of genetically distinct cells, to the extent that two identical cells can hardly be found (4, 5). Such amazing genetic diversity can be explained by constant, life-long exposure of human cells to a multitude of mutagens originating both inside the body and in the surrounding environment, which results in somatic mosaicism exemplified in its extreme form in tumors.

In view of this, genetic instability can no longer be considered a unique property of cancer cells but the one inherent to all somatic cells, to some extent, which makes it necessary to revise numerous generally accepted fundamental concepts in oncology. In particular, the phenomenon of genetic mosaicism makes us view carcinogenesis as collective rather than individual “guilt” and put the blame on the whole cellular community rather than on a single cell.

In this Perspective, I discuss stochastic changes occurring only in the genome of somatic cells rather than programmed mosaicism of germ and immune cells.

The term “somatic mosaicism” refers to the presence of two or more genetically diverse cell populations in an organism originating from a single fertilized egg, the zygote. Mosaicism is a natural consequence of generalized and continuous life-long mutagenesis. Random mutations are unavoidable during cell division because of errors in replication, repair, and mitosis. In addition, mutations can be caused by some environmental factors. Both somatic and germline mutations are widely dispersed throughout the genome, appearing at many locations in every chromosome; however, the rate of somatic mutations is almost two orders of magnitude higher than that of germline mutations (6).

Genetic mosaicism is determined by two fundamental parameters: the human diploid genome size of 6 × 10⁹ base pairs and a point mutation rate of approximately 0.5 to 3.0 × 10⁻⁹ per site per generation (6–8). The latter is characteristic for normal human cells with full-fledged repair systems, as it should be understood that DNA replication cannot (and should not) be absolutely error-free because without mutations, there would be no genetic variability and hence, no evolution. When DNA repair systems are affected, as often happens in tumor cells, the mutation rate typically increases manifold, giving rise to the so-called mutator phenotype. Considering the genome size and mutation rate, every cell division can produce several (from 1 to 10) new mutations in the genome of each daughter cell, suggesting that the two new cells are not genetically identical and mosaicism already occurs after the first division of the zygote, which is magnified in subsequent cell cycles. The number of mutations, which distinguishes any two cells, indicates the time since they had a common ancestor (the founder cell), thus reflecting the process of speciation (9). After approximately 40 cell cycles that an average cell has undergone from the zygote stage to birth, a newborn can carry over 120 mutations in each cell (4). Calculations show that by the age of 15 years, an individual has approximately 3.5 × 10¹³ cells (10), each accumulating from 100 to 1,000 point mutations only in the coding genes, which constitute 1% to 2% of the genome (4). Structural rearrangements such as deletions, insertions, and chromosome aberrations, which occur with lesser frequency than point mutations (9) but are more functionally significant, further increase the mutational burden on proliferating cells and the whole organism (4, 5, 11). Cumulatively, these data suggest that all cells in the human body are genetically damaged, albeit to different degrees. Genome-wide association studies suggest that approximately 10% of genome defects are manifested phenotypically (12); further cell fate is determined by fitness selection (neutral, negative, or positive), which results in the disappearance of unsuccessful cells and clonal expansion of successful ones.

Mosaicism is a dynamic process as DNA mutations accumulate during both embryonic and postnatal development. The earlier a mutation appears, the higher is its representation in the body, and if it is potentially oncogenic, the greater is the cancer risk. Mutations in germ cells spread to all cells of the newborn and may be passed on to the progeny, whereas postzygotic somatic mutations are not inherited and disappear from the population with the death of a particular carrier (5). As for the rate of this process, it has been shown that each human hematopoietic stem/progenitor cell (HSPC) from healthy individuals contains mutations accumulating with age at a rate of 0.13 ± 0.02 exonic mutations per year (7). Whole-genome sequencing using organoid technology shows that mutations in human adult stem cells (ASC) of the small intestine, colon, and liver are added steadily with time, at a rate of about 40 novel point mutations per genome per year. It is noteworthy that mutation spectra of driver genes in cancer are highly similar to the tissue-specific ASC mutation spectra, suggesting that the intrinsic mutational process can initiate tumorigenesis (13).

It is currently suggested that mosaicism can be involved in processes such as inflammation, neurodegeneration, and atherosclerosis, in which postzygotic genetic variation was not thought to exist. In terms of cancer, the role of mosaicism is established beyond any doubt, as growing evidence indicates that the genetic profile underlying carcinogenesis has been shaping over the years prior to the final diagnosis (5). It was estimated that every healthy child is born with at least one cell clone carrying an oncogenic mutation, suggesting that many solid tumors are initiated at the embryonic stage (4). Consistent with this notion, genetic defects associated with malignant transformation have often been detected in the genome of normal cells (14). For example, NOTCH1, NOTCH2, NOTCH3, and TP53 mutations have been found in 18% to 32% of normal skin cells (driver mutation density, ∼140/cm²), which indicate clonal expansion of partially transformed cells long before clinical manifestations. A study described mutations and large structural rearrangements in FGFR3, HRAS, and KRAS oncogenes in sun-exposed healthy skin, and the authors concluded that aged skin represents a patchwork of thousands of evolving clones, with many cells carrying driver mutations while maintaining the physiologic functions (15). Mutations in FGFR2, FGFR3, and HRAS genes have often been found in the spermatogonia of old men (16), indicating age-related accumulation. Genome sequencing of acute myeloid leukemia (AML) samples with normal karyotype or with a known leukemia-initiating mutation (PML-RARA, promyelocytic leukemia/retinoic acid receptor alpha fusion protein), and exome sequencing of HSPCs from healthy people suggest that most mutations in AML genomes are random events that occurred in HSPCs before they acquired the initiating mutation (7).

Live imaging of mouse skin epithelium carrying Wnt/β-catenin–activated stem cells shows that healthy skin cells surround and expel neighbors with driver mutations, suggesting a conserved mechanism (17), which may also exist in humans. If so, this mechanism is evidently incapable of preserving tissue homeostasis throughout the lifespan, as evidenced by sequencing data of 74 cancer genes in physiologically aged normal human skin, which revealed 2 to 6 mutations per megabase per cell (15). Thus, the potential of the biological principle of “the order in the large over heterogeneity in the small,” i.e., the ability of a mass of contiguous normal cells to correct the appearance and behavior of abnormal cells (18), is apparently exhausted with time. As a result, similar to a perfect design sliding into a state of disrepair over time because of the erosion of its constituent elements, the multicellular organism succumbs to aging and various pathologies, including cancer, because of genetic mosaicism (11, 19).

As observed by morphologic imaging, a cancer focus (which violates the normal cell communications and tissue homeostasis, thus behaving as a “criminal”) arises in the background of apparently healthy tissue. The phenomenon of somatic mosaicism reveals the real situation beyond this seeming normality: because all cells of the body appear genetically damaged to some extent, it can be suggested that a cancer cell always emerges in a somewhat abnormal environment and can be viewed as a mutant among mutants. It is possible that the degree of abnormality/mutagenesis in the environment (degree of tissue disorganization) defines whether carcinogenesis would be promoted or inhibited, which can explain the fact that the emergence of cancerous cells is relatively uncommon. Although at the population level, one may speak of a cancer pandemic, at the cellular level, malignant transformation is extremely rare: only some people develop mostly single tumors, although the human body consists of 10 to 30 × 10¹² cells, each carrying multiple mutations, including cancer-driving mutations. It seems likely that cancer cells are capable of revealing their tumorigenic potential and giving rise to a cancer only in the apparently very rare situation when the microenvironment is fully compatible, which is clearly demonstrated by the phenomenon of cancer in situ (20).

Almost 130 years ago, Paget proposed a concept of “seeds and soil” to explain the organ specificity of breast cancer metastases (21). This terminology also seems applicable to cancer initiation triggered by the interaction between the transforming cell (seed) and the microenvironment (soil). At the dawn of oncology, the priority in this pair was given to the “seed,” whereas the “soil” was assigned a passive role of selecting the fittest clone; in contrast, the current opinion is that the “soil,” i.e., tissue where the tumor emerges, plays a critical role (14, 22–24).

Impressive evidence of the leading role of the “soil” in carcinogenesis is provided by a seminal study of Mintz and Illmensee (25), who showed that malignant teratocarcinoma cells behaved differently depending on the environment. In the course of an 8-year experiment, ascitic teratocarcinoma cells of black mice (agouti strain) underwent almost 200 transplant generations in syngeneic hosts (the animals died in 3 to 4 weeks); five tumor cells were then injected into a blastocyst from parents of the non-agouti brown strain, which was subsequently transferred to the uterus of a pseudopregnant foster mother. The striped skin and isozyme pattern of the internal organs in the newborn mice proved full participation of teratocarcinoma cells in normal embryonic development, and the chimeric mice eventually produced normal progeny. These findings are supported by more recent experiments (26).

The critical importance of the environment in the development and metastasis of the established tumor is well documented (27–30). Moreover, there is growing evidence that an appropriate microenvironment is capable of initiating tumorigenesis in normal cells. Mutations in tumor stroma cells were found at very early stages of carcinogenesis, and destabilized stroma was shown to induce genetic instability in the adjacent epithelium, which subsequently underwent malignant transformation (31–33).

Normal cells can be induced to carcinogenesis by tissue injury or neighboring aging cells, which have been shown to promote inflammation and tissue remodeling and reprogramming of differentiated cells into pluripotent cells through their secretome and paracrine regulation, ultimately resulting in malignant growth (34–36). Inflammatory cellular mediators of procarcinogenic effects include tumor-associated macrophages and fibroblasts, myofibroblasts, neutrophils, and adipocytes (27, 37), whereas the active factors include cytokines, chemokines, and transcription factors, such as TGFβ, TNF, and NF-κB (38), as well as exosomes and miRNA (39, 40). Epigenetic modifications occurring in response to tissue-level stress can change the behavior of entire cell populations and precede genetic aberrations; it was suggested that this abnormal signaling is the rate-limiting factor in progression of precancerous lesions (14, 41).

Functional and structural tissue integrity is ensured, to a large extent, by biomechanical interactions. Forces existing at the tissue level descend via mechanically responsive sensors to subcellular structures and pathways, thus directing cell fate during embryogenesis and development. Disruption of tensional tissue homeostasis by mechanical stress or oncogenic mutations disturbing mechanosensitive signaling stimulates malignant transformation (42). It is plausible that mosaicism can contribute to cell transformation not only through the well-known pathways but also through deteriorating biomechanical interactions in a tissue.

Somatic genetic mosaicism introduces into this picture a new element, namely, the heterogeneity of tissue structure. This idea is supported by the very phenomenon of mosaicism, which, however, is not formally proved. Indeed, although the theoretical assumption that every cell in the body is genetically unique is supported by deep sequencing of bulk DNA and genome sequencing of individual cells (6, 43, 44), it cannot be validated by a direct experiment: no matter how effective modern technologies are, it is not possible (and probably will never be) to sequence the genome of each of trillion cells composing the body. Hence, we see only the “tip of the iceberg” in the mosaicism phenomenon, i.e., the heterogeneity of cell clones rather than individual cells, because only genetic variants carried by many cells can generate signals that exceed the background and have a chance to be detected by sequencing (45).

Thus, the ascertainment of genetic mosaicism would simultaneously provide evidence for clonal heterogeneity (“cellularity”) of apparently homogeneous tissue in which the morphologic normality may conceal functional, structural, and biomechanical disorder. Mosaicism gives rise to myriad combinations of random mutations that can deregulate cell–cell communications, ultimately generating an overwhelming spectrum of “seed/soil” relationships. By loosening strictly ordered tissue structure and variegating its cellularity, mosaicism can create unique “seed/soil” variants, which are capable of interactive coevolution, leading to the onset of cancer. In reality, such complete compatibility does not exist for the majority of “seed/soil” pairs evolving in a tissue, because the foci of clonal expansion are suppressed at different growth stages, producing forms that are abortive and dormant in situ (20).

The correlation between the lifetime risk of cancer and the total number of divisions undergone by normal self-renewing cells has recently been revealed by Tomasetti and Vogelstein. The authors used it as a basis for their widely discussed “bad luck” hypothesis (46), which assumes that approximately 70% of cancer variability can be attributed to random errors arising during DNA replication in normal, noncancerous stem cells. This “mutation-centric” model of cancer was criticized on the grounds that it does not consider tissue context, whereas “connecting cancer to its causes requires incorporation of effects on tissue microenvironments” (24). The last statement is doubtless, and most probably, it is the genetic mosaicism affecting both “seeds” and “soil” that links cancer to its causes.

For long, it was believed that the decisive role in carcinogenesis belongs to a transformed cell, which overcomes the resistance of the normal environment, multiplies, and produces clones that evolve and colonize the body. The discovery of genetic mosaicism suggests that a significant, if not main, part of the blame for carcinogenesis lays on the environment, which triggers malignant transformation of a normal cell and favors cancer development.

From a practical viewpoint, this discovery prompts us to prioritize intrinsically oriented cancer prevention efforts, i.e., those directed toward normalization of the microenvironment, as a promising strategy for the war on cancer (24, 27, 47). There is every reason to believe that as our knowledge of tumor/host relationships deepens, the strategies aimed at the prevention of chronic inflammation, obesity, neoangiogenesis, and tissue hypoxia, as well as the removal of senescent cells, will become increasingly effective.

No potential conflicts of interest were disclosed.

This work was supported by funding from the N.N. Blokhin Cancer Research Center. I thank Samuil Umansky and Valery Kobliakov for helpful discussions. I would also like to thank Editage (www.editage.com) for English language editing.