Mammalian tumor-like organs. 1. The role of tumor-like normal organs and atypical tumor organs in the evolution of development (carcino-evo-devo)

The prostate gland has many biological similarities with the mammary gland. Like the mammary gland, the prostate originated in placental mammals [ 140 ]. (The prostate glands of male marsupials are disseminated [ 141 ]). In the course of evolution, both glands developed many similarities in physiology, endocrinology, and oncology. Similarities of prostate and breast cancer are outstanding and include common epidemiological, biochemical, and genetic features [ 142 – 144 ]. Both types of cancer are hormone-dependent [ 144 ]. Germline mutations in breast cancer predisposition genes 1/2 (BRCA1/2), especially in the BRCA2 gene, are predictive factors for prostate cancer also [ 145 , 146 ]. On the other hand, prostate-specific antigen (PSA) is found in normal breast tissues and fluids, in breast tumors, and in benign breast disease [ 147 ]. PBOV1 gene is overexpressed in breast and prostate cancer [ 148 ]. This gene originated de novo in humans; its expression is connected with a favorable clinical outcome of breast cancer [ 149 ]. On the other hand, PBOV1 rs6927706 polymorphism is associated with an increased risk of developing breast cancer [ 150 ]. Like the mammary gland, the prostate involutes upon deprivation of hormonal factors, e.g., upon castration [ 151 ]. Similar to the mammary gland, the prostate gland demonstrates the correlation between evolutionary novelty and the highest incidence of cancer [ 78 ]. The GBD 2015 study reported that for men, the most common type of cancer globally was prostate cancer [ 77 ]. In the US, one in eight men will be diagnosed with prostate cancer during their lifetime. Prostate cancer is the leading cancer type for new cancer cases, and the second leading cause of cancer lethality in men [ 152 ]. Age-specific prostate cancer prevalence, determined by autopsy studies, reaches 59% by age > 79 years [ 153 ]. In older men, the prevalence of benign prostatic hyperplasia may reach 100% (reviewed in [ 154 ]), which produces the impression that the prostate is a benign tumor slowly growing throughout the life of an individual. The prostate also has similarities with the placenta. PSA is synthesized and excreted by the placenta [ 155 ]. On the other hand, placenta-specific protein 1 (PLAC1) is expressed in prostate adenocarcinoma [ 156 ]. Like the placenta and mammary gland, the prostate has a regulated invasion stage in its organogenesis. At the earliest stages of prostate development, prostate epithelial buds invade into surrounding mesenchyme. Genes expressed during prostate cancer progression overlap with genes expressed at the most initial stages of prostate development [ 157 ]. This evidence indicates the tumor-like nature of the prostate gland. Recapitulation of neoplastic features at the earliest stage of prostate development, when its identity is first becoming established, points at the possible origin of a prostate from the tumor. Prostate accumulates the amount of zinc almost an order of magnitude higher than other tissues. The high amounts of zinc and citric acid in prostatic fluid (their metabolism is linked to the prostatic gland) are important for the functioning of spermatozoa. In prostate tumors, the concentration of zinc is lower due to the downregulation of zinc transporters ZIP1, ZIP2, and ZIP3 [ 181 – 183 ]. We may guess that the primary adaptation provided by the benign tumor ancestor of the prostate could be an initial accumulation of zinc, which was selected for in evolving prostate organs because it supported the viability of spermatozoa. Downregulation of zinc concentration in prostate tumors may recapitulate the initial evolutionary condition of the ancestor tumors. Hereditary/familial prostate cancer is described [ 158 , 159 ]. Prostate cancer predisposition genes include ATM, BRCA1, BRCA2, CHEK2, EPCAM, HOXB13, MLH1, MSH2, MSH6, NBN, PALB2, PMS2, RAD51D, and TP53 genes. BRCA2, ATM, CHEK2, and HOXB13 mutate more often. The multigene panel is suggested as the primary germline testing for hereditary prostate cancer [ 184 ].

Placenta has recurrently originated in different mammalian orders after infection with different retroviruses, which became endogenous to their hosts (reviewed in [ 2 ]). The ancestral retrovirus env gene has been replaced by new retroviruses and their env genes through independent germline infections in different mammalian lineages [ 24 – 26 ]. This is supported by the discovery of retroviral envelope gene capture and syncytin exaptation for placentation in marsupials [ 27 ]. Syncytin is a domesticated retroviral gene that plays a role in placental biology. Human syncytin is the env gene of human endogenous retrovirus HERV-W. It is expressed in multinucleated placental syncytiotrophoblast and mediates placental cytotrophoblast fusion (reviewed in [ 2 ]). Eutherian placenta shares many features with tumors. Multinucleated cells and syncytia formation are known features of tumors. Placenta and tumors rely on glycolysis for energy production; the DNA of the placenta demonstrates the genome-wide hypomethylation; some cancer/testis antigens are expressed in the placenta; placental trophoblast is capable to create large ploidies; tumor markers present in the serum of cancer patients include placental proteins; placenta produces angiogenic factors secreted also by tumors (e.g. VEGF), and placenta causes maternal immunosuppression. Parallels between cancer and pregnancy in growth, invasion, and immune modulation have been reviewed, shared characteristics of trophoblast cells and tumor cells, their proliferation and invasion, vasculogenic mimicry and angiogenesis, immunologic similarities of the fetomaternal interface and tumor microenvironment, systemic immune modulation, hypoxia and HIF pathway and other links between cancer and placenta development have been extensively discussed [ 2 , 28 – 30 ]. A special conference has been devoted to common features between placental development and cancer growth [ 29 ]. The placenta may be considered a regulated tumor [ 2 ]. The robust regulation of placental trophoblast by TGF-β efficiently inhibits its malignant properties [ 31 ]. The origin of evolutionarily novel genes is connected with the origin of evolutionary innovations and morphological novelties [ 2 , 32 ]. Domesticated retroviral genes expressed in the placenta were evolutionarily novel to their hosts at the time of infection and acquired new functions in the placenta (syncytins). There are also other evolutionarily novel genes expressed in the placenta. Transposable elements have donated genes now expressed in the placenta to their ancestral hosts. For example, two Ty3/Gypsy retrotransposon-derived genes, Peg10 and Peg11 , are predominantly expressed in the placenta of both humans and mice and are essential for placental development, at early and late stages, respectively (reviewed in [ 33 , 34 ]). Many placenta-associated novel genes are found only in certain mammalian species and are not conserved in all Mammalia [ 33 ] . This is in correspondence with their relatively recent evolutionary origin, i.e., in mammals. For instance, mouse trophoblast-specific protein ( Tpbp ) a and b genes and the related rat gene, SSP ( spongiotrophoblast-specific protein ), are specifically expressed in placenta, have been found only in rodents, and are novel. The placenta-specific protein 1 ( PLAC1 ) gene is annotated in the cow, rat, mouse, and human. Endothelin B receptor , Early placenta insulin-like peptide (INSL4) , Midline 1 , and Pleiotrophin genes are found only in humans and New World monkeys. PLAC1 gene, annotated in different mammals, was initially thought of as specifically expressed only in the placenta. But later studies [ 35 , 36 ] have shown that in humans, it is also expressed in testis and various tumors. LTRs of different HERVs are expressed in the placenta and various tumors [ 37 ]. Syncytin-1 is upregulated in different tumors and participates in their pathogenesis [ 38 – 40 ]. Early placenta insulin-like peptide , INSL4 , found in humans and New World monkeys, is present in breast cancer cells and enhances their invasiveness and motility [ 41 ]. The retrotransposon-derived pleiotrophin ( PTN ) gene is expressed in both tumors and the placenta. In the placenta, its expression is determined by the LTR promoter [ 37 , 42 ]. Chorionic gonadotropin (CG) is produced by the placenta to sustain pregnancy [ 43 , 44 ] and ectopically by a variety of tumors. In early studies, human chorionic gonadotropin has been found in lung carcinoma [ 45 , 46 ], in other cancers [ 47 ], and HeLa cell cultures [ 48 ]. CG consists of two subunits, α and β [ 33 , 43 , 44 ]. Subunit α is encoded by a single conservative gene, expressed in the pituitary and placenta, and enters into the composition of all glycoprotein hormones [ 33 , 43 ]. The different β subunits are encoded by separate genes of the luteinizing hormone ( LHB )/chorionic gonadotropin ( CGB ) gene cluster. In humans, this cluster contains one LHB gene, four β subunit genes ( CGB , CGB5 , CGB8 , and CGB7 ), and two genes ( CGB1 and CGB2 ), which encode novel proteins as a result of frameshifting. All genes of the cluster appear to have originated as a result of duplication of the ancestral LHB gene in anthropoid primates. The LHB / CGB gene cluster is specific to primates. It is an evolutionarily young and unstable genome region [ 49 – 52 ]. The products produced by the human LHB / CGB gene cluster include the regular CG consisting of two subunits, hyperglycosylated CG, free β subunit, hyperglycosylated free β subunit, and other variants of CG, mostly degradation products, altogether 12 common variants of human CG. A hyperglycosylated free β subunit is produced by almost all nontrophoblastic human malignancies, including cervical cancer, breast, bladder, ovarian, brain, colorectal, uterine, and lung malignancy cell lines [ 44 , 53 ]. A hyperglycosylated free β subunit may be a molecule with an independent function consisting of the promotion of nontrophoblastic cancer cells' growth and malignancy [ 44 ]. Human chorionic gonadotropin beta subunit genes CGB1 and CGB2 are transcribed in ovarian cancer tissues [ 54 ], in epithelial cancer cell lines [ 55 ], as well as in testis [ 56 ], pituitary [ 57 ], transgenic mouse brain [ 58 ], and in the placenta [ 59 , 60 ]. CGB1 and CGB2 have always been considered pseudogenes since the publication of Talmage and co-authors [ 61 ], but accumulated expression evidence suggested their functional role. In silico study predicted that as a result of frameshifting, they might encode an utterly novel protein [ 50 ]. CGB1 and CGB2 genes originated in the lineage-specific to humans and African great apes and are conserved between humans and chimps. In the gorilla, insertion and deletion mutations disrupt the predicted protein. After duplication in the common ancestor of African great apes, CGB1 and CGB2 genes may have evolved towards a novel functional gene in humans and chimps and to pseudogenization in gorillas [ 50 ]. These many examples of evolutionarily novel genes acting in both placenta and tumors add to the similarity of placenta and tumors and support the hypothesis that the placenta originated from the tumor, induced by oncogenic retrovirus in ancestral eutherian [ 62 – 64 ]. These data also support the broader concept that ancestral hereditary tumors might evolve into functional organs through the expression of evolutionarily novel genes, i.e. evolution by tumor neofunctionalization [ 2 ]. Thus, the eutherian placenta represents the first identified example of a tumor-like organ.

The human brain, the most recently evolved organ, has many features recapitulating those of tumors. Besides a disproportional increase in size, these features include production of excessive neurons during development; aneuploidy connected with recombination-related genes; many additional copies of L1 transposable elements; genetic mosaicism; high level of gene expression; the involvement of many proto-oncogenes and tumor suppressor genes in brain evolution and development (reviewed in [ 2 ]). Evolutionarily novel genes, including cancer/testis/brain genes, are expressed in the brain [ 2 , 160 ]. The pattern of directional and accelerating evolution towards larger brain size has been described within hominins. The human brain is exceptionally large for primates, "238% larger than the size expected for a primate of similar body mass and phylogenetic position" [ 175 ]. During normal development, prolongation of the high prenatal rate of brain growth into early childhood results in a human-size brain, which is much larger than the monkey’s brain. Studies of the ontogenic allometry have shown that prenatal brain-body curves for humans and monkeys are identical, but humans extend their curve into postnatal ontogeny, until two years after birth (reviewed in [ 2 ]). That is probably why the human brain demonstrates more of its tumor-like nature during childhood and infancy. Brain cancers are the most common type of solid organ tumor in children and are the leading cause of cancer death in children [ 161 – 163 ]. The theory of carcino-evo-devo suggests that it was a heritable benign tumor-like process that supplied evolving human ancestors with additional cell masses for brain evolution. Indeed, expansion of the human cerebral cortex may be a result of selection for tumor growth connected with the human-specific loss of tumor suppressor gene GADD45G enhancer [ 164 ]. Tumor suppressor gene BRCA1 (breast cancer susceptibility gene 1) is rapidly evolving under positive selection in primates and humans and participated in the evolution of brain size in humans [ 2 , 108 – 110 ]. Brain tumors possess mechanisms of neural plasticity. Many gliomas molecularly and phenotypically resemble oligodendrocyte precursor cells. Gliomas functionally integrate into electrically active neural circuits through neuron-to-glioma signaling [ 165 ].

Embryonic and neoplastic development have many common features: intensive cell proliferation, invasiveness, cell migration, the convergence of signaling pathways, important roles of proto-oncogenes and tumor suppressor genes, similarities in gene expression and differentiation, cell adhesion and apoptosis, epithelial-mesenchymal transition, and cell plasticity [ 1 , 2 ]. Reactivation of embryonic signaling pathways is characteristic of tumors [ 1 , 3 , 4 ]. Tumors may originate from embryonic cells [ 5 – 9 ]. Tumors are widespread and are represented throughout the phylogenetic tree (reviewed in [ 2 , 10 , 11 ]). Proto-oncogenes and tumor suppressor genes are the oldest gene classes [ 12 ]. Tumors and cancer-related genes originated during the early period of evolution of multicellular organisms [ 2 , 13 , 14 ]. Many similarities between normal and neoplastic development, and the ancient origin of tumors and proto-oncogenes, assume the long history of the coevolution of normal and neoplastic development. Earlier I hypothesized that hereditary tumors might participate in the evolution of multicellular organisms [ 15 – 17 ]. I formulated the hypothesis of evolution by tumor neofunctionalization (below I will call it "the main hypothesis"), which suggested that the evolutionary role of hereditary tumors might consist in supplying evolving multicellular organisms with extra cells masses for expression of evolutionarily novel genes and the origin of new cell types, tissues and organs [ 15 – 17 ]. Several non-trivial predictions of the main hypothesis have been confirmed in my laboratory. The possibility of selecting tumors for a new function in the organism we confirmed using the novel model of the “hoods” of goldfishes. We performed macroscopic and microscopic studies of adult hoods and the dynamics of the hood growth. We proved histologically that these hoods are benign papillomas [ 2 , 18 ]. The prediction about the expression of evolutionarily novel genes in tumors was confirmed in many publications from our lab. As a result, a new class of genes—tumor specifically expressed, evolutionarily novel (TSEEN) genes—have been described (reviewed in [ 19 ]; see also [ 12 , 20 ]). The predicted parallel evolution of oncogenes, tumor suppressor genes, and differentiation genes was confirmed using in silico genomic analysis [ 12 ]. In the same paper, the predicted correspondence of the number of oncogenes to the number of differentiated cell types was verified [ 12 ]. The acquisition of progressive functions, not encountered in fishes, by human orthologs of fish TSEEN genes, was proven using a transgenic zebrafish inducible tumor model [ 20 ]. This paper is considered by many specialists as the direct confirmation of the main hypothesis. We also have shown that the PBOV1 gene, which is overexpressed in breast and prostate cancer, originated de novo in humans and its expression is connected with a favorable clinical outcome of breast cancer [ 149 ]. The accumulating evidence assumed the fundamental nature of the main hypothesis. It was generalized in my book “Evolution by tumor neofunctionalization” [ 2 ], which contained more than one thousand references. The book was translated to Russian [ 21 ] and Chinese [ 22 ]. The main hypothesis started to acquire the shape of the theory. I called it the carcino-evo-devo theory. In a recent publication, this theory was further developed [ 23 ]. Its relationships with existing biological theories have been examined. The conclusion was reached that carcino-evo-devo theory does not contradict the existing biological theories but fills the lacunas between them and explains questions not wholly understood or not explained by current theories. Non-trivial explanations, suggested by the new theory, include the possible role of tumor-bearing organisms as transitional forms in progressive evolution; the role of tumors as the general mechanism to overcome developmental constraints in the origin of major morphological novelties and complex evolutionary innovations; and explanation of the relationships of tumors with embryonic development and evo-devo . The diagram describing carcino-evo-devo relationships has been suggested (Fig.  1 ) [ 23 ]. Fig. 1 Carcino-evo-devo diagram: devo—normal ontogenies. carcino—ontogenies with neoplastic development. evo—progressive evolution of ontogenies. Arrows indicate participation in the corresponding process or essential connections. From A. P. Kozlov (2019) Acta Naturae 11: 65–72, with permission Carcino-evo-devo diagram: devo—normal ontogenies. carcino—ontogenies with neoplastic development. evo—progressive evolution of ontogenies. Arrows indicate participation in the corresponding process or essential connections. From A. P. Kozlov (2019) Acta Naturae 11: 65–72, with permission Another non-trivial prediction of the new theory is that evolutionarily novel organs if they indeed originated from hereditary tumors or tumor-like structures, should recapitulate some features of tumors in their development. That is why I was looking for the data that might confirm this prediction in the literature, and also performed some experiments in my lab. This paper reviews the evidence that evolutionarily novel organs indeed have many features of tumors that supports the main hypothesis of the possible evolutionary role of tumors and the carcino-evo-devo theory [ 2 , 23 ]. Below I will call evolutionarily new normal organs, which have many tumor features, the tumor-like organs for brevity.

We see that evolutionarily novel organs such as the eutherian placenta, mammary gland, prostate, infantile human brain, and hoods of goldfishes have many features of tumors and may be considered as normal but tumor-like organs. Tumor-like organs might originate from hereditary tumors and atypical tumor organs and represent the part of carcino-evo-devo relationships, i.e., coevolution of normal and neoplastic development, and involvement of hereditary tumors in the evolution of development. During subsequent evolution, tumor-like organs may lose the features of tumors and the high incidence of cancer and become normal organs without (or with almost no) tumor features. However, the proneness of different normal organs to cancer development, although with varying rates of incidence, is the tumor feature. According to the carcino-evo-devo theory, this may be a recapitulation of the origin of new organs from the ancestral hereditary tumors.

Tumor-like organs and atypical tumor organs may occupy intermediate, transitory positions on the carcino-evo-devo diagram (Fig.  2 ). Normal tumor-like organs, from one side, and tumors as atypical organs, from the other side, thus help to fill the gaps in the description of the origin of evolutionarily new organs from hereditary tumors. Fig. 2 Carcino-evo-devo diagram with tumor-like organs and atypical tumor organs: devo—normal ontogenies. carcino—ontogenies with neoplastic development. evo—progressive evolution of ontogenies. devo’—tumor-like organs. carcino’—atypical tumor organs Carcino-evo-devo diagram with tumor-like organs and atypical tumor organs: devo—normal ontogenies. carcino—ontogenies with neoplastic development. evo—progressive evolution of ontogenies. devo’—tumor-like organs. carcino’—atypical tumor organs The origin of the eutherian placenta, mammary gland, and prostate in ancestral eutherians may be represented in the following way (Fig.  3 ). As follows from Fig.  3 , the ancestral ontogenesis (Devo 1) has evolved three additional processes of organogenesis (Devo 2’, Devo 2’’ and Devo 2″’) through the participation of three different hereditary tumors/atypical tumor organs (Carcino 1’, Carcino 1’’ and Carcino 1″’). Fig. 3 Carcino-evo-devo diagram illustrating the origin of the eutherian placenta, mammary gland, and prostate in ancestral eutherians Carcino-evo-devo diagram illustrating the origin of the eutherian placenta, mammary gland, and prostate in ancestral eutherians Devo 2, which includes Devo 1 and three novel organs (Devo 2 = Devo 1 + Devo2’ + Devo 2″ + Devo 2’’’), can further evolve with the origin of younger organs like a human brain with its neocortex (e.g., Devo 5 at Fig.  4 ) with the help of the other hereditary tumors/atypical tumor organs (Carcino 4 at Fig.  4 ). Fig. 4 Carcino-evo-devo diagram showing several successive steps in the progressive evolution of ontogenies. From A. P. Kozlov (2019) Acta Naturae 11: 65–72, with permission Carcino-evo-devo diagram showing several successive steps in the progressive evolution of ontogenies. From A. P. Kozlov (2019) Acta Naturae 11: 65–72, with permission Thus, the concepts of tumor-like normal organs and tumors as atypical organs help to understand better the carcino-evo-devo relationships and the role of hereditary tumors as the transitory condition in the evolution of normal development.

Infantile in situ neuroblastomas, detected by screening, represent an interesting example of tumors fundamentally different from symptomatic tumors. Their natural history is mostly unknown. Neuroblastomas may regress or differentiate into benign cells in older children. Infant screening programs for neuroblastoma demonstrated increasing incidence rates of early-stage tumors. Still, they did not show any increase in more advanced tumors and deaths due to neuroblastoma (reviewed in [ 2 ]). It was suggested to call such lesions the “pseudodisease” [ 166 , 167 ]. Tumor-like conditions exist in different normal organs. Tumor-like conditions can be defined as conditions that macroscopically and/or microscopically may appear as neoplasms but are not truly neoplastic [ 176 ]. Many organs (bones, pleura, lung, heart, brain, etc.) have tumor-like conditions which complicate the diagnostics of the tumors. Tumor-like malformations (e.g. hamartomas and choristomas) occur anywhere in the body and may be confused with true neoplasms [ 177 , 178 ]. Breast lumps may be non-neoplastic, benign, and malignant. Tumor-like conditions of the mammary gland discussed above may never progress to cancer. The borders between tumor-like conditions, benign and malignant tumor processes in the mammary gland are difficult to draw, and the prognosis is difficult to make. The borderline tumors (tumors of low malignant potential) may be associated with tumor-like conditions. For example, borderline ovarian tumors may be related to ovarian endometriosis—the tumor-like condition of endometrial cells growing outside the uterus in about 10% of reproductive-age women [ 179 , 180 ]. The existing evidence produces the impression that normal organs are not fixed and stable entities, but fluctuating and relatively unstable in terms of their cellular composition and proliferative processes, sometimes resembling tumors. This may reflect the role of hereditary tumors in their origin, which may consist in providing cellular material for the natural selection of new or improved organs.