Programming of Cell Resistance to Genotoxic and Oxidative Stress

Different organisms, cell types, and even similar cell lines can dramatically differ in resistance to genotoxic stress. This testifies to the wide opportunities for genetic and epigenetic regulation of stress resistance. These opportunities could be used to increase the effectiveness of cancer therapy, develop new varieties of plants and animals, and search for new pharmacological targets to enhance human radioresistance, which can be used for manned deep space expeditions. Based on the comparison of transcriptomic studies in cancer cells, in this review, we propose that there is a high diversity of genetic mechanisms of development of genotoxic stress resistance. This review focused on possibilities and limitations of the regulation of the resistance of normal cells and whole organisms to genotoxic and oxidative stress by the overexpressing of stress-response genes. Moreover, the existing experimental data on the effect of such overexpression on the resistance of cells and organisms to various genotoxic agents has been analyzed and systematized. We suggest that the recent advances in the development of multiplex and highly customizable gene overexpression technology that utilizes the mutant Cas9 protein and the abundance of available data on gene functions and their signal networks open new opportunities for research in this field.


Introduction
Genotoxic stress, including oxidative stress, causes DNA damage. The evolutionary conservative cellular mechanisms of DNA-damage prevention and response (DNA repair, defense against reactive oxygen species, cell cycle checkpoints, and apoptosis) protect cells from mutations and tissues from acquiring malignancy [1,2]. On the one hand, genotoxic stress can induce carcinogenesis, on the other hand, it is used to treat cancer. The advancement of knowledge on regulation of stress-resistance in cells and organisms is extremely important for increasing the effectiveness of cancer treatment. In particular, the creation of new in vitro models of upregulated cell resistance to genotoxic and oxidative stresses allows for expanding the spectrum of in vivo models for studies of genetic regulation of carcinogenesis. In addition, it was suggested multiple times that gene therapy of normal tissues surrounding tumor can be used for increasing their resistance to genotoxins. This can help to minimize the negative side effects of cancer treatment by chemotherapy and radiation therapy [3][4][5]. This technology can also be used for gene therapy and gene prophylaxis of diseases that are associated with increased sensitivity to DNA-damaging agents [6]. Understanding the mechanisms of cellular stress resistance, and especially resistance to oxidative stress, is one of the most important tasks in studies of lifespan extension [7,8]. Knowledge of stress-resistance is also important when creating new genetically modified varieties of Due to the risk of carcinogenesis, the mechanisms described above cannot be used as practical targets for induction of cellular stress-resistance. However, stress resistance of tumor cells is often formed by the mechanisms that are not associated with initiation of malignant transformation. As mentioned above, alteration in components of genome stability machinery could lead to an increase in mutation rate in tumors, and result in an increased genetic heterogeneity of cells. This heterogeneity facilitates the rapid selection of cells subpopulations that are resistant to stress [23]. The possibility of this selection-based mechanism of resistance has been repeatedly confirmed in direct selection experiments [30][31][32]. However, there is also evidence that stress-resistance can be induced at the epigenetic level, independently from the selection process [33]. The resistance that is developed by selection or independently of it often results from the overexpression of the genes encoding transporter proteins, which support enhanced drug efflux [24]. In many cases, overactivation of DNA damage recognition and repair as well as detoxification of free radicals are also observed. For example, Rad51 gene, which is involved in homologous recombination is overexpressed in a variety of human cancer types. This often leads to chemo-resistance of these tumors [34]. An inverse correlation was observed between the expression of the excision repair gene ERCC1 and the sensitivity to platinum treatment of various types of tumors [35]. An enhancement of excision repair activity in lung cancer cells can also be associated with a SIRT1 dependent increase in XPA sensitivity to DNA damage [36]. Expression of the antioxidant defense gene-MnSOD-correlates with resistance to doxorubicin and mitomycin C in gastric carcinoma cells [37]. RPA1 gene, which is involved in DNA replication and repair is overexpressed as a result of selection of a radioresistant clone in esophageal carcinoma cell line TE-1. Inhibition of RPA1 in that radioresistant clone restored the normal sensitivity to ionizing radiation [38].
There are many other examples of an established link between genotoxic stress resistance and overexpression of genes involved in DNA repair, xenobiotic detoxification, or efflux. However, the diversity of possible mechanisms of resistance seems to be even larger. This is supported by the studies comparing transcriptomes of similar cell lines that differ in sensitivity to genotoxic agents. For example, a comparison of ten microarray studies performed on cancer cells with different degrees of resistance to ionizing radiation did not identify any commonly overexpressed genes [39][40][41][42][43][44][45][46][47][48]. We could not find a gene that would be significantly overexpressed in three or more comparison pairs. Approximately 95% of the total number of overexpressed genes were observed in only one study and were absent in others ( Figure 1). Interesting, that among the genes overexpressed in two different studies most are interferone induced genes, which involved in response to virus infection [49]. This fact shows once again that different systems can be involved in the regulation of resistance to genotoxic stress.
Thus, the diversity of pathways leading to resistance in cancer cells, allows for us to suggest a wide range of possibilities for increasing resistance of normal cells to genotoxic and oxidizing agents. We suppose, that if we exclude all of the targets that affect cell cycle control, apoptosis, proliferation, and differentiation, we can enhance stress-resistance without the risk of increasing malignancy. Moreover, the increased efficiency of cellular defense systems should in theory lead to a decrease in carcinogenesis. This assumption is supported by the fact that the activity of DNA repair systems inversely correlates with the risk of neotransformation [50]. In addition, a decrease in alkylating agent-induced carcinogenesis has been repeatedly demonstrated upon overexpression of the gene O 6 -methylguanine-DNA methyltransferase (MGMT), which is responsible for DNA damage recognition and repair [51][52][53][54][55][56].
pairs. Approximately 95% of the total number of overexpressed genes were observed in only one study and were absent in others ( Figure 1). Interesting, that among the genes overexpressed in two different studies most are interferone induced genes, which involved in response to virus infection [49]. This fact shows once again that different systems can be involved in the regulation of resistance to genotoxic stress.  CXCL10,  IFI44, IFIH1, IFITM1, STAT1, DDX60, HERC6, IFI27, PLSCR1, IFIT1, IFI35, IFIT3; c -ISG15; d -ERP70. Numbers in parenthesis is the quantity of transcripts analyzed. *-Genes that are involved in apoptosis, DNA repair, cell cycle control, cell proliferation and other mechanisms of stress response. **-only tumor-related genes.

Genotoxic Stress Resistance in Experimental Models with Gene Overexpression
Change in gene transcription is only one of the existing ways of readjusting the mechanisms of stress resistance. Another way of establishing stress resistance is a pharmacological targeting of proteins and signaling cascades, which seem more acceptable for clinical applications. However, accumulation of experimental data on the effects of overexpression of individual genes and their combinations is required to develop pathways of stress-resistance regulation that might help finding new pharmacological targets. The literature on the effects of overexpression of stress-responsive genes on the resistance of cells and organisms to genotoxins is overwhelmingly broad. However, we attempted to systematize such published experimental data based on overexpressed genes, on the effect on stress resistance and on genotoxicants. Being mindful of the scale and the variety of the published studies, in our analysis we chose a simple algorithm of grouping the target genes by their function. The resulting lists of reviewed published reports are presented in Tables 1 and 2 for in vitro and in vivo studies, respectively. One interesting, but not totally surprising, finding of our analysis was that most studies driven by a targeted hypothesis (about involvement of a particular gene in stress resistance based on previous experimental evidence) found that overexpression of the gene did increase stress resistance. On the other hand, it seems that in case of randomly selected targets, the predominant outcome would be sensitization to stress, likely due to a disruption of normal gene activity regulation.
As suggested above, the two most promising gene categories to enhancing resistance by overexpression are the genes involved in DNA damage recognition and repair, as well as the genes that are responsible for efflux and detoxification of xenobiotics. Overexpression of these genes tends to be the most successful strategy of enhancing resistance to genotoxic stresses without the risk of increasing the frequency of neoplastic transformations. However, overexpression of these targets does not always lead to an expected/desired outcome. Firstly, an increase in survival can mask the decrease in DNA repair quality. For example, overexpression of the gene encoding DNA polymerase β in CHO cells lead to an increase in survival after treatment with cisplatin, melphalan, or mechlorethamine. However, it also dramatically increased the frequency of mutations in surviving cells. DNA polymerase β, the most error prone eukaryotic DNA polymerase [57][58][59][60] has been repeatedly shown to be the cause of the phenomenon mentioned Therefore, the required outcome and endpoints used should be selected carefully. Secondly, the effect of overexpression of various single elements of a repair or detoxification system/pathway can sometimes produce an effect that is opposite of the expected one. At the cellular level, the two main groups of reasons for this are (a) the imbalance between the elements of the protective systems; and, (b) the absence of the expected relationship between the level of gene transcription and the activity of the gene product. The latter primarily applies to all of the proteins whose activity depends on post-translational modifications. The mismatch between the mRNA levels and the protein function may also arise when a gene encodes only one subunit of multisubunit protein complexes. For example, stability of the DNA repair protein XPC depends on the levels of HR23A and HR23B proteins [61], therefore the overexpression of XPC gene may not be sufficient to enhance nucleotide excision repair. As consistent with this, an averaged quantitative relationship between the levels of mRNA and corresponding protein tends to be weak [62]. However, estimations of this correlation are still the subject of discussion and differ widely in the range from 0.21 to 0.9 [63]. In exceptional cases, for example, in the case of ribosomal proteins, mRNA can be a repressor of translation of its own product. This phenomenon is known to occur for the RpS3 protein that is involved in stress responses [64].
The imbalance of protective systems resulting from overexpression of individual genes may be caused by several different mechanisms. First, it can be driven by the imbalance in productivity of successive stages of a single cascade. For example, a wide range of modified bases in S. cerevisiae is excised using MAG1 (3-methyladenine DNA glycosylase). The abasic sites that are generated by MAG1 are processed normally by the major yeast APN1-encoded AP endonuclease. Disproportionately high expression of MAG1 when compared to the AP endonuclease increases spontaneous mutation by up to 600-fold in S. cerevisiae and by 200-fold in E. coli [65]. CHO cells with overexpressed MPG gene are more sensitive to alkylating agent N-methyl-N'-nitro-N-nitroso-guanidine (MNNG) that is also associated with excessive accumulation of abasic sites [66].
Secondly, there are situations when an increase in resistance to one agent is accompanied by sensitization to others. For example, overexpression of APE1 increases the resistance of CHO cells to dioxolane cytidine [67], but it sensitizes cells to agents, which are activated by reduction reactions. This happens because the product of APE1 gene has a RedOx function in addition to AP endonuclease activity [68]. Another mechanism is a shift in balance between the two competing processes. For example, the overexpression of XRCC1 required for base excision repair (BER) slows gap-filling, because of the competition of BER with nucleotide excision repair for the PCNA protein [69].
The listed nuances of regulation of resistance to genotoxic stress explain the opposite outcomes observed during the overexpression of the same genes in different experiments (Tables 1 and 2). The same opposite outcomes are observed on the level of functional groups of gene, as obtained using PANTHER classification system [70,71]. The classification shows that researchers mainly chose the genes encoding nucleic acid binding proteins and proteins that catalyze redox reactions. This is expected, since the many proteins of these groups are involved in DNA repair and oxidative stress defence, respectively. At the same time, if we divide the experiments that are based on the direction of the effect on stress-resistance, the ratio of the functional groups does not change significantly ( Figure 2). This means that we cannot say that in fact overexpression of the genes of one of these functional groups increases the stress resistance more effectively than the overexpression of the genes of the other group. At the level of the whole organism, potential disruptions of functional interactions between cells, tissues, organs, and organ systems are added to the intra-cellular mechanisms of imbalance listed above. But improvements in survival, decrease in frequency of mutations, fewer incidence of cancer, and some others desirable outcomes are still observed as a result of overexpression of stress-responsive genes in a number of studies, which holds promise ( Table 2).  Table 1 were divided into two groups, depending on the effect of their overexpression on the resistance of cells ("In vitro"). The same division was performed for orthologues of genes listed in Table 2 ("In vivo"). Each groups was classified using PANTHER Protein class ontology [70,71]. *-number of analyzed genes/total number of hits to "PANTHER protein class" classification.

Prospects
The decrease in stress-resistance of cells in the variety of experiments described above is largely caused with the multicomponent nature of stress response mechanisms that the studied genes participate in. Numerous experimental data that support the high efficiency of overexpression of the MGMT gene support confirm this assumption (Tables 1 and 2). Product of this gene solely performs recognition and repair of damaged DNA bases, in contrast to most other elements of cell protective systems that operate in cooperation with many other gene products [73]. When considering the accumulated detailed knowledge of such interactions, the development of multiplex gene activation systems with mutant RNA-guided Cas9 protein opens up the widest opportunities for studying the regulation of stress resistance. Multiplex activation using one large [74] or a number of small [16] plasmids, using activators with different degrees of efficiencies, allows for selecting the appropriate range of activation. To some extent, the level of overexpression of individual genes can be adjusted by selecting sgRNA for sequences that are located at different distances from the transcription start site.   Table 1 were divided into two groups, depending on the effect of their overexpression on the resistance of cells ("In vitro"). The same division was performed for orthologues of genes listed in Table 2 ("In vivo"). Each groups was classified using PANTHER Protein class ontology [70,71]. *-number of analyzed genes/total number of hits to "PANTHER protein class" classification.
In addition to the above, there are, apparently, many other factors that can radically change the influence of overexpression of certain genes on cellular stress-resistance. This is supported by the cell line specific effect of overexpression of the proto-oncogene HER2/neu in human breast and ovarian cancer cells. In six different cell lines, overexpression led to either a decrease, or an increase in sensitivity to chemotherapeutic agents of different classes [72]. These experimental data provide additional evidence in favor of the need for further studies of genetic regulation of stress resistance in normal and cancerous cells, as well as the stress-resistance of an organism as a whole.

Prospects
The decrease in stress-resistance of cells in the variety of experiments described above is largely caused with the multicomponent nature of stress response mechanisms that the studied genes participate in. Numerous experimental data that support the high efficiency of overexpression of the MGMT gene support confirm this assumption (Tables 1 and 2). Product of this gene solely performs recognition and repair of damaged DNA bases, in contrast to most other elements of cell protective systems that operate in cooperation with many other gene products [73]. When considering the accumulated detailed knowledge of such interactions, the development of multiplex gene activation systems with mutant RNA-guided Cas9 protein opens up the widest opportunities for studying the regulation of stress resistance. Multiplex activation using one large [74] or a number of small [16] plasmids, using activators with different degrees of efficiencies, allows for selecting the appropriate range of activation. To some extent, the level of overexpression of individual genes can be adjusted by selecting sgRNA for sequences that are located at different distances from the transcription start site.  Genes involved in control of proliferation and cell cycle CCND1 (595) Human adenocarcinoma cells (MCF7) γ-ray ↓ [145] p21 (1026) Glioma cells (T-98G, U-251MG with mutant p53 allele and U-87MG with wild-type p53). Medulloblastoma cells MED-3.

Genes involved in detoxification and efflux of free radicals and xenobiotics
Gclc (