2. ETS Transcription Factors Expression in Cancers
2.1. Expression and Involvement in Cancers
Since their discovery in 1983 as part of the gag-myb-ets transforming fusion protein of an avian replication-defective retrovirus (E26), ETS transcription factors have been associated with carcinogenesis
[33][34]. Indeed, the original ETS family member v-ets, a homolog of Ets-1, can transform fibroblasts, myeloblasts, and erythroblasts and cause mixed erythroid–myeloid and lymphoid leukaemia in chicken
[35]. Thereafter, a correlation between ETS genes expression level and tumour progression has been established in a wide range of human neoplasias such as thyroid, pancreas, ovarian, liver, colorectal, or lung carcinomas and a great number of studies showed the complex and essential function of ETS proteins in the progression and prognosis of breast and prostate carcinomas, and Ewing’s sarcoma, as well as in diverse leukaemias
[3][6]. This extensive involvement can be explained by their roles at all stages of carcinogenesis processes. Indeed, ETS transcription factors promote transformation, invasion, angiogenesis, and inflammation, as well as metastasis through a wide range of molecular and cellular mechanisms including metabolism, tumour microenvironment, histone modifications, cancer self-renewal and survival and DNA repair (reviewed in more detail in
[6][36] and summarized in
Figure 1). Indeed, ETS factors have an impact on the metabolism of steroids and nucleotides, which are necessary for the survival of tumour cells. TMPRSS2-ERG modulates the expression of AKR1C3, the androgen biosynthesis enzyme, decreasing dihydrotestosterone (DHT) synthesis
[37]. ETS2 and p53-GOF activate together the expression of deoxycytidine kinase (DCK), an enzyme that phosphorylates deoxyribonucleosides
[38] (
Figure 1A). ETS factors modulate angiogenesis, inflammation, and extracellular matrix (ECM) modification through the direct transcriptional regulation of collagenases, serine proteases, and matrix metalloproteinases (MMPs)
[39] (
Figure 1B). Modulating chromatin dynamics and epigenetics is a third molecular mechanism through which ETS factors mediate tumorigenesis. Indeed, ETS2 cooperates with p53-GOF mutants to regulate the histone acetyltransferase MOZ, the histone methyltransferases mixed-lineage leukaemia 1 (MLL1), and MLL2. Analysis of The Cancer Genome Atlas demonstrated high expression of MLL1, MLL2, and MOZ in p53-GOF patient-derived tumours
[40] (
Figure 1C). ETS factors are also involved in cancer self-renewal and survival by inhibiting the epithelial–mesenchymal transition (EMT) in mammary gland development and breast cancer metastasis. Elf5, an ETS transcription factor, represses the transcription of Snail2, an inducer of EMT
[41] (
Figure 1D). The modulation of DNA repair is the latest molecular mechanism by which ETS factors regulate tumorigenesis (
Figure 1E). The ETS fusion protein, TMPRSS2-ERG is an interaction partner of PARP-1 and DNA-dependent protein kinase catalytic subunit (DNA-PKcs)
[28]. PARP-1 sequestration prevents DNA repair by homologous recombination (HR). TMPRSS2-ERG-PARP-1 interaction inhibits the phosphorylation of DNA-PKcs, thus blocking DNA repair by the non-homologous end joining (NHEJ)
[42]. Moreover, EWS-FLI1 and EWS-ERG directly interact with PARP-1 and DNA-PKcs, blocking DNA repair in Ewing sarcoma and prostate cancer
[27].
Figure 1. Mechanisms driven by ETS factors in solid tumours. (A) TMPRSS2–ERG activates the expression of AKR1C, the androgen biosynthesis enzyme. ETS2 complexed to the p53-GOF mutant regulates the transcription of DCK, an enzyme that phosphorylates deoxyribonucleosides. (B) ETS factors modulate the expression of numerous factors (e.g., collagenases, serine proteases and MMPs) implicated in angiogenesis, inflammation, and ECM modification. (C) Synergistic cooperation between ETS2 and p53-GOF mutants to activate the transcription of MLL1, MLL2 and MOZ, enzymes of histone modifications. (D) ELF5, an ETS transcription factor represses the expression of SNAI2, an inducer of EMT and metastasis. (E) ETS fusion proteins (TMPRSS2-ERG and EWS-FLI1) interact with PARP-1 and DNA-PKcs. PARP-1 sequestration prevents DNA repair by HR and DNA-PKcs sequestration inhibits DNA repair by NHEJ.
It is possible to differentiate several groups of ETS transcription factors regarding their specific role in particular types of neoplasia. A report by Wei et al. demonstrated that we can separate ETS proteins from each other by analysing their DNA-binding profiles and their affinity for slight base variation within the EBS
[4].
2.2. ETS Fusions and Cancers
Expression or deregulation of ETS transcription factors in cancer cells is mostly due to the activation of ETS genes by amplifications or punctual mutations, but also by chromosomal translocations. In this case, ETS genes are often involved in chromosomal translocations which result in a fusion with other proteins
[6]. There are different types of ETS fusions. The most studied are ETS gene fusions in Ewing’s sarcoma and prostate cancer
[43][44][45]. Other fusions involve Etv6 and several proteins in a wide range of leukaemias, but we will not approach them here (reviewed in
[46]).
Ewing’s sarcoma is a rare cancer affecting bones and soft tissues occurring in teenagers and young adults. Cytogenetically, this cancer is defined by its signature translocations that produce fusion proteins with strong carcinogenic potential. In 90% of cases, the fusion occurs between the encoding regions of the N-terminal portion of EWS, an RNA binding protein, and the C-terminal portion of Fli1, an ETS transcription factor. In the remaining 10%, the fusion occurs between EWS and other ETS factors from the same group, Erg
[43].
The resulting fusion proteins act as transcription factors and regulate a wide range of genes involved in proliferation, carcinogenesis, and tumour progression
[43][47]. EWS-ETS fusions are considered as the keystone of Ewing’s sarcoma development. Thus, their targeting is a high priority to treat this disease
[43][47].
In prostate cancer, numerous gene fusions are found involving different ETS transcription factors such as Erg, Etv1, Etv4, and Etv5
[44]. However, none of them gained as much importance in clinical practice as the rearrangement involving TMPRSS2 and Erg genes. This rearrangement, found in approximately 50% of prostate cancers, results in a gene fusion, TMPRSS2:Erg, which places ERG expression under the transcriptional control of androgen and oestrogen receptors
[48]. TMPRSS2:Erg fusion might be associated with the transition to invasive cancer by over-regulating Erg target genes involved in migration, invasion, and metastasis
[44][49][50].
3. PARP-1 Inhibition in Cancer Therapy
3.1. The Plethoric Roles of PARP-1 in Cancer Cells
PARP-1, one of the most abundant proteins in the cell nucleus, is the founding member of the PARP family of enzymes. Its catalytic activity is characterised by the addition of ADP-ribose polymers, by consuming cellular NAD
+, on the target proteins with which PARP-1 interacts. This post-translational modification is called poly(ADP-ribosyl)ation (PARylation)
[51]. Since its discovery in 1963, PARP-1 has mainly been associated with DNA repair processes, notably in the Base Excision Repair (BER) mechanism where PARP-1 facilitates the association and dissociation of the repair complexes after recognition of single strand breaks (SSBs). Furthermore, PARP-1 is also involved as a backup enzyme for the repair of diverse single-strand and double-strand DNA lesions like in HR, NHEJ, and Nucleotide Excision Repair (NER)
[52][53].
In a sense, PARP-1 is situated at the crossroads of signalisation pathways involved in the sensing of genotoxic, metabolic, and oncogenic stresses. Thus, its function could be described as a sensor of the cellular stresses, which explains its plethoric roles in the cell
[54]. Moreover, PARP-1 is integrated in the structure of chromatin. This localisation also allows PARP-1 to be involved in transcription processes since this enzyme controls chromatin remodelling and interacts with transcription factors on the promoter of many genes
[55]. It has been argued that PARP-1 could be involved in the regulation of 3.5% of all genomes in embryonic cells. Even if it is very difficult to know if this effect is due to transcription regulation or chromatin structure defects, it shows the importance of PARP-1 in transcription processes
[56].
Finally, PARP-1 catalytic activity assures the control of energetic resources, since this activity consumes the stock of NAD
+. Therefore, only cells with a high metabolism, like cancer cells, can maintain a sustained PARylation activity at the risk of depleting all NAD
+, which could lead to cell death
[54].
3.2. PARP-1 Inhibition in Cancer Cells and Clinical Trials
First, PARP-1 was considered as a tumour suppressor due to its plethoric roles in DNA repair. It was argued that suppression of its expression and/or its activity could be involved in cancer development. However, transgenic mice deficient in PARP-1 expression do not show any spontaneous tumour development, even though they are more sensitive to alkylating agents which provoke in these mice liver and colorectal cancers with a greater frequency
[51][57].
Indeed, once the tumour is formed, PARP-1 expression tends to be increased. PARP-1 expression is higher in breast
[58][59], liver
[60], colorectal carcinomas
[61], and melanomas
[62]. Furthermore, studies showed that the PARylation level is increased in cancer tissues
[63][64]. We may suppose that tumour cells can hijack PARP-1 activity to promote DNA repair, and therefore cancer progression, without activating cell death pathways which are often defective in these cancers.
3.3. Limitation of PARP-1 Inhibitors in Cancer Therapy
The different phases of clinical trials using PARPi have given encouraging results. Nevertheless, resistance to PARPi is developed in cancer cells and clinical trials through several mechanisms (reviewed in
[65][66]). The first mechanism of PARPi resistance is the upregulation of drug-efflux transporters such as the ATP-binding cassette transporter, ABCB1, which prevents the intracellular accumulation of PARPi
[67][68]. Mutations in PARP-1 that reduce the binding affinity of PARPi to the catalytic domain of PARP-1 or reduce the affinity of PARP-1 to DNA also produce resistance to PARPi
[69][70].
Cancer cell resistance to PARPi might also occur through restoration of HR upon reactivation of BRCA1/2 function or loss of DNA end protection. BRCA1/2 function is restored through reversion mutations
[71] or epigenetic modifications
[72][73]. Loss of DNA end protection occurs only in BRCA1-deficient cells upon loss of 53BP1
[68][74], an NHEJ factor, and the downstream factors such as REV7, RIF1, and the shieldin complex
[75][76][77].
Another mechanism of PARPi resistance is the restoration of replication fork stability by loss of PTIP, EZH2, or RADX expression
[78][79][80] or loss of cell-cycle checkpoint arrest in BRCA1/2-deficient cells
[81].
4. Molecular Interplay between ETS Transcription Factors and PARP-1 Enzyme
4.1. Regulation of PARP-1 Expression and Activity by ETS Transcription Factors
The interplay between PARP-1 and ETS proteins starts with the regulation of the PARP1 gene upon cellular stresses. Indeed, in Ewing’s sarcoma cells, analysis of the PARP1 promoter showed its up-regulation in response to DNA damage, and this is carried out by ETS transcription factors, Ets-1 and Fli1
[82][83]. Furthermore, the depletion of Fli1 in Ewing’s sarcoma cells leads to the disappearance of PARP-1 expression
[27]. This control of PARP-1 expression allows Ewing’s sarcoma cells to resist ionizing radiation and genotoxic agents
[82][83]. Another study showed that in ovarian cancer cells, Ets-1 activates PARP-1 expression synergistically with histone modification H3K9 by binding to the hypomethylated EBS present in the PARP1 promoter
[84]. To keep the balance of the NAD
+ stock, ETS transcription factors also up-regulate the expression of the poly(ADP-ribose) glycohydrolase (PARG) enzyme which degrades PAR to reform the NAD
+ level
[85]. Thus, ETS transcription factors promote the global process of PARylation/dePARylation and therefore increase the efficiency of DNA repair in cancer cells.
However, the interplay between PARP-1 and ETS factors is not limited to the control of PARP-1 expression. Indeed, work by ourselves and others demonstrated protein–protein interaction between PARP-1 and ETS family members, Ets-1, and Erg (WT and fusion)
[28][29]. This interaction is triggered by direct contact between PARP-1 and the ETS domain common to all ETS factors
[86].
Thus, ETS transcription factors promote both PARP-1 expression and activity to allow cancer cells to improve the efficiency of DNA repair and overcome genotoxic stress.
4.2. Control of ETS Transcription Factors Functions by PARP-1
Since it has been reported by Cohen-Armon et al., it has been known that PARP-1 could positively regulate ETS factors transcriptional activity, in this case, Elk-1 factors activity, by promoting their phosphorylation by ERK-2
[87]. Given the fact that PARP-1 is known to have numerous roles as a transcription cofactor, we and others have tried to find out whether PARP-1 could directly regulate transcriptional activity of the ETS factors on target gene promoters. The results showed that PARP-1 depletion leads mainly to a decrease in ETS factors transcription activity
[28][29]. This would tend to prove that PARP-1 is an essential component of transcription platforms, as has already been observed for other transcription factors.
Nevertheless, although PARP-1 is needed as an interaction partner, its catalytic activity is not always necessary to ensure transcription. Furthermore, a possible role of PARylation is to promote a proper dissociation of protein–protein interactions to avoid blockade in the processes and ensure the dynamism and renewal of the complexes
[51][55]. PARylation of the binding partner could then be a determinant factor in the role of PARP-1 in transcription. In the case of Erg, PARP-1 depletion and inhibition both gave the same decrease in Erg transcriptional activity, but there is no evidence of Erg PARylation
[28][88]. On the contrary, in the case of Ets-1, PARP-1 knock-out provoked a decrease in Ets-1 transcriptional activity, whereas PARP-1 inhibition caused an increase
[29]. This could be explained by the fact that Ets-1 is PARylated and that this PARylation is needed to dissociate Ets-1 from the promoter.
5. Cellular Consequences of PARP-1 Inhibition on ETS-Expressing Tumour Cells
5.1. PARP-1 Inhibition Slows down ETS-Expressing Tumour Growth by Inhibiting Invasion and Metastasis and Decreasing Cell Survival
Given the involvement of PARP-1 on ETS-driven transcription, studies have tried to elucidate whether PARP-1 inhibition could counteract ETS factors’ role in tumour growth and progression. In vitro invasion assays showed that PARP-1 inhibition attenuates Erg- and Etv1-dependent invasion in prostatic cancer cell lines
[28]. The same observations were made for Ewing’s sarcoma cells expressing EWS-Fli1
[27]. Furthermore, PARP-1 inhibition leads to a significant decrease in the intravasation capacity of Erg-expressing cancer cells, which is more proof of the reduced ability of ETS-expressing cells to commit invasion under PARP-1 inhibition. But, very surprisingly, the authors did not notice any effect on the proliferation and survival of Erg-expressing cells
[28].
5.2. PARP-1 Inhibition Causes the Accumulation of Unrepaired DSB in ETS-Expressing Cells
Although the decrease in invasion and metastasis after PARP-1 inhibition could easily be explained by disruption of ETS-driven transcriptional activity, the impact on tumour growth and cell survival is more surprising. Indeed, how can we explain that ETS-positive cells are more sensitive to PARP-1 inhibition than ETS-negative cells, when we could have expected that PARP-1 inhibition would just have neutralised ETS factors’ oncogenic effects? This supposes deleterious effects of ETS expression.
It is known that ETS transcription factors cause genomic instability, even if the mechanism is poorly understood
[89]. Indeed, overexpression of ETS factors in cancer cells causes the formation of numerous DSBs
[28][88]. Yet, PARP-1 inhibition impressively increases this ETS-dependent formation of DSB in cancer cells. This increase has been observed for treated cells expressing Ets-1, Erg (WT and fusion), and EWS-Fli1 by evaluating the level of γH2AX, 53BP1, and Rad51 foci, which are well-known DSB markers, and by neutral comet assay
[27][28][29][88]. Of course, the increase in DSB in these cells is dependent on the expression of ETS factors, since ETS depletion abrogates DSB formation during PARP-1 inhibition.
Thus, it seems as if ETS-dependent sensitivity to PARP-1 inhibition is mainly due to an increase in DSB which is highly toxic for the cells. However, the cause of this increase is still under investigation.