WO2022058472A1

WO2022058472A1 - Method for predicting the risk of radiation-induced toxicity

Info

Publication number: WO2022058472A1
Application number: PCT/EP2021/075565
Authority: WO
Inventors: Sarah LIBERT; Thierry MAILLET; Giovanna MUGGIOLU; Sylvie Sauvaigo; Bertrand TREILLARD
Original assignee: Lxrepair
Priority date: 2020-09-17
Filing date: 2021-09-17
Publication date: 2022-03-24

Abstract

The invention relates to an in vitro method for predicting the risk of radiotoxicity such as inflammation-induced fibrosis of a subject, in particular a cancer patient treated by radiotherapy, comprising determining the Base Excision Repair (BER) activity in a subject sample, and deducing therefrom if the subject is at risk or not of radiotoxicity.

Description

METHOD FOR PREDICTING THE RISK OF RADIATION-INDUCED TOXICITY

FIELD OF THE INVENTION

The present invention relates to a method for predicting the risk of radiation-induced toxicity such as inflammation-related fibrosis in a subject, in particular a subject treated by radiotherapy or potentially exposed to ionising radiations or radiomimetics, comprising analysing the Base Excision Repair (BER) activity in a subject sample.

BACKGROUND OF THE INVENTION

About one in two patients with cancer receive radiotherapy (RT) treatment. While today the majority of patients tolerates standard radiotherapy, up to 10-15% of patients suffer from adverse effects: acute and late effects (Bentzen SM. et al. Semin Radiat Oncol., 2003, 13:189- 202; Popanda, JU. et al. Mutat Res. 2009, 667:58-69) arising from the intrinsic sensitivity of normal tissues.

Early or acute effects occur within days/weeks of treatment and late effects occur several months or years following completion of the RT. Inflammation is the hallmark of acute effects which translates into dermatitis, cystitis, mucositis, etc. Late effects are more diverse and encompass fibrosis, atrophy, vascular damage, gastrointestinal and genitourinary syndromes, second malignancies, etc. (Barnett GC. et al. Nat Rev cancer, 2009, 9:134-142; De Ruysscher D. et al. Nat Rev Dis Primers. 2019, 5(1): 13).

Late effects would be related to fibrogenesis associated with activated cytokine cascade, involving TGFp. This pro-fibrotic cytokine, would stimulate in turn an uncontrolled extracellular matrix remodelling and collagen deposition, leading to tissue hypoxia (Bentzen SM., 2006, Nat Rev 6, 702-713).

Early and late side effects limit radiation dose and might affect long-term health-related quality of life of the patient. An identification of the individuals at risk of developing adverse effects would allow a personalization of the RT regimen and an adjustment of the RT dose (Yaromina A. et al., Mol. Oncol., 2012, 6:211-221). These various adverse side effects of radiotherapy are at least partially associated with altered genetic factors involved in the DNA Damage Response (DDR). The role of DNA doublestrand break (DSB) repair in radiotoxicity is notably illustrated by the fact that Fanconi Anemia patients have high rate of complication to radiation-therapy (Kennedy RD. et al. J Clin Oncol. 2006, 24:3799-808; Birkeland AC. et al. Arch Otolaryngol Head Neck Surg. 2011, 137:930-934). DNA damage initiates signalling response which is dependent on the activation of ataxia telangiectasia mutated (ATM) or ATM and Rad3 -related (ATR) proteins. ATM is activated by DNA double-strand break and ATR by DNA single-strand break. ATM- deficient cells and cells lacking proteins involved in ATM signalling (H2AX, MDC1, 53BP1, etc.) also exhibit dramatic radiosensitivity (Goodarzi AA. et al. Mutat. Res. 2012, 736:39-47).

DNA double-strand breaks (DSBs) are considered as the most lethal form of DNA damage and a primary cause of cell death. DSBs are induced by ionizing radiation (IR) during radiotherapy and it was long believed that defects in DSBs repair was the sole responsible for the pathogenesis of late normal tissue effects in radiotherapy patients. For these reasons researches have essentially focused on double-strand break repair pathways (Yaromina A., Mol Oncol, 2012, 6:211-221).

The invention described in WO2015/121596 aims at characterizing the radiosensitivity and the tissue reaction of a patient to therapeutic ionizing radiation. It requires the amplification of cells taken from subjects and the in vitro irradiation of these samples. It is related to the detection of the residual amounts of markers selected to detect double-strand breaks (DSB) (ATM, yH2AX, MRE11). This invention focuses only on double-strand breaks indirect biomarkers.

US 2016/0054337 discloses a method for the determination of the radiosensitivity of a subject on a biological sample comprising T lymphocytes from the subject which requires the isolation, culturing and irradiation of T lymphocytes from the test sample, prior to measuring the presence or increased level of a protein chosen from mitochondrial isocitrate dehydrogenase 2 (IDH2), DNA-(apurinic or apyrimidic site) lyase (APEX1), Heat shock cognate protein 71 kda (HSC70), adenylate kinase (AK2), annexin 1 (ANX1), or nucleic acid encoding the same in the test sample. As an improvement of the previous method, the patent application WO2018/229439 described the use of a single biomarker (pATM) detected at basal level, in the absence of in vitro irradiation, on subject cells, as a predictor of the subject radiosensitivity. This method overcomes the inconvenience of the need of the in vitro irradiation of the samples, but is restricted to the detection of the basal expression level of a unique double-strand break biomarker. The method is based on the detection of the activated ATM protein, under its phosphorylated form (pATM). However, it is known that DSB can be repaired by several repair pathways, and that the presence of pATM cannot represent the whole DSB repair mechanisms. Moreover, the authors affirmed that ATM is under the regulation of BRCA1 that triggers ATM phosphorylation. Consequently, absence or deficiency of BRCA1 would prevent any further phosphorylation of ATM and thus renders the method inappropriate. The method proposed in the patent application WO2018/229439 thus presents a major drawback as it focuses on only one activated protein that is under the regulation of frequently mutated signalling pathway and it cannot be generalized.

A lot of researches focus on the potential yH2AX foci analysis as a marker for double-strand break formation and repair. This histone is phosphorylated following activation of the ATM protein in response to double-strand breaks (Martin OA. et al. DNA repair (AMST), 2013, 12:844-55; Olive PL. Radiother Oncol., 2011, 101:18-23). Its use is based on the rational that defects in DNA Repair signalling following formation of DSB are at the origin of overreactions to radiotherapy. However, a multicentric study conducted on in vitro irradiated peripheral Blood Lymphocytes (PBLs) of head and neck patients highlighted that yH2AX, like the comet assay, was unable to detect differences in radiosensitivity among individual cancer patients (Greve B. et al. Pios One. 2012, 7:e47185).

Although early adverse effects of radio toxicity, in particular caused by radiotherapy, are mostly transient, late effects tend to be irreversible or even progressive in severity. As late effects are associated with both human suffering and direct health costs, treatment modifications would be justified in patients at high risk of developing late toxicity if they could be reliably identified by an assay. At the same time, the identification of predictive markers could point to new interventional targets for ameliorating late effects (De Ruysscher D. et al., Nat Rev Dis Primers. 2019 Feb 21; 5(1):13). Thus, currently, no method has demonstrated sufficient robustness, power or applicability to permit a significant development for clinical purpose. It remains a need to develop predictive method to determine the risk of radiotoxicity in cancer patients.

SUMMARY OF THE INVENTION

Although the relationship between BER and inflammation is not completely known, the inventors have focused instead on Base Excision Repair (BER) system which is the primary pathway responsible for DNA repair during inflammation since it deals with single-base lesions resulting from oxidation, deamination and alkylation, created by the inflammation state.

Most inflammation-induced DNA damage is caused by reactive oxygen and nitrogen species (RONS), which are produced endogenously as a consequence of ionising radiations. Many RONS are potent oxidizing agents and can produce a variety of DNA lesions. The most common oxidation RONS-induced is the oxidation of guanine. 8-oxo-guanine (8-oxoG) and nitro-guanine (which is unstable and quickly becomes an abasic site). Once produced they can lead to T-G transversions. RONS can deaminate DNA bases which alter the pattern of hydrogen bonds leading to base mispairing and producing chemistry products as Uracile (U) and hypoxanthine (Hx). Abasic resulting from the hydrolysis of a nucleotide's glycosidic bond can also be created (Sage E. & Shikazono N. Free Radic Biol Med. 2017 Jun; 107:125-135). In addition, RONS can also cause indirectly damage by creating reactive species from other biomolecules. For example, encountering polyunsaturated fatty acids they cause lipid peroxidation resulting in DNA exocyclic adducts of two (etheno-) or three (propano-) carbons onto a base. All etheno- lesions can block replication to some extent, which can lead to larger- scale mutations (Kay J. et al, DNA Repair, 2019, 83:102673).

An orchestrated active biological response is activated upon inflammation, proteins involved in BER are activated and they promote many cytokine cascades by creating a positive loop with pro-inflammatory genes. In case of radiation therapy, the balance between inflammation and BER may become deregulated and may contribute to a systemic or loco-regional toxicity also in healthy tissues. Reasons for this possible deregulation in some individuals are not known, but could involve altered BER. For example, AAG-initiated BER at alkylated DNA bases activates an inflammatory response that amplifies the original damage (Alloca M. et al. Sci Signal. 2019; 12(568):eaau9216) To confirm this hypothesis, the inventors have performed a prospective study on a cohort of patients suffering from various types of cancers. The inventors have analysed the BER activity in normal tissue of the cancer patients before or at early stages of radiotherapy treatment using a panel of various DNA substrates that are recognized and cleaved by distinct DNA glycosylases and the AP-endonuclease of the BER system. For all types of cancer, they observed that patients who developed severe adverse effects following radiotherapy have an altered BER activity as compared to cancer patients who did not. Using the repair activities towards the various DNA substrates before or at early stages of radiotherapy treatment as distinct biomarkers, they have obtained combination of biomarkers that allow to predict the risk of radiotoxicity in a cancer patient with a high specificity and sensitivity. This demonstrates that the analysis of BER activity allows the prediction of radiotoxicity risk in a subject. In particular, the analysis of BER activity profile allows the prediction of radiotoxicity risk in a subject, by combining biomarkers obtained from the subject’s samples collected before or at early stages of radiotherapy treatment. While some prior art methods require the culturing and/or irradiation of cells from a subject sample, these steps are absent in the method of the invention which is therefore more simple and usable in clinical routine.

The present invention relates to an in vitro method for predicting the risk of radiotoxicity of a subject, comprising determining the BER activity in a subject sample, in particular the BER activity profile before or after few sessions of radiotherapy, and deducing therefrom if the subject is at risk or not at risk of radio toxicity.

In some embodiments, the radiotoxicity is from radiotherapy.

In some embodiments, the radiotoxicity is late radiotoxicity and/or includes inflammation- induced fibrosis.

In some embodiments, the subject is a cancer patient, preferably chosen from breast, head and neck, lung or prostate cancer and melanoma patient.

In some embodiments, the method comprises determining the presence or absence of an altered BER activity in the subject sample, wherein the presence of an altered BER activity indicates that the subject is at risk of radiotoxicity; preferably wherein the altered BER activity is a decreased or increased activity compared to a reference. In some embodiments, the BER activity is determined by: contacting the sample with at least one damaged DNA comprising lesion(s) that are repaired by BER pathway proteins and further comprising a detectable marker which is released during BER reaction and, measuring the variation of signal from the marker; preferably wherein the BER activity involves at least one enzyme selected from the group consisting of UDG glycosylases UNG2, SMUG1, TDG and MBD4; OGGI, NTH1, NEIL, MUTYH and AAG glycosylases; APE1 endonuclease and combination thereof.

In some preferred embodiments, the damaged DNA comprises at least one lesion selected from the group consisting of: (i) 8-oxoG paired with cytosine (8oxoG-C), (ii) adenine paired with 8oxoG (A-8oxoG), (iii) thymine glycol paired with Adenine (Tg-A), (iv) hypoxanthine paired with thymine (Hx-T), (v) ethenoadenine (EthA) paired with thymine (EthA-T), (vi) abasic site analogue paired with adenine (THF-A), (vii) uracil paired with adenine (U-A), (viii) guanine (U-G) and (ix) T-G mispair.

In some preferred embodiments, the method further comprises contacting the sample with a lesion-free DNA further comprising a detectable marker which is released upon cleavage of the DNA and, measuring the variation of signal from the marker; wherein a decreased or increased activity compared to a reference indicates that the subject is at risk of radio toxicity.

In some preferred embodiments, the method further comprises obtaining a BER activity profile of the subject by combining the BER activities determined toward a panel of damaged DNA fragments comprising distinct lesions repaired by various BER pathway proteins.

In some embodiments, the radiotoxicity is of grade 2 and higher; and/or the radiotoxicity is of grade 3 and higher; and wherein the grade is according to the Common Terminology Criteria for Adverse Events scale.

In some embodiments, the sample is a blood sample, preferably a protein extract of PBMCs, preferably wherein the sample is collected before radiotherapy and/or during the first five sessions of radiotherapy, in particular the first and/or fifth radiotherapy session.

In some preferred embodiments, the BER activity toward DNA substrate comprising Hx-T or T-G prior to radiotherapy; toward DNA substrate comprising 8oxoG-C, Hx-T or T-G after the first session of radiotherapy; or toward DNA substrate comprising T-G after five sessions of radiotherapy allow to discriminate the group of breast cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3); preferably combination of BER activities toward DNA substrates comprising respectively Hx-T and T-G, after the first session of radiotherapy allows to discriminate the group of breast cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3).

In some preferred embodiments, the BER activity toward DNA substrate comprising Hx-T prior to radiotherapy; toward DNA substrate comprising EthA-T, Hx-T, THF-A or U-G substrate after the first session of radiotherapy; or toward DNA substrate comprising 8oxoG- C, Hx-T, T-G, Tg-A, THF-A or U-G after five sessions of radiotherapy allow to discriminate the group of prostate cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3); preferably combination of BER activities toward EthA-T and THF-A substrates after the first session of radiotherapy and BER activities toward DNA substrates comprising respectively EthA-T, Hx-T, Tg-A and THF-A, after five sessions of radiotherapy allows to discriminate the group of prostate cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3).

In some preferred embodiments, the BER activity toward DNA substrate comprising U-G prior to radiotherapy; toward DNA substrate comprising 8oxoG-C, Hx-T, T-G, THF-A or U- G after the first session of radiotherapy; toward DNA substrate comprising 8oxoG-C, Hx-T or T-G after five sessions of radiotherapy; or the difference of BER activity toward a DNA substrate comprising 8oxoG-C or U-A between the first session of radiotherapy and prior radiotherapy allow to discriminate the group of breast cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2); preferably the combination of BER activities toward DNA substrates comprising respectively EthA-T, Tg- A and U-G, prior to radiotherapy and BER activities toward DNA substrates comprising respectively 8oxoG-C, EthA-T, Hx-T, T-G, Tg-A, THF-A and U-G, after the first session of radiotherapy allow to discriminate the group of breast cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2).

In some preferred embodiments, the BER activity toward DNA substrate comprising T-G, U- A or U-G prior to radiotherapy; toward DNA substrate comprising THF-A, U-A or U-G after the first session of radiotherapy; or toward DNA substrate comprising T-G or U-A after five sessions of radiotherapy; or the DNA cleavage activity toward Lesion_Free_ODN after the first session of radiotherapy allow to discriminate the group of prostate cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2); preferably combination of BER activities toward DNA substrates comprising respectively 8oxoG-C, Hx-T, Tg-A, T-G and THF-A, after the first session of radiotherapy and BER activities toward DNA substrates comprising respectively T-G and THF-A, after five sessions of radiotherapy allow to discriminate the group of prostate cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2).

In some preferred embodiments, the BER activity toward DNA substrate comprising THF-A prior to radiotherapy; or toward DNA substrate comprising THF-A after the first or the fifth session of radiotherapy; or the difference of BER activity toward DNA substrate comprising U-G between the fifth and the second session of radiotherapy allow to discriminate the group of breast cancer patients presenting fibrosis from patients who do not; preferably combination of BER activities toward DNA substrates comprising respectively THF-A and U-G, after the first session of radiotherapy; BER activities toward DNA substrates comprising respectively THF-A and U-G, after five sessions of radiotherapy; and DNA cleavage activity toward Lesion_Free_ODN after five sessions of radiotherapy allow to discriminate the group of breast cancer patients presenting fibrosis from patients who do not.

DETAILED DESCRIPTION OF THE INVENTION

The inventors surprisingly showed that the analysis of Base Excision Repair (BER) activity allows the prediction of radiotoxicity risk in a subject. In particular, the inventors showed that the analysis of BER activity profile allows the prediction of radiotoxicity risk including late radiotoxicity risk and inflammation-induced fibrosis, with high specificity and sensitivity in a subject by combining biomarkers obtained from the subject’s samples collected before or at early stages of radiotherapy treatment.

The present invention relates to an in vitro method for predicting the risk of radio toxicity, in particular late radiotoxicity risk and inflammation-induced fibrosis of a subject, comprising determining the Base Excision Repair (BER) activity in a subject sample, and deducing therefrom if the subject is at risk (radiosensitive subject) or not at risk (normal subject) of radio toxicity. In some particular embodiments, the method comprises determining the BER activity profile in subject samples collected before and after few sessions of radiotherapy. As used herein, “radiotoxicity” refers to the clinical manifestation following exposure to a given dose of ionising radiation, in particular following radiotherapy. The radiotoxicity corresponds to any adverse effects such as unfavorable and involuntary sign, symptom or disease associated in time with ionising radiation. An adverse event is a unique representation of a specific event used for medical documentation and in scientific analyses. Clinical manifestation following irradiation can be early toxicity or late toxicity. Earlier toxicity occurs during or shortly after the completion of irradiation, in particular radiotherapy treatment and can be for example dermatitis, mucositis, cystitis, proctitis, hair loss, bone marrow suppression. Late toxicity manifests 6 months to several years after irradiation, in particular radiotherapy and can be for example fibrosis such as inflammation-induced fibrosis, tissue necrosis, atrophy, vascular damage, infertility, hormone deficiencies, gastrointestinal and genitourinary syndromes, and development of a second disease.

The CTCAE (Common Terminology Criteria for Adverse Effect) provides a brief definition of each adverse event in order to clarify the meaning of the adverse event. This scale, valid for other genotoxic stresses (for example: bum wounds) is particularly used in radiotherapy.

The grade refers to the severity of the adverse event. The CTCAE has 5 severity grades (from 1 to 5) with unique clinical severity descriptions for each adverse event, described in table 1 below. Each severity grade is defined by specific tissue reactions.

Table 1 represents the most recent version of the Common Terminology Criteria for Adverse Events (CTCAE) scale published by the National Cancer Institute of the United States of America on Nov. 27, 2017).

In some embodiments, the radio toxicity, in particular late radio toxicity, includes fibrosis such as inflammation-induced, in particular skin fibrosis.

The radiotoxicity may be from previous natural, accidental or medical exposure to ionising radiations or radio-mimetic chemicals. In some embodiments, said radiotoxicity is following accidental exposure to ionising radiations or radio-mimetic chemicals. Accidental refers in particular to radiation emergency, such as caused by radioactive contamination. In some other preferred embodiments, said radiotoxicity is from exposure to medical radiations (radiotherapy).

In particular embodiments, the severity of toxicity symptoms is classified according to grades. In particular embodiments, the method according to the invention allows predicting the risk of late radio toxicity, in particular following radiotherapy. In some preferred embodiments, the method according to the invention allows predicting the risk of radiotoxicity of grade 2 and higher, preferably late toxicity, in particular following radiotherapy. In some more preferred embodiments, the method according to the invention allows predicting the risk of radiotoxicity of grade 3 and higher, preferably late toxicity, in particular following radiotherapy, wherein the grade is according to the Common Terminology Criteria for Adverse Events (CTCAE) scale as disclosed above.

The term “predicting radiotoxicity” as used herein, refers to an ability to assess whether the exposure to ionising radiations, in particular the radiotherapy of a subject will induce or not adverse effects in the subject. In particular, such ability to assess whether the exposure to ionising, in particular the radiotherapy, will induce or not radiotoxicity is performed before and/or during exposure to ionising radiations, in particular before and/or during the first sessions of radiotherapy in the subject, in particular the first five sessions of radiotherapy in the subject. Predictive methods of the invention can be used clinically to make treatment decisions by choosing the most appropriate treatment modalities for any particular subject.

The term “radiotherapy” refers to multiple types of radiation therapy including internal and external radiation therapy, radioimmunotherapy, and the use of ionising radiation. Ionising radiation includes but is not limited to X-rays, gamma-rays, alpha particles, beta particles, photons, electrons, neutrons, radioisotopes. The radiotherapy is used to treat many types of cancer. Preferably, the radiotherapy involves the use of X-rays or gamma-rays. Said ionising radiation is defined by the absorbed dose (parameter called D and expressed in Gray). In the context of the present invention, the absorbed dose D per session is between 0.5 Gy and 4 Gy, preferably between 1 Gy and 3 Gy, more preferably between 1.7 Gy and 2.3 Gy, and even more preferably 2 Gy. These ranges typically correspond to an individual radiotherapy treatment session, the number of sessions depending upon the location, type and stage of advancement of the tumor.

The terms "subject" and "patient" are used interchangeably herein and refer to human.

“a”, “an”, and “the” include plural referents, unless the context clearly indicates otherwise. As such, the term “a” (or “an”), “one or more” or “at least one” can be used interchangeably herein; unless specified otherwise, “or” means “and/or”.

As used herein, the term “cancer” refers to any type of cancer, including solid or liquid cancers, primary cancers, cancer metastases and relapse of cancer. Cancers are classified by the type of cell that the tumor resembles and, therefore, the tissue presumed to be the origin of the tumor. For example, carcinomas are malignant tumors derived from epithelial cells. This group represents the most common cancers, including the common forms of breast, prostate, lung, and colon cancer. Lymphomas and leukemias include malignant tumors derived from blood and bone marrow cells. Sarcomas are malignant tumors derived from connective tissue or mesenchymal cells. Mesotheliomas are tumors derived from the mesothelial cells lining the peritoneum and the pleura. Gliomas are tumors derived from glia, the most common type of brain cell. Germinomas are tumors derived from germ cells, normally found in the testicle and ovary. Choriocarcinomas are malignant tumors derived from the placenta. Head and Neck cancers are tumors derived from the squamous cells that line the mucosal surface inside the head and neck. Preferably, a patient according to the invention is a human cancer patient, more preferably chosen from breast, prostate, head and neck or lung cancer, and melanoma patients.

To predict the risk of radiotoxicity in a subject, base excision repair (BER) activity, in particular BER activity profile is characterized in at least one sample of the subject.

Base excision repair (BER) is the major cellular pathway responsible for correcting small lesions that do not significantly distort the DNA structure such as damaged bases (base fragmentation, base methylation, oxidized bases, etheno-bases and the like) resulting from oxidation, alkylation, deamination, and other damage mechanisms and abasic sites. The initial steps of the BER pathway include sequential action of a substrate-specific DNA iV-glycosy lase that removes the target base by N-glycosidic bond hydrolysis and generates an abasic (apurinic/apyrimidic or AP) site, which is recognized by an AP endonuclease that hydrolyses the phosphodiester bond 5’ to the lesion. Some DNA glycosylases are endowed with an intrinsic AP-lyase activity that directly nicks DNA by P -elimination. The complete restoration of DNA integrity requires DNA polymerase, DNA ligase and a number of accessory proteins. Repair can proceed either by the short-patch BER pathway, which involves the incorporation of a single nucleotide into the gap by the DNA polymerase followed by strand ligation by DNA ligase, or by the long-patch BER pathway, which involves the incorporation of several nucleotides followed by cleavage of the resulting 2-to 8- nucleotide flap and ligation. A backup pathway for BER, nucleotide incision repair (NIR), bypasses the glycosylase step and starts with AP endonuclease cutting at damaged base-containing nucleotide.

Depending on the type of lesion, BER is initiated by distinct DNA glycosylases. Each glycosylase possesses its own repertoire of substrates; there is some overlap in substrate recognition between the different glycosylases, to ensure a robust repair (Kay J., et al. DNA Repair (2019), 83:102673). Abasic sites are repaired by AP endonuclease, poly(ADP-ribose) polymerase, and DNA polymerase beta, among others (Dianov GL., Mutat Res. 2003 Oct 29; 531(1-2): 157-63).

Cytosine deamination and incorporation of dUTP instead of dTTP during replication gives rise to U-A pairs. Spontaneous cytosine deamination results from the instability of the base and introduces U-G mismatch. Sequences containing dUTP instead of dTTP are rapidly recognized in mammalian cells by uracil DNA glycosylases (UDG) superfamily including UNG2, SMUG1, TDG and MBD4 enzymes. MBD4 is also able to remove T in the mispair T-G.

The most prevalent oxidative lesions are 8oxoG and the oxidized pyrimidine glycols.

8oxoG is due to the oxidation of G, and thus is generally paired with C. If not repaired, during replication, the polymerases insert an A opposite 8oxoG, leading to the A-8oxoG mispair. In humans, OGGI is in charge of removing the 8oxoG lesions; MUTYH is the specific enzyme that removes the adenine base when paired with 8oxoG.

Thymine and cytosine glycols are processed by the DNA glycosylases NTH1 and NEIL.

Another family of DNA glycosylases is known as HhH superfamily. Among the glycosylases belonging to this family only one, AAG glycosylase is present in human cells. This alkylbase DNA glycosylase recognizes and removes damaged bases such as deaminated adenine (inosine also called hypoxanthine) and cyclic ethenoadducts.

Upon recognition of a substrate and removal of the damaged base by one or several of the glycosylases mentioned above, an intermediate abasic site (AP site) is created. The AP sites are incised and processed by APE1 (also called HAP1 and Apex), the major AP-endonuclease present in humans (Krokan HE. & Bjpras M., Cold Spring Harb Perspect Biol. 2013; 5(4):a012583).

As used herein “Base Excision Repair (BER) activity” refers to the activity of the proteins of the BER pathway, including in particular the enzymes of said BER pathway such as the various DNA glycosylases and the AP endonuclease of the BER pathway as disclosed above. The BER activity present in a subject sample may be determined by using different assays that are well-known in the art such as oligonucleotide cleavage assays as described in WO 01/90408 and EP 1 283 911 Bl; Sauvaigo S. et al. Anal. Biochem., 2004, 1, 333, 182-92; Guemiou et al. Biochimie, 2005, 87, 151-159; Pons B. et al., Meeh. Aging Dev., 2010, 131, 11-12; Favrot C. et al. 2018, Oxid Med Cell Longev, 2018:5895439.

BER activity, in particular BER activity profile of a subject sample is characterized by contacting said sample with at least one damaged DNA that is repaired by the BER pathway, herein referred to as “BER substrate” or “BER matrix”, and analysing the BER activity of said sample by measuring the signal of a detectable marker present on the damaged DNA.

As used herein, the term “BER substrate”, “BER matrix”, “DNA substrate”, “repair matrix”, “substrate” or “matrix”, refers to a damaged single-strand or double-strand DNA comprising lesion(s) that are repaired by the BER pathway proteins. The lesion is advantageously incorporated into a DNA fragment and released (e.g. cleaved by excision or incision) during BER reaction. The BER substrate according to the invention comprises advantageously a detectable label or marker which is released upon cleavage of the lesion(s) by the base excision repair enzymes present in the sample. For ease of determination, the BER substrate is advantageously immobilized on a support and the marker is released or eliminated from the support during BER reaction. BER activity is then determined by the variation of signal from the marker and is proportional to the decrease of the signal from the marker.

As used herein, the term “sample” of a patient refers to any biological sample comprising cells obtained from said patient including any cell fraction or extract thereof. The biological sample may be body fluid, tissue, solid or liquid tumor sample. In particular embodiments, the sample is a blood sample, tissue sample or a tumor biopsy; preferably it is a blood sample. In particular embodiments, said sample is a cell extract comprising BER activity, preferably a cell lysate or protein extract, more preferably nuclear lysate, still more preferably a cell nuclear protein extract. The cell extract according to the present disclosure may be prepared from a fraction of isolated cells of the sample, in particular from Peripheral Blood Mononuclear Cells (PBMCs). In some preferred embodiments, the sample is a blood sample of patient diagnosed with cancer or protein extract thereof, preferably a cell protein or nuclear protein extract thereof, in particular a protein extract or nuclear protein extract of PBMCs. In the method according to the invention, the sample(s), preferably blood sample(s), more preferably from breast, head and neck, lung or prostate cancer and melanoma patient(s), may be collected before and/or after radiotherapy session(s), in particular first and/or fifth radiotherapy session of the cancer patient. The method of the invention does not comprise the irradiation of the biological test sample or culture of cells from the biological test sample prior to determining the BER activity.

To perform the method of the invention, the cells are advantageously treated to extract the proteins, in particular nuclear proteins which comprise DNA repair enzymes including BER enzymes. Nuclear proteins extract is prepared by methods well-known in the art. For example, cells are lysed to isolate the nucleus then the nucleus is lysed to extract nuclear proteins. Nuclear proteins extract are prepared according to methods described in in vitro transcription or in vitro DNA repair test such as described in Dignam JD. et al. (Nucleic Acids Research, 1983, 11:1475-1489); Zerivitzet K. and Akusjarvi G. (Gene Analysis Techniques, 1988, 5, 22-31); Hliakis G. et al. (Methods Mol. Biol., 2006, 314:123-31); Luo Y. et al. (BMC Immunol., 2014, 15:586). The nuclear extracts can be further fractionated or dialysed by conventional methods. All the steps of nuclear proteins extract are performed under conditions that do not impair enzymatic activities and proteins comprised in the cells, such as BER activity initially presents in the sample is conserved in nuclear extract analysed in the method of the invention.

In some particular embodiments, the sample comprises less than lOOpg/ml, preferably less than 90, 80, 70, 60, 50, 40, 30, 20; more preferably less than 15, 10, 1, 0.9, 0.8, 0.7, 0.6, 0.5pg/mL of protein. The concentrations disclosed herein refer to final protein concentrations in the repair reaction. In some preferred embodiments, the sample is a nuclear protein extract of a blood sample comprising less than 50pg/mL of proteins.

In some embodiments, the BER activity, in particular the BER activity profile of the sample is characterized by contacting the sample with at least one damaged DNA comprising lesion(s) that are repaired by the BER pathway proteins (e.g., a BER substrate) and further comprising a detectable marker which is released during BER reaction and the BER activity is determined by measuring the variation of signal from the marker. For example, such embodiments may be used to determine the activity of incision or excision enzymes towards the damaged DNA lesions. In this case, the marker may be present on one end of the damaged DNA. Thus, an incision or excision eliminates a damaged DNA fragment carrying the marker and causes loss of the signal. The ratio of the signal emitted by the marker present on one end of the damaged DNA before the DNA was repaired or in the absence of sample to the signal emitted by the marker after the BER reaction is correlated to the BER activity. In this case, the damage repair is evaluated by the disappearance of a signal specific to the damage.

Markers or labels that can be used to determine the BER activity may be of different types, provided that they emit a signal that can be detected or that can be revealed by emitting a detectable signal. Said marker or label can be a fluorescent compound, an antibody, a hapten, a biotin, a radioactive molecule, magnetic molecule or any other detectable molecule.

Said markers may also be introduced in-vitro at the extremity of the DNA fragment following the activity of incision or excision of repair enzymes on the damaged DNA. Such introduction can be done by exogenous enzymes such as terminal deoxynucleotidyl transferase or polymerase by supplying a labelled nucleotide or nucleotide analogue. Such labelling is known as TUNEL assay (terminal deoxynucleotidyl transferase (TdT) dUTP nick-end labelling assay; Figueroa-Gonzalez G. & Perez-Plasencia C. Oncol Lett. 2017 Jun; 13(6): 3982-3988) or nick-translation labelling (Christopher G.P. Mathew, 1983, Techniques in Molecular Biology, pp 159-16).

Said marker may also be an antibody, which reveals the lesion present on the damaged DNA. After lesion repair, said antibody is no longer fixed and a loss of signal of the marker representative of the activity of the repair protein may be observed. Thus, the ratio of the signal emitted by specific antibodies before the DNA was repaired (or in the absence of sample) to the signal emitted by the antibodies after the protein reaction, for example an enzymatic reaction, is correlated to the BER activity.

The signal of the marker present on the damaged DNA and excised during BER reaction, in particular fluorescent, radioactive, magnetic or colorimetric signal can be detected by any methods well-known in the art. Such methods include but are not limited to: fluorescent microscopy, flow cytometry, gammagraphy, nuclear magnetic resonance and mass spectrometry.

The damaged DNA according to the invention that serves as a repair matrix for BER enzymes present in the sample may be a short single or double-strand nucleic acid (less than 100 nucleotides).

In particular, the damaged DNA is a short oligonucleotide comprising between 15 to 100 nucleotides, preferably between 15 to 50 nucleotides as described in WO 01/90408; Pons B. et al., Meeh. Aging Dev., 2010, 131, 11-12; Candeias S., et al., Mutat. Res., 2010, 694, 53-9.

In particular embodiments, said damaged DNA is an oligonucleotide and the marker is present on the damaged DNA, preferably on one end of the damaged DNA.

The damaged DNA can be synthesized chemically to choose the exact sequences containing and surrounding the damage.

In preferred embodiments, the lesion is incorporated into a double-strand DNA fragment, preferably a short double-strand DNA fragment according to the present disclosure, and released during BER reaction as described in WO 01/90408. Preferably, the damaged DNA according to the method of the invention comprises at least one purine and/or pyrimidine lesion as defined above, in other terms the damaged DNA does not comprise single or doublestrand break. The damaged DNA according to the invention can be an abasic site. According to the invention, each damaged DNA comprises predefined lesion(s), different from those present in another damaged DNA (predefined and distinct lesions). Consequently, the damaged DNA lesions of each damaged DNA are previously characterized.

In some embodiments of the method according to the invention, the marker is present on a damaged DNA which comprises a base lesion or an AP site. In particular said base lesion is located on specific sites and allows the study of damage recognition and excision by glycosylases/AP endonuclease which belongs to the base excision repair (BER) pathway. The BER pathway enzymes are advantageously selected from the group consisting of UDG glycosylases UNG2, SMUG1, TDG and MBD4; OGGI, NTH1, NEIL, MUTYH and AAG glycosylases; APE1 endonuclease and combination thereof. In some particular embodiments, the BER pathway enzymes are DNA glycosylases selected from the group consisting of UDG glycosylases UNG2, SMUG1, TDG and MBD4; OGGI, NTH1, NEIL, MUTYH and AAG glycosylases and combination thereof.

In preferred embodiments, the damaged DNA comprises at least one lesion selected from the group consisting of: (i) 8-oxoG paired with cytosine (8oxoG-C), (ii) adenine paired with 8oxoG (A-8oxoG), (iii) thymine glycol paired with Adenine (Tg-A), (iv) hypoxanthine paired with thymine (Hx-T), (v) ethenoadenine (EthA) paired with thymine (EthA-T), (vi) abasic site analogue paired with adenine (THF-A), and a series of misrepairs (vii) uracil paired with adenine (U-A), (viii) guanine (U-G) and (ix) T-G mispair. The damaged DNA may comprise one or more of the above disclosed lesions (from 1 to 9 specific lesions). In particular embodiments, the damaged DNA comprises at least one lesion repaired by the above disclosed DNA glycosylase(s), preferably selected from the group consisting of: (i) 8-oxoG paired with cytosine (8oxoG-C), (ii) adenine paired with 8oxoG (A-8oxoG), (iii) thymine glycol paired with Adenine (Tg-A), (iv) hypoxanthine paired with thymine (Hx-T), (v) ethenoadenine (EthA) paired with thymine (EthA-T), and a series of misrepairs (vi) uracil paired with adenine (U-A), (vii) guanine (U-G) and (viii) T-G mispair.

These damaged DNA are disclosed in Sauvaigo S. et al. 2004, Anal Biochem; 333(1): 182-92; Favrot C. et al. 2018, Oxid Med Cell Longev, 2018:5895439.

These substrates are recognized and cleaved by specific glycosylases and endonucleases: 7,8-dihydro-8-oxo-guanine (8oxoG) paired with C (8oxoG-C) is a substrate of OGGI; adenine (A) paired with 8oxoG (A-8oxoG) is a substrate of MUTYH; thymine glycols (Tg) paired with A (Tg-A) is a substrate of NTH1 and NEIL; tetrahydrofuran (THF), paired with A (THF-A) is a substrate of APE1; hypoxanthine (Hx) paired with thymine (T) (Hx-T) is a substrate of AAG; ethenoadenine (EthA) paired with T (EthA-T) is a substrate of AAG;

- uracil (U) in front of G (U-G) or A (U-A) is a substrate of UNG2, SMUG1, TDG, MBD4; T-G mispair is a substrate of MBD4.

The method according to the above embodiment, advantageously quantifies the cleavage activities of UDG superfamily, including UNG2, SMUG1, TDG and MBD4, the OGGI, NTH1, NEIL, MUTYH, AAG glycosylases and the endonuclease activity of APE1 through the recognition and repair of specific substrates. The readout of the functional assay reflects the multiple processes that regulate BER and its signalling, at different molecular levels (genetic, epigenetic, transcription, splicing, messenger RNA stability, translation, protein stability, protein interactions, post-translation modifications,...) and the consequences of possible defects and dysregulations (Almeida KH. & Sobol RW. DNA Repair, 2007, 6, 695- 711; Carter RJ. & Parsons JL. Mol Cell Biol. 2016, 36, 1426-1437; Sevilya Z. et al. Carcinogenesis. 2015, 36, 982-991; Limpose KL. et al. DNA Repair, 2017, 56, 51-64).

In some embodiments, the method of the invention comprises characterizing the BER activity profile of a subject sample by contacting said sample with different damaged DNA, in particular comprising different lesions according to the present disclosure, and analysing the BER activity of the different damaged DNA by measuring a signal of a marker present on the damaged DNA. The results of the analysis of the different DNA repair activity profiles using different damaged DNA and markers can be used independently or combined to simultaneously identify BER defects associated with risk of radio toxicity.

In some preferred embodiments, the method of the invention further comprises obtaining a BER activity profile of the subject by combining the BER activities determined toward a panel of damaged DNA fragments comprising distinct lesions repaired by various BER pathway proteins. In some more preferred embodiments, the different damaged DNA comprise the 9 specific lesions (listed (i) to (ix)) recognized by specific glycosylases and endonucleases according to the present disclosure. In particular embodiments, the method according to the invention further comprises a step of contacting said sample with a lesion-free DNA and measuring the signal of the marker present on the lesion-free DNA.

In some preferred embodiments, the cleavage of lesion-free DNA is used as control to eliminate DNA cleavage background in the oligonucleotide cleavage assay, in particular to correct the non-specific cleavage activities which can be present in the sample of the subject.

The inventors also showed that the degradation of lesion-free DNA in particular Lesion_Free_ODN analysed by oligonucleotide cleavage assay is surprisingly high in certain patients.

Thus, in some other preferred embodiments, the cleavage activity of the subject sample toward lesion-free DNA, in particular analysed by oligonucleotide cleavage assay, is used to predict the risk of radiotoxicity of the subject.

In preferred embodiments, said damaged DNA and lesion-free DNA are oligonucleotides (ODN). In more preferred embodiments said ODN comprise a detectable marker on one end; more preferably a marker which is released upon cleavage of the ODN.

In more preferred embodiments, the method of the invention further comprises contacting the sample with a lesion-free DNA, preferably a lesion-free oligonucleotide, further comprising a detectable marker which is released upon cleavage of the DNA and, measuring the variation of signal from the marker; wherein a decreased or increased activity compared to a reference indicates that the subject is at risk of radio toxicity.

The DNA repair reaction is performed in conditions allowing lesions repair by BER enzymes in the sample which generally comprises recognition of the damage, a cleavage of the damaged base, and an incision of the DNA chain and elimination of a nucleotide or nucleic acid fragment. These conditions are well-known in the art and are described in WO 01/90408, Pons B. et al., Meeh. Aging Dev., 2010, 131, 11-12; and Candeias S. et al. 2010, Mutat Res, 694( 12-):53-59. The reaction is preferably at a temperature promoting DNA repair, preferably between 30°C and 37°C.

In some preferred embodiments, the damaged DNA is fixed on a solid support. The attachment of the damaged DNA is performed using conventional processes and can be done directly or via an intermediate molecule (biotin type antigen, nucleic acid, etc.) or a chemical function (NH₂, OH, etc).

Solid support is an appropriate support to immobilise the nucleic acid, in particular DNA. Said solid support can be glass support, polypropylene, polystyrene, silicone, metal, nitrocellulose and nylon. In some preferred embodiments, the solid support is a biochip.

The damaged DNA may also be fixed on the solid support by any methods compatible with maintaining the integrity of the damaged DNA fragment, for example by hybridising with another DNA fragment immobilised on the support, by immobilisation by an affinity molecule, or by direct fixation on the chip by deposition or by any other means.

The support may be for example a biochip comprising DNA fragments on which damaged DNA is fixed by hybridization using an oligonucleotide comprising a part complementary to the damaged DNA and a part complementary to the DNA fragment of the biochip.

A damaged DNA may be synthesised directly on the solid support by means of automatic synthesisers that can for example incorporate modified bases. Preferably, a biochip will be used as a solid support on which DNA fragments comprising different or identical but perfectly identified lesions, are fixed at determined locations. In some more preferred embodiments, the BER activity or profile is determined using an oligonucleotide cleavage assay (as disclosed in WO 01/90408) which determines the excision/incision enzyme activity of the BER pathway using a damaged DNA fragment immobilized on a solid support, said damaged DNA fragment comprising a first oligonucleotide comprising a lesion according to the present disclosure and a marker on one end (lesion-ODN or O) hybridized to an oligonucleotide (cHcO) comprising a part complementary to the lesion-ODN (O) and a part complementary to an oligonucleotide (H) immobilized on the support, wherein the marker is eliminated from the damaged DNA fragment by excision/incision of the lesion during the BER reaction. Preferably, the BER activity or profile is determined using at least one lesion- ODN and at least one lesion-free ODN according to the present disclosure. More preferably, the BER activity or profile is determined using a panel of lesion-ODNs comprising the 9 specific lesions (listed (i) to (ix)) recognized by specific glycosylases and endonucleases according to the present disclosure; and preferably further comprising a lesion-free ODN according to the present disclosure. In some embodiments, the method comprises determining the presence or absence of an altered BER activity in the subject sample, wherein the presence of an altered BER activity indicates that the subject is at risk of radio toxicity, in particular of late radiotoxicity and/or inflammation-induced fibrosis. The altered BER activity is determined by comparison with a reference.

In some preferred embodiments, the altered BER activity is a decreased or increased activity compared to a reference.

According to the present invention, the reference may be a predetermined value; a value obtained with control sample(s) from normal subject(s) tested simultaneously. The method may eventually comprise additional controls such as control sample(s) from radiosensitive subject(s), preferably of predetermined CTCAE grade, tested simultaneously; or a value from a subject sample tested at another time point.

As used herein, the term "predetermined value" refers to the BER activity level in biological samples obtained from a selected group or population of subjects who showed no adverse effect following radiotherapy, as determined using the Common Terminology Criteria for Adverse Events (CTCAE) scale, herein referred to as normal subjects. The predetermined value can be a threshold value, or a range. Typically, a "control value" or "threshold value" or "cut-off value" can be determined experimentally, empirically, or theoretically. A threshold value can also be arbitrarily selected based upon the existing experimental and/or clinical conditions, as would be recognized by a person skilled in the art. For example, the predetermined value can be established based upon comparative measurements between normal subject group who showed no adverse effect following radiotherapy and radiosensitive subject group who showed no adverse effect following radiotherapy, as determined using the Common Terminology Criteria for Adverse Events (CTCAE) scale. As shown in the examples, the comparison is performed with groups having different radiosensitivity according to CTCAE grade scale: grade 0 (normal); grade 1 (mild radiotoxicity/mild radiosensitivity); grade 2 (moderate toxicity /moderate radiosensitivity); grade >3 (severe radiotoxicity /high radiosensitivity) .

The normal and radiosensitive subjects are advantageously cancer patients, preferably affected by the same cancer type as the tested subject(s). In some embodiments, the subject has breast cancer or prostate cancer. In some preferred embodiments, a decreased BER activity in the breast cancer patient compared to a reference indicates that the subject is at risk of radiosensitivity. In some preferred embodiments, an increased BER activity in the prostate cancer patient compared to a reference indicates that the subject is at risk of radiosensitivity.

In preferred embodiments, the BER activity toward DNA substrate comprising Hx-T or T-G (represented as lesion-paired base) prior to radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for Hx-T and 15pg/mL of protein for T-G; toward DNA substrate comprising 8oxoG-C, Hx-T or T-G after the first session of radiotherapy; preferably with a sample comprising 15pg/mL of protein; or toward DNA substrate comprising T-G after five sessions of radiotherapy; preferably with a sample comprising 15pg/mL of protein; allow to discriminate the group of breast cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3). Preferably, combination of BER activities toward DNA substrates comprising respectively Hx-T and T-G, after the first session of radiotherapy; preferably with a sample comprising 15pg/mL of protein for Hx-T and 0.5pg/mL of protein for T-G; allows to discriminate the group of breast cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3). As shown in the examples, highly radiosensitive breast cancer patients (grade >3) have a decreased BER activity compared to less radiosensitive or normal breast cancer patients.

In other preferred embodiments, the BER activity toward DNA substrate comprising Hx-T prior to radiotherapy; preferably with a sample comprising 15pg/mL of protein; toward DNA substrate comprising EthA-T, Hx-T, THF-A or U-G substrate after the first session of radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for THF-A and U-G and 15pg/mL of protein for EthA-T and Hx-T; or toward DNA substrate comprising 8oxoG- C, Hx-T, T-G, Tg-A, THF-A or U-G after five sessions of radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for THF-A and U-G and 15pg/mL of protein for Hx- T, T-G, 8oxoG-C and Tg-A allows to discriminate the group of prostate cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3). Preferably combination of: (i) BER activities toward EthA-T and THF-A substrates after the first session of radiotherapy; preferably with a sample comprising 0.5pg/mL of protein and (ii) BER activities toward DNA substrates comprising respectively EthA-T, Hx-T, Tg-A and THF-A after five sessions of radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for THF-A and 15pg/mL of protein for EthA-T, Hx-T and Tg-A allows to discriminate the group of prostate cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3). As shown in the examples, highly radiosensitive prostate cancer patients have an increased BER activity compared to less radiosensitive or normal prostate cancer patients.

In other preferred embodiments, the BER activity toward DNA substrate comprising U-G prior to radiotherapy; preferably with a sample comprising 0.5pg/mL of protein; toward DNA substrate comprising 8oxoG-C, Hx-T, T-G, THF-A or U-G after the first session of radiotherapy; preferably with a sample comprising 15pg/mL of protein; or toward DNA substrate comprising 8oxoG-C, Hx-T or T-G after five sessions of radiotherapy; preferably with a sample comprising 15pg/mL of protein; or the difference of BER activity toward DNA substrate comprising 8oxoG-C or U-A between the first session of radiotherapy and prior radiotherapy; preferably with a sample comprising 0.5pg/mL of protein; allows to discriminate the group of breast cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2). Preferably the combination of: (i) BER activities toward DNA substrates comprising respectively EthA-T, Tg-A and U-G, prior to radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for U-G and 15pg/mL of protein for EthA-T and Tg-A and (ii) BER activities toward DNA substrates comprising respectively 8oxoG-C, EthA-T, Hx-T, T-G, Tg-A, THF-A and U-G after the first session of radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for U-G and 15pg/mL of protein for 8oxoG-C, EthA-T, Hx-T, T-G, Tg-A, THF-A allows to discriminate the group of breast cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2). As shown in the examples, moderately and severe radiosensitive breast cancer patients have a decreased BER activity compared to mildly radiosensitive or normal breast cancer patients.

In other preferred embodiments, BER activity toward DNA substrate comprising T-G, U-A or U-G prior to radiotherapy; preferably with a sample comprising 0.5pg/mL or 15pg/mL of protein for U-G and 15pg/mL of protein for T-G and U-A; toward DNA substrate comprising THF-A, U-A or U-G after the first session of radiotherapy; preferably with a sample comprising 0.5pg/mL or 15pg/mL of protein for U-G and THF-A and 15pg/mL of protein for U-A; or toward DNA substrate comprising T-G or U-A after five sessions of radiotherapy; preferably with a sample comprising 0.5pg/mL of protein for U-G and 15pg/mL of protein for T-G; or DNA cleavage activity toward Lesion_Free_ODN after the first session of radiotherapy; preferably with a sample comprising 15pg/mL of protein, allows to discriminate the group of prostate cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2). Preferably combination of : (i) BER activities toward DNA substrates comprising respectively 8oxoG-C, Hx-T, Tg-A, T-G and THF-A, after the first session of radiotherapy; preferably with a sample comprising 0.5 or 15pg/mL of protein for T-G, 0.5pg/mL of protein for Tg-A and 15pg/mL of protein for 8oxoG-C, Hx-T and THF- A; and (ii) BER activities toward DNA substrates comprising respectively T-G and THF-A, after five sessions of radiotherapy; preferably with a sample comprising 0.5pg/mE of protein for THF-A and 15pg/mE of protein for T-G; allows to discriminate the group of prostate cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2). As shown in the examples, moderately radiosensitive prostate cancer patients have an increased BER activity and an increased Eesion_Free_ODN degradation compared to mildly radiosensitive or normal prostate cancer patients.

In other preferred embodiments, BER activity toward DNA substrate comprising THF-A prior to radiotherapy; or toward DNA substrate comprising THF-A after the first or the fifth session of radiotherapy; or the difference of BER activity toward DNA substrate comprising U-G between the fifth and the second session of radiotherapy; preferably with a sample comprising 0.5pg/mE of protein for U-G and 15pg/mE of protein for THF-A; allows to discriminate the group of breast cancer patients presenting fibrosis from patients who do not. Preferably, combination of: (i) BER activities toward DNA substrates comprising respectively THF-A and U-G, after the first session of radiotherapy; (ii) BER activities toward DNA substrates comprising respectively THF-A and U-G after five sessions of radiotherapy, and (iii) DNA cleavage activity toward Eesion_Free_ODN after five sessions of radiotherapy; preferably with a sample comprising 0.5pg/mE of protein for U-G, 15pg/mE of protein for Eesion_Free_ODN and 0.5 or 15pg/mE of protein for THF-A after five sessions of radiotherapy; and 15pg/mE of protein for THF-A after the first session of radiotherapy; allows to discriminate the group of breast cancer patients presenting fibrosis from patients who do not. As shown in the examples, cancer patients presenting fibrosis have a decreased BER activity compared to patients who do not. In some embodiments, the method is a high throughput method comprising testing several subject samples simultaneously; preferably samples from different subjects; more preferably using a multiplex assay comprising a panel of damaged DNA substrates according to the present disclosure.

In some embodiments, the method comprises a further step of classifying the subject(s) into radiosensitive group or normal group based on the BER activity determined in the subject sample(s). This allows to identify patients at risk of developing serious adverse effects following radiotherapy, in particular inflammation-induced fibrosis, who will require closer monitoring to anticipate occurrence of adverse effects.

Depending on the risk of radiotoxicity in a subject, the physician can optimize the radiotherapy procedure to the subject, for instance by selecting the radiation dose and/or by selecting the most appropriate delivery means. The physician can also use another cancer therapy instead of radiotherapy. In a particular embodiment, the method according to the invention further comprises selecting the suitable irradiation dose for the subject. By “suitable irradiation dose”, it is intended a dose when applied to a subject for treating a cancer is sufficient to effect a treatment but not to induce adverse effects. The suitable dose will vary depending on the disease and its severity and the age, weight, physical conditions, concomitant treatment and responsiveness of the subject to be irradiated.

The method disclosed herein may further include administering/applying the suitable irradiation dose to the subject following the determination of the BER activity profile of the sample indicating the risk of radiotoxicity of the subject. If a subject sample is found to have a risk of developing radio toxicity, the subject may benefit from the administration of lower irradiation dose.

The invention encompasses the use of BER activity toward a damaged DNA or BER activity profile toward a panel of damaged DNA according to the present disclosure as a biomarker for predicting the risk of radiotoxicity of a subject.

The invention also relates to a method of treating cancer by radiotherapy, comprising: determining the risk of radiotoxicity of a patient using the prediction method according to the invention; and administering normal irradiation dose if the patient is not found at risk of radiotoxicity and administering a lower irradiation dose, another cancer therapy or a combination thereof if the patient is found at risk of radio toxicity.

The other cancer therapy may be chemotherapy, immunotherapy, surgery or a combination thereof.

The method according to the invention allows the identification of patients at risk of developing serious adverse effects following radiotherapy who will require closer monitoring to anticipate any occurrence of adverse effects.

The present invention will now be further described by the following non-limiting examples, with reference to the attached drawings in which:

FIGURE LEGENDS

Figure 1: Box-plots of the Lesion_ODNs cleavage rates for which significant differences between breast cancer patients suffering from radiation-induced moderate and severe late adverse effects (Yes category) and the others (No category) at different time points and time point combinations are shown, p: p value of the Mann-Whitney U test.

Figure 2. Box-plots of the Lesion_ODNs cleavage rates for which significant differences between breast cancer patients suffering from radiation-induced severe late adverse effects (Yes category) and the others (No category) at different timepoints are shown, p: p value of the Mann-Whitney U test.

Figure 3. Box-plots of the Lesion_ODNs cleavage rates for which significant differences between breast cancer patients suffering from radiation-induced fibrosis (Yes category) and the others (No category) at different timepoints and time point combinations are shown, p: p value of the Mann-Whitney U test.

Figure 4. Box-plots of the Lesion_ODNs cleavage rates, or residual Lesion_Free_ODN, for which significant differences between prostate cancer patients suffering from radiation- induced moderate or severe late adverse effects (Yes category) and the others (No category) at different timepoints are shown, p: p value of the Mann- Whitney U test.

Figure 5. Box-plots of the Lesion_ODNs cleavage rates for which significant differences between prostate cancer patients suffering from radiation-induced severe late adverse effects 1

(Yes category) and the others (No category) at different timepoints and time point combinations are shown, p: p value of the Mann- Whitney U test.

EXAMPLES

1. Material and methods

1.1 Sampling

A prospective study was conducted on a cohort of patients suffering from various types of cancers (breast cancer and prostate cancer, essentially). The patients suffered from non- metastatic cancer. The study was approved by ethical committee and all patients had given their informed consent for the study.

Blood sampling was performed before any treatment (JI). Typically, the patients received 2 Gy per dose of treatment. Two more blood samplings were performed after both the first (J2) and the fifth (J8) session of treatment. Blood samples of 8 mL were collected using BD vacutainer CPT™ tubes (reference BD362782; sodium citrate) at each time point. The peripheral blood mononuclear cells (PBMCs) were retrieved as described by the manufacturer.

The PBMCs were frozen in serum/DMSO (90/10% v/v) according to standard protocols used for cells cry opreservation. The cells were stored at -80°C.

Nuclear extracts were prepared from thawed cells as described in Pons B. et al. Meeh Ageing Dev. 2010; 131( 11- 12):661-5, with some minor modifications. Aliquots of 5pL were prepared and stored at -80°C until use.

The protein concentration in each sample was determined using the Pierce BCA Protein Assay Kit.

The cleavage rates for each lesion substrate, in each experimental condition, were established using a multiplex ODN cleavage assay, or multiplex ODN excision assay, as described in WO 01/90408 EP 1 283 911 B l and Glyco-SPOT at www.lxrepair.com. In addition the degradation of Lesion_Free_ODN was also considered. 1.2 ODN cleavage assay

For the excision assay, a panel of DNA duplexes, each containing a different lesion repaired by BER, were immobilized at specific sites on glass slides, forming 24 identical pads. On each pad, an oligonucleotide (ODN) without any lesion (Lesion_Free_ODN) and nine lesioncontaining ODN (Lesion_ODNs) were available in duplicate: 7,8-dihydro-8-oxo-guanine (8oxoG) paired with C (8oxoG-C), adenine (A) paired with 8oxoG (A-8oxoG), thymine glycols (Tg) paired with A (Tg-A), tetrahydrofuran (THF), as an AP site substrate equivalent, paired with A (THF-A), hypoxanthine (Hx) paired with thymine (T) (Hx-T), ethenoadenine (EthA) paired with T (EthA-T), uracil (U) in front of either G (U-G) or A (U-A) and T-G mispair. The lesion-containing ODNs and the Lesion_Free_ODN were labeled at their 5 ’end by a Cy3.

Each slide was incorporated into a microarray frame to materialize 24 wells (Grace Bio-labs Proplates).

Each extract was applied in 2 wells, for each final protein concentration condition. Each slide also comprised two wells incubated with the excision buffer only. Cleavage of the lesions by the enzymes contained in the extracts led to the elimination of the cleaved ODNs and to the release of the attached Cy3 fluorochrome.

The repair capacity of samples was expressed as the percentage of excision of each lesion compared to a reference. To calculate the percentage of excision of each lesion, the inventors measured the fluorescence of the spots in the wells incubated with the excision buffer only. These signals served as reference (100% fluorescence corresponding to the absence of cleavage of each substrate). An additional normalization step took into account the possible degradation of the Lesion_Free_ODN in each well incubated with the extract. The percentage of excision of each lesion was then calculated using the following formula: (100 x (1- percentage of fluorescence of Lesion_ODN/percentage of fluorescence of Lesion_Free_ODN)) .

This calculation of the residual Lesion_ODN after cleavage takes into account a possible degradation of the Lesion_Free_ODN by nucleases in the medium. The correcting factor that is typically applied takes the Lesion_Free_ODN degradation into account, to quantify only specific cleavage activities. The residual Lesion_Free_ODN was calculated using the formula: (100 x percentage of fluorescence of Lesion_Free_ODN after cleavage/percentage of fluorescence of Lesion_Free_ODN before cleavage). Typically, the degradation level of this Lesion_Free_ODN in biological samples is between 0 and 10%. Unexpectedly, the inventors found that in some blood samples of patients with cancer it could raise up to 40-80%.

As this degradation is attributed to the presence of non-specific nuclease and that nucleases are the hallmark of apoptosis, the inventors decided to consider the residual Lesion_Free_ODN as another biomarker present on the biochip.

The samples were tested in duplicate at 2 different final protein concentrations (0.5pg/mL and 15pg/mL) using the ODN multiplex cleavage assay. The blood samples from about 160 breast cancer patients and from about 50 prostate cancer patients were analyzed using the ODN multiplex cleavage assay.

The excision reactions were conducted at 30°C for 60 min in 80 mL of excision buffer (lOmM Hepes/KOH pH 7.8, 80mM KC1, ImM EGTA, O.lmM ZnCl₂, ImM DTT, 0.5mg/mL BSA). The slides were subsequently rinsed at room temperature (RT) 3x 5 min with 80 mL of washing buffer (PBS containing 0.2M NaCl and 0.1% Tween20) and dried for 10 min at 30°C. The fluorescence quantification was performed using the Innoscan 710 scanner from Innopsys (Toulouse, France). The results between the replicates (4 spots) were normalized using the Normalized software as described in Millau JF. et al. Lab Chip. 2008; 8:1713-22 and S. Sauvaigo S. et al., J. Invest. Dermatol. 2010; 130 1739-1741.

1.3 Clinical characterization of the adverse effects

The patients suffering from various drawback effects were graded according to the Common Terminology Criteria for Adverse Events (CTCAE) scale published by the National Cancer Institute of the United States of America on Nov. 27, 2017). Only toxicities that were directly linked to radiotherapy were considered. They concerned skin toxicity for breast cancer patients, genitourinary and gastrointestinal toxicity for prostate cancer patients. Only late toxicities, occurring beyond 6 months after the beginning of treatment were considered in the following analysis.

In breast cancer group, 101 patients had severity CTCAE grade 0; 47 had severity grade 1; 25 had severity grade 2; 3 had severity grade >3.

In prostate cancer group, 13 patients had severity CTCAE grade 0; 37 had severity grade 1; 16 had severity grade 2; 4 had severity grade >3. 2. Data analysis

Each sample was characterized by the cleavage rate obtained for the 9 lesions and by the residual Lesion_Free_ODN at 0.5pg/mE and 15pg/mE final protein concentrations. For each lesion, cleavage rates were considered at JI, J2, J8 and also the combinations (differences between the cleavage rates) between these days (J2-J1, J8-J1 and J8-J2) were taken into account.

2.1 Box plots

ODN multiplex cleavage assay values collected at different time-points (or their differences) were represented in box-plots. Patients were grouped according to the presence of moderate and severe (grade >2) or only severe (grade>3) late adverse effects for both breast and prostate cancers. In a supplementary analysis, breast cancer patients were grouped according to the presence of fibrosis. Mann-Whitney U tests were performed for each lesion to test if the cleavage rates in patients who developed severe, moderate late adverse effects or fibrosis compared to patients who did not were different. A p-value cutoff of 0.05 was used to select the most significant biomarkers.

To facilitate an understanding of results, a series of contractions were used. The name of each lesion is followed by “-5” or “-15” indicating the proteins concentration of 0.5 and 15pg/mE used for the ODN multiplex cleavage assay, respectively.

2.2 Classifiers

The set of cleavage rates obtained were used to train Random Forest models with stratified sampling to try to predict the existence of adverse effects by using the ODN multiplex cleavage data with different combinations of days ((JI), (J2), (J8), (JI, J2, J2-J1), (JI, J8, J8- Jl), (J2, J8, J8-J2), or (JI, J2, J8, J2-J1, J8-J1, J8-J2)). In addition to the obtained cleavage rates, all clinical available variables (Sex, Age, Height, Weight, Total Dose, ...) were included to run the Random Forest. The Boruta algorithm (with 100000 trees and a maximum of 1000 iterations) was used to select the most significant variables (Kursa M., Rudnicki W., 2009, Journal of Statistic Software, 36:29870). Then using the Boruta- selected variables, the confusion matrix computed (at the 0.5 threshold) over the repeated cross validations (entries are the average percentages of the held-out donors) was obtained for each combination of days. R software packages were used for all data and statistical analysis.

3. Results

3.1 Breast cancer population: comparison of late adverse effects of both moderate and severe grade with the other grades.

3.1.1. Moderate/severe grade (Grade >2) versus other grades (Grade <2)

The Mann- Whitney U test was used to compare the cleavage rates of the Lesion_ODNs in breast cancer patients suffering from moderate or severe late adverse effects (grade >2) with the Lesion_ODNs cleavage rates in breast cancer patients who did not suffer from these adverse effects (grades <2).

Lesion_ODNs which cleavage rate significantly differentiate the two groups (p value < 0.05) are shown in Table 2 below and as Box-Plots in Figure 1:

Table 2. The most predictive discriminative repair activities, identified by their substrates, that differentiated the late grades >2 group from the other group are summarized.

By using the Boruta-selected variables (U-G-5_J2, U-G-5_J1, EthA-T-15_J2, EthA-T-15_Jl, T-G-15_J2, 8oxoG-C-15_J2, THF-A-15_J2, Tg-A-15_J2, Tg-A-15_Jl, and Hx-T-15_J2) at the combination of days (JI), (J2), (J2-J1), the method of the inventors was able to calculate the appearance of moderate late adverse effects with a specificity of 82.5% whereas sensitivity of this test was 59%. The negative predictive value of detection was 90.5%.

3.1.2 Severe grade (Grade >3) versus other grades (Grade <3)

To identify the patients who developed severe late adverse effects following radiotherapy in breast cancer patients the ODN multiplex cleavage assay was performed at two protein concentrations, 0.5 and 15|lg/mL. The Mann- Whitney U test was used to determine if the cleavage rates of lesions in breast cancer patients suffering from severe adverse effects (grade >3) were significantly different from the cleavage rates of lesions in breast cancer patients suffering from less severe adverse effects (grades <3).

Significant results (p value < 0.05) are shown as Box-Plots in Figure 2. Mean ± standard deviation of the cleavage rates, and the median values are summarized in Table 3 below.

Table 3. The most predictive discriminative repair activities, identified by their substrates that differentiated the late grade >3 group from the other group are summarized. The ability of the cleavage assay to predict (predictive value) the appearance of severe late adverse effects in a cohort of 121 breast cancer patients with a prevalence of 2.5% was calculated with the Boruta- selected variable Hx-T-15_J2. All patients who had a severe late adverse effect were classified as positives showing an assay sensitivity of 100% and a negative predictive value of 100%. The specificity was 68.5%.

To improve the predictivity performance of the test with the best markers identified, the inventors processed to a second round of patients selection using the T-G substrate. The inventors identified the patients classified within the first quarter of all patients on the basis of Hx-T-15_J2 excision rate values (using increasing value order), then the patients classified within the first quarter of all patients on the basis of T-G-5_J2 excision rate values (using increasing value order). By combining the patients that were positive for both criteria, the inventors determined a new classification: this time, among 147 samples, 27 were identified as positive, 3 were true positive, 120 were identified as negative and 24 were false positive.

Using these 2 markers (Hx-T-15_J2 and T-G-5_J2), the assay had a sensitivity of 100%, a negative predictive value of 100% and a specificity of 83%.

3.1.3. Fibrosis

The Mann- Whitney U test was used to test the differences between the cleavage rates of lesions in breast cancer patients suffering from fibrosis compared to the cleavage rates of lesions in breast cancer patients who did not developed fibrosis.

The most significant results (p value < 0.05) are shown in Figure 3. Mean ± standard deviation and the median values are summarized in Table 4 below.

Table 4. The most predictive discriminative repair activities, identified by their substrates that differentiated the fibrosis group from the other group are summarized. By using the Boruta- selected variables (THF-A-5_J8, U-G-5_J8, LesionFree-15_J8, THF-A- 15_J8, THF-A-15_J1, U-G-5-J1) at the combination of days (JI), (J8), (J8-J1), the method of the inventors was able to calculate the appearance of fibrosis with a specificity of 80.0% whereas sensitivity of this test was 61.9%. The negative predictive value of detection was 95.1%.

As a conclusion, the DNA repair cleavage assay performed before the treatment and/or after two and/or five sessions of radiotherapy allows to discriminate among breast cancer patients who are susceptible of develop radiation-induced severe adverse late effects with 100% sensitivity (grade >3

grade <3). Different combinations of selected biomarkers allow discriminating between patients with moderate and severe late adverse effect from patients who did not show these adverse effects (grade >2

grade <2). A different combination of biomarkers was founded to be discriminant between patients who developed a fibrosis and patients who did not.

3.2 Prostate cancer population: comparison of late adverse effects of both moderate and severe grade with the other grades.

3.2.1 Moderate and severe grades (Grade >2) versus other grades (Grade <2)

To identify patients susceptible to develop severe late adverse effects following radiotherapy in a cohort of prostate cancer patients, the inventors performed the ODN cleavage assay at two protein concentrations, 0.5 and 15|lg/inL as explained above. The Mann-Whitney U test was used to calculate the significance between the mean cleavage rates of the lesions for the prostate cancer patients suffering from moderate or severe adverse effects (grade >2) compared to the mean cleavage rates of the lesions for the prostate cancer patients who showed no adverse effects or very mild adverse effects (grades <2).

The most significant results (p value < 0.05) are shown in the box-plots in Figure 4. The Mean ± Standard deviation for excision rates and residual Lesion_Free_ODN, and the Median values are shown in the Table 5 below.

Table 5. The most predictive discriminative repair activities, identified by their substrates that differentiated the late grades >2 group from the other groups are summarized.

In this cohort composed of prostate cancer patients our method was able to predict the appearance of moderate late adverse effects. By using the Boruta-selected variables (T-G-

5_J2, THF-A-5J8, Tg-A-5_J2, T-G-15.J2, T-G-15.J8, 8oxoG-C-15_J2, THF-A-15J2, and Hx-T-15_J2) at the combination of days (J2), (J8), (J8-J2), the ODN cleavage assay was able to discriminate patient who did not suffer from moderate severe adverse effects with a specificity of 90.5% and the positive predictive of 80.4%. The sensitivity of this test was 68% and the negative predictive value of detection was 83%.

3.2.2 Severe grade (Grade >3) versus other grades (Grade <3)

As done previously, the ODN cleavage assay was performed at two protein concentrations, 0.5 and 15pg/mL to predict which patient would develop severe late adverse effects following radiotherapy in prostate cancer patients. The Mann-Whitney U test was used to calculate the significance between the mean cleavage rates of the lesions for the prostate cancer patients suffering from severe adverse effects (grade >3) compared to the mean cleavage rates of the lesions for the prostate cancer patients suffering from less severe adverse effects or from no effect (grades <3). The most significant results are shown (p value < 0.05) in Figure 5. Mean ± standard deviation and median values are shown in the Table 6 below.

Table 6. The most predictive discriminative repair activities, identified by their substrates, that differentiated the late grade >3 group from the other groups are summarized The predictive value of the cleavage assay to predict the appearance of severe late adverse effects in a cohort of 53 prostate cancer patients was calculated with the combination of days (J2), (J8) and (J8-J2). In this population the Boruta-selected variables were EthA-T-5_J2, THFA-5_J2, THF-A-5_J8, EthA-T-15_J8, Tg-A-15_J8 and Hx-T-15_J8. All patients who had a severe late adverse effect were classified as positives showing sensitivity and a negative predictive value of 100%. The positive predictive value was 44% and the specificity was 87%.

Claims

37 CLAIMS

1. An in vitro method for predicting the risk of radiotoxicity of a subject, comprising determining the Base Excision Repair (BER) activity in a subject sample, and deducing therefrom if the subject is at risk or not of radio toxicity.

2. The method of claim 1, wherein the radiotoxicity is from radiotherapy.

3. The method of any preceding claim, wherein the radiotoxicity is late radiotoxicity and/or includes inflammation-induced fibrosis.

4. The method of any preceding claim, wherein the subject is a cancer patient, preferably chosen from breast, head and neck, lung or prostate cancer and melanoma patient.

5. The method of any preceding claim, wherein the method comprises determining the presence or absence of an altered BER activity in the subject sample, wherein the presence of an altered BER activity indicates that the subject is at risk of radio toxicity; preferably wherein the altered BER activity is a decreased or increased activity compared to a reference.

6. The method of any preceding claim, wherein the BER activity is determined by: contacting the sample with at least one damaged DNA comprising lesion(s) that are repaired by BER pathway proteins and further comprising a detectable marker which is released during BER reaction, and measuring the variation of signal from the marker; preferably wherein the BER activity involves at least one enzyme selected from the group consisting of UDG glycosylases UNG2, SMUG1, TDG and MBD4; OGGI, NTH1, NEIL, MUTYH and AAG glycosylases; APE1 endonuclease and combination thereof.

7. The method of claim 6, wherein the damaged DNA comprises at least one lesion selected from the group consisting of: (i) 8-oxoG paired with cytosine (8oxoG-C), (ii) adenine paired with 8oxoG (A-8oxoG), (iii) thymine glycol paired with Adenine (Tg-A), (iv) hypoxanthine paired with thymine (Hx-T), (v) ethenoadenine (EthA) paired with thymine (EthA-T), (vi) abasic site analogue paired with adenine (THF-A), (vii) uracil paired with adenine (U-A), (viii) guanine (U-G) and (ix) T-G mispair.

8. The method of any one of claim 6 or 7, further comprising contacting the sample with a lesion-free DNA further comprising a detectable marker which is released upon cleavage of 38 the DNA, and measuring the variation of signal from the marker; wherein a decreased or increased activity compared to a reference indicates that the subject is at risk of radio toxicity.

9. The method of any one of claims 6 to 8, further comprising obtaining a BER activity profile of the subject by combining the BER activities determined toward a panel of damaged DNA fragments comprising distinct lesions repaired by various BER pathway proteins.

10. The method of any preceding claim, wherein the radiotoxicity is of grade 2 and higher; and/ or the radiotoxicity is of grade 3 and higher; and wherein the grade is according to the Common Terminology Criteria for Adverse Events scale.

11. The method of any preceding claim, wherein the sample is a blood sample, preferably a nuclear protein extract of PBMCs, preferably wherein the sample is collected before radiotherapy and/or during the first five sessions of radiotherapy, in particular the first and/or the fifth radiotherapy session.

12. The method of any preceding claim, wherein the BER activity toward DNA substrate comprising Hx-T or T-G (represented as lesion-paired base) prior to radiotherapy; toward DNA substrate comprising 8oxoG-C, Hx-T or T-G after the first session of radiotherapy; or toward DNA substrate comprising T-G after five sessions of radiotherapy allow to discriminate the group of breast cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3); preferably combination of BER activities toward DNA substrates comprising respectively Hx-T and T-G, after the first session of radiotherapy allows to discriminate the group of breast cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3).

13. The method of any one of claims 1 to 11, wherein the BER activity toward DNA substrate comprising Hx-T prior to radiotherapy; toward DNA substrate comprising EthA-T, Hx-T, THF-A or U-G substrate after the first session of radiotherapy; or toward DNA substrate comprising 8oxoG-C, Hx-T, T-G, Tg-A, THF-A or U-G after five sessions of radiotherapy allow to discriminate the group of prostate cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3); preferably combination of BER activities toward EthA-T and THF-A substrates after the first session of radiotherapy and BER activities toward DNA substrates comprising respectively EthA-T, Hx-T, Tg-A and THF-A, after five sessions of radiotherapy allows to discriminate the group of prostate cancer patients presenting late severe radiotoxicity (grade >3) from the other group (grade <3).

14. The method of any one of claims 1 to 11, wherein the BER activity toward DNA substrate comprising U-G prior to radiotherapy or toward DNA substrate comprising 8oxoG- C, Hx-T, T-G, THF-A or U-G after the first session of radiotherapy; toward DNA substrate comprising 8oxoG-C, Hx-T or T-G after five sessions of radiotherapy; or the difference of BER activity toward DNA substrate comprising 8oxoG-C or U-A between the first session of radiotherapy and prior radiotherapy allow to discriminate the group of breast cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2); preferably the combination of BER activities toward DNA substrates comprising respectively EthA-T, Tg-A and U-G, prior to radiotherapy and BER activities toward DNA substrates comprising respectively 8oxoG-C, EthA-T, Hx-T, T-G, Tg-A, THF-A and U-G, after the first session of radiotherapy allow to discriminate the group of breast cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2).

15. The method of any one of claims 1 to 11, wherein the BER activity toward DNA substrate comprising T-G, U-A or U-G prior to radiotherapy; toward DNA substrate comprising THF-A, U-A or U-G after the first session of radiotherapy; or toward DNA substrate comprising T-G or U-A after five sessions of radiotherapy; or the DNA cleavage activity toward Lesion_Free_ODN after the first session of radiotherapy allow to discriminate the group of prostate cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2); preferably combination of BER activities toward DNA substrates comprising respectively 8oxoG-C, Hx-T, Tg-A, T-G and THF-A, after the first session of radiotherapy and BER activities toward DNA substrates comprising respectively T-G and THF-A, after five sessions of radiotherapy allow to discriminate the group of prostate cancer patients presenting late moderate and severe radiotoxicity (grade >2) from the other group (grade <2).

16. The method of any one of claims 1 to 11, wherein the BER activity toward DNA substrate comprising THF-A prior to radiotherapy; or toward DNA substrate comprising THF- A after the first or the fifth session of radiotherapy; or the difference of BER activity toward DNA substrate comprising U-G between the fifth and the second session of radiotherapy allow to discriminate the group of breast cancer patients presenting fibrosis from patients who do not; preferably combination of BER activities toward DNA substrates comprising respectively THF-A and U-G, after the first session of radiotherapy; BER activities toward DNA substrates comprising respectively THF-A and U-G, after five sessions of radiotherapy; and DNA cleavage activity of Lesion_Free_ODN after five sessions of radiotherapy allow to discriminate the group of breast cancer patients presenting fibrosis from patients who do not.