In vitro radiation-induced expression of XPC mRNA as a possible biomarker for developing adverse reactions during radiotherapy

Authors


Abstract

Repair of radiation-induced DNA damage is believed to play a critical role in developing adverse reactions during radiotherapy. Ionizing radiation induces transcription of several DNA repair genes including XPC as a part of the p53-transmitted stress response. XPC gene induction was measured to analyze whether it predicts occurrence of therapy-related acute side effects. Prostate cancer patients (n = 406) receiving radiotherapy were monitored for development of acute adverse effects using common toxicity criteria. For gene induction analysis, lymphocytes from 99 patients were selected according to their observed grade of clinical side effects. Cells were irradiated in vitro with 5 Gy and analyzed after 4 hr for XPC gene induction using reverse transcription and quantitative real-time PCR. Analysis of modulation of XPC induction by personal, clinical or lifestyle factors was included. Inter-individual induction of XPC expression by ionizing radiation varied up to 20-fold (0.29–5.77) and was significantly higher in current or exsmokers than in never-smokers (p value: 0.008). Patients with XPC induction above the 90th percentile compared to those with lower induction levels were at increased risk of suffering from adverse reactions during radiotherapy (odds ratio 5.3, 95% confidence interval 1.2–24.5; adjusted for smoking). In summary, XPC mRNA levels induced by ionizing radiation were shown for the first time to be strongly affected by smoking and to be associated with an approximately 5-fold increased risk for developing acute side effects of radiotherapy. The predictive value of DNA damage-induced XPC levels as a possible biomarker for radiosensitivity has to be further investigated. © 2007 Wiley-Liss, Inc.

Ionizing radiation (IR) is an important tool in cancer therapy; its application, however, is limited by side effects developing in the irradiated normal tissue. Interindividual variation in sensitivity of normal tissue to IR in cancer patients is high and can not be reliably predicted prior to therapy.1, 2, 3, 4, 5 Thus, for an effective killing of tumor cells during therapy, it is accepted that 5–10% of all radiotherapy patients suffer from severe acute or late side effects.3, 6 Developing predictive assays of radiosensitivity suitable for clinical use would allow to adjust radiotherapy for patients with different sensitivities7 and consequently to improve the therapeutic ratio.8

The mechanisms for hypersensitive reactions to IR remain unclear as they include both endogenous and exogenous factors.9, 10 Among those, hereditary factors are known to influence the cellular response to IR and patient-to-patient variability has been estimated to depend for up to 70% on genetic factors.3, 5, 11, 12 Different processes appear to be involved in the cellular response to IR like generation and scavenging of free radicals as well as DNA repair, apoptosis and inflammation. However, repair of DNA damage after IR is believed to be a major protective mechanism as nuclear DNA is the most susceptible target in living cells.13 This supposition is also based on several genetic syndromes with defects in DNA repair which share an extremely enhanced radiosensitivity, such as Ataxia telangiectasia or the Nijmegen Breakage Syndrome.14

DNA repair after IR exposure is controlled by mechanisms inducing mRNA expression of various DNA repair genes. Several reports showed expression changes in cells after irradiation under different experimental conditions, mainly performed in cultured cells and measured by array technology.15, 16, 17, 18, 19, 20 Expression changes after in vivo irradiation of cancer patients21 or chronic radiation exposure of radiation workers,22 measured in peripheral blood lymphocytes, have also been described. Investigations on the induction of gene expression in patients scheduled for radiotherapy and on the association with IR sensitivity are, however, lacking.

The XPC gene is one of the transcriptionally regulated DNA repair genes induced by both UV and IR16, 23 as part of the p53-transmitted stress response.24, 25 Further DNA repair genes induced by IR include CDKN1A, PCNA, GADD45A and DDB2.17, 23 The XPC protein is involved as an early factor in nucleotide excision repair (NER) where it plays a role in damage recognition.26 XPC in association with its heterodimer partner hHR23B mainly recognizes bulky, helix-distorting DNA adducts during global genome repair. These adducts can result from exposures to e.g. UV-light or polycyclic aromatic hydrocarbons present in tobacco smoke but also to reactive oxygen species generated by IR, e.g. purine cyclodeoxynucleosides.27 Although regulation, DNA lesion binding and DNA repair properties of XPC have been partially characterized,25, 26, 28 there is only little information available on basal and inducible XPC mRNA levels, especially in primary human cells. There is, in addition, increasing evidence that this gene plays an important and multifaceted role in protection of cells from oxidative and radiation-induced DNA damage.29

In our pilot study, we therefore investigated induction of XPC mRNA expression in peripheral blood lymphocytes of prostate cancer patients after γ-irradiation in vitro using reverse transcription of mRNA and quantitative real-time PCR. We further analyzed possible clinical and personal confounders in these patients such as age, body mass index or smoking that could modulate XPC induction. Finally, XPC induction was correlated with the incidence of acute side effects, which the patients developed during or directly after radiotherapy. These analyses were performed as a nested case–control study within a prospective clinical study in prostate cancer patients that was aimed to investigate the role of DNA repair in the development of radiotherapy related side effects.

Material and methods

Study subjects

Between 2001 and 2005, 406 unselected prostate cancer patients receiving radiotherapy at the Department of Clinical Radiology at the University of Heidelberg were enrolled into a prospective epidemiological study to evaluate determinants of acute toxicity associated with radiotherapy. Only patients were included. The study was approved by the Ethical Committee of the Medical Faculty at the University of Heidelberg (Ref. 206/2000). Patients gave informed written consent to take part in this study. Data on histopathology of the tumour, therapy and radiation-related side effects were collected from patients' hospital records. Demographic and epidemiologic information was collected using a self-administered patient questionnaire. A blood sample was taken from all patients prior to radiotherapy. Here, we present data of a nested case–control study including 99 patients of the overall study cohort. A detailed epidemiological description of study design and the complete cohort will be published elsewhere.

All patients were treated by 3D conformal radiotherapy with a target dose of 61–72 Gy. Adverse side effects were weekly documented for a period of 6 weeks using a modified system of the common toxicity criteria (CTC), version 2.0 of the US National Institute of Health (http://ctep.cancer.gov/reporting/ctc.html). The following grades were used for documentation: 0 (no adverse event), 1 (mild adverse event), 2 (moderate adverse event), 3 (severe adverse event), 4 (life-threatening or disabling adverse event). Patients were checked for acute side effects in different categories including blood and bone marrow (haemoglobin, platelets), gastrointestinal tract (nausea, vomiting, diarrhea, constipation, proctalgia, hemorrhage) and urogenital tract (urinary urgency/frequency, dysuria, hematuria, incontinence, bladder spasms). Only diarrhea and urinary urgency were observed as side effects in our study cohort. For this analysis, patients (n = 31) showing CTC grade 3 or 4 side effects in these two categories were considered as hypersensitive and comprised the “case” group. Seventeen (55%) of these patients suffered severe diarrhea and 18 (59%) patients urinary urgency. Four patients (13%) showed both symptoms. The “control” group consisted of age-matched patients with normal acute clinical sensitivity to radiotherapy (CTC score 0, 1 or 2; n = 68).

Tumor characteristics such as tumor size, involvement of lymph nodes and metastases showed no differences between hypersensitive and normal sensitive patients. The type of therapy (prostatectomy or adjuvant hormone therapy) had no influence on patients' radiosensitivity (clinical data not shown). No significant differences between hypersensitive and normal sensitive patients were observed for the parameters body mass index, family status and alcohol consumption (Table I). The self-administered questionnaire included detailed information on lifetime smoking habits. The participants were asked when they began smoking, the type of product consumed, the amount smoked daily and the date of cessation or change in their smoking habits. In the case of a change, the same questions were asked again for the following phase, allowing for up to three different phases. Exsmokers were persons who reported to have smoked regularly in the past, and current smokers had reported regular smoking at time of completion of the questionnaires. In addition, pack-years were computed as the average number of packs per day times the total number of years of smoking, assuming 20 cigarettes per pack. Patients were stratified into two groups according to the median of the variable pack-years: (i) never or light smokers (0–5 pack-years) and (ii) heavy smokers (≥6 pack-years). Light smokers seemed to be more sensitive to radiation than heavy smokers as 62% of all hypersensitive, but only 41% of all normal sensitive patients were in the group of light smokers. This difference in distribution was, however, not statistically significant (p = 0.06, Table I).

Table I. Characteristics of the Study Cohort
CharacteristicsClinical radiosensitivity
Normal sensitive n = 68 (%)hypersensitive n = 31 (%)p value
  • 1

    Fisher's exact test.

  • 2

    Data missing for 6 patients.

  • 3

    Data missing for 7 patients.

Age (years)
 Mean ± SD66.5 ± 7.867.3 ± 6.9 
 Range48–7950–79 
Age group (years)
 <6013 (19)3 (10) 
 60–6410 (15)7 (23) 
 65–6917 (25)10 (32) 
 70–7418 (27)7 (23) 
 ≥7510 (15)4 (13)0.641
Smoking status2
 Never25 (40)17 (57) 
 No smoking in the last 20 years21 (33)4 (13) 
 Smoking in the last 20 years17 (27)9 (30)0.11
Tobacco consumption3
Pack-years, median
 0–526 (41)18 (62) 
 ≥637 (59)11 (38)0.06
Body mass index (kg/m2)
 Normal weight (<25)18 (27)8 (26) 
 Overweight (≥25–29.9)37 (54)17 (55) 
 Obesity (≥30)13 (19)6 (19)1.00
Family status
 Single9 (13)1 (3) 
 Married59 (87)30 (97)0.171

Lymphocyte isolation

A venous blood sample of 50 ml was obtained in EDTA-layered monovettes from each patient prior to radiotherapy. Peripheral blood lymphocytes were isolated through gradient centrifugation by overlaying whole blood on Lymphoprep™ (Axis-Shield, Oslo, Norway) in Leucosep® tubes (Greiner, Frickenhausen, Germany), centrifugation at 800g for 15 min at room temperature, removal of lymphocytes containing layers, and 2 washing steps with phosphate-buffered saline. Aliquots with 10 × 106 lymphocytes were frozen at a rate of −1°C/min in cryomedium (RPMI 1640 medium (Invitrogen, Karlsruhe, Germany) containing 50% of heat inactivated fetal calf serum (PAA, Pasching, Austria) and 10% dimethylsulfoxide) and stored at liquid nitrogen until use. Lymphocytes isolated from a buffy-coat of a male healthy individual (obtained from the blood bank of the University of Heidelberg) served as a standard control.

Irradiation of lymphocytes in vitro

Lymphocytes of selected patients were irradiated in vitro and analyzed for induction of gene expression. Two aliquots of lymphocytes per patient were thawed and incubated at 37°C and 5% CO2 in 1 ml culture medium per 106 cells. Culture medium was RPMI 1640 supplemented with 2 mM L-glutamine (Invitrogen), 10% fetal calf serum, 100 U/ml penicillin and 100 μg/ml streptomycin (Invitrogen). After 72 hr of incubation, cells were pelleted by centrifugation and resuspended in a density of 105 cells/ml of fresh culture medium supplemented with 10 mM HEPES-buffer (Invitrogen). The first aliquot of cells was not irradiated; the second aliquot was irradiated with 5 Gy of 137Cs γ-rays at a dose rate of 10.1 Gy/min (Gammacell 1000, MDS Nordion, Ottawa, Canada). After irradiation, both the irradiated and the nonirradiated cell samples were incubated for 4 hr at 37°C prior to RNA extraction. Experimental conditions were optimized in separate experiments with lymphocytes from a healthy donor (data not shown). Induction of XPC after irradiation was already visible after 2 Gy and 2 hr of repair interval, but we selected 5 Gy and a 4 hr interval as these conditions showed the most wide spread radiation response which allowed a clear distinction between high and low responders and processing of cells from irradiation to RNA isolation within one experimental day. Samples were analyzed without knowledge of their clinical status.

RNA isolation

Total RNA was isolated by a modified single-step method30 using Trizol® Reagent (Invitrogen), according to the manufacturer's protocol. Cells were collected by centrifugation and lysed by addition of 0.8 ml Trizol. To reduce viscosity, genomic DNA was sheared with 2 passes through a 26-gauge needle. Aqueous and organic phases were separated by addition of chloroform and subsequent centrifugation. After transfer of the aqueous phase, 10 μg of glycogen (Invitrogen) were added as a carrier and RNA was recovered by precipitation with 100% isopropanol and centrifugation. Following washing with 75% ethanol, pellets were dried and dissolved in RNase-free H2O. RNA integrity and concentration were measured using an Agilent 2100 Bioanalyzer with the RNA 6000 Nano LabChip® Kit (Agilent Technologies, Palo Alto, USA).

Reverse transcription and quantitative real-time PCR

Total RNA (500 ng) were reverse transcribed using the Superscript II system (Invitrogen), 0.5 mM dNTPs (MBI Fermentas, St. Leon-Rot, Germany), 40 U RNasin (Promega, Mannheim, Germany) and 0.5 μg oligo dT primers. Reactions were incubated at 42°C for 50 min followed by incubation at 70°C for 15 min.

Concentrations of XPC and ACTB transcripts were measured by quantitative real-time PCR on a LightCycler I ® (Roche Diagnostics, Mannheim, Germany) using the QuantiTect SYBR Green PCR Kit (Qiagen, Hilden, Germany). The following gene-specific primers, designed with Primer 0.5 software (Whitehead Insitute for Biomedical Research), were used: XPC_for 5′-ACC TGG TGA AGT GGT TCA TTG-3′, XPC_rev 5′-TGC AGG TTA TCT TGT TCA CTG G-3′, ACTB_for 5′-GGC ATC CTC ACC CTG AAG TA-3′ and ACTB_rev 5′-GGG GTG TTG AAG GTC TCA AA-3′. PCR conditions were as follows: 15 min at 94°C preincubation followed by 50 cycles of denaturation at 94°C for 30 sec, annealing at 56°C (XPC and ACTB) or 55°C (ATLTP6) for 30 sec and polymerization at 72°C for 30 sec. Reaction products were characterized by determination of melting point (40–97°C with 1°C/sec).

Each reverse transcription reaction was spiked with 0.05 ng ATLTP6 mRNA from Arabidopsis thaliana (Stratagene, La Jolla, USA). Recovery of this spike by quantitative real-time PCR served as a control to measure efficiency of the reverse transcription step (primers: ATLTP6_for 5′-CAG GCT CAA ACT TCT GTG GA-3′, ATLTP6_rev 5′-GGG AGA TCG ACA CCA CAT TT-3′).

LightCycler data were analysed by calibrator-normalized relative quantification with efficiency correction using the LightCycler Software 4 (Roche Diagnostics). Normalized ratios were determined for each sample, using ACTB as reference gene. Induction factors were determined as the ratio of the mRNA amount of XPC (given as normalized ratio) in the irradiated sample divided by the mRNA amount (normalized ratio) in the corresponding sample of the same patient which was not irradiated.

Statistical evaluation

Univariate comparisons were evaluated with χ2 test or, where appropriate, with Fisher's exact test. Variation between different experimental runs was determined using a lymphocyte sample of a healthy donor which was independently measured in 11 experiments. Experimental variability was calculated as the variation coefficient which is the quotient of the standard deviation by the mean value. The Kruskal–Wallis test was used to compare the XPC induction factors between subgroups. Odds ratios and 95% confidence intervals for enhanced radiosensitivity were computed using multivariate unconditional logistic regression analysis. Estimates were produced by the PHREG procedure of the statistical software package, SAS release 9.1 (SAS Institute, Cary, North Carolina).

Results

XPC mRNA induction was measured in lymphocytes of 99 patients 4 hr after in vitro irradiation with 5 Gy. Induction factors ranged from 0.29 to 5.77 fold, the median was 2.49. The mean of all induction factors was 2.59 with a standard deviation of 1.13 (Table II). From these values, a coefficient of variation of 44% was calculated. In addition, XPC mRNA induction factors were measured for one lymphocyte standard sample from a healthy blood donor in 11 independent experiments to determine experimental variability. We obtained a mean induction factor of 2.31, a standard deviation of 0.57 and a coefficient of variation of 25%. Thus, the variation of mRNA induction in prostate cancer patients was nearly twice as high as the variation between different experimental runs of the same blood sample indicating a considerable inter-individual variation of XPC induction after IR between the patients investigated.

Table II. Induction of mRNA Expression of the XPC Gene after Irradiation with 5 Gy Depending on Smoking Status and Clinical Radiosensitivity
 n (%)XPC induction factorsp value (Kruskal–Wallis test)
Mean ± SDMedian (range)
  • 1

    Dichotomized by median value, data missing for 7 patients.

All99 (100)2.59 ± 1.132.49 (0.29–5.77) 
Smoking status
 Never42 (42)2.23 ± 0.892.23 (0.29–4.29) 
 Current or ex-smoker57 (58)2.87 ± 1.222.87 (0.49–5.77)0.008
Tobacco consumption1 (pack-years)
 0–544 (48)2.26 ± 0.892.27 (0.29–4.29) 
 ≥648 (52)2.83 ± 1.292.72 (0.49–5.77)0.04
XPC induction
 low to medium (<4.10)89 (90)2.34 ± 0.872.36 (0.29–4.10) 
 high (≥4.14)9 (10)4.83 ± 0.544.89 (4.14–5.77) 
Clinical radiosensitivity
 normal sensitive (CTC 0-2)68 (69)2.56 ± 1.12.51 (0.29–5.77) 
 hypersensitive (CTC 3+4)31 (31)2.66 ± 1.22.36 (1.09–5.12)0.98

The XPC mRNA induction factors were not associated with patients' age, body mass index, alcohol consumption and family status (p values > 0.05; data not shown) but they were strongly associated with smoking habits. Current or exsmokers had higher induction factors than never-smokers (p = 0.008). This difference was also observed when tobacco consumption was dichotomized by the median value of pack-years smoked (Table II).

As there were some patients with a remarkably high XPC induction, patients were divided into two groups for further analysis: (i) those with XPC induction above the 90th percentile (4.14–5.77) and (ii) those with low to medium induction (0.29–4.10) which subsumes the remaining 90% of all patients. Both groups significantly differed by smoking status (Table IIIA), when patients were categorized as current or exsmokers or never-smokers (p = 0.04). 89% of all patients with high but only 48% of all patients with low to medium XPC induction were found to be heavy smokers with more than 6 pack-years (p = 0.03).

Table IIIA. Association of High XPC mRNA Induction with Smoking and Tobacco Consumption
 XPC induction 
 Low to medium (<4.10) n (%)High (≥4.14) n (%)p value1
  • 1

    Fisher's exact test.

  • 2

    Data missing for 7 patients.

All89 (90)10 (10) 
Smoking status
 Never41 (46)1 (10) 
 Current or ex-smoker48 (54)9 (90)0.04
Tobacco consumption2 (Pack-years, median)
 0–543 (52)1 (11) 
 ≥640 (48)8 (89)0.03

XPC induction factors were analyzed for their association with clinical radiosensitivity. Patients with normal sensitivity to radiotherapy showed induction factors from 0.29 to 5.77 with the median at 2.51. XPC induction factors in hypersensitive patients were similar: 1.09–5.12 with the median being 2.36 (Table II).

The two groups of patients with low to medium (<4.10) and high XPC induction factors (>4.14) were tested for association with clinical radiosensitivity (Table IIIB). Severe adverse side effects occurred in 29% of patients in the group with low to medium induction and in 56% of patients in the high XPC induction group (p = 0.12). Compared to patients showing low XPC induction levels, patients with high XPC induction after IR were at increased risk to suffer severe side effects, with a crude odds ratio of 3.0 (confidence interval 0.8–12.2; p = 0.05). As tobacco consumption was shown to strongly affect XPC induction factors, a multivariate logistic regression model was used to adjust the association of XPC induction factors and radiosensitivity for tobacco consumption. After adjustment, having high XPC induction factors was found to be associated with a significantly increased risk of developing severe acute side effects of radiotherapy (odds ratio 5.3; p = 0.03; 95% confidence interval 1.2–24.5).

Table IIIB. Association of High XPC mRNA Induction with Development of Acute Side Effects in Patients
 Clinical radiosensitivityp value1Odds ratio
Normal sensitive (CTC 0-2) n (%)Hypersensitive (CTC 3+4) n (%)(95% confidence limits)
  • 1

    Fisher's exact test.

  • 2

    Dichotomized by median value.

  • 3

    Odds ratio adjusted for pack-years (2 categories: 0–5, ≥6 pack-years).

Tobacco consumption2 (pack-years, median)
 0–526 (59)18 (41) 1.0
 ≥637 (77)11 (23)0.070.4 (0.2–1.1)
XPC induction
 Crude OR
  low to medium (<4.10)59 (71)24 (29) 1.0
  high (≥4.14)4 (44)5 (56)0.123.0 (0.8–12.2)
 Adjusted for smoking3
  low to medium (<4.10)59 (71)24 (29) 1.0
  high (≥4.14)4 (44)5 (56)0.035.3 (1.2–24.5)

Discussion

Induction of XPC mRNA after ionizing irradiation

To find a biomarker for clinical radiosensitivity, XPC mRNA induction factors were determined in peripheral blood lymphocytes of 99 prostate cancer patients using reverse transcription of mRNA coupled with quantitative real-time PCR which, at present, is estimated as the most reliable method for RNA quantification.31, 32, 33 To prevent most of the typical shortcomings of RNA quantification, the quantification protocol was optimized by controlling carefully for RNA and cDNA quality and for experimental differences by normalizing all measurements by a calibrator. Our XPC induction levels were in accordance to other studies, which measured expression changes in cells after irradiation with similar doses and repair intervals in human cell lines or primary cells.16, 17, 20, 24 The XPC-inducing irradiation dose of 5 Gy has been used in many studies investigating induction of DNA damage and repair in human lymphocytes.34, 35 This dose is in the same range as single doses used in radiotherapy regimens, e.g., 1.5–2 Gy per fraction.36 Furthermore, it has been shown that lymphocytes can repair DNA damage induced by 5 Gy quite well, although some cells will undergo apoptosis at later time points.34, 37, 38

At a first sight, induction of XPC expression in response to treatment with IR is surprising as the XPC protein is involved in nucleotide excision repair26 and is therefore not primarily responsible for the repair of IR-induced DNA damage. However, besides DNA single- and double-strand breaks, IR generates reactive oxygen species (ROS) which can induce certain helix-distorting DNA modifications and protein-DNA cross links.27 Furthermore, there is evidence that also smaller ROS-induced lesions may be substrates for NER.39 IR-induced upregulation of XPC mRNA is part of the cellular DNA damage response pathway which acts via ATM and p53 as initial DNA damage sensors. It is suggested to occur by a similar mechanism as induction of XPC mRNA by UV light which is more widely investigated.25, 40 In analogy to the UV-induced response, activated p53 protein can act as a transcription factor for several DNA damage response genes, thus inducing nucleotide excision repair after ionizing radiation. In addition, XPC protein was also found to be active in the base excision repair pathway by interacting e.g., with thymine DNA glycosylase.41

Inter-individual variation of XPC mRNA induction after ionizing irradiation

Inducibility of XPC mRNA expression by irradiation varied about 20-fold among our patients. The high variability in XPC induction could be due to prior exposure to DNA damaging agents, such as tobacco smoke20 but also to other physiological factors which were not amenable to our analysis. We could however show that high induction factors were associated with smoking. These results might suggest an adaptive response induced by smoking as more radiosensitive individuals belong to the non-smoking or light-smoking group. To the best of our knowledge, this is the first report of an association of XPC mRNA up-regulation with tobacco smoking. In contrast, constitutive expression levels of XPC mRNA were reported to be not affected by smoking.42 The p53-transmitted DNA damage response might be adaptive and higher in individuals with a longer smoking history than in non-smokers. Such an adaptive repair response would elicit an increased repair activity against DNA damage by smoking. This in turn could improve the repair response to subsequent radiation damage. The exact underlying molecular mechanisms remain to be investigated and involve complex signaling pathways and many proteins.43

Lymphocytes of some patients showed XPC induction factors below 1.0 indicating that the mRNA message is not inducible within 4 hr in these patients, as also observed by others.25, 44 Adimoolam and Ford25 suggested that an initial temporary decrease in levels of XPC mRNA is likely to be due to DNA damage-induced, transcription-blocking lesions in the large XPC gene which is spanning a region of about 30 kb. Only after transcription-coupled repair was completed, XPC mRNA expression could be restarted so as to provide sufficient XPC protein for inducing repair of the global genome. In our population, an initial decrease in mRNA levels in some individuals could indicate a delay in other repair pathways. The analysis of later time points was, however, not possible within the framework of our study as the amount of lymphocytes available from patients was limited.

A further explanation for the inter-individual differences in XPC mRNA levels after irradiation might be genetic variation, e.g. caused by single nucleotide polymorphisms in the XPC gene promoter region. Data on possible associations between single nucleotide polymorphisms in the XPC gene and its expression levels are, however, too scarce to support this assumption.42 Genetic variants not only of the XPC gene but also of other interacting genes of the DNA damage signaling pathway could be involved and should be investigated in a comprehensive genomic analysis.

Radiation-induced XPC mRNA up-regulation and incidence of acute side effects of radiotherapy

An association between cellular XPC induction after irradiation in vitro and the occurrence of clinical side effects after radiotherapy in patients was not detectable using crude expression data. We adjusted for tobacco consumption since XPC induction levels were strongly correlated with smoking in our study population and smoking was associated with a non-significantly lower risk of acute toxicity. After this adjustment which accounts for differences between subjects with low and high XPC induction levels with respect to smoking, high induction factors above the 90th percentile were still found to be associated with a more than 5-fold increased risk for developing acute side effects of radiotherapy. The effect was not detectable when the study population was divided by tertiles suggesting that only the highest induction levels are linked to this effect. It is conceivable that factors other than smoking which can increase XPC induction might also contribute to radiosensitivity (e.g. xenobiotic exposures, inflammatory processes).

High levels of XPC protein might interfere with vital cellular processes as it was observed that RAD4 and its mammalian equivalent XPC are highly toxic when expressed at too high levels in yeast.45, 46 This might be explained by the ability of XPC to bind to a wide range of minimally distorted DNA structures in case no DNA damage is present. These DNA structures include backbone modifications caused by phosphorothiolate or methylphosphonate treatment or DNA mismatches which occur during regular cellular transcription and replication.47, 48, 49 As such an “unnecessary” DNA binding of XPC could lead to “spontaneous” DNA turnover, and thus could impede normal DNA metabolism, a tight regulation of cellular XPC protein concentrations is required, such as the transcriptional control by p53. This is in agreement with a previously reported finding that overexpression of the glycosylases hNTH1 and OGG1 increased radiation lethality and mutant frequency.50 We suggest that in patients with particularly strong IR-induced XPC upregulation, the resulting high XPC protein levels might not be counterbalanced by an adequate amount of DNA damage suitable for nucleotide excision repair, as most IR-induced DNA damage will need to be processed by more specialized DNA repair pathways such as base excision repair. Thus, the highly unbalanced XPC levels could cause cellular toxicity and cell loss in normal tissue. Cell death by mitotic catastrophe and apoptosis has been recognized as an important response to ionizing radiation in many cell types.51, 52 If radiation-induced loss of cells is, however, too extensive, the repopulating activity of the normal tissue during therapy might be insufficient resulting in acute side effects in the normal tissue.53, 54 Further in vitro experiments investigating the role of apoptosis in the development of side effects are however needed to support these explanations. In addition, it will be necessary to prove that XPC induction after irradiation measured in the surrogate tissue, peripheral blood lymphocytes, occurs similarly in the affected target tissues of colon and bladder where the side effects, diarrhea and urinary urgency, develop. This requires, however, that target tissue specimens can be obtained from patients before and after radiotherapy by invasive procedures.

We are aware that the statistical power of our pilot study is limited, only 5 hypersensitive cases and 4 controls are in the high XPC mRNA induction cohort. We nevertheless were able to demonstrate the feasibility of this type of gene induction study. We minimized sources of bias in our study by restricting tumor type (prostate cancer) and type of side effects. We collected data on side effects at defined time-points during radiotherapy using a predefined classification system comparable to other studies. By choosing CTC score 3 and 4 toxicity (requiring clinical intervention because of side effects) as an indicator of increased acute toxicity, we selected an indicator that should be less prone to variability in classification.35

In the current study, we restricted the analysis to acute toxicity whereas other studies on tissue radiosensitivity have also included late reactions after radiotherapy. Early and late reactions are not necessarily related to each other and may be influenced differently by genetic predisposition.55 More comprehensive analyses including late effects of radiotherapy are therefore warranted.

Moreover, high induction of XPC mRNA expression after IR in vitro cannot entirely explain the incidence of side effects during radiotherapy, as all other regulatory mechanisms of cellular XPC protein levels have to be considered (e.g. posttranslational modifications, protein stability and interaction with other components of the DNA damage signaling and repair pathways). Tissue characteristics, such as propensity for inflammatory reactions or cytokine production, may play a major role in the induction of these side effects in the irradiated normal tissue.6 Thus, further mechanistic analyses as well as molecular-epidemiological studies are necessary to establish XPC mRNA induction in vitro as a predictive biomarker for radiosensitivity.

Acknowledgements

We thank O. Zelezny, P. Waas and R. Gliniorz (Division of Toxicology and Cancer Risk Factor, German Cancer Research Center) for their excellent technical assistance, R. Haselmann (Department of Clinical Radiology, University Hospital Heidelberg,) and K. Smit (Clinical Epidemiology, German Cancer Research Center) for their important contribution to data collection and management. We are grateful to all the patients who participated in the study.

Ancillary