Ac

PRa

b

c

d

a

ARRAA

KDFHGNNUX

1

r

SlT

ST

t

1d

DNA Repair 10 (2011) 603–610

Contents lists available at ScienceDirect

DNA Repair

journa l homepage: www.e lsev ier .com/ locate /dnarepai r

modified host-cell reactivation assay to quantify DNA repair capacity inryopreserved peripheral lymphocytes

edro Mendeza,b,∗, Miquel Tarona,b,c,∗∗, Teresa Morana, Marco A. Fernandeza,d, Gerard Requenaa,d,afael Rosell a,b,c

Health Science Research Institute Germans Trias i Pujol, Badalona, Barcelona, SpainCatalan Institute of Oncology, Hospital Germans Trias i Pujol, Badalona, Barcelona, SpainUSP Dexeus University Institute, Barcelona, SpainFlow Cytometry Core Unit, Health Science Research Institute Germans Trias i Pujol, Badalona, Barcelona, Spain

r t i c l e i n f o

rticle history:eceived 9 November 2010eceived in revised form 30 March 2011ccepted 1 April 2011vailable online 5 May 2011

ey words:NA repair capacitylow cytometryost-cell reactivation assayreen fluorescent proteinucleofectionucleotide excision repairV-lighteroderma pigmentosum

a b s t r a c t

The host-cell reactivation assay (HCRA) is a functional assay that allows the identification of thegenes responsible for DNA repair-deficient syndromes, such as Xeroderma pigmentosum, by cross-complementation experiments. It has also been used in molecular epidemiology studies to correlatethe low nucleotide excision repair pathway function in peripheral blood lymphocytes with an increasedrisk of bladder, head and neck, skin and lung cancers. Herein, we present the technical validation of anewly modified HCRA, where nucleofection is used for the transfection of the pmaxGFP plasmid into cry-opreserved peripheral blood lymphocytes (PBLs) or lymphoblastoid cell lines. In each sample, 20–24 hafter transfection, the relative DNA repair capacity (DRC) was quantified by flow cytometry, comparingthe transfection efficiency of nucleoporated cells with undamaged plasmid to those transfected with UV-light damaged plasmid in the seven cell lines that were characterized by different DNA repair phenotypes.Dead cells were excluded from the analysis. We observed a high reproducibility of the relative DRC, trans-fection efficiency and cell viability. The inter-experimental normalization of the flow cytometry resultedin an increased data accuracy and reproducibility. The amount of cells required for each transfectionreaction was reduced fourfold, without affecting the final relative DRC. Furthermore, our HCRA demon-strated strong discrimination power in the UV-light dose–response, both in lymphoblastoid cell lines andcryopreserved PBLs. We also observed a strong correlation of the relative DRC data, when samples were

measured against two independent batches of both damaged and undamaged plasmid DNA. The relativeDRC variable shows a normal distribution when analyzed in the cryopreserved PBLs from a cohort of 35lung cancer patients and a 5.59-fold variation in the relative DRC is identified among our patients. Themitotic dynamic was discarded as a confounding factor for the relative DRC measurement in this cohortof patients. The results indicate that our method is highly sensitive, reliable and reproducible, and thus,

-base

it suitable for population

. Introduction

A deficiency in the DNA repair capacity (DRC) is an independentisk factor for cancer [1–3].

∗ Corresponding author at: Molecular Biology Laboratory, Medical Oncologyervice-ICO, Hospital Germans Trias i Pujol, Fundació Investigació en Ciències dea Salut Germans Trias i Pujol, Ctra Canyet s/n, 08916-Badalona, Spain.el.: +34 93 497 86 82; fax: +34 93 497 89 50.∗∗ Corresponding author at: Molecular Biology Laboratory, Medical Oncologyervice-ICO, Hospital Germans Trias i Pujol, Ctra Canyet s/n, 08916-Badalona, Spain.el.: +34 93 497 89 25; fax: +34 93 497 89 50.

E-mail addresses: pmendez00@gmail.com (P. Mendez),aron.miquel@gmail.com (M. Taron).

568-7864/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.dnarep.2011.04.001

d studies to quantify in vitro DNA-repair deficiencies.© 2011 Elsevier B.V. All rights reserved.

It has been described that the DRC, measured as a continu-ous variable in a healthy population, distributes normally with anapproximate 4.7-fold variation in the DRC values, ranging from 3.8to 17.9 [4]. In addition, the median DRC values from lung cancercases are significantly lower than those in matched controls, sug-gesting that an inefficient systemic (host) DRC might account foran increase of lung cancer risk in some population subsets [5,6].

Platinum-based chemotherapy is the cornerstone in the treat-ment of non-small cell lung cancer (NSCLC) patients. Although themechanism of action of platinum (Pt) drugs is poorly understood,it is accepted that a major cause of its cytotoxicity arises from its

covalent binding to DNA, which inhibits some important cellularfunctions, such as transcription and replication [7,8]. The cellularresistance to Pt drugs, despite its multifactorial nature, is believedto result from an increased DNA repair of Pt-DNA adducts and plays

6 Repa

ab

tt

epetafotsrssvr

bpcrflv

2

2

iSviPeoEt

unpSwps

2

vtNtpiPaXlt

04 P. Mendez et al. / DNA

central role in the inhibition of the therapeutic efficacy of Pt drugs,oth in vitro [9,10] and in vivo [11–14].

For this reason, we have developed a technically reliable assayo measure the host-DRC levels in populations of cancer patientshat may be treated with DNA-damaging agents.

The HCRA is based on the independent transfection of periph-ral blood lymphocytes (PBLs) with either damaged or undamagedlasmid DNA. At 20 to 24 h after transfection, the reporter proteinxpression is used for relative quantitation from both experimen-al conditions. Previous versions of the HCRA utilized transfectiongents, such as DEAE-dextran [4,15], adenovirus [16] or lipo-ectamine [17], which are either, toxic, biologically hazardousr less effective, respectively, when compared with nucleofec-ion technology. The reporter genes previously used in the HCRA,uch as chloramphenicol[15], luciferase [4] or B-galactosidase[16],equired the use of radioactivity, a short detection time for enzyme-ubstrate interaction (especially for luciferin-luciferase) or lowerensitivity, respectively, when compared with the non-radioactive,ersatile and sensitive use of green fluorescent protein (GFP) as aeporter gene.

Our modified HCRA method presented here relies on the com-ination of nucleofection technology for the transfection of themaxGFP plasmid into cryopreserved peripheral blood lympho-ytes or lymphoblastoid cell lines. At 20-24 h post-transfection, theelative DNA repair capacity (DRC) of each sample is quantified byow cytometry. This report extensively describes the methodology,alidation and advantages of our HCRA.

. Material and methods

.1. Plasmid preparation and UV irradiation

The 4671 base pair (bp) pmaxGFP plasmid is supplied as a pos-tive control in the Human T cell nucleofection kit (Lonza, Basel,witzerland). The plasmid is a eukaryotic mammalian expressionector that encodes the green fluorescent protein, maxGFP, whichs an improved CopGFP variant of the protein cloned from copepodontellina plumata. The CMV promoter controls the maxGFP genexpression. The DH5� transformants were selected with 30 �g/mlf kanamycin. Plasmid preparations were carried out using thendo-Free Plasmid Maxi Kit (QIAGEN, Valencia, CA, USA), followinghe manufacturer’s instructions.

Because of the lability of the DNA adducts induced by cisplatinnder physiological conditions [18,19], we decided to explore theucleotide excision (NER) function by damaging the pmaxGFPlasmid by UV-irradiation. To produce the UV-photolesions, atratalinker® UV Crosslinker 1800 (Stratagene, La Jolla, CA, USA)ith 5 UV-light bulbs (254 nm, 8 watts each) was used. A detailedrotocol for plasmid damaging with UV-light is provided as aupplementary method.

.2. Cell lines

The 7-lymphoblastoid cell lines transformed by Epstein-Barrirus and used in this study were obtained from the Coriell Insti-ute for Medical Research, Camden, NJ, USA. As controls of severeER deficiency, we included the XPA−/− and XPC−/− lymphoblas-

oid cell lines and the PBLs from an Xeroderma pigmentosum (XP)atient. To test the degree of the assay’s discrimination power, we

ncluded an XPF−/−-derived cell line that was kindly provided byrofessor Leon H. Mullenders, Leiden, Netherlands. The XPF vari-

nt of the disease is characterized as having one of the mildestP phenotypes, and its cells have residual NER DRC activity. Cell

ines GM07544 and GM03798 were used as NER-proficient con-rols. Finally, we also added the FANCD1−/− and FANCD2−/− cell

ir 10 (2011) 603–610

lines. For cell culture conditions, see subsection 2.5 PBMCs cul-ture.

2.3. Cryopreserved peripheral blood mononuclear cells (PBMCs)from NSCLC patients and donors

PBMCs from seven apparently healthy donors and from oneclinically, but not molecularly, diagnosed XP patient were cryopre-served as described below and were used to validate the technicalexperiments.

Strictly from the technical point of view and independent ofits clinical implications, there are three important questions tobe addressed in the PBLs from NSCLC patients with our modifiedHCRA: (1) what is the range of relative DRC values among NSCLCpatients, (2) does the relative DRC values obtained from the NSCLCpatients distribute normally and (3) what is the impact of mitoticdynamics on DRC, because it may be a potential confounding factorfor DRC assessment?

To address these issues, blood samples were collected and thePBLs from 35 newly diagnosed NSCLC patients, regardless of theirclinical or pathological characteristics, were cryopreserved beforeany therapeutic intervention. All participants signed an informedconsent form.

2.4. PBMCs isolation and cryopreservation

From each subject, 16–24 ml of peripheral blood was drawn inheparinized tubes (BD Biosciences, Franklin Lakes, NJ, USA). A ficollgradient was used, as previously described [1], to isolate the PBMCs.The cells were resuspended at a concentration of 1 × 107 cells ml−1

in freezing media [containing 90% fetal bovine serum (FBS) and10% dimethyl sulfoxide (DMSO)] and gradually cryopreserved byimmersion in 2-propanol at −80◦ C. To increase the cell viabilityafter long-term cryopreservation, the storage of the PBLs in thegaseous phase of liquid nitrogen is recommended.

2.5. PBMCs culture

The PBMCs were washed once in thawing media (50% RPMI-1640, 40% FBS and 10% d-glucose). Cells were cultured in standardconditions for 48 h in upright flasks, at a concentration of0.4 × 106 cells ml−1 in RPMI-1640 culture media, supplementedwith 20% FBS, 56 �g/ml of phytohemagglutinin-P (Sigma, St. Louis,MO, USA) and 1× of penicillin, streptomycin and glutamine (Invit-rogen). After renewing the culture media, an additional 48 h ofculturing in the same culture flasks was required for optimalgrowth.

2.6. T lymphocyte transfection

After 96 h in culture, the viable cells were counted using the try-pan blue (0.4%) staining method. In each transfection, between 0.3and 0.8 × 106 viable PBMCs were resuspended in 20 �l of HumanT cell Nucleofection solution (Lonza) and then mixed with 200 ngof either damaged or undamaged plasmid. In one sample, tworeplicates were performed for each experimental condition. Addi-tionally, for each sample, an additional transfection condition wasperformed without plasmid to obtain GFP-negative control cells.For transfection of both cryopreserved PBLs and lymphoblastoidcell lines, the FF138 nucleofection condition was applied in the 96-well shuttle Nucleofector II Device (Lonza). After transfection, the

cells were resuspended in 1 ml of complete pre-warmed (37 ◦C)DMEM culture media (Invitrogen), without PHA supplementation.The cells were incubated in standard conditions for 20–24 h priorto the assessment of the relative DRC by flow cytometry.

Repai

2

tttonfli2aaTle

2

fitat(t

2

d1lfpddhNGFws

2

(wvFS1u3ope

utctt

P. Mendez et al. / DNA

.7. Optimization of nucleofection conditions

To optimize the nucleofection process, 41 different nucleopora-ion conditions of varying pulse strength and pulse length, coded byhe individual pre-programmed parameters, were tested for generansfer into the GM02246 lymphoblastoid cell line and in cry-preserved PBLs from a healthy control patient. We selected theucleofection condition that produced a combination of high trans-

ection efficiency and cell viability simultaneously in both the celline and PBLs. As described in the literature [20], DNA is introducednto the cells via a voltage pulse of 10–100 s, with a field strength of–10 kV/cm and a current density of at least 2 A/cm2, followed byn uninterrupted current flow of 1–100 ms (preferably up to 50 ms)nd a current density of 2–14 A/cm2 (preferably up to 5 A/cm2) [21].he specific parameters are confidential information and not pub-ished. The combination of programs is unique and reproducible forvery Nucleofector device (www.lonzabio.com).

.8. Optimization of cell numbers per transfection

We used cryopreserved PBMCs from the same donor to performve independent HCRAs, where the only difference was the ini-ial amount of cells introduced in the transfection. The followingmounts were used: 0.2, 0.4, 0.6, 0.8 × 106 and 106 cells per reac-ion. As indicators of robustness, the mean ± standard deviationSD) was calculated for the three variables: relative DRC, transfec-ion efficiency and cell viability.

.9. UV-dose–response curves

As a substrate for the HCRA, the pmaxGFP plasmid was irra-iated with different UV-light doses (0, 400, 600, 800, 900 and000 J/m2) and dose–response curves were performed for the

ymphoblastoid cell lines or the cryopreserved PBLs. As controlsor severe NER deficiency, we included XPA−/− and XPC−/− lym-hoblastoid cell lines and PBLs from an XP patient. To test theegree of the assay’s discrimination power, we included an XPF−/−

erived cell line. The XPF variant of the disease is characterized asaving one of the mildest XP phenotypes, and its cells have residualER DRC activity. As NER-proficient controls, cell lines GM07544,M03798 and PBLs from an apparently healthy donor were used.inally, we also added the FANCD1−/− and FANCD2−/− cell lines,hich, although DNA repair-deficient in the FA/BRCA pathway,

hould in theory behave in the same way as NER-proficient cells.

.10. Relative DRC analysis by flow cytometry

Relative DRC analysis was performed using a BD FACSCanto IIBD Biosciences) flow cytometer. Before beginning any experiment,e routinely performed an inter-experimental normalization of the

oltage values of the photo-multiplier channels, Forward Scatter,ITC (GFP) and PE for PI staining, by loading 5000 fluoro-labeledPHERO Rainbow calibration particles (BD Biosciences) diluted:50 in PBS (Fig. S1). For cell viability assessment during DRC eval-ation, cells were incubated for the last 15 min in media containing0 �M of propidium iodide (PI) (Sigma). For each patient’s sample,r cell line, we recorded 15,000 cells that were transfected withoutlasmid to first exclude dead cells from the DRC analysis and tostablish the threshold for GFP-negative cells.

From cells transfected with either damaged (pmaxGFP-DAM) orndamaged (pmaxGFP-UND) plasmid, 2000 GFP-positive or 40,000

otal cells were recorded in each replicate. The relative DRC was cal-ulated only in samples with a variation coefficient equal or lowerhan 15% between replicates, using the following formula: rela-ive % DRC = (x pmaxGFP-DAM/x pmaxGFP-UND) × 100, where x

r 10 (2011) 603–610 605

represents the weighted average from the transfection efficienciesduplicates.

2.11. NDI calculation

After 96 h of PBMC culture, and in parallel to relative DRC quan-tification, we evaluated the NDI, as an indicator of mitotic dynamics,in each clinical sample, as previously described [22].

2.12. HCRA reproducibility

We evaluated the reproducibility of the three variables mea-sured (relative DRC, transfection efficiency and cell viability) usingour method in XP and NER-proficient lymphoblastoid cell lines andPBLs. Seven HCRA independent experiments, from eight consecu-tive cell line passages, were carried out.

2.13. Validation of HCRA for prospective studies

The main technical source of variability in the HCRA methodcomes from the preparation of different plasmid batches as DNArepair substrates. We prepared two independent batches of bothUV-light damaged and undamaged pmaxGFP plasmids that wereseparated by a time interval of four months. Five cell lines, with awide relative DRC spectrum, and seven cryopreserved PBLs fromdonors were used to measure the relative DRC with both batchesof plasmid. We conducted a linear regression analysis to assessthe degree of correlation between the relative DRC values obtainedfrom both plasmid batches.

2.14. Statistics

The relative DRC and NDI data were analyzed as a continuousvariable. We studied the relationship between both variables bylinear regression analysis. For the evaluation of the normal dis-tribution, we applied the Kolmogorov–Smirnov and Shapiro–Wilktests to the results obtained in 35 cryopreserved PBLs from NSCLCpatients. For evaluation of the relative DRC, transfection efficien-cies and cell viability reproducibility variables, the mean ± standarddeviation was used as an indicator of robustness. All statistical testswere accepted as significant when the alpha values were <0.05. Theunpaired t-test was used to compare the differences in the relativeDRC between WT and XP samples. All of the statistical analyseswere performed with GraphPad Prism v 5.0 software.

3. Results and discussion

A diagram of an example of the relative DRC analysis by flowcytometry is represented in Fig. 1.

3.1. Optimization of nucleofection conditions

Two transfection-related variables were recorded for analysis:cell viability and transfection efficiency. Transfection efficiencywas measured by flow cytometry as the % of live cells (PI neg-ative), transfected with undamaged plasmid that expressed theGFP reporter gene (Fig. 1F). Among the 41 conditions tested, wefound a wide range of results and an inverse correlation betweenthem, using cell viability as an independent variable in the lin-ear regression analysis (R2 = 0.9283; slope = −0.817; P < 0.0001 forGM02246 cell line; R2 = 0.7952; slope = −0.977; P < 0.0001 for cry-opreserved PBLs). Each of the 41 experimental conditions tested

are represented on the X-axis of Fig. 2A and B. The condition FF138produced a transfection efficiency of 71.60% and 50.5% and a cellviability of 34.46% and 15.33% after dead cell exclusion, in cell linesand PBLs, respectively (Fig. 2). For this reason, FF138 was selected

606 P. Mendez et al. / DNA Repair 10 (2011) 603–610

Fig. 1. An example of the relative DRC analysis of one sample by flow cytometry used in this study. Each row represents only one replicate of each of the three experimentalconditions used. (A–C) Flow cytometry plots of transfected cells without plasmid, (D–F) undamaged pmaxGFP plasmid transfected cells and (G–I) 1000 J/m2 UV-light damagedpmaxGFP plasmid transfected cells. Dead cells are doubly excluded: first by Forward Scatter plot (A, D and G), and then, by excluding the PI positive cells (B, E and H). Live,transfected cells without plasmid (C) are used to establish the threshold of GFP+ and GFP− cells, where 15,000 events were recorded. The threshold is adjusted in each samplea ed pmo remai≤ e DRC

fpredPm

3

fosccbwcft

nd used to obtain the percentage of live GFP+/− cells in the undamaged and damagr 40,000 total events, were recorded in each replicate (two replicates in the two15% between replicates were valid. The following formula was used for the relativ

or use throughout the entire validation procedure. The nucleo-oration technology produced high intra- and inter-experimentaleproducibility of transfection efficiency and cell viability, whenight independent experiments were evaluated (Fig. 3E). The repro-ucibility of the test behaved equally in cell lines and cryopreservedBLs, as we demonstrated by performing five independent experi-ents on a low relative DRC PBL sample from the same patient.

.2. Optimization of cell number per transfection

The relatively high number of cells per transfection requiredor most HCRA protocols, when used to measure the relative DRCf cryopreserved PBLs, can limit its application. Thus, the idealituation would be to have stable relative DRC, transfection effi-iency and cell viability values, regardless of the initial amount ofells introduced in the transfection. Therefore, we tested the sta-ility of these three variables in cryopreserved PBLs from a donor,

here the only experimental difference was the initial amount of

ells used per transfection. The results obtained were consistentor the three variables: the mean ± standard deviation for the rela-ive DRC, transfection efficiency and cell viability were 8.56 ± 0.75,

axGFP transfected cells (F and I). For the relative DRC quantitation, 2000 GFP+ cells,ning experimental conditions). Only samples with variation coefficient that werecalculation: relative DRC = (pmaxGFP-DAM/pmaxGFP-UND) × 100.

50.26 ± 1.68 and 34.26 ± 3.02, respectively excepting for the DRCvalue, when 106 cells were used (Fig. 4).

3.3. UV-dose–response curves

Because the presence of even one UV-induced pyrimidine dimercan block the reporter gene transcription [23], we performeddose–response curves in both lymphoblastoid cell line (Fig. 3A) andcryopreserved PBLs (Fig. 3B), and we screened the UV-light dosethat better discriminated the relative DRC between NER-proficientand -deficient cells. We found that 1000 J/m2 was the dose thatbest discriminated the NER-proficient from the NER-deficient phe-notype. The selected dose is similar to that previously reportedby other investigators [24,25]. In addition, cyclobutane pyrimidinedimers (CPDs) and 6–4 pyrimidone photoproducts, representedapproximately 75% and 25%, respectively, of the photolesions gen-

erated by UV-light radiation [26]. Within the coding sequences(CDS) for the CMV promoter and the maxGFP reporter gene, 79and 80 isolated target sequences for CPDs formation, respectively,were found.

P. Mendez et al. / DNA Repair 10 (2011) 603–610 607

Fig. 3. Discrimination power and reproducibility of our HCRA. The pmaxGFP UV-light dose–response curves were calculated in the NER-deficient (XP) and NER-proficientlymphoblastoid cell lines (A) and the cryopreserved primary cultures from the PBLs (B). Panel (C) represents the relative DRC ratio between GM07544 and other cell lines andPBLs. Reproducibility of relative DRC at 1000 J/m2 (D) and transfection efficiency and cell viability (E); each bar represents the mean ± standard deviation of eight independentexperiments, except for the low relative DRC PBLs, where the bar represents the mean ± standard deviation of five independent experiments. The details to calculate the%DRC are included in the subheading 2.10. For the unpaired t-test **P < 0.01; ***P < 0.001 comparing relative DRC values from WT and XP cell lines (D) or PBLs (B) at 1000 J/m2.

Fig. 2. 41 nucleoporation conditions in GM02246 cell line and PBLs. In each transfection reaction, 0.6 × 106 cells were mixed with 200 ng of WT pmaxGFP plasmid. Thetransfection efficiency (striped columns) and cell viability (solid black columns) were assayed by flow cytometry 24 h after transfection in both (A) the GM02246 cell lineand (B) the PBLs from a donor. FF138 was selected as the nucleoporation condition that produced the better equilibrium between transfection efficiency and cell viability inboth the GM02246 cell line and PBLs.

608 P. Mendez et al. / DNA Repair 10 (2011) 603–610

F ive DRe nsfecT ss.

ltvrwPticXXXXfinP

3

tiifll1r8542

3

fpspocsRod

any relationship between the relative DRC and NDI variablesby linear regression analysis. The correlation coefficient wasR2 = −0.0016, and the slope was 0.0009, which were without sta-tistical significance (P = 0.800) (Fig. 6E).

Fig. 5. Correlation of relative DRC values obtained from two independent batches

ig. 4. Using a different cell number per transfection does not affect the final relatxperiments, where the only difference was the number of cells introduced in the trahe transfection efficiency and cell viability (B) were used as indicators of robustne

Interestingly, when we calculated the WT/XP DRC ratio in cellines, using GM07544 as a reference, it was directly correlated withhe clinical severity of the XP disease. The WT/XPA−/− DRC ratioalue was 35.60, the WT/XPC−/− ratio value was 10.68, and theatio of WT/XPF−/− was 2.30. The ratio value for the XP patientithout molecular diagnosis was 13.42. Finally, the values for WT

BLs and for the GM03798 WT cell lines were 1.50 and 1.08, respec-ively (Fig. 3C). Our HCRA test was able to efficiently discriminate,n a highly reproducible manner, between both the NER-proficientell lines and the XP cells with residual relative DRC (GM07544 vs.PF−/−, P < 0.01; GM03798 vs. XPF−/−, P < 0.01) and also from theP cells with a severe phenotype (XPA−/− vs. GM07544, P < 0.001;PA−/− vs. GM03798, P < 0.001; XPC−/− vs. GM07544, P < 0.001;PC−/− vs. GM03798, P < 0.001; Fig. 3D). These findings were con-rmed in the PBLs (Fig. 3B) (WT PBLs vs. XP PBLs, P < 0.01), butot when WT cell lines were compared (GM07544 vs. GM03798= 0.801).

.4. HCRA reproducibility

Good reproducibility of relative DRC (Fig. 3D), transfec-ion efficiency and cell viability (Fig. 3E) were observed, asndicated by their dispersion in eight and five independent exper-ments for cell lines and PBLs, respectively. The results obtainedor the relative DRC ± SD in XPA−/−, XPC−/−, XPF−/−, WT cellines (GM03798 and GM07544), FANCD1−/−, FANCD2−/− andow relative DRC PBLs were 0.58 ± 0.24, 1.23 ± 0.36, 6.45 ± 0.40,2.02 ± 3.20, 12.45 ± 1.58, 32.87 ± 6.09, 7.19 ± 1.04 and 3.93 ± 0.54,espectively. The transfection efficiency values were 51.37 ± 6.82,6.74 ± 2.95, 69.22 ± 6.37, 62.30 ± 7.74, 66.28 ± 4.98, 68.88 ± 6.77,1.39 ± 4.41 and 19.97 ± 1.76. The cell viability values were5.44 ± 10.47, 44.96 ± 3.96, 28.50 ± 3.08, 32.02 ± 4.81, 27.10 ± 2.70,2.65 ± 4.83, 42.70 ± 2.75 and 13.63 ± 1.69.

.5. Validation of HCRA for prospective studies

The technical sources of variability must be strictly controlledor any biomarker that is implemented in large and prospectiveopulations studies. In the HCRA, the main source of variabilitytems from the preparation of different batches of the damagedlasmid. The linear regression analysis of relative DRC valuesbtained in a group of five lymphoblastoid cell lines and sevenryopreserved PBLs from the two independent plasmid batches

howed a strong correlation, with a coefficient of determination of2 = 0.986 and a slope of 1.168 (P < 0.0001) (Fig. 5), indicating thatur proposed method for plasmid damage by UV-light is repro-ucible.

C value. PBLs from the same donor were used to perform five independent HCRAtion step. The mean ± standard deviation was calculated for the relative DRC values.

3.6. HCRA analysis in NSCLC patients

Strictly from the technical point of view, and independent of itsclinical implications, there are three important questions we haveaddressed in the NSCLC PBLs.

First, we explored the range of NDI and relative DRC valuesamong the NSCLC patients. The median relative DRC value was10.57, with values ranging from 3.04 to 16.99 (Fig. 6B). A 5.59-foldvariation in relative DRC was observed among our patients. TheNDI variable had a median value of 1.76 with a range of 1.38–1.95(Fig. 6C).

Second, we verified that the relative DRC values obtainedfrom NSCLC patients, using our DRC assay, distribute nor-mally. The relative DRC data was normally distributed forboth Kolmogorov–Smirnov (KS) (P > 0.10) and Shapiro–Wilk (SW)(P = 0.629) tests among the 35 NSCLC samples (Fig. 6D). In addition,the NDI values were distributed normally for either KS (P > 0.10) orSW (P = 0.782).

Finally, because the mitotic dynamics has been described as apotential confounding factor for DRC assessment, we discounted

of plasmid. Five lymphoblastoid cell lines and seven PBLs from donors were used tocompare the relative DRC values obtained from two independent batches of plasmid(New vs. Old). A strong correlation of relative DRC data was observed (R2 = 0.986,slope of 1.168; P < 0.0001) by linear regression analysis. Each point represents onesample.

P. Mendez et al. / DNA Repair 10 (2011) 603–610 609

Fig. 6. Description of transfection efficiency, relative DRC and NDI values in PBLs from NSCLC patients. (A) Transfection efficiency (striped columns) and cell viability (solidb s fromi to a nD of 0.0

ao

4

Dmuatrgqp

ofi

lack columns) are shown for the 35 primary cultures from the cryopreserved PBLn PBLs from NSCLC patients. A histogram of the relative DRC variable adjustment inRC and NDI was observed by linear regression analysis (panel E; R2 = 0.0016, slope

On the basis of these findings, future work to explore our HCRAs a potential biomarker of response or resistance in the treatmentf cancer patients, with DNA-damaging agents, is warranted.

. Advantages of our HCRA method

The HCRA assay is a very powerful technique for exploring theRC in either in vitro or population-based studies. However, ourodified HCRA method provides important advantages: First, the

se of nucleofection avoids the need to use adenoviruses or haz-rdous chemical compounds without affecting the sensitivity ofhe assay. Second, the maxGFP reporter gene does not requireadioactivity for quantitation. Finally, the expression of the reporterene is detected in living cells, and subsequent to the relative DRCuantitation, living cells can be sorted for additional genomics or

roteomics experiments.

In addition, we were able to reduce up to fourfold the numberf cells needed in each transfection reaction without affecting thenal relative DRC value.

the NSCLC patients used in the current study (B) the relative DRC and (C) the NDIormal distribution is represented in panel D. A lack of correlation between relative009; P = 0.80).

The use of flow cytometry to calculate relative DRC confers twoimportant advantages: dead cells are excluded from relative DRCanalysis, and the use of fluoro-labeled calibration particles to nor-malize the detection signal from photo-multiplier channels of thecytometer generates homogeneous inter-experimental DNA repairdata.

Our HCRA method can be implemented in large prospectivepopulation studies, as demonstrated by the strong correlation ofrelative DRC data obtained in a group of samples, measured withtwo independent batches of damaged and undamaged plasmid.Finally, to induce damage in plasmid DNA in the present study, weused a 254 nm UV-light wavelength, which specifically providesfunctional information from the NER pathway. However, depend-ing on the DNA-damaging agent used for plasmid DNA insult,different DNA repair pathways can be studied by the same tech-nique [4,17,27].

Conflict of interest statement

The authors declare that there are no conflicts of interest.

6 Repa

A

omSfaaeI

A

t

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

10 P. Mendez et al. / DNA

cknowledgements

We thank M.R. Spitz and Q. Wei, Department of Epidemiol-gy and M. D. Anderson Cancer Center, for comments on theanuscript. We also thank Drs. Antoni Matilla Duenas and Ivelisse

ánchez for their useful comments and suggestions. We thank Pro-essor J.J. Sanchez, Autonomous University of Madrid, for statisticalssistance, Kate Williams for assistance in drafting the manuscriptnd the oncology nursing staff for their help with patient bloodxtractions. This work has been funded by the Fundación para lanvestigación Clínica y Molecular del Cáncer de Pulmón.

ppendix A. Supplementary data

Supplementary data associated with this article can be found, inhe online version, at doi:10.1016/j.dnarep.2011.04.001.

eferences

[1] A.S. Neumann, E.M. Sturgis, Q. Wei, Nucleotide excision repair as a marker forsusceptibility to tobacco-related cancers: a review of molecular epidemiolog-ical studies, Mol. Carcinog. 42 (2005) 65–92.

[2] L.E. Wang, Z. Hu, E.M. Sturgis, M.R. Spitz, S.S. Strom, C.I. Amos, Z. Guo, Y. Qiao,A.M. Gillenwater, J.N. Myers, G.L. Clayman, R.S. Weber, A.K. El-Naggar, L. Mao,S.M. Lippman, W.K. Hong, Q. Wei, Reduced DNA repair capacity for removingtobacco carcinogen-induced DNA adducts contributes to risk of head and neckcancer but not tumor characteristics, Clin. Cancer Res. 16 (2010) 764–774.

[3] L.E. Wang, C. Li, S.S. Strom, L.H. Goldberg, A. Brewster, Z. Guo, Y. Qiao, G.L. Clay-man, J.J. Lee, A.K. El-Naggar, V.G. Prieto, M. Duvic, S.M. Lippman, R.S. Weber,M.L. Kripke, Q. Wei, Repair capacity for UV light induced DNA damage associ-ated with risk of nonmelanoma skin cancer and tumor progression, Clin. CancerRes. 13 (2007) 6532–6539.

[4] Y. Qiao, M.R. Spitz, Z. Guo, M. Hadeyati, L. Grossman, K.H. Kraemer, Q. Wei,Rapid assessment of repair of ultraviolet DNA damage with a modified host-cell reactivation assay using a luciferase reporter gene and correlation withpolymorphisms of DNA repair genes in normal human lymphocytes, Mutat.Res. 509 (2002) 165–174.

[5] H. Shen, M.R. Spitz, Y. Qiao, Z. Guo, L.E. Wang, C.H. Bosken, C.I. Amos, Q. Wei,Smoking, DNA repair capacity and risk of nonsmall cell lung cancer, Int. J. Cancer107 (2003) 84–88.

[6] Q. Wei, L. Cheng, C.I. Amos, L.E. Wang, Z. Guo, W.K. Hong, M.R. Spitz, Repairof tobacco carcinogen-induced DNA adducts and lung cancer risk: a molecularepidemiologic study, J. Natl. Cancer Inst. 92 (2000) 1764–1772.

[7] S. Ahmad, Platinum-DNA interactions and subsequent cellular processes con-trolling sensitivity to anticancer platinum complexes, Chem. Biodivers. 7 (2010)543–566.

[8] S.G. Chaney, S.L. Campbell, E. Bassett, Y. Wu, Recognition and processing ofcisplatin- and oxaliplatin-DNA adducts, Crit. Rev. Oncol. Hematol. 53 (2005)3–11.

[9] M. Selvakumaran, D.A. Pisarcik, R. Bao, A.T. Yeung, T.C. Hamilton, Enhancedcisplatin cytotoxicity by disturbing the nucleotide excision repair pathway inovarian cancer cell lines, Cancer Res. 63 (2003) 1311–1316.

10] X. Wu, W. Fan, S. Xu, Y. Zhou, Sensitization to the cytotoxicity of cisplatinby transfection with nucleotide excision repair gene xeroderma pigmentosun

[

ir 10 (2011) 603–610

group A antisense RNA in human lung adenocarcinoma cells, Clin. Cancer Res.9 (2003) 5874–5879.

11] B. Koberle, M.T. Tomicic, S. Usanova, B. Kaina, Cisplatin resistance: preclinicalfindings and clinical implications, Biochim. Biophys. Acta 1806 (2010) 172–182.

12] K.A. Olaussen, A. Dunant, P. Fouret, E. Brambilla, F. Andre, V. Haddad, E. Taran-chon, M. Filipits, R. Pirker, H.H. Popper, R. Stahel, L. Sabatier, J.P. Pignon, T.Tursz, T. Le Chevalier, J.C. Soria, DNA repair by ERCC1 in non-small-cell lungcancer and cisplatin-based adjuvant chemotherapy, N. Engl. J. Med. 355 (2006)983–991.

13] R. Rosell, K.D. Danenberg, V. Alberola, G. Bepler, J.J. Sanchez, C. Camps, M.Provencio, D. Isla, M. Taron, P. Diz, A. Artal, Ribonucleotide reductase mes-senger RNA expression and survival in gemcitabine/cisplatin-treated advancednon-small cell lung cancer patients, Clin. Cancer Res. 10 (2004) 1318–1325.

14] R. Rosell, L. Perez-Roca, J.J. Sanchez, M. Cobo, T. Moran, I. Chaib, M. Provencio,M. Domine, M.A. Sala, U. Jimenez, P. Diz, I. Barneto, J.A. Macias, R. de Las Penas,S. Catot, D. Isla, J.M. Sanchez, R. Ibeas, G. Lopez-Vivanco, J. Oramas, P. Mendez,N. Reguart, R. Blanco, M. Taron, Customized treatment in non-small-cell lungcancer based on EGFR mutations and BRCA1 mRNA expression, PLoS ONE 4(2009) e5133.

15] W.F. Athas, M.A. Hedayati, G.M. Matanoski, E.R. Farmer, L. Grossman, Develop-ment and field-test validation of an assay for DNA repair in circulating humanlymphocytes, Cancer Res. 51 (1991) 5786–5793.

16] M.A. Francis, A.J. Rainbow, UV-enhanced reactivation of a UV-damaged reportergene suggests transcription-coupled repair is UV-inducible in human cells,Carcinogenesis 20 (1999) 19–26.

17] R.J. Slebos, J.A. Taylor, A novel host cell reactivation assay to assess homol-ogous recombination capacity in human cancer cell lines, Biochem. Biophys.Res. Commun. 281 (2001) 212–219.

18] M.F. Anin, F. Gaucheron, M. Leng, Lability of monofunctional cis-platinumadducts: role of DNA double helix, Nucleic Acids Res. 20 (1992) 4825–4830.

19] C. Perez, M. Leng, J.M. Malinge, Rearrangement of interstrand cross-links intointrastrand cross-links in cis-diamminedichloroplatinum(II)-modified DNA,Nucleic Acids Res. 25 (1997) 896–903.

20] C. Hagemann, C. Meyer, J. Stojic, S. Eicker, S. Gerngras, S. Kuhnel, K. Roosen,G.H. Vince, High efficiency transfection of glioma cell lines and primary cellsfor overexpression and RNAi experiments, J. Neurosci. Methods 156 (2006)194–202.

21] K. Rothmann-cosic, H. Wessenfork, J. Helfrich, C. Thiel, G. Riemen, H. Brosterhus,Buffer solution for electroporation and a method comprising the use of the same(2002) Patent # 1390518.

22] M. Fenech, Cytokinesis-block micronucleus cytome assay, Nat. Protoc. 2 (2007)1084–1104.

23] M. Protic-Sabljic, K.H. Kraemer, One pyrimidine dimer inactivates expressionof a transfected gene in xeroderma pigmentosum cells, Proc. Natl. Acad. Sci.U.S.A. 82 (1985) 6622–6626.

24] M. D’Errico, A. Calcagnile, I. Iavarone, F. Sera, G. Baliva, L.M. Chinni, R. Corona,P. Pasquini, E. Dogliotti, Factors that influence the DNA repair capacity of nor-mal and skin cancer-affected individuals, Cancer Epidemiol. Biomarkers Prev.8 (1999) 553–559.

25] Q. Wei, J.E. Lee, J.E. Gershenwald, M.I. Ross, P.F. Mansfield, S.S. Strom, L.E. Wang,Z. Guo, Y. Qiao, C.I. Amos, M.R. Spitz, M. Duvic, Repair of UV light-induced DNAdamage and risk of cutaneous malignant melanoma, J. Natl. Cancer Inst. 95(2003) 308–315.

26] R.P. Rastogi, Richa, A. Kumar, M.B. Tyagi, R.P. Sinha, Molecular mechanisms of

ultraviolet radiation-induced DNA damage and repair, J. Nucleic Acids (2010)592980.

27] L. Wang, Q. Wei, Q. Shi, Z. Guo, Y. Qiao, M.R. Spitz, A modified host-cell reacti-vation assay to measure repair of alkylating DNA damage for assessing risk oflung adenocarcinoma, Carcinogenesis 28 (2007) 1430–1436.