The Promise of Pharmacogenomics: Gemcitabine and Pemetrexed

November 2, 2004

Although no overall differences in survival have been observed betweenthe many chemotherapy combinations in non–small-cell lungcancer, the clinical application of mRNA expression levels of amplifiedgenes may disclose many genetic influences on cytotoxic drug sensitivityand enable clinicians to tailor chemotherapy according to eachindividual’s gene profile. Specifically, the assessment of ribonucleotidereductase subunit M1 and thymidylate synthase mRNA expression levelsmight select patients who benefit from gemcitabine (Gemzar) orpemetrexed (Alimta) combinations. Until recently, clinical prognosticfactors such as performance status, weight loss, and lactate dehydrogenasewere the only parameters used to predict chemotherapy responseand survival. However, accumulated data indicate that overexpressionof genes involved in cancer glycolysis pathways plays an important role,and might be an independent mechanism of chemoresistance. Thedysregulation of glycolytic genes is affected by growth signals involvingthe PI3K/Akt pathway and downstream genes such as hypoxiainduciblefactor-1-alpha. One can thus envision that substantial improvementsin therapeutic outcome could benefit from the integrationof tailored ribonucleotide reductase-dependent chemotherapy, ribonucleotidereductase antisense therapy, and targeted therapy.

ABSTRACT: Although no overall differences in survival have been observed betweenthe many chemotherapy combinations in non–small-cell lungcancer, the clinical application of mRNA expression levels of amplifiedgenes may disclose many genetic influences on cytotoxic drug sensitivityand enable clinicians to tailor chemotherapy according to eachindividual’s gene profile. Specifically, the assessment of ribonucleotidereductase subunit M1 and thymidylate synthase mRNA expression levelsmight select patients who benefit from gemcitabine (Gemzar) orpemetrexed (Alimta) combinations. Until recently, clinical prognosticfactors such as performance status, weight loss, and lactate dehydrogenasewere the only parameters used to predict chemotherapy responseand survival. However, accumulated data indicate that overexpressionof genes involved in cancer glycolysis pathways plays an important role,and might be an independent mechanism of chemoresistance. Thedysregulation of glycolytic genes is affected by growth signals involvingthe PI3K/Akt pathway and downstream genes such as hypoxiainduciblefactor-1-alpha. One can thus envision that substantial improvementsin therapeutic outcome could benefit from the integrationof tailored ribonucleotide reductase-dependent chemotherapy, ribonucleotidereductase antisense therapy, and targeted therapy.

Host-specific factors modulatesusceptibility to tobaccoinducedcarcinogenesis, includingvariations in DNA repair thatmay influence the rate of removal ofboth DNA damage and fixation ofmutations. In a seminal study by Weiet al,[1] DNA repair capacity (DRC)was measured in peripheral blood lymphocytes(PBLs) with the host-cell reactivationassay that measures cellularreactivation of a reporter gene from aplasmid damaged by exposure to 75μM benzo(a)pyrene diol epoxide, andtransfected into cultured cells. Theplasmids undergo intracellular repair,leading to increased gene expression.Because benzo(a)pyrene-DNA lesionsare repaired via nucleotide excisionrepair (NER), the degree of reportergene expression is believed to reflectthe activity of this repair pathway.The mean level of DRC in lung cancercases (3.3%) was significantlylower than in controls (5.1%; P < .01).Younger cases (< 65 years old) andsmokers were more likely than controlsto have reduced levels of DRC.[1]The investigators confirmed thesefindings in a case-control study of316 newly diagnosed lung cancer patientsand 316 cancer-free controls.Case patients who were younger atdiagnosis, female, lighter smokers, orwith a family history of cancer exhibited the lowest DRC, suggesting thatthese subgroups are especially susceptibleto lung cancer.[2] ReducedDRC and increased DNA adduct levelsare thus associated with an increasedrisk of lung cancer.NER is a principal pathway forrepair of DNA adducts that are inducedby smoking-related carcinogens,and also the primary mechanismfor removing cisplatin adducts.[3] TheNER molecular machinery includesproteins that are mutated in xerodermapigmentosum (XP) and Cockaynesyndrome patients. In the global genomerepair pathway, the protein complexXPC-HHR23B, which appearsto be essential for the recruitment ofall subsequent NER factors in the preincisioncomplex, binds to damagedDNA. Then, the multicomponent transcriptionfactor TFIIH that is responsiblefor unwinding the damagedregion of the DNA is recruited. Next,XPG nuclease cleaves the DNA on its3' end. Following DNA cleavage,XPA/RPA proteins join the complexand recruit the ERCC1-XPF complex,which cleaves the 5' end.[3] Polymorphismsof a number of DNA repair genes that are involved in theNER pathway potentially affect proteinfunction and, subsequently, DRC.In lung cancer patients, reducedlevels of XPG and CSB expressionhave been observed in peripheral lymphocytes.[4] Moreover, ERCC1 andXPD mRNA levels in lymphocytescorrelate with DRC, and could thusbe useful surrogates of DRC.[5] Extensivereviews of DRC-related geneshave been presented elsewhere.[6-11]Prognostic FactorsIn a retrospective analysis of 2,531non-small-cell lung cancer (NSCLC)patients with extensive disease (eitherdistant metastases or locoregionalrecurrence after definitiveradiotherapy) treated in a SouthwestOncology Group (SWOG) trial, Albainet al performed Cox modelingand recursive partitioning and amalgamation(RPA) to determine independent,predictive outcomefactors.[12] Patients received treatmentbetween 1974 and 1988. Performancestatus (PS) was defined as good (SWOG 0-1, no symptoms or withsymptoms but fully ambulatory) orpoor (SWOG 2-4, nonambulatory).Good PS, sex (female), and age ≥ 70years were significant independentpredictors.In a second Cox model for patientswith good PS, hemoglobin levels ≥11 g/dL, normal calcium, and a singlemetastatic site were significantly favorablefactors. The use of cisplatinwas also an additional independentpredictor of improved outcome. AnRPA performed in 904 patients frommore recent SWOG trials, almost allof whom were treated with cisplatin,revealed three distinct subsets basedon PS, age, hemoglobin, and lactatedehydrogenase (LDH) in which 1-yearsurvivals were 27%, 16%, and 6%.Until 1980, additional variables, suchas weight loss (< 10 or ≥ 10 pounds),were not included in final analysis.In the multivariate survival analysesby prognostic variables and therapydiscriminants, the significantlyfavorable SWOG predictive factorswere PS 0-1, sex (female), age ≥ 70years, (≥ 45 years in female patients),a single metastatic lesion, < 10 poundsof weight loss, normal LDH, normalalkaline phosphatase, and hemoglobin≥ 11 g/dL. Median survival was 1to 3 months better for the good PS,female, single-lesion, and cisplatinbasedtherapy categories.[12] Intriguingly,LDH was important in the poorPS subset; patients with a normal LDHlevel and poor PS had a survival outcomesimilar to other subsets with agood PS.In an analysis of 1,052 patientsincluded in clinical trials conductedby the European Lung Cancer WorkingParty (ELCWP), Paesman and coworkersused a Cox regression modelto determine following predictive variables:Karnofsky PS (≥ 80 = SWOG0-1; ≤ 70 = SWOG 2-4), neutrophilcounts, metastatic involvement ofskin, serum calcium level, age andgender, as well as disease extent; patientswith stage I-III disease wereincluded in the analysis.[13] An RPAdefined the best subgroup of patientsas females with limited disease and aKarnofsky PS ≥ 80.In a third, smaller, analysis thatincluded a homogenous group of stageIII unresectable or inoperable patientsreceiving cisplatin at 120 mg/m2 plusvinca alkaloid combination chemotherapy,a multivariate analysis byO'Connell and colleagues highlightedthe following outcome parameters(and associations): initial PS, with patientswith good PS (increased objectiveresponse/survival); bonemetastases (decreased response rateand survival); elevated LDH and sex(male) (shortened survival); and ≥ 2extrathoracic metastatic organ sites(shortened survival).[14] When objectiveresponse following chemotherapywas included in their analysis, theauthors reported a strong associationwith increased survival.Multiple clinical characteristics influenceprognosis, such as PS, hemoglobinlevel, and bone, liver, or skinmetastases. Gender, LDH, and weightloss also influence survival. New piecesof information should shed light onthese prognostic factors. First, thereare interindividual differences in DRCwhich are accentuated according toage and gender, with females havinga reduced DRC.[2] Based on DRC,females would thus have greaterchemosensitivity than males. Also,data indicate that in vitro intrinsic cisplatinresistance is associated with elevatedlevels of DRC in NSCLCcells.[15] DRC is a surrogate of theNER pathway that eliminates cisplatinadducts,[3,5] and NSCLC patientswith effective systemic (host) DRCare reported to achieve poorer survivalthan patients with suboptimalDRC.[16]In a study (N = 375) by Boskin etal, patients in the top DRC quartile ofthe group (DRC > 9.2%) had a risk ofdeath exceeding two times that of subjectsin the bottom quartile (DRC < 5.8%; P = .01).[16] Median survivalwas 8.9 months for patients in the topDRC quartile, compared with 15.8months for those in the bottom quar-tile (P = .04).Earlier findings suggest that theformation and persistence of cisplatinor carboplatin adducts either in buccalcells or leukocytes predict betterresponse.[17,18] Schaake-Kong et alexamined cisplatin-DNA adducts inthe nuclei of buccal cells in 27 patientswith unresectable NSCLC receivingradical radiotherapy and dailyadministration of low-dose cisplatin.[19] Nuclear staining was performedin buccal cells collected 1 hourafter cisplatin on the fifth treatmentday (after five daily doses of cisplatinat 6 mg/m2). Cisplatin-DNA adductstaining remained a significant independentpredictor of survival: patientswith low levels of induced DNA adductshad a meager median survivalof 5 vs 30 months for patients withelevated DNA adduct levels.[20]Thus, beyond stratification for genderin future clinical trials, measuringDRC by either functional assays orsurrogates like ERCC1 or XPDmRNA in peripheral lymphocytes[5]might help to predict treatment responders,as reported previously.[14]Alberola and colleagues reportedin a prospective Spanish Lung CancerGroup (SLCG) phase III studyincluding 557 NSCLC patients treatedwith cisplatin-containing regimensthat PS, gender and weight loss weresignificant prognostic factors.[21]Note that weight loss in lung cancerpatients was also previously reportedto be predictive of both reduced outcomeand survival.[12]Upregulation of Genes in theGlycolysis and Krebs CyclePathwayUnlike normal mammalian cellsthat require oxygen in energy-generatingpathways, cancer cells rely onanaerobic pathways of glycolysis.Lung cancer patients with weight losshave elevated levels of 3-phosphoglycerateand phosphoenolpyruvate-both components of the glycolyticpathway (glyconeogenesis).[22] Furthermore,c-Myc and HIF-1 overexpressionderegulates glycolysisthrough the activation of the glucosemetabolic pathway, which regulateslactate dehydrogenase and induces lactateoverproduction.[23]Note that O'Connell et al reportedthat elevated serum LDH was associatedwith shortened survival and remissionduration.[14] Elevated mRNAlevels of phosphofructokinase, glyceraldehyde-3-phosphate dehydrogenase,phosphoglycerate kinase, andenolase were also reported, indicatingthe activation of several componentsin the glucose metabolicpathway.[23]Chen et al[24] reported the systematicidentification of lung adenocarcinomaproteins using twodimensionalpolyacrylamide gel electrophoresis(PAGE) and mass spectrometry,and found that at least fourproteins (phosphoglycerate kinase 1,phosphoglycerate mutase, alpha enolase,and pyruvate kinase M1-allcomponents of glycolysis) had increasedexpression levels. They werealso associated with poor survival inresected lung adenocarcinoma.[24]Expression of phosphoglyceratekinase 1, the sixth enzyme of the glycolyticpathway, reflects increasedglycolysis in tumor cells, and is relatedto the induction of a multidrugresistant (MDR) phenotype distinctfrom MDR1. The hypoxic nature ofthe solid tumor triggers vascular endothelialgrowth factor (VEGF) expressionthat both stimulatesangiogenesis and activates glycolyticenzymes, including phosphoglyceratekinase, thus facilitating anaerobic productionof adenosine triphosphate(ATP).[25]However, cancer cells also maintainhigh aerobic glycolytic rates andproduce high levels of lactate andpyruvate, the end products of glycolysis(Warburg effect); thus, preferentialreliance on glycolysis is correlatedwith disease progression as reportedby Lu et al.[26] The selective adaptationof cancer cells to hypoxia is alsomediated via HIF-1, which upregulatesa series of genes involved inglycolysis, angiogenesis, cell survival,and erythropoiesis. Glycolytic metabolismsubsequent to glyceraldehydes-3-phosphate dehydrogenaseleads to pyruvate in cells withactive pyruvate kinase.[26] Similarly,elevated expression of hexokinaseand phosphofructokinase is a hallmarksof glycolytic pathways of cancercells.Interestingly, phosphofructokinaseactivity is stimulated by insulin, insulinlikegrowth factor-1, and epidermalgrowth factor receptor (EGFR),all of which induce HIF-1 under normoxiainvolving the PI3K signalingpathway.[26] Phosphoglycerate mutaseand enolase activities have alsobeen found elevated in lung cancer(as opposed to colon, liver, and nonendocrinelung tumors).[27] Accordingly,HIF-1-alpha (and HIF-2-alpha)overexpression is a common event inthe natural history of NSCLC, alsorelated to the upregulation of variousangiogenic factors and with poor prognosis.[28] Interestingly, signal transductionfrom HER2 to PI3K, Akt, andFKBP-rapamycin-associated protein(FRAP) increases the rate of HIF-1-alpha synthesis, whereas hypoxia andloss of p53 activity decrease the rateof HIF-1-alpha degradation.[29] Akt/protein kinase B (PKB) is constituitivelyactive in NSCLC cells and promoteschemoresistance (and resistanceto radiation-induced apoptosis).[30]Brabender et al reported that in 83surgically resected NSCLC patients,high levels of HER2-neu mRNA expressionas well as high EGFR/HER2-neu mRNA coexpression weresignificant, independent prognosticfactors, ie, their expression definedlow- and high-risk groups for treatmentfailure in curatively resectedNSCLC.[31]In Figure 1 we postulate a modelillustrating the significance and potentiallyclose relationship of increasedmRNA or protein expression of glycolyticgenes and HIF-1-alpha/VEGFto clinical outcome, beyond the influenceof serum LDH and weight loss.Clearly, this molecular-based modeland its hypotheses require validationin clinical trials.A surrogate marker of the cancerglycolytic pathway could be positronemission tomography (PET), whoseimaging capability can detect the bio-chemical and physiologic processesthat occur in tissues. The most frequentlyused positron emitting radiopharmaceuticalis 18-fluor, labeled2-deoxy-D-glucose (18F-FDG), a radioactivelylabeled glucose analog.Because cancer cells presumably exhibita higher glycolytic rate than nonneoplastictypes, the clinicalapplication of 18F-FDG-PET is promisingand potentially warranted. It isreasonable that higher tumor uptakeof the radiolabeled glucose analogmight be a surrogate of the glycolyticpathway. Using this rationale/technology,clinicians have employed PETimaging to assess changes in intracellulartumor glucose utilization duringthe course of chemotherapy.In the NSCLC arena (n = 57), Weberet al reported that median time toprogression (P = .0003) and overallsurvival (P = .005) were significantlylonger for 18FDG6-PET metabolic responders(vs nonresponders) in theinterval before and after the first chemotherapycycle.[32] One-year survivalrates were 44% and 10%.Overexpression of ERCC1 mRNAand other NER genes has been associatedwith repair of cisplatin-inducedDNA damage and clinical resistanceto cisplatin. In a small study of advancedNSCLC patients (n = 56 stageIIIb/IV) receiving a cisplatin/gemcitabineregimen, Lord and collaborators[33] reported that PS, weightloss, and low ERCC1 mRNA expressionwere independent prognostic factors.Indeed, ERCC1 mRNA levelswere more significant factors than PSin a Cox proportional hazards multivariableanalysis.[33] Median survivalfor patients with low ERCC1mRNA expression was significantlylonger (61.6 weeks; 95% confidenceinterval [CI] = 42.4-80.7 weeks) comparedto those exhibiting high expression(20.4 weeks, 95% CI = 6.9-33.9weeks).Ribonucleotide Reductase andThymidylate Synthase mRNAExpressionRibonucleotide reductase (RR),which is increased in cancer cells, isanother potentially new predictive andprognostic marker. The M2 (or RRM2) subunit is directly involvedin many cellular signaling pathways.[34] For example, the active chemotherapeuticagent gemcitabine(Gemzar) decreases ribonucleotidereductase activity. In a retrospectiveanalysis of tissue using quantitativepolymerase chain reaction (PCR) analysis(n = 75) as part of a randomizedtrial, our group reported better time toprogression (P = .05) and survival(P = .0028) in gemcitabine/cisplatintreatedmetastatic NSCLC patients(n = 22) who had low ribonucleotidereductase subunit M1 (or RRM1)mRNA expression; the other studyarms received vinorelbine (Navelbine)/cisplatin (n = 25) and paclitaxel/carboplatin (Paraplatin) (n =28).[35]

Interestingly, retinoblastoma proteinis sequentially phosphorylated bycyclin D-CDK4/6 and cyclin E/CDK2during G1/S cell-cycle transitionphase. This modification leads to thedissociation of retinoblastoma fromE2F/DP heterodimers, leaving themin a transcriptionally active state thatregulates several DNA synthesis enzymesalso involved in chemotherapyresponse, such as dihydrofolatereductase (DHFR), thymidylate synthase(TS), and RR.[36] In small-celllung cancer (SCLC) and NSCLC, p16/INK4A and retinoblastoma are reciprocallyinactivated, resulting in theinactivation of the same p16/INK4A/RB pathway.[37] Therefore, in prospectivestudies, the predictive and/orprognostic value of certain transcriptsshould be kept in mind. For example,in ribonucleotide-dependent chemotherapycombinations, the role ofRRM1 and TS mRNA levels mightinfluence response and survival, asproposed in the model illustrated inFigure 1.Several studies have shown thatthe status of TS is a predictive markerfor response to 5-FU in different primarycancers. The pioneering studyby Lenz et al[38] reported that TSmRNA levels were associated withresponse and survival in 65 gastriccancer patients treated with 5-FU andcisplatin. Thymidylate synthase wasexamined by quantitative PCR(QPCR), with beta-actin gene expressionas an internal control. Tumor tissuewas obtained from endoscopicbiopsies of gastric adenocarcinomas,and TS expression was examined fromfresh tissue. The mean gastric cancerTS mRNA level in responding andresistant patients was statistically significant(P < .001).Patients with high TS mRNA levels(> 4.6) had a median survival of 6months, while for those with lowerlevels (< 4.6), median survival wasnot yet reached at 43 months of follow-up.[38] The mean pretherapy TSmRNA level in all 65 patients was4.6. When TS values were brokendown by the ethnicity of the patients,the mean was 4.3 for Hispanics, 5.6for Asians, 3.8 for Caucasians, and6.1 for African-Americans. Due to thesmall number of patients, these differenceswere not statistically significant,but they provided an indicationthat racial differences should be takeninto account in these and futureanalyses. Differences in response andsurvival according to TS levels havealso been observed in colorectal, lung,and breast cancer (reviewed elsewhere[39]). More recently, TS mRNAlevels have been examined from microdissectedparaffin-embedded tissuesamples, facilitating the routine useof QPCR.Since the combination of pemetrexed(Alimta) plus cisplatin has becomethe standard treatment inmalignant pleural mesotheliomas (alsoapproved by the US Food and DrugAdministration for that indication),[40] one potential approach toimprove therapeutic outcome wouldbe to identify pemetrexed target geneexpressionchanges in cancerous mesothelialcells. The diagnosis ofmesothelioma often requires a thoracoscopicor open pleural biopsy thatcan provide a rich source of tumortissue for RNA isolation. TS is theprincipal marker of pemetrexed sensitivity,although pemetrexed also inhibitsdihydrofolate reductase and thepurine biosynthetic enzyme glycinamideribonucleotide formyltransferase.In an MDA 435 breast cancer-derived cell line, Longley et al reportedthat TS upregulation highly inhibitedpemetrexed while the growthinhibitory effects of irinotecan, cisplatin,oxaliplatin, and paclitaxel wereunaffected by TS upregulation.[41](The pemetrexed IC50 dose was unobtainablewhen TS was overexpressed,indicating that pemetrexed toxicitywas highly sensitive to increased TSexpression.) In vitro and in vivo studieshave also demonstrated a strongassociation between increased TS expressionand the development of resistanceto fluorouracil (5-FU) andraltitrexed (Tomudex).

Metzger and coworkers reportedthat TS mRNA levels and ERCC1plus TS mRNA levels strongly predictedsurvival in 5-FU/cisplatin-treatedgastric cancer patients (n = 38).[42]When both ERCC1 and TS mRNAlevels were below their medians, 85%(11 of 13) of patients responded; withboth ERCC1 and TS mRNA levelsabove, responses were 20% (2 of 10)(P = .003). Other upstream determinantsof 5-FU chemosensitivity includethe 5-FU-degrading enzymedihydropyrimidine dehydrogenaseand 5-FU-anabolic enzymes, such asorotate phosphoribosyl transferase.However, it is likely that events downstreamof TS inhibition, such as activationof DNA damage responsepathways (eg, ERCC1), also play keyroles in determining the cellular responseto TS inhibitors.[42]Customized ChemotherapyIn addition to the individual predictiverole of ERCC1 mRNA [33]and RRM1 mRNA [35], in a thirdstudy, we examined both ERCC1 andRRM1 mRNA expression in pretreatmentbronchial biopsies from gemcitabine/cisplatin-treated advancedNSCLC patients who were part of alarge randomized trial.[21] There wasa strong correlation between ERCC1and RRM1 mRNA expression levels(R = 0.4; P < .001). Patients with lowRRM1 mRNA levels had significantlylonger median survival than thosewith high levels (13.7 vs 3.6 months;P = .009). There were no significantdifferences according to ERCC1mRNA levels, although there was atendency toward longer median survivalamong patients who expressedlow ERCC1 levels.[43]Finally, in a fourth study of neoadjuvantgemcitabine/cisplatin-treatedstage III NSCLC patients, those patientswith the lowest RRM1 levels (in the bottom quartile) had a decreasedrisk of death compared withthose in the top quartile (risk ratio =0.30; P = .033). Median survival forthese 17 patients in the bottom quartilewas 52 months, whereas the 15 inthe top quartile had a median survivalof 26 months (P = .018).[44] Theseobservations have been recently confirmedat the preclinical level. Fromexpression profiling of two gemcitabine-resistant NSCLC cell lines, increasedexpression of RRM1 mRNAwas detected in both. QuantitativePCR analysis demonstrated that resistantcells had a greater than 125-foldincrease in RRM1 mRNA expression.[45] The accumulated evidenceindicates that RRM1 may serve as apredictive marker of gemcitabine andcisplatin response.We have also analyzed BRCA1mRNA expression in processed formalin-fixed, paraffin-embedded tissuesfrom resected lung cancerpatients and demonstrated thatBRCA1 expression can be accuratelyassessed. BRCA1 gene expression wasdetectable in all 55 samples analyzedin this study. Patients in the bottomquartile of BRCA1 mRNA levels (< 0.61) obtained the maximum benefitof neoadjuvant gemcitabine/cisplatinchemotherapy, while those in the topquartile (> 2.45) had the poorest outcome.[46] These findings support thehypothesis that BRCA1 mRNA expressionlevels could be an indicatorof differential cisplatin sensitivity inNSCLC, which is consistent with findingsin preclinical models in breastcancer.[47]For instance, the HCC1937 cell linefrom a primary breast carcinoma witha germline BRCA1 mutation wastransfected with either wild-typeBRCA1 or an empty vector to testresponse to antimicrotubule drugs (paclitaxeland vinorelbine) and DNAdamagingdrugs (cisplatin, bleomycin[Blenoxane], and etoposide). Reconstitutionof wild-type BRCA1 functioninto HCC1937 resulted in a1,000-fold increase in sensitivity topaclitaxel and a 10,000-fold increasein sensitivity to vinorelbine. Conversely,it resulted in a 2-fold increase inresistance to bleomycin, a 20-fold increasein resistance to cisplatin, and agreater than 100-fold increase in resistanceto etoposide.[47] Interestingly,BRCA1 failed to modulateresistance or sensitivity to the antimetabolite5-FU, perhaps reflecting thedistinct mode of action of antimetabolites.[47]Several cisplatin-based doubletshave demonstrated similar survivalin several randomized studies of advancedNSCLC. Furthermore, otherstudies have found no survivaldifferences between cisplatin aloneand cisplatin/paclitaxel or betweendocetaxel [Taxotere] alone and docetaxel/cis-platin. Based on our results[46] and on preclinicaldata[47], we can reason that patientswith low BRCA1 mRNA levels canbenefit from gemcitabine/cisplatin,whereas those with high levels couldbenefit from single-agent docetaxelor paclitaxel. In contrast, highBRCA1 levels may diminish thesynergism between the taxanes andcisplatin or carboplatin. While sensitivityto antimetabolites, such asgemcitabine, may not be affectedby BRCA1 levels, gemcitabine/cisplatinsynergism may be partiallyabrogated in tumors with highBRCA1 mRNA levels; on the otherhand, these tumors may benefit fromthe synergism observed between thetaxanes and gemcitabine.To date, no other clinical studyhas assessed BRCA1 mRNA expressionas a predictive marker of chemotherapyresponse in lung cancer.If further research validates ourfindings, BRCA1 mRNA assessmentwill provide an important toolfor customizing NSCLC chemotherapyin order to improve survival.Figure 2 illustrates the schema ofthe new Spanish Lung Cancer Groupcustomized chemotherapy trial inadvanced NSCLC.ConclusionsCustomized chemotherapy opensnew avenues of translational research,where the main difficulty is the availabilityof tumor tissue for gene expressionanalyses. In vitro and clinicalevidence point out that customizedchemotherapy is a promising approachto be validated.


The authors have nosignificant financial interest or other relationshipwith the manufacturers of any productsor providers of any service mentioned in thisarticle.



Wei Q, Cheng L, Hong WK, et al: ReducedDNA repair capacity in lung cancer patients.Cancer Res 56:4103-4107, 1996.


Wei Q, Cheng L, Amos CI, et al: Repairof tobacco carcinogen-induced DNA adductsand lung cancer risk: A molecular epidemiologicstudy. J Natl Cancer Inst 92:1764-1772,2000.


Furuta T, Ueda T, Aune G, et al: Transcription-coupled nucleotide excision repair asa determinant of cisplatin sensitivity of humancells. Cancer Res 62:4899-4902, 2002.


Cheng L, Spitz M, Hong WK, et al: Reducedexpression levels of nucleotide excisionrepair genes in lung cancer: A case-controlanalysis. Carcinogenesis 21:1527-1530, 2000.


Vogel U, Dybdahl M, Frentz G, et al: DNArepair capacity: Inconsistency between effectof over-expression of five NER genes and thecorrelation to mRNA levels in primarylymphocyites. Mutat Res 461:197-210, 2000.


Rosell R, Lord RVN, Taron M, et al: DNArepair and cisplatin resistance in non-small-celllung cancer. Lung Cancer 38:217-227, 2002.


Rosell R, Taron M, Barnadas A, et al:Nucleotide excision repair pathways involvedin cisplatin resistance in non-small-cell lungcancer. Cancer Control 10:297-305, 2003.


Rosell R, Taron M, Alberola V, et al: Genetictesting for chemotherapy in non-smallcelllung cancer. Lung Cancer 41(suppl):S97-S102, 2003.


Rosell R, Crino L, Danenberg K, et al:Targeted therapy in combination withgemcitabine in non-small-cell lung cancer.Semin Oncol 30(suppl 10):19-25, 2003.


Rosell R, Taron M, O’Brate A: Predictivemolecular markers in non-small-cell lungcancer. Curr Opin Oncol 13:101-109, 2001.


Rosell R, Monzo M, O’Brate A, et al:Translational oncogenomics: Toward rationaltherapeutic decision-making. Curr Opin Oncol14:171-179, 2002.


Albain KS, Crowley JJ, LeBlanc M, etal: Survival determinants in extensive-stagenon-small-cell lung cancer: The SouthwestOncology Group experience. J Clin Oncol9:1618-1626, 1991.


Paesmans M, Sculier JP, Libert P, et al:Prognostic factors for survival in advanced nonsmallcell lung cancer: Univariable and multivariateanalyses including recursive partitioningand amalgamation algorithms in 1052 patients.J Clin Oncol 13:1221-1230, 1995.


O'Connell JP, Kris MG, Gralla RJ, et al:Frequency and prognostic importance of pretreatmentclinical characteristics in patientswith advanced non-small-cell lung cancertreated with combination chemotherapy. J ClinOncol 4:1604-1614, 1986.


Zeng-Rong N, Paterson J, Alpert L, etal: Elevated DNA repair capacity is associatedwith intrinsic resistance of lung cancer to chemotherapy.Cancer Res 55:4760-4764, 1995.


Bosken CH, Wei Q, Amos CI, et al: Ananalysis of DNA repair as a determinant of survivalin patients with non-small-cell lung cancer.J Natl Cancer Inst 94:1091-1099, 2002.


Blommaert FA, Michael C, TerheggenPMAB, et al: Drug-induced DNA modificationin buccal cells of cancer patients receivingcarboplatin and cisplatin combination chemotherapy,as determined by an immunocytochemicalmethod: Interindividual variationand correlation with disease response. CancerRes 53:5669-5675, 1993.


Schellens JH, Ma J, Planting AS, et al:Relationship between the exposure to cisplatin,DNA-adduct formation in leucocytes and tumourresponse in patients with solid tumours.Br J Cancer 73:1569-1575, 1996.


Schaake-Koning C, Van Den Bogaert W,Dalesio O, et al: Effects of concomitantcisplatin and radiotherapy on inoperable nonsmall-cell lung cancer. N Engl J Med 326:524-530, 1992.


Van de Vaart PJ, Belderbos J, De JongD, et al: DNA-adduct levels as a predictor ofoutcome for NSCLC patients receiving dailycisplatin and radiotherapy. Int J Cancer 89:160-166, 2000.


Alberola V, Camps C, Provencio M, etal: Cisplatin plus gemcitabine versus acisplatin-based triplet versus nonplatinum sequentialdoublets in advanced non-small-celllung cancer: A Spanish Lung Cancer Groupphase III randomized trial. J Clin Oncol21:3207-3213, 2003.


Leij-Halfwerk S, Ven Den Berg JW,Sijens PE, et al: Altered hepatic glucogenesisduring L-Alanine infusion in weight-losing lungcancer patients as observed by phosporus magneticresonance spectroscopy and turnovermeasurements. Cancer Res 60:618-623, 2000.


Osthus RC, Shim H, Kim S, et al: Deregulationof glucose transporter 1 and glycolyticgene expression by c-Myc. J Biol Chem275:21797-21800, 2000.


Chen G, Gherib TG, Wang H, et al: Proteinprofiles associated with survival in lungadenocarcinoma. Proc Natl Acad Sci USA100:13537-13542, 2003.


Lay AJ, Jiang XM, Kisker O, et al: Phosphoglyceratekinase acts in tumour angiogenesisas a disulphide reductase. Nature 408:869-873, 2000.


Lu H, Forbes RA, Verma A: Hypoxiainduciblefactor 1 activation by aerobic glycolysisimplicates the Warburg effect in carcinogenesis.J Biol Chem 277:23111-23115, 2002.


Durany N, Joseph J, Campo E, et al:Phosphoglycerate mutase, 2-3-bisphosphoglyceratephosphatase and enolase activityand isoenzymes in lung, colon and liver carcinomas.Br J Cancer 75:969-977, 1997.


Giatromanolaki A, Koukourakis MI,Sivridis E, et al: Relation of hypoxia induciblefactor 1α and 2α in operable non-small celllung cancer to angiogenic/molecular profile oftumours and survival. Br J Cancer 85:881-890,2001.


Laughner E, Taghavi P, Chiles K, et al:HER2 (neu) signaling increases the rate of hypoxia-inducible factor 1α (HIF-1α) synthesis:Novel mechanism for HIF-1-mediated vascularendothelial growth factor expression. MolCell Biol 21:3995-4004, 2001.


Brognard J, Clark AS, Ni Y, et al: Akt/protein kinase B is constitutively active in nonsmallcell lung cancer cells and promotes cellularsurvival and resistance to chemotherapyand radiation. Cancer Res 61:3986-3997, 2001.


Brabender J, Danenberg KD, MetzgerR, et al: Epidermal growth factor receptor andHER2-neu mRNA expression in non-small celllung cancer is correlated with survival. ClinCancer Res 7:1850-1855, 2001.


Weber WA, Petersen V, Schmidt B, et al:Positron emission tomography in non-smallcelllung cancer: Prediction of response to chemotherapyby quantitative assessment of glucoseuse. J Clin Oncol 21:2651-2657, 2003.


Lord RVN, Brabender J, Gandara D, etal: Low ERCC1 expression correlates with prolongedsurvival after cisplatin plus gemcitabinechemotherapy in non-small cell lung cancer.Clin Cancer Res 8:2286-2291, 2002.


Lee Y, Vassilakos A, Feng N, et al: GTI-2040, an antisense agent targeting the smallsubunit component (R2) of human ribonucleotidereductase, shows potent antitumor activityagainst a variety of tumors. Cancer Res63:2802-2811, 2003.


Rosell R, Scagliotti G, Danenberg KD,et al: Transcripts in pretreatment biopsies froma three-arm randomized trial in metastatic nonsmall-cell lung cancer. Oncogene 22:3548-3553, 2003.


Sowers R, Toguchida J, Qin J, et al:mRNA expression levels of E2F transcriptionfactors correlate with dihydrofolate reductase,reduced folate carrier, and thymidylate synthasemRNA expression in osteosarcoma. Mol CancerTher 2:535-541, 2003.


Osada H, Takahashi T: Genetic alterationsof multiple tumor suppressors andoncogenes in the carcinogenesis and progressionof lung cancer. Oncogene 21:7421-7434,2002.


Lenz HJ, Leichman CG, Danenberg KD,et al: Thymidylate synthase mRNA level inadenocarcinoma of the stomach: A predictorfor primary tumor response and overall survival.J Clin Oncol 14:176-182, 1995.


Sarries C, Haura EB, Roig B, et al:Pharmacogenomic strategies for developingcustomized chemotherapy in non-small-celllung cancer. Pharmacogenomics 3:763-780,2002.


Vogelzang NJ, Rusthoven JJ,Symanowski J, et al: Phase III study ofpemetrexed in combination with cisplatin versuscisplatin alone in patients with malignantpleural mesothelioma. J Clin Oncol 21:2636-2644, 2003.


Longley DB, Ferguson PR, Boyer J, etal: Characterization of a thymidylate synthase(TS)-inducible cell line: A model system forstudying sensitivity to TS- and non-TS-targetedchemotherapies. Clin Cancer Res 7:3533-3539,2001.


Metzger R, Leichman CG, DanenbergKD, et al: ERCC1 mRNA levels complementthymidylate synthase mRNA levels in predictingresponse and survival for gastric cancerpatients receiving combination cisplatin andfluorouracil chemotherapy. J Clin Oncol16:309-316, 1998.


Rosell R, Danenberg KD, Alberola V, etal: Ribonucleotide reductase messenger RNAexpression and survival in gemcitabine/cisplatin-treated non–small-cell lung cancerpatients. Clin Cancer Res 10:1318-1325, 2004.


Rosell R, Felip E, Taron M, et al: Geneexpression as a predictive marker of outcomein stage IIB-IIIA-IIIB non–small-cell lung cancerafter induction gemcitabine-based chemotherapyfollowed by resectional surgery. ClinCancer Res 10:4215S-4219S, 2004.


Davidson JD, Liandong M, Flagella M,et al: An increase in the expression of ribonucleotidereductase large subunit 1 is associatedwith gemcitabine resistance in non–small-celllung cancer cell lines. Cancer Res 64:3761-3766, 2004.


Taron M, Rosell R, Felip E, et al: BRCA1mRNA expression levels as an indicator ofchemoresistance in lung cancer. Hum MolGenet 13:2443-2449, 2004.


Quinn JE, Kennedy RD, Mullan PB, etal: BRCA1 functions as a differential modulatorof chemotherapy-induced apoptosis. CancerRes 63:6221-6228, 2003.