(CANCERRESEARCH58, 5340-5343, DecemberI. 19981

Advances in Brief

Catalytic Efficiencies of Allelic Variants of Human Glutathione S-transferase P1—itoward Carcinogenic anti-Diol Epoxides of Benzo[c]phenanthreneand Benzo[g]chrysene'

Xun Hu, Hong Xia, Sanjay K. Srivastava,Ajai Pal, YogeshC. Awasthi,Piotr Zirnniak, and ShivendraV. Singh2Cancer Research Laboratory, Mercy Cancer Institute. Mercy Hospital of Pittsburgh. Pittsburgh, Pennsylvania 15219 IX. H.. H. X., S. K. S.. A. P., S. V. S.!: Department of HwnanBiological Chemistry and Genetics. The University of Texas Medical Branch. Galveston, Texas 77555 fY. C. A.): and Departments of Medicine. and Biochemistry and MolecularBiology. University ofArkansas for Medical Sciences and McClellan Veterans Affairs Hospital Medical Research, Little Rock. Arkansas 72205 (P. ZJ

Abstract

Four allelic variants of glutathione (GSH) S-transferase P1—i(hGSTP1—1)that differ in their structures at amino acid(s) In position(s)104 and/or 113 are known to exist in human populations. However, thephysiological significance of hGSTP1—1 polymorphism is not fully understood. In this communication, we report that the 1104,A113 ailele ofhGSTP1—1,which is most frequent in human populations, is also mostefficient in the GSH conjugation of carcinogenic anti-diol epoxides ofbenzo[g]chrysene and benzo(c]phenanthrene (anti-BGCDE and antiBCPDE,respectively).Thecatalyticefficiencyof hGSTP1—i(1i04,Aii3)isoform toward anti-BGCDE, 0.36 mM'@s@1, was —1.7-fold higher(P < 0.05) compared with hGSTP1-i(Vi04,V113). Interestingly, the Irequency ofcodon i04-valine alleles is significantly higher in certain cancerscompared with codon i04-isoleucine alleles. Like anti-BGCDE, the catslytic efficiency of hGSTP1—1(1i04,A113) Isoform toward anti-BCPDE washigher by about 1.4- to 2.2-fold (P < 0.05) than those of other hGSTP1—ivarIants. These observations are interesting because we have shown previously (Hu, X. ci aL, Biochem. Biophys Res. Commun., 238: 397—402,1997) that the V104,V113 variant, not the 1104,A1i3 isoform, Is mostefficient in the GSH conjugation of bay.reglon anti-diol epoxide of benzo(a)pyrene (anti-BPDE), which, unlike anti-BGCDE or anti-BCPDE, Is aplanar molecule. In conclusion, our results suggest that hGSTP1—ipolymorphism may be an important factor in differential susceptibility ofhumans to cancers where polycyclic aromatic hydrocarbons are etiological factors and that 1104,A1i3 variant may play a major role in thedetoxification of nonplanar, stericaily hindered fjord-region diol epoxldes(e.g., anti-BGCDE).

Introduction

PAHs3 are widespread environmental pollutants that are abundantin cigarette smoke and other substances and are suspected humancarcinogens (1). It is well known that the diol epoxides of the PAHs,rather than the parent compounds, are the ultimate carcinogens ofthese environmental pollutants (2, 3). Covalent interaction of thePAH-diol epoxide with nucleophilic sites in DNA is an importantevent in PAH-induced carcinogenesis. Although a number of differentmechanisms can detoxify PAH-diol epoxides, the most importantmechanism in their detoxification is the GST-catalyzed conjugation

Received 7/I 5/98; accepted 10/I3/98.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

@ Supported in part by USPHS Grants ES 09140, CA 55589, and CA 76348 (toS. V. S.); ES 07804 (to P. Z.); and CA 27967 (to Y. C. A.).

2 To whom requests for reprints should be addressed, at Cancer Research Laboratory,

The Mercy Hospital, 1400 Locust Street,Pittsburgh,PA 15219.3 The abbreviations used are: PAH, polycyclic aromatic hydrocarbon; BPDE, 7,8-

dihydroxy-9, I0-epoxy-7,8,9, lO-tetrahydrobenzo(a)pyrene; BCPDE, 3,4-dihydroxy-1,2-epoxy-l,2,3,4-tetrahydrobenzo[c]phenanthrene; BGCDE, I l.12-dibydroxy-13,14-epoxyI I, I2, 13, l4-tetrahydrobenzo[g@chrysene; CDE, I.2-dihydroxy-3,4-epoxy- I.2,3,4-tetrahydrochrysene; GSH, glutathione; GST, glutathione S-transferase; TKE buffer, 50 instTns-HCI(pH 7.5) containing2.5 mst KCIand0.5 msi EDTA;HPLC,high-performanceliquid chromatography.

with GSH (4—9).Cytosolic GSTs are a superfamily of detoxificationisoenzymes that have been categorized into five known major classes—Alpha, Mu, N, Theta, and Zeta—based on their physicochemicalproperties (10—13).Isoenzymes of these classes differ significantly intheir substrate specificities (10, 11). For example, the Pi class isoenzyme (hGSTP1—l)is relatively more efficient than other classes ofGSTs in the GSH conjugation of anti-BPDE (8), which is the ultimatecarcinogen of widespread environmental pollutant benzo(a)pyrene

(2, 3).The hGSTPJ gene has been shown to be polymorphic in humans

(14—20).Fouralleic variantsof hGSTP1—l[hGSTP1—1(1l04,A113),hGSTP1—l(Il04,Vl 13), hGSTP1—l(V104,Al 13) and hGSTP1—l(Vl04,Vl 13)], differing in their structures by the amino acid(s) inposition(s) 104 (isoleucine or valine-ATC or GTC in nucleotideposition 313—315;including the initiator methionine some publications refer this residue as 105) and/or 113 (alanine or valine-GCG orGTG in nucleotide position 340—342;including the initiator methionine some publications refer this residue as 114) have now beenidentified in human populations (16—20).These studies have shownthat, although 1104,A113allele is most frequent in human populations,the 1104,V113 is a rare allele (16—20).The frequency of codon 104valine has been shown to be significantly higher in cancer patients ascompared with codon 104 isoleucine (16, 17, 19, 20). However, themolecular basis and physiological implications of these epidemiological observations are not fully known.

We have demonstrated previously (21) that the allelic variants ofhGSTP1—l differ significantly in their kinetic properties. More recently, we have shown (22, 23) that the hGSTP1—lforms with valinein position 104 [(V104,Al 13) and (V104,Vl 13)] are relatively moreeffective toward (+)-anti-BPDE than the isoenzymes with isoleucinein position 104. Subsequently, Sundberg et a!. (24) have confirmedour observations and demonstrated that the hGSTP1—l(Vl04,A 113)variant (referred to as GSTP1—l/V-105 in Ref. 24) is significanflymore active than hGSTP1—l (1l04,Al13) form (referred to asGSTP1—1/1-l05 in Ref. 24) in conjugation reactions with bulky PAH

diol epoxides including anti-BPDE. These observations led us tohypothesize that subjects homozygous for the V104,A113 orV104,V113 alleles may be at a relatively lower risk for PAH-inducedcarcinogenesis than individuals with other hGSTP1—1genotypes.Recent epidemiological studies, however, have shown that the codon104 valine alleles, not the 1104,A113 allele, is strongly associated withvarious cancers including bladder, testicular, and breast cancer (16,17, 19, 20). Although BPDE is the prototypical and best studied diolepoxide, other members of this family such as BGCDE and BCPDEare potent carcinogens and occur in environment at comparable levels(1, 25, 26). Therefore, to understand the role of hGSTP1—1polymorphism in PAH-induced carcinogenesis, it is essential that the speci

5340

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

5PECtFIC@ES OF hG5TPI-I VARIANTS TOWARD PAH.DIOL EPOXIDES

ficities of hGSTP1—1variants toward diol epoxides other than BPDEbe determined.

In this communication, we report specificities of hGSTP1—l vanants toward racemic anti-BGCDE and anti-BCPDE that, unlike antiBPDE,arenonplanarandstericallyhinderedPAH-diolepoxides.Theresults of the present study indicate that the 1104,A1 13 variant ofhGSTP1—1 is most efficient in the GSH conjugation of both antiBGCDE and anti-BCPDE. The present study, taken together withprevious studies from our laboratory (22, 23) suggest that althoughV104,V1 13 variant may play a major role in the detoxification ofplanar bay-region diol epoxides (e.g., anti-BPDE), the I104,A1 13isoform is likely to be important in the GSH conjugation of nonplanarfjord-region diol epoxides, which are biologically more active than thebay-region diol epoxides (25, 26).

Materials and Methods

Materials. GSH and epoxy-activated Sepharose 6B were purchased fromSigma (St. Louis, MO). The diol epoxides were obtained from the NationalCancer Institute Chemical Carcinogen Reference Standard Repository (Midwest Research Institute, Kansas City, MO).

Expression and Purification of hGSTP1-i Isoforms. Expression vectorspBT9a carrying the cDNAs for the four variants of hGSTP1—1were preparedas described by us previously (23). The four variants of hGSTP1—lwerepurified by GSH-affinity chromatography according to the method of Simonsand Vander Jagt (27) with some modifications described by us previously (28).The purified hGSTP1—lisoforms were dialyzed against 50 msi Tris-HC1(pH7.5) containing 2.5 mM Kcl and 0.5 nmi EDTA (TKE buffer), and stored at—20°Cuntil used. The activity of each variant toward model substratel-chloro-2,4-dinitrobenzene was determined by the method ofHabig et a!. (29)immediately before activity measurements toward diol epoxides.

Determination of GST ActIvity toward Racemic anti-BGCDE and antiBCPDE. The OST activity toward anti-BGCDE and anti-BCPDE was determined by the method ofJernströmet aL (30) with some modifications. Briefly,the reaction mixture in a fmal volume of 0.1 ml contained TKE buffer, 2 msiGSH,the desiredconcentrationof the diolepoxide,and 30 @.tg/m1hGSTP1—1isoform protein. The reaction was initiated by adding the diol epoxide substrate; the reaction mixture was then incubated for 30 s at 37°C.The GSTactivity was measured as a function of varying protein concentration andincubationtimeto optimizethe assayconditions.Thereactionwasterminatedby rapid mixingwith 0.1 ml cold acetone.Unreacteddiol epoxidesubstratewas removed by extraction with ethyl acetate. The GSH-diol epoxide conjugates in the aqueous phase were quantified by reverse-phase HPLC using aWaters Nova-Pak C18(3.9 mm X 150 mm) column. The GSH conjugates of(+)- and (—)-anti-BCPDEwere elutedisocratically(18.6% acetonitrilein

0.1%trifluoroaceticacid).Undertheseconditions,theGSHconjugatesof(+)and (—)-anti-BCPDEwere eluted at retention times of about 6.6 and 5.8 mm,respectively.The separationof GSHconjugatesof(+)- and(—)-anti-BGCDEwas also achieved by isocratic elution (25.4% acetonitrile in 0.1% trifluoroacetic acid). The retention times of the GSH conjugates of (+)- and (—)-antiBGCDB were 6.2 and 4.8 mm, respectively. The GST activity was measuredas a function of varying diol epoxide concentration at a fixed saturatingconcentration of GSH to determine the kinetic parameters. For anti-BGCDE,the K@and V@,,,,,valuesweredeterminedby nonlinearregressionanalysisoftheexperimentaldatapointsusingtheMichaelis-Menten(hyperbolic)equationas the model. For calculation of the catalytic efficiency, the Michaelis-Mentenequation

[S]vV

Km + [5]

Michaelis-Menten equation as the model directly yields the catalytic efficiencyparameter together with its associated asymptotic SE. The latter constitutes themajor advantage of the procedure described above over a simple division ofkcat by Km. Although both approaches result in the same value for catalytic

efficiency, there is no simple way to arrive at the SE of a ratio of two numbers.The catalytic efficiencies of hGSTP1—lvariants toward anti-BCPDE werecalculated from activity measurements at low substrate concentrations asdescribed by Kolm et a!. (31) because saturation curves could not be obtainedfor this diol epoxide substrate. Poor solubility of anti-BCPDE at higherconcentrations was a limiting factor in activity measurements at concentrationsof >320 @.&M.

Results and Discussion

Fig. 1 depicts the reverse-phase HPLC analysis of the water-solubleproducts resulting from the reaction of 2 nmi GSH and 160 ,.LMracemic anti-BGCDE in the presence of 30 p.g/ml hGSTP1—l variantprotein. The nonenzymatic GSH conjugation of anti-BGCDE wasnegligible (data not shown in Fig. 1), whereas this reaction wasincreased significantly in the presence of hGSTP1—l protein. All fourof the GSTP1—lvariants were virtually inactive toward the (+)-enantiomer of anti-BGCDE. When GST activity was measured as afunction of varying anti-BGCDE concentrations (10—160 @LM)at afixed saturating concentration of GSH (2 mtvi), all four variants ofhGSTP1—1 followed Michaelis-Menten kinetics (plots not shown).

The kinetic constants for hGSTP1—lvariants toward anti-BGCDE aresummarized in Table 1. The catalytic efficiency of hGSTP1—

1(1l04,A1l3) variant was approximately 1.7-fold higher than that ofthe V104,Vl 13 isoform (P < 0.05). This was mainly due to a 1.6-foldhigher V@ for the 1l04,A1l3 variant as compared with V104,Vl 13isoform. The catalytic efficiencies of the other two variants of hGSTP1—1toward anti-BGCDE were intermediate but statistically notsignificantly different from either the 1104,Al 13 or the V104,V113isoform.

As shown in Fig. 2, unlike anti-BGCDE, the concentrations ofanti-BCPDE approaching saturation of the enzyme could not beattained because of poor solubility of this diol epoxide at higher

concentrations. Therefore, a reparameterized Michaelis-Menten rateequation

kcatv_._[E0][S]

EI')

1-I0

1l04,Vl13

V104,Al

was rearranged to yield:

2 4Retention Time ( mm )

[5]vK .Q.

Km + [5]

Fig. 1. Reverse-phase HPLC analysis of the water-soluble products resulting from thewhere Q = V@/Km. Q @Sequivalent to the catalytic efficiency (after appro- reactionof 2 instGSHand 160@ racemicanti-BGCDEin thepresenceof 30 pg/miofpriate change of units for V,,...@).Nonlinear regression using the rearranged hGSTP1-1variants.

5341

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

Table I Kinetic constants for hGSTPI-1 variants in the GSH conjugation ofanti-BGCDETheactivity of each enzyme was measured in triplicate or quadruplicate at five different substrate concentrations. Kinetic parameters were estimated by nonlinear(hyperbolic)regression

and are given ±asymptotic error of the mean. The catalytic efficiencies of the I104,Al 13 and the V104,Vl13 variants of the enzyme are significantly different (P <0.05);thedifferences in catalytic efficiency between the remaining enzymes do not reach statistical significance.Catalytic

efficiencyhGSTPI-l variant Km (lAM) V,,.@(nmol/min/mg)@ (s') (mM@.I

l04,A I I3 203 ±61 95 ±I8 0.074 ±0.014 0.36 ±0.04Il04,Vl13 231 ±90 85 ±23 0.066 ±0.018 0.28 ±0.04V104,Al13 197 ±75 78 ±20 0.060 ±0.015 0.31 ±0.04V104,Vl13 223 ±52 58 ±20 0.045 ±0.015 0.21 ±0.03

0

0.

atE

E

0E

4,4.,

4,I-

04.,

(.1(U4,

200

I 60

120

80

40

0

SPECIFICITIES OF hGSTP1-l VARIANTS TOWARD PAH-DIOL EPOXIDES

Furthermore, we have demonstrated that the catalytic efficiency of theV104,V1 13 variant is significantly higher compared with I104,A113

isoform in the GSH conjugation of anti-CDE (23), another planar diolepoxide of the bay-region class. These observations led us to postulatethat the hGSTP1—lpolymorphism with disparate enzyme activitiestoward carcinogenic PAH-diol epoxides plays a role in determiningoverall susceptibility toward certain carcinogens. The results of thepresent study reveal that hGSTP1—l(Il04,Al13) variant is most efficient in catalyzing the GSH conjugation of fjord-region diol epoxides.Taken together, our studies suggest that although V104,V113 variantmay play an important role in the detoxification of planar bay-regiondiol epoxides (e.g., anti-BPDE or anti-CDE), the catalytic efficiencyranking is reversed for the nonplanar fjord-region diol epoxides (e.g.,anti-BGCDE or anti-BCPDE toward which the 1104,All3 isoform ismost active. Thus, the molecular shape of the diol epoxide substrateseems to be an important determinant for the variable activity ofhGSTP1—l isoforms toward these environmentally relevant carcinogens.

The results of present studies suggest that the putative role ofhGSTP1—1polymorphism in humans with regards to the susceptibilityof individuals to PAH-induced carcinogenesis may be more complexthan currently realized and should be carefully re-evaluated. It hasbeen shown that hGSTP1—1polymorphs significantly differ in theircatalytic efficiencies in the detoxification of (+)-anti-BPDE (22—24),which suggests differential capabilities of individuals for detoxification of the “ultimate―carcinogenic metabolite of benzo(a)pyrene.However, the frequency of V104,A113and/or V104,V113 alleles hasbeen shown to be higher in cancer patients compared with 1104 ,A113

allele, which has lesser activity toward (+)-anti-BPDE (17, 19, 20,

0

0 70 140 210 280 350

0.3

0.2

0.1

U)

E

UC4,U

LU

U

(U4.,

(UC)

1104 1104 V104 V104A113 V113 A113 V113

Fig. 3. Catalytic efficiencies ofhGSTP1-l variants in the GSH conjugation of racemicanti-BCPDE. The data are means ±SE of six measurements. The catalytic efficiencieswere obtained directly from initial rate measurements at low substrate concentrations (see“Materialsand Methods―).By one-way ANOVA, differences exist between the catalyticefficiencies of the four enzyme variants (P 0.0003). Several multiple comparisonprocedures (Bonferroni, Duncan, Newman-Keuls, and Tukey-Kramer) indicate that. atP < 0.05. the catalytic efficiency of 1l04,Al 13 is different from that of the remainingthree enzymes but that the latter three enzymes are not different from each other.

5342

ant/-BCPDE (jiM)Fig. 2. The GSH conjugation of anti-BCPDE in the presence of hGSTP1—lvariants as

a function of varying diol epoxide concentration.

described by Kolm et al. (31) was used to determine the catalyticefficiencies of hGSTP1—l variants toward anti-BCPDE; the resultsare shown in Fig. 3. Like anti-BGCDE, the 1l04,Al13 variant ofhGSTP1—l was most efficient in the GSH conjugation of anti

BCPDE. The catalytic efficiency of the hGSTP1—l(1104,Al13) vanant toward anti-BCPDE was higher by approximately 2.2-, 1.4-, and1.8-fold as compared with hGSTP1—l(1l04,Vl 13), hGSTP1—l(Vl04,Al 13), and hGSTP1—l(Vl04,Vl 13), respectively (P < 0.05;

Fig. 3), whereas the catalytic efficiencies of the latter three proteinswere not significantly different from each other.

The existence of the four allelic forms of hGSTP1—lin humanpopulations has been recently confirmed by analyses of the genomicDNA and/or mRNA (16—20).Although hGSTPJ—1(1104,A113)alleleis most frequent in human populations, the 1104, V113 seems to be arare allele (16—20). In addition, the frequency of the hGSTPJ—1alleles has been shown to differ between healthy controls and cancerpatients (17, 19, 20). For example, the frequency of V104,A113 and/or

V104, V113 alleles is significantly higher in certain cancers as compared with the 1104,A113 allele (16, 17, 19, 20). The results of thepresent study indicate that the 1l04,Al13 variant of hGSTP1—l isapproximately 1.7- to 1.8-fold more efficient than the Vl04,V1l3isoform in the GSH conjugation of fjord-region anti-diol epoxides ofbenzo[g]chrysene and benzo[c]phenanthrene, which are nonplanardiol epoxides. These results are in contrast with our previous findingswith the planar bay-region diol epoxide (+)-anti-BPDE (22, 23), inwhich we have demonstrated that the Vl04,Vl 13 variant of hGSTP1—lis significantly more efficient than 1l04,A1l3 isoform in theGSH conjugation of this diol epoxide and that the transition of All3Vin the presence of valine 104 causes a pronounced increase in thecatalytic efficiency of the enzyme toward (+)-anti-BPDE (22, 23).

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

5PEClFlC@ES OF hG5TPI-1 VARIANTS TOWARD PAH.DIOL EPOXIDES

22—24).On the basis of these findings, it has been suggested thathGSTP1—1polymorphism may not be a determinant of the susceptibility of individuals to PAH-induced carcinogenesis (24). The resultsof the present study, however, clearly demonstrate that the catalyticefficiencies of hGSTP1—l variants vary remarkably with the nature ofthe PAH-diol epoxide and that these polymorphs exhibit differentialactivities toward the planar and nonplanar PAH-diol epoxides. Thus,it may be possible that although the individuals homozygous forhGSTPJ—1(V104,V113) allele may have a better protection whenbay-region diol epoxides such as (+)-anti-BPDE are the initiatingagents, the homozygotes for the hGSTPI—I(1104,A113) allele mayhave advantage in the detoxification of the nonpianar PAH-diol epoxides such as anti-BGCDE and anti-BCPDE, which are also abundant among the environmental PAHs (1). In addition, the fjord-regiondiol epoxides are significantly more mutagenic and carcinogenic inanimal models than the diol epoxides of the bay-region class (25, 26).Because PAIl mixtures in cigarette smoke, industrial emissions, andother environmental sources are likely to be different, our resultsindicate that, depending on the chemical composition of the PAHmixture, individuals with a given hGSTPJ—1genotype may be at adifferential risk for carcinogenesis.

References1. IARC. Monographs on the Evaluation of the Carcinogenic Risk of Chemicals to

Humans: Polynnclear Aromatic Compounds. Part 1. Chemical, Environmental, andExperimentalData.IARCScientificPubI.,Vol.32. Lyon,France:IARC,1983.

2. Sims, P., andcirover, P. L Epoxides in polycyclic aromatic hydrocarbon: metabolismandcarcinogenesis.Adv.CancerRes.,20: 165—274,1974.

3. Thakker@D. R., Yagi, H., Levin, w., Wood, A. W., Conney. A. H., and Jerina, D. M.Polycycicaromatichydrocarbons:metabolicactivationto ultimatecarcinogens.In:M. W. Andes-s (ed.), Bioactivation of Foreign Compounds, pp. 177-242. New York:AcademicPress,1985.

4. Yang, S. K., and Gelboin, H. V. Nonenzymatic reduction of benzo(a)pyrene diolepoxidesto trihydroxypentahydmbenzo(a)pyrenesby reducedmcotinamideadeninedinucleotide phosphate. Cancer Res., 36: 4185—4189, 1976.

5. Dock, L, Wacin, F., Martinez, M., Grover, P. L, and Jemstrom, B. Studies on thefurtheractivationofbenzo[a]pyrenediolepoxidesbyratlivermicrosomesandnuclei.Chcm.-Biol. Jnteract., 58: 301—318,1986.

6. Thakker, D. R., Yagi, H., Levin, W.. Lu, A. Y. H., Conney, A. H., and Jerina, D. M.Stereospecificity of microsomal and purified epoxide hydrase from rat liver. J. Biol.Chain.,252:6328—6334,1977.

7. Robertson, I. G. C., Jensson, H., Mannervilc, B., and Jernstrom, B. Glutathionetransferases in rat lung: the presence of transferase 7—7,highly efficient in theconjugation of glutathione with the carcinogenic (+)-7@,8a-dihydmxy-9a,l0a-oxy7,8,9,lO.tetrahydrobenzo(a)pyrene. Carcinogenesis (Lond.), 7: 295—299,1986.

8. Robertson,I. G.C.,Guthenberg,C.,Mannervik.B.,andJernström,B.Differencesinstereoselectivity and catalytic efficiency of three human glutathione transferases inthe conjugation of glutathione with 7@3,8a-dihydroxy-9a,l0a-oxy-7,8,9,l0-tetrahydrobenzo(a)pyrene. Cancer Res., 46: 2220—2224,1986.

9. Sundberg, K., Widersten, M., Seidel, A., Mannervik, B., and Jernström,B. Glutathione conjugation of bay- and fjord-region diol epoxides of polycyclic aromatichydrocarbons by glutathione transferases Ml—land P1—I.Chem. Rca. Toxicol.. 10:1221—1227,1997.

10. Mannervik, B. AIm, P., Guthenberg, C., iensson, H., Tahir, M. K., Warholm, M., andJörnvall,H. Identification of three classes of cytosolic glutathione transferase coinmonto severalmanunalianspecies:correlationbetweenstructeraldataandenzymaticproperties.Proc.NatI.Aced.Sci.USA,82: 7202—7206,1985.

1!. Hayes, 1. D., and Pulford, D. J. The glutathione S-transferase supergene family:regulationofGST'@andthecontributionof theisoenzymestocancerchemoprotectionanddrugresistance.Crit.Rev.Biochem.Mol.Biol.,30: 445-600, 1995.

12. Meyer, D. J., Coles, B.. Pemble, S. E., Gilmore, K. S., Fraser, G. M., and Ketterer,B. Theta, a new class of glutathione transferases purified from rat and man. Biochem.1., 274: 409—414,1991.

13. Board, P. G., Baker, R. T., Chelvanayagam, G., and Jeriniin, L. S. Zeta, a novel classof glutathione transferases in a range of species from plants to humans. Biochem. J.,328: 929—935,1997.

14. Board, P. G., Webb, G. C., and Coggan, M. Isolation of a cDNA clone andlocalization of the human glutathione S-transferase 3 genes to chromosome bandsllql3 and 12q13—14.Ann. Hum. Genet., 53: 205—213,1989.

15. Ahmad, H., Wilson, D. E., Fritz, R. R., Singh, S. V., Medh, R. D., Nagle, G. T.,Awasthi, Y. C., and Kurosky, A. Primary and secondasy structural analyses ofglutathione S-transferase it from human placenta. Arch. Biochem. Biophys., 278:398—408,1990.

16. Ali-Osman, F., Akande,0., Antoun,0., Mao, J-X., and Buolamwini,J. Molecularcloning, characterization, and expression in Escherichia coli of full-length cDNAs ofthree human glutathione S-transferase P1 gene variants: evidence for differentialcatalytic activity of the encoded proteins. J. Biol. Chem., 272: 10004-10012, 1997.

17. Ryberg, D., Skaug. V., Hewer, A., Phillips, D. H., Harries, L W., Wolf, R. C., Øgreid,D., Ulvik, A., Vu, P., and Haugen, A. Genotypes of glutathione transferase Ml andP1 andtheirsignificancefor lung DNA adductlevels andcancerrisk.Carcinogenesis(Lond.), 18: 1285—1289,1997.

18. Watson, M. A., Stewart, R. K., Smith, G. B. 3., Massey, T. E., and Bell, D. A. Humanglutathione S-transferase P1 polymorphisms: relationship to lung tissue enzyme activity and population frequency distribution. Carcinogenesis (Land.), 19: 275—280,1998.

19. Harries,L. W., Stubbins,M. J., Forman,D., Howard,G. C. W., and Wolf, C. R.Identificationof genetic polymorphismsat the glutathioneS-transferaseP1 locus andassociation with susceptibility to bladder, testicular and prostate cancer. Carcinogenesis (Lond.), 18: 641—644,1997.

20. Helzlsouer, K. J., Selmin, 0., Huang, H-Y., Strickland, P. T., Hoffman, S., Alberg,A. J., Watson, M., Comstock, G. W., and Bell, D. Association between glutathione5.@jar@sfe@a@eMl, P1, and TI genetic polymosphisms and development of breastcancer. J. NatI. Cancer Inst., 90: 512—518,1998.

21. Zimniak, P., Nanduri, B., Pikula, S., Bandorowicz-Pikula, J., Singhal, S. S.,Srivastava, S. K., Awasthi, S., and Awasthi, Y. C. Naturally occurring humanglutathione S-transferase GSTPI—lisoforms with isoleucine and valine in position104 differ in enzymic properties. Eur. J. Biochem., 224: 893—899,1994.

22. Hu, X., O'Donnell, R., Srivastava, S. K., Xia, H., Zimniak, P., Nanduri, B., Bleicher,R. J., Awasthi, S., Awasthi, Y. C., ii, X., and Singh, S. V. Active site architecture ofpolymorphic forms of human glutathione S-transferase P1—Iaccounts for theirenantioselectivity and disparate activity in the glutathione conjugation of 713,8a-dihydroxy-9a,l0a-oxy-7,8,9,10-tetrahydrobenzo(a)pyrene. Biochem. Biophys. Res.Commun., 235: 424—428.1997.

23. Hu, X., Xia, H., Srivastava, S. K., Herzog, C., Awasthi, Y. C., Ji, X., Zimniak, P., andSingh, S. V. Activity of four allelic forms of glutathione S-transferase hGSTPI—lfordiol epoxides of polycyclic aromatic hydrocarbons. Biochem. Biophys. Rca. Commun., 238: 397—402,1997.

24. Sundberg, K., Johansson, A-S., Stenberg, G., Widersten, M., Seidel, A., Mannervik,B., and JemsirOm, B. Differences in the catalytic efficiencies of allelic variants ofglutathione transferase P1—Itowards carcinogenic diol epoxides of polycyclic aromatic hydrocarbons. Carcinogenesis (Land.), 19: 433—436,1998.

25. Glatt, H., Pieé,A., Pauly, K., Steinbrecher, 1., Schrode, R., Oesch, F., and Seidel, A.Fjord- and bay-region diol-epoxides investigated for stability, SOS induction inEscherichia coli, and mutagenicity in Salmonella typhimuriwn and mammalian cells.Cancer Res., 51: 1659—1667,1991.

26. Levin, W., clang, R. L., Wood, A. W., Thakker, D. R., Yagi, H., Jerina, D. M, andConney. A. H. Tumorigenicityof optical isomers of the diastereomericbay-region3,4-diol-l,2-epoxides of benzo(c)phenanthrene in murine tumor models. Cancer Rca.,46: 2257—2261, 1986.

27. Simons, P. C., and Vander Jagt, D. L. Purification of glutathione S-transferases fromhuman liver by glutathione-affinity chromatography. Anal. Biochem., 82: 334—341,1977.

28. Singh, S. V., Leal, T., Ansari, 0. A. S., and Awasthi, Y. C. Purification andcharacterization of glutathione S-transferases of human kidney. Biochem. 3., 246:179—186,1987.

29. Habig, W. H., Pabst, M. J., and Jakoby, W. B. Glutathione S-transferases. the firstenzymatic step in mercapturic acid formation. J. Biol. Chem., 249: 7130—7139,1974.

30. Jernstrom, B., Funk, M., Frank, H., Mannervik, B., and Seidel, A. GlutathioneS-transferase Al—l-catalyzedconjugation of bay and fjord region diol epoxides ofpolycyclic aromatic hydrocarbons with glutathione. Carcinogenesis (Lond.), 17:1491—1498,1996.

31. Kolm, R. H., Danielson, U. H., Zhang, Y., Talalay, P., and Mannervik, B. lsothiocyanates as substrates for human glutathione transferases: structure-activity studies.Biochem. J., 311: 453—459,1995.

5343

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

[CANCERRESEARCH58, 5344-5347. Decembert. 9981

Advances in Brief

Absent or Low Expression of the@ Chain in T Cells at the Tumor Site Correlateswith Poor Survival in Patients with Oral Carcinoma'

Torsten E. Reichert, Roger Day, Eva M. Wagner, and Theresa L. Whiteside2

Department of Maxillofacial Surgery. University of Mainz. 55131 Mainz. Germany IT. E. R., E. M. W.J; and Departments of BiostatisticsfR. DI, Pathology IT. L WI, andOtolarvngologv fT L WI. Universirt of Pittsburgh, and University of Pittsburgh Cancer Institute fR. D., T. L WI, Pittsburgh. Pennsylvania 15213

Abstract

Immunohistology for expression of the CD3@and CD3e chains In TILwas performed in 138 paraffin-embedded primary oral squamous cellcarcinoma tissues and 10 nontumor, inflammatory lesions. Semiquantitative analysis of the staining intensity for @chainexpression and number of

Cchainexpression-positivecellsdistinguishedtumorswithabsentorlowCexpression(42of132)fromthosewithnormal@ expression(90of132).Cchainexpressionwasinverselycorrelatedwiththetumorstage.Survivalwas significantly lower in patients whose ilL had absent or low@ expression, controlling for stage (P 0.003) and lymph node status (P 0.0005).The prognostic value of@ chain was restricted to patients with stage ifi orIv tumors(P OA)03).Thedataindicatethatabsentor decreased@expression in TIL combined with tumor stage or nodal status defines agroup of patients with oral squamous cell carcinoma who have an extremely poor prognosis.

Introduction

Alterations in the expression of signaling molecules in T cells ofpatients with cancer have been reported by many investigators (1—5).These alterations, including a decrease in expression or absence of thesignal-transducing@ chain, are thought to be responsible for functionalimpairments of immune cells at the tumor site and in the peripheralcirculation of patients with malignancies (reviewed in Ref. 6). Decreased@ chain expression has been observed in the circulating T cellsof patients with melanoma, renal cell carcinoma, ovarian carcinoma,cervical carcinoma, colon carcinoma, and other cancers (1—9).Absentor low expression of the@ chain is especially pronounced in TIL@(1,10). This observation and the evidence for a significant decrease orabsence of the@ chain in T cells coincubated with human tumor cellsin vitro (1 1) suggest that this alteration may be tumor induced. Themechanism(s) responsible for a decrease in@ expression are unknown,and various explanations have been advanced, including the possibility that it is an artifact induced by monocytes during isolation ofperipheral blood mononuclear cells ( 12). More recently, decreasedexpression of the@ chain has been linked to the process of apoptosisinduced in T cells by tumor-derived4 or other apoptotic signals (1 1).Overall, however, the biological significance of@ chain abnormalitieshas remained unclear.

To address the presence and biological significance of the@ chainalterations in T cells present in the tumor microenvironment, a retro

Received 8/17/98: accepted 10/19/98.The costs of publication of this article were defrayed in part by the payment of page

charges. This article must therefore be hereby marked advertisement in accordance with18 U.S.C. Section 1734 solely to indicate this fact.

I Supported in part by NIH Grant POI-DE 12321 (to T. L. W.).

2 To whom requests for reprints should be addressed, at University of Pittsburgh

Cancer Institute, Wl04 I BST. 21 1 Lothrop Street. Pittsburgh, PA I52 13-2582. Phone:(412) 624-0080: Fax: (412) 624-0264; E-mail: Whitesidetl@msx.upmc.edu.

3 The abbreviations used are: IlL, tumor-infiltrating lymphocytes; Ab, antibody; mAb,

monoclonal Ab.4 Gastman, B. R., Johnson, D. E., Whiteside, T. L., and Rabinowich, H. Activation of

caspase-3 and cleavage of Bcl.2 in tumor-induced apoptosis of T lymphocytes. submittedfor publication.

spective immunohistochemical study was performed on tumor biopsies obtained from 138 patients with oral carcinoma over a period ofseveral years. Semiquantitative evaluation of@ chain expression inCD3@ T cells at the tumor site established correlations betweenstaining intensity for the@ chain in TIL and prognosis or patientsurvival. The results indicated that patients with stage III or IV tumorsand absent or low expression of the@ chain in ilL have extremelypoor survival compared to patients with either normal@ chain expression or early-stage oral SCC.

Materials and Methods

Patients and Specimens. Tissue samples were collected retrospectivelyfrom 138patients with 5CC ofthe oral cavity. Of these, clinical follow-up datawere available for 132. The sites of primary tumors were as follows: floor ofthe mouth, 56 tumors; maxilla including palate, 15 tumors; retromolar tnigone,7 tumors; gingiva of the mandible, 24 tumors; buccal mucosa, 5 tumors; and

tongue, 25 tumors. The patients underwent potentially curative tumor resectionbetween 1980 and 1993. No preoperative radio- or chemotherapy had beenperformed. Among the patients, 101 were male and 31 were female.

The median age was 58 years (range, 36—89years). Control specimens wereobtained from lesions in 10 nontumor patients (3 patients with chronic inflammation of the maxillary sinus; 3 patients with periapical cyst; 3 patients withtonsilitis; and I patient with chronic sialadenitis of the parotid gland). Of 138patients, whose specimens were studied, only 11 had no history of smokingand/or alcohol abuse. All others were in a high-risk group for oral carcinoma.

The biopsy specimens were fixed in 4% formalin, embedded in paraffin,sectioned, and stained with H&E. The ThM staging categories were determined according to the criteria proposed by the International Union AgainstCancer.Clinicaldefinitionsof the tumorstage wereconfirmedby histopathology and were as follows: stage I, 30 patients; stage II, 33 patients; stage HI, 7patients; and stage IV, 62 patients. Histopathology indicated that 80 tumorswere N0, 11 were N1, and 41 were N2. Tumor size was defined as the largestdiameter of the tumor.

Abs. The unlabeled mAb to CD3C (6B 10.2) was purchased from SantaCruz Biotechnology (Heidelberg, Germany). The CD3€mAb and the vimentin

mAb were purchased from DAKO (Glostrup, Denmark). Secondary Abs,which were biotinylated and included in the LSAB2 kit from DAKO, wereused to perform the immunohistological avidin-biotin technique. In all experiments, isotype Abs purchased from DAKO were used as controls.

Immunohistochemistry. Paraffin sections (5 @tmthick) were conseculively cut from each specimen and mounted on electrostatically prechargedslides (Superfrost Plus; Menzel-Glaser, Frankfurt, Germany). The sectionswere deparaffinized and rehydrated. After endogenous peroxidase quenching(0.3% H2O2in PBS for 30 mm), antigens were retrieved by boiling the sectionsin 1 mMEDTA-NaOH solution (pH 8.0) in a microwave oven with the powerset at 750 W for three cycles (5 mm each). The same number of slides (n = 20)was always placed in a rectangular plastic staining jar for microwave treatmenu. After the completion of the third cycle, slides were allowed to cool atroom temperature for 30 mm.

Immunohistology was performed by use of an avidin-biotin technique.Optimal working dilutions of the primary Abs were as follows: CD3@, 1:800;CD3€,1:200;andvimentin, 1:200.The workingdilutionswere determinedinpreliminary titration experiments performed with human tonsils. Staining was

developed with peroxidase and amino-9-ethylcarbazole. The mAb againstvimentin was used as an internal control to ascertain the quality of antigen

5344

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

Cchainexpression°TumorstageTotalspecimensIIIIIIIVAbsent/low7512942

(32)―Normal232763490(68)Total3032763132(100)

CD3CCHAIN EXPRESSIONIN ilL

and 72 were alive as of July 1998. Median follow-up time was 41months. Estimated median survival was 78 months.

CChainExpressioninTumororControlTissues.Usingexpression of vimentin as a measure of antigen preservation in formalinfixed, paraffin-embedded SCC tissues, it was possible to confirm thatall tissues examined were uniformly positive (data not shown). Thiswas an essential control step because many antigens undergo dena

turation during fixation, embedding, or microwave treatment andbecome nonreactive with specific mAbs (13). Staining of nontumortissues containing inflammatory infiltrates was another essential control that we included. This staining showed that T lymphocytes in all10 inflammatory lesions had strong expression of the@ chain (Fig.

q 1A), which was comparable in intensity to CD3€ staining in the serial

i_ section of the same tissue (data not shown). In contrast, a large

proportion (42 of 132) of oral SCCs showed absent or low expressionof the@ chain in ilL, as compared to staining for CD3€in the same

specimen (Fig. 1, B and C) or to@ staining in control sections (Fig.

IA). The distinction between normal@ expression (Fig. 1E, group 1)and “low―@ expression (Fig. 1C, group 2) was based on two separatecriteria: (a) visual comparison of experimental and control sectionsand (b) the number of@ T cells in serial sections stained for CD3C orCD3€.In 9 of the tumors, no@ expression was detected, and it wasjudged to be low in 33 SCC tissues in comparison to control specimens. Because of the small number of samples (n 9) in the “absent―category (designated 0), it was combined with the low category(designated 1) in subsequent analyses.

Association of@ Chain Expression with the Tumor Stage. Therelationship between advanced tumor stage and low@ expression wasstrong (odds ratio = 3.3, P = 0.003). Absent/low expression of the@chain was observed in only 19% of the stage I—I!tumors, but it wasobserved in 43% of stage IH—IVtumors (Table 1).

Association between@ Chain Expression and Tumor Size,Nodal Involvement, or Distant Metastases. Primary tumor size (T),the presence and type of tumor spread to the local lymph nodes (N),and the presence or absence of distant metastases (M) are all impor(ant staging and prognostic factors in oral SCC. Our data indicatedthat low@ expression was associated with lymph node involvement(odds ratio = 3.4; P = 0.002), worse primary tumor status (oddsratio = 3.8; P = 0.0009), and the presence of distant metastases (oddsratio infinite; P 0.03).

To clarify what role@ expression might play in the process of tumorprogression, the prediction of nodal involvement by@ expression,controlling for actual tumor dimension, was investigated. It was foundthat low/no@ expression increases the risk of nodal involvement by anodds ratio of 1.68 (P = 0.01).

Association between@ Chain Expression and Patient Survival.Absent or low@ expression was associated with poor survival ofpatients with oral SCC, even after data were controlled for tumor stage(Table 2; hazard ratio = 1.6, P = 0.0003). The interaction term wassignificant as well (Table 2 and Fig. 2A), indicating that the prognostic

value of@ expression is greater in more advanced disease. The datapresented in Fig. 2A strongly suggest that the concurrent presence of

Table 1 Association between@ chain expression in TILcand tumor stage in humanoral SCC

V@

Fig.@ . lniniuiiopeni@ida@estaining with anti-CD3@ mAb of a paraffin section of a.tir(inic iiiIliniiiiitir\ lesion in the maxillary sinus (A, positive control). Note strongly

@iaincdT I@iiipliuc@ic@in the connective tissue underlying the epithelium of the maxillary@uius.lii It @iikl( @.immunoperoxidase staining of serial paraffin sections of an oral SCC

with ifiti(1)@f ..\h (RI and anti-CD3CmAb (C). Note the near absenceofstaining for@iii C. In I) and L. @eri.iI sections ofanotheroral SCC with normal expression ofboth a(D)

and (L) in Til .. Original magnification, X400.

Preserv@It@@and uniformity of staining in the formalin-fixed tissue sections( 13.

Evaluation of Stained Sections. All immunostainedsections were examned by light microscopy by one of us (T. E. R.). Only those sections that

@l1OWCda uiiitorni and intense vimentin staining in cells of the mesenchymal

()rivtn were iiiclu(k-d in the study. To evaluate CD3Cexpression in i'lL, at least10 high-po@@er fields were randomly selected for microscopic examinations at

the border 1)1the tumor and within the peritumoral tissue. The staining intensity

‘@as as'@e@sed with coded slides. Only two groups were distinguished in respect

to staining inteimit\ : group 1, normal CD3@ expression; and group 2, reduced

: orabsentCl)3@expression.StainingintensityinTILwasjudgedrelativetol@mphoc@te-@8djaceflt to normal squamous epithelium within the same speciflien and reLitive to lymphocytes in the nontumor tissues used as controls,‘,.hich @vere stained and examined in parallel. In addition, the number of

positiveI@ stained cells was determined, to provide a semiquantitative estimation of (‘D3i@ in TIL. Cells were counted in five high-power (X400)fields@ ith an ocular grid. The number of CD3C+ cells was compared to thenunther of (I)3f-@ cells. In cases in which CD3C+ cells were <50% ofCD3e + celk. the stainingwas judged to be reduced(group 2).

Statistical Analysis. Survival data were analyzed using proportional hazards reore'@@ion and Kaplan-Meier survival estimation. Nodal status as an

OUtCOHiC wa@ analyzed using logistic regression. Categorical associations were

e\ abated using Fisher exact tests. Where appropriate, sparse categories were

combined. For this reason, the four stages were collapsed into stage I—ilandstage lll—lV. Because only 7 patients had stage ifi disease, it cannot bedetermined whetlici stage ifi resembles stage IV or stage I—il.

Results

Patient Follow-up Data. Of the group of 132 patients with oralscc whose tumor specimens were included in the study, 60 had died

a Expression of the@ chain in TlLs was defined microscopically, as described in

“Materialsand Methods.―b Numbers in parentheses represent percentages.

5345

.@ 4

:4d@':@@

, .@@

4,',, &@. @t..@ @.

4 %•‘I,' ;,*t@@ ‘.@ P@

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from

1996;56:3244-3249. Cancer Res   Klaus Becker, Jörg Dosch, Cornelia M. Gregel, et al.   Tumor Initiation in Two-Stage Skin CarcinogenesisMethyltransferase (MGMT) in Transgenic Mice Protects against

-Methylguanine-DNA6OTargeted Expression of Human

  Updated version

  http://cancerres.aacrjournals.org/content/56/14/3244

Access the most recent version of this article at:

   

   

   

  E-mail alerts related to this article or journal.Sign up to receive free email-alerts

  Subscriptions

Reprints and

  .pubs@aacr.orgDepartment at

To order reprints of this article or to subscribe to the journal, contact the AACR Publications

  Permissions

  Rightslink site. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC)

.http://cancerres.aacrjournals.org/content/56/14/3244To request permission to re-use all or part of this article, use this link

Research. on October 1, 2020. © 1996 American Association for Cancercancerres.aacrjournals.org Downloaded from