DIFFERENTIAL EXPRESSION OF MULTIDRUG RESISTANCE GENE PRODUCT,P-GLYCOPROTEIN, IN NORMAL, DYSPLASTIC AND MALIGNANT ORALMUCOSA IN INDIAVibhor JAIN,1 Satya N. DAS,1 Kalpana LUTHRA,2 Nootan K. SHUKLA3 and Ranju RALHAN2*1Department of Biotechnology, All India Institute of Medical Sciences, New Delhi, India2Department of Biochemistry, All India Institute of Medical Sciences, New Delhi, India3Institute Rotary Cancer Hospital, All India Institute of Medical Sciences, New Delhi, India

Multidrug resistance (MDR) in human cancer is oftenassociated with over-expression of the mdr-1 gene, whichencodes a 170-kDa transmembrane protein, termedP-glycoprotein (P-gp).We evaluated the immunoreactivity ofP-gp in oral tissues at different stages of tumorigenesis in theIndian population by flow cytometry, using the MRK-16monoclonal antibody, which recognizes an external epitopeof P-gp. The expression of P-gp was studied in human oralnormal tissues (12 cases), dysplastic lesions (13 cases), pri-mary untreated squamous-cell carcinomas (12 cases) andrecurrent tumors (18 cases). Quantitative flow-cytometricanalysis of P-gp expression showed a significant increase inP-gp levels in untreated primary oral tumors (pF 0.01) andin dysplastic lesions (pF 0.05) as compared with normal oraltissues. A marked significant increase in P-gp expression wasobserved in recurrent oral carcinomas as compared withnormal oral tissues (p F 0.001) and dysplastic lesions(pF 0.01). Among recurrent tumors, a significant increase inthe level of P-gp was observed in T4-stage tumors as com-pared with T3-stage tumors (pF 0.01). We conclude thatP-gp is differentially expressed during oral tumorigenesis, andmay be an indicator of the biological behavior of oralmalignan-cies. Int. J. Cancer 74:128–133.r 1997 Wiley-Liss, Inc.

The mechanisms causing resistance to chemotherapeutic drugsin oral-cancer patients are poorly understood. New strategies forcircumvention of drug resistance can be designed by understandingthe biochemical basis of multidrug resistance (MDR). A largenumber of oral-cancer patients show poor or partial response tochemotherapy, so that drug resistance remains an enigma for oraloncologists. MDR has been extensively studied at the cellular andmolecular levels (Juliano and Ling, 1976; Riordanet al., 1985;Gottesman and Pastan, 1988, 1993; Gottesman, 1993; Biedler,1994; Zhouet al., 1996). In human cancers, the most commonlyobserved indicator of MDR is the over-expression of the multidrugresistance gene (mdr-1) product, P-glycoprotein (P-gp). This170-kDa transmembrane phosphoglycoprotein serves as an ATP-dependent efflux pump, enabling tumor cells to circumvent thetoxic effects of natural lipophilic drugs (Gottesman and Pastan,1993). Intrinsic over-expression of themdr-1 gene is found inhuman cancers derived from kidney, liver, colon, pancreas andadrenal glands (Goldsteinet al.,1989). Basal levels of P-gp are alsodetected in normal tissues (Cordon-Cardoet al.,1990). Though thepresence of P-gp in various cancers has been unequivocallyestablished, the correlation with clinical stage, response to chemo-therapy and prognosis is still speculative (Goldsteinet al., 1989;Moscowet al.,1989; Chanet al.,1990).Oral squamous-cell carcinoma ranks as the sixth most common

malignancy globally. It is the major cause of cancer-related death inIndian males (Sanghavi, 1981). Chronic tobacco- and betel-chewing habits account for the high incidence of oral cancer in theIndian population (Jussawala and Deshpande, 1971). The oraltumors that develop in this tobacco-abusing population are oftenpreceded by a well-defined pre-cancerous stage, termed leukopla-kia, which appears as a white plaque in the oral cavity and whichserves as a good model system for studying the various processesinvolved in oral tumorigenesis (Daftary, 1990; Bonneet al.,1991).Since P-gp expression in clinically distinct forms of oral tumors in

the context of the Indian population has not been reported, thepresent study was designed mainly to investigate (i) the expressionof P-gp in distinct stages of oral-cancer development,i.e., pre-malignant leukoplakic lesions showing histological evidence ofdysplasia and primary untreated and recurrent squamous-cellcarcinomas, also in normal oral tissues; and (ii) the correlation ofP-gp expression with the clinicopathological features.

MATERIAL AND METHODS

Tissue specimens

Biopsy or surgically resected tissue specimens from untreatedprimary as well as recurrent oral carcinomas and leukoplakiclesions as well as normal oral tissues were obtained from InstituteRotary Cancer Hospital, All India Institute of Medical Sciences,New Delhi, India. Tissue specimens were collected in Dulbecco’smodified Eagle’s medium (DMEM) supplemented with 10% FCS.No treatment had been given prior to removal of tissues except inrecurrent cases. The clinical history of recurrent oral-squamous-cell-carcinoma patients revealed that they had received combinedmodality therapy involving surgery, radiation and chemotherapyfor the management of primary tumors. Surgery on primary tumorswas followed by radiation therapy (5 to 6 Gy over a period of 50 to60 days), either alone or in combination with chemotherapy(cisplatin, 50 mg/m2, bleomycin, 15 mg/m2, and vincristine, 2mg/m2, i.v. injections over a period of 30 to 45 days). At the end oftherapy the tumors had regressed and there was no clinicalevidence of residual disease. However, the tumors had recurredwithin a period of 2 to 5 years and were subsequently analyzed forP-gp expression.Apiece of tissue specimenwas used for histopatho-logical examination. The clinical and pathological data recorded atthe time of enrollment of the subjects in the cancer clinic includedage, gender, tobacco- and betel-chewing history, site of tumor,histopathological differentiation and tumor staging. The TNMstaging of the tumors was conducted according to the 1987classification of the UICC (Hermanek and Sobin, 1987). Thevarious sites of untreated primary oral squamous carcinomasincluded buccal mucosa (5 cases), lower alveolus (4 cases), tongue(2 cases) and floor of the mouth (1 case). The site distribution ofrecurrent oral tumors comprised buccal mucosa (6 cases), loweralveolus (6 cases), tongue (3 cases) and floor of the mouth (3cases). Betel is an Indian masticatory consisting of a piece of arecanut rolled in a betel leaf coated with aqueous lime. This preparation

Abbreviations:MDR, multidrug resistance; P-gp, P-glycoprotein; scc,squamous-cell carcinoma; GSTpi, glutathione S-transferase-pi; MRP, mul-tidrug-resistance associated protein.

Contract grant sponsor: Department of Science and Technology of India.

*Correspondence to: Department of Biochemistry, All India Institute ofMedical Sciences,Ansari Nagar, New Delhi 110029, India. Fax: 91-11-686-2663.

Received 27 September 1996

Int. J. Cancer (Pred. Oncol.):74,128–133 (1997)

r 1997 Wiley-Liss, Inc.

Publication of the International Union Against CancerPublication de l’Union Internationale Contre le Cancer

is commonly known as ‘‘pan’’. Oral leukoplakia as well asmalignant tissue specimens were obtained from habitual pan and/ortobacco consumers. These patients had been consuming 10 to 20bidis/cigarettes or the equivalent amount of chewing tobacco for aperiod of 5 to 10 years. The betel chewers consumed 10 to 20 betel(pan) per day for a period of 5 to 10 years. Punch-biopsy specimensof normal oral tissues (12 cases) used as controls were obtainedfrom cancer-free patients who abstained from the use of tobaccoand were not habitual pan chewers. Four tissue specimens wereobtained from an ipsilateral site adjacent to the tumor of cancerpatients who were tobacco and betel (pan) consumers. Histopatho-logical examination of these tissue sections showed normal epithe-lium.

Histopathological gradingFormalin-fixed paraffin-embedded tissue sections (3–5 µm thick)

were stained with hematoxylin and eosin. The histological gradingof dysplastic lesions and squamous-cell carcinomas was carried outby 2 pathologists independently. Leukoplakic lesions showinghistological evidence of dysplasia were included in this study. Theuntreated primary and recurrent oral tumors were all squamous-cellcarcinomas.

Monoclonal antibodies (MAbs)Murine MAb MRK-16 was used to detect P-gp. It was a

generous gift from Dr. M.M. Gottesman (NCI, Bethesda, MD). It isan IgG2a isotype antibody which recognizes an external epitope ofP-gp and was used for FACS analysis of P-gp-bearing cells.

Flow-cytometric analysis of P-glycoprotein expressionFresh tissue specimens (normal, pre-malignant lesions, both

primary and recurrent tumors) were suspended in PBS (0.01 M, pH7.2) and immediately processed for FACS analysis. Single-cellsuspension was obtained by chopping, mincing the tissues into1- to 2-mm3 fragments by mechanical disintegration and forcingthe fragments through a fine nylon-wire-mesh sieve (mesh size 70µm) as described (Kauret al., 1994). The cell suspension waspassed through a 221⁄2-gauge needle to obtain a single-cell suspen-sion. Cells were washed with PBS, and viability was determined bytheTrypan-blue-dye-exclusionmethod. Viable cell aliquots (13 106cells/100 µl of PBS containing 1% BSA) were incubated with antiP-gp MAb, MRK-16 (5 µg/ml), or with an equivalent amount ofisotypic control mouse-IgG2a antibody for 30 min at 4°C. Cellswere washed twice with cold PBS containing 1% BSA andsubsequently incubated with 100 µl of PBS-BSA containing 50µg/ml fluorescein-isothiocyanate(FITC)-conjugated goat anti-mouse total immunoglobulins (National Institute of Immunology,New Delhi, India) in the dark at 4°C for 30 min. Cells were thenwashed 3 times with PBS and pelleted by centrifugation. The

pellets were re-suspended in 1% paraformaldehyde and 10,000events were acquired in a flow cytometer (FACSCAN, BectonDickinson, Mountain View, CA) using LYSIS II Software at anexcitation wavelength of 494 nm with a 517-nm-band pass filter.The machine was calibrated with calibrite beads and the patient’speripheral-blood mononuclear cells (PBMC). Heparinized periph-eral-blood samples were obtained from all the subjects included inthis study. PBMC were separated on Ficoll-paque (Lymphoprep,Dakopatts, Copenhagen, Denmark). Heparinized blood (10 µl)diluted with an equal volume of PBS (0.01 M, pH 7.2) wascarefully layered on lymphoprep in 1:1 ratio (vol/vol), andcentrifuged at 400g for 30 min at 18 to 20°C. Interphase cells werewashed with PBS (0.01 M, pH 7.2) 3 times before use for FACSanalysis. PBMC from the same patients were used as controls. Thetumor cells treated with mouse IgG2aantibody and goat anti-mousetotal immunoglobulins conjugated to FITC served as isotypecontrol for assessment of non-specific binding. Analysis was doneby LYSIS II Software. Tumor-cell populations were gated by usingforward-scatter (FSC) and side-scatter (SSC) parameters. Thecontaminating lymphocytes were removed by the gate constructedand confirmed by PBMC of the same patient. The number of cellsspecifically stained for P-gp was calculated by subtracting thenumber of cells stained with the mouse isotype-control antibodyfrom the total number of antibody-stained cells. For each sample10,000 events were analyzed.

Statistical analysisStatistical analysis of the data was done with Microstat Software.

The data were analyzed by Kruskal-Wallis (non-parametric) one-way analysis of variance and the Chi-squared test. Correlationsbetween P-gp expression and tumor variables such as tumor stagewere determined by Wilcoxon rank-sum test.p values less than0.05 were considered significant.

RESULTS

Flow-cytometric analysis of P-glycoprotein expression

Our study was conducted on 13 histologically confirmed dysplas-tic lesions, 12 primary and 18 recurrent malignant lesions of theoral cavity. Twelve normal oral tissues were also studied forcomparison of the data. Four tissue specimens obtained from anipsilateral site adjacent to the tumor and reported to be histologi-cally normal were also included in the study.Quantitative differences in P-gp expression in oral lesions at

different stages of neoplasia were analyzed by flow cytometry. Theclinical features and immunofluorescence analysis of P-gp expres-sion in normal and dysplastic oral tissues are shown in Table I.

TABLE I – P-GLYCOPROTEIN EXPRESSION IN NORMALAND DYSPLASTIC ORAL TISSUES

Patientnumber

Normal tissues P-GP expression1Patientnumber

Dysplastic tissues P-GP expression1

Gender/age (years) MF/CV %1ve cells Gender/age (years) MF/CV %1ve cells

1 M/56 9.29 59.18 1 M/58 30.08 71.112 M/64 16.68 21.38 2 M/56 10.4 53.83 F/55 12.82 22.03 3 M/60 7.26 10.724 F/60 9.84 25.9 4 M/64 44.32 205 F/52 6.66 79.83 5 F/55 21.17 26.926 M/35 28.85 37.02 6 F/40 11.55 1.047 M/39 7.18 1.95 7 M/50 24.78 4.818 M/50 15.31 9.78 8 M/60 9.07 50.369 M/60 8.62 39.6 9 F/36 15.59 13.4410 M/58 14.28 3.8 10 F/63 39.02 44.3311 M/45 13.32 8.04 11 M/44 34.11 1912 M/55 1.98 6.44 12 M/70 20.4 20.8512 M/55 276.28 74.77 13 F/18 112.88 55.622 F/50 38.47 54.3832 M/36 47.88 67.5342 F/70 182.88 92.21

1MF/CV, mean fluorescence/channel value.–2Tissue taken from oral-cancer patients from the same sideas that of the tumor showing histological features characteristic of normal epithelium.

129P-GLYCOPROTEIN EXPRESSION IN ORAL CANCER

A significant increase in the level of P-gp expression, fluores-cence intensity 27.806 3.56 (mean6 SE) was observed in oraldysplastic lesions as compared with normal oral-tissue specimensobtained from cancer-free patients, fluorescence intensity 12.0661.95 (mean6 SE) (p, 0.01) as shown in Table II. However, therewas no significant increase in the percentage of P-gp-positive cellsin dysplastic lesions as compared with normal tissues.P-gp expression was also studied in 4 tissue specimens that

showed histological features of normal epithelium obtained froman ipsilateral site adjacent to the oral tumors. A significant increasein P-gp expression was observed in these tissue specimens, asshown by a higher percentage of P-gp-positive cells 72.26 7.88(mean6 SE) as well as a higher level of P-gp expression:136.376 57.12 (mean fluorescence intensity6SE) as comparedwith normal tissues obtained from cancer-free patients (p5 0.018andp5 0.004 respectively), as shown in Table III.The clinical and histopathological data as well as immunofluores-

cence analysis of P-gp expression in untreated primary oralsquamous-cell carcinomas and recurrent oral carcinomas aresummarized in Tables IV and V respectively. The correlation ofP-gp expression with clinical and pathological features of thepatients including tumor site, age, gender, tumor stage andhistopathological grading was also examined. No significant corre-lation was observed between P-gp expression and tumor site, age orgender of the patients. Since most of the primary tumors were welldifferentiated, no correlation between the histologic grade of tumorand the level of P-gp expression could be established.A significant increase in level of P-gp expression (mean fluores-

cence) was observed in stage-IV recurrent tumors as comparedwith stage-III recurrent tumors (p5 0.01) (Table VI). Howeverthere was no significant difference in the number of P-gp positivecells observed in these 2 groups.Representative flow cytograms of normal tissue, dysplastic

lesion, primary and recurrent tumors of oral cavity are shown inFigure 1,a,b,c,andd respectively.The results of FACScan analysis of normal oral tissues and those

of different groups of neoplasms are summarized in Table II. Therewas a significant increase in P-gp expression (146.786 51.50,mean fluorescence6SE) in primary untreated oral carcinomas incomparison with normal tissues (p, 0.01). Percentage of P-gp-positive cells also showed a significant increase in primary tumorsin comparison with normal tissues (p, 0.05) and with dysplastic

lesions (p, 0.05). Recurrent-squamous-cell-carcinoma patientsexpressed significantly higher levels of P-gp than those withnormal tissues (p, 0.001), and dysplastic cases (p, 0.01). Asignificant increase in the percentage of P-gp-positive cells wasobserved in recurrent tumors as compared with normal tissues(p, 0.01) and with dysplastic lesions (p, 0.01). An increase inP-gp expressionwas observed in recurrent tumors (383.626 103.50,mean fluorescence6SE) as compared with primary tumors

TABLE II – FLOW-CYTOMETRIC ANALYSIS OF P-GLYCOPROTEIN EXPRESSIONIN ORAL TISSUES

Sample Numberof cases

Meanfluorescence

(channel value)6 SE

Mean percentpositivecells6 SE

Normal (N) 12 12.066 1.95 26.246 6.95Dysplasia (D) 13 27.806 3.56 29.466 6.32Primary oral carcinoma(P)

12 146.786 51.50 67.936 10.25

Recurrent oral carci-noma (R)

18 383.626 103.50 81.236 2.83

A comparison of P-gp expression in various groups showed:fluorescence intensity, N/D, N/P, R/D,p, 0.01; N/R,p, 0.001; P/R,p , 0.08; percentage of P-gp positive cells, N/P, D/P,p , 0.05; N/R,D/R,p, 0.01; P/R,p, 0.15.

TABLE III – COMPARISON OF P-GLYCOPROTEIN EXPRESSION IN NORMALORAL TISSUES FROM CANCER-FREE PATIENTSAND HISTOLOGICALLY

NORMAL TISSUEADJACENT TO TUMOR IN CANCER PATIENTS

Sample Numberof cases

Meanfluorescence

channel value6 SE

Mean percentpositivecells6 SE

Cancer-free subjects 12 12.066 1.95 26.246 6.95Cancer patients 4 136.376 57.12 72.226 7.88

Mean fluorescence,p5 0.004; percentage positive cells,p5 0.018.

TABLE IV – P-GLYCOPROTEIN EXPRESSION IN PRIMARYORAL CARCINOMAS

Patientnumber

Gender/age

(years)

Tumorstage Grading1

P-gp expression

(mean fluorescencechannel value)

Percentpositive cells

1 M/55 T1 M 1.08 9.152 M/58 T1 M 170.05 8.483 M/35 T2 W 43.06 32.034 M/40 T3 P 8.20 60.175 F/50 T3 M 85.83 14.706 M/60 T3 P 23.82 94.677 M/38 T4 W 56.27 62.818 M/65 T4 M 99.64 89.229 M/45 T4 W 428.19 97.81102 F/70 T4 W 371.59 96.8511 F/36 T4 M 192.90 95.3712 M/45 T4 M 280.83 57.12

1Histopathological grading of oral carcinoma. P, poorly differenti-ated; M, moderately differentiated; W, well differentiated.–2Cancerpatient whose surrounding tissue showing histological features ofnormal oral mucosa was also taken for P-gp analysis.

TABLE V – P-GLYCOPROTEIN EXPRESSION IN RECURRENT ORALCARCINOMAS

Patientnumber

Gender/age

(years)

Tumorstage Grading1

P-gp expression

(mean fluorescencechannel value)

Percentpositive cells

1 M/50 T1 W 78.79 76.072 F/50 T2 P 77.59 61.563 F/52 T2 W 39.53 77.224 F/55 T3 W 52.15 63.005 F/60 T3 W 117.34 68.846 M/39 T3 W 363.66 97.837 F/60 T3 W 70.41 77.088 M/38 T3 W 336.65 95.379 M/55 T3 W 359.97 95.9410 M/51 T3 W 141.99 87.2011 F/75 T3 W 165.72 92.3112 M/36 T4 W 402.85 74.9913 M/55 T4 W 276.28 74.7014 M/51 T4 W 1551.30 69.0415 M/65 T4 W 416.10 98.5316 F/60 T4 W 1446.26 90.3917 M/50 T4 W 202.10 75.3818 F/60 T4 W 610.64 86.86

1Histopathological grading: P, poorly differentiated;W, well differen-tiated.–2Oral-cancer patients in whom tissue taken from the same sideas that of the tumor showed histological features characteristic ofnormal epithelium.

TABLE VI – CORRELATION BETWEEN P-GLYCOPROTEIN EXPRESSIONANDTUMOR STAGE IN RECURRENT ORAL SQUAMOUS-CELL CARCINOMAS

Numberof patients

Tumorstage

Pgp-expression

Mean fluorescencechannel value(mean6 SE)

Percentpositive cells(mean6 SE)

1 T1 78.79 76.072 T2 58.566 19.03 69.396 7.838 T3 200.996 46.49 84.706 4.747 T4 700.796 211.90 81.416 4.02

Mean fluorescence intensity: T3/T4,p5 0.01.

130 JAIN ET AL.

(146.786 51.50 mean fluorescence,6SE). However this increasefailed to reach borderline significance (p5 0.08). There was nosignificant difference in the percentage of P-gp-positive cells inprimary and in recurrent tumors.

DISCUSSION

P-gp expression was assessed in tissue specimens at differentstages of oral-tumor progression, with a view to investigating itspossible role in oral carcinogenesis. Our study demonstrates thedifferential expression of P-gp in normal, dysplastic and malignantoral tissues, suggesting its association with a more progressedmalignant phenotype in oral tumorigenesis. P-gp expression hasbeen shown in a variety of normal tissues (liver, kidney, pancreaticductules, small and large intestines; Thiebautet al.,1987). The highlevel of P-gp expression observed on the surface of adrenocorticalcells, capillaries in brain and testis, peripheral mononuclear cellsand trophoblasts of human placenta suggests its involvement intrans-epithelial transport of toxic endogenous metabolites andxenobiotics (Thiebautet al., 1989). The possibility that P-gp canact as an efflux pump for compounds in diet or encountered in theenvironment has been suggested (Rothenberg and Ling, 1989;Bradleyet al.,1990). Yehet al.(1992) have shown that the cellularburden of certain carcinogens [benzo(a)pyrene] may be mitigatedby P-gp, suggesting its role in cellular defence against carcinogens.

The oral cavity is constantly exposed to environmental insultsand toxic chemicals. Thus P-gp may play a major role as axenobiotic transporter in this tissue. The oral cavity of habitualtobacco chewers is constantly exposed to carcinogens, tobacco-specific nitrosoamines, thus the presence of P-gp may affect theirsusceptibility to transformation. Significant increase in the level ofP-gp expression (p, 0.01) was observed in oral dysplastic lesionsas compared with normal tissue, indicating that alterations in P-gpexpression may be one of the important early events in oraltumorigenesis.Furthermore, higher expression of P-gp was observed in oral

mucosa adjacent to the tumor in cancer patients as compared withnormal oral mucosa from cancer-free subjects. The increased P-gpexpression in oral mucosa adjacent to the tumor site in thesepatients may be amanifestation of field cancerization, the predispo-sition of an entire field of tissue to the development of multiplecancers through repeated carcinogenic insults to that field (Slaugh-teret al.,1953).Detection of P-gp in primary oral tumors suggests that these

tumors are likely to exhibit innate drug resistance and couldpossibly be the rationale for the limited use of chemotherapy inoral-cancer management during the early stages. Significant expres-sion of P-gp in untreated primary oral tumors may serve as amolecular marker for diagnosis of chemoresistance in the clinical

a b

cd

FIGURE 1 – Flow-cytometric analysis of P-glycoprotein expression in oral tissues. Cells (13 106) isolated from oral-tissue specimens weretreated with anti-P-gp MAb MRK-16 and secondary antibody (FITC-conjugated goat anti-mouse total immunoglobulins) as described in‘‘Material and Methods’’. P-gp expression was analyzed by a FACScan flow cytometer (Becton Dickinson) using Lysis II Software; 10,000 eventsacquired were gated for the required cell population, and the corresponding mean fluorescence/channel value was determined(a)Normal tissue;(b) dysplastic lesion (leukoplakia);(c) untreated primary squamous-cell carcinoma;(d) recurrent squamous-cell carcinoma.

131P-GLYCOPROTEIN EXPRESSION IN ORAL CANCER

samples. The mechanisms for the resistance of oral tumors areunknown, but may be multifactorial. We have reported a significantincrease in glutathione-S-transferase-pi (GST-pi) expression inrecurrent oral tumors of habitual tobacco consumers as comparedwith matched normal tissues (data not shown). In the present study,a significant increase in P-glycoprotein expression was observed inuntreated primary (p, 0.01) as well as in recurrent oral tumors(p, 0.001) in habitual tobacco consumers. Studies aimed atdetermining the correlation of over-expression of P-glycoproteinand/or GST-pi within vitro drug resistance of oral tumors arecurrently in progress in our laboratory. Over-expression ofP-glycoprotein and glutathione-S-transferase pi (GST-pi) has beenshown in drug-resistant non-small-cell lung carcinomas of smokers(Volm et al.,1991). Significant correlations were observed betweendoxorubicin resistance of these tumorsin vitro and the expressionof P-glycoprotein (p, 0.0001) or of GST-pi (p, 0.0001). Keithet al. (1990) found a weak correlation between the expression ofmdr-1and GST-pi and drug resistance in human breast carcinomas.The functional multidrug-resistance phenotype associated withcombined over-expression of P-gp/mdr-1andmultidrug-resistance-associated protein (MRP), together with 1-b-D-arabinofuranosylcy-tosine sensitivity, has been suggested as useful in predictingclinical response in acute myeloid leukemia (Schuurhuiset al.,1996). Bellet al. (1985) proposed that P-gp could be a potentialmolecular marker for detection of multidrug-resistant tumor cellsin ovarian cancers. High levels of P-gp have also been observed insolid tumors, such as carcinoma of colon, liver or kidney, suggest-ing that these tumors exhibit innate drug resistance (Weinsteinetal., 1991; Graciadel Moralet al.,1995). The hallmark of our studywas the marked elevation of P-gp levels observed in recurrent oraltumors (p, 0.001) in comparison with normal tissues. Over-expression of P-gp has also been shown during the course ofchemotherapy in leukemia, lymphomas, breast and renal carcino-mas (Sauerbreyet al.,1994; Pastanet al.,1991).

Alternatively, over-expression of P-gp in untreated primary oralcarcinomasmay be related to the inherent process of oral tumorigen-esis. Over-expression of themdr-1 gene in cancers derived fromtissues not normally expressing high levels of P-gp could also beexplained on the basis that the process of malignant transformationper secan activatemdr-1gene expression. Chinet al. (1992) haveshown thatrasandmutantp53can activate themdr-1promoter.Wehave shown that alterations inp53 expression are an early eventduring oral tumorigenesis: detectable levels of mutantp53 wereobserved not only in primary oral tumors but also in oral dysplasticlesions among the Indian population (Kauret al.,1994). Saranathet al. (1991) have shownras mutations in oral malignancies inIndian population. Activation of themdr-1gene in oral tumors maybe partly under the control of mutantp53.Specificp53mutationshave been shown to be associated withde novo resistance todoxorubicin in breast-cancer patients (Aaset al.,1996). Differen-tial P-gp expression has also been observed during rat livercarcinogenesis, associated with a more malignant phenotype.Differential expression of Pgp in dysplastic lesions and in primaryand recurrent oral tumors, as shown in our study, also in studies ondifferent malignancies (Bradleyet al., 1992), suggests that P-gpmay be more than a marker for drug resistance. Further studies arebeing carried out, in order to understand the influence of P-gp onthe biological behavior of oral tumors and on prognosis, and to planthe treatment of oral cancer.

ACKNOWLEDGEMENTS

This work was supported by a research grant fromDepartment ofScience and Technology of India. We are grateful to Dr. I. Nath forgiving permission for the use of the FACScan flow cytometer. Theskillful help of Dr. N. Nath in preparation of the manuscript isappreciated.

REFERENCES

AAS, T., BOOR ESEN, A.L., GEISLER, S., SORENSEN, B.G., JOHNSEN, H.,VARHAUG, J.E., AKSLEN, L.A. and LONNING, P.E., Specificp53mutations areassociated withde novoresistance to doxorubicin in breast-cancer patients.Nature Med.,2,811–814 (1996).

BELL, D.R., GERLACH, J.H., KARTNER, N., BUICK, R.N. and LING, V.,Detection of P-glycoprotein in ovarian cancer: a molecular marker associ-ated with drug resistance.J. clin. Oncol.,3,311–315 (1985).

BIEDLER, J.L., Drug resistance: genotypeversusphenotype. Thirty-secondG.H.A. Clowes Memorial Award Lecture.Cancer Res.,54, 666–678(1994).

BONNE, C.W., KELLOFF, G.J. and STEELE, V., Histopathology of thepre-malignant process.Cancer Bull.,43,481–484 (1991).

BRADLEY, G., GEORGES, E. and LING, V., Sex-dependent and -independentexpression of the P-glycoprotein isoforms in Chinese hamster.J. cell.Physiol.,145,398–408 (1990).

BRADLEY, G., SHARMA, R., RAJALAKSHMI , S. and LING, V., P-glycoproteinexpression during tumor progression in the rat liver.Cancer Res.,52,5154–5161 (1992).

CHAN, H., THORNER, P.S., HADDAD, G. and LING, V., Immunohistochemicaldetection of P-glycoprotein: prognostic correlation in soft-tissue sarcoma ofchildhood.J. clin. Oncol.,8,689–704 (1990).

CHIN, K.V., UEDA, K., PASTAN, I. and GOTTESMAN, M.M., Modulation ofactivity of the promoter of the humanMDR-1by rasandp53. Science,255,459–462 (1992).

CORDON-CARDO, C., O’BRIEN, J.B., BOCCIA, J., CASALS, D., BERTINO, J.R.and MELAMED, M.R., Expression of the multidrug-resistance-gene product(P-glycoprotein) in human normal and tumor tissues.J. Histochem.Cytochem.,38,1277–1287 (1990).

DAFTARY, D.K., The situation in high-risk areas of the world.In: N.W.Johnson (ed.),Risk markers for oral diseases,Vol. 2,Cambridge UniversityPress, Cambridge (1990).

GOLDSTEIN, L.J., GALSKI, H., FOJO, A., WILLINGHAM , M., LAI, S.L.,GAZADAR, A., PIRKER, R., GREEN, A., CRIST, W., BRODEUR, G., GRANT, C.,LIEBER, M., COSSMAN, J., GOTTESMAN, M.M. and PASTAN, I., Expression of

a multidrug resistance gene in human cancers.J. natl. Cancer Inst.,81,116–124 (1989).

GOTTESMAN, M.M., How cancer cells evade chemotherapy. SixteenthRichard and Hinda Rosenthal Foundation Award Lecture.Cancer Res.,53,747–754 (1993).

GOTTESMAN, M.M. and PASTAN, I., The multidrug transporter, a double-edged sword.J. biol. Chem.,263,12163–12166 (1988).

GOTTESMAN, M.M. and PASTAN, I., Biochemistry of multidrug resistancemediated by the multidrug transporter.Ann. Rev. Biochem.,62, 385–427(1993).

GRACIADEL MORAL, R., O’VALLE, F., ANDUJAR, M., AGUILAR, M., LUCENA,M.A., LOPEZ-HIDALGO, J., RAMIREZ, C., MEDINA-CANO, M.T., AGUILAR, D.and GOMEZMORALES, M., Relationship between P-glycoprotein expressionand cyclosporin A in kidney.Amer. J. Pathol.,1246,398–408 (1995).

HERMANEK, P. and SOBIN, L.H. (eds.), UICC: TNM classification ofmalignant tumors,4th ed., Springer, Berlin (1987).

JULIANO, R.L. and LING, V., A surface glycoprotein modulating drugpermeability in Chinese-hamster ovary cell mutants.Biochim. biophys.Acta,455,152–156 (1976).

JUSSAWALA, D. and DESHPANDE, V., Evaluation of cancer risk in tobaccochewers and smokers: an epidemiologic assessment.Cancer,28, 244–252(1971).

KAUR, J., SRIVASTAVA , A. and RALHAN , R., Over-expression of p53 proteinin betel- and tobacco-related human oral dysplasia and squamous-cellcarcinoma in India.Int. J. Cancer,58,340–345 (1994).

KEITH, W.N., STALLORD, S. and BROWN, R., Expression ofMDR-1and GSTin human breast tumors: comparison toin vitro chemosensitivity.Brit. J.Cancer,61,712–716 (1990).

MOSCOW, J.A., FAIRCHILD, C.R., MADDEN, M.J., RANSOM, D.T., WIEAND,H.S., O’BRIEN, E.E., POPLACK, D.G., COSSMAN, J., MYERS, C.E. andCOWAN, K.H., Expression of anionic glutathione-S-transferase andP-glycoprotein genes in human tissues and tumors.Cancer Res.,49,1422–1428 (1989).

PASTAN, I., WILLINGHAM , M.C. and GOTTESMAN, M.M., Molecular manipu-

132 JAIN ET AL.

lations of the multidrug transporter: a new role for transgenic mice.FASEBJ.,5,2523–2528 (1991).RIORDAN, J.R., DEUCHARS, K., KARTNER, N., ALON, N., TRENT, J. and LING,V., Amplification of P-glycoprotein genes in multidrug-resistant mamma-lian cell lines.Nature (Lond.),316,817–819 (1985).ROTHENBERG, M. and LING, V., Multidrug resistance: molecular biology andclinical relevance.J. nat. Cancer Inst.,81,907–920 (1989).SANGHAVI , L.D., Epidemiologic and intervention studies. Screening: cancerepidemiology: the Indian scene.J. Cancer Res. clin. Oncol.,9,1–14 (1981).SARANATH, D., CHANG, S.E., BHOITE, L.T., PANCHAL, R.G., KERR, I.B.,MEHTA, A.R., JOHNSON, N.W. and DEO, M.G., High-frequency mutation incodons 12 and 61 of H-ras oncogene in chewing-tobacco-related humanoral carcinoma in India.Brit. J. Cancer,63,573–578 (1991).SAUERBREY, A., ZINTL, F. and VOLM, M., P-glycoprotein and glutathioneS-transferase in childhood acute lymphoblastic leukemia.Brit. J. Cancer,70,1144–1149 (1994).SCHUURHUIS, G.J., BROXTERMAN, H.J., OSSENKOPPELE, G.J., BAAK , J.P.,EEKMAN, C.A., KUIPER, C.M., FELLER, N., VAN HEIJNINGEN, T.H.M.,KLUMPER, E., PIETERS, R., LANKELMA , J. and PINEDO, H.M., Functionalmultidrug-resistance phenotype associated with combined over-expressionof pgp/MDR-1 and MRP together may predict clinical response in acutemyeloid leukemia.Clin Cancer Res.,1,81–93 (1996).SLAUGHTER, D.L., SOUTHWICK, H.W. and SMEJKAL, W., ‘‘Field canceriza-tion’’ in oral stratified squamous epithelium: clinical implications ofmulticentric origin.Cancer,6,963–968 (1953).

THIEBAUT, F., TSURUO, T., HAMADA , H., GOTTESMAN, M.M., PASTAN, I. andWILLINGHAM , M.C., Cellular localization of the multidrug-resistance-geneproduct. P-glycoprotein in normal human tissues.Proc. nat. Acad. Sci.(Wash.),84,7735–7738 (1987).THIEBAUT, F., TSURUO, T., HAMADA , H., GOTTESMAN, M.M., PASTAN, I. andWILLINGHAM , M.C., Immunohistochemical localization in normal tissues ofdifferent epitopes in the multidrug-transport protein, P170: evidence forlocalization in brain capillaries and cross-reactivity of one antibody with amuscle protein.J. Histochem. Cytochem.,37,159–164 (1989).VOLM, M., MATTERN, J. and SAMSEL, B., Over-expression of P-glycoproteinand glutathione-S-transferase pi in resistant non-small-cell lung carcinomasof smokers.Brit. J. Cancer,64,700–704 (1991).WEINSTEIN, R.S., JAKATE, S.M., DOMINQUEZ, J.M., LEBOVITZ, M.D., KOUK-OULIS, J.K., KUSZAK, J.R., KLUSENS, L.F., GROGAN, T.M., SACLARIDES, T.J.,RONINSON, I. and COON, J.S., Relationship of the expression of themultidrug-resistance-gene product (P-glycoprotein) in human colon carci-noma to local tumor aggressiveness and lymph-node metastasis.CancerRes.,51,2720–2726 (1991).YEH, G.C., LOPACZYNSKA, J., POORE, C.M. and PHANG, J.M., A newfunctional role for P-glycoprotein: efflux pump for benzo (a) pyrene inhuman breast-cancer MCF-7 cells.Cancer Res.,52,6692–6695 (1992).ZHOU, D.C., RAMOND, S., VIGUIE, F., FAV SSAT, A.M., ZITTOUN, R. andMARIE, J.P., Sequential emergence ofMRP- and MDR-1-gene over-expression as well asMDR-1-gene translocation in homoharringtonine-selected K562 human leukemia cell lines.Int. J. Cancer,65, 365–371(1966).

133P-GLYCOPROTEIN EXPRESSION IN ORAL CANCER