Genome-wide association studies and Mendelian randomization analyses provide insights into the causes of early-onset colorectal cancer.

BACKGROUND: The incidence of early-onset colorectal cancer (EOCRC; diagnosed <50 years of age) is rising globally; however, the causes underlying this trend are largely unknown. CRC has strong genetic and environmental determinants, yet common genetic variants and causal modifiable risk factors underlying EOCRC are unknown. We conducted the first EOCRC-specific genome-wide association study (GWAS) and Mendelian randomization (MR) analyses to explore germline genetic and causal modifiable risk factors associated with EOCRC. PATIENTS AND METHODS: We conducted a GWAS meta-analysis of 6176 EOCRC cases and 65 829 controls from the Genetics and Epidemiology of Colorectal Cancer Consortium (GECCO), the Colorectal Transdisciplinary Study (CORECT), the Colon Cancer Family Registry (CCFR), and the UK Biobank. We then used the EOCRC GWAS to investigate 28 modifiable risk factors using two-sample MR. RESULTS: We found two novel risk loci for EOCRC at 1p34.1 and 4p15.33, which were not previously associated with CRC risk. We identified a deleterious coding variant (rs36053993, G396D) at polyposis-associated DNA repair gene MUTYH (odds ratio 1.80, 95% confidence interval 1.47-2.22) but show that most of the common genetic susceptibility was from noncoding signals enriched in epigenetic markers present in gastrointestinal tract cells. We identified new EOCRC-susceptibility genes, and in addition to pathways such as transforming growth factor (TGF) ß, suppressor of Mothers Against Decapentaplegic (SMAD), bone morphogenetic protein (BMP) and phosphatidylinositol kinase (PI3K) signaling, our study highlights a role for insulin signaling and immune/infection-related pathways in EOCRC. In our MR analyses, we found novel evidence of probable causal associations for higher levels of body size and metabolic factors-such as body fat percentage, waist circumference, waist-to-hip ratio, basal metabolic rate, and fasting insulin-higher alcohol drinking, and lower education attainment with increased EOCRC risk. CONCLUSIONS: Our novel findings indicate inherited susceptibility to EOCRC and suggest modifiable lifestyle and metabolic targets that could also be used to risk-stratify individuals for personalized screening strategies or other interventions.

Host immune genetic variations influence the risk of developing acute myeloid leukaemia: results from the NuCLEAR consortium.

The purpose of this study was to conduct a two-stage case control association study including 654 acute myeloid leukaemia (AML) patients and 3477 controls ascertained through the NuCLEAR consortium to evaluate the effect of 27 immune-related single nucleotide polymorphisms (SNPs) on AML risk. In a pooled analysis of cohort studies, we found that carriers of the IL13rs1295686A/A genotype had an increased risk of AML (PCorr = 0.0144) whereas carriers of the VEGFArs25648T allele had a decreased risk of developing the disease (PCorr = 0.00086). In addition, we found an association of the IL8rs2227307 SNP with a decreased risk of developing AML that remained marginally significant after multiple testing (PCorr = 0.072). Functional experiments suggested that the effect of the IL13rs1295686 SNP on AML risk might be explained by its role in regulating IL1Ra secretion that modulates AML blast proliferation. Likewise, the protective effect of the IL8rs2227307 SNP might be mediated by TLR2-mediated immune responses that affect AML blast viability, proliferation and chemorresistance. Despite the potential interest of these results, additional functional studies are still warranted to unravel the mechanisms by which these variants modulate the risk of AML. These findings suggested that IL13, VEGFA and IL8 SNPs play a role in modulating AML risk.

Association of microsomal epoxide hydrolase polymorphisms and lung cancer risk.

Microsomal epoxide hydrolase (mEH) plays a dual role in the detoxification and activation of tobacco procarcinogens. Two polymorphisms affecting enzyme activity have been described in the exons 3 and 4 of the mEH gene, which result in the substitution of amino acids histidine to tyrosine at residue 113, and arginine to histidine at residue 139, respectively. We performed a hospital-based case-control study consisting of 277 newly diagnosed lung cancer patients and 496 control subjects to investigate a possible association between these two polymorphisms and lung cancer risk. The polymorphisms were determined by polymerase chain reaction/restriction fragment length polymorphism and TaqMan assay using DNA from peripheral white blood cells. Logistic regression was performed to calculate odds ratios (ORs), confidence limits (CL) and to control for possible confounders. The exon 3 polymorphism of the mEH gene was associated with a significantly decreased risk of lung cancer. The adjusted OR, calculated relative to subjects with the Tyr113/Tyr113 wild type, for the His113/His113 genotype was 0.38 (95% CL 0.20-0.75). An analysis according to histological subtypes revealed a statistically significant association for adenocarcinomas; the adjusted OR for the His113/His113 genotype was 0.40 (95% CL 0.17-0.94). In contrast, no relationship between the exon 4 polymorphism and lung cancer risk was found. The adjusted OR, calculated relative to the His139/His139 wild type, was for the Arg139/Arg139 genotype 1.83 (0.76-4.44). Our results support the hypothesis that genetically reduced mEH activity may be protective against lung cancer.

Genetic polymorphisms of CYP1A1 and GSTM1 and lung cancer risk.

Susceptibility to lung cancer may, in part, be determined by interindividual differences in the cytochrome P450-catalysed bioactivation and the glutathione S-transferase-catalysed detoxification of procarcinogens. Therefore a lung cancer case-control study was set up to investigate the association of three polymorphisms of the CYP1A1 gene (CYP1A1*2A, CYP1A1*2B, CYP1A1*4) and GSTM1*0 genotype with lung cancer risk in Austrian Caucasians. Genomic DNA was isolated from the peripheral blood lymphocytes of 134 male lung cancer patients and 134 age-matched controls with nonmalignant conditions and PCR-based analyses were performed. There was no significant difference in risk between cases and controls, either for the CYP1A1*2A (OR=1.09, 95%CI=0.46-2.58), CYP1A1*2B (OR=1.09, 95%CL=0.46-2.58) or for the CYP1A1*4 polymorphism (OR=0.49, 95%CL=0.20-1.16). The prevalence of the GSTM1*0 genotype in the lung cancer group (47.8%) was comparable to that found in the control group (49.3%) and also had no effect on lung cancer risk (OR=0.94, 95%CL=0.54-1.57). Further, in a subgroup of male ever-smokers (n=126), no significant influence on the relative risk was found for these polymorphisms. Our results suggest that these investigated polymorphisms can not be considered as genetic susceptibility markers for lung cancer within the Austrian Caucasian population.

Polymorphisms of glutathione-S-transferase genes (GSTP1, GSTM1 and GSTT1) and prostate-cancer risk.

Several polymorphic glutathione-S-transferase (GST) enzymes are involved in the metabolism of a number of potential prostate carcinogens and are thought to engage in the transport of steroid hormones. A case-control study was conducted to determine the association of the GSTP1, GSTM1 and GSTT1 polymorphisms and prostate-cancer risk. The study population consisted of 166 patients with previously untreated, histologically proven prostate cancer and 166 age-matched control patients with benign prostatic hyperplasia (BPH), all of them Caucasians. In the GSTP1 gene, 2 polymorphic alleles, GSTP1*B and GSTP1*C, have been described in addition to the wild-type allele, GSTP1*A. Both polymorphic GSTP1 alleles have an A-to-G transition in exon 5, causing an isoleucine-to-valine change. The GSTP1*C allele has an additional transition from C to T. For GSTM1 as well as GSTT1, the polymorphic allele is a deletion of the gene. The proportion of individuals homozygous for the GSTP1 variant alleles (GSTP1*B/*B, GSTP1*B/*C and GSTP1*C/*C) was significantly lower in prostate-cancer patients (4.8%) than in BPH controls (14.5%), and the odds ratio (OR) was 0.24 [95% confidence interval (CI) = 0.09-0.61). The heterozygous genotypes (GSTP1*A/*B and GSTP1*A/*C) were also lower in the cancer group, though this was not significant. On the contrary, no significant effect on prostate-cancer risk was detectable for either GSTM1 (OR = 0.86, 95% CI = 0.55-1.36) or GSTT1 (OR = 0.78, 95% CI = 0.43-1.42). Of the polymorphic GSTs, GSTP1 is the most interesting candidate as a biomarker for prostate-cancer risk as we found a 76% reduced risk in men homozygous for the polymorphic GSTP1 alleles compared to those with wild-type GSTP1.

Elevated plasma levels of crosslinked fibrinogen gamma-chain dimer indicate cancer-related fibrin deposition and fibrinolysis.

Cancer-related fibrin deposition and fibrinolysis were investigated by two-dimensional gel electrophoresis of human solid tumor and effusion specimen in addition to plasma samples. Fibrinogen gamma-chain dimer indicating fibrin deposition and plasmin-generated fibrinogen beta-chain fragments were identified in various solid tumor types by amino acid sequencing, mass spectrometry analysis and Western blotting. In tumor-associated effusions, these techniques allowed to observe plasmin-generated fragments of fibrinogen alpha, beta and gamma-chains in addition to elevated levels of acute-phase proteins. Similar observations were made in case of inflammation-associated effusions. No fibrin degradation product was observed in plasma samples, however, high amounts of fibrinogen gamma-chain dimer crosslinked by transglutaminase were detected in plasma from tumor patients, but not in plasma from controls and patients suffering acute infections and/or inflammations. This finding demonstrated that high transglutaminase activity may be associated with cancer. The presented data indicate that the amount of crosslinked fibrinogen gamma-chain dimer in plasma may correlate with tumor-associated fibrin deposition. The tumor-biological relevance of this potential marker protein is discussed.

Association of vitamin D receptor and 17 hydroxylase gene polymorphisms with benign prostatic hyperplasia and benign prostatic enlargement.

OBJECTIVES: To determine whether polymorphisms in 17 hydroxylase (CYP17) and vitamin D receptor (VDR) genes have an association to prostate volume/histology and endocrine patterns in elderly men with lower urinary tract symptoms (LUTS). METHODS: Elderly men with LUTS underwent the following investigations: International Prostate Symptom Score (IPSS), uroflowmetry, serum prostate-specific antigen (PSA) assessment of prostate volume, and an endocrine study. Polymorphisms of CYP17 (T-->C substitution in the 5' promoter region) and VDR (T1055C) genes were detected by polymerase chain reaction followed by restriction-length polymorphism analysis, using DNA from peripheral white blood cells. Clinical and endocrine parameters and the prostate stroma/epithelial ratio were correlated to CYP17 and VDR genotypes. RESULTS: A total of 148 (mean +/- SD, 67.0 +/- 9.7 years) patients were analyzed. IPSS (17.8 +/- 7.0), prostate volume (41.9 +/- 17.9 cc), maximum flow rate (10.9 +/- 5.8 mL/s), and PSA (4.7 +/- 4.7 ng/mL) indicate a typical LUTS population. Mean endocrine levels were consistently within age-specific reference values. Neither CYP17 nor VDR gene polymorphisms revealed an association to prostate size, PSA, clinical parameters, and endocrine parameters. Men who had the A1/A1 CYP17 genotype had on average a greater stromal/epithelial ratio than men with the A1/A2 or A2/A2 genotypes, yet after adjusting for multiple testing, this significance disappeared. CONCLUSIONS: Gene polymorphisms of CYP17 and VDR have no association to prostate volume, clinical parameters, and endocrine parameters in elderly men. The association of CYP17 polymorphism and prostate histology warrants further studies. Assessment of gene polymorphisms might provide new insights into the pathogenesis of benign prostatic hyperplasia and benign prostate enlargement and may hold promise as genetic biomarkers of this disease.

A polymorphism in the CYP17 gene is associated with prostate cancer risk.

CYP17 encodes the enzyme cytochrome P-450c17 alpha, which mediates both 17 alpha-hydroxylase and 17,20-lyase in the steroid biosynthesis pathway. A polymorphism in the 5; promoter region of the CYP17 gene has been described. Steroid hormones, especially androgens, are believed to play a key role in the etiology of prostate cancer. Therefore, polymorphisms in genes involved in the androgen metabolism may affect the risk of prostate cancer. We conducted a case-control study of 63 patients with untreated histologically proven prostate cancer and 126 age-matched control men with benign prostatic hyperplasia (BPH) to determine whether a polymorphism in the CYP17 gene is associated with prostate cancer risk. This polymorphism was investigated by PCR/RFLP using DNA from lymphocytes. The transition (T-->C) in the risk allele (A2) creates a new recognition site for the restriction enzyme MspAI, which permits designation of the wildtype (A1) and the risk allele (A2). The prevalence of the A2/A2 genotype was significantly higher (P = 0.03) in the cancer group (23.8%) than in the BPH control group (9.5%). We found an increased risk in men carrying 2 A2 alleles (OR = 2.80, 95%CI = 1.02-77.76). For carrier with at least 1 A2 allele, the OR was 0.90 (95%CI = 0.43-1.89). After stratification by median age (66 years) at time of diagnosis, a marked increased risk was found in carriers of the A2/A2 genotype older than 66 years (OR = 8.93, 95%CI = 1.78-49.19, P = 0.01). Although the sample size is rather small and the controls are BPH patients, our results suggest that the CYP17A2/A2 genotype may be a biomarker for prostate cancer risk, especially for older men.

Interphase fluorescence in situ hybridization improves the detection of malignant cells in effusions from breast cancer patients.

In diagnostic evaluation of effusions, difficulties are encountered when atypical reactive mesothelial cells have to be differentiated from malignant cells. We tested the impact of fluorescence in situ hybridization (FISH) to identify metastatic cells in breast cancer effusions by detection of numerical chromosomal changes. Pleural and ascitic fluid samples (n=57) from 41 breast cancer patients were concomitantly evaluated by routine cytology and FISH, using centromere-specific probes representing chromosomes 7, 11, 12, 17 and 18. After setting stringent cut-off levels deduced from non-malignant control effusions (n=9), the rates of cells with true aneuploidy were determined in each effusion sample from breast cancer patients. The occurrence of aneuploid cells, as detected by FISH and indicative of malignancy, was correlated with the cytological findings. Routine cytology revealed malignancy in 60% of effusions. Using FISH, aneuploid cell populations could be observed in 94% of cytologically positive and in 48% of cytologically negative effusions, thus reverting diagnosis to malignancy. To confirm malignancy in cases with a low frequency of aneuploid cells, two-colour FISH was additionally performed and indeed showed heterogeneous chromosomal aneuploidy within single nuclei. We conclude that FISH is a valuable tool in the diagnosis of malignancy and may serve as an adjunct to routine cytological examination, as demonstrated here for breast cancer effusions.

Involvement of P-glycoprotein in the transmembrane transport of interleukin-2 (IL-2), IL-4, and interferon-gamma in normal human T lymphocytes.

The physiological role of the multidrug resistance P-glycoprotein (P-gp), which is expressed by normal human T lymphocytes, is still largely unknown. To investigate whether or not P-gp is involved in the transport of cytokines, peripheral blood lymphocytes were stimulated with phytohemagglutinin (PHA) in the absence or presence of P-gp inhibitors, and concentrations of cytokines (interleukin-2 [IL-2], IL-4, IL-6, interferon-gamma [IFN-gamma]) in the supernatants of these cultures were quantitated by enzyme-linked immunosorbent assay. P-gp inhibitors included verapamil (Ver), tamoxifen (Tmx), and the P-gp specific monoclonal antibody UIC2. Release of IL-2 was significantly suppressed by these inhibitors at concentrations that were also effective in blocking efflux of Rhodamine-123 from normal T lymphocytes. IL-2 mRNA expression in lymphocytes was not different between PHA control and the cultures with P-gp inhibitors. Ver and Tmx did not interfere with T-cell activation as determined by CD25 and CD69 expression. In a nonhematological model, the P-gp expressing HCT-8 adenocarcinoma cell line, exogenously added IL-2 was shown to exert an inhibitory effect on P-gp mediated Rhodamine-123 efflux. In addition, transepithelial transport of IL-2 by electrophysiologically tight and polarized HCT-8 monolayers was examined. A time-dependent flux of IL-2 across dense monolayers, which was partially inhibited by Ver, was observed. We also investigated whether or not P-gp inhibitors suppressed release of other cytokines produced by activated T cells (IL-4, IL-6, IFN-gamma). Release of IL-4 and IFN-gamma was significantly inhibited by Ver, Tmx, and UIC2; however, release of IL-6 remained unaffected. These data show P-gp mediated transmembrane flux of IL-2 in T lymphocytes and HCT-8 cells. We conclude that P-gp participates in the transport of cytokines (IL-2, IL-4, and IFN-gamma) in normal peripheral T lymphocytes.

Multiple myeloma: high incidence of chromosomal aneuploidy as detected by interphase fluorescence in situ hybridization.

Because metaphase cytogenetic studies in multiple myeloma (MM) are hampered by a low proliferative activity of myeloma cells in vitro, interphase cytogenetics by means of fluorescence in situ hybridization (FISH) should improve the detection of chromosomal abnormalities in MM. We therefore investigated chromosomal aneuploidy in 36 patients with MM using interphase FISH and alpha-satellite DNA probes for chromosomes 1, 3, 7, 8, 11, 12, 16, 17, 18, and X. By FISH, myeloma cells from 32 patients (88.9%) were aneuploid for at least one of the chromosomes examined. In 24 patients (66%), aberrations of > or = 3 chromosomes were observed. Aneuploidy was predominantly characterized by a gain of chromosome numbers, with involvement of chromosomes 3, 7, and 11 occurring in > 50% of patients. Loss of a centromeric signal suggesting monosomy was most frequently observed for chromosomes 17 (22.2% of patients) and X (monosomic in 42.3% of female patients, but loss of chromosome X was never observed in males, P < 0.05). Dual-color FISH studies provided evidence for marked heterogeneity of aneuploid cells in 8 patients (22.8%). Occurrence of chromosomal aneuploidy was independent of stage and pretreatment status. Gain of chromosome 3 was significantly correlated with an IgA paraprotein (P < 0.05). In 12 patients, the direct comparison of metaphase cytogenetics and FISH showed that FISH detected aneuploidy of chromosomes in 9 patients that was missed by metaphase analysis. In conclusion, interphase FISH, by which chromosomal aneuploidy was detected in almost 90% of patients with MM, represents an approach for evaluating the clinical significance of specific chromosomal abnormalities in MM.

Glucocorticoid-mediated immunomodulation: hydrocortisone enhances immunosuppressive endogenous retroviral protein (p15E) expression in mouse immune cells.

To define glucocorticoid (GC)-regulated genes contributing to the anti-inflammatory and immunosuppressive effects of GC, previous work from our laboratory revealed up-regulation of transcripts from endogenous type B mouse mammary tumour virus (Mtv) and type C murine leukaemia virus (Emv) loci by high dose GC treatment of P388D1 macrophage-like cells. This study demonstrates enhancement of expression from Mtv and Emv loci in P388D1 cells by more physiological hydrocortisone concentrations (1 microM), and shows direct transcriptional mode of regulation by blocking GC-mediated signal transduction at different levels. Furthermore, we found up-regulation of Emv mRNA steady-state levels in murine lymphoid lineage cells (T-like EL4 and BW5147 cells; B-like X63 cells) upon GC treatment. The Emv transcripts shown by us to be GC-up-regulated encode for the transmembrane envelope protein TM/p15E which is highly conserved in several retroviruses. TM/p15E and the p15E-like products found in humans exert immunosuppressive effects in different test systems. Thus, our findings raise the possibility that immunomodulation by GC might be mediated in part by enhanced expression of p15E(-like) products.

Dexverapamil as resistance modifier in acute myeloid leukaemia.

In order to evaluate dexverapamil as a resistance modifier in acute myeloid leukaemia, we have added dexverapamil (4 x 300 mg/d orally) to DA chemotherapy (daunorubicin, cytosine arabinoside) in six patients with acute myeloid leukaemia. Two patients (1 first and 1 second relapse) achieved complete remission and two patients (1 refractory disease, 1 third relapse) showed some improvement. One patient in first relapse died due to disease progression and one drug-refractory patient remained refractory. The peak plasma levels of dexverapamil and nordexverapamil ranged from about 600 to 4100 ng/ml and from 450 to 1130 ng/ml, respectively. Major sideeffects were hypotension and sinus bradycardia. These results show the need for further evaluation of dexverapamil as a resistance modifier in acute myeloid leukaemia.

P-glycoprotein expression as unfavorable prognostic factor in acute myeloid leukemia.

In order to further define the role of the MDR1 gene in acute myeloid leukemia (AML), we determined the association between the presence of P-glycoprotein on leukemic cells and the efficacy of therapy in patients with AML. Immunocytochemistry with monoclonal antibody C219 was performed to demonstrate the presence of P-glycoprotein. Positive staining ranged from 0 to 60% of the leukemic cells. For further analysis, patients were assigned into groups with 0-5% staining cells (group 1, n = 33) and with > 5% staining cells (group 2, n = 19). The complete remission rate of induction chemotherapy was 76% for group 1 but only 32% for group 2 (p = 0.002). The median duration of overall survival was 19 months for patients in group 1 as compared to 3 months for patients in group 2 (p = 0.007). The data indicate that P-glycoprotein expression is associated with an unfavorable prognosis in patients with AML.

MDR1 gene expression in myelodysplastic syndrome and in acute myeloid leukemia evolving from myelodysplastic syndrome.

The expression of the MDR1 gene, a multidrug resistance gene, was determined in patients (N = 24) with myelodysplastic syndrome or acute myeloid leukemia evolving from it. MDR1 RNA expression of the mononuclear cells was detected in 14 (58%) patients. Among patients who had progressed to acute myeloid leukemia (sAML) (N = 14), MDR1 RNA expression was seen in 8 (57%) patients. Expression was observed in the 2 patients who had been pretreated with MDR drugs. These findings indicate that the MDR1 gene is frequently expressed in MDS or sAML and suggest that multidrug resistance might be involved in the clinical drug resistance of these diseases.

MDR1 gene expression in chronic lymphocytic leukemia.

In order to assess the clinical role of the MDR1 gene in chronic lymphocytic leukemia (CLL), we determined its expression in the leukemic cells of 39 patients with CLL and compared this with other clinical and laboratory parameters. MDR1 RNA expression was detected in 29 patients. MDR1 RNA transcripts were independent of age, treatment status of the patients and the clinical stage of CLL, but correlated with the white blood cell count and MDR2 RNA transcripts. Expression of the tumor suppressor gene p53 was found in 30 out of 37 patients and was associated with MDR1 RNA expression (P < 0.001). Immunocytochemistry using the monoclonal antibody C219 was performed in 38 patients, and in 28 cases, more than 5% of the leukemic cells were found to express cell surface P-glycoprotein. P-glycoprotein expression correlated with the expression of MDR1 RNA (P = 0.048).

MDR1 RNA expression as a prognostic factor in acute myeloid leukemia: an update.

In order to confirm our initial report on the negative impact of MDR1 gene expression on the outcome of de novo acute myeloid leukemia (AML), we present an update of our prospective study with a larger number of patients and a longer duration of follow-up. At diagnosis, MDR1 RNA expression of the leukemic cells was negative in 37% and positive in 63% of the patients (N = 79). The complete remission rate of induction chemotherapy was 76% for MDR1 RNA negative and 54% for MDR1 RNA positive patients (p = 0.05). At a median observation duration of 33 months, the duration of overall survival was 19 months for the MDR1 RNA negative patients but only 8 months for the patients with MDR1 gene expression (p = 0.02). Thus the long-term data also indicate that MDR1 gene expression is an unfavourable prognostic factor in AML.

MDR1 gene expression in primary colorectal carcinomas.

The expression of the MDR1 gene, a multidrug resistance gene, was prospectively determined in 113 primary colorectal carcinoma specimens and correlated with clinical data including survival durations of the patients. MDR1 RNA was detected in 65% of the carcinomas. No expression of the MDR2 gene was seen, MDR1 gene expression was independent of age and sex of the patients, size and histologic grading of the tumour, lymph node involvement and distant metastasis. Kaplan-Meier analysis revealed that the durations of both relapse-free survival and overall survival were not different between patients with MDR1 RNA positive tumours and those with MDR1 RNA negative tumours.

MDR1 gene expression and its clinical relevance in primary gastric carcinomas.

BACKGROUND: Drug resistance remains a major problem in gastric carcinomas. To evaluate the mechanisms involved in this resistance, the authors determined the expression of the MDR1 gene, a multidrug resistance gene, in primary gastric carcinomas. METHODS: MDR1 RNA levels of gastric carcinoma specimens (n = 22) were determined by slot blot analysis. An MDR1 cDNA (probe 5A) was used for the hybridization. RESULTS: MDR1 RNA was detected in 41% of the gastric carcinomas, with high levels in 18% of the specimens. No expression of the MDR3 gene was observed in these tumors. MDR1 gene expression was independent of patient age, tumor localization, and lymph node involvement. However, MDR1 RNA expression was less frequent in locally advanced tumors and was absent in the primary tumors of all six patients who had distant metastases. CONCLUSIONS: The data indicate that multidrug-resistant cells are present in primary gastric carcinomas and suggest that multidrug resistance might contribute to the clinical drug resistance of these tumors.