T lymphomagenesis is determined by a dominant host gene thymic lymphoma susceptible mouse-1 (TLSM-1) in mouse models.

Susceptibility to T lymphomas in mice is determined by a number of viral and host genetic factors. We analyzed the types and latent period of lymphomas spontaneously occurring in crosses between AKR/Ms, a T lymphoma-prone mouse strain, and SL/Kh, a pre-B lymphoma-prone strain. The incidence of T lymphomas in the F1 hybrids backcross to SL/Kh as well as F2 generation mice indicated that a dominant host gene thymic lymphoma susceptible mouse-1 (Tlsm-1) of AKR/Ms determined the type of lymphomas to be thymic. Linkage analysis with microsatellite markers assigned Tlsm-1 to the map position 61 cM from centromere of the chromosome 7. Close scrutiny of this region of AKXD recombinant inbred strains for spontaneous T lymphomas revealed the presence of Tlsm-1-like gene most likely between D7MIT71 (map position 62) and D7MIT13 (map position 70). On the other hand, a SL/Kh-derived recessive allele at a major histocompatibility complex (MHC)-linked locus accelerated development of both T and B lymphomas.

Expression of leukemogenic recombinant viruses associated with a recessive gene in HRS/J mice.

HRS/J inbred mice carry a mutant autosomal recessive gene (hr), which in homozygotes coincides with susceptibility to spontaneous thymic leukemia. Unlike their heterozygote (hr/+) littermates, hr/hr homozygotes express high levels of xenotropic virus during the preleukemic period, and viruses with a broadened host range (termed polytropic viruses) can be isolated from their preleukemic and leukemic tissues. Because hr/hr and hr/+ mice are otherwise genetically identical, the virological differences between them support the role of polytropic viruses in the generation of thymic leukemia. In the present report we show that the HRS/J polytropic viruses are env gene recombinants with unique oligonucleotide and peptide maps. These polytropic viruses appear to arise by recombination between ecotropic virus and an unidentified genome related, but not identical to, the endogenous xenotropic viruses. Moreover, polytropic viruses not only accelerate leukemogenesis in HRS/J mice, but also induce thymic leukemia in the low leukemia strain CBA/J. By contrast, cloned ecotropic and xenotropic viruses have no leukemogenic action.

Autoimmune dilated cardiomyopathy in PD-1 receptor-deficient mice.

Dilated cardiomyopathy is a severe pathology of the heart with poorly understood etiology. Disruption of the gene encoding the negative immunoregulatory receptor PD-1 in BALB/c mice, but not in BALB/c RAG-2-/- mice, caused dilated cardiomyopathy with severely impaired contraction and sudden death by congestive heart failure. Affected hearts showed diffuse deposition of immunoglobulin G (IgG) on the surface of cardiomyocytes. All of the affected PD-1-/- mice exhibited high-titer circulating IgG autoantibodies reactive to a 33-kilodalton protein expressed specifically on the surface of cardiomyocytes. These results indicate that PD-1 may be an important factor contributing to the prevention of autoimmune diseases.

Overexpression of a transcription factor LYL1 induces T- and B-cell lymphoma in mice.

LYL1, a member of the class II basic helix-loop-helix transcription factors, is aberrantly expressed in a fraction of human T-cell acute lymphoblastic leukemia. Here, we generated transgenic mice ubiquitously overexpressing LYL1 using a construct expressing full-length cDNA driven by a human elongation factor 1alpha promoter. Four independent lines exhibiting high LYL1 expression were established. Of these transgenic mice, 96% displayed loss of hair with a short kinked tail. Furthermore, 30% of them developed malignant lymphoma, with an average latent period of 352 days. In these mice, histological examination revealed tumor cell infiltration in multiple organs and immunohistochemical analysis showed that the infiltrated tumor cells were either CD3 or CD45R/B220-positive; fluorescence-activated cell sorter analysis indicated that each tumor consisted either of mainly CD4, CD8 double-positive T cells or mature B cells; the clonality of LYL1-induced lymphoma was confirmed by T-cell receptor rearrangement and immunoglobulin heavy-chain gene rearrangement analyses. Mammalian two-hybrid analysis and luciferase assay suggested that excess LYL1 blocked the dimerization of E2A and thus inhibited the regulatory activity of E2A on the CD4 promoter. Reverse transcription-polymerase chain reaction results showed that the expression of certain E2A/HEB target genes was downregulated. Taken together, our results provide direct evidence that aberrant expression of LYL1 plays a role in lymphomagenesis.

Developmentally regulated expression of a human "finger"-containing gene encoded by the 5' half of the ret transforming gene.

We isolated and sequenced a cDNA clone of the human gene encoded by the 5' half of the ret transforming gene. The nucleotide sequence indicates that it encodes a protein with "finger" structures which represent putative metal- and nucleic acid-binding domains. Transcription of this gene was detected at high levels in a variety of human and rodent tumor cell lines, mouse testis, and embryos. In addition, a unique transcript was observed in testis RNA. When the expression of the unique transcript was examined at different stages of spermatogenesis, a striking increase in mRNA levels accompanied progression from meiotic prophase pachytene spermatocytes to postmeiotic round spermatids. This finger-containing gene may thus function in male germ cell development.

Temperature-dependent rosette formation by mouse lymphoma cells as a result of viral hemadsorption.

Cells from several mouse lymphomas formed rosettes with nonsensitized foreign erythrocytes through C-type virus particles clustered on the cell surface in serum-free medium held at 4 degrees C. This type of rosetting was found most typically in a lymphoma induced by Rauscher leukemia virus in tissue culture (RD-12), but it also occurred in 23 of 61 spontaneous thymic lymphomas in AKR mice. Chemically or X-ray-induced leukemias and spontaneous reticulum cell sarcomas did not form rosettes. The nature of the rosette formation may be interpreted as viral hemadsorption, with a possible relationship to hemagglutination by murine leukemia viruses. The receptor on virus particles was trypsin sensitive and showed high affinity to serum inhibitors (RIF). Serum rosette-inhibiting activity was assessed by a quantitative rosette inhibition test; rosette inhibition proved widely distributed among species. Physicochemical properties of serum RIF and their function both in vivo and in vitro were described. Rosette formation with similar temperature requirements, previously reported in a mouse lymphoma carrying membrane-bound heterophile cold hemagglutinin, was readily distinguished from viral hemadsorption by its insensitivity to mouse serum RIF.

Growth stimulation of microenvironment-dependent mouse leukemias by tumor-promoting phorbol esters.

The effects of 12-O-tetradecanoylphorbol 13-acetate (TPA) and related tumor-promoting phorbol esters on 23 symbiotic cell lines from AKR thymic leukemias were studied. The phorbol derivatives with in vivo cocarcinogenic activity increased the viable yield of leukemia cells of most symbiotic cell lines in the absence of growth-supporting adherent cells. Particularly, in 10 cell lines remarkable growth stimulation was observed with 1 to 10 ng TPA/ml. The effect of TPA was reversible, since the leukemia cells exposed to TPA for 48 hours failed to grow without TPA or adherent cells. The leukemia cells from a symbiotic cell line could be maintained in TPA-containing medium for as long as 8 weeks without losing dependence. The leukemia cells were refractory to growth stimulation by TPA after they acquired the capability to grow independently. Furthermore, by affecting the leukemia cells, TPA and active tumor promoters inhibited pseudoemperipolesis (i.e., localization of leukemia cells under or between adherent cells), which is the basic cell interaction in symbiotic complexes. These effects of tumor-promoting phorbol esters are discussed with a special reference to the multistage hypothesis of leukemogenesis.

Epigenetic control of endogenous ecotropic virus expression in SL/Ni mice.

SL/Ni mice were found to be highly polymorphic in the expression of endogenous ecotropic virus (ETV). By selective mating of the mice with either high-virus or virus-free phenotypes, the following stably virus-positive and virus-negative sublines were obtained: SL/Ni-Eco+ and SL/Ni-Eco-, respectively. This polymorphism was produced by an epigenetic factor transmitted by SL/Ni-Eco- female mice via milk. F1 hybrids between SL/Ni-Eco- females with males of high-virus strains did not express ETV, whereas reciprocal F1 hybrids did. On the other hand, F1 mice between females of low-virus strains or of NFS mice lacking the ETV proviral genome and SL/Ni-Eco- males expressed a high level of ETV. Foster-nursing of newborn mice of high-virus strains by SL/Ni-Eco- foster-mothers or injection of pooled sera of SL/Ni-Eco mice resulted in intense inhibition of virus expression. On the contrary, nursing of SL/Ni-Eco- newborns by NFS/N foster-mothers resulted in high virus expression. These observations strongly support the hypothesis that failure to express ETV by SL/Ni-Eco- mice is due to a milk-transmitted maternal resistance factor, but probably not due to genetic heterogeneity among SL/Ni sublines. This factor caused strong, long-lasting, and selective suppression of endogenous ETV, but it did not confer resistance to exogenous infection of ETV. This activity was present also in the sera of SL/Ni-Eco- mice, since neonatal injection of the sera into high-virus strains of mice, SL/Ni-Eco+, SL/Kh, and AKR/Ms, caused strong selective suppression of ETV expression. By this procedure, the spontaneous occurrence of nonthymic lymphomas in SL/Kh mice was suppressed.

Mouse lymphoid leukemias: symbiotic complexes of neoplastic lymphocytes and their microenvironments.

Of 17 primary lymphoid leukemias of the mouse, 15 symbiotic cell lines were isolated by the explantation of leukemia tissues from which free leukemia cells had been mechanically removed. In vitro survival and growth of symbiotic leukemia cells depended on close association with the adherent cells from the initial explants or other sources. Pseudo-emperipolesis was a remarkable morphologic manifestation of symbiosis common to all cell lines, i.e., the leukemia cells were beneath the adherent cells in close contact. Cell interaction in symbiotic leukemias was studied with a representative symbiotic leukemia AKRL-3 and a cell line B6TE-A from normal thymic epithelium. Failure of the culture supernatant of the adherent cells to support the growth of leukemia cells indicated that the function of the adherent cells was mediated by close cell contact. During the culture, many symbiotic cell lines changed growth patterns and eventually grew independently. Consistent isolation of symbiotic cell lines from most primary leukemias, as well as consideration of the role of the thymus in leukemogenesis, may indicate that the lymphoid leukemias are basically symbiotic complexes of neoplastic lymphocytes and their microenvironments in their natural history. Similar lymphoepithelial cell complexes were isolated recently from normal murine thymus.

Formation of symbiotic complex by microenvironment-dependent mouse leukemias and thymic epithelial reticular cells.

Developing thymic leukemias of the mouse have been assumed to form symbiotic complexes with thymic microenvironments. This symbiosis is morphologically based on pseudoemperipolesis (PEMP). The mechanism of the association of microenvironment-dependent leukemia cells with thymic epithelial reticular cells (TER) was analyzed in vitro by scanning electron microscopy, microcinematography, and a quantitative assessment of PEMP. PEMP was a consequence of active locomotion of the leukemia cells, with TER passively accepting the leukemia cells "crawling" under their cytoplasm. The integrity of the cytoskeletal system of both cells was essentially required for PEMP, since cytochalasins and colchicine were highly inhibitory to PEMP. The mechanism of action of these compounds was probably dual: inhibition of the locomotive movements of the leukemia cells. A similar inhibition of PEMP was also observed with the tumor promoter 12-O-tetradecanoylphorbol 13-acetate.

Genetically determined susceptibility of Fischer 344 rats to propylnitrosourea-induced thymic lymphomas.

Administration of propylnitrosourea p.o. by our protocol induced a high incidence of hematolymphatic neoplasms in all six rat strains studied. Remarkable strain differences in susceptibility to thymic lymphomas were observed. The incidence of thymic lymphomas was high in Fischer 344 (98%) and Wistar/Furth (71%) but low in Sprague-Dawley (29%), ACI/Ms (23%), Donryu (24%), and Long-Evans (10%) strains. Segregation of thymic lymphoma incidence among crosses between highly susceptible Fischer and poorly susceptible Long-Evans rats indicated that the increased susceptibility to thymic lymphomas of Fischer rats was determined by a dominant gene TIs-1 (thymic lymphoma susceptible) and that this gene was linked to the coat color loci, p and c, in Linkage Group I in the order of TIs-1 - c - p. The presence of another independently assorting dominant gene, TIs-2, was also suggested to accelerate the thymic lymphoma-genesis. Expression of the group-specific antigen of murine leukemia virus as well as infectious viruses was not detected in nine propylnitrosourea-induced thymic lymphomas of Fischer rats.

Differentiation-associated cellular complex formation of murine thymocytes with thymic stromal cells.

Properties of normal murine thymocytes forming in vitro cellular complexes with thymic epithelial-like stromal cells in the form of pseudoemperipolesis were studied. The complex-forming cells were low-buoyant-density blasts primarily localized to the subcapsular zone. After transition into small cortical lymphocytes, their capacity for complex formation was lost. The complex-forming cells were relatively resistant to cortisone acetate and low-dose (170 rads) whole-body X-irradiation. Their number increased sharply in the early stage of thymic regeneration, corresponding to an increase in the percentage of large thymic lymphocytes 4 to 5 days after cortisone treatment or X-irradiation. However, after the thymus was repopulated with small lymphocytes, the percentage of complex-forming lymphocytes decreased rapidly to the normal level. A possible relationship between a step in thymic leukemogenesis and intrathymic T-cell differentiation is discussed.

Genetic resistance to urethan-induced pulmonary adenomas in SMXA recombinant inbred mouse strains.

Development of pulmonary adenomas (PAs) in mice is under the genetic control of multiple host genes. We have established a new set of SMXA recombinant inbred strains from PA-susceptible A/J and PA-resistant SM/J mice. The number of urethan-induced PAs was variable among substrains of the SMXA recombinant inbred strains, indicating the involvement of multiple genes. SMXA24 mice were highly resistant to PA, although they had susceptible alleles at all four known susceptibility genes, including kras2 and MHC. To identify the resistance gene in SMXA24, progeny of reciprocal F1 crosses and progeny of backcrosses to A/J were given urethan at 4 weeks of age and examined for induced PA at the age of 5 months. In reciprocal F1 cross progeny, the incidence of PA was very low, indicating that the resistance was a semidominant trait. Quantitative trait analysis of the backcross generation revealed significant linkages to loci on chromosome 12 (logarithm of odds score, 6.47) and chromosome 11 (logarithm of odds score, 4.35). To date, two PA resistance (PAR) genes, Par1 (located on chromosome 11) and Par2 (located on chromosome 18), have been reported. From the map position, one of the resistance genes on chromosome 11 was indistinguishable from Par1. However, another resistance gene on chromosome 12 was new, and we named this gene Par3. A likely candidate gene for Par3 is nPKCn, which is expressed exclusively in skin and lung and is down-regulated in PA. Par1 and Par3 seemed to act synergistically.

Two dominant host resistance genes to pre-B lymphoma in wild-derived inbred mouse strain MSM/Ms.

To explore possible host genes suppressing spontaneous B-lymphomagenesis in the mouse, expression of ecotropic murine leukemia virus (E-MuLV) and lymphoma development were observed in crosses between the pre-B lymphoma-prone SL/Kh and low-lymphoma strains of mice. E-MuLV expression was intensely inhibited in F1 hybrids with the strains either with the Fv-1b allele (BALB/C, C57BL/10, and A/J) or with the Fv-1nr allele (NZB). In these F1 mice, no lymphoma developed by 18 months of age. On the other hand, F1 hybrids with the strains with the Fv-1n allele [C3H/He, CBA/N, SJL, DBA/2, and MSM/Ms (hereafter referred to as MSM)], high or intermediate levels of E-MuLV expression were observed. Lymphoma incidence in these F1 hybrids, however, was low. This observation suggests the presence of non-Fv-1 dominant resistance genes in these strains. In an attempt to characterize such host genes, we analyzed crosses between SL/Kh mice and a wild mouse-derived inbred strain, MSM/Ms. The latter was susceptible to N-tropic virus expression, but (SL/Kh x MSM)F1 hybrids, did not develop and lymphomas. Of 60 SL/Kh x (SL/Kh x MSM)F1 hybrids, 14 B-lineage lymphomas, including 13 pre-B and 1 follicular center cell lymphoma, developed by 18 months of age. This was compatible with the hypothesis of two independently segregating dominant genes of MSM suppressing lymphomagenesis. By scanning all chromosomes for linkage of lymphoma susceptibility with polymorphic microsatellite loci, one significant linkage disequilibrium was found in the proximal segment of chromosome 17, containing D17MIT44 (map position 15.0) to D17MIT150 (position 33.3), and another linkage disequilibrium, in the midproximal segment of chromosome 18, containing D18MIT90 (map position 28.0) and D18MIT140 (37.0). All 13 pre-B lymphoma-bearing backcross mice were homozygous for SL/Kh-derived alleles at these loci. We named the gene on chromosome 17 Msmr1 (for MSM resistance 1) and that on chromosome 18 Msmr 2 (for MSM resistance 2).

Bone marrow pre-B-1 (Bomb1): a quantitative trait locus inducing bone marrow pre-B-cell expansion in lymphoma-prone SL/Kh mice.

Abnormalities of regulatory genes in early B-cell development often lead to lymphomagenesis. Our previous study showed that there is an abnormal transient expansion of bone marrow (BM) pre-B cells in lymphoma-prone SL/Kh strain mice. Such expansion is a genetic property of SL/Kh stem cells rather than BM microenvironments. Using the percentage of BP1+ B220+ pre-B cells in total BM lymphoid cells as a quantitative parameter, we studied the genetic control of BM pre-B cells in 159 F2 offspring of crosses between SL/Kh and NFS/N mice and 334 back-crosses to SL/Kh mice. A highly significant quantitative trait locus was identified on the distal segment of chromosome 3, showing logarithm of odds scores of 22.7 in the F2 cohort and 10.7 in back-cross mice. This quantitative trait locus, named bone marrow pre-B-1, colocalized with lymphoid enhancer factor-1, which encodes a high mobility group DNA-binding protein that is expressed in T and pre-B cells.

Abnormal bone marrow B-cell differentiation in pre-B lymphoma-prone SL/Kh mice.

SL/Kh mice spontaneously develop pre-B lymphomas with surface phenotypes of B220+, BP-1+, Thy-1-, and surface immunoglobulin negative. The immunoglobulin heavy chain of lymphoma is clonally rearranged but the light chain gene remains in germline configuration. Studying prelymphoma stage SL/Kh bone marrow (BM), we found unusual multiclonal expansion of BP-1+ pre-B cells [34.8 +/- 5.8% (mean +/- SD)] by 4 weeks of age, whereas there were far fewer of such cells in most other laboratory strains (8 +/- 5%). The BP-1+ cells did not express surface immunoglobulin, Thy-1.1, or c-kit. Therefore, they seemed to belong to the pre-B II category. Increased numbers of BP-1+ cells were seen in F1 hybrids between SL/Kh and NFS/N; thus it was apparently a dominant heritable property of SL/Kh mice. Emergence of this population was independent of expression of endogenous ecotropic virus, since they were present in BMs of the F1 hybrid to C4W (Fv-4') and were not inhibited by neonatal injection of maternal resistance factor. In the radiation chimeras SL/Kh-->BALB/c, BP-1+ cells appeared abundantly (29.0 +/- 3.8%), whereas in the reciprocal chimeras BALB/c-->SL/Kh, for fewer (5.5 +/- 2.3%) appeared. Therefore, expansion of BP-1+ cells in prelymphomatous BM is a property of SL/Kh stem cells rather than BM microenvironments.

Genetic predisposition to pre-B lymphomas in SL/Kh strain mice.

Genetic predisposition of SL/Kh mice to spontaneous pre-B lymphomas was investigated in crosses between SL/Kh and NFS/N, another inbred strain of mice lacking endogenous ecotropic provirus and spontaneous lymphoma. (SL/Kh x NFS/N) F1 hybrids developed lymphomas similar to those in SL/Kh but at a lower frequency and with a longer latent period. Of 83 backcross mice to NFS/N, 22 developed hemopoietic tumors: 8 were diffuse lymphoblastic lymphomas; 2 were myeloid leukemias arising by 12 months of age; and 12 were follicular center cell lymphomas found later in life. Of 6 endogenous ecotropic proviruses in SL/Kh, 2 were expressed in (SL/Kh x NFS) F1 backcrossed to NFS. One, encoded by a 27-kilobase EcoRI fragment, was closely linked to Gpi-1a on chromosome 7 and its expression seemed to be a prerequisite for the occurrence of all types of hemopoietic tumors. Microsatellite analysis of the backcross generation revealed multiple host genetic factors determining susceptibility to tumors. An allele derived from SL/Kh, mapped in the major histocompatibility locus on chromosome 17, was essential for development of early onset tumors. This locus was designated as Esl-1 (early lymphoma of SL-1). On the other hand, follicular center cell lymphomas developed mostly in the backcross mice homozygous for the NFS/N derived allele at the D4MIT17-linked locus, designated as foc-1 (follicular center cell lymphoma-1), on chromosome 4.

Tumor promoter-dependent mouse leukemia cell line.

From a spontaneous AKR/Ms thymic leukemia symbiotically cultured with thymic epithelial reticular cells, a tumor promoter-dependent cell line A65T was established by passaging the cells in medium containing 12-O-tetradecanoylphorbol-13-acetate (10 ng/ml). The in vitro growth of A65T was strictly dependent on the presence of active tumor promoters. Their action was reversible, since withdrawal of 12-O-tetradecanoylphorbol-13-acetate resulted in rapid decrease in viability of the cells. Three classes of chemically unrelated compounds sharing tumor-promoting activity in mouse skin could support the in vitro growth of A65T: plant diterpene esters; indole alkaloids; and polyacetates. Their growth effect on A65T cells quantitatively correlated well with the tumor-promoting activity in mouse skin. However, other growth stimulators of epidermal cells such as cholera toxin and epidermal growth factor failed to support the growth of A65T. It is suggested that lymphokines such as interleukin-2 and interleukin-3 were not responsible for 12-O-tetradecanoylphorbol-13-acetate-stimulated growth of A65T because concanavalin A-stimulated spleen cell-conditioned medium containing both interleukin-2 and interleukin-3 activities as well as WEHI-3 cell culture supernatant containing potent interleukin-3 activity did not stimulate the proliferation of A65T cells. Furthermore, 12-O-tetradecanoylphorbol-13-acetate did not induce production of any significant amount of either activity in A65T cells. This cell line is useful for the screening of tumor promoters in environments although, so far, all the compounds capable of stimulating A65T growth have been limited to those competing with phorbol esters for the cellular receptor. Also, the cell line provides a potential model for analyzing growth requirements of developing mouse thymic leukemias.

Expression of metastasis-related nm23-H1 and nm23-H2 genes in ovarian carcinomas: correlation with clinicopathology, EGFR, c-erbB-2, and c-erbB-3 genes, and sex steroid receptor expression.

To verity the role of metastasis-related nm23 genes in carcinogenesis and progression of ovarian carcinoma, we analyzed the mRNA levels of the nm23 genes of both isoforms, -H1 and -H2, together with those of the epidermal growth factor receptor, the c-erbB-2, and the c-erbB-3 genes in 45 ovarian carcinomas and 5 benign cystadenomas. Expressions of nm23 gene products/nucleoside diphosphate kinases, epidermal growth factor receptor, erbB-2 protein, and sex steroid receptor status in ovarian carcinomas were also examined by immunohistochemistry. The mRNA levels of nm23-H1 and nm23-H2 were higher in carcinoma tissues compared with benign tumors (H1, P < 0.01). The mRNA levels of c-erbB-2 and c-erbB-3 were also elevated in carcinoma tissues, and there was a positive correlation between mRNA levels of the nm23-H1 and the c-erbB-2 genes (r = 0.58; P < 0.05). Correlation of immunohistochemical staining between nucleoside diphosphate kinases and erbB-2 protein was also observed in ovarian carcinoma tissues. Sex steroid receptor positivity was related to a higher expression of nucleoside diphosphate kinases. Expression levels of the nm23 genes in ovarian carcinomas were not related to either histological subtype or local extension and peritoneal dissemination. Among stage III ovarian carcinomas, however, tumors possessing lymph node metastasis showed significantly lower nm23-H1 mRNA levels than those without nodal involvement (P < 0.05). Stage IV carcinomas also exhibited lower nm23-H1 and nm23-H2 expression levels compared with other stages (P < 0.05). These results suggest that expression of the nm23 genes, especially nm23-H1, is activated, accompanied by c-erbB-2 and c-erbB-3 overexpressions, in early stages of the carcinogenic process of ovarian carcinoma and reduction of nm23-H1 expression occurs in association with lymph nodal and/or distant metastasis.

Expression of multidrug resistance gene and localization of P-glycoprotein in human primary ovarian cancer.

Resistance to chemotherapy is the major obstacle to controlling malignant tumors. To characterize multidrug resistance phenotype in human primary ovarian cancer without chemotherapy, expressions of the mdr1 gene in 52 cases of ovarian cancer (44 common epithelial, 5 nonepithelial, and 3 metastatic cancers) were analyzed by polymerase chain reaction of RNA after reverse transcription. Furthermore, localization of P-glycoprotein, which is encoded by the mdr1 gene, was studied immunohistochemically. Although overall expression of the mdr1 gene was relatively low, its expression level was the highest in well-differentiated cancer tissues. Serous and mucinous adenocarcinomas showed higher levels of expression compared with clear cell and endometrioid carcinomas. P-glycoprotein was positive on luminal surfaces of lining cells of ovarian cancer and on those of inclusion cysts from which epithelial ovarian cancer is considered to develop. Thus, some ovarian cancer cases before chemotherapy are intrinsically multidrug resistant, which can be determined by mdr1 gene expression, and this phenotype should be taken into account for effective chemotherapy of ovarian epithelial carcinomas.