Gene Literature Dashboard

Estrogen receptor (ER) analysis in breast cancer has been used in three clinical situations: to select patients with advanced breast cancer for hormonal therapy, as a prognostic parameter, and for selection of women with early breast cancer to adjuvant hormonal treatment. ER has traditionally been measured using labelled hormone in binding assays--often in dextran-coated charcoal assays (DCC). Monoclonal antibodies to ER has permitted development of a solid phase enzyme immunoassay (ER-EIA) used for quantitative determination of ER in tissue homogenates, and have also been used for determination of ER using an immunohistochemical assay in frozen sections (ER-ICA) or in formalin-fixed, paraffin-embedded tissue (ER-PAR). A large number of studies has compared ER-EIA with ER-DCC assays. There is a good linear correlation between the two types of assay but ER-EIA measure more ER and classify a larger fraction of tumors ER-positive than conventional ER assays. Lack of clinical data makes the significance of this uncertain. Numerous studies have reported on the correlation between ER-ICA and ER-DCC or ER-EIA. There is a good correlation among the assays on classification of ER status with a median 86% concordance, but a somewhat poorer correlation between semiquantified ER of immunohistochemical assays and ER determined by the quantitative methods (median coefficient of correlation 0.67). There is a large variation in the cut-off level for definition of ER-positive in immunohistochemical assays emphasizing the need for quality control studies. The major problem involved in ER analysis in paraffin-embedded tissue is a considerable loss of immunoreactivity compared to sections from frozen tissue. This can partly be overcome by modifications of the immunohistochemical technique using enzyme pretreatment and other amplification systems, but the sensitivity of ER-PAR remains lower than ER-ICA despite these modifications, and the ER status is less reliably determined in tumors with low ER contents (< 100 fmol). The prognostic value of ER-PAR was evaluated with a multivariate analysis. The endpoint was disease-free interval in systemically untreated patients with early breast cancer, and the variables used were: ER-DCC, ER-PAR, age, tumor size, tumor grade, and nodal status. A total of 133 patients from the Danish Breast Cancer Cooperative Group's (DBCG) 77c protocols had a complete set of variables. The analysis showed that only nodal status, ER-DCC, and tumor grade were significant and independent prognostic variables. An overview of larger multivariate studies on mainly node-negative patients failed to show independent prognostic significance of ER-DCC.(ABSTRACT TRUNCATED AT 400 WORDS)

In 9 dogs with severe hemorrhagic shock, the effect of hemocorrector "Lactoprotein" on the indices of coagulative hemostasis and rheologic blood properties was studied. It was established that in intravenous transfusion of lactoprotein at a dose of 10 ml/kg permitting to lead the animals out from the state of shock, the syndromes of disseminated intravascular coagulation and high viscosity of the blood were cupped off. However, by the end of 48 hours of observation, fibrinogen level in the blood of the animals increased sharply, while antithrombin-III concentration and hematocrit decreased. Together with recommendation to use lactoprotein in the complex of shock therapy at all the stages of medical evacuation, the conclusion about necessity to perform at the early postshock period the differential component hemo- and anticoagulative therapy has been made.

Carcinogenesis in human large intestine is a result of multiple, heterogeneous and random genetic changes. Deletion of tumor suppressor genes and activation of oncogenes appear to be important molecular events. These compromise the loss of chromosomes 5, 17, 18 or functional inactivation of FAP, p53 and DCC genes. Activation of Ki-ras and c-myc oncogenes seems to be crucial for both cell immortalization and morphology modification. Identification of genes involved in this process enables both a screening and a new classification. Also it is an important step towards a gene therapy.

Myelodysplastic syndromes (MDS) include hemopoietic cytopenias of different origin, which are usually refractory to treatment. Therefore MDS patients should generally be treated conservatively. Transfusions of packed red cells (given in a strict regimen to minimize the risk for secondary hemochromatosis) may be sufficient to maintain a good quality of life. Indications for cytotoxic treatment include signs of progression of the disease. In patients with symptomatic cytopenias low-dose cytarabine (ara-C) should be tried. It is essential then to monitor each patient individually and to avoid fixed treatment schedules. Standard (high-dose) chemotherapy in MDS, is associated with a high mortality and a low response rate, and should be considered only in younger patients with advanced MDS. Allogeneic bone marrow transplantation (BMT) may be offered to younger MDS patients, when a suitable donor is available. Treatment with differentiation inducers has not met with expectations and should not be used outside clinical trials at the present. The use of recombinant hemopoietic growth factors (GF) seems promising. GF, like GM-CSF, G-CSF, IL-3, and erythropoietin, can be used either alone or in combinations, to support failing peripheral blood values, and decrease the risk for lethal complications. GF can also be given together with chemotherapy, in an effort to make the leukemic clonogenic cells more susceptible to cytotoxic drugs. Other treatments for MDS include: IFN-alpha and etoposide, with responses primarily in chronic myelomonocytic leukemia; hem arginate, whose role is still not clear; and corticosteroids, but only in carefully selected cases.

Retinoblastoma (RB) and the familial adenomatous polyposis/colorectal cancer (FAP/CRC) complex provide well-characterised examples of multistage carcinogenesis and inheritance of a predisposition to cancer. Retinoblastoma appears to conform to the simple two-step model first proposed by Knudson. The gene responsible for RB, now called Rb1, has been located in chromosome region 13q14. The Rb1 gene has been cloned and subjected to extensive analysis. It is probable that the Rb1 gene product has a role in the regulation of transcription. The familial form of RB occurs as the result of a germline mutation of one of the copies of the Rb1 gene. Colorectal cancer, in contrast, appears to be the result of four or five steps involving both activation of oncogenes and inactivation of antioncogenes. The FAP gene has been located in chromosome region 5q21 by genetic linkage, and a candidate gene, MCC (mutated in colon cancer), has been cloned. Other mutations in previously-identified genes that have been identified as important in the genesis of CRC include the activation of p53 and of Ki-ras. A gene lying in chromosome region 18q which is deleted in colorectal cancer, and hence named DCC has been cloned. Its protein product has sequence homology to neural cell adhesion molecules and other related cell-surface glycoproteins. Delineation of the genes involved in the development of tumours such as RB and CRC provides insight into the mechanisms by which sequential mutations result in carcinogenesis.

It is apparent that multiple genetic events occur in the development and progression of breast cancer. From the limited data available, no consistent temporal pattern of mutational events is required. This conclusion is consistent with data in colorectal carcinoma, where the number of mutational events, and not the order, appears to be relevant. Several authors have questioned whether the multiple mutational events occur independently or whether significant associations were evident. Cropp et al. postulated that two sets of mutational events occurred simultaneously in a higher degree of breast tumors than expected based on chance: Set 1 consisted of deletions on 11p, 17p, 18q, and int-2 and myc amplifications; set 2 consisted of 17q, 1p, and 3p deletions. Sato et al., analyzing another tumor cohort for simultaneous mutations, noted a correlation of 17p and 16q deletions, 13q and 17p deletions, and 17p deletion with erbB-2 amplification. Clearly, concordant data on this issue will require the use of large breast tumor cohorts for a comprehensive set of probes. The reasons why mutations to specific genes on different chromosomes tend to occur coordinately is unknown, but may involve common flanking and/or intron sequences at high risk for certain types of mutational events. Another interesting question is the degree to which alterations, but not homozygous inactivation, of suppressor genes occur and its phenotypic consequences. In this chapter, evidence was presented for the amplification of a DCC allele in breast cancer and for variable RB protein expression in breast tumors as a consequence of allelic deletion. For many of the metastasis suppressor genes, a simple reduction in their expression, or an alteration in their expression over the normal cellular regulatory controls, may be sufficient to fuel the metastatic process. The data suggest a more complex regulation of the cancer phenotype by suppressor genes than by recessive inactivation alone. Why do many sporadic cancers, including breast cancer, appear to require alterations to multiple suppressor genes, as compared to diseases such as retinoblastoma, where a single suppressor gene appears to control the cancer phenotype? The answer to this question is unknown, but most theories are based on the hypothesis that suppressor genes act to control cellular responses to either other cells or signals in the microenvironment. In retinoblastoma all cells can carry a germ-line mutation. Cells carrying the RB mutation can interact with both the embryonic and differentiated microenvironments; the specific interaction of mutated cells with the embryonic retinal microenvironment may trigger the onset of retinoblastoma.(ABSTRACT TRUNCATED AT 400 WORDS)

The patterns of expression of keratins K1 and K8, filaggrin, and the proliferation-associated protein, proliferating cell nuclear antigen (PCNA), were studied in normal and human papillomavirus (HPV) 16 or 18 immortalized keratinocyte cell lines grown on organotypic raft cultures. Normal keratinocytes produced an epithelial sheet that closely resembled epidermis in vivo, characterized by lack of K8 expression, PCNA expression restricted to the basal layer, and K1 and filaggrin expression in the suprabasal layers. Although morphologically abnormal in many respects, some HPV-immortalized cell lines produced cornified epithelial layers and approximated the normals in their patterns of expression of keratins and filaggrin. Other HPV-immortalized cell lines produced poorly differentiated epithelial layers that were characterized by loss of filaggrin expression, and the single tumorigenic cell line, 18-11, was distinguished by abundant K8 expression. All of the HPV-immortalized cell lines were distinguished from normal keratinocytes by a common pattern of full-thickness PCNA expression in the epithelial layers they produced, suggesting that maintenance of the proliferative state may be an important contribution made by HPV 16 or 18 sequences toward the production of a malignant phenotype.

The steroid receptor (estrogen (ER) and progesterone receptor (PR] status was studied in 94 cases of invasive breast carcinoma from three separate institutions. All cases had fresh tissue examined for ER and PR by the dextran-coated charcoal cytosolic assay (DCC), and each case was examined immunohistochemically for ER and PR from archival formalin-fixed, paraffin-embedded tissue. Immunohistochemical assays (IH) were reviewed blinded to the DCC results and scored in a semiquantitative fashion prior to comparison to the DCC results. Overall, there was agreement between DCC and IH in 89% of ER and in 87% of PR assays. Some 50% of the ER discorrelations were of the IH-positive DCC-negative type, while 27% of the PR discorrelations were of this type. In four cases, both ER and PR did not correlate between IH and DCC determinations, with two being IH (ER and PR) positive and DCC negative, and two of the opposite type. The results of the study show that steroid receptor assays performed on routinely processed formalin-fixed archival material are reliable and closely recapitulate the results of traditional biochemical assays. Results suggest that, in the cases where IH is positive while DCC is negative, the IH result may actually provide a more reliable receptor status of the tumor than does the DCC result. Semiquantitation of fixed tissue IH assays shows a trend toward quantitative correlation with DCC results, but this correlation is weak, and factors concerning fixation and processing are most likely to be responsible.(ABSTRACT TRUNCATED AT 250 WORDS)

Immunocytochemical assay (ICA) of the progesterone receptor (PgR) was performed on 152 patients with stage I-II breast cancer. We employed the rat monoclonal antibody KD-68 and a peroxidase/antiperoxidase displaying system. The results obtained by ICA (Pg(RICA)) were compared with those by the biochemical dextran-coated charcoal assay (PgRDCC). Comparing the two methods we found an overall agreement (accuracy) of 77.5%, a PgR(ICA) sensitivity of 83.5% and a specificity of 73%. Both methods were significantly associated with oestrogen receptor expression, detected by DCC (P less than 0.001 for PgRDCC and P = 0.0014 for PgR(ICA)). No significant association was found between PgR(ICA) or PgRDCC and the other clinicopathological features analysed. After a median follow-up of 36 months, the overall survival probability was 91% in PgRDCC-positive versus 81.5% in PgRDCC-negative patients (log-rank test, chi 2 = 0.91) compared to 87.5% in PgR(ICA)-positive versus 82% in PgR(ICA)-negative ones (log-rank test, chi 2 = 0.93). Disease-free survival probability was 74.5% in both PgRDCC-positive and PgRDCC-negative patients (log-rank test, chi 2 = 0.02) compared to 78% in PgR(ICA)-positive versus 71.5% in PgR(ICA)-negative cases (log-rank test, chi 2 = 0.37). The present study demonstrates that ICA is a reliable method to detect PgR, correlating well with the DCC assay. Moreover, the ICA assay seems to provide clinical information complementary to the biochemical method. The definition of its prognostic value in operable breast cancer needs additional studies, particularly in node-negative patients.

A human mammary carcinoma originating from a postmenopausal patient was successfully transplanted into nude mice. According to the adopted criteria the tumour proved to be oestradiol- and progesterone-receptor-positive. Histological studies of the patient tumour revealed a ductal invasive mammary carcinoma with 80% tubular growth pattern. Following transplantation the adenoid structures decreased to 30%; the mitosis rate and grade of malignancy increased. Treatment of the nude mice with 20 micrograms oestradiol benzoate/mouse caused a loss of the oestradiol receptor of the mammary carcinoma. The mammary carcinoma 3366 can be used for testing of antineoplastic substances, antihormones and for studies in regard to down-regulation or blocking of hormone receptors and possible consequences for therapies.

Mitogenic properties have been demonstrated in vitro for the lysosomal acidic protease cathepsin-D (cath-D). We investigated possible relationships between cath-D cytosol cell content and tumor proliferative activity in a series of 129 operable breast cancer patients. For total cytosol cath-D evaluation, a solid phase two-site immunoradiometric assay was utilized on tumor cell cytosol obtained for hormone receptor assay (DCC method). The percentage of S-phase cells was analyzed by 3H-thymidine autoradiographic assay. Median 3H-thymidine Labeling Index (3H-Tdr-LI) of the series was 2.7%; median cath-D content resulted 57 pmol/mg of protein cytosol and was significantly higher in node-positive with respect to the node-negative subgroup (p < 0.03). When classified in low, intermediate or high tumor cath-D content and slow or fast proliferative activity (cut-off: median values of the series), no significant agreement was found between the two variables. Statistical analysis, however, showed that a significant inverse correlation existed in node positive tumors between cath-D and 3H-Tdr-LI values which was even more evident in N-positive high estrogen receptor-positive (ER+) cases (coefficient of correlation = 0.6828; p = 0.0001). Cytosol cath-D content cannot be generally proposed as a direct marker of proliferative activity for operable breast cancer.

Oral contraceptives caused increased fibrinogen, FVII, FX, and fibrinolysis. The latter was associated with elevated FXII and PKK, while C1-INH was decreased, ATIII and alpha 2M were unchanged; it could not be accounted for by changes in t-PA, u-PA, PAI, plasminogen, alpha 2-AP, proteins C or S. HCII and alpha 1-PI were increased and may regulate the availability of thrombin and FXIa. The increased FXII/PKK dependent fibrinolytic potential and HCII may offset any increase in thrombin generation, while alpha 1-PI limits intrinsic coagulation.

We investigated the imbalance between thrombin and plasmin activity in vivo with various grades of severity of disseminated intravascular coagulation (DIC) in relation to the underlying diseases. Plasma thrombin-antithrombin-III complex (TAT) and plasmin-alpha 2-antiplasmin complex (PAP) levels were measured in 133 blood samples obtained from patients with DIC. The TAT/PAP ratio was higher in patients with sepsis or solid cancer than in those with hematologic malignancies. In acute promyelocytic leukemia (APL), the TAT levels were the highest, but the PAP levels were even higher and the TAT/PAP ratio was the lowest. As for the severity of DIC, in mild DIC, both thrombin and plasmin activities were increased. In moderate DIC, the TAT/PAP ratio increased, and thrombin activity was much more predominant. However, in severe DIC, the ratio decreased, and plasmin activity became excessive. In 3 patients with acute myeloblastic leukemia, APL and pancreatic cancer, respectively, the PAP level remained high during heparin therapy although the TAT level was decreased. When tranexamic acid was given, the PAP level was selectively reduced, and the TAT/PAP ratio was markedly decreased along with clinical improvement. These results indicate that monitoring of the TAT/PAP ratio may contribute to decisions regarding the institution and performance of combination therapy for DIC using anticoagulants and antifibrinolytic agents.

Tumorigenesis is thought to be a multistep process in which genetic alterations accumulate to bring about the neoplastic phenotype. Colorectal tumors appear to arise as a result of the mutational activation of oncogenes coupled with the inactivation of several tumor suppressor genes. We have found frequent allelic deletions of specific portions of chromosomes 5, 17, and 18 which presumably harbor suppressor genes. The target of allelic loss events on chromosome 17 has been shown to be the p53 gene, which is frequently mutated not only in colon cancer but in several other tumor types as well. Candidate suppressor genes have also recently been identified on chromosomes 18 and 5. The DCC gene on chromosome 18q encodes a protein with significant sequence similarity to neural cell adhesion molecules and other related cell surface glycoproteins. Alterations of this gene may interfere with normal cell growth and differentiation by disrupting cell-cell or cell-substrate interactions. Two genes (MCC and APC) on chromosome 5q have also recently been identified and partially cloned. These genes are located in a region tightly linked to familial adenomatous polyposis (FAP). While MCC mutations have been found only in sporadic colon tumors, APC mutations have been identified in sporadic tumors as well as the germline of patients with FAP. Studies are currently in progress to increase our understanding of how alterations of these genes affect colorectal tumor cell growth.

Estrogen (ER) and progesterone receptor (PR) contents in primary tumors from 127 advanced breast cancer patients were measured by DCC method. The patients were followed for 2 years and the prognostic value of the receptor levels was evaluated and compared with other tumor and patient characteristics. No relation was found between receptor levels and tumor, first relapse site as well as short-term survival (2 years) which might be due to the advanced stage of disease at diagnosis.

Heparin cofactor II (HCII) is a thrombin inhibitor in human plasma which displays similarities with antithrombin III (ATIII). Hereditary HCII deficiency was recently reported to be associated with thrombophilia. Since thromboembolism constitutes one of the main complications of the nephrotic syndrome (NS), both activities and antigen concentrations of HCII and ATIII were measured in the plasma and urine of 33 adult patients with nephrotic syndrome. The mean HCII plasma level was significantly increased whereas the ATIII level was decreased. Plasma HCII was significantly correlated with proteinuria and with the fibrinogen level, suggesting that HCII could act as an acute phase reactant in patients with NS. HCII antigen was detectable in 16 of the 24 available urine samples, whereas ATIII antigen was present in all of them. In addition, functionally active HCII was detected in most of the urine samples containing HCII antigen, while ATIII was only present in the inactive form. In conclusion, these findings suggest that HCII is submitted to a metabolic pathway different from that of ATIII in patients with NS.

There are many genetic abnormalities in colorectal cancers, and they can schematically be studied according to 3 approaches. 1. The quantitative abnormalities of the DNA content of the nucleus and the cell cycle are studied with flow cytometry. 2. Karyotypic abnormalities relating to the loss and/or gain of chromosomes or structural abnormalities are studied by cytogenetics. 3. Oncogene or anti-oncogene mutations carried out by these chromosomal segments are studied by molecular biology. When compared to the clinical data, some of these abnormalities have a prognostic value. They allow an insight into the fundamental mechanisms of colorectal carcinogenesis. Finally, they may allow predicting and assessing the efficacy of some adjunctive therapies, especially that of medication.

Steroid receptor binding activity was evaluated in specimens of 28 human cerebral meningiomas using a dextran coated charcoal (DCC) assay. Oestrogen receptor (ER) binding activity at significant levels (greater than 10 fmol/mg protein) could be detected only in three postmenopausal females (11% positive cases) at low levels. Positive progesterone receptor (PR) binding was detected in eighteen of the twenty-eight analysed meningioma tissues (64% positive cases). A significantly higher level of PR in male than in female patients could be demonstrated. The degree of peritumoural oedema could be evaluated from CT scans. There was no significant correlation between lack and amount of peritumoural oedema and quantity of cytosolic PR binding activity. Therefore we conclude, that peritumoural oedema is related to other factors and a possible role of PR activity in development of peritumoural oedema and growth control of meningiomas could not be demonstrated.

In the present study we investigated several factors of hemostasis and fibrinolysis during the normal pregnancy in 52 women. Whereas the Thrombin-Time did not show any change, the increase of the Hepato-Quick and the decrease of the Partial Thromboplastin-Time was significant. During pregnancy we observed a significant increase of the activity of fibrinogen, factor XII and prekallikrein as well as of the von-Willebrand-factor-antigen (VIIIR:Ag). The activity of factor II and X remained constantly. The increased activity of factors of hemostasis was accompanied by an increase of activity and concentration of antithrombin-III. In contrast to the activity of the hemostatic system we could not find any significant alteration of the fibrinolytic system. The cause must be searched in the observed increase of plasminogen-activator-inhibitor (PAI)--likly in combination with a decrease of the tissue-type plasminogen activator (t-PA). Our data indicate, that the highest risk for thrombosis already might exist in week 24 of pregnancy, because a distinct increase of hemostatic activity is combined with a nearly unchanged fibrinolytic activity.

<h4>Background and methods</h4>In contrast to hemochromatosis, which in white populations is inherited through a gene linked to the HLA locus, iron overload in sub-Saharan Africa is believed to result solely from increased dietary iron derived from traditional home-brewed beer. To examine the hypothesis that African iron overload also involves a genetic factor, we used likelihood analysis to test for an interaction between a gene (the hypothesized iron-loading locus) and an environmental factor (increased dietary iron) that determines transferrin saturation and unsaturated iron-binding capacity. We studied 236 members of 36 African families chosen because they contained index subjects with iron overload. Linkage to the HLA region was tested with use of lod scores.<h4>Results</h4>In the index subjects, increased iron was present in both hepatocytes and cells of the mononuclear-phagocyte system. Among family members with increased dietary iron due to the consumption of traditional beer, transferrin saturation in serum was distributed bimodally, with 56 normal values (less than 60 percent saturation) and 44 elevated values; the mean serum ferritin concentration was five times higher in the subjects with elevated transferrin saturation (P less than 0.005). The pedigree analysis provided evidence of both a genetic effect (P less than 0.005) and an effect of increased dietary iron (P less than 0.005) on transferrin saturation and unsaturated iron-binding capacity. In the most likely model, increased dietary iron raised the mean transferrin saturation from 30 to 81 percent and lowered the mean unsaturated iron-binding capacity from 38 to 13 mumol per liter in subjects heterozygous for the iron-loading locus. The hypothesis of tight linkage to HLA was rejected.<h4>Conclusions</h4>Iron overload in Africa may be caused by an interaction between the amount of dietary iron and a gene distinct from any HLA-linked gene.

The capacity of breast tumor cells to proliferate is considered a potential prognostic factor together with other histopathologic parameters. The authors determined the proliferation index on a large panel of human primary breast tumors by measuring the levels of incorporation of bromodeoxyuridine (BrdU) by fresh tumor specimens in culture. Previous analysis showed that the percentage of cells entering the S-phase of the cell cycle strongly correlates with tumor grade, tumor size, and estrogen and progesterone receptor status. The capacity of tumor cells to proliferate might be associated with specific genetic mutations in primary tumors. To test this hypothesis, a panel of 96 human breast carcinomas, for which the BrdU labeling index (LI) was known, were tested for loss of heterozygosity (LOH) or increased copy number (ICN) at chromosomes 1q, 3p, 13q, 17p, and 18q. On chromosome 17p, LOH and ICN were observed in 27% and 12%, respectively, of the informative breast tumors. The LOH on chromosome 17p was significantly associated with tumors having an elevated BrdU proliferation index (P = 0.022). No association (P = 0.45) was observed between BrdU LI and tumor size (T2 + T3 compared with T1), tumor grade, and lymph node status. Increased copy number on chromosome 17p, LOH or ICN on 1q, and LOH on 13q14, 18q, and 3p also showed no significant correlation with cell kinetic parameters. These data are consistent with the presence of a gene or genes on chromosome 17p13 near the YNZ22.1 locus whose normal functioning is necessary for controlling breast tumor cells proliferation in vivo.

Paracrine influences from fibroblasts derived from different sources of breast tissue on epithelial breast cancer cell growth in vitro were investigated. Medium conditioned (CM) by fibroblasts derived from tumours, adjacent normal breast tissue, and normal breast tissue obtained from reduction mammoplasty or from skin tissue significantly stimulated the growth of the steroid-receptor positive cell lines MCF-7 and ZR 75.1. The proliferation index (PI) on MCF-7 cells with CM from fibroblasts derived from breast tumour tissue was significantly higher than that obtained with fibroblasts derived from adjacent normal breast tissue (2p less than 0.05, n = 8). The PI obtained with CM from normal fibroblast cultures from reduction mammoplasty tissue, like normal tissue adjacent to the tumour, fell in the lower range of values. Skin fibroblast, like tumour tissue derived fibroblast, CM caused a high range PI. MDA-MB-231 and Evsa-T, two steroid-receptor negative cell lines, showed only a minor growth stimulatory responses with some of the fibroblast CM's. Evsa-T was occasionally inhibited by CM's. In conclusion, stromal factors play a role in the growth regulation of human breast cancer cells. The effects on cancer cell growth are, however, varying depending on the source of the stroma and the characteristics of the epithelial tumour cells.