Reliable and sensitive analysis of occult bone marrow metastases using automated cellular imaging.

The presence of occult bone marrow metastases (OM) has been reported to represent an important prognostic indicator for patients with operable breast cancer and other malignancies. Assaying for OM most commonly involves labor-intensive manual microscopic analysis. The present report examines the performance of a recently developed automated cellular image analysis system (ACIS; ChromaVision Medical Systems, Inc.) for identifying and enumerating OM in human breast cancer specimens. OM analysis was performed after immunocytochemical staining. Specimens used in this study consisted of normal bone marrow (n = 10), bone marrow spiked with carcinoma cells (n = 20), and bone marrow obtained from breast cancer patients (n = 39). The reproducibility of ACIS-assisted analysis for tumor cell detection was examined by having a pathologist evaluate montage images generated from multiple ACIS runs of five specimens. Independent ACIS-assisted analysis resulted in the detection of an identical number of tumor cells for each specimen in all instrument runs. Additional studies were performed to analyze OM from 39 breast cancer patients with two pathologists performing parallel analysis using either manual microscopy or ACIS-assisted analysis. In 17 of the 39 cases (44%), specimens were classified by the pathologist as positive for tumor cells after ACIS-assisted analysis, whereas the same pathologist failed to identify tumor cells on the same slides after analysis by manual microscopy. These studies indicate that the ACIS-assisted analysis provides excellent sensitivity and reproducibility for OM detection, relative to manual microscopy. Such performance may enable an improved approach for disease staging and stratifying patients for therapeutic intervention.

Simultaneous measurement of cell cycle and apoptotic cell death.

Understanding the dynamics of cell death in conjunction with those of cell cycle can be illuminating in the investigation of various cellular behaviors. Robust assays for measuring such parameters are invaluable. Many assays of apoptosis and/or cell cycle use flow cytometry. This chapter describes two different assays to measure apoptosis and cell cycle simultaneously using flow cytometry. The first involves the use of terminal transferase (the "TUNEL" assay) together with propidium iodide for identification of cell cycle. The second uses fluorescently labeled annexin V, together with propidium iodide as an indicator of cell membrane integrity; and additionally Hoechst 33342 for determination of cell cycle. Each assay has positive and negative attributes. The terminal transferase assay is performed using fixed cells and is therefore useful in the analysis of samples collected over time. The annexin V assay is performed using unfixed cells, and thus provides information regarding membrane integrity. Other practical aspects of both assays are discussed.

BCL-2 expression correlates with lower proliferative activity in the intermediate- and high-grade non-Hodgkin's lymphomas: an Eastern Cooperative Oncology Group and Southwest Oncology Group cooperative laboratory study.

An inverse relationship between BCL-2 expression and cell cycle transition has been suggested by recent studies in murine models. To investigate the clinical relevance of these laboratory studies, a group of 116 paraffin-embedded non-Hodgkin's lymphoma (NHL) biopsy specimens (Working Formulation Groups D-H, and J) from a cooperative group study of cellular DNA content were analyzed for the 14;18 translocation using polymerase chain reaction (PCR)-based methods and, if sufficient tissue remained, for BCL-2 and BAX expression by immunohistochemistry. The results of these studies were then compared with the results of the previously performed flow cytometric analysis of ploidy and proliferative activity (S-phase-fraction). BCL-2 expression was inversely associated with proliferative activity (P = .001; n = 41), but there was no association between staining for Bax and %S-phase. Ploidy was not associated with either BCL-2 or BAX expression. The t(14;18) was detected in 21 of the 54 cases in which PCR-amplifiable DNA was recovered; 20 of these occurred at the major breakpoint region and 1 at the minor breakpoint region. High levels of BCL-2 or BAX expression occurred independently of t(14;18). There was no association between t(14;18) and either ploidy or proliferative activity. The inverse relationship between BCL-2 expression and proliferative activity in the intermediate- and high-grade NHLs is consistent with recent studies suggesting that Bcl-2 both retards entry into the cell cycle and inhibits apoptosis.

Effects of temperature shift on cell cycle, apoptosis and nucleotide pools in CHO cell batch cultues.

Temperature reduction in CHO cell batch culture may be beneficial in the production of recombinant protein and in maintenance of viability. The effects on cell cycle, apoptosis and nucleotide pools were studied in cultures initiated at 37°C and temperature shifted to 30 °C after 48 hours. In control cultures maintained at 37 °C, viable cells continued to proliferate until the termination of the culture, however, temperature reduction caused a rapid decrease in the percent of cells in S phase and accumulation of cells in G-1. This was accompanied by a concurrent reduction in U ratio (UTO/UDP-GNAc), previously shown to be a sensitive indicator of growth rate. Culture viability was extended following temperature shift, as a result of delayed onset of apoptosis, however, once initiated, the rate and manner of cell death was similar to that observed at 37 °C. All nucleotide pools were similarly degraded at the time of apoptotic cell death. Temperature reduction to 30 °C did not decrease the energy charge of the cells, however, the overall rate of metabolism was reduced. The latter may be sufficient to extend culture viability via a reduction in toxic metabolites and/or limitation of nutrient deprivation. However, the possibility remains that the benefits of temperature reduction in terms of both viability and productivity are more directly associated with cultures spending extended time in G-1.

Apo-3, a new member of the tumor necrosis factor receptor family, contains a death domain and activates apoptosis and NF-kappa B.

BACKGROUND: Two receptors that contain the so-called "death domain' have been described to date: tumor necrosis factor receptor 1 (TNFR1) and Fas/Apo-1 (CD95); both belong to the TNFR gene family. The death domain of TNFR1 mediates the activation of programmed cell death (apoptosis) and of the transcription factor NF-kappa B, whereas the death domain of CD95 only appears to activate apoptosis. RESULTS: We have identified an additional member of the TNFR family, which we have named Apo-3. Apo-3 is a transmembrane protein of approximately 47 kDa that has similarity of members of the TNFR family in its extracellular, cysteine-rich domains. In addition, Apo-3 resembles TNFR1 and CD95 in that it contains a cytoplasmic death domain. The Apo-3 gene mapped to human chromosome 1p36.3, and Apo-3 mRNA was detected in several human tissues, including spleen, thymus, peripheral blood lymphocytes, small intestine and colon. Ectopic expression of Apo-3 in HEK293 or HeLa cells induced marked apoptosis. CrmA, a poxvirus inhibitor of Ced-3-like proteases which blocks death signaling by TNFR1 and CD95, inhibited Apo-3-induced apoptosis. Ectopic expression of Apo-3 also induced the activation of NF-kappa B. Apo-3 did not specifically bind to the Apo-2 ligand, suggesting the existence of a distinct ligand for Apo-3. CONCLUSIONS: These results identify Apo-3 as a third member of the TNFR family that activates apoptosis, and suggest that Apo-3, TNFR1 and CD95 engage a common apoptotic cell-death machinery. Apo-3 resembles TNFR1 because it can stimulate NF-kappa B activity and regulate apoptosis. Apo-3 mRNA is expressed in various tissues, consistent with the possibility that this receptor may regulate multiple signaling functions.

Prognostic implications of ploidy and proliferative activity in the diffuse, aggressive non-Hodgkin's lymphomas.

The International Index is a powerful predictor of outcome in the aggressive non-Hodgkin's lymphomas that is based solely on clinical features. Proliferative activity (% S-phase) measured by flow cytometry has been reported to have prognostic significance in many series and may represent a biologic correlate of clinical behavior that further defines prognosis. Flow cytometric analysis of cellular DNA content and proliferative activity (% S-phase) was performed on fixed paraffin-embedded biopsy specimens from 242 previously untreated patients with diffuse, aggressive non-Hodgkin's lymphomas entered on phase III intergroup clinical trials. The International Index was calculated for each patient based on stage, lactate dehydrogenase, performance status, number of extranodal sites, and age, as previously reported. The International Index consistently predicted response to therapy (P = .027) and survival (P = .007) in this series. DNA aneuploidy was shown in 57% of cases, but was not predictive of clinical outcome. The median % S-phase was 9.9 (median coefficient of variation, 3.6%), which was highly correlated with mitotic index (P = .0001). Although a trend associating low proliferative activity with good early survival and very high S-phase with a shortened survival was shown, International Index risk was the only significant predictor of survival in the multivariate analysis. Although proliferative activity quantitated by flow cytometric analysis of nuclei extracted from paraffin-embedded specimens is probably predictive of survival, it is a less powerful prognostic indicator than clinical parameters represented by the International Index and provides no additional prognostic information.

Aequorea green fluorescent protein: simultaneous analysis of wild-type and blue-fluorescing mutant by flow cytometry.

Aequorea green fluorescent protein (GFP) has been expressed in a variety of cell lines and host organisms. A recent report (Heim et al.: Proc Natl Acad Sci USA 91:12501-12504, 1994) has documented that a GFP mutant with a single amino acid substitution (tyrosine 66 to histidine; Y66H-GFP) elicits altered spectral properties. Whereas wild-type GFP emits with a maximum at approximately 509 nm (green fluorescence), Y66H-GFP fluoresces with a maximum at approximately 448 nm (blue fluorescence). In this study we employed available argon and krypton ion laser lines to investigate the impact of laser excitation wavelength on the detection of Y66H-GEP by flow cytometry. Using transiently transfected 293 cells, a cellular subpopulation with elevated blue fluorescence was detectable with excitation at 407 nm, but not with ultraviolet (UV), 458 nm, or 488 nm excitation. The blue-fluorescing cells were further documented to express Y66H-GFP by immunoblot analysis of sorted cells. Finally, we demonstrated the simultaneous analysis of both wild-type and Y66H-GFP in cotransfected 293 cells using 407 nm excitation while collecting blue fluorescence at 460 +/- 20 nm (Y66H-GFP) and green fluorescence at 525 +/- 25 nm (wild-type GFP). These studies illustrate the potential for assessing differential gene expression by simultaneously analyzing multiple GFP species with multiparameter flow cytometry.

Activation of apoptosis by Apo-2 ligand is independent of FADD but blocked by CrmA.

A new member of the tumor necrosis factor (TNF) cytokine family, designated Apo-2 ligand (Apo-21) [1] or TRAIL [2], has been shown recently to induce apoptosis in various tumor cell lines; however, its biological role is unknown. Here, we show that Apo-21, activated apoptosis in T-cell-enriched cultures of peripheral blood lymphocytes stimulated by interleukin-2 (IL-2), but not in unstimulated cells. This finding suggests that, like Fas/Apo-1 ligand and TNF [3-5], Apo-2L may play a role in regulating post-stimulation apoptosis of mature lymphocytes. Studies on the mechanism of Apo-2L action demonstrated marked membrane blebbing, a hallmark of apoptosis, within a few minutes of the addition of Apo-2L to tumor cells. Ectopic expression of a dominant negative mutant of FADD, a cytoplasmic protein that mediates death signalling by Fas/Apo-1 and by TNF receptor type 1 (TNFR1) [6-9], inhibited the induction of apoptosis by anti-Fas/Apo-1 antibody, but had little effect on Apo-2L function. In contrast, expression of CrmA, a cowpox virus-derived inhibitor of the Ced-2-like proteases ICE [10] and CPP32/Yama [11,12], blocked the induction of apoptosis by either Apo-2L or anti-Fas/Apo-1 antibody. These results suggest that Apo-2L activates a rapid, FADD-independent pathway to trigger a cell-death programme that requires the function of cysteine proteases such as ICE or CPP32/Yama.

Heparin induces dimerization and confers proliferative activity onto the hepatocyte growth factor antagonists NK1 and NK2.

Hepatocyte growth factor (HGF) is a potent epithelial mitogen whose actions are mediated through its receptor, the proto-oncogene c-Met. Two truncated variants of HGF known as NK1 and NK2 have been reported to be competitive inhibitors of HGF binding to c-Met, and to function as HGF antagonists (Lokker, N.A., and P.J. Godowski. 1993. J. Biol. Chem. 268: 17145-17150; Chan, A.M., J.S. Rubin, D.P. Bottaro, D.W. Hirschfield, M. Chedid, and S.A. Aaronson. 1991. Science (Wash. DC). 254:1382-1387). We show here, however, that NK1 acts as a partial agonist in mink lung cells. Interestingly, NK1, which is an HGF antagonist in hepatocytes in normal conditions, was converted to a partial agonist by adding heparin to the culture medium. The interaction of NK1 and heparin was further studied in BaF3 cells, which express little or no cell surface heparan sulfate proteoglycans. In BaF3 cells transfected with a plasmid encoding human c-Met, heparin and NK1 synergized to stimulate DNA synthesis and cell proliferation. There was no effect of heparin on the IL-3 sensitivity of BaF3-hMet cells, and no effect of NK1 plus heparin in control BaF3 cells, indicating that the response was specific and mediated through c-Met. The naturally occurring HGF splice variant NK2 also stimulated DNA synthesis in mink lung cells and exerted a heparin-dependent effect on BaF3-hMet cells, but not on BaF3-neo cells. The activating effect of heparin was mimicked by a variety of sulfated glycosaminoglycans. Mechanistic studies revealed that heparin increased the binding of NK1 to BaF3-hMet cells, stabilized NK1, and induced dimerization of NK1. Based on these studies, we propose that the normal agonist activity of NK1 and NK2 in mink lung cells is due to an activating interaction with an endogenous glycosaminoglycan. Consistent with that model, a large portion of the NK1 binding to mink lung cells could be blocked by heparin. Moreover, a preparation of glycosaminoglycans from the surface of mink lung cells induced dimerization of NK1. These data show that the activity of NK1 and NK2 can be modulated by heparin and other related glycosaminoglycans to induce proliferation in cells expressing c-Met.

Growth regulation of human breast and ovarian tumor cells by heregulin: Evidence for the requirement of ErbB2 as a critical component in mediating heregulin responsiveness.

Alterations in the expression of the epidermal growth factor (EGF) receptor ErbB family are frequently encountered in a number of human cancers. Two of these receptors, ErbB3 and ErbB4, are known to bind a family of related proteins termed heregulins (HRGs) or neu differentiation factors. In biologically relevant systems, interaction of HRG with ErbB3 or ErbB4 results in the transactivation of ErbB2. In this report, we show that ErbB2 is a critical component in mediating HRG responsiveness in a panel of human breast and ovarian tumor cell lines. Because HRGs have been reported to elicit diverse biological effects on cultured cells, including growth stimulation, growth inhibition, and induction of differentiation, we systematically examined the effect of rHRG beta 1 on tumor cell proliferation. HRG binding studies were performed with a panel of breast and ovarian tumor cell lines expressing a range of levels of ErbB2. The biological responses to HRG were also compared to EGF and to the growth-inhibitory anti-ErbB2 antibody, 4D5. In most cases, HRG stimulation of DNA synthesis correlated with positive effects on cell cycle progression and cell number and with enhancement of colony formation in soft agar. On each cell line tested, the HRG effects were distinguishable from EGF and 4D5. Our findings indicate that HRG induces cell proliferation in a number of tumor cell lines. In addition, we show that methods for measuring cell proliferation, as well as experimental conditions, are critical for determining HRGs effect on tumor cell growth in vitro.

Aequorea green fluorescent protein analysis by flow cytometry.

The isolation and expression of the cDNA for the green fluorescent protein (GFP) from the bioluminescent jellyfish Aequorea victoria has highlighted its potential use as a marker for gene expression in a variety of cell types (Chalfie et al.: Science 263: 802-805, 1994). The longer wavelength peak (470 nm) of GFP's bimodal absorption spectrum better matches standard fluorescein filter sets; however, it has a considerably lower amplitude than the major absorption peak at 395. In an effort to increase the sensitivity of GFP with routinely available instrumentation, Heim et al. (Nature 373:663-664, 1995) have generated a GFP mutant (serine-65 to threonine; S65T-GFP) which possesses a single absorption peak centered at 490 nm. We have constructed this mutant in order to determine whether it or wild-type GFP (wt-GFP) afforded greater sensitivity when excited near their respective absorption maxima. Using the conventionally available 488 nm and ultraviolet (UV) laser lines from the argon ion laser as well as the 407 nm line from a krypton ion laser with enhanced violet emission, we were able to closely match the absorption maxima of both the S65T and wild-type forms of Aequorea GFP and analyze differences in fluorescence intensity of transiently transfected 293 cells with flow cytometry. The highest fluorescence signal was observed with 488 nm excitation of S65T-GFP relative to all other laser line/GFP pairs. The wt-GFP fluorescence intensity, in contrast, was significantly higher at 407 nm relative to either 488 nm or UV. These results were consistent with parallel spectrofluorometric analysis of the emission spectrum for wt-GFP and S65T-GFP. The relative contribution of cellular autofluorescence at each wavelength was also investigated and shown to be significantly reduced at 407 nm relative to either UV or 488 nm.

Effects of omeprazole on cell kinetics of carcinogen-induced colon tumours in rats.

This study was designed to evaluate the effects of hypergastrinaemia induced via suppression of gastric acid by omeprazole on carcinogen-induced colon cancer in rats. The carcinogen methylazoxymethanol (MAM), 30 mg/kg, was administered intraperitoneally at 6-weekly intervals to Sprague-Dawley rats. Four weeks after the last MAM injection, the first daily dose of omeprazole, 40 mg/kg, was given by gastric gavage to one group of rats, and the rest were given buffered methylcellulose vehicle. After 10 weeks of daily omeprazole or vehicle, the rats were anaesthetized with ether, blood samples obtained, and animals sacrificed. Gastrin levels in serum from omeprazole-treated rats were elevated nearly six-fold. DNA and RNA levels in gastric mucosa were unchanged by omeprazole, but protein content was somewhat reduced. No biochemical or histological changes related to omeprazole treatment were observed in normal colon. The number of tumours, tumour volumes, and total tumour burden were not significantly different in colons of vehicle- or omeprazole-treated rats. Analysis by flow cytometry revealed that the S phase fraction was lower in tumour cells from omeprazole-treated animals; and that the frequency of DNA aneuploidy was also reduced. The results indicate that while omeprazole-induced suppression of stomach acid in rats elevates levels of gastrin in serum, it does not substantially alter the biochemical or cellular characteristics of carcinogen-induced colon tumours.

Apoptosis in CHO cell batch cultures: examination by flow cytometry.

Chinese hamster ovary cells grown under conditions which are optimal for the production of a genetically engineered protein in batch culture, lose significant viability shortly after entering the stationary phase. This cell death was investigated morphologically and was found to be almost exclusively via apoptosi. Furthermore, cells were analyzed by flow cytometry using a fluorescent DNA end-labeling assay to label apoptotic cells, in conjunction with cell cycle analysis using propidium iodide. Apoptotic cells could be detected by this method, and by the radioactive end-labeling of extracted DNA, on all days of culture from day 1 to day 7; however, the degree of apoptotic cell death increased dramatically when the cells entered the stationary phase, rising to 50-60% of the total cell number at the termination of the culture. Flow cytometric analysis showed that the majority of cells underwent apoptosis whilst in G(1)/G(0) and formed an apoptotic population with high DNA FITC end-labeling and hypodiploid propidium iodide binding. Additionally, the ability or inability to secrete specific protein products did not appear to interfere with the development of the apoptotic population with time.

Evidence for molecular heterogeneity in human ganglioneuroblastoma.

Ganglioneuroblastomas are tumors of sympathetic cell origin that contain both primitive neuroblastomatous and mature ganglioneuromatous elements. It is thought that these tumors arise from a single cellular clone and that the morphologically distinct components of ganglioneuroblastomas represent cells in different stages of differentiation. Two pathologic variants of this tumor, composite and diffuse, have been described; metastasis is more commonly seen with composite ganglioneuroblastomas. We analyzed a composite ganglioneuroblastoma for N-myc copy number at initial resection and 2 years later after progressive disease. In the second sample the more differentiated portion of the tumor was analyzed separately from the neuroblastic foci for N-myc copy number and DNA ploidy. The DNA content and N-myc copy number differed in the two morphologically discrete areas of the tumor, suggesting that at least two clones were present. More composite ganglioneuroblastomas need to be examined to determine whether these tumors are largely composed of tumor cell populations with molecular heterogeneity.

Effect of recombinant aprotinin on platelet activation in patients undergoing open heart surgery.

The hemostatic effect of recombinant (r)-aprotinin was studied in 41 patients undergoing cardiopulmonary bypass surgery. Flow cytometry was used to measure the expression of glycoprotein 1b (GP1b) and alpha-granule membrane protein 140 (GMP-140) on platelets. Additional parameters studied were beta-thromboglobulin (beta-TG), fibrinogen, fibrinopeptide A, plasminogen, platelet count, and the amount of postoperative chest tube drainage. Chest tube drainage was significantly less in the r-aprotinin-treated patients (p < 0.001). The percentage of platelets expressing GMP-140 increased to 9.7% in r-aprotinin patients and to 12.1% in controls during the surgery. The difference between both groups was not significant. Both values returned to baseline the next day. GP1b expression decreased in both groups by 33-38% during the surgery and returned to baseline the next day. GP1b expression and the plasma concentrations of fibrinogen, fibrinopeptide A, beta-TG, and plasminogen were not different in r-aprotinin and control patients. We conclude that r-aprotinin reduces blood loss during cardiopulmonary bypass surgery, but the primary mechanism is likely not through a protective effect on platelets.

DNA flow cytometric analysis of clinical stage I endometrial carcinomas with lymph node metastases.

Twenty-one (8%) of 264 consecutive evaluable patients with clinical stage 1 endometrial carcinoma had histologic evidence of pelvic and/or para-aortic lymph node metastases. DNA flow cytometry was performed on both the primary tumor and nodal metastasis. Seventeen of 21 sets could be analyzed. Overall, 11 (65%) of the primary carcinomas were aneuploid. Nine of 17 (53%) had consistent ploidy patterns when the primary tumor and lymphatic metastasis were compared. The remaining 8 (47%) had aneuploid primaries with diploid nodal metastases. Five (83%) of the 6 patients with diploid primary tumors were alive without evidence of disease compared to 3 of 11 (27%) patients with aneuploid tumors (P < 0.05). Other predictors of disease outcome included tumor histology, lymph vascular space invasion, and depth of myometrial invasion. Ploidy status of the lymphatic metastasis was not important in terms of overall survival. All 8 patients with para-aortic nodal metastases had aneuploid primary carcinomas compared to 4 (44%) of 9 patients with pelvic node involvement only (P < 0.01). Mean survival was 31 months for patients with para-aortic node metastases compared to 51 months for patients with only pelvic node metastases. Comparison of survival curves among these two groups demonstrated a significant survival advantage in patients with regional nodal metastases (P = 0.032). S-phase fraction of both the primary tumor and lymphatic metastasis did not correlate with survival or predict disease outcome. DNA index of the primary tumor, as a continuous variable, was inversely proportional to survival, demonstrating poorer survivorship with incremental increases of DI. Ploidy status of the lymph node metastasis was an inconsistent reflection of the primary tumor's expression and behavior and, therefore, little additional information was gained by knowledge of the lymphatic ploidy status.

Growth factor receptor and related oncogene determination in mesenchymal tumors.

BACKGROUND: Sarcoma are rare malignant neoplasms that originate from a mesenchymal cell line. The epidermal growth factor receptor (EGF-R) has been identified in these malignant neoplasms by immunohistochemical techniques. METHODS: This investigation has evaluated the gene amplification and expression of EGF-R and the homologous oncogene c-erbB-2 in soft tissue and osseous sarcomas by Southern and northern blot analysis. RESULTS: Amplification of EGF-R and c-erbB-2 was identified in 2 of 117 (1.7%) and 6 of 105 (5.7%) of the sarcomas, respectively. Increased expression of EGF-R and c-erbB-2 was identified in 21 of 43 (49%) and 35 of 94 (37%) sarcomas, respectively. CONCLUSIONS: The expression of these two genes in sarcomas appears to occur independently and not be associated with tumor histologic characteristics, grade, size, DNA content, or proliferative activity.

Effect of retinoic acid on the proliferation and secretory activity of androgen-responsive prostatic carcinoma cells.

We studied the effect of retinoic acid on the growth and secretory activity of the androgen-responsive prostatic carcinoma cell line LNCaP. Our data showed that retinoic acid at 0.01 microM. stimulated the proliferation of LNCaP cells but inhibited their growth at 0.1 microM. under androgen-free conditions. In the presence of 0.1 nM. dihydrotestosterone (DHT), LNCaP cell proliferation was inhibited by 10 microM. retinoic acid but not by lower concentrations of retinoic acid. Retinoic acid reduced LNCaP cell growth at concentrations of 0.1 microM. in the presence of 10 nM. DHT. Retinoic acid (10 microM.) also reduced the growth response of LNCaP cells to epidermal growth factor and transforming growth factor alpha and potentiated the inhibitory effect of transforming growth factor beta. In additional studies, retinoic acid induced a dose-dependent increase in prostate specific antigen (PSA) secretion at concentrations of 0.1 to 1 microM. Dihydrotestosterone (10 nM.) also enhanced the secretion of PSA by LNCaP cells, and this effect was potentiated in a dose-dependent fashion by the addition of retinoic acid at 0.1-10 microM. Competitive binding studies showed that retinoic acid did not bind to androgen receptors. Overall, retinoic acid had a biphasic effect on LNCaP proliferation and promoted the secretion of PSA. The biphasic effect of retinoic acid on LNCaP growth should be considered in designing in vivo studies to determine the impact of retinoic acid on solid prostatic tumor growth. In addition, the ability of retinoic acid to increase PSA secretion may complicate the interpretation of serum PSA levels used for diagnostic and prognostic purposes.