Biomarker and in vitro chemoresistance profiles of undifferentiated ovarian cancers

2532

Background: To determine the biomarker expression and drug resistance patterns of undifferentiated ovarian cancers. Methods: Biomarker expression was determined by immunohistochemistry. In vitro drug resistance profiles were analyzed by assays exposing tumors to suprapharmacologic doses of chemotherapy. Results: Of 177 ovarian cancer specimens, 63 were classified as undifferentiated tumors and 114 well-differentiated cancers were used for comparison. Of the undifferentiated tumors, 46.7% were primary cancers and the remainder were recurrent cancers. Most of the well-differentiated tumors were obtained from primary (82%) vs. recurrent disease specimens. The average DNA index and S-phase fraction were significantly higher in the undifferentiated vs. well-differentiated tumors (1.42 vs. 1.23, p=0.025 for DNA index; 1.69 vs. 3.06, p<0.0001 for S phase fraction). In addition, the percent of tumors with p53 mutations was also higher in the undifferentiated compared to the well-differentiated tumors (47.8% vs. 17.8%; p=0.005). The mean percent cell inhibition was significantly lower in the undifferentiated tumors compared to well-differentiated cancers after exposure to etoposide (54%±21 vs. 80%±18; p=0.0002), doxorubicin (59%±28 vs. 80.0%±21; p=0.07), and topotecan (61.7%±20 vs. 71%±21; p=0.015). Compared to well-differentiated tumors, undifferentiated tumors tended to be more resistant to etoposide (40% vs. 11%), doxorubicin (22% vs. 9%), and topotecan (19% vs. 16%). However, the drug resistance patterns after exposure to taxanes (paclitaxel and taxotere) and platinums (carboplatin and cisplatin) were similar. Conclusions: Undifferentiated tumors of the ovary have a significantly different biomarker expression and drug resistance profiles compared to well-differentiated tumors. Given that undifferentiated cancers provide a treatment challenge for the clinician, this data can provide additional prognostic and therapeutic information, and warrants further investigation.

No significant financial relationships to disclose.

Angiogenesis index (AI) as a predictor of tumor burden in lymph nodes in colon cancer (Cca) patients (pts) undergoing sentinel lymph node (SLN) mapping (M)

14509

Background: Lymph Node (LN) status is the most important prognostic factor in colon cancer (Cca). Angiogenesis Index (AI) has been studied as a prognostic marker in various solid tumors with conflicting results. Hence, a retrospective analysis was done to evaluate the role of AI as a prognostic marker in Cca. Methods: Pts with Cca who underwent SLNM to determine LN status were included. A portion of tumor was sent for tumor marker analysis including p53, Thrombospondin-1 and CD31 by IHC. AI was derived for each specimen by summing the biomarker specific score for the three tumor markers. AI of -ve 6 was taken as cut off for significance based on previous studies on solid tumors. Metastatic foci in SLNs and non-SLNs were measured in greatest diameter by ocular micrometer; and were added for each pt to estimate overall tumor burden in SLNs and non-SLNs. Results: A total of 111 consecutive pts with Cca were included in the study. SLNM was successful in 100% pts. Pts with distant metastasis (mets) (n=18), Tis (n=1) and skip mets (n=9) were excluded from final analysis. Out of the remaining 83 pts, an AI of -6 or less was found in 22 pts (26.5%) while 61 pts (73.5%) had AI of more than (>) -6. Of the 61 pts with AI > -6, 37.7% pts were SLN +ve while 62.3% pts were SLN -ve (p=0.05)( Table 1A ). Of the 22 pts with AI less than or -6, 45.5% pts were SLN +ve and 54.5% pts were SLN -ve (p=0.65). Size of the metastatic tumor burden in lymph nodes was available in 69.7% of SLN +ve pts. Total average tumor burden for pts with AI > -6 (n=15) was 2.04cm as compared to 1.48cm in pts with AI of -6 or less (n=8)(p=0.66). Average SLN met size was 0.73cm in pts with AI > -6 and 0.63cm in pts with AI of -6 or less (p=0.66)( Table 1B ). Conclusions: AI did not correlate with nodal positivity or tumor burden in LNs in pts with Cca. LN status remains the most important prognostic marker in Cca. Further larger trials are required to determine the role of AI as a prognostic marker in Cca.

[Table: see text] [Table: see text]

No significant financial relationships to disclose.

Correlation between immunohistochemical biomarker expression and in vitro drug response factors in stage III-IV ovarian cancer

16055

Background: The reliability of molecular biomarkers as predictors of treatment outcome remains unclear and, in relation to drug resistance even less is known. Immunohistochemical (IHC) expression of biomarkers and in vitro drug resistance (IVDR) relative to clinical outcome were evaluated in patients with stage III-IV epithelial ovarian cancer (OC) who received combination carboplatin (CAR) plus paclitaxel (TAX) chemotherapy. Methods: We correlated clinical outcome with histopathology and biomarker expression of MIB-1, p53, BCL2, EGFR, ER and PR in 98 OC patients with tumors tested for IVDR by extreme drug resistance (EDR) assay. IVDR was determined following exposures to single agents CAR, cisplatin (CP), TAX, taxotere (TXT), gemcitabine (GCB), topotecan (TP), liposomil doxorubicin, and cytoxan (CT). Percentage cell growth inhibition (PCI) for each drug was compared with PCI from untreated control cultures. Results: Tumors with prior chemotherapy (n=29) were more resistant to CAR (unpaired t-test, p=0.046), and CT (p=0.038), and more sensitive to liposomil doxorubicin (p=0.009), compared to treatment naive tumors (n=53). For 65 tumors, there was a positive relationship between MIB-1 vs p53 (r=0.419; p=0.001), and between ER vs PR (r=0.339; p=0.011). Correlation between biomarkers and drug response showed a significantly inverse relationship between MIB-1 vs liposomil doxorubicin, p53 vs liposomil doxorubicin, EGFR vs CAR, CP, GCB and TAX; and between ER vs TXT and CT, and PR vs CT and TP. With a 53% (33/61) response rate to CAR+TAX, time to progression (TTP) for responders (R) with EGFR+ tumors was longer (median TTP 15.6 mo); and shorter for R with EGFR- tumors (p=0.008). For R and non-R, BCL2+ tumors had longer overall survival (OS) (median OS 23.3 mo; p=0.006). No other significant correlations were seen. Conclusion: Prior chemotherapy increased IVDR to CAR and CT, and increased sensitivity to liposomil doxorubicin. Increased expression of proliferation biomarkers, particularly EGFR, correlated with resistance to several chemotherapeutic agents. EGFR expression in responders was associated with longer TTP. The significance of in vitro correlates of prognostic biomarker expression and IVDR for predicting chemoresistance warrants prospective studies to assess its value in clinical settings.

[Table: see text]

In vitro drug resistance profiles correlate with DNA ploidy and p53 expression in 3,106 endometrioid and papillary serous adenocarcinomas of the uterus

5037

Background: Although chemotherapy is widely used to treat advanced and recurrent endometrial cancer, no patient-tailored protocols were developed, resulting in intermittent and incomplete clinical response. The utility of the Extreme Drug Resistance (EDR) assay to predict clinical resistance to chemotherapy was proven in several cancer types, including ovarian cancer. This study is the first attempt to evaluate the efficacy of the EDR assay in endometrioid (E) and papillary serous (PS) tumors and correlate it with the expression of two major biomarkers, DNA ploidy and p53. Methods: 2520 E and 586_PS adenocarcinomas were tested in the EDR assay using in vitro incubation in suprapharmacologic exposures of 4-HC (the active species of cyclophosphamide), cisplatin, carboplatin, and doxorubicin. The results were defined as EDR (one standard deviation more resistant than the database median), intermediate drug resistance (IDR - between the median and EDR), and low drug resistance (LDR - lower than the median). DNA ploidy was assessed in flow cytometry; p53 expression was quantitated on paraffin sections by immunohistochemistry. Results: Compared to E, the more clinically aggressive PS tumors were significantly associated with high (IDR + EDR) in vitro resistance to all EDR-tested drugs: 4-HC (p = 0.0010; χ2 statistics), cisplatin (p = 0.0064), carboplatin (p = 0480), doxorubicin (p = 0.0001). PS tumors showed significantly higher non-diploid status (p < 0.0001) and p53-positivity (p < 0.0001). For all EDR-tested drugs, increased in vitro resistance closely correlated with the non-diploid and high p53 status. Conclusions: These data document close correlation between the EDR assay results and biological behavior of endometrial tumors in terms of their clinical aggressiveness and biomarker expression and justify further studies on combined EDR and biomarker testing in assay-guided chemotherapy.

[Table: see text]

No significant financial relationships to disclose.

Evidence for multidrug resistance-1 P-glycoprotein-dependent regulation of cellular ATP permeability

The mechanisms responsible for regulating epithelial ATP permeability and purinergic signaling are not well defined. Based on the observations that members of the ATP-binding cassette (ABC)1 family of proteins may contribute to ATP release, the purpose of these studies was to assess whether multidrug resistance-1 (MDR1) proteins are involved in ATP release from HTC hepatoma cells. Using a bioluminescence assay to detect extracellular ATP, increases in cell volume increased ATP release approximately 3-fold. The MDR1 inhibitors cyclosporine A (10 microm) and verapramil (10 microm) inhibited ATP release by 69% and 62%, respectively (p < 0.001). Similarly, in whole-cell patch-clamp recordings, intracellular dialysis with C219 antibodies to inhibit MDR1 decreased ATP-dependent volume-sensitive Cl- current density from -33.1 +/- 12.5 pA/pF to -2.0 +/- 0.3 pA/pF (-80 mV, p < or = 0.02). In contrast, overexpression of MDR1 in NIH 3T3 cells increased ATP release rates. Inhibition of ATP release by Gd3+ had no effect on transport of the MDR1 substrate rhodamine-123; and alteration of MDR1-substrate selectivity by mutation of G185 to V185 had no effect on ATP release. Since the effects of P-glycoproteins on ATP release can be dissociated from P-glycoprotein substrate transport, MDR1 is not likely to function as an ATP channel, but instead serves as a potent regulator of other cellular ATP transport pathways.

P-Glycoprotein conformational changes detected by antibody competition

Conformational changes accompanying P-glycoprotein (Pgp) mediated drug transport are reflected by changes in the avidity of certain monoclonal antibodies (mAbs). More of the UIC2 mAb binds to Pgp-expressing cells in the presence of substrates or modulators [Mechetner, E.B., Schott, B., Morse, S.B., Stein, W., Druley, T., Dvis, K.A., Tsuruo, T. & Roninson, I.B. (1997) Proc. Natl Acad. Sci. USA 94, 12908-12913], while the binding of other mAbs (e.g. MM12.10, MRK16, 4E3) is not conformation sensitive. Pre-staining of Pgp+ cells with UIC2 decreased the subsequent binding of MM12.10 mAb by about 30-40%, suggesting that there are Pgp molecules available for both UIC2 and MM12.10, and others accessible only for MM12.10. In the presence of certain substrates/modulators such as vinblastin, cyclosporin A or valinomycin, the MM12.10 reactivity was completely abolished by preincubation with UIC2. However, verapamil, Tween-80 and nifedipine did not influence the ratio of bound mAbs significantly. This is the first assay to our knowledge, sharply distinguishing two classes of modulators. The conformational changes accompanying the mAb competition phenomenon appear to be closely related, though not identical to those accompanying the UIC2-shift, as suggested by the simultaneous assessment of the two phenomena.

Conformational heterogeneity of P-glycoprotein

P-glycoprotein (P-gp) acts as an active efflux mechanism for a large number of cytostatics and seems to be involved in the frequent failure of cancer chemotherapy. The molecular events of substrate recognition and transport still are not understood completely. We show here that the percentage of P-gp epitopes available for labeling with UIC2 monoclonal antibody is increased significantly after methanol permeabilization/fixation of cells. At the same time, binding of the MRK16 and 4E3 anti-P-gp antibodies is changed only moderately. Confocal microscopical images of UIC2-PE-labeled cells show that the epitopes becoming available after fixation are situated mainly in the plasma membrane. Thus, only a minority of P-gp molecules are accessible for UIC2 in the cell membrane of live cells, and methanol treatment can expose a large pool of previously plasma membrane-embedded, cryptic UIC2 epitopes. The UIC2-reactive P-gp molecules do not appear to be sequestered spatially, as suggested by the high fluorescence resonance energy transfer efficiency measured between the fluorescently labeled competing UIC2 and MRK16 antibodies, suggestive of P-gp dimerization and oligomerization on live cells.

Chemotherapy immediately following autologous stem-cell transplantation in patients with advanced breast cancer

Most patients relapse after high-dose chemotherapy (HDCT) with autologous stem-cell transplantation (ASCT) for metastatic breast cancer. Further chemotherapy immediately after hematopoietic recovery from ASCT is not given for fear of irreversibly damaging the newly engrafted stem cells. In a pilot chemoprotection trial, autologous CD34+ cells from patients with metastatic breast cancer were exposed to a replication-incompetent retroviral vector carrying MDR-1 cDNA and then reinfused after HDCT. Immediately on recovery, patients received multiple courses of escalating dose paclitaxel. All of the 10 patients tolerated reinfusion of modified cells without any toxicity and had myeloid engraftment within 12 days (range, 11-14). The bone marrow cells of three patients contained vector MDR-1-positive cells only at the time of the first course of posttransplant paclitaxel, indicating that the MDR-1 vector-modified cells had only short-term engrafting potential. A total of 83 courses of paclitaxel were administered starting at a median of 30 (range, 21-32) days from ASCT. The median dose of paclitaxel was 225 mg/m2 and the median interval between paclitaxel cycles of therapy was 21 (range, 20-41) days. Five of the six CR patients were able to receive all of the 12 courses of paclitaxel. Three patients who had achieved less than a complete response to the HDCT (2 patients) and partial response (1 patient) were converted to complete clinical responses during the 12 cycles of paclitaxel. No delayed toxicity or bone marrow failure was noted in these patients with a median follow-up of 2 years from ASCT. This is the first study of chemotherapy immediately after transplantation with autologous CD34+ cells. These data indicate that paclitaxel can be safely administered immediately after ASCT without any delayed toxicities. Paclitaxel given immediately after ASCT can further improve the response to pretransplant chemotherapy in patients with advanced breast cancer.

Levels of multidrug resistance (MDR1) P-glycoprotein expression by human breast cancer correlate with in vitro resistance to taxol and doxorubicin

To determine whether multidrug resistance (MDR1) P-glycoprotein (Pgp) expression correlated with clinical MDR1-related drug resistance, we established a protocol for quantitative measurement of Pgp expression and in vitro drug resistance in doxorubicin resistant MCF7 breast cancer cell lines and 359 freshly resected specimens of breast carcinoma. Pgp expression was detected with 4E3, UIC2, and JSB-1 monoclonal antibodies using flow cytometry and immunohistochemistry (IHC). Pgp function was determined using PSC833 in a drug resistance-reversal assay and with a three-dimensional agarose-based extreme drug resistance assay. MCF7 calibrator cell lines expressed Pgp, which was functional and in proportion to the degree of drug resistance. Flow cytometry, UIC2 shift assays, IHC scores, and determination of absorbance products by image analysis were all highly correlated (r > 0.9). Overall Pgp expression increased from 11% in untreated patients to 30% in patients who had previously received chemotherapy. Compared with Pgp-negative tumors, a significant increase in doxorubicin and Taxol resistance was seen for breast cancers that expressed Pgp, regardless of prior treatment. A strong correlation between the degree of Pgp expression and in vitro resistance to Taxol and doxorubicin (but not to 5-fluorouracil) was found when either IHC scores or image analysis-based methods were used to quantify Pgp expression (n = 185, P < 0.0001). The degree of Pgp expression strongly correlated with the degree of drug resistance in the clinical specimens studied. These data suggest that (a) Pgp contributes to clinical MDR1-related drug resistance, and (b) both intrinsic and acquired expression of Pgp in breast cancer may contribute in part to therapeutic failure and relapse.

Evaluation of newer prognostic markers for adult soft tissue sarcomas

PURPOSE

In addition to tumor size, grade, location, and the presence of metastases, other factors may be useful in prognostication for adults with soft tissue sarcoma (STS). This study examines the relationship of MDR-1 mRNA, p-glycoprotein (P-gp), Ki-67 expression, and DNA content expression to clinical outcome in adults with STS.

PATIENTS AND METHODS

Snap-frozen STS specimens from 65 patients were analyzed and compared with clinical outcomes. Immunohistochemistry was performed for the Ki-67 antigen and P-gp. DNA content was determined using the Feulgen reaction and quantitated using image analysis. MDR-1 mRNA expression was determined using a reverse-transcriptase polymerase chain reaction (RT-PCR)-based assay.

RESULTS

P-glycoprotein expression was found by immunohistochemistry in 48% of cases with 5-year overall (54% v 14%, P = .07) and disease-free survival rates (32% v 18%, P = .039) higher in high-grade tumors that did not express P-gp. MDR-1 mRNA was detected in 51% of cases and no patient with high levels of MDR-1 mRNA expression was a long-term survivor. Patients with diploid tumors had significantly better survival than those with nondiploid tumors (51% v 31%, P = .03). High levels of Ki-67 were associated with poorer overall survival (46% v 31%, P = .04). On multivariate analysis, American Joint Committee on Cancer (AJCC) staging, DNA content, Ki-67, and P-gp staining were significant prognostic factors for 5-year overall and disease-free survival.

CONCLUSION

P-gp expression, high-level Ki-67 expression, and nondiploid DNA content are independent prognostic indicators that correlate with poor outcomes in STS patients. However, MDR-1 mRNA was not found to be predictive of survival. These newer markers are useful additions to AJCC staging for prognostication for patients with STS. Such markers may be useful in selecting high-risk STS patients who could benefit from systemic adjuvant therapy.

Expression of Fas/CD95 and Bcl-2 by primitive hematopoietic progenitors freshly isolated from human fetal liver

The cell-surface expression and the functional status of the CD95/Fas antigen on primitive hematopoietic progenitors (PHPs) freshly isolated from human fetal liver (FL) were studied. PHPs were phenotypically defined as CD34++ CD38 -/+ cells. The most immature subfractions of PHPs, CD34++CD38- and CD34+2CD38+ FL cells, expressed CD95, whereas the more mature CD34++CD38++ and CD34+CD38++2 FL cells displayed low CD95 expression. Combinations of cytokines, such as kit ligand (KL) + interleukin-3 or KL + granulocyte-macrophage colony-stimulating factor (GM-CSF) upregulated the expression of CD95 on PHPs upon in vitro culture. Tumor necrosis factor-alpha (TNF-alpha) and interferon-gamma (IFN-gamma) further increased the CD95 expression induced by KL+GM-CSF. The hematopoietic potential of sorted CD34++lineage (lin)- CD95+ versus CD34++ lin-CD95-FL cells was compared by colony-forming unit-culture (CFU-C) assays performed in serum-deprived medium. Lin+ cells were composed of erythrocytes, monocytes, T cells, B cells, and natural killer cells. Our results indicated that both CD95- and CD95+ subsets contained pluripotent progenitors, generating myeloid and erythroid progenitors. The functional status of CD95 and the effects of TNF-alpha and IFN-gamma, cytokines known to induce CD95-mediated apoptosis, were analyzed by incubation of PHPs in the presence of anti-CD95 monoclonal antibodies (MoAbs). The effect of anti-CD95 MoAbs was measured by viable cell counting, flow cytometry, and CFU-C assays. A decrease of CFU-C numbers was observed in the presence of anti-CD95 MoAbs and TNF-alpha and/or IFN-gamma. However, whereas growth factor deprivation induced apoptosis of PHPs, cross-linking of CD95 did not lead to apoptosis of PHPs measured by flow cytometry and viable cell counting. The correlation of increased intracytoplasmic levels of bcl-2 with high levels of cell-surface CD34 and the presence of CD95 on fresh FL cells suggests that bcl-2 may be involved in protecting against CD95-mediated apoptosis of FL PHPs.

Chemotherapy resistance to taxol in clonogenic progenitor cells following transduction of CD34 selected marrow and peripheral blood cells with a retrovirus that contains the MDR-1 chemotherapy resistance gene

A retrovirus containing the multiple drug resistance (MDR-1) cDNA, was used to transduce cultures of CD34 selected human marrow cells, on stromal monolayers in the presence of hematopoietic growth factors IL-3 and IL-6, following collection from patients recently recovered from chemotherapy-induced myelosuppression. In one experiment, these CD34 selected cells were grown in Dexter cultures for 35 days or more following MDR-1 transduction, and then plated in methylcellulose. Polymerase chain reaction (PCR) analysis of colonies picked after 10-14 days of methylcellulose culture, using a set of primers that are specific for the endogenous or the retrovirally transduced MDR-1, showed that the long-term culture initiating cells (LTCICs) were transduced by the MDR-1 virus. Analysis of the colonies from the CD34 selected MDR-1 transduced cells, with a reverse transcription (RT) PCR assay that could distinguish viral MDR-1 mRNA from endogenous MDR-1 mRNA, showed that the viral MDR-1 mRNA levels were much higher than that of the MDR-1 mRNA from the endogenous MDR-1 gene in the transduced CD34 selected cells. Fluorescence activated cell sorting (FACS) analysis of the CD34 selected transduced marrow cells within 48 h after the transduction, using the C219 and UIC2 monoclonal antibodies for p-glycoprotein, showed that the transduction frequency under these conditions varied from 7 to 20%. Rhodamine efflux studies showed that this additional p-glycoprotein was functional and that the frequency of cells with high p-glycoprotein levels was higher in the transduced cells than in the non-transduced cells. The resistance to taxol of the CD34 selected transduced cells, as judged by the plating efficiency of clonogenic progenitor cells derived from these cells by growth in methylcellulose supplemented with taxol was much higher in the transduced cells than in untransduced cells. In order to test the reproducibility of the transduction frequency of the retroviral supernatants from PA317 MDR-1 viral producer cells on CD34 selected cells, the virus produced from 12 different lots of supernatants from the PA317 MDR-1 producer cell line was used to transduce CD34 selected marrow cells from four different patients, and to transduce the peripheral blood cells of two additional patients collected following chemotherapy-induced myelosuppression. The supernatant lots used for these transduction experiments were tested by Microbiological Associates (Rockville, MD, USA), by the Mus dunni co-cultivation and amplification tests in the S+L-assay and found to be negative for replication-competent retrovirus, and later approved for human use by the Food and Drug Administration.(ABSTRACT TRUNCATED AT 400 WORDS)