Second-line chemotherapy with pemetrexed after gemcitabine failure in patients with advanced pancreatic cancer: a multicenter phase II trial.

BACKGROUND: A standard second-line chemotherapy regimen has yet to be defined for patients with gemcitabine (Gem)-refractory advanced pancreatic cancer (PC). PATIENTS AND METHODS: In this multicenter phase II trial, patients with unresectable or metastatic PC who had progressed on single-agent Gem or a Gem-containing regimen received pemetrexed 500 mg/m(2) as a 10-min infusion every 3 weeks until disease progression or occurrence of unacceptable toxicity. The primary end point was the 3-month survival rate. RESULTS: A total of 192 treatment cycles were given to 52 patients. The overall response rate was 3.8% (two partial responses); 10 patients (19.2%) experienced stable disease, nine of them for >12 weeks. At least one CA 19-9 reduction > or =50% occurred in 12 patients (23.1%). The 3-month survival rate was 75% (95% confidence interval 63.2% to 86.8%), the median time to tumor progression was 7 weeks (range 1-62 weeks) and the median overall survival time was 20 weeks (range 1-84 weeks). Grade 3/4 hematological toxic effects included (percent of patients): neutropenia (17.3%), thrombocytopenia (5.8%) and anemia (3.8%). The most frequent non-hematological toxic effects were diarrhea, nausea and stomatitis/pharyngitis (23.1% each). CONCLUSION: Pemetrexed is a safe treatment option with moderate activity in patients with advanced PC after failure of Gem.

In vitro activity of liposomal N4octadecyl-1-beta-D-arabinofuranosylcytosine (NOAC), a new lipophilic derivative of 1-beta-D-arabinofuranocylcytosine on biopsized clonogenic human tumor cells and hematopoietic precursor cells.

N4-octadecyl-1-beta-D-arabinofuranosylcytosine (NOAC) is a new lipophilic derivative of 1-beta-D-arabinofuranosylcytosine (ara-C) with potent antitumor activity against leukemias and solid tumors. In this study the activity of NOAC against freshly explanted clonogenic cells from human tumors was determined and compared with conventional antitumor agents. NOAC was used in two liposomal preparations, a stable lyophilized and a freshly prepared liquid formulation. Both formulations inhibited tumor colony formation equally in a concentration-dependent fashion in both short- (1 h) and long-term (21-28 d) exposure experiments. NOAC (100 microM, long-term exposure) had a significantly better activity compared to the clinically used drugs cisplatin, doxorubicin, 5-fluorouracil, gemcitabine, mitomycin C and etoposide. The comparison of NOAC with ara-C in the long-term exposure experiment showed that ara-C was more effective at 4 and 10 microM, whereas at 1 and 100 microM there was no difference between the two drugs. NOAC was less toxic in a hematopoietic stem cell assay than ara-C and doxorubicin by factors ranging from 2.5 to 200, indicating that this drug is well tolerated at high doses. The antitumor activity of NOAC (NSC 685096) was confirmed by the NCI in vitro drug screening program where the drug was found to be active against several types of human tumors. Further development of NOAC in phase II studies is warranted.

Preliminary results of a phase I study with MTA (LY231514) in combination with cisplatin in patients with solid tumors.

MTA (multitargeted antifolate, LY231514) is a novel antimetabolite resulting from structure/activity studies of the lometrexol-type antifolates. It has been shown to inhibit various enzymes of folate pathways and has broad antitumor activity in a variety of in vitro and in vivo tumor models. Clinical phase 1 studies have been performed using different administration schedules, and subsequently the every-21-days schedule has been selected for further development. We report the preliminary findings from a combination phase I study of MTA and cisplatin administered every 21 days. In the first cohort (34 patients), both agents were administered on day 1 with a starting dose of 300 mg/m2 MTA and 60 mg/m2 cisplatin. In a second cohort (10 patients), MTA (500 or 600 mg/m2) was administered on day 1 followed by cisplatin (75 mg/m2) on day 2. The maximum tolerated doses were reached at 600 mg/m2 MTA/100 mg/m2 cisplatin (cohort 1) and 600 mg/m2 MTA/75 mg/m2 cisplatin (cohort 2). In cohort 1, dose-limiting toxicities consisted of reversible myelosuppression with leukopenia and neutropenia. In addition, delayed fatigue also was of clinical significance. Pharmacokinetic analyses indicated that hydration administered before the administration of cisplatin did not influence the major pharmacokinetic parameters of MTA. Eleven objective remissions were observed, including one complete response in a patient with relapsed squamous cell carcinoma of the head and neck and partial responses in four of seven patients with mesothelioma In contrast, the dose-limiting toxicities in patient cohort 2 consisted of neutropenic sepsis, diarrhea, and skin toxicity with two possibly treatment-related deaths on study. No objective remissions are presently observed in cohort 2. We conclude that the combination of MTA and cisplatin is feasible and clinically active when both agents are administered on day 1 and that it should be pursued for further clinical development.

Clinical and pharmacokinetic phase I study of multitargeted antifolate (LY231514) in combination with cisplatin.

PURPOSE: Multitargeted antifolate (MTA; LY231514) has broad preclinical antitumor activity and inhibits a variety of intracellular enzymes involved in the folate pathways. This study was designed to (1) determine the maximum-tolerated dose (MTD), dose-limiting toxicities (DLT), and pharmacokinetics of MTA combined with cisplatin; (2) determine a recommended dose for phase II studies; and (3) collect anecdotal information on the antitumor activity of MTA combined with cisplatin. PATIENTS AND METHODS: Patients with solid tumors received MTA intravenously over 10 minutes and cisplatin over 2 hours once every 21 days. In cohort 1, both agents were administered on day 1 starting with MTA 300 mg/m(2) and cisplatin 60 mg/m(2). In cohort 2, MTA (500 or 600 mg/m(2)) was administered on day 1, followed by cisplatin (75 mg/m(2)) on day 2. RESULTS: In cohort 1, 40 assessable patients received 159 courses of treatment. The MTD was MTA 600 mg/m(2)/cisplatin 100 mg/m(2). DLTs were reversible leukopenia/neutropenia and delayed fatigue. Hydration before cisplatin therapy did not influence MTA pharmacokinetics. Eleven objective remissions included one complete response in a patient with relapsed squamous cell head and neck carcinoma, and partial responses in four of ten patients with epithelial pleural mesothelioma. In cohort 2, 11 assessable patients received 23 courses of treatment. The MTD was MTA 600 mg/m(2) and cisplatin 75 mg/m(2). DLTs were neutropenic sepsis, diarrhea, and skin toxicity. Two patients died of treatment-related complications during the study. Two patients had objective remissions (one mesothelioma patient, one colon cancer patient). CONCLUSION: The combination of MTA and cisplatin is clinically active, and administering both agents on day 1 is superior to a split schedule. Further development of this combination for mesothelioma is warranted.

Preclinical activity of 17 beta-[N-[N'-(2-chloroethyl)-N'-nitrosocarbamoyl]-L-alanyl]-5 alpha-dihydrotestosterone (E91) against tumour colony forming units and haematopoietic progenitor cells.

E91 (17 beta-[N-[N'-(2-chloroethyl)-N'-nitrosocarbamoyl]-L-alanyl]-5 alpha-dihydrotestosterone) (CNC-ala-DHT) is a newly synthesised alkylating compound consisting of N-[N'-(2-chloroethyl)-N'-nitrosocarbamoyl]-L-alanine (CNC-ala) as the alkylating moiety and of 5 alpha-dihydrotestosterone (DHT) as a steroid carrier molecule. We studied the antitumour activity of E91 (final concentrations: 0.1, 1, 10 and 30 mumol/l) against freshly explanted human tumours, using an in vitro soft agar cloning system. A total of 54 tumour samples was evaluated using 1 h-exposure and 51 tumour specimens were studied using a continuous exposure for 21-28 days. In addition, the compound's activity was compared with other clinically used anticancer agents. After short-term exposure, 49 of 53 evaluable specimens (92%) had adequate colony formation, as compared with 49 of 50 (98%) after long-term exposure. After short-term exposure, E91 exhibited only marginal antitumour activity. However, in long-term exposure experiments, E91 had marked and concentration-dependent antitumour activity (P < 0.001). At concentrations of > 10 mumol/l, E91 was as active as the other clinically used antineoplastic agents and at 30 mumol/l, E91 was significantly more active than 5-fluorouracil (P = 0.041). E91 showed activity against a wide spectrum of tumour types. The highest activity was observed against colorectal carcinomas (3/4 tumour specimens inhibited at 30 mumol/l). Sensitivity was also high remarkable in breast cancer specimens with 3/6 specimens inhibited at 30 mumol/l. In vitro myelotoxicity was less than that of doxorubicin. At 30 mumol/l, E91 induced a reduction of colony forming units-granulocyte macrophage (CFU-GM) to only 53% of control and of CFU-GEMM to 20% of control. We conclude that because of broad activity and reduced myelotoxicity further clinical development of E91 appears warranted.

In vitro activity of aplidine, a new marine-derived anti-cancer compound, on freshly explanted clonogenic human tumour cells and haematopoietic precursor cells.

Aplidine is a new marine anti-cancer depsipeptide isolated from the Mediterranean tunicate Aplidium albicans. We have evaluated its antiproliferative action against a variety of freshly explanted human tumour specimens. Concentration ranges of 0.01-1.0 microM and 0.0001-1.0 microM were used in short- and long-term exposure schedules respectively. After exposure for 1 h in 49 evaluable specimens, aplidine showed a clear concentration-dependent anti-tumour effect. At 0.05 microM, 85% of the specimens were markedly inhibited. Continuous exposure for 21-28 days in 54 tumour specimens also led to a concentration-dependent activity relationship. Fifty per cent and 100% tumour inhibitions were achieved with 0.001 microM and 0.05 microM respectively. A head to head evaluation assessing short vs continuous exposure was carried out, resulting in evidence of an activity-time of exposure relationship. Breast, melanoma and non-small-cell lung cancer appear to be sensitive to low concentrations of aplidine. In addition the evaluation of the effects of aplidine on haematopoietic cells showed a concentration-dependent toxicity. However, under continuous exposure, active concentrations induced mild bone marrow toxicity, indicating that a therapeutic window at marginally myelotoxic concentrations might exist.

Dose intensity phase I/II trial with carboplatin, ifosfamide, etoposide and vincristine combined with filgrastim in patients with small-cell lung cancer.

BACKGROUND: The purpose of the study was to evaluate the feasibility of increasing dose intensity by a stepwise reduction of the time intervals between chemotherapy cycles in separate patient cohorts with small-cell lung cancer. Patients received up to 6 courses of combination chemotherapy with carboplatin, etoposide, ifosfamide and vincristine followed by support with filgrastim. Dose intensity, incidence, duration and severity of neutropenic fever and infections, objective response to chemotherapy, and safety of filgrastim were determined. PATIENTS AND METHODS: 29 patients with small-cell lung cancer (limited disease: 2, extensive disease: 27) were treated with a combination of carboplatin 250 mg/m2 i.v. day 1, ifosfamide 2 g/m2 and etoposide 120 mg/m2 i.v. days 1 and 2, etoposide 120 mg/m2 orally day 3, and vincristine 1.4 mg/m2 day 14. Initially, filgrastim (5 micrograms/kg) was administered subcutaneously from day 7 to 16. With shorter treatment intervals, filgrastim was administered on days 4-16 or 4-14. RESULTS: An overall increase in dose intensity by a factor of 1.44 was achieved after reducing the treatment interval from 27 to 17 days. Further reduction to 14 days was not feasible due to persistent thrombocytopenia. Six patients (21%) developed a total of 9 febrile episodes, and 14 patients (48%) had to be withdrawn from the study before the completion of six cycles of chemotherapy. The median duration of infectious episodes was 6 days. Overall, a total of 22 of 27 evaluable patients had an objective response. Longer treatment intervals resulted in a lower probability for objective response (> or = 23 days: 10/14 patients vs. < or = 17 days: 7/7 patients). CONCLUSION: Filgrastim allows for the reduction of treatment intervals in patients with small-cell lung cancer and increased dose intensity with acceptable hematologic and nonhematologic toxicities.

Gemcitabine and etoposide in small cell lung cancer: phase I and II trials.

Gemcitabine and etoposide have both shown single-agent activity against multiple tumor types in clinical trials, including small cell lung cancer, but have not been previously used together. Forty-four patients with small cell and non-small cell lung cancer or other tumor types were enrolled in a phase I dose-finding trial using this drug combination. Gemcitabine 1,000 mg/m2 was given intravenously on days 1, 8, and 15 of a 28-day cycle, and etoposide (dose escalated from 20 to 80 mg/m2) was given on days 8, 9, and 10. Leukopenia, thrombocytopenia, and anemia were the major toxicities noted. Objective responses were observed in five of 44 patients. The maximum tolerated dose of etoposide was determined to be 80 mg/m2. On the basis of these results, a phase II trial of gemcitabine and etoposide in patients with small cell lung cancer has been initiated. Twelve patients have been enrolled in this ongoing trial, and toxicity to date has been manageable.

Preclinical activity of trans-indazolium[tetrachlorobisindazoleruthenate(III)] (NSC 666158; IndCR; KP 1019) against tumour colony-forming units and haematopoietic progenitor cells.

Trans-indazolium[tetrachlorobisindazoleruthenate(III)] (KP 1019) is a new heavy metal complex with promising activity against tumour cell lines and in animal models. We studied the antineoplastic effects of KP 1019 (final concentrations: 1, 10, 100 micrograms/ml) on in vitro proliferation of clonogenic cells from freshly explanted human tumours in a capillary soft agar cloning system, and compared the activity of KP 1019 with conventional antineoplastic agents. 53 of 75 specimens (71%) showed adequate growth in controls. KP 1019 inhibited tumour colony formation in a concentration-dependent manner in both short- (1 h) and long-term (21 d) exposure experiments. KP 1019 at 100 micrograms/ml with 1 h exposure was as active as bleomycin, cisplatin, doxorubicin, etoposide, 5-fluorouracil, methotrexate, mitomycin-C and vinblastine, with only paclitaxel more active than KP 1019 (P = 0.002). The antitumour activity of KP 1019 was more pronounced after long-term exposure, indicating the potential schedule dependency of KP 1019. Activity was observed against non-small cell lung, breast and renal cancer. We conclude that if appropriate plasma levels can be achieved in patients, KP 1019 may have significant clinical activity against a variety of different tumour types.

Efficacy of 5'-nor-anhydrovinblastine (vinorelbine), against freshly explanted clonogenic human tumor cells in vitro.

Vinorelbine (5'-nor-anhydrovinblastine) is a semisynthetic vinca alkaloid currently undergoing extensive clinical evaluation. We have studied the antitumor effect of vinorelbine (final concentrations: 8.4-1000.0 ng/ml) against freshly explanted clonogenic cells from 102 human tumors using a capillary soft agar cloning system and have compared the compound's activity with vinblastine, vincristine, vindesine, paclitaxel, docetaxel, and other clinically used anticancer agents. Four specimens were excluded from further analyses (3 bacterial or fungal contamination, 1 benign histology). Fifty-one of the remaining 98 (52%) specimens had adequate colony formation in control capillaries. Vinorelbine showed concentration-dependent antitumor activity against a variety of solid tumors. At clinically relevant concentrations (0.1 x peak plasma concentrations in humans) vinorelbine inhibited 21 of 49 specimens (43%) and was as active as vinblastine, vincristine, vindesine, bleomycin, doxorubicin, 5-fluorouracil, mitomycin-C, cisplatin, methotrexate, and etoposide. However, paclitaxel (71% inhibition, p = 0.006) and docetaxel (78% inhibition, p = 0.002) were significantly more active than vinorelbine. Moreover, vinorelbine showed antitumor activity against several tumor types and in particular against breast cancer but also in non-small cell lung cancer. We conclude that vinorelbine has a wide spectrum of in vitro activity against freshly explanted human tumors and that the clinical activity of this compound against breast cancer and non-small cell lung cancer is reflected in vitro.

Characterization and mutual ligand-induced modulation of epidermal growth factor and interferon-alpha receptors on renal carcinoma cells in vitro.

We have determined the binding of epidermal growth factor (EGF) and interferon (IFN)-alpha to their specific receptors on four renal carcinoma cell lines. CaKi-2, a-498 and ACHN cell lines express high numbers, and CaKi-1 expresses low number of EGF receptors (EGFRs). On all four renal carcinoma cell lines, we have also detected specific IFN-alpha binding sites. EGF and IFN-alpha binding to their receptors caused modulation of the other ligand's receptor binding. Scatchard analyses of binding data showed that IFN-alpha treatment leads to an increase of EGFR number in three out of four cell lines and to a decrease of EGFR number in one out of four (CaKi-1). No significant changes in EGF binding affinities were detected. EGF induced a reduction in IFN-alpha receptor number in all four cell lines without significant changes in IFN-alpha binding affinities. We hypothesize that presence of EGF in the microenvironment of renal cancer cells may modulate the biological effects of IFN and consequently decrease its antiproliferative activity.

Activity of NK 611, a new epipodophyllotoxin derivative, against colony forming units from freshly explanted human tumours in vitro.

NK 611 is a new semisynthetic analogue of etoposide, which presumably also acts through inhibition of topoisomerase II, and has been found to be more potent against several cancer cell lines in vitro than etoposide. The objectives of our study were to determine the activity of NK 611 against freshly explanted clonogenic cells from human tumours and compare this agent with etoposide and other clinically useful agents. After exposure for 1 h in 45 evaluable tumour specimens, NK 611 showed clear concentration-dependent antitumour activity. At 51 microM, 49% of specimens were markedly inhibited. Using a long-term (21-28 day) exposure at 6.8 microM, 58% of 50 evaluable specimens were profoundly inhibited. At equimolar concentrations, NK 611 was as active as etoposide. Across all tumour types studied, NK 611 was as active as vinblastine, bleomycin, doxorubicin, 5-fluorouracil, mitomycin-C and cisplatin. Our results showed cross resistance to etoposide in the majority of specimens. Activity of NK 611 was greater with long-term exposure than with short-term exposure indicating schedule dependency. We conclude that NK 611 has a wide spectrum of in vitro antitumour activity. Since preliminary clinical information suggests that this drug is well tolerated at high doses, further development of this agent in Phase II trials with multiple dosing schedules is warranted.

Interleukin 1 modulates growth of human renal carcinoma cells in vitro.

We have investigated the influence of interleukin 1 (IL-1) on growth of human renal carcinoma cells in vitro. Using a capillary soft-agar cloning system, 18% of freshly explanted renal carcinomas were stimulated to grow by IL-1 and 4% were inhibited. Subsequent experiments with established renal cancer cell lines demonstrated that two out of four cell lines (Caki-2, A-498) were sensitive to IL-1. [3H]Thymidine incorporation as well as monolayer growth was enhanced in Caki-2 cells in the presence of high (10%) and low (1%) serum concentrations. Although clonogenic growth of A-498 cells was stimulated by IL-1, overall [3H]thymidine incorporation and monolayer proliferation were decreased. Using radioligand experiments, 250 cell-surface receptors of high affinity (KD 4.5 x 10(-11) M) and 2500 receptors of low affinity (KD 1.3 x 10(-9) M) were detected on A-498 cells. IL-1 binding was reduced under the influence of IL-1. Competition experiments with inhibiting antibodies against IL-1 receptor type I and type II revealed that signal transduction was performed via type I receptors. After cross-linking to IL-1, receptor type I was immunoprecipitated using anti-IL-1 antibodies. We hypothesise that, since IL-1 modulates in vitro growth of a subgroup of human renal cancer cells, interference with its mechanism of action may be of potential value in order to modulate tumour proliferation.

Investigation of the comparative effects of 2-chlorodeoxyadenosine on tumor colony forming units in vitro.

2-CdA is a deaminase-resistant purine analogue which has shown clinical activity against various hematological tumors, and is currently undergoing clinical phase II trials. The objectives of our study were to determine the activity of 2-CdA against freshly explanted clonogenic cells from non-hematological human tumors and compare this agent with other clinically useful anticancer agents. We also compared short-term (1 hour) and long-term (21-28 days) exposures. For short-term exposure (1-hour), final concentrations were 0.57, 5.7, 57, and 114 ng/ml. Inhibition of tumor specimens was concentration-dependent: 0.57 ng/ml: 1/51 (2%), 5.7 ng/ml: 4/52 (7%), 57 ng/ml: 11/52 (21%), 114 ng/ml: 27/50 (54%). At concentrations > or = 57 ng/ml, 2-CdA was as active as cisplatin, doxorubicin, 5-fluorouracil, mitomycin-C, vinblastine, and etoposide. For long-term exposures (21-28 days), final concentration of 2-CdA were 0.57, 5.7, and 57 ng/ml. At 0.57 ng/ml, 2-CdA was active in 4/54 (7%) specimens [5.7 ng/ml: 13/54 (24%), 57 ng/ml: 40/54 (74%)]. A head-to-head comparison with short-term exposures demonstrated greater activity if the drug exposure time was extended. Using the strategy for testing other standard agents (in vitro dose of 1/10th achievable peak plasma concentration), one would predict clinical response rates for single agent bolus or short-term administration of 2-CdA to be in the neighborhood of 7%. Longer durations of infusion or multiple doses might increase the response rate to about 24%. If higher peak plasma concentration could be achieved, dose-dependent increases in clinical responses might be achievable. We conclude that 2-CdA is active against clonogenic cells from freshly explanted non-hematological human tumor specimens at high concentrations.

Effects of vinorelbine on epidermal growth factor-receptor binding of human breast cancer cell lines in vitro.

Epidermal Growth Factor (EGF) is a mitogenic peptide that binds to surface membrane receptors (EGFR) of breast cancer cells. After binding, secondary transmitter molecules are activated by tyrosine phosphorylation of the intracellular receptor domaine. The activity of the EGF/EGFR system can be modulated by a variety of chemically unrelated compounds including cytostatic agents. The purpose of our present study was to determine the effects of vinorelbine, a novel semisynthetic vinca alkaloid on EGF receptor binding on human breast cancer cells. We have found that MDA-231 and MDA-468 cells bind substantially more [125I]-EGF after preincubation with vinorelbine. This effect was concentration- and time-dependent reaching a maximum at 100 ng/ml and 24 h incubation. Subsequent experiments showed an increase in the rate of EGF binding as well as maximal binding capacity. Scatchard analysis of binding experiments under equilibrium conditions indicated that this was mainly due to an increase in the number of apparent EGF binding sites. Modulation of EGF receptor binding by vinorelbine was not detectable when isolated membranes were used indicating that intact cytoplasmatic mechanisms are required for the upregulation of EGF receptors.

Effects of the microtubule-disturbing agents docetaxel (Taxotere), vinblastine and vincristine on epidermal growth factor-receptor binding of human breast cancer cell lines in vitro.

Epidermal growth factor (EGF) is a mitogenic peptide that binds to surface membrane receptors (EGFR) of breast cancer cells. After binding, secondary transmitter molecules are activated by tyrosine phosphorylation of the intracellular receptor domaine. The activity of the EGF/EGFR system can be modulated by a variety of chemically unrelated compounds including cytostatic agents. The purpose of our present study was to determine the effects of mitotic inhibitors on EGF receptor binding on human breast cancer cells. We found that MDA-231 and MDA-468 cells bind substantially more [125I]EGF after preincubation with docetaxel, vinblastine and vincristine. This effect was concentration- and time-dependent, reaching a maximum at 3000 ng/ml and 48 h incubation for docetaxel, and 100 ng/ml and 48 h incubation for vinca alcaloids. Subsequent experiments showed an increase in the rate of EGF binding as well as maximal binding capacity. Scatchard analysis of binding experiments under equilibrium conditions indicated that this was due to an increase in the number of apparent EGF binding sites. Modulation of EGF receptor binding by docetaxel, vinblastine, and vincristine was not detectable when isolated membranes were used, indicating that intact cytoplasmatic mechanisms are required for the upregulation of EGF receptors.

Preclinical activity of taxotere (RP 56976, NSC 628503) against freshly explanted clonogenic human tumour cells: comparison with taxol and conventional antineoplastic agents.

Taxotere (TER) and taxol (TA) are new antitumour agents currently undergoing clinical evaluation. We studied the antineoplastic effects of these agents (final concentrations: 4.0, 0.4, 0.04 mumol/l) on the in vitro proliferation of clonogenic cells from freshly explanted human tumours using a capillary soft agar cloning system. We also compared the activity of these new compounds to conventional antineoplastic agents (bleomycin, cisplatin, dacarbazine, doxorubicin, etoposide, 5-fluorouracil, vinblastine, interferon-alpha 2). Using a 21-28-day continuous drug exposure, 54/81 specimens (67%) were evaluable for comparisons, and using a 1-h drug exposure followed by 21-28 days incubation, 50/80 specimens (63%) were similarly evaluable. With both schedules, TA and TER showed concentration-related antitumour activity. At 0.4 mumol/l, median colony survival was 0.61 x control (range 0.09-0.96) for TA and 0.51 x control (0.15-0.81) for TER in the 1-h incubation (P = 0.0002). Median colony formation was also reduced significantly more by TER as compared to TA in the long-term incubation schedule. Statistical analysis indicated that TER but not TA was significantly more active than cisplatin (P = 0.02), doxorubicin (P = 0.01), 5-fluorouracil (P = 0.01) and interferon-alpha 2 (P = 0.01). We conclude that TER and TA are more active against in vitro tumour colony formation from freshly explanted human tumours. TER appears to be slightly more active than taxol and promises to be active against tumours resistant to conventional antineoplastics.

Effects of cytokines on in vitro colony formation of primary human tumour specimens.

Although under study to alleviate chemotherapy-induced bone marrow toxicity, cytokines can stimulate in vitro growth of solid human tumour cell lines. The effects of granulocyte-macrophage colony-stimulating factor (GM-CSF), granulocyte colony-stimulating factor (G-CSF) and interleukin-3 (IL-3) on in vitro colony formation of primary human tumours was studied in a capillary soft-agar cloning system. Of 108 tumour specimens from 100 patients, 85 specimens were tested against all three factors at concentrations ranging from 0.1 to 1000 ng/ml. 44 of 100 tumours showed adequate growth in controls. 8 out of 43 (19%) specimens were significantly stimulated by GM-CSF, 6 of 40 (15%) by G-CSF and 10 of 44 (23%) by IL-3. Sensitivity to all three cytokines was observed in 4 of 44 (9%) specimens. By light microscopy the appearance of colonies from stimulated specimens was identical to that of controls. Sensitivity to cytokines was independent from sensitivity to epidermal growth factor, transferrin or insulin. Sensitivity to GM-CSF, G-CSF and IL-3 may be aberrantly expressed in a subgroup of solid human tumours.