Three-dimensional computer-aided design based design sensitivity analysis and shape optimization of the stem using adaptive p-method.

The number of stem designs for total hip arthroplasty is increasing, and occasionally design changes have yielded unexpected clinical results. At present, we are not able to clearly identify which parameter of the stem is most important, and the optimum value of many parameters. The goals of this study were to identify which parameter is most important, to understand the effect of design change, and to find the optimum stem shape. For this purpose, we used adaptive p-method together with three-dimensional computer-aided design software program for the design sensitivity analysis (DSA) and shape optimization of the stem. The results suggested that increasing the lateral and medial width of the distal cross-section together with decreasing the medial-lateral width and the medial radius of the distal cross-section from the default value would lead to a decrease in the largest maximum principal stress of the distal cement. The medial width of middle cross-section, however, was not so simple. The result of DSA suggested that decreasing this parameter from the default value decreased the stress in the distal cement, but the optimum shape was obtained by increasing this parameter. The method used in this study will assist our engineers and surgeons in the process of modifying and optimizing the stem design.

Dual modulation by l-leucovorin and recombinant human interferon alpha 2a of 5-fluorouracil antitumor activity against the human colon carcinoma xenograft Co-4.

We investigated the dual modulation by l-leucovorin (LV) and recombinant human interferon-alpha 2a (IFN-alpha 2a) of 5-fluorouracil (5-FU) antitumor activity against human colon carcinoma cells (Co-4) using a nude mouse system. 5-FU was administered intraperitoneally (IP) at 10 or 90 mg/kg. 5-FU (10 mg/kg) was administered daily for 10 days, and 90 mg/kg was administered once. LV was administered IP 1 and 0 h before 5-FU treatment at 200 mg/kg. IFN-alpha 2a was administered subcutaneously (SC) daily for 14 days at 60,000 IU/mouse. When 5-FU was administered at 10 or 90 mg/kg with these two modulators, the antitumor effect was increased significantly, with T/C ratios of 18.1 and 6.1, respectively. These modulatory effects were assessed as synergistic, without associated severe side effects or death during the experimental period. LV augmented the antitumor activity of 5-FU through increment of thymidylate synthetase (TS) inhibition, and IFN-alpha 2a showed a modulatory effect in elevating the intratumoral concentration of fluorouridine without change in TS inhibition. These results suggest that 5-FU antitumor activity against human colon carcinoma could be significantly potentiated without severe side effects by these two modulators, which possess different modes of action.

Differential subcellular localization of caspase family proteins in the adult rat brain.

The distribution of the caspase family (caspase-2, -3, -6, -7, -8, -9, -10) was assessed using immunochemical detection of subcellular fractions of 8-week-old rat brain tissues. The present study demonstrated that the relative protein level of caspase-2, -3, -6, -8 and -10 was highest in the soluble cytosolic fraction, while that for caspase-9 was highest in the nucleus. We also found that caspase-3 and -6 were present at high levels and caspase-2, -8 and -9 at moderate levels in the nerve endings fraction as well as in the soluble cytosolic fraction. These results suggest that rat brain caspases are differentially expressed in the subcellular fractions of the rat brain, and that caspases not only contribute to the regulation of neuronal death, but also to synaptic plasticity.

Alteration of myo-inositol monophosphatase in Alzheimer's disease brains.

myo-Inositol monophosphatase (E.C.3.1.3.25) catalyzes the hydrolysis of myo-inositol 1-phosphate in the presence of Mg2+ at a physiologic pH to form free myo-inositol, maintaining a supply that represents the precursor for inositol phospholipid second messenger signaling systems. In the present study the activity and protein level of myo-inositol monophosphatase were investigated in samples from normal human and Alzheimer's disease (AD) postmortem brains. The separation profile on Sephadex G-100 gel filtration chromatography revealed one major form of myo-inositol monophosphatase in crude extracts from both normal human and AD brains. In AD brains myo-inositol monophosphatase activity and its protein level were significantly higher than in control brains. The activity of myo-inositol monophosphatase per enzyme molecule was similar in control and AD brains. These results suggest that myo-inositol monophosphatase is upregulated in AD, probably reflecting compensatory mechanisms concerned with phospholipid metabolism.

Antitumor activity of hexamethylmelamine on human tumor xenografts serially transplanted in nude mice.

The antitumor activity of hexamethylmelamine (HMM) was evaluated using four human tumor xenografts serially transplanted in nude mice. HMM was dissolved in 0.2 ml of 1% hydroxypropyl cellulose per mouse and administered perorally daily, except on Sunday, for 4 weeks, giving an estimated maximum tolerated dose (MTD) of HMM of 75 mg/kg. The MX-1 cell line showed dose-dependent sensitivity to HMM and was completely eradicated by treatment at the MTD. The minimum effective dose of HMM against MX-1 was calculated to be 22.1 mg HMM/kg, resulting in the chemotherapeutic index of 3.4. The demethylated derivatives of HMM, pentamethylmelamine and tetramethylmelamine, were also effective against MX-1, whereas trimethylmelamine was ineffective. The effect of HMM was more marked when the drug was administered on day 1 after tumor inoculation, compared with administration during the exponential growth phase. HMM is thought to be a promising agent for the treatment of several types of human carcinoma, producing active metabolites in vivo after peroral administration.

Interferon beta increases antitumor activity of 5-fluorouracil against human colon carcinoma cells in vitro and in vivo.

The modulating effect of human fibroblast-derived interferon beta (IFN-beta) on the antitumor effect of 5-fluorouracil (5-FU) against human colon carcinoma cells in vitro and in vivo was investigated. The 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl-2H tetrazolium bromide (MTT) assay was carried out in vitro using the cultured human colon cancer cell line C-1. IFN-beta at concentrations of 50, 500, 5,000 and 50,000 IU/ml was added to the cultured tumor cells with or without 5-FU at concentrations of 10, 50 and 500 micrograms/ml. The antitumor activity of 5-FU with or without IFN-beta was assessed using Co-4, a human colon carcinoma xenograft in nude mice, with reference to thymidylate synthetase inhibition. IFN-beta was administered subcutaneously daily for 14 days at doses of 6,000, 60,000 and 600,000 IU/mouse. The combined antitumor effect with 5-FU was evaluated by simultaneous intraperitoneal administration of 5-FU at doses of 10 and 20 mg/kg daily for 10 days. The antitumor activity of IFN-beta alone increased in a dose-dependent manner against Co-4 in nude mice, whereas its antitumor activity in vitro against C-1 was limited. The synergistic effect of 5-FU and IFN-beta was observed both in vitro and in vivo, and the in vivo synergism was obtained without any enhancement of thymidylate synthetase inhibition or side effects in terms of death rate and body weight loss. These results suggest that the mechanism of the combined effect of 5-FU and IFN-beta is not related to enhancement of thymidylate synthetase inhibition or the host immune system, since human fibroblastoid IFN-beta is species-specific to humans. The clinical usefulness of this combination method for the treatment of advanced colorectal carcinoma is expected.

Chemosensitivity of breast cancer lymph node metastasis compared to the primary tumor from individual patients tested in the histoculture drug response assay.

Lymph node metastasis is often the first indication of the aggressiveness of breast cancer. Effective chemotherapy in breast cancer depends on targeting the metastatic component of the disease. In order to optimize chemotherapy in the metastatic target of breast cancer, the histoculture drug response assay (HDRA) was performed on surgical specimens of primary tumor and axillary lymph node metastasis from 30 breast cancer patients. The surgical specimens were cut into approximately 10 mg pieces, and placed onto the collagen gel sponges in the medium containing previously-determined cutoff concentrations of doxorubicin (DXR), 5-fluorouracil (5-FU), cisplatin (DDP), and mitomycin C (MMC). After incubation for 7 days, the chemosensitivity of the tumor fragments was evaluated with the 3-(4,5-dimethythiazol2yl)-2,5-diphenyl-2H tetrazolium bromide (MTT) endpoint. The lymph node metastases were more resistant than the primary tumor for DXR, 5-FU, and MMC (p < 0.05) but not for CDDP. The data suggest that both primary tumor and metastases from individual patients should be tested in the HDRA to enhance clinical efficacy of chemotherapy.

Immunochemotherapy prevents human colon cancer metastasis after orthotopic onplantation of histologically-intact tumor tissue in nude mice.

A metastatic model of human colon cancer has been previously established using orthotopic onplantation of histologically intact in tissue nude mice. In this study, effects of immunochemotherapy using OK-432, 5-fluorouracil (5-FU) and mitomycin C (MMC) on Col-2-JCK, a human colon cancer xenograft, were evaluated using this model. When 5-FU and MMC were administered without OK-432, liver metastases were not reduced even at maximum tolerated doses of both drugs, although cecal tumor growth was significantly reduced. On the other hand, when combined with OK-432, both 5-FU and MMC reduced liver metastases with synergistic reduction of cecal tumor growth, demonstrating the potential of combining immunotherapy with chemotherapy against metastases.

Antitumor activity of (2''R)-4'-O-tetrahydropyranyl adriamycin on human gastric cancer cell lines in vitro and in vivo.

The antitumor activity of (2''R)-4'-O-tetrahydropyranyl adriamycin (pirarubicin; THP) was assessed using human gastric cancer cell lines in vitro and in vivo. The cytotoxicity of THP on MKN-28 and MKN-45 was superior to that of adriamycin (ADM) as detected by a growth assay with an MTT colorimetric endpoint. When the same doses of THP and ADM were administered intraperitoneally to nude mice bearing St-15, St-40 and SC-1-NU, the antitumor activity of THP was almost equivalent to ADM in terms of relative mean tumor weight. However, the adverse effects of THP were also significantly lower than those of ADM in terms of death rate, body weight loss and spleen weight loss. This was also confirmed in THP or ADM combination chemotherapy with mitomycin C and 5-fluorouracil on St-15 and MKN-45. These results indicated that THP is a candidate anthracycline to replace ADM for combination cancer chemotherapy in gastric carcinoma.

Combined effect of 5-fluorouracil and carboplatin against human gastric cancer cell lines in vitro and in vivo.

The antitumor activity of a sequential combination of 5-fluorouracil (5-FU) and carboplatin (JM-8) was evaluated using gastric cancer cell lines in vitro and in vivo. In the in vitro study, the sequence of 5-FU followed by JM-8 showed higher antitumor activity than that of the reverse sequence. The sequence of 5-FU at 5 micrograms/ml for 24 h followed by 5 micrograms/ml JM-8 for 24 h showed antitumor activity almost equivalent to that of 10 micrograms/ml 5-FU for 24 h and higher activity than that of 10 micrograms/ml JM-8 for 24 h on two cell lines. To evaluate the antitumor activity and toxicity of 5-FU and JM-8 in vivo, BALB/cA nu/nu mice bearing human gastric cancer xenografts St-15, St-40 and SC-1-NU were administered 5-FU and JM-8 intraperitoneally. The sequence of 5-FU prior to JM-8 showed higher antitumor activity than that of the reverse sequence on all the xenografts, and simultaneous administration of 5-FU and JM-8 showed the most potent antitumor activity on St-40 and SC-1-NU. On the other hand, the sequence of 5-FU before JM-8 showed the lowest toxicity in all the treated groups, in terms of death rate, body weight loss and spleen weight loss. This combination is thought to be a promising chemotherapy regimen, showing high antitumor activity without an increment of toxicity.

A newly developed hexamethylmelamine derivative, SAE9 with both antitumor and aromatase-inhibitory activity.

Hexamethylmelamine (HMM) has previously been shown to be active against ovarian, breast and small cell lung cancer. However HMM dose not have aromatase-inhibitory activity. A newly developed HMM derivative, 2-N,N-dimethylamino-4, 6-bis (1-H-imidazol-1-yl)-1,3,5-triazine (SAE9), was found to have direct antitumor activity as well as aromatase-inhibitory activity. The direct antitumor activity on breast carcinoma cell lines (MCF-7, R-27 and MDA-MB-231) was assessed using the 3-(4,5-dimethylthiazol-2yl)-2, 5-diphenyl tetrazolium bromide (MTT) on cells growing in monolayer culture. The 50% inhibitory concentrations (IC50) of SAE9 were found to be approximately 10(-4) M for each cell line, roughly equivalent to those of HMM. When the aromatase-inhibitory effect was assessed using a human placental aromatase-inhibitory assay, the IC50 of SAE9 was 5.5 x 10(-7) M, which was superior to that of aminoglutethimide (AG) (3.8 x 10(-5) M). In a rat uterine growth model treated with androstenedione as the in vivo aromatase inhibition assay, SAE9 had an effect equivalent to that of AG. Since SAE9 has both antitumor and aromatase-inhibitory activity on breast carcinoma cell lines with estrogen dependency, this and similar non-steroidal aromatase inhibitors are thought to be promising for further study.

Different chemo- and endocrino-sensitivity of MCF-7 cells with or without estradiol supplement in vitro.

The sensitivity of MCF-7 cells to tamoxifen (TAM) and mitomycin C (MMC) was assessed in rapidly and slowly growing cells with or without estradiol supplementation, respectively. The growth of MCF-7 was inhibited by MMC in a concentration-dependent manner with or without estradiol (E2) supplementation. Preincubation with MMC suppressed subsequent E2 stimulated growth of MCF-7. TAM inhibited the growth of MCF-7 supplemented with E2 and preincubation with TAM prevented subsequent E2 stimulated growth of MCF-7. However, TAM did not inhibit the growth of MCF-7 cells in E2 free medium. These results suggested that MMC may be more effective than TAM on breast cancer cells in the dormant or slow-growth phase.

Site-specific chemosensitivity of human small-cell lung carcinoma growing orthotopically compared to subcutaneously in SCID mice: the importance of orthotopic models to obtain relevant drug evaluation data.

We have developed a novel in vivo model of human small-cell lung carcinoma (SCLC) using orthotopic reconstitution by injecting human SCLC in the tail vein of severe combined immunodeficient (SCID) mice whereby the SCLC grows in the lung and other organs. Cisplatin (DDP) had significant antitumor effects on the SCLC growing orthotopically in the lung whereas mitomycin C (MMC) did not, thereby reflecting the clinical situation. However, the opposite effects were found when the SCLC was growing subcutaneously, where the tumors responded to MMC and not to DDP. This suggests that the tumors growing orthotopically reflect the clinical effects of drugs on human SCLC more closely than the tumors growing subcutaneously. Therefore, this orthotopic reconstitution model of human SCLC in SCID mice is thought to be useful for studies on the treatment of human SCLC and emphasizes the need for orthotopic models for relevant cancer drug evaluation.

Modulation by l-leucovorin of 1-hexylcarbamoyl-5-fluorouracil antitumor activity on human gastric and colon carcinomas serially transplanted into nude mice.

Experimental biochemical modulation of 1-hexylcarbamoyl-5-fluorouracil (HCFU) with l-leucovorin (LV) was carried out using human gastric (H-111) and colon (Co-4) carcinoma xenografts serially transplanted into nude mice. Thirty-five or 70 mg/kg HCFU dissolved in 0.2 ml of 1% hydroxymethyl cellulose was administered po daily for 3 weeks except Sundays, and 50, 100, 200 or 300 mg/kg LV dissolved in 0.2 ml physiological saline was administered po 30 min before administration of HCFU. The biochemically modulated antitumor activity was evaluated in terms of actual tumor weight, the relative mean tumor weight and the degree of inhibition of thymidylate synthetase (TS) in the tumors at the end of the experiments, assayed according to the method of Spears et al. Although 35 mg/kg HCFU was ineffective against gastric carcinoma H-111, combination with 200 or 300 mg/kg LV resulted in a positive antitumor effect of HCFU on this strain without any increase of side effects in terms of body weight loss and mouse mortality. The colon carcinoma strain Co-4 showed marginal sensitivity to HCFU (35 mg/kg) alone, but 50 or 100 mg/kg LV modulated the antitumor activity of HCFU on Co-4 to produce a significant positive effect without any increase in toxicity, and HCFU administered with 100 mg/kg LV was more effective than the maximum tolerated dose of HCFU (70 mg/kg) alone. The TS inhibition rate was closely related to the biochemical modulation of HCFU antitumor activity by LV, suggesting that the modulation involves an increase of the ternary complex of TS, 5,10-methylene tetrahydrofolate from LV and 5-fluorodeoxyuridine 5'-monophosphate (FdUMP). Combination of HCFU and LV is therefore thought to be useful in increasing the antitumor activity of HCFU on gastrointestinal carcinomas without enhancing its toxicity.

Recombinant human interferon alpha-2a increases the antitumor activity of 5-fluorouracil on human colon carcinoma xenograft Co-4 without any change in 5-FU pharmacokinetics.

We investigated the modulating effect of recombinant human interferon alpha-2a (IFN-alpha) on the antitumor activity of 5-fluorouracil (5-FU) against a human colon carcinoma xenograft (Co-4) in nude mice with reference to changes in the pharmacokinetic pattern of 5-FU. Mice bearing Co-4 received 5-FU ip at a dose of 90 mg/kg once with or without IFN-alpha, which was administered sc at a dose of 60.000 IU/mouse daily for 7 days before 5-FU treatment. When the area under the curve (AUC) and peak plasma concentration (Cmax) of 5-FU with or without IFN-alpha were measured as pharmacokinetic parameters, the pharmacokinetics of 5-FU was not changed by IFN-alpha administration. This result suggests that the modulating effect of IFN-alpha on 5-FU does not involve augmentation of 5-FU pharmacokinetic parameters.

Orthotopic reconstitution of human small-cell lung carcinoma after intravenous transplantation in SCID mice.

We have constructed an orthotopically reconstituted model of human small-cell lung carcinoma (SCLC) by intravenous transplantation in severe combined immunodeficient (SCID) mice. Two human SCLC xenografts, H-69 and Lu-130, were disaggregated and injected through the tail vein of SCID mice. Human SCLCs were orthotopically reconstituted with multi-focal lung tumor growth in all SCID mice after intravenous injection of 5 x 10(6) tumor cells per mouse. The heart and liver were also seeded with actively growing SCLC. This orthotopic reconstitution model of human SCLC in SCID mice should be useful for further studies on the biological behavior and treatment of human SCLC.

Pirarubicin might partly circumvent the P-glycoprotein-mediated drug resistance of human breast cancer tissues.

BACKGROUND: Anthracyclines are the first line antitumor agents against breast cancer, and P-glycoprotein (Pgp) is thought to be the main resistance mechanism against these agents. We have evaluated the chemosensitivity of fresh surgical specimens of breast cancer and compared them with their Pgp-expression. MATERIALS AND METHODS: The in vitro chemosensitivity of 65 surgical specimens obtained from 63 patients with advanced breast cancer was assessed by the histoculture drug response assay (HDRA) using doxorubicin (DXR), pirarubicin [(2"R)-4'-tetrahydropyranyladriamycin: THP], and epirubicin (EPIR). Breast cancer tissues were plated onto collagen gel matrix and incubated with 15 micrograms of DXR or EPIR, or 17 micrograms of THP per ml for 7 days with MTT assessed at the endpoint. The efficacy of the agents was evaluated by the inhibition index (I.I.) of the optical density detected by ELISA reader. RESULTS: When 60% or more I.I. was regarded as in vitro sensitive at each cut-off concentration of the drugs, the overall efficacy rates were 60.7%, 48.6%, and 78.6% for DXR, EPIR, and THP, respectively. Fifty-one surgical specimens were evaluated for the immunohistochemical analysis of Pgp and the correlation between the sensitivity to anthracyclines and the expression of Pgp was compared. Pgp was expressed in 23.5% (12/51) specimens and the efficacy of anthracyclines was reduced in Pgp-positive breast cancer tissues, although this reduction was low in THP with a statistically significant difference when comparing with DXR and EPIR. CONCLUSION: The present results suggest that THP might partly circumvent the mdr1/PgP-mediated drug resistance mechanism in human breast cancer tissue and would have some different antitumor spectra on breast cancer comparing with DXR and EPIR.

Acquisition of multidrug resistance in recurrent breast cancer demonstrated by the histoculture drug response assay.

Recurrent breast cancer has a very poor response rate to chemotherapy. To understand the degree of acquisition of multidrug resistance in recurrent disease, 24 recurrent breast tumors and 127 primary tumors were evaluated and compared for chemosensitivity in the histoculture drug response assay (HDRA). The evaluation rate was 98.8%. The HDRA utilizes 3-dimensional culture of human tumors on collagen-gel rafts. Doxorubicin (DXR), 5-fluorouracil (5-FU) and mitomycin C (MMC) were tested as standard agents and cisplatin (CDDP) as a candidate agent on surgical specimen of breast cancer in the HDRA. In vitro drug exposure in the HDRA was for 7 days. At the end of the assay, tumor response was assessed by the reduction of 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). The mean inhibition rates of primary tumors vs. recurrent tumors were 57.9% and 38.6% for DXR (p<0.0005); 59.9% and 42.8% for MMC (p<0.01); 49.0% and 33.4% for 5-FU (p<0.01); and 34.5% and 16.0% for CDDP (p<0.005), respectively. The recurrent cases were pretreated clinically with CAF (cyclophosphamide, DXR and 5-FU), CEF (cyclophosphamide, epirubicin and 5-FU) or CMF (cyclophosphamide, methotrexate and 5-FU). In the CAF and CEF group, the HDRA sensitivity to CDDP was significantly lower in recurrent disease (p<0.005) than that of primary breast cancer suggesting that one agent can induce resistance to another. This is further suggested by the fact that 64.7% of the recurrent cases were resistant to all 4 agents tested as opposed to 27% of the primary cases and that only 5.9% of the recurrent cases were sensitive to three or more agents as opposed to 18% of the primary cases. The correlation of the HDRA results to clinical outcome in the study was 80.0% with 15 cases evaluated consisting of 5 true positives, 3 false positives, 7 true negatives and no false negatives. Thus, the HDRA gives useful clinical information, in particular for the specific individualized treatment design necessary to overcome the multidrug resistance problem of recurrent breast cancer.

Disappearance curves for tumor markers after resection of intrathoracic malignancies.

Determination of the standard elimination kinetics of tumor markers will be helpful in the diagnosis of malignancies. We analyzed the disappearance curves for serum tumor marker levels after resection of intrathoracic malignancies. Serum levels of CEA, SLX, AFP, CA 19-9, SCC, TPA and CYFRA were measured several times after surgery in a total of 40 patients. To obtain precise biological half-lives, we applied non-linear least square analysis, taking into consideration the possibility of residual tumor cells. Disappearance curves were monophasic for CEA, SCC, TPA, CYFRA and SLX and biphasic for CA 19-9 and AFP. Temporary elevation of serum levels after surgery was observed for SCC, TPA and CYFRA. The average half-lives of CEA, SLX, SCC, TPA and CYFRA were 1.5 days, 2.7 days, 2.2 hours, 2.5 hours and 1.5 hours, respectively. The average half-life of CA 19-9 was 0.5 days in the first compartment and 4.3 days in the second compartment, while that of AFP was 1.0 days and 6.3 days, respectively. These values will be helpful in the interpretation of serum tumor marker levels after surgery.

Resistant mechanisms of anthracyclines--pirarubicin might partly break through the P-glycoprotein-mediated drug-resistance of human breast cancer tissues.

Juliano and Ling initially reported the expression of a 170 kDa glycoprotein in the membrane of Chinese hamster ovarian cells in 1976, and named this glycoprotein P-glycoprotein (P-gp) based on its predicted role of causing "permeability" of the cell membrane. After much research on anthracycline-resistance, this P-gp was finally characterized as a multidrug-resistant protein coded by the mdr1 gene. Multidrug resistance associated protein (MRP) was initially cloned from H69AR, a human small cell-lung carcinoma cell line which is resistant to doxorubicin (DXR) but does not express P-gp. MRP also excretes substrates through the cell membrane using energy from ATP catabolism. The substrate of MRP is conjugated with glutathione before active efflux from cell membrane. Recently, membrane transporter proteins were re-categorized as members of "ATP-Binding Cassette transporter"(ABC-transporter) superfamily, as shown at http://www.med.rug.nl/mdl/humanabc.htm and http://www.gene.ucl.ac.uk/nomenclature/genefamily/abc.html. A total of ABC transporters have been defined, and MDR1 and multidrug resistance associated protein 1 (MRP1) were reclassified as ABCB1 and ABCC1, respectively. Their associated superfamilies include 11 and 13 other protein, in addition to ABCB and ABCC, respectively. Lung resistance-related protein (LRP) is not a member of the superfamily of ABC transporter proteins, because it shows nuclear membrane expression and transports substrate between nucleus and cytoplasm. LRP was initially cloned from a non-small cell lung carcinoma cell line, SW1573/2R120 which is resistant to DXR, vincristine, etoposide and gramicidin D and does not express P-gp. The mechanisms of resistance remains unclear, and why some resistant cell lines express P-gp and others express MRP and/or LRP is likewise unclear.