Abstract
Background: The goal of this study was to determine whether liver, gastric, or colonic cancer may be suitable targets for chemosaturation therapy with percutaneous hepatic perfusion (CS-PHP) and to assess the feasibility of utilizing other cytotoxic agents besides melphalan in the CS-PHP system. Materials and Methods: Forty human cell lines were screened against three cytotoxic chemotherapeutic agents. Specifically, the dose-dependent effect of melphalan, oxaliplatin, and paclitaxel on proliferation and apoptosis in each cell line was evaluated. These agents were also evaluated for their ability to induce apoptosis in normal primary human hepatocytes. A high-dose short-term drug exposure protocol was employed to simulate conditions encountered during CS-PHP. Results: The average concentration of melphalan required for inducing significant apoptosis was 61 μM, or about 3-fold less than the theoretical concentration of 192 μM, achieved in the hepatic artery during CS-PHP dosing with melphalan. Additionally, we found that gastric cancer cell lines were 2-5 fold more sensitive to apoptosis than liver cancer cell lines to all three compounds, suggesting that in addition to colonic and gastric cancer metastases to the liver, primary gastric cancer may also be amenable to management by CS-PHP using an appropriate therapeutic agent. Significantly, at concentrations that are predicted using the CS-PHP system, these agents caused apoptosis of colonic, gastric, and liver cancer cells but were not toxic to primary human hepatocytes. Conclusion: The compounds tested are potential candidates for use in the CS-PHP system to treat patients with gastric and colonic metastases, and primary cancer of the liver.
Chemotherapeutic molecules exert beneficial clinical effects by inhibiting cell growth or by inducing cell death via apoptosis. They can be divided into several categories based on their mechanisms of action. The chemotherapeutic agent melphalan hydrochloride, which has been approved by the US Food and Drug Administration and is used in the treatment of multiple myeloma and ovarian cancer, is a derivative of nitrogen mustard that acts as a bifunctional alkylating agent. Melphalan causes the alkylation of DNA at the N-7 position of guanine and the N-3 position of adenine (1). Thus, the binding of melphalan to DNA can result in cross-linking between bases, particularly G-G and G-A, on complementary strands, which leads to double-stranded DNA breaks and cell death through a caspase-mediated apoptotic pathway (2-5). Taxanes, such as paclitaxel, exert their effects by inhibiting microtubules (6, 7). Disruption of microtubules during cell division leads to cell-cycle arrest and subsequent cell death by apoptosis. Paclitaxel has been used clinically in treatment regimens for a number of cancer types such as breast cancer and non–small-cell lung cancer (8). Platinum agents, such as oxaliplatin, bind to and damage DNA, which leads to disruption of transcription and replication and to cell-cycle arrest. If this DNA damage is not repaired, apoptosis will ultimately result (9). These agents comprise part of the treatment regimens for cancer such as colorectal, ovarian, and bladder cancer (10).
At the doses required for clinical efficacy, chemotherapeutic agents are often associated with significant side effects. For example, because these compounds target rapidly dividing cells, they often induce myelosuppression, which contributes significantly to patient morbidity (11). Thus, a system that enables the delivery of chemotherapeutic compounds specifically to the organ of interest while reducing systemic exposure offers significant advantages. Isolated hepatic perfusion (IHP) and percutaneous hepatic perfusion (PHP) offer this type of delivery system.
The administration of melphalan via IHP and PHP has been studied for the treatment of liver cancer, as well as for metastases to the liver from colorectal cancer and metastatic melanoma, among others (12-15). IHP permits intensification of chemotherapy to the liver, thereby improving efficacy while limiting extrahepatic toxicity. However, its efficacy is restricted because the treatment cannot be repeated since post-surgical intra-abdominal adhesions preclude further surgery. By contrast, chemosaturation therapy with PHP (CS-PHP), represented in Figure 1, is a repeatable percutaneous procedure for the intra-arterial administration of melphalan hydrochloride to the liver with extracorporeal filtration. The CS-PHP device is approved for clinical use in Europe where cases are being performed routinely in several countries. The proprietary CS-PHP system delivers melphalan to the liver via the proper hepatic artery and employs a double-balloon aspiration catheter in the retrohepatic inferior vena cava to isolate and collect hepatic venous outflow and direct the effluent to an extracorporeal circuit containing a proprietary hemofiltration system that reduces the concentration of melphalan in the blood dramatically and returns the filtered blood to the systemic circulation through the jugular vein. Thus, when combined with an appropriate filtration system, this system could potentially be used to deliver other chemotherapeutic agents specifically to the liver.
In this study, we screened six gastric, seven liver, and 27 colon cancer cell lines for their sensitivity to melphalan, oxaliplatin, and paclitaxel. We determined whether any of these agents effectively induces cell death of cancer cells in vitro and, thus, could be used with the CS-PHP device. This study is notably relevant to clinical settings because of the short time of exposure (50 min) of cells to high doses of the compound of interest followed by a wash-out step to remove the agent from the cell milieu, simulating CS-PHP. A multiplex system was used to simultaneously measure cell proliferation and apoptosis. Finally, in order to address the safety of these agents in a potential CS-PHP to treat liver metastases, we performed assays on primary human hepatocytes and measured the relative attached cell count and apoptosis induction.
Materials and Methods
Cell lines and human hepatocytes. The 40 cell lines used in this study and their tissue of origin are shown in Table I. All cell lines were obtained from the American Type Culture Collection (Manassas, VA, USA), Leibniz-Institut Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (Braunschweig, Germany), and the Japanese Collection of Research Bioresources (National Institute of Health Sciences, Osaka, Japan) and maintained as specified by the vendor. For the multiplexed cytotoxicity assays, cell lines were grown in Roswell Park Memorial Institute 1640 medium (RPMI 1640) or Dulbecco's modified Eagle's medium (DMEM), 10% fetal bovine serum (FBS), 2 mM L-alanyl-L-glutamine, and 1 mM sodium pyruvate in a humidified atmosphere of 5% CO2 at 37°C. Cells were seeded into 384-well plates and incubated in 5% CO2 at 37°C.
Human primary hepatocytes (Invitrogen, Carlsbad, CA, USA) were thawed in hepatocyte culture medium with supplements and growth factors [hepatocyte basal medium (HBM™) and hepatocyte culture medium (HCM™) SingleQuots® from LONZA, Walkersville, MD, USA], treated with a Percoll (Sigma-Aldrich, St. Louis, MO, USA) separation step, and immediately seeded into 384-well collagen I-coated optical plates (BD Biosciences, San Jose, CA, USA). Cells were incubated at 37°C in 5% CO2 for 24 h, after which the medium was exchanged and the cells were treated with test compounds.
Cell treatment with test compounds. Melphalan, oxaliplatin, paclitaxel and mitoxantrone were purchased from Sigma-Aldrich or Calbiochem (EMD Millipore, Billerica, MA, USA). For the multiplexed cytotoxicity assays, compounds were serially diluted 3.16-fold and assayed over a range of 10 concentrations; the highest and lowest concentration for each compound is specified in Tables I-III. The final dimethyl sulfoxide (DMSO) concentration in the assay did not exceed 0.1%. Compounds were added 24 h after cell seeding. At the same time, a time-zero untreated cell plate was generated. The compound treatment period was 50 min. After incubation with a compound, the cells were spun, the supernatant was aspirated to remove the compound, and the medium was exchanged three times and cells incubated for an additional 72 h at 37°C.
For primary hepatocytes, at 24 h after plating, the medium was exchanged and compounds were added. Each plate of tested compounds included one reference compound (mitoxantrone) and vehicle control. Compounds were serially diluted 3.16-fold and assayed over a range of 10 concentrations (0.008 μM - 250 μM for mitoxantrone) similar to those described for the cell lines. The final DMSO concentration in the assay did not exceed 0.5%. The compound treatment period was 1 or 4 h, followed by a wash-out. Fresh medium was added to the cells, and the cells were incubated for an additional 17 or 14 h, respectively, at 37°C prior to being assayed for an attached live count and for caspase-3 activation at 18 h.
Cell fixation and labeling. Following the final incubation step, the cells were fixed and stained with fluorescently labeled antibodies and nuclear dye to allow visualization of nuclei and apoptotic cells. Cells were fixed with paraformaldehyde (3.33% final concentration). Plates were spun briefly and incubated for 15 min at room temperature. The supernatant was aspirated using a BioTek Elx405 plate washer (BioTek, Winooski, VT, USA) and 80 μl of block solution was added [0.25% bovine serum albumin (BSA), 0.05% Triton X-100, and 0.25% goat serum in tris-buffered saline (TBS) buffer] for 10 min. A rabbit polyclonal anti-active caspase-3 (Cell Signaling Technology, Danvers, MA, USA) was added to the plates after aspiration of the block solution and cells then incubated for 30 min. Plates were then washed twice with PBS and secondary antibody was added for 30 min (Alexa Fluor 488 goat anti-rabbit, Invitrogen). The plates were then washed twice, stained with 2 mg/ml amidino phenyl indole (DAPI; Invitrogen) and imaged using a General Electric (GE) Healthcare (Waukesha, WI, USA) IN Cell Analyzer 1000. Twelve-bit TIFF images were acquired using the IN Cell Analyzer 1000 3.2 and analyzed using Developer Toolbox 1.6 software (GE Healthcare, Piscataway, NJ, USA).
Summary of melphalan sensitivity. Cell lines are ranked from sensitive to resistant to the induction of 5-fold apoptosis. The tissue of origin of each cell line is indicated as gastric (G), liver (L) or colonic (C).
Cell proliferation assessment (EC50). Cell proliferation was assessed using a cell image-based technique. Cells were stained with DAPI, a nuclear dye, to visualize nuclei. Cell proliferation was measured
by the signal intensity of the incorporated nuclear dye. The cell proliferation assay output is the relative cell count. To determine the cell proliferation endpoint of EC50, data output was transformed to a percentage that of the control (POC) using the following formula:
Relative cell count (EC50) is the test compound concentration at the curve inflection point or half the effective response. EC50 values were calculated using nonlinear regression analysis to fit the data to a sigmoidal 4-point, 4-parameter, one-site dose response model, where y (fit)=A + [(B − A) ÷ (1 + [C/x]D)].
Summary of results for oxaliplatin sensitivity. Cell lines are ranked from sensitive to resistant to the induction of 5-fold apoptosis. The tissue of origin of each cell line is indicated as gastric (G), liver (L) or colonic (C).
Summary of results for paclitaxel sensitivity. Cell lines are ranked from sensitive to resistant to the induction of 5-fold apoptosis. The tissue of origin of each cell line is indicated as gastric (G), liver (L) or colonic (C).
Apoptosis induction by tissue of origin.
Apoptosis assay. Apoptotic cells were detected using an antibody to anti-activated caspase-3 (rabbit polyclonal anti-active caspase-3; Cell Signaling Technology). Automated fluorescence microscopy was carried out using a GE Healthcare IN Cell Analyzer 1000, and images were collected with a 4 × objective.
For apoptosis output (indicated by activated caspase-3 marker), each intensity was first normalized to the nuclear intensity in their respective wells. This ratio was then normalized to that of the control wells and expressed as fold induction over vehicle:
where Cx=caspase intensity for well X, Nx=nuclear intensity for well X, Cc=caspase intensity for control well, and Nc=nuclear intensity for control well. Wells with concentrations higher than the relative cell count (IC95) were eliminated from the caspase-3 induction analysis.
The output is reported as the fold increase in apoptotic cells over vehicle background normalized to the relative cell count for each well. Significant induction of apoptosis was defined based on vehicle mean plus six standard deviations (SDs). When the vehicle mean plus six SDs was compared for multiple cell lines, the overall range of apoptosis induction, as measured by activation of caspase-3, was 2.5-fold to 4.5-fold. Hence, 5-fold apoptosis induction was selected as being significant, and concentrations of each agent that result in 5-fold apoptosis induction are reported.
Primary hepatocyte cytotoxicity assay: Staining. For primary hepatocytes, following the 18-h incubation period, the cells were incubated for 30 min at 37°C with multiplexed fluorescent dyes and then imaged. Nuclei were labeled with DAPI for cell quantification. The cells were then fixed with 3% paraformaldehyde and imaged. Automated fluorescence microscopy was carried out using a GE Healthcare IN Cell Analyzer 1000, and images were collected with a 4 × objective.
Attached Live Cell Count: For primary human hepatocytes, live cell attachment was measured by the degree of nuclear segmentation of the incorporated nuclear dye. The live cell attachment assay output is referred to as the percentage of attached live cells. To determine the live cell attachment end point, the nuclear segmentation data output was transformed to POC using the following formula:
The EC50 for attached live cells is the concentration of test compound that produces 50% of the maximum effective response at the curve inflection point.
The chemosaturation therapy with percutaneous hepatic perfusion (CS-PHP) system. The CS-PHP system enables the isolation of the blood supply to the liver through the use of a catheter system, thus allowing, via repeated procedures over time, the direct delivery of chemotherapeutic agents to the liver. The blood containing free chemotherapeutic agent is captured between the two balloons in the isolation aspiration catheter and is directed to an extracorporeal circuit containing a proprietary filter that filters the chemotherapeutic agent from the blood, significantly reducing the concentration of the chemotherapeutic agent in the blood before it is returned to the systemic circulation through the jugular vein. This allows the patient to be treated with relatively high doses of chemotherapy agents over a 60-minute period while limiting extrahepatic exposure to melphalan, thereby greatly reducing systemic toxicity. The figure shown is the property of Delcath Systems, Inc.
Results
Cancer cell line sensitivity to melphalan. The 40 cell lines that were tested in this study had a range of sensitivities to melphalan, defined by the induction of apoptosis and EC50 (Table I). The concentration required for 5-fold induction of apoptosis ranged from 2.89 μM to 450 μM. The cancer cell lines that were most sensitive to melphalan-induced apoptosis were the HCT-8 colorectal adenocarcinoma and the RKO-AS45-1 colon carcinoma cell lines. Among the cell lines in which apoptosis could be quantified, the most resistant were the colorectal adenocarcinoma lines NCI-H747 and the SW1116. In three cell lines, namely the HS 746T gastric carcinoma, SNU-423 hepatocellular carcinoma, and SW948 colorectal adenocarcinoma, 5-fold induction of apoptosis was not achieved at any of the concentrations tested. The range for EC50 was 0.68 to 167 μM. When assessed by EC50, the colorectal adenocarcinoma cell lines Colo 205 and LS-174T were the most sensitive, and the colorectal adenocarcinoma cell line SW1417 and the gastric carcinoma cell line SNU-5 were the most resistant.
Cancer cell line sensitivity to oxaliplatin. As shown in Table II, 5-fold induction of apoptosis was achieved at oxaliplatin concentrations ranging from 15.5 μM to 446 μM. Colo 320 HSR and HCT-116 colorectal carcinoma cells exhibited the greatest sensitivity to oxalipatin. Among the cell lines for which apoptosis could be quantified, the most resistant were the colorectal adenocarcinoma lines LS1034 and SW1116. Maximum apoptosis induction was less than 5-fold in 12 cell lines. The range of EC50 values was 4.71 μM to 466 μM. The most sensitive cell lines were AGS and KATO III, both of which are gastric carcinomas. Among the cell lines with measurable EC50 values, the colorectal carcinoma NCI-H747 and gall bladder carcinoma OCUG-1 cell lines were highly resistant to oxaliplatin. The most resistant cell line was SNU-423, a hepatocellular carcinoma, for which an EC50 could not be determined.
Concentration required for a 5-fold induction of apoptosis in all cell lines for all compounds tested. Note that for cell lines for which 5-fold induction of apoptosis was not achieved, the value used to construct the graph was the highest concentration of that agent tested.
Cancer cell line sensitivity to paclitaxel. Quantifiable apoptosis was detected in 32 out of the 40 cell lines (Table III). Concentrations required for 5-fold induction of apoptosis were 0.013 μM to 28.9 μM. The cells requiring the lowest concentration of paclitaxel for 5-fold induction of apoptosis were OCUG-1 gall bladder carcinoma cells and SNU-1 gastric carcinoma cells. The cells most resistant to paclitaxel-induced apoptosis were the colorectal adenocarcinoma lines SW1463 and Colo 320 DM. The cells most sensitive to paclitaxel, as determined by EC50, were SNU-16 and SNU-1 gastric carcinoma cells. The colorectal adenocarcinoma cell lines Colo 320 DM and SW1116 were the least sensitive.
Comparison of sensitivity of cancer cell Lines to all agents tested. Shown in Figures 2 and 3 are radar graphs of the sensitivity of each cell line to the three compounds tested in this study. Paclitaxel was the most potent of the three drugs, with apoptosis induction and EC50 concentration ranges in the double-digit nanomolar range for sensitive cell lines. For melphalan and oxaliplatin, these concentrations were found to be several orders of magnitude higher and were calculated to be in the single- and double-digit micromolar range, respectively. Among the cell lines in which no measurable drug-induced apoptosis was detected, 10 (HCT-8, RKO-AS-45-1, SW48, WiDr, SW403, HuCCT1, SNU-5, SW1417, NCI-H747, and SW1116) were resistant to one drug, two (T84 and Colo 201) were resistant to two drugs, and three (Hs746T, SNU-423, and SW948) were resistant to all three drugs. When compared with the other cell lines used in this study, the three cell lines resistant to all three drugs were also relatively resistant when assayed for the antiproliferative EC50 endpoint. In fact, in the case of SNU-423 cells, an EC50 value could not be calculated for oxaliplatin. In the oxaliplatin-treated group, there were 12 cell lines (the most among the three drugs) for which 5-fold apoptosis was not detected. The paclitaxel- and melphalan-treated groups had eight and three cell lines, respectively, in which 5-fold apoptosis was not detected.
Relative cell count. EC50 in all cell lines, for all compounds tested.
The 40-cell–line panel used in this study consists of cell lines derived from gastric (n=6), liver (n=7), and colon (n=27) cancer. Table IV the average concentration of each drug required to induce apoptosis according to tissue of origin. The average concentration of melphalan required to induce apoptosis across all tissue types tested was 60.9 μM. For the liver-specific subset, this dropped to 51 μM, whereas for colon cancer cell lines, it was 70.1 μM. The average melphalan concentration required for apoptosis induction in the gastric cell lines was more than 2-fold lower at 24.9 μM,with the concentration being in the range of 14 to 18 μM for four (AGS, KATO III, SNU-16, and SNU-1) of the five gastric cancer cell lines (Table I). A similar trend was observed for oxaliplatin and paclitaxel, where the average concentrations required for the induction of apoptosis in gastric cancer cell lines were lower than those required for liver and colonic cancer cell lines.
Human hepatocyte cytotoxicity and apoptosis upon exposure to high-dose compound for 1 and 4 hours: (A) melphalan, (B) oxaliplatin, (C) paclitaxel, and (D) mitoxantrone.
Hepatic toxicity assay using human primary hepatocytes. A multiplexed hepatic toxicity assay was performed to evaluate the effect of short-term exposure of the three agents on primary human hepatocytes. The hepatic CS-PHP procedure performed in the clinic consists of a 30-min drug infusion period followed by a 30-min wash-out period. In order to simulate these conditions, we exposed the cells to drugs for 1 h then washed off the drug, replenished the cells with a fresh drug-free medium, and carried out assays for apoptosis and cytotoxicity at 18 h. We also assessed the potential safety window of CS-PHP by simulating conditions in which hepatocytes were exposed to high doses of drug for 4 h.
The IC50 was used as a measure of cell detachment or cytotoxicity and caspase-3 activation as a measure of apoptosis. The results of these assays are shown in Figure 4. None of the three compounds at any of the concentrations or time points tested showed any measurable cytotoxicity or caspase-3 activation. However, the reference control agent, mitoxantrone, was found to be cytotoxic at 1 and 4 h, with an EC50 of approximately 8 μM. Mitoxantrone also induced measurable apoptosis at the second highest concentration, 79 μM. The highest concentration of mitoxantrone, 250 μM, was found to be severely cytotoxic (<5% of the cells were viable) and, thus, accurate measurements of apoptosis were not possible. It was also determined that melphalan, oxaliplatin, and paclitaxel, at the concentrations used in this study, do not cause mitochondrial damage or the induction of reactive oxygen species (data not shown). Taken together, these data suggest that melphalan, oxaliplatin, and paclitaxel, at the concentrations and time employed in these studies, are not toxic to primary human hepatocytes and may be well tolerated by the liver.
Discussion
In this study, we used a high-dose, short-term drug exposure protocol to measure the chemosensitivity of six gastric, seven liver, and 27 colon cancer cell lines to three commonly used chemotherapeutic agents, namely melphalan, oxaliplatin, and paclitaxel. Specifically, we tested the ability of these drugs to induce apoptosis and inhibit cell proliferation. We also tested the ability of any of these agents to cause apoptosis and/or cytotoxicity in primary human hepatocytes at high doses. The treatment protocol was designed to assess the feasibility of extending the applicability of Delcath's proprietary CS-PHP system by simulating conditions encountered during the clinical procedure. Given the high doses of agents employed over a short time during CS-PHP, induction of apoptosis is likely to be the key determinant of cell death. The most common origin of metastases confined to the liver is colorectal cancer (12), making colonic, gastric, and liver cancer cells logical targets for the CS-PHP system.
In this study, the average melphalan concentration of 60.9 μM required for induction of 5-fold apoptosis in cells from gastric, colonic, and liver cancer was lower than the theoretical concentration achieved in the hepatic artery at the site of drug infusion during CS-PHP. In a phase I dose escalation trial in patients with unresectable hepatic malignancies, the maximum tolerated dose (MTD) for melphalan delivered via CS-PHP was determined to be 3 mg/kg (15). In a subsequent phase III trial using melphalan at 3 mg/kg, up to a maximum dose of 220 mg, in 92 patients with hepatic metastases from malignant melanoma, the investigators reported a hepatic progression-free survival (H-PFS) of 245 days, which was significantly longer than the H-PFS of 49 days seen in the control group (16). Assuming that approximately one fourth of the total hepatic circulation comes from the hepatic artery, a 220-mg dose of melphalan infused over 30 min at an extracorporeal circuit flow rate of 500 ml/min would result in a theoretical concentration of 192 μM (220 mg in 3.75 l of blood) in the hepatic artery. This represents a 3-fold increase over the average concentration required to induce apoptosis in the cell lines used in our study. A similar argument can be made for the use of paclitaxel through CS-PHP to treat patients with metastases to the liver. Clinical trials with intravenous (i.v.) paclitaxel to treat patients with breast cancer have used doses ranging from 175 mg/m2 (17) to 250 mg/m2 (18), which for an average individual with a body surface area of 1.73 m2 translates to 303 to 433 mg. Intrahepatic delivery of paclitaxel at this dose via CS-PHP to treat patients with cancer affecting the liver would result in a theoretical concentration in the range of 94 to 135 μM, which is several orders of magnitude greater than the average of 3.5 μM (Table IV) required to kill the majority of the cancer cell lines in our study. The MTD for i.v. oxaliplatin was determined to be 135 mg/m2 in a dose escalation phase I clinical trial in patients with advanced cancer (19). A subsequent phase II study with oxaliplatin as a single agent at 130 mg/m2 to treat patients with colorectal carcinoma patients reported modest response rates of 20% (20). Intrahepatic delivery of oxaliplatin via CS-PHP at 130 mg/m2 would achieve a theoretical concentration of 151 μM, which is almost identical to the average concentration of 130 μM required to induce apoptosis in our study. Interestingly, neither of these two studies reported any hepatotoxicity in patients who were given i.v. oxaliplatin at 130 mg/m2, which indicates that even higher doses might be tolerated by the liver if the drug were delivered specifically to the liver via CS-PHP. Indeed, a more recent study demonstrated that the MTD of oxaliplatin delivered via hepatic arterial infusion was 140 mg/m2, and this dose, when used in combination with other systemic therapies had antitumor activity in patients with metastatic liver disease (21). These data indicate that oxaliplatin delivered via CS-PHP may be effective in treating patients with liver metastasis.
The effect of melphalan, oxaliplatin, and paclitaxel on tumor cell killing is all the more striking when viewed in light of the fact that none of these drugs had a deleterious effect on hepatocyte viability or toxicity, as assessed in our study. This is an important consideration, because high doses of drug can be repeatedly administered via CS-PHP to achieve tumor ablation. The maintenance of hepatocyte cell viability and a lack of apoptosis induction even after extended exposure periods of one and four hours clearly demonstrate that these drugs are well tolerated by hepatocytes. There are several lines of clinical and in vitro evidence that support our data. Firstly, although systemic cytoreductive therapy using i.v. high-dose melphalan at 100 mg/m2 has been shown to result in hepatotoxicity, manifested as veno-occlusive disease (22), the aforementioned phase I clinical trials using high doses of melphalan via CS-PHP found no hepatotoxicity associated with doses below the MTD but at which antitumor effects were observed (15). Secondly, melphalan induced caspase-3-dependent apoptosis of murine hepatocytes in vitro was shown to occur in the range of 100 to 200 μg/ml (327-654 μM) but required an overall exposure time of 9 to 12 h (5). Moreover, there was no increase in alanine aminotransferase (ALT), a marker of liver health and function, in mouse livers perfused with 150 mg/kg of melphalan until 360 min (5). Intrahepatic delivery of melphalan via CS-PHP results in a far lower theoretical concentration (192 μM) over a much shorter time period (30 min) and, thus, would be predicted not to result in significant apoptosis of hepatocytes. Thirdly, oxaliplatin at i.v. doses of 130 mg/m2 was shown to be equally well tolerated by patients regardless of hepatic function, which suggests that it does not result in hepatotoxicity (23). Fourthly, clinical studies with 250 mg/m2 i.v. paclitaxel given as a 3-h infusion to treat patients with breast cancer did not show any evidence of drug-related hepatotoxicity, which indicates that it is safe and well tolerated by the liver (17). Hepatotoxicity to paclitaxel has been reported in patients with impaired hepatic function (24). Consequently, any use of paclitaxel via CS-PHP in these patients needs to be considered carefully. Finally, even though our primary hepatocyte cytotoxicity assays showed that the three compounds tested are not cytotoxic towards these cells, it is important to note the limitations of this study: hepatocytes represent only one cell type in a complex organ. The liver contains many other cell types, such as Kupffer cells, which may respond differently to these agents. Because hepatocytes are nondividing cells, and cells such as Kupffer cells are dividing cells, they may be more sensitive to the agents tested.
The potential use of chemotherapeutic drugs in hepatic arterial infusion has been investigated in several studies using ex vivo and in vitro cell-based assays (25-27). A 2-h exposure to oxaliplatin was found to result in the dose-dependent inhibition of proliferation of the HT-29 colon cancer cell line as well as nine out of 10 tumors derived from colorectal and pancreatic cancer metastasis to the liver (25). The average IC50 for the inhibition of colony formation of the metastatic cells was reported to be less than 10 μg/ml (25), which corresponds to 25 μM and is in general agreement with our average EC50 values of 23 and 45 μM for gastric and colonic cancer cell lines, respectively (as calculated from the values for the respective cell lines shown in Table II). Two other recent studies have also investigated in vitro dosing paradigms similar to our study (26, 27). Boulin et al. tested 11 chemotherapeutic drugs, including oxaliplatin and paclitaxel, on three human hepatocellular carcinoma cell lines to select a candidate for transarterial chemoembolization (26). The only cell line common to our study and theirs was HepG2. However, they report substantially higher IC50 values for oxaliplatin when compared to our study (262 μM vs. 36 μM) and paclitaxel (241 μM vs. 0.5 μM.) Moreover, in their study, there does not appear to be a difference in the IC50 for oxaliplatin and paclitaxel, whereas we found a clear difference (36 μM vs. 0.5 μM). In another study, investigators used an in vitro cell line screen and evaluated the potential use of oxaliplatin in combination with melphalan for the treatment of patients with colorectal cancer to identify dosing schedules that could be employed during isolated liver perfusion (27). Five cell lines-SW48, RKO, DLD-1, HT-29, T84, and SW480 -in their panel of 13 lines were also tested by us in this study. Again, a majority of the IC50 values reported by them are significantly higher than the EC50 values identified in our study, although it is important to point out that the rank order of sensitivity of these lines to each drug is similar to what we found. One possibly important reason for the differences seen in our study versus both of the recent studies could be the state of plated cells prior to drug exposure. In both of these studies, cells were plated and allowed to reach confluence for three to four days prior to drug exposure. In our study, we plated cells in the log phase and allowed them to adhere overnight before drug exposure. This ensured that the added drug encountered rapidly dividing cells in their most sensitive state for drug efficacy. It is likely that in the absence of a rapidly dividing cell population in their experimental set up, Boulin et al. (26) and van Iersel et al. (27) would have failed to see effects at the lower concentrations. Moreover, neither of these studies investigated apoptosis as an endpoint in their study, whereas apoptosis, according to us, represents an important component of the cell death process. This is especially relevant in CS-PHP, where tumor cell killing due to short-term exposure to a high dose of a therapeutic agent would most likely be a result of apoptosis.
An unexpected and interesting finding of our study is that gastric cancer cell lines were 2- to 4-fold more sensitive than liver and colonic cancer cell lines across all three drugs. This greater sensitivity of gastric cancer cell lines to melphalan, oxaliplatin, and paclitaxel is indicative of a drug-specific response and not a general increase in sensitivity to all drugs, because for certain other drugs the average concentration required to induce apoptosis in gastric cancer cell lines is equal to or greater than that required for apoptosis to be induced in liver cancer cell lines (data not shown). To our knowledge, this is the first demonstration that melphalan is more effective against cancer cell lines of gastric origin than those of liver origin when applied in a simulated CS-PHP dosing regimen. This has important implications, not just for treating patients with gastric cancer metastases to the liver, but also for treating patients with primary gastric cancer via techniques similar to CS-PHP. Indeed, intraperitoneal chemotherapy using abdominal perfusion for gastric carcinomas has been reported for mitomycin C, both as a single agent (28) and in combination with 5-fluorouracil and cisplatin (29). Given our current data, it would be interesting to speculate that melphalan, oxaliplatin, and paclitaxel could be used via chemosaturation to treat primary gastric cancer. Such evaluations would be predicated on these drugs not being cytotoxic to cells in the normal gastric epithelium, similar to what we have demonstrated in this study for normal human hepatocytes.
Recent advances in molecular medicine have led to a greater understanding of the genetic heterogeneity of cancer. Personalized medicine allows for cancer therapies to be developed for and employed in a specific set of patients who are stratified based on the expression of certain biomarkers. Notable examples of such approved targeted therapeutics include the monoclonal antibody trastuzumab (Herceptin®; Roche, Madison, WI, USA), which is used for the treatment of patients with human epidermal growth factor receptor 2 (HER-2)–positive metastatic breast cancer; the small-molecule inhibitor imatinib (Gleevec®; Novartis, Carlsbad, CA, USA), which is used to treat patients with chronic myelogenous leukemia harboring the breakpoint cluster region-Abelson leukemia (BCR-ABL) translocation and gastrointestinal stromal tumors carrying a tyrosine protein kinase Kit (c-KIT) mutation (30-32). A need to identify biomarkers predictive of chemotherapy response has led to large screenings of multiple human cancer cell lines against multiple drugs (33). Previous work in this area has generally focused on single-gene mutations, but some studies have also analyzed the association of gene expression profiles with the response of different cell lines to individual agents. Oncogenes whose products are direct targets of relevant drugs and inactivating mutations of tumor-suppressor genes have been associated with sensitivity and resistance, respectively, to cancer therapies (33). Many such genes code for proteins that are implicated in regulation of the cell cycle, such as the products of TP53, p21, transcription factor E2F (E2F), and retinoblastoma tumor suppressor (PRB), whose dysregulation can have a major influence on the response to cytotoxic agents because of their effects on tumor cell apoptosis and survival (34). Notably, loss of the TP53 gene has been associated with resistance to multiple agents; however, at least one large screen failed to discern this association, perhaps because the near universality of p53 inactivation in cancer lines rendered this feature undifferentiating (33, 34). A detailed molecular definition of the cell lines used in our study is beyond the scope and intent of this discussion; however, we too have not observed any correlation between mutations in common tumor suppressor genes such as TP53 and adenomatous polyposis coli (APC) and resistance to drug-induced apoptosis. For example, the colorectal cancer cell lines Colo 320 HSR and HCT-15 contain mutations in the TP53 and APC genes (Sanger Institute Core Cell Line Viewer, http://www.sanger.ac.uk), but the concentration of melphalan required to induce significant apoptosis is lower (4.83 μM and 5.05 μM respectively) than the average (70.1 μM) observed for all colonic cell lines. A similar trend is seen for these two cell lines in the case of oxaliplatin-induced apoptosis (see Tables II and IV). By contrast, the gastric cancer cell line Hs 746T, for which the only reported mutation has been in the C-MET gene (35), was resistant to apoptosis induction by all three drugs tested. Thus, biomarker signatures that can be used to predict sensitivity and resistance to more general chemotherapeutics, such as those used in our study, may not be inherently obvious as mutations and may require more systematic approaches based on expression profiling.
Conclusion
Our results clearly demonstrate that melphalan, oxaliplatin, and paclitaxel are good candidates for further study in the CS-PHP system. These agents, at concentrations that are feasible using CS-PHP, induced significant apoptosis in liver, colonic, and gastric cancer cells but were not toxic to primary human hepatocytes, even at high doses and extended exposure times of 1 and 4 h. Thus, while cell line testing in this format is not always indicative of actual results in humans, this study provides compelling evidence that the compounds tested could be applied at effective concentrations using the CS-PHP system, without inducing liver toxicity, to treat primary and metastatic liver cancer.
- Received March 8, 2013.
- Revision received April 9, 2013.
- Accepted April 10, 2013.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
