Doxorubicin and mitoxantrone drug eluting beads for the treatment of experimental peritoneal carcinomatosis in colorectal cancer


  • Statement: A.L. is director of the company that sponsored the study and provided the materials. No restrictions were placed upon publication of any results.


We investigated the therapeutic efficiency of sulfonate-modified polyvinyl alcohol beads loaded with doxorubicin, irinotecan or mitoxantrone in vitro and in vivo in a model of experimental peritoneal carcinomatosis (PC). In vitro, cell proliferation was efficiently impaired by doxorubicin drug eluting bead (DEB) treatment while mitoxantrone DEBs were less effective than. Irinotecan showed little effect for both DEBs and free drug. Apoptosis was not different between free mitoxantrone and the DEB form while more apoptosis induction was observed in cells incubated with free doxorubicin and irinotecan. Experimental PC was produced in mice. The therapeutic efficiency of either mitoxantrone and doxorubicin DEB or free drugs were compared. Mice were treated either once on day 12 or by 3 repetitive applications on days 7, 10 and 12. Mice treated by DEBs showed less weight loss and mortality. Therapeutic effect was determined by measuring tumor volume and tumor load on the day 15 after tumor inoculation. For the single application on the day 12, an advantage could be observed for the free drugs. After 3 repeated injections of both free and mitoxantrone DEB no difference in tumor load or tumor volume could be observed. Least tumor load and tumor volume was observed in mice that received 3 repeated injections of doxorubicin DEB. No animal survived 3 injections of free doxorubicin. We conclude that bead encapsulation of chemotherapeutic drugs may show the advantage of less toxicity in peritoneal spread of colorectal cancer. © 2008 Wiley-Liss, Inc.

Peritoneal carcinomatosis (PC) of colorectal origin is a significant cause of morbidity and mortality in colorectal cancer.1 In an estimated 25% of patients, no other tumor locations can be found, even when a detailed diagnostic work-up is performed. Sugarbaker has suggested that PC of colorectal origin should probably not be equated with generalized disease, but can be a 1st step of dissemination, not unlike the situation with liver metastases of colorectal origin.2, 3 Based on this concept, attempts have been made to achieve long-term survival in patients with PC by combining surgery and intraperitoneal chemotherapy to eradicate microscopic residual disease.1

In clinical practice, PC frequently represents the terminal stage of adenocarcinomas of the gastrointestinal tract. The majority of patients with PC can be expected to die within 6 months.4 In the case of surgically unresectable PC, chemotherapy remains palliative.5 Patients under chemotherapy are reported to experience an average of 20 symptoms related to side effects, 13 of which were physical and 7 psychosocial.6 These symptoms are caused by the systemic application of the chemotherapy and should be expected to be less severe in targeted therapy where the drug is directly delivered to the organ of interest.

An elegant approach to further lessen systemic side effects of targeted therapy is loading of drugs or precursors within hydrogel beads. Previously we have used beads which contained microencapsulated human embryonic kidney cells expressing a transfected ifosfamide-converting enzyme that prevents the host immune system from attacking the transfected cells and allows implantation of non-HLA matched cells into patients.7 These beads have been shown to be effective in vivo using intraperitoneally injected C-26 cells which were stably transfected with enhanced green fluorescent protein (EGFP). This model has the advantage of easy tumor detection and the amount of EGFP detected correlates well to the tumor load.7

Up to now, novel drug eluting encapsulation systems, drug eluting bead (DEB) systems including beads loaded with doxorubicin (Dox) and other positively charged drugs like mitoxantrone (Mitox) and irinotecan (Iri) have so far been used as embolisation systems in the treatment of malignant hypervascularized tumors.8 These beads are composed of a sulphonate-modified polyvinyl alcohol based polymer synthesized by a suspension polymerisation reaction and formulated into beads of varying sizes ranging from 100–900 μm.9 Drug release is by ion exchange and is sustained over many days in vitro.10, 11 In PC, these bead systems could provide a precisely controlled release of the chemotherapeutic agent into the tumor bed, because the amount of chemotherapeutic agent loaded into the beads can be controlled.12 Furthermore, the potential benefits of such beads with a sustained release of chemotherapy over time are a delivery of large amounts of drugs to tumors for a prolonged time, thereby decreasing plasma levels.13 This should improve drug targeting of tumors, maximize drug potency, and minimize systemic toxicity.

The rationale of the present study was to determine the effectiveness of these novel DEBs in a model of experimental PC, and to compare the systemic side effects with the corresponding free drugs.

Material and methods



DMEM, DPBS, streptomycin, penicillin and trypsin were purchased from PAN Biotech (Aidenbach, Germany) and FCS purchased from PAA Laboratories GmbH. DMSO, Doxorubicin hydrochloride, irinotecan hydrochloride, mitoxantrone dihydrochloride were purchased from Sigma Aldrich (Steinheim, Germany). Light Cycler FastStart DNA Master SYBR Green I was obtained from Roche Molecular Biochemicals (Mannheim, Germany) and always stored at −20°C until use. Alexia Fluor 488-conjugated annexin V, Propidium iodide (PI), and annexin binding buffer were provided by Molecular Probes (Eugene, OR) inside the Vybrant Apoptosis Assay Kit#2. DC Bead Drug Delivery Embolisation Systems were kindly provided by Biocompatibles UK (Farnham, UK).

Bead loading technique

Saline was removed from the DC Bead vials by using syringes with a small gauge needle. Reconstituted drug solution (4 ml of 10 mg/ml solution of mitoxantrone dihydrochloride or 2 ml of 25 mg/ml solution of doxorubicin hydrochloride) was added directly to the vial of the DC beads. DC Bead solution was agitated gently and allowed to stand 120 min until the beads were fully loaded. Beads were then centrifuged and the supernatant was discarded. This resulted in 20 mg/ml of mitoxantrone and 37.5 mg/ml of doxorubicin loaded beads. Irinotecan Beads (50 mg/ml) were provided preloaded by Biocompatibles (Farnham, UK). Beads were stored at 2–8°C until further use.

Tumor cells

Wild-type C-26 murine colon-carcinoma cells (referred to as C-26 cells), syngenic to BALB/c mice, were provided by the American Type Culture Collection™ (Manassas VA). Cells were cultivated in DMEM medium supplemented with 10% (v/v) heat-inactivated fetal bovine serum, 100 U/ml penicillin, 100 μg/ml streptomycin and 1% (v/v) glutamine in a humidified atmosphere of 95% air and 5% CO2 at 37°C. Stock cultures were stored in liquid nitrogen and used for in vitro experiments within 5 passages. The EGFP transfected C-26 cells14 were maintained under the same conditions as the wild-type cells. Although for in vitro studies wild type C-26 cells were used, EGFP-C-26 cells were used for in vivo studies. Here, cells were harvested from subconfluent cultures. Cell viability determined by trypan blue exclusion. The number of the cells was adjusted to 1 × 106/ml for intraperitoneal injection to produce peritoneal metastasis in mice7 using a Beckman Coulter Analyzer (Coulter Cooperation, Miami, FL).

In vitro assays

Cell proliferation assay

Cell viability and proliferation was measured using a CellTiter 96 Aqueous One Solution Cell Proliferation Assay (Promega, Germany). For measurements cells were grown in 25 mm tissue culture flasks as described earlier, rinsed with PBS, trypsinized and diluted in DMEM medium to 2 × 103 cells/100 μl. Cells were thereafter plated into each well of a 96 well plate by adding 100 μl of above prepared suspension and incubated overnight at normal maintenance conditions to enable them to seed well. The next day, the cells were exposed to different concentrations of doxorubicin, mitoxantrone and irinotecan DEBs or the free drugs for 24, 48 and 72 hr. For negative controls cells were incubated in normal medium. By the end of the incubation time 20 μl of CellTiter 96 Aqueous One Solution reagent was added in each well and the plate placed in cell culture incubator for 3 hr. Thereafter the absorbance was recorded at 490 nm using a 96 well Elisa plate reader (Bio-Rad, Germany).

Flow cytometry

Cells were harvested from culture plates by trypsinizing and centrifuged (1200 RPM for 5 min). Cells (2.5 × 105) were added to each well and incubated for 24 hr. On the next day different concentrations of doxorubicin, mitoxantrone and irinotecan DEBs or free drugs were added into the wells and incubated for 48 hr. As a control, cells were incubated with camptothecin (cam) 4 μg/ml for 48 hr. After the incubation time supernatants which contained already detached cells were collected and added to labelled FACS tubes. Cells were washed with cold PBS and trypsinized. The harvested cells were added to the supernatant. The tubes were centrifuged (1200 RPM for 5 min) and supernatants were carefully aspirated. Cells were resuspended with 1× Annexin Binding Buffer and adjusted to 1 × 106 cells/ml. Five microliter Alexa Flour 488 conjugated Annexin V and 1 μl of the 100 μg/ml PI working solution were added into each FACS tubes and then 100 μl of the cell suspensions were added. The cells were incubated at room temperature for 15 min before 400 μl of 1× Annexin Binding Buffer were added into the cell suspensions and mixed gently. The tubes were placed on ice and analyzed by flow cytometry with a FACS Calibur flow cytometer (Becton Dickinson, Germany). Cells (5 × 104) were counted. Analysis was carried out using the Cell Quest software. Only Annexin V stained cells that were PI negative were considered to be apoptotic.


For measurements cells were grown in 25 mm tissue culture flask as described earlier, rinsed with PBS, trypsinized and diluted in DMEM medium to 6,000 cells/100 μl. Cells were thereafter plated into each well of a 96 well plate by adding 100 μl of the prepared suspension provided in the US-Biological DNAss Apoptosis Bioassay ELISA Kit (US-Biologicals) and incubated overnight at normal maintenance condition to let them seed well. The next day, the cells were exposed to different concentrations of doxorubicin, mitoxantrone and irinotecan DEBs for 48 hr. Later cells were fixed by centrifugation and adding 80% methanol. Then the fixative was removed and the plate was dried. For denaturation of DNA, formamide was used and the plate was heated at 75°C for 20 min. For binding nonspecific sites we used nonfat 3% milk. Afterward the plate was incubated with the provided apoptosis antibody mixture, washed with buffer, and the reaction stopped. For positive control 0.3 μg/ml single stranded DNA was used. Thereafter the absorbance was recorded at 405 nm using a 96 well ELISA plate reader.

In vivo experiments

Animal experiments

Experiments were performed in accordance with German legislation on protection of animals and the Guide for the Care and Use of Laboratory Animals.15

A 10–12 week old female BALB/c mice were purchased from Janvier (France). Before experiments mice rested for a week while caged in groups of 8. Mice were maintained in macrolon cages (size 3) in the temperature (21°C) and humidity-controlled (56%) 12 hr light-cycle environment of the Animal Care Faculty of the Universitätsklinikum Mannheim. Peritoneal metastases were generated by injecting EGFP-C-26 cells (1 × 106 cells in 1 ml PBS) into the peritoneal cavity.7, 14 For the drug treatments 8 animals were used in each group. Control groups were treated with unloaded beads on the day 7, 10 and 12 respectively.

Two groups were treated with a single dose of beads or free drug on the day 12 day or the compounds were applied on day 7, 10 and 12. By the end of day 15 animals were anesthetized with Xylazin (Rompun®) 16 mg/kg and Ketamine-Hydrochloride (Hostaket®) 120 mg/kg i.p. To document intraperitoneal tumor growth, we used intravital microscopy. EGFP-expressing C-26 cells emit green light when excited with blue light.16 This green fluorescence was used to identify in the peritoneal cavity. A FITC filterset (515 and 525 nm) (Carl Zeiss GmbH, Germany) was used for excitation of GFP and a 100 W mercury lamp was used for epiillumination (Carl Zeiss GmbH, Germany).

Animals were sacrified under narcosis by cervical translocation. After midline laparotomy the intra abdominal tumor spread was documented. Intra abdominal tumor spread was quantified using a modified Sugarbaker-index as described in our previous publication.7 In brief we classified as follows:

0, no macroscopic tumor; 1, tumor nodules ≤2 mm; 2, tumor nodules 2–5 mm; 3, tumor nodules ≥5 mm. Total tumor volume in each mouse was calculated as the sum of volumes of all peritoneal tumor nodules found in each animal. Peritoneal cancer usually presented as half spherical nodules adjacent to the peritoneum, the serosal surface of organs. The volume ‘V’ of each tumor nodule was estimated as half of a spherical volume with a radius ‘r’ using Segner's method: V = 1/2 (4/3 r3π). Total tumor volume per animal is given as the sum of tumor-nodule volumes (mm3). Thereafter whole omentum was taken and tissue samples were frozen in liquid nitrogen and stored at −80°C.

Tissue preparation and DNA purification

To purify genomic DNA from tissues a modified QIAmp DNA Mini tissue protocol was used. First, the wet weight of the whole organ was determined, then the organ was dissected to small pieces and 60 μl of PBS were added for each 25 mg of tissue. After mechanical homogenization 60 μl homogenate were added to 120 μl of ATL buffer and from this point the QIAmp DNA Mini tissue protocol proceeded according to the manufacturer's instruction. The concentration and purity of the DNA samples were determined a GeneQuant pro RNA/DNA (Pharmacia Biotech, Cambridge, England). The ratio of absorbance at 260 and 280 nm of the DNA samples were higher than 1.7. Quantitative Real Time PCR for EGFP was performed on a LightCycler apparatus (Roche Molecular Biochemicals, Mannheim, Germany) using 100 ng of genomic DNA derived from the omentum. The reaction mixture consisted of 2 μl of LightCycler FastStart DNA Master SYBR Green I (FastStart Taq DNA polymerase, reaction buffer, dNTP mixture (with dUTP instead of dTTP), SYBR Green I dye and 10 mM MgCl2), 2.4 μl of 25 mM MgCl2, 0,5 mM of EGFP specific primers, 5′-TAC GGC AAG CTG ACC CTG AAG TTC-3′ (sense) and 5′-CGT CCT TGA AGA AGA TGG TGC-3′ (antisense)14, 17 and 2 μl of DNA derived from different preparations.

The cycling program consisted of 600 sec initial denaturation at 95°C, 7 sec 95°C denaturation, 64°C annealing for 5 sec and 72°C extension for 10 sec with a transition rate of 20°C/sec between temperature plateaus for a total of 35 cycles. Quantification data was analysed using the LightCycler analysis software version 3.5. As standard the plasmid EGFP cDNA was used. The 2nd derivative maximum method of the LC software was used for generating the standard curve which was designed by using 10-fold serial dilutions of EGFP plasmid DNA. The standard curve from the 1st run exported and later imported into the different runs of experiments. The error point was <0.1, the slope <3.3 Cp and the regression coefficient was r = −1.00. PCR products were analyzed on a 1% Agarose gel to ensure specifity.


Data are displayed as box-plots showing means and standard deviation or as heat map (MTT-tests). All in vitro experiments were reproduced five times. In vivo, 8 animals per groups are used. We compared each treatment group versus controls and other groups by one-way analysis of variance (ANOVA). If ANOVA indicated a significant difference between groups, pair wise multiple comparison of all means and posthoc testing using Tukey's method were employed to determine significant differences between groups and controls. Differences were considered significant if p < 0.05. SPSS 10.0.7 software (SPSS 1989–1999) was used for statistical analysis.


We evaluated the therapeutic efficiency of the compounds in vitro and in vivo. A flow chart of the experimental set up is shown in Figure 1.

Figure 1.

Flow chart of the experimental protocol.

In vitro, in the cell proliferation assays were performed for the free compounds and the doxorubicin, irinotecan and mitoxantrone DEBs (Fig. 2), different drug concentrations were applied. The effect of drugs on cell growth was determined at 24, 48 and 72 hr. As negative control we used cells which were incubated with PBS or unloaded beads at the same time points. An advantage could be shown for doxorubicin DEBs with a significant inhibition of cell growth (p < 0.05 for 1, 5, 10 and 25 ng/μl when compared to free doxorubicin on the 1st day). No effect could be found for free doxorubicin on the 1st day. At later time points no differences could be demonstrated between the free and encapsulated drug. For irinotecan we could show a significant inhibition of cell growth for the free drug when compared to beads and the control (p < 0.05). For Mitoxantrone, the free drug inhibited cell growth more efficiently than mitoxantrone DEBs used at the same concentrations. On the 1st day, less cell growth was detected using 1–10 ng/μl of the free drug when compared to the bead group (p < 0.05). At later time points at higher concentrations, the effect was less pronounced.

Figure 2.

Inhibition of cell proliferation was monitored using a MTT-Test. Data is given as heat map. C-26 cells were incubated with free doxorubicin and doxorubicin DEBs (a,b) free irinotecan and irinotecan DEBs (c,d) or free mitoxantrone and mitoxantrone DEBs (e,f) over the time spans indicated.

Flow cytometry was performed to determine the proportion of apoptotic cells after incubation with the beads or the drugs over 48 hr (Fig. 3). For positive control and to set the gates for FACS, cells were treated with camptothecin at 4 μg/ml. Free doxorubicin led to apoptosis in more cells at all concentrations used here, while bead treatment resulted in apoptosis in only 38% of the cells (p < 0.005). For irinotecan DEBs and free drug only a slight effect could be demonstrated (p < 0.005 for beads when compared to the control group). For mitoxantrone, incubation with 5–10 ng/μl of the free compound resulted in apoptosis in significantly more cells than after incubation with mitoxantrone DEBs (p < 0.005 against control and mitoxantrone DEBs. No differences were found for lower concentrations between mitoxantrone DEBs and free drug.

Figure 3.

Apoptosis induction by incubation with free or encapsulated drugs over 48 hr. (a) apoptosis induced by doxorubicin, (b) apoptosis induction after incubation with irinotecan and (c) apoptosis induction by mitoxantrone.

To determine the optimal concentration of the encapsulated drug compounds for later in vivo experiments we investigated denaturated DNA which was detected by using a monoclonal antibody binding to ssDNA (Fig. 4). Doxorubicin showed a dose-response curve in which higher doses seemed to induce more apoptosis. Irinotecan showed little effect at all. 100 ng/μl mitoxantrone induced apoptosis in only 20% of the cells, higher doses let to a significantly higher apoptotic activity.

Figure 4.

Apoptosis as detected by ELISA for ss-DNA. (a) apoptosis induced by doxorubicin, (b) apoptosis induction after incubation with irinotecan and (c) apoptosis induction by mitoxantrone.

In vivo, peritoneal metastasis of colorectal cancer in BALB/c mice was generated by intraperitoneal injection of 1 × 106 C-26 cells into the peritoneal cavity. Metastatic growth could be detected on the intestinum, on the peritoneum of the lateral and posterior abdominal wall, on the greater and the lesser omentum, and on the mesenterium (Fig. 5, Tables I–II). All mice which received C-26 developed peritoneal metastasis. On day 15 all mice had developed a disseminated PC.

Figure 5.

15 days after intraperitoneal inoculation of EGFP-transfected C-26 cells. Using intravital microscopy, upon excitation by green light single solid metastases (a) or disseminated tumor growth (b) can be detected (200×).

Table I. Sugarbaker Index for Animals Treated by Free Doxorubicin and Doxorubicin DEBs
Ctr 10 (5–15)9.75 (5–19) 
db1 14.3 (10–20)5.2 (4–8) 
df1 7.4 (5–10)5 
db3 5  
db100 8 (5–10)  
Table II. Sugarbaker Index for Animals Treated by Free Free Mitoxantrone and Mitoxantrone DEBs
Ctr 12.8 (7–22)16 (7–37)1
mb1 11.5 (10–15)13 (5–23) 
mf1 27.88 (20–37)7.75 (3–18)2
mb3 8.8 (2–20)  
mf3 3.9 (2–5)  
mb100 11.9 (5–30)5 (4–6) 

Because Irinotecan had shown no substantial inhibition of cell growth or apoptosis induction in vitro in the doses applied here while these doses were well above the IC50 concentrations described for other colorectal cancer cell lines,18 we excluded irinotecan in the in vivo eperiments and used doxorubicin and mitoxantrone in this model. Both loaded DEB formulations were shown to be easily injectable for subsequent i.p administration in vivo.

To determine the side effect of the compounds, the body weight of animals was measured on day 15 after cell application (Fig. 6). For doxorubicin, no difference in body weight could be observed in the animals which received DEBs (25 mg/kg) or free drug (10 mg/kg) only once on day 12 (df1 when compared to db1). No mortality was observed in the group treated by DEBs while 30% of the animals died in the df1 group. In the groups which received doxorubicin DEBs (free drug ad 10 mg/kg and doxorubicin DEBs at 25 mg/kg) trice on day 7, day 10 and day 12 (df3 and db3 respectively), all animals died in the df3 group when compared to 10% mortality in the db3 group. In the group which received 100 mg/kg doxorubicin beads (db100) a mortality of 10% was observed. This indicated more toxicity of the free drug.

Figure 6.

Weight loss observed in tumor bearing animals treated with doxorubicin (a) or mitoxantrone (b) ctr-treated with unloaded beads on days 7, 10 and 12 day; db1-treated with 25 mg/kg of doxorubicin DEBs on day 12; db3-treated with 25 mg/kg doxorubicin DEBs on days 7, 10 and 12. df1-treated with 10 mg/kg doses of free doxorubicin on day 12; (df3)-treated with 10 mg/kg of free doxorubicin on days 7, 10 and 12 (all animals died in this group) and db100-treated with 100 mg/kg doses of doxorubicin DEBs on day 12. The groups treated by mitoxantrone are: ctr-treated with unloaded beads on days 7, 10 and 12; mb1-treated with 20 mg/kg doses of mitoxantrone DEBs on day 12; mb3-treated with 20 mg/kg doses of mitoxantrone DEBs on days 7, 10 and 12; mf1-treated with 10 mg/kg doses of free mitoxantrone on day 12; mf3-treated with 10 mg/kg doses of free mitoxantrone on days 7, 10 and 12; mb100-treated with 100 mg/kg doses of mitoxantrone DEBs on day 12.

In parallel, more weight loss was observed in the groups treated with free mitoxantrone. Mice which received free mitoxantrone (10 mg/kg) once on day 12 (mf1 group) showed a weight average of 18 ± 1.7 g and a mortality of 20%. Mice which received the compound trice (mf3) the avarage body weight was 14 ± 0.77 g (p < 0.05 between control and mb3). Here 40% of the animals died. Mice which were treated by mitoxantrone beads (20 mg/kg) either once on day 12 (mb1) or trice on day 7, day 10 and day 12 (mb3) showed no mortality. Weight average in the animals treated with beads remained stable when compared to the controls. Only in the mb100 group (Mitoxantrone beads 100 mg/kg applied once on day 12) weight was reduced significantly (p < 0.05 between controls and mb100). Here 10% of all animals died. Therefore, again more toxicity could be observed for the free mitoxantrone than the DEBs.

As variables of the tumor load, the Sugarbaker Index for peritoneal carcinoma, the tumor volume and the number of EGFP-DNA-copies were determined. Total tumor volume for each treatment group was calculated by the sum of the tumor volumes for each animal (Fig. 7). For doxorubicin beads and free drug treated groups, total tumor volume in the db1 group was not different from the control group. Less tumor volume was detected for the df1, db3 and db100 groups when compared to the controls (p < 0.05, Table I). For these 3 treatment groups no difference could be observed in the total tumor volume. As all of animals of df3 group died before day 15, tumorload could not be determined in this group. We determined the amount of cDNA copies using RTPCR (Fig. 8). After doxorubicin treatment, most DNA copies could be determined in the control group. In df1 and db3 only a very low amount of EGFP-DNA copies could be found (p < 0.05 for both of groups when compared to the controls). No difference could be observed between EGFP-DNA copies in the db1 and the db100 group. Both groups differed significantly from the controls (p < 0.05). When the Sugarbaker index was determined for the animals, no SB3 tumor nodules were found in any treatment group. For doxorubicin all treatment groups showed less tumor nodules than the controls. Least tumor nodules could be counted in the groups with repeated application of the doxorubicin beads. Also, beads loaded with a higher dose of doxorubicin proved to be effective.

Figure 7.

Intraperitoneal tumor volume on day 15 after inoculation of EGFP-C-26 cells and i.p treatment by free and encapsulated doxorubicin (a) or mitoxantrone (b).

Figure 8.

Intraperitoneal tumor load in 25 mg of mesenteric tissue on day 15 after tumor cell inoculation as determined by real time PCR for EGFP.

In the animals treated by mitoxantrone, less tumor volume was detected in mb3 and mf3 (p < 0.05 between control and these groups). No significant differences in tumor volumes could be found between the groups treated by beads or free drugs treatment. For mitoxantrone (Table 2) SB3 nodules could be found in the controls, but also in the group treated by a single application of the free drug (mf1). More SB1 nodules could be counted in this group while the group treated by mitoxantrone beads showed less tumor nodules than the control and mf1. Least nodules could be found in the mf3 group followed by the mb3 and mb100 groups. For groups treated with mitoxantrone beads and free drug most tumor load could be determined in the control group. Mf1, mb3 and mb100 groups demonstrated almost the similar EGFP-DNA copies. The number of DNA copies in mb1 group was not significantly different from the control group. Least DNA copies were found in the mf3 group.


PC is a fatal short-term condition amenable only to palliative treatment. It is generally considered as a systemic disease at clinical presentation, and is resistant to standard treatments.19 One treatment approach is cytoreductive surgery combined with the intraperitoneal administration of cytotoxic agents.20 This may diminish any residual tumor after macroscopic excision and may overcome the pharmacokinetic limits of systemic chemotherapy. Intraperitoneal chemotherapy gives high response rates within the abdomen because the “peritoneal plasma barrier” provides dose-intensive therapy21 since high-molecular-weight substances are confined to the abdominal cavity for long time periods. This means that the exposure of peritoneal surfaces to pharmacologically active molecules can be increased considerably by giving the drugs via the intraperitoneal route rather than the intravenous route. However, the use of intraperitoneal chemotherapy in the past has met with limited success and acceptance by oncologists. There have been 3 major impediments to greater success: (a) intracavitary instillation allows very limited penetration of drug into tumor nodules, (b) nonuniform drug distribution by adhesions and (c) by gravity.22 Intraperitoneal chemotherapy utilizing DEBs may here represent an alternative treatment modality since the drug is delivered constantly over time thus allowing a more complete intraperitoneal distribution while minimizing systemic side effects.23 Here, DEB loaded with doxorubicin, irinotecan or mitoxantrone have been evaluated in vitro and in vivo for peritoneally metastasized colorectal carcinoma.

Up to now, only the use of all 3 free (not encapsulated) compounds has been described in advanced colorectal cancer. Irinotecan is used in the regimen FOLFIRI which consists of infusional 5-fluorouracil, leucovorin, and irinotecan.24 Irinotecan is activated by hydrolysis to SN-38, an inhibitor of topoisomerase I. This is then inactivated by glucuronidation by uridine diphosphate glucoronosyltransferase 1A1 (UGT1A1). The inhibition of topoisomerase I by the active metabolite SN-38 eventually leads to inhibition of both DNA replication and transcription.25 Doxorubicin, the 2nd compound used here, is known to interact with DNA by intercalation and inhibition of macromolecular biosynthesis.26 This inhibits the progression of the enzyme topoisomerase II, which unwinds DNA for transcription. Doxorubicin stabilizes the topoisomerase II complex after it has broken the DNA chain for replication, thereby stopping replication. Mitoxantrone, the 3rd drug used to load the DC beads, is an Anthracenedione (not an anthracycline) antineoplastic agent used in the treatment of certain types of cancer, mostly metastatic breast cancer, acute myeloid leukemia, and nonHodgkin's lymphoma. In clinical combination studies for colorectal cancer have shown that mitroxantrone can be combined at full doses with other agents without unexpected toxicities,27 however also without impressing effect.

Here we focused on the comparison between free and encapsulated chemotherapeutic compounds. Irinotecan showed little effect in vitro while mitoxantrone and doxorubicin demonstrated a dose-dependent reduction in cell proliferation whether free or delivered by DEB. The antitumoral effect of the drug compounds, measured by Annexin V-staining, is mainly apoptotic, as demonstrated before.28 No difference in apoptosis induction was observed for DEB and the free compounds. Although recently, irinotecan has been shown to be effective in an animal model of experimental liver metastasis,29 we found it largely without effect in the in vitro assays. The largely missing therapeutic effect of irinitecan is most appalling. Irinotecan is subject to extensive metabolic conversion by various enzymatic systems in the body. The conversion of irinotecan to the more active form SN-38 requires carboxylesterase. A clear relationship between carboxylesterase level and the chemosensitivity of human small-cell and nonsmall-cell lung cancer cell lines has been demonstrated in vitro, where irinotecan resistance was encountered in cell lines with low carboxylesterase expression.30 Furthermore irinotecan inhibits topoisomerase I - unlike doxorubicin and mitoxantrone which iact on topoisomerase II. Topoisormase I expression has been described to be proportional to irinotecan cytotoxic effects in yeast systems and mammalian cell lines.31 So far, no information is available on the expression of these enzymea in C-26 cells.

The present study deals with 2 aspects of experimental oncology, treatment and drug toxicity. In vivo administration of free drug and DEB was by either a single dose on day 12 or 3 repeated doses on day 7, 10 and 12. A small decrease in body weight was observed for the single administration of mitoxantrone DEBs and a more significant decrease associated with 3 × free mitoxantrone administration. 3 × free doxorubicin administration was lethal to the mice, whereas 3 × doxorubicin/mitoxantrone DEBs was well tolerated. Single or multiple administrations of DEBs were also effective in reducing tumor volume, multiple administrations of either doxorubicin or mitoxantrone DEBs efficiently reduced tumor volume to between 5 and 15% of that of the control. The free drugs, when introduced intraperitoneally, had efficiency against tumor cells in vivo, but multiple applications of free drug were lethal or had significant effect on body weight indicating poor tolerability.

We conclude that multiple applications of DEBs were well tolerated and had a significant effect on reducing tumor volume. We find that the application of DEBs into the peritoneal cavity is a safe and well tolerated procedure in this animal model and may help to target chemotherapeutic agents specifically to metastatic peritoneal cancer.


The authors thank Ms. Professor H. Allgayer head of Department of Experimental Surgery, Medical Faculty Mannheim, University Heidelberg for her helpful discussions.