The antineoplastic antibiotic taurolidine promotes lung and liver metastasis in two syngeneic osteosarcoma mouse models and exhibits severe liver toxicity

Authors


Abstract

Osteosarcoma (OS) is the most frequent primary bone tumor. Despite multiagent neoadjuvant chemotherapy, patients with metastatic disease have a poor prognosis. Moreover, currently used chemotherapeutics have severe toxic side effects. Thus, novel agents with improved antimetastatic activity and reduced toxicity are needed. Taurolidine, a broad-spectrum antimicrobial, has recently been shown to have antineoplastic properties against a variety of tumors and low systemic toxicity. Consequently, we investigated in our study the antineoplastic potential of taurolidine against OS in two different mouse models. Although both OS cell lines, K7M2 and LM8, were sensitive for the compound in vitro, intraperitoneal application of taurolidine failed to inhibit primary tumor growth. Moreover, it enhanced the metastatic load in both models 1.7- to 20-fold and caused severe liver deformations and up to 40% mortality. Thus, systemic toxicity was further investigated in tumor-free mice histologically, by electron microscopy and by measurements of representative liver enzymes. Taurolidine dose-dependent fibrous thickening of the liver capsule and adhesions and atrophies of the liver lobes were comparable in healthy and tumor-bearing mice. Liver toxicity was further indicated by up to eightfold elevated levels of the liver enzymes alanine transaminase, aspartate transaminase and GLDH in the circulation. Ultrastructural analysis of affected liver tissue showed swollen mitochondria with cristolysis and numerous lipid vacuoles in the cytoplasm of hepatocytes. The findings of our study question the applicability of taurolidine for OS treatment and may suggest the need for caution regarding the widespread clinical use of taurolidine as an antineoplastic agent.

Osteosarcoma (OS), a highly malignant bone tumor, is the second most frequent cause of cancer-related death in children and young adolescents. Metastasis, predominantly to the lung and bone, is detected in ∼ 20% of OS patients at diagnosis.1 Up to 50% of patients develop metastases after completion of initial therapy. (Neo)-adjuvant chemotherapy introduced in the late 1970s has improved the survival rate of patients with localized disease by more than 50% compared to surgery alone, but patients with metastatic disease continue to have a 5-year survival rate below 30%.1, 2 Current chemotherapeutics approved for clinical use, for example, methotrexate, doxorubicin, cisplatin, etoposide and ifosfamide, cause serious side effects including neurotoxicity, cardiotoxicity, hepatotoxicity, nephrotoxicity, hearing loss, sterility and induction of secondary malignancies (reviewed in Refs.1 and2). Consequently, novel compounds for more effective treatment of OS with minimal systemic toxicity are urgently needed.

Taurolidine [bis(1,1-dioxoperhydro-1,2,4-thiadizinyl-4)methane], a derivative of the amino acid taurin and broad-spectrum antibiotic, has originally been used for prophylactic intraoperative wound lavage and the treatment of severe surgical infections like peritonitis (reviewed in Ref.3). Taurolidine has sustained antiadherence activity against bacteria, is effective against multiple-antibiotic-resistant strains such as Staphylococcus aureus and has anti-inflammatory activity presumably due to its endotoxin-neutralizing effect (reviewed in Ref.4). More recently it has been considered as a promising anticancer agent. Antineoplastic activity of taurolidine has been demonstrated in various tumor types, including brain, colon, esophageal and ovarian cancer, as well as fibrosarcoma, glioma and melanoma (reviewed in Ref.3). We recently reported cytotoxic and antiadhesive activity of taurolidine in vitro in ten human OS cell lines with distinct genetic defects (e.g., p53, Rb) and varying metastatic potential.5 Importantly, taurolidine was found to be less cytotoxic in non-neoplastic than in tumor cells in vitro5–7 and exhibited low systemic toxicity in vivo with some tolerable side effects like vein irritation3 and respiratory depression8, 9 during i.v. infusion. Based on all these findings, taurolidine was considered an attractive new chemotherapeutic and initial clinical study with glioblastoma and gastric cancer patients revealed promising results.10, 11 Surprisingly, the number of preclinical studies with taurolidine in animal cancer models is low and some failed to demonstrate tumor suppressive properties.8, 9

In our study, the antineoplastic potential of taurolidine and drug-related side effects were investigated in a murine K7M2 OS cell line model of experimental metastasis in syngeneic BALB/c mice12 and in C3H mice that, on subcutaneous inoculation of syngeneic LM8 OS cells, develop primary tumors that metastasize to the lung and the liver.13 K7M2 and LM8 cells were transduced with a lacZ gene for sensitive detection by X-gal staining of metastatic lesions in lung and liver down to the single cell level as reported.13

Abbreviations

ALT: alanine transaminase; AST: aspartate transaminase; GLDH: glutamate dehydrogenase; H&E: hematoxyline and eosine; i.p.: intraperitoneally; OS: osteosarcoma; SEM: standard error of the mean; TEM: transmission electron microscopy; X-Gal: 5-bromo-4-chloro-3-indolyl-β-D-galactoside

Material and Methods

Cell culture and reagents

The mouse OS cell lines K7M2 and LM8 were provided by Chand Khanna (National Cancer Institute, Bethesda, MD) and Takafumi Ueda (Osaka University Graduate School of Medicine, Osaka, Japan), respectively. Both cell lines were stably transduced with a lacZ reporter gene and cultured as recently reported.13 Sensitivity of the cell lines for taurolidine was tested in a WST-1 assay as described before.5

Taurolidine (Taurolin®) was purchased from Geistlich Pharma AG (Wolhusen, Switzerland) and Polyvinylpyrrolidone (Kollidon 17 PF) was generously provided by F. Andreesen, Sales Pharma Ingredients & Services, BASF ChemTrade GmbH (Burgbernheim, Germany).

Experimental OS lung metastasis model

Eight-week-old female BALB/cAnNCrl mice (20 g average body weight) were obtained from Charles River Laboratories (Sulzfeld, Germany) at least 10 days prior to experimental commencement. Housing conditions and experimental protocols were in accordance with the guidelines of the Swiss “Federal Veterinary Office, FVO” and approved by the authorities of the canton Zurich. On day 0, 5 × 105 lacZ-transduced syngeneic K7M2 (K7M2-lacZ) cells in 200 μl phosphate buffered saline (PBS) were injected into the tail vein of the mice. On Day 1 after tumor cell inoculation, the mice were randomly distributed into a therapy (n = 10) and a control group (n = 10). The therapy group was treated intraperitoneally (i.p.) on day 1 and day 2 with 20 mg/mouse (equivalent to 1,000 mg/kg) taurolidine. Taurolidine at this dose was not well tolerated (see Results section) and the dosage was therefore reduced to 10 mg/mouse (equivalent to 500 mg/kg) and the mice were treated every second day from day 4 until the end of the study. The control group was treated with the corresponding volume of vehicle (5% [w/v] polyvinylpyrrolidone). The mice were sacrificed on day 10 and the lungs dissected. The left lobe was fixed in 2% formaldehyde/0.2% glutaraldehyde/PBS for 1 hr and subsequently incubated with 5-bromo-4-chloro-3-indolyl-β-D-galactoside (X-Gal [Axxora, Lausen, Switzerland]) as described.13 X-Gal stained metastases on the lung surface were quantified as described below.

Subcutaneous osteosarcoma mouse model

Eight-week old female C3H/HeNCrl mice (20 g average body weight) were obtained and housed as described above. On day 0, 107 lacZ-transduced syngeneic LM8 (LM8-lacZ) cells in 200 μl PBS were injected s.c. into the right flank. Two different protocols for taurolidine treatment were used: in Protocol 1, i.p. treatment of the mice with 750 mg/kg taurolidine (n = 10) or vehicle (n = 10) in 2-day intervals was started 1 day after tumor cell injection. In Protocol 2, primary tumor growth, to a measurable size (length × width2/2), was allowed for 7 days prior to treatment. Based on tumor size, mice were then divided into two indistinguishable groups of ten animals and then treated with vehicle or taurolidine according to Protocol 1. The health of the mice was monitored daily and the tumor size was measured once a week. On day 25, the mice were sacrificed under anesthesia and the lungs perfused with PBS and then fixed under inflation with 3% paraformaldehyde as described.13 The left lobe of the lungs, the two upper and the middle lobes of the livers were X-Gal stained and the lacZ expressing metastases quantified as described below.

Quantification of lacZ-expressing metastases in lungs and livers

Photographs of dissected lungs and livers were taken with a Kappa PS 20 C digital camera (Kappa Opto-Electronics GmbH, Gleichen, Germany) connected with an OpMi-1 binocular microscope (Carl Zeiss Jena GmbH, Jena, Germany) and imported as TIF files into Power Point® software. Macrometastases on organ surfaces, defined as indigo-blue stained foci > 0.1 mm in diameter, were counted. Close-ups were taken with the same camera but connected to an Eclipse E600 microscope (Nikon, AG, Egg, Switzerland) and micrometastases (<0.1 mm) were counted in ten randomly selected close-ups of 1 mm2 organ surface.

Taurolidine toxicity in healthy C3H mice

The study included 8-week old female “adolescent” (20 g average body weight) and 20-week old “adult” (40 g average body weight) mice. Mice of the same age were divided into five groups and then treated as follows. Groups 1 and 2 of each age group were treated in 2-day intervals with 750 mg/kg (high dose) and 500 mg/kg (medium dose) taurolidine, respectively. Group 3 in each age group was treated daily with 250 mg/kg (low dose) taurolidine, which resulted in the same accumulative dose of 6,000 mg/kg in all mice of Groups 2 and 3. Two control groups (Groups 4 and 5) of each age group were either treated with vehicle (5% [w/v] polyvinylpyrrolidone [Kollidon 17 PF, BASF GmbH] in aqua ad iniectabilia [B. Braun Medical AG, Sempach, Switzerland]) or left untreated. The health of all mice in the study was monitored daily. At the end of the study on day 25, the mice were sacrificed and analyzed as follows.

Histology

Formaldehyde fixed tissue samples from liver and the main visceral organs were embedded in paraffin and 3-μm thin sections were deparaffinized and stained with hematoxyline and eosine (H&E). For all tumor-free mice, additional tissue sections were stained for glycogen (periodic acid Schiff's reagent), collagen (Van Gieson stain) and reticulin fibers (Gomori's reticulin stain). Frozen sections of fixed metastases-free liver specimens were stained for lipid (Oilred O). The slides were examined and photographed under a microscope (Olympus “Vanox” [Olympus Schweiz AG, Volketswil, Switzerland]).

Electron microscopy

Small liver samples from three tumor-free taurolidine-treated mice (one high dose; one medium dose; one low dose) and from three controls (two vehicle-treated, one untreated) were fixed in 2.5% glutaraldehyde for 6 hr, then stored in 0.1 M Na–P buffer (pH 7.4) at 4°C and embedded in Epon (Sigma-Aldrich, Buchs, Switzerland). Ultrathin sections were examined by transmission electron microscopy (TEM; Philips CM 10 [Philips Electronics, Eindhoven, Netherlands]).

Measurement of liver enzyme plasma levels

Blood was collected from the caudal vena cava during autopsy, mixed with heparin and plasma was obtained after centrifugation. Aspartate transaminase (AST) and alanine transaminase (ALT) activity were measured with a Modular P instrument (Roche, Rotkreuz, Switzerland) according to the International Federation of Clinical Chemistry at 37°C with activation by pyridoxal phosphate. Glutamate dehydrogenase (GLDH) activity was determined with a KoneLab 30i instrument (ThermoScientific, Reinach, Switzerland) in a standardized enzymatic reaction (Roche, Rotkreuz, Switzerland).

Statistical analysis

Throughout the study, results are presented as the mean ± standard error of the mean (SEM) of original data or as % of control. Statistical analysis was performed with the GraphPad Prism® 5.01 software. Normal distributed (Kolmogorov–Smirnov test) unpaired datasets were analyzed with the two-tailed unpaired t-test and other unpaired data with the two-tailed Mann–Whitney rank sum test. Normal distributed paired data were analyzed with the two-tailed paired t-test and other paired data with the two-tailed Wilcoxon signed rank test. Data were considered significantly different when p < 0.05.

Results

Taurolidine promotes experimental OS lung metastasis

BALB/c mice treated by i.p. bolus injections of 1 g/kg taurolidine (20 mg/mouse) on days 1 and 2 after injection of K7M2-lacZ cells into the tail vein showed signs of severe toxicity on experimental day 3, including a shaggy and mat fur, reduced food and water intake, dimmed eyes and reduced physical activity and contact reaction. This was not observed in vehicle-treated mice. Consequently, the taurolidine dosage was reduced from daily 20 mg/mouse to 10 mg/mouse every other day, which was well tolerated up to day 10 when the mice were sacrificed and the lungs dissected for inspection of metastases (Fig. 1, upper panel). Unexpectedly, the mean number of lung macrometastases in taurolidine-treated mice (n = 9) was found significantly (p = 0.025) increased by 73% compared to that observed in vehicle-treated control mice (n = 9; Fig. 1, left graph), and the number of micrometastases was 31% higher in taurolidine-treated than in control animals (Fig. 1, right graph, p = 0.20). These findings were contradictory to the observed in vitro sensitive of the K7M2-lacZ for taurolidine (IC50 of 63 μM, data not shown). Consequently, we tested taurolidine in a second, independent OS model based on the LM8-lacZ cell line, which was similar sensitive for taurolidine in vitro (IC50 of 71 μM, data not shown).

Figure 1.

Lung metastasis of i.v. injected K7M2-lacZ OS cells is not inhibited by i.p. administration of taurolidine. Top panel, representative whole mounts of X-gal stained lungs of control mice (left) and animals treated with taurolidine (right). Bottom panel, mean (±SEM) number of macrometastases (left) and micrometastases (right) per lung control mice (set to 100%, equivalent to 33.0 ± 7.5 macrometastases and 4.5 ± 0.5 micrometastases) and taurolidine-treated animals. *p < 0.05 compared to control.

Taurolidine has no inhibitory effect on OS primary tumor growth and enhances spontaneous metastasis to lung and liver

LM8-lacZ cells inoculated subcutaneously in the flank of C3H mice grow to local primary tumors (Fig. 2a, top panel) and metastasize spontaneously to lung and liver. The growth of primary tumors in mice treated by i.p. bolus injections of 750 mg/kg taurolidine in 2-day intervals, beginning on day 7 after LM8-lacZ cell inoculation (day 7 treated mice), was indistinguishable from that in corresponding vehicle-treated mice until sacrifice on day 25 (Fig. 2a, bottom panel). Surprisingly, primary tumors in mice that received the same taurolidine dose, but from day 1 until sacrifice (day 1 treated mice), grew even slightly faster than those in the corresponding vehicle-treated mice.

Figure 2.

Subcutaneous primary tumor growth and spontaneous lung and liver metastasis of LM8-lacZ OS cells are not inhibited by i.p. taurolidine treatment. Treatment was started on day 1 (day 1 mice) or on day 7 (day 7 mice) after tumor cell injection. (a) Representative images of tumor-bearing mice 25 days after tumor cell injection (top) and mean (±SEM) tumor growth over time in mice treated as indicated (bottom). (b) and (c) Representative images of metastases-bearing lungs and livers (top panels) and mean (± SEM) numbers (percent of respective controls) of macrometastases (>0.1 mm) and micrometastases (<0.1 mm) on lung (b) and liver (c) surfaces. Mean (±SEM) numbers of metastases in control day 1 mice were 24.8 ± 7.5 lung macrometastases, 9.9 ± 2.1 lung micrometastases, 6.2 ± 3.6 liver macrometastases, 3.3 ± 0.8 liver micrometastases, and in control day 7 mice 23.1 ± 6.8 lung macrometastases, 9.1 ± 1.3 lung micrometastases, 9.8 ± 2.4 liver macrometastases and 7.1 ± 2.4 liver micrometastases. *p < 0.05, **p < 0.01, ***p < 0.001 compared to the respective control.

Moreover, metastasis to the lung and the liver was enhanced by taurolidine in both treatment regiments (Figs. 2b and 2c; upper panels). The mean number of lung macrometastases in day 1 treated mice was 248.8 ± 47.2% (p = 0.0286, n = 6) and in day 7 treated mice 283.1 ± 60.6% (p = 0.0136, n = 8) of that in respective vehicle-treated (control) mice (n = 10; Fig. 2b: bottom left). The mean number of lung micrometastases increased to 236.7 ± 39.3% (p = 0.0088) in day 1 and to 191.2 ± 36.3% (p = 0.0277) in day 7 treated mice compared to respective control animals (Fig. 2b; bottom right). In the liver, the mean number of macrometastases in day 1 and day 7 treated mice was found increased to 879.0 ± 188.7% (p = 0.0028) and 358.2 ± 153.1% (p = 0.0259) compared to respective control mice (Fig. 2c; bottom left). The increase in the number of micrometastases compared to control to 2,021.2 ± 297.00% (p < 0.0001) and to 784.5 ± 173.2% (p = 0.0008) in day 1 and day 7 treated mice, respectively, was even more dramatic (Fig. 2c, bottom right).

Unexpectedly, morphological examination of the liver of taurolidine-treated mice showed severe toxic side effects of the compound, indicated by liver atrophy and partial fusion of the left and right medial lobes (Fig. 3a) not observed in control animals. This could be histologically confirmed in paraffin sections (Figs. 3b and 3c). Liver toxicity of taurolidine was most pronounced in day 1 treated mice, suggesting dose dependency. The deformations of the liver lobes were accompanied by a strong fibrous thickening of the liver capsule, particularly in the atrophic region (Fig. 3c). Severe toxicity of taurolidine was also indicated by a significant (p = 0.0148) loss of body weight of 17.4% in day 1 treated mice between day 0 (tumor cell injection) and day 21, which was not observed in any group of control mice (Fig. 3d). Taurolidine toxicity causing weight loss appeared to be also dependent on the duration of treatment and the total cumulative dose as suggested by the more pronounced decrease in body weight of day 1 compared to day 7 treated mice, which only showed a significant (p < 0.0001) loss of weight of 8% between day 14 and day 21. Toxic side effects of taurolidine depending on the duration of the treatment and the accumulative dose became also evident when 40% of day 1 and 20% of day 7 treated mice became moribund and had to be sacrificed ahead of time while all mice in the control groups survived until the end of the study.

Figure 3.

Side effects of taurolidine in tumor-bearing mice. (a) Representative image of liver deformations in taurolidine-treated mice. Representative H&E stained tissue section of fused liver lobes (b) and of atrophic regions (c). (d) Mean (±SEM) changes of body weight over time of mice treated as indicated. *p<0.05, ***p < 0.001 compared to the respective control.

Side effects of taurolidine in tumor-free C3H mice

The toxic side effects of taurolidine observed in tumor-bearing mice raised the question whether they are exclusively drug-related and not potential health deficiencies caused by growing tumors and metastases. Thus, tumor-free adolescent and adult mice were i.p. treated as outlined in materials and methods with different doses per injection and at varying time intervals, resulting in different accumulative doses. Two of the high-dose (750 mg/kg at 2-day intervals) treated mice had to be sacrificed 3 days before the end of the experiment on day 22. Mice treated with the medium dose (500 mg/kg, 2-day interval) of taurolidine showed milder symptoms. Animals in the low-dose (250 mg/kg, daily) group and vehicle-treated control mice had no symptoms.

Livers dissected from mice treated with the high dose of taurolidine showed the same severe deformations (Fig. 4b) as those seen in the therapy study. In addition, peritoneal adhesions of the abdominal organs occurred (Fig. 4a). The livers of mice of the medium-dose treatment group showed similar, but less severe anatomical changes of the liver (Fig. 4c). The livers of low-dose (not shown) or vehicle-treated mice looked normal (Fig. 4d). Interestingly, these observations indicated that 250 mg/kg taurolidine i.p. administered daily was less toxic than 500 mg/kg injected in 2-day intervals although the cumulative dose of taurolidine per mouse at the end of the study was the same in the two treatment groups.

Figure 4.

Adhesions of abdominal organs and liver deformations in taurolidine-treated tumor-free mice. (a)–(d) Representative images at autopsy of abdominal organs and of dissected livers. Adhesions of abdominal organs in mice treated with 750 mg/kg taurolidine (a, arrows). Liver deformations were observed in mice treated with 750 mg/kg (b) and 500 mg/kg (c) taurolidine, whereas animals treated with 250 mg/kg taurolidine or vehicle-treated mice displayed no abnormalities (d). (e)–(h) Representative H&E sections of liver tissue of mice treated with 750 mg/kg taurolidine show fibrous thickening of the liver capsule (arrow) (e), covered with fibrin tags and marked vacuolation (lipid) of the cytoplasm (f) compared to corresponding liver sections obtained from vehicle-treated mice with a normal liver capsule (g) and fine granular vacuolation (glycogen storage) of the cytoplasm (h).

Histological analysis of liver sections revealed intracytoplasmic vacuolation of hepatocytes in various degrees in all investigated taurolidine-treated mice (n = 13) but also in vehicle-treated (n = 7) and untreated mice (n = 2). These vacuoles were attributed to glycogen or lipid storage. Glycogen storage was equal in both treated and untreated mice and lipid accumulation was related to inappetence. H&E staining revealed no significant histological lesions in the liver parenchyma of the taurolidine-treated mice, with the exception of focal areas of coagulation necrosis in two animals. In particular, obvious signs of hepatocellular degeneration or regeneration were absent in light microscopy. In the areas with gross atrophy of the lobes, the histological lesions consisted of a diffuse fibrous thickening of the liver capsule, which was most pronounced in mice treated with high doses of taurolidine (Fig. 4e). The thickening of the capsule was accompanied by a slight increase in number and size of reticulin fibers between subcapsular hepatocytes (Fig. 4f). The thickened capsule contained immature collagen fibers (VG-negative) and occasionally a few scattered inflammatory cells (polymorphonuclear leukocytes). Focally, there were tags of fibrin on the surface of the liver capsule. In contrast, vehicle-treated mice displayed normal liver capsules and no histological alterations (Figs. 4g and 4h). All other organs, such as myocardium, lungs, kidney, spleen and intestine were normal.

Ultrastructural analysis revealed in the high-dose taurolidine group numerous lipid vacuoles as well as swollen mitochondria with cristolysis in the cytoplasm of hepatocytes (Figs. 5a and 5b). Neither ribosome disintegration nor dilation of the rough endoplasmic reticulum was recognized. In mice treated with medium or low doses, mitochondrial swelling was less pronounced and lipid accumulation was absent (not shown). In vehicle-treated as well as in nontreated mice, the intracellular structures of the hepatocytes appeared normal (Figs. 5c and 5d).

Figure 5.

Ultrastructural changes of hepatocytes and elevated plasma levels of liver enzymes in taurolidine-treated mice. Representative TEM images of liver tissue from mice treated with 750 mg/kg taurolidine (a) and (b) show swelling of mitochondria with cristolysis (mi) and lipid storage (li) but no displacement of the ribosomes (ri). (c) and (d) Representative TEM images of liver tissue from vehicle-treated mice with normal mitochondria and rough endoplasmic reticulum and moderate glycogen storage (gl). (e)–(g) Taurolidine dose-related elevation of AST, ALT and GLDH plasma levels compared to kollidon (vehicle) control. *p < 0.05, ***p < 0.001 compared to vehicle control.

Plasma levels of the liver enzymes AST, ALT and GLDH in low-dose (n = 5), vehicle (n = 10) and nontreated (n = 2) mice were indistinguishable and considered as normal (Figs. 5e, f and g). In contrast, in mice of the high-dose treatment group (n = 5) AST, ALT and GLDH were increased 2.4- (p = 0.0209), 8.8- (p = 0.0003) and 6.2-fold (p = 0.0001), respectively, compared to normal levels in kollidon-treated mice. In mice treated with the medium taurolidine dose (n = 9), the increase of enzyme levels was moderate and only statistically significant for ALT (2.3-fold, p = 0.0489).

Discussion

Taurolidine is an antimicrobial and anti-inflammatory compound and considered as a promising antineoplastic agent. Several studies demonstrated cytotoxicity in numerous tumor cell types in vitro while normal cells were less affected (reviewed in Refs.3 and4). In colon, ovarian and prostate cancer, the antineoplastic activity of taurolidine in vitro was correlated with therapeutic efficacy in vivo.6, 14–16 Based on these findings, we investigated in our study the antineoplastic potential of the compound in two syngeneic OS mouse models. Unexpectedly, the findings in mice were contradictory to the taurolidine sensitivity of the OS cell lines in vitro and i.p. administered taurolidine exhibited severe liver toxicity in tumor-bearing and healthy mice.

Repetitive i.p. administration of taurolidine at doses shown to be effective in other in vivo-studies6, 14 failed to inhibit the growth of primary subcutaneous tumors in the LM8 model although the mouse OS cell lines were sensitive for the compound in vitro at comparable concentrations as reported for human OS cells.5 An effect of the lacZ-tagging on in vivo responsiveness of the LM8 cells to the therapy could be excluded by a control experiment with the standard cytostatic drug gemcitabine (article in preparation). Treatment with gemcitabine inhibited the growth and the formation of metastases of LM8-lacZ and nonmanipulated LM8 cell-derived tumors with indistinguishable efficacy that was comparable to that previously reported by Ando et al.17 in the same OS mouse model.

Interestingly, also a few other studies failed to show a therapeutic effect of taurolidine in vivo although they could demonstrate an antineoplastic effect on the respective tumor cell lines in vitro.9, 18, 19 In a recently published study, orthotopic bladder tumor growth in rats was even enhanced by instillation of taurolidine.8 In our study, taurolidine treatment promoted significantly experimental and spontaneous metastasis in two different OS mouse models, and this effect increased with the duration of the treatment. Potential explanations for these adverse effects include the poor health condition of the mice as a result of the taurolidine treatment, which likely disabled protective mechanisms and allowed unrestricted tumor growth and metastasis. Along these lines, a direct fatal effect of taurolidine on the immune system has to be considered. The putative transfer of methylol groups to amino groups of proteins by the taurolidine metabolites20 might impair the function of components required for a tumor-fighting immune-response. In addition, taurolidine may have facilitated the extravasation of the circulating tumor cells and the invasion of target organs by disrupting the tissue integrity and, consequently, normal tissue barriers. This would also explain the up to ten times larger increase of the metastatic burden in the liver compared to the lungs as a consequence of the severe liver toxicity of taurolidine demonstrated here.

As already mentioned above, our study also showed life-threatening toxic side effects of i.p. administered taurolidine in the mouse OS models. In the K7M2 OS model, the dosage for further treatment had to be reduced after 2 days because of the severity of the toxic side effects of taurolidine. Accordingly, mice in the LM8 OS model were treated only every other day with a lower dose of 750 mg/kg/day, but life-threatening toxicity was still evident and reflected by a significant loss of body weight of up to 17.4% and by a mortality of 40%. These observations were comparable to those of a study in a xenograft model of DU145 human prostate cancer where similar treatment (700 mg/kg/day, 3 days per week for 3 weeks) resulted in 19.6% body weight loss and 50% mortality.14 The doses used in both studies were at least four times lower than the half-maximal median lethal dose (LD50) of 4,000 mg/kg taurolidine reported to cause acute death on i.p. injection in mice.21

Acute respiratory suppression, reported to occur after fast i.v. injection of taurolidine,9 was not observed in our study and can be excluded as a cause of death. It is more likely that the mortality was associated with the hepatotoxicity of taurolidine resulting in impressive changes of liver morphology that, to our knowledge, is reported here for the first time. To further investigate the observed severe toxic side effects of taurolidine in mice, we performed an additional toxicity study in healthy mice. Importantly, the toxic side effects of taurolidine in the liver were comparable in healthy and tumor-bearing mice, indicating that tumor cells and tumor burden did not contribute to the liver phenotype caused by i.p. taurolidine treatment. Interestingly, the autopsy of the tumor-free taurolidine-treated mice showed, in addition to the liver phenotype, abnormal peritoneal adhesion of the abdominal organs and between the individual organs themselves. This was a surprising finding because taurolidine was previously shown to have antiadhesive properties in postoperative adhesion prevention studies.22–24

All toxic side effects of taurolidine reported here were independent of the age and the weight of the mice. However, when mice were treated with highest dose of taurolidine in the toxicity study, the heavier mice displayed taurolidine-induced effects earlier than mice of lower body weight. This effect may be attributed to adjustment of the injected dose to the body weight of individual animals. Further support for this interpretation is provided by the findings in mice of the medium- and low-dose groups, which indicated that the severity of liver deformations was related to the applied dosages per injection in the different treatment groups, but not to the accumulative dose at the end of the study, which was the same in the medium- and the low-dose groups.

Histological and ultrastructural analysis of the liver tissue in taurolidine-treated mice indicated a dose-dependent damage. This was supported by significant and dose-dependent elevation of AST, ALT and GLDH in the circulation of taurolidine-treated mice. Increased levels of these moderately (AST) to highly (ALT, GLDH) liver-specific enzymes indicate liver cell destruction in humans and mice.25 With all these findings, we can only speculate about the mode of the liver-damaging action of taurolidine on the basis of what is known so far. The formulation of 2% taurolidine used in our study is strongly hypotonic (145 mOsm/kg)21 and tissue swelling and in the worst case cytolysis of cells exposed to the solution has to be considered. The liver capsule usually protects the hepatocytes against mechanical forces in the peritoneal cavity. Thus, the observed thickening of the capsule in the high- and medium-dose taurolidine treatment groups could be a protective reaction on the repeated i.p. hypotonicity caused by the taurolidine bolus injections. However, such a hypotonic environment should also affect other abdominal organs, which was not observed in our study. This suggests that the liver damage is related to products of taurolidine decomposition. Interestingly, the liver was shown to take up 5–12 times higher amounts of taurolidine and/or its decomposition products than other organs, such as lung, spleen, pancreas and kidney,21 which might explain why the liver is mainly affected.

Taurolidine affects and induces several processes in (tumor) cells3, 26 but the detailed molecular mechanisms of action are largely unknown. In microorganisms, effects of taurolidine are at least in part attributed to its hydrolyzation metabolites N-methylol-taurultam and N-methylol-taurinamide, which transfer methylol groups to amino groups of proteins in the cell wall or the membrane of microbes.3 There is evidence that this effect is not restricted to microorganisms, but could also affect multicellular organisms by a transfer of methylol groups to residues such as arginine and histidine, which may impair the function of, for example, components of the complement and the fibrinolysis systems.20 Degradation of taurolidine to toxic formaldehyde and formic acid (drug information sheet), partially prevented by polyvinylpyrrolidone in the formulation, has also been proposed.27 Repetitive exposure of hepatocytes to formaldehyde or formic acid over a longer period of time is likely cytotoxic. In this context, it is important to note that, in the vast majority of in vivo taurolidine treatment studies, the compound was only administered for less than 14 days, which might explain why toxic side effects, including weight loss and increased mortality, were rarely reported.8, 9, 14

In conclusion, our study demonstrates that taurolidine is not effective as an antineoplastic agent in two experimental OS mouse models. Repetitive i.p. bolus administration of taurolidine failed to suppress primary tumor growth and even promoted experimental and spontaneous metastasis. Furthermore, taurolidine caused severe, life-threatening hepatotoxic side effects in tumor-bearing as well as healthy mice. These findings, taken together, question a future application of taurolidine as novel compound for more effective treatment of OS with less severe side effects than currently used chemotherapeutics. Moreover, recently published clinical trials with patients suffering from resectable colorectal, gastric and pancreatic cancers or advanced Stage IV melanoma also failed to demonstrate a therapeutic effect of taurolidine treatment.28, 29

Acknowledgements

The authors thank F. Andreesen of BASF ChemTrade GmbH for providing the Kollidon 17 and Mazda Farshad of the Balgrist University Hospital Zurich for statistical support. Bruno Fuchs has been supported by the Krebsliga of the Canton Zurich, the Walter L. and Johanna Wolf Foundation (Zurich), the Lydia Hochstrasser Foundation (Zurich), the Balgrist Foundation and the University of Zurich.

Ancillary