Twenty-four nude mice (6 to 8-week-old females) of the NMRI strain obtained from Harlan Winkelmann GmbH (Borchen, Germany) were maintained in groups of six animals per cage in laminar flow hoods in a pathogen-free environment. They were allowed access to food and water ad libitum. The study was reviewed and approved by the Animal Care Committee of the local government in accordance with the national guidelines for animal care (German Law for the Protection of Animals).
We used cells of the human gastric adenocarcinoma cell line 23132/87 [32], which were kindly provided by Prof. H.P. Vollmers, Institute of Pathology, University of Würzburg. The cell line is available from the German Collection of Microorganisms and Cell Cultures (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMS), Braunschweig, Germany). The tumour cells were cultured as a monolayer in RPMI 1640 medium supplemented with 100 U/ml penicillin, 100 μg/ml streptomycin, 2 mmol/L glutamine, 1 mmol/L sodium pyruvate, 10% heat-inactivated fetal calf serum (all products from Invitrogen-GIBCO, Germany). The cells were routinely tested for mycoplasma contamination to ensure that only negative cells were used [33].
Cells of the carcinoma cell line 23132/87 were tested in vitro for their ability to metabolise glucose to lactate in the presence of oxygen. This aerobic glycolytic activity of glucose metabolism may correlate with a unique tumour phenotype characterised by a higher metastatic and invasive potential [6]. The glucose uptake by the tumour cells was monitored with the fluorescent deoxyglucose analog 2-(N-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)amino)-2-deoxyglucose (2-NBDG) [34]. Since no corresponding benign gastric cancer cell line was available, HUVEC cells (Promocell, Heidelberg, Germany) were used as control cells and tested in parallel. In brief, cells were seeded at 100,000/0.5 ml/well in a 24-well plate and after 24 h of culture the cells were washed with PBS and cultured in glucose-free DMEM medium (PAA Laboratories, Linz, Austria) for 15 min. The cells were then incubated with 0.01, 0.1 and 1 mmol/l 2-NBDG (Invitrogen/Molecular Probes, Karlsruhe, Germany) for 10, 30, and 60 min. at 37°C. The 2-NBDG uptake was stopped by washing the cells twice with ice-cold PBS. Negative controls were cells incubated with 2-NBDG on ice and cells incubated without 2-NBDG at 37°C. Following trypsination, cells were resuspended in 100 μl PBS/10% FCS and counterstained with 1 μg/ml Propidium Iodide (PI). Ten thousand PI-negative cells were measured in a FACS Scan flow cytometer (Becton Dickinson, Heidelberg, Germany) and data were analysed by the free WinMDI 2.8 software package (The Scripps Research Institute, USA). Lactate production was assessed in cell supernatants of 20,000 cells/100 μl/well in a 96-well plate incubated for 24 h with the L-lactic acid detection kit (Roche/R-Biopharm, Darmstadt, Germany) following the manufacturer’s instructions. The colour reaction was measured in an absorption plate reader (Sunrise Absorbance Reader; Tecan, Crailsheim, Germany) at 340 nm. For this purpose the measuring volume was scaled down to 200 μl.
For the in vivo experiments a freshly thawed tumour cell aliquot was cultured for up to three passages in vitro not more than 3 weeks prior to injection into nude mice. All mice received tumour cells from the same cell passage. The cultured cells (nearly 70–80% confluence) were harvested with trypsin EDTA (PAA) and the viability of the detached cells was routinely checked with the trypan blue exclusion test. All tumour cells prepared for inoculation were highly viable (>90%). They were inoculated subcutaneously in both hind flanks (2.5 × 106 cells per flank). Tumour nodules appeared approximately 8–10 days following cell injection and the larger of the two was selected for analyses. All data presented is based on these tumour nodules. No animal died from tumour growth. The tumour size was measured with calipers and the tumour volume VT (mm3) was calculated using the ellipsoid formula A
2 × B × π/6, where A represents the smaller diameter. Endpoint for the experiments was attainment of a tumour volume between 600 and 700 mm3 (target tumour volume), with the interval between subcutaneous tumour cell inoculation and the endpoint defined as the survival time. Tumours reaching the target tumour volume were dissected and the final target volume and wet weight were determined. Subsequently, they were cut through the median, one part was fixed in formalin and embedded in paraffin, the other part was embedded in Tissue-Tek (Sakure Finetek Europe B. V., Zoeterwoude, The Netherlands) and snap frozen in liquid nitrogen.
All mice received a nutritionally balanced diet (altromin 1430) provided by the special animal feed manufacturer Altromin GmbH & Co. KG, Lage, Germany, prior to the inoculation of tumour cells. This standard diet was supplied in pellets delivering 12.8 kJ/g gross energy and consisting of 7.0% fat, 23.8% protein, and 36.4% carbohydrates (Table 1). The ketogenic diet consists of a mixture of fresh, high quality food homogenized into a paste using a standard food processor. A similar diet is being examined in an ongoing clinical trial at the University of Würzburg Hospital for the treatment of incurable tumour patients. The paste consists of 40.7% curd cheese (40% fat), 19.9% mackerel, 8.1% blue veined cheese, 8.1% white veined cheese, 8.1% bacon, 4.0% Tavarlin bread (Tavartis, Otzberg, Germany), 8.1% flaxseed, and 2.7% sesame seed. The paste (737 g) enriched with 300 ml of Tavarlin oil and 100 ml Tavarlin lactate drink (both Tavartis) was autoclaved, aliquoted in petri dishes under sterile conditions and stored at -20°C. Three dishes per six animals were thawed overnight at +4°C prior to feeding. This diet delivers 15.4 kJ/g gross energy and consists of 35.5% fat, 13.0% protein, and 0.2% carbohydrates (Table 1) and contains 21.45% MCT. The Tavarlin oil mixture consists of 28.7% saturated fatty acids, 35.7% unsaturated fatty acids, and 35.6% polyunsaturated fatty acids with a ratio of omega-6/omega-3 of 1.77:1. Chemical analyses of the nutrient contents were performed by the Chemical Laboratory Hameln (Dr. Kaiser & Dr. Woldmann GmbH, Hameln, Germany). Following tumour cell injection on day 0 the animals (n = 24) were randomly split into two equal feeding groups: standard diet (SD) and ketogenic diet (KD). Tumour size and body weight of all animals were measured every second/third day.
Blood glucose and β-OHB levels were measured on the day of tumour cell injection (day 0) and every week thereafter until the last day of the experiments before tumour resection. Measurements were done with a blood glucose and ketone monitoring system (Precision Xtra, Abbott Laboratories, Abbott Park, Illinois, U.S.A.) and corresponding test strips (Abbott GmbH & Co. KG, Wiesbaden, Germany) using 2 μl of peripheral blood collected from a snipped tail vein of each animal.
Tumour tissue sections 2-μm thick were deparaffinized with xylene and rinsed in decreasing concentrations of ethanol prior to unmasking by heating for 5 min with 10 mmol/L sodium citrate buffer in a microwave oven at 600 W. After irrigating in distilled H2O, the endogenous peroxidase was quenched with 3% hydrogen peroxide in methanol for 10 min. Cryosections (5 μm) were fixed in acetone and subsequently air-dried. All sections were then washed with PBS, blocked for 15 min in 1% goat serum, and incubated with primary monoclonal mouse antibodies for 60 min. The following antibodies, diluted in a commercial antibody diluent (DAKO, Hamburg, Germany), were used: mouse-anti-transketolase like enzyme 1 (TKTL1; clone JFC12T10, Linaris GmbH, Wertheim, Germany), final dilution 1:400; mouse anti-pan cytokeratin (clone KL1, Immunotech, Marseille, France), final dilution 1:100; anti-Ki-67 antigen (clone MIB-1, DAKO), final dilution 1:50; and rabbit anti-glucose transporter type 1 (Glut-1; G3900-01, US Biologicals, Swampscott, MA, USA), final dilution 1:100. The slides were washed in PBS, incubated with biotinylated anti-mouse and anti-rabbit immunoglobulins (LASB-kit, DAKO), and treated with streptavidin-peroxidase (LASB-kit, DAKO) according to the manufacturer’s protocol. After development in 5% 3,3′-diaminobenzidine (DAKO) and counterstaining with haematoxilin, the sections were dehydrated in graded ethanol and embedded in Vitro Clud (Langenbrinck, Emmendingen, Germany).
Cryosections of dissected tumours were stained with a rat anti-mouse CD34 antibody (RAM34, BD Pharmingen, Heidelberg, Germany), final dilution 1:100. Microvessel density was qualitatively assessed by examining the entire vital cellular zone of the tumours with a light microscope at 100× magnification.
Sections corresponding to the median line were stained with haematoxyline-eosine and photographed at a magnification of 4× with a digital camera. Images of each whole section were imported into Microsoft PowerPoint and all areas with morphologically well-defined necrosis were circled using the free-hand drawing function of the program at high magnification. The complete area of necrosis per section was quantified using the “analyse particles” option of the public domain Java image processing program ImageJ 1.34 s (downloaded from the National Institutes of Health (NIH), Bethesda, MD, USA) and expressed as percentage of the section’s total area.
GraphPad Prism 4.0 software (Statcon, Witzenhausen, Germany) was used for statistical analyses. Body weight, tumour growth, plasma glucose, blood ketone levels and necrotic areas were analysed by Mann-Whitney U test to show significant differences between the KD and SD groups after the nonparametric rank order test of Puri and Sen [35]. Probability values below 0.05 were considered significant.
