Ethical approval

The study was conducted at the Institute of Clinical Medicine, Aarhus University, Aarhus, Denmark. Approval was obtained from the Danish Animal Experiments Inspectorate (license no. 2017/15–0201-01184) and carried out in accordance with existing laws and ARRIVE guidelines. All chemical analyses were performed at the Department of Forensic Medicine, Aarhus University, Aarhus, Denmark.

Microdialysis

Microdialysis is a catheter-based method enabling dynamic collection of small samples, from any tissue of interest, called dialysates. This allows for concentration quantification of virtually any unbound drug in the extracellular matrix with a size below the membrane cut-off. The basic principle behind microdialysis is passive concentration-driven diffusion, which happens across a semipermeable membrane located at the tip of the microdialysis catheter. An equilibrium between the membrane and the surrounding tissue will never occur as the catheter is connected to a precision pump that continuously perfuses the system with a perfusion fluid at a set flow rate. This means that the concentration measured in the dialysate only represents a fraction of the absolute concentration in the tissue of interest. The fraction is referred to as the relative recovery and can be determined by several calibration methods [15]. In the present study, calibration by drug was applied meaning that after the observation period, a known concentration of a doxorubicin solution replaced the initial perfusion fluid, and two 40 min dialysates were collected under this setup.

The relative recovery (RR) could then be determined by the following equation [16, 17]:

$$}= \frac_}}- _}}}_}}}$$

Cperfusate is the concentration of doxorubicin in the perfusate during the calibration period, while Cdialysate is the concentration of doxorubicin in the dialysate during the calibration period.

The relative recovery was then used to calculate the absolute tissue concentration of unbound doxorubicin (Ctissue):

Cdialysate is the concentration of doxorubicin in the dialysate during the sampling period.

In a standard microdialysis setup, dialysates are collected in 200 μl microvials. However, in the present study, the collection was done in 1.5 ml LoBind Eppendorf tubes (Eppendorf, Hamburg, Germany) as thorough prior in vitro and in vivo experiments have shown that undesirable adsorption of doxorubicin was caused almost solely by the standard polystyrene/santoprene vials [18].

The microdialysis equipment was purchased from M Dialysis AB (Stockholm, Sweden). All microdialysis catheters had a cut-off of 20 kDa. The catheters were type 70 with membrane lengths of 20 mm and 30 mm and type 67 intravenous catheters with 30 mm membranes. The flow rate was 1 μl/min, and the perfusion fluid was saline. The concentration of the doxorubicin solution used under the calibration period was determined as the mean of three samples from the solution.

Animals and anaesthesia

Sixteen female pigs (Danish Landrace, mean weight 77 kg (range 73–83 kg), age 5 months) were included in the study and randomized into two groups of eight. The animals were kept in pens in groups of minimum two animals with a light cycle of 12 h. Straw was used as bedding, and they had access to ad libitum water. Feeding was restricted (farm pig ration) to limit weight gain. Before transportation to the surgical facility, the animals were sedated with zoletil mix ((25 mg/ml tiletamine + 25 mg/ml zolazepam) + 6.25 ml xylazine (20 mg/ml) + 1.25 ml ketamine (100 mg/ml) + 2.5 ml butorphanol (10 mg/ml) 1 ml/10 kg)). Upon arrival, the animals were placed under general anaesthesia and kept so until euthanasia with an overdose of pentobarbital at the end of the sampling period. The anaesthesia consisted of a combination of continuous intravenous infusion of propofol (40 ml/h) (Fresenius Kabi, Bad Homburg, Germany) and fentanyl (25 ml/h) (B. Braun, Melsungen, Germany). Arterial blood gas samples were taken and analysed every 2 h to monitor pH, which was within a range of pH 7.38–7.57. Body temperature was controlled with a rectal thermometer and regulated by room temperature, ventilation, cooling fluids and covers.

To minimize the risk of the anaesthesia and long observation period to affect the distribution of doxorubicin, a minimal accepted mean arterial pressure (MAP) of 65 mmHg was opted for. For animals going below this value, continuous infusion of norepinephrine (concentration: 0.1 mg/mL, start infusion rate: 0.3 mL/h) was started. The infusion rate was increased by 0.3 mL/h according to need. The animals were continuously provided with fluid to control both the level of glucose and urine production. The blood loss during the surgery was minimal.

Randomization

Before any surgical intervention, the animals were block-randomized in pairs of two to receive either bolus or continuous administration of doxorubicin. Randomization was done by drawing a note indicating Group 1 (bolus administration) or Group 2 (continuous administration) from a non-translucent envelope.

Surgical procedures

After induction of anaesthesia, surgical procedures were initiated. With the pig in a supine position, a central venous catheter was placed ultrasound-guided in a jugular vein. Via an approximately 5–6 cm midline incision starting from 2–3 cm cranial to manubrium sterni, an arterial sheath was placed in the internal carotid artery on the opposite side.

With an anteromedial incision starting approximately 2 cm below the tibial plateau and continuing to the midpoint of the anterior crest, the right tibial bone was assessed. A drill hole, 35 mm in depth and ∅2 mm, was made in the cancellous bone approximately 10 mm distal to the epiphyseal line. Overheating of the bone was prevented by frequent pausing and continuous cooling with saline. A microdialysis catheter, with a membrane length of 30 mm, was placed in the cancellous drill hole (Fig. 1).

Fig. 1

Location of microdialysis catheters; (1) cancellous bone, (2) subcutaneous tissue, (3) synovial fluid of the knee joint, (4) muscle tissue and (5) intravenously

A 30-mm catheter was placed in the synovial fluid of the right knee joint using a splittable introducer. Also, by the use of splittable introducers and ultrasound, a 30-mm catheter was placed in the muscle tissue (mean depth 104 mm, range 94–116 mm) and subcutaneous tissue on the right front leg, respectively. Finally, a 30 mm intravenous catheter (type 67) was placed on the right front leg to assess unbound plasma concentrations. All microdialysis catheters were sutured to the skin for fixation. The positions of the synovial fluid of the knee joint and cancellous bone catheters were verified intraoperatively with the use of fluoroscopic imaging. After euthanasia, the drill holes in the cancellous bone were verified by computed tomography (CT).

After the placement of all catheters, each catheter was connected to a precision pump. To fill the entire microdialysis system with perfusion fluid (saline), flushing was performed until no air bubbles seemed to be trapped within the system.

Administration of doxorubicin

To lower the risk of extravasation, 500 mL of saline was administered through the central venous catheter over 30 min before the administration of doxorubicin, as in accordance with clinical guidelines. Hereafter, Group 1 received an intravenous bolus administration of 150 mg doxorubicin over 10–15 min, while Group 2 received a continuous administration of 150 mg doxorubicin over 6 h. Both administrations were followed by an administration of a minimum of 100 mL saline.

Every animal received a dosage of 150 mg of doxorubicin, and dosage was thus not determined by body surface area due to a lack of existing formulas for the specific porcine breed [19, 20].

Sampling

The initiation of administration of doxorubicin is defined as time zero (T = 0) (Fig. 2). The overall sampling period was 24 h. In both groups, dialysates were collected every 30 min from time 0 to 120 min, every 60 min from time 120 min to 360 min and every 120 min from time 360 min to 840 min. At time 960 min, 1200 min and 1380 min, a LoBind Eppendorf tube was placed for collection of dialysates over 60 min. A total of 15 dialysates were collected from each compartment in each animal. After the collection of the last dialysate, calibration was performed with the collection of two calibration dialysates. Calibration was performed with a solution containing 10 μg/mL doxorubicin hydrochloride. Blood samples were drawn from a central venous catheter at the midpoint of each dialysate sampling interval. A total of 15 blood samples were taken. During the entire sampling period, the lights in the operation room were switched off due to the risk of photodegradation of doxorubicin.

Fig. 2Handling of samples

All dialysates were stored immediately after collection at -80 °C until analysis. Venous blood samples (EDTA 1.8 mg/mL) were stored for a maximum of 2 h before being centrifuged at 3000g, for 10 min at 5 °C. After centrifugation, plasma samples were stored at -80°°C until analysis.

Quantification of doxorubicin in microdialysates and plasma samples by ultra-high performance liquid chromatography and tandem mass spectrometry (UHPLC-MS/MS)

Doxorubicin and doxorubicinol were quantified in microdialysate and blood plasma samples by UHPLC-MS/MS using a previously described and validated method [18]. The method utilizes stable isotope-labelled doxorubicin (13CD3-doxorubicin) as internal standard and a linear calibration model based on matrix-correct calibrator samples spiked with reference standard compounds (see Supplemental Fig. 1 for selected method documentation and further details in [18]).

The lower limit of quantification for doxorubicin and doxorubicinol was estimated to be 0.002 (dialysate) and 0.003 µg/mL (plasma), respectively, and standard requirements for the method precision (CV < 15%) and trueness (bias < 15%) were met.

Pharmacokinetic analysis and statistics

All concentrations quantified in the dialysate are of unbound doxorubicin, while the plasma samples represent the total concentrations (bound + unbound). For all animals and each compartment, the following pharmacokinetic parameters were calculated by non-compartmental analysis using Stata (version 16.0, StataCorp, College Station, Texas, USA): area under the concentration–time curve (AUC0–24 h) from time zero until 24 h, peak drug concentration (Cmax), time to peak drug concentration (Tmax) and tissue penetration AUCtissue/AUCplasma. The AUC0–24 h was calculated by the use of the linear up-log-down trapezoidal method. The Cmax was calculated as the mean peak concentration of doxorubicin in each compartment. The Tmax was estimated as time until Cmax. All measured dialysate doxorubicin concentrations were attributed to the midpoint of each sampling interval. The pharmacokinetic parameters for doxorubicin between the two groups were compared using mixed models for repeated measurements taking into account multiple compartments per animal, followed by post-hoc tests for pairwise comparisons. All model assumptions were tested by visual inspection of residuals, fitted values and estimates of random effects. Due to a lack of normal distribution, log-transformed Tmax for doxorubicin was analysed, and the results were back-transformed giving medians and ratio of medians for comparisons.

The pharmacokinetic parameters for doxorubicinol between the two groups were compared by t-test.

A p-value < 0.05 was regarded as statistically significant.

Target evaluation

IC50 is the concentration capable of inhibiting 50% of a tumour cell line. Time above IC50 for two selected osteosarcoma cell lines (HOS: 0.016536531 μg/mL and NOS-1: 0.046344245 μg/mL) was calculated by linear interpolation in Microsoft Excel [21].