Effect of repeated bolus and continuous doxorubicin administration on bone and soft tissue concentrations– a randomized study evaluated in a tumour-free porcine model

Eighteen pigs were included and randomized into two groups of nine, receiving either two bolus or continuous administrations on day 1 and 22. On day 22, twelve hours of microdialysis sampling of doxorubicin bone and soft tissue concentrations was performed.

The study was conducted at the Department of Clinical Medicine, Aarhus University, Aarhus, Denmark. Approval was obtained from the Danish Animal Experiments Inspectorate (license No. 2021-15-0201-01079) and carried out in accordance with existing laws. All chemical analyses of doxorubicin were performed at the Department of Forensic Medicine, Aarhus University Hospital, Aarhus N, Denmark. Kidney and liver values were analysed by the Department of Clinical Biochemistry, Aarhus N, Denmark. The study was carried out in accordance with the ARRIVE-guidelines [30].

Microdialysis is a catheter-based sampling tool that allows for a dynamic and continuous sampling of molecules below the cut-off value of the semi-permeable membrane located at the tip of the catheter. The semi-permeable membrane allows concentration-dependent diffusion. As the catheter is connected to a perfusion pump continuously perfusing the catheter at a set flow rate, an equilibrium between the membrane and the surrounding extracellular space will not occur. This means that the concentration of the molecule of interest, quantified in the dialysate (collected in the collecting tube), only constitutes a fraction of the absolute concentration in the surrounding tissue. This fraction is referred to as the relative recovery, which must be calculated and corrected to estimate absolute tissue concentrations. In the present study, relative recovery was estimated using retrodialysis by drug with doxorubicin performed at the end of the study. Formulas for calculation can be found elsewhere [29, 31].

All microdialysis equipment was purchased from M Dialysis AB (Stockholm, Sweden). Based on results from a previous study, the equipment used for the collection of doxorubicin were the type 70 catheter with a 30 mm membrane and the type 67 intravenous catheter also with a membrane length of 30 mm (cut-off of 20 kDa) [32]. For all catheters, the flow rate was set at 1 µl/min, and the perfusion fluid was saline 0.9%. Sample collection was done in 1.5 ml LoBind Eppendorf tubes (Eppendorf, Hamburg, Germany) [32].

Eighteen female pigs (Danish Landrace) were included in the study and randomized into two groups of nine. Randomization was done before the first intervention as block randomization of two (as one for the first two animals), by drawing a note indicating Group 1 (bolus infusion) or Group 2 (continuous infusion) from a non-translucent envelope.

Sample size was, due to a lack of relevant clinical assumptions based on results, from a study measuring doxorubicinol concentrations (metabolite of doxorubicin) in plasma as well as the hepatic system. Using a two-independent means power calculation with a mean doxorubicinol AUC (min×mM) of 3.8 (SD 1.4) in plasma (n = 4), a mean doxorubicinol AUC of 2.0 (SD 0.8) in the hepatic system (n = 4), a power of 90% and a significance level of 5%, a group size of n = 8 was found [4].

The animals arrived at the research facility a minimum of 14 days before the first intervention, providing sufficient time for acclimatization and human contact training. They were kept singularly in pens with a light cycle of 12 h. Straw was used as bedding, and the animals had access to water ad libitum. Feeding was restricted (farm pig ration) to control weight gain.

Doxorubicin dosage is traditionally administered based on body surface area. However, due to a lack of a suitable formula for the pig breed used in the present study, a dosage based on weight (2 mg/kg) was opted for. This approach has previously provided clinically relevant plasma concentrations [32]. Animals were fed to reach a mean weight of 66 kg on day 1 and 75 kg on day 22 resulting in dosages of 132 mg and 150 mg of doxorubicin, respectively. Pigs gain weight much faster than humans, wherefore, the dosage on day 22 had to be increased compared to the dosage on day 1. The actual attained mean weight was 65 kg on day 1 and 73 kg on day 22.

The start of doxorubicin administration indicated time zero (Fig. 3). The overall sampling time was 12 h. The room light was off for the entire study period due to the risk of photodegradation of doxorubicin. Sampling was identical for animals receiving bolus and continuous infusion, respectively.

Dialysates were collected every 30 min from time 0 to 120 min, every 60 min from time 120 to 360 min, and every 120 min from time 360 to 720 min. A total of 11 dialysates were collected from each compartment plus an additional three retrodialysis calibration samples. Venous blood samples (total concentration) were taken at the midpoint of each of the above-mentioned sampling intervals and additionally after 60 and 360 min. Calibration was performed with an 18.74 µg/mL doxorubicin saline solution.

All venous blood samples taken for the quantification of the total concentration of doxorubicin was collected in EDTA 1.8 mg/mL 4 mL tubes and stored at 4–5 °C for a maximum of two hours. Samples were then centrifuged at 3,000 g for 10 min at 5 °C. Venous samples taken for the evaluation of kidney and liver status were collected in lithium heparin 5 mL tubes and centrifuged at 2,000 g for 10 min at 20 °C. All plasma samples and dialysates were immediately stored at -80 °C until analysis.

To ensure that variations in kidney and liver function between the two groups did not influence the results, a total of ten blood samples were taken from each animal. Samples were taken on day 1 before as well as 1 h and 6 h after the start of doxorubicin administration (time 0), respectively. One sample was taken on days 3, 7 and 10. On day 22, a sample was taken before as well as 1 h, 6 h and 12 h after the start of doxorubicin administration, respectively. Except for hemoglobin, all values were attained after analysis on an Atellica CH (Siemens Atellica Solution, Siemens Healthineers, Erlangen, Germany). Hemoglobin was analyzed on an ABL90 Flex Plus (Radiometer Medical Aps, Brønshøj, Denmark).

The concentration of doxorubicin measured in dialysates represents the unbound concentration, while the concentration measured in plasma from venous samples is the total concentration. Doxorubicin concentrations in microdialysates/plasma were measured by ultra-high performance liquid chromatography and tandem mass spectrometry (UHPLC-MS/MS) as previously described in detail [32]. Briefly, microdialysate samples were prepared for analysis by mixing 10 µL dialysate with 190 µL internal standard solution (13CD3-doxorubicin stable isotope (Clearsynth, Mumbai, India) at 0.025 µg/mL in water: methanol (85:15) with 0.1% formic acid) in a 96-well plate (1mL Eppendorf LoBind). Blood plasma samples were prepared in 96-well plates (1 mL Eppendorf Lo-bind) by mixing 50 µL plasma with 50 µL saline (water with 0.9% sodium chloride) and 300 µL internal standard/protein precipitation solution (13CD3-doxorubicin at 0.05 µg/mL in 100% acetonitrile), followed by vortex-mixing and centrifugation (5000 ×g, 5 min). A 100 µL aliquot from the resulting supernatant was diluted with 200 µL water supplemented with 0.1% formic acid to yield the final sample. Separate samples for calibration were prepared in matched matrices at concentrations of 0, 0.001, 0.004, 0.012, 0.037, 0.111, 0.333 and 1 µg/mL using reference compound doxorubicin hydrochloride (European Pharmacopoeia Reference Standard CRS batch 7 supplied from Sigma-Aldrich, Germany). Samples were analyzed with an UHPLC-MS/MS system (Acquity UPLC coupled to a TQS mass spectrometer from Waters, Milford, Massachusetts, USA) with a C18 column (Waters UPLC HSS-C18, 1.8 μm, 100 × 2.1 mm) and using multiple reaction monitoring mode as described [32]. Calibration curves were constructed by linear regression of the peak area ratio (analyte/internal standard) versus the nominal analyte concentrations of the calibrant samples using 13CD3-doxorubicin as internal standard. The lower limit of quantification for doxorubicin was estimated to 0.002 (dialysate) and 0.003 µg/mL (plasma) and the standard requirements for precision (CV < 15%) and trueness (bias < 15%) were met.

For all animals and each compartment, the following pharmacokinetic parameters were determined for doxorubicin by non-compartmental analysis using Stata (version 16.0, StataCorp, College Station, Texas, USA). The area under the concentration-time curve (AUC_{0 − 12 h}) from time zero until the end of the sampling period of 12 h was calculated using the linear up-log-down trapezoidal method. Peak drug concentration (C_max) was calculated as the mean peak concentration of doxorubicin in each compartment. Penetration ratio was calculated by AUC_tissue/AUC_iv. For dialysates, all concentrations were assigned to the midpoint of each sampling interval.

The pharmacokinetic parameters for the animals receiving bolus infusion were compared to the animals receiving continuous infusion. The data was modelled separately for each compartment data to compare the groups (regression analysis with fixed effect) and analysed separately within groups to compare the compartments within groups (regression analysis with random effect). Mean plasma concentrations after 1 and 6 h on days 1 and 22 were compared in Excel (Microsoft version 16.78.3) using a t-test. A p-value < 0.05 was regarded as statistically significant. The p-values were not adjusted for multiple comparisons.

Except for two animals receiving continuous infusion, all animals survived the entire study period. One animal was euthanized on day 10 due to continuous refusal to eat. The second animal died suddenly during the induction of anaesthesia on day 22. For both animals, all attained samples were included.

The ranges of mean relative recovery (SD) were 34% (8)– 80% (11) for the bolus group and 25% (9)– 79% (23) for the continuous group.

Figure 4 contains mean doxorubicin plasma (total) concentration-time profiles for both groups from day 1 to 22. On day 1, no clear peak was seen for the bolus group, probably because the first sample was taken after 1 h. For the continuous group, there are no plasma concentrations for ‘Day 1 1 h’ because the samples were taken through the same venous catheter as the infusion was administered– thus, reliable samples were not possible during doxorubicin infusion. Comparing the concentrations measured after 1 and 6 h after bolus administration on days 1 and 22, no intra-group differences were shown. For the continuous group after 6 h concentrations were higher after the first administration (p = 0.037).

Figure 4 also contains mean doxorubicin concentration-time profiles for each of the five microdialysis compartments. Individual concentration-time profiles for each animal on day 22 can be seen in supplementary Figures S1 and S2. A tendency of a prompt peak or a gradual increase was seen for all compartments after both bolus and continuous infusion. The exception is the intravenous (unbound concentration) compartment after bolus infusion, which seems heterogeneous, looking at the peak drug concentrations with a range of 0.02–0.41 µg/mL.

Pharmacokinetic parameters for day 22 are presented in Table 1. Except for AUC_{0 − 12 h} for plasma (total concentration) in the continuous group and intravenous C_max in the bolus, which were significantly higher, no difference in AUC_{0 − 12 h} and C_max on day 22 was observed for the investigated compartments between the two groups. No difference was found between the two groups for AUC_tissue/AUC_iv. Mean C_max were generally very low, with values below 0.10 µg/mL. The only exceptions were plasma (total concentration) with mean C_max values of 0.82 and 0.65 µg/mL after bolus and continuous infusion, respectively. The intravenous compartment in the bolus group was the only compartment with a mean unbound C_max value of doxorubicin above 0.10 µg/mL (0.11 µg/mL).

AUC_{0 − 12 h} and C_max for plasma (total concentration) were higher than the other compartments after both continuous and bolus infusion (Table 2). Intravenous (unbound concentrations) were generally higher than the other compartments except the synovial fluid of the knee joint for both groups and the subcutaneous tissue after continuous infusion. Overall, the distribution to the different compartments was heterogeneous, with the lowest values of AUC_{0 − 12 h} and C_max found in the cancellous compartment for both groups.

Supplementary Fig. 3 shows the time-concentration curve for nine different markers of liver and kidney status, which appears comparable between the two groups. Potassium, calcium and creatinine seem to be highly affected. However, the similar profiles indicate that both groups were affected equally by the doxorubicin administration. Comparing the results for the time of the two administrations, they were similar except for creatinine and sodium, which seemed much higher and lower at the end of day 22.

The effect of doxorubicin treatment is based on clinical and radiological evaluations, and no correlated pharmacokinetic targets exist. The closest to an effect parameter is the tumour cell line half-maximal inhibitory concentration (IC50), which is defined as the concentration capable of inhibiting 50% of a tumour cell line. However, IC50 is defined in vitro and can, therefore, not uncritically be extrapolated to a clinical setting. The lowest IC50 value for an osteosarcoma cell line is 0.017 µg/mL, which is higher than the C_max values found in cancellous bone after both administration forms on day 22 [33]. Additionally, the time needed above or equal to IC50 is also unknown.

Measuring the target intra-tumoral concentration therefore seems to be of utmost interest. Microdialysis presents itself as a usable tool for intra-tumoral measurements, and the risk of tumour cell seeding after implantation of the microdialysis catheter has been estimated to be similar to that of fine needle biopsies of < 0.005% [34, 35]. With the use of microdialysis, Müller et al. evaluated the concentrations of the chemotherapeutic drug 5-Fluororacil intratumorally and in nearby healthy subcutaneous tissue in 10 patients with breast cancer during their first cycle of neoadjuvant chemotherapy [6]. They found a correlation between intra-tumoral AUC and clinical tumour response. In comparison, a correlation could not be found when looking at plasma or subcutaneous AUC. Interestingly, the plasma AUCs for the non-responders in the study by Müller et al. were some of the highest values found [6]. This may be caused by differing tumour vascularisation and microenvironment.

Doxorubicin is known to cause systemic inflammation affecting all cells of the body, which potentially could also affect the penetration into the tumoural tissue. The level of intratumoural inflammation could be evaluated with the use of microdialysis, and a correlation to clinical effect could be investigated.

Doxorubicin is administered several times during treatment, but clinical studies have reported that repeated administrations impact the plasma concentrations [24‐26, 28]. Two clinical studies (n = 45, n = 12) reported an increase in plasma concentration, whereof one of them attributed it to a decrease in the central volume of distribution [25, 26]. This is contradicted by two other clinical studies (n = 19, n = 15) finding decreased plasma concentrations [24, 28]. None of the studies had investigated effects in tissue.

Comparing mean plasma concentrations measured after 1 and 6 h, on day 1 and day 22, in the present study, the only significant difference was found in the continuous group after 6 h. However, this may very likely be due to a small sample size for these measurements (n = 6). When comparing the tissue data found on day 22 in the present study with a previous study performed in our group measuring the tissue concentrations on day 1 in a comparable animal group of 16 pigs (8 receiving continuous infusion and 8 receiving bolus infusion with the same dosage regime of 150 mg doxorubicin) only minor differences are seen [31].

Many clinical studies have found a high level of inter- and intraindividual differences in doxorubicin plasma (total concentrations) pharmacokinetics [11, 27]. This could be caused by heterogeneous study populations due to varying cancers, liver- and kidney function, age, etc. In the study population evaluated in the present porcine study, the liver- and kidney function was equal, and normal, between the two groups and between days 1 and 22 [36]. This observation provides reassurance that variations in results between the two groups cannot be attributed to differences in these organ functions. The only exceptions were creatinine and sodium, which increased and decreased (within the normal range), respectively, at the end of day 22 compared to day 1. However, with the low fraction excreted through the kidneys (5–15%) as well as the overall short amount of time for which this occurred, the clinical relevance seems small [5, 12, 37, 38].

Looking at the blood compartments (plasma; total concentration and intravenous; unbound concentration), a tendency of higher variance in the time-concentration profiles was seen for intravenous in the bolus group, which consonants with the findings of high intra- and interindividual concentrations in previous studies of only the plasma compartment [11, 24‐28].

None of the animals in the present study had a tumour, wherefore any potential effects due to tumour tissue on the doxorubicin concentration after repeated administration could not be evaluated. Pigs are generally considered a good experimental model due to their anatomical and physiological resemblance to humans [39]. Also, the animals in the present study were 5 months of age with an active growth plate, which potentially could affect the translational potential of the concentrations to adult bone. During the study, the animals also gained a lot of weight, which follows the animals’ natural growth, but could potentially affect the distribution. However, for other drugs (antibiotics), a similar PK profile in bones between humans and young pigs has been found following continuous infusion [40, 41].