Refined tamoxifen administration in mice by encouraging voluntary consumption of palatable formulations

Abstract

Drug administration in preclinical rodent models is essential for research and the development of novel therapies. Compassionate administration methods have been developed, but these are mostly incompatible with water-insoluble drugs such as tamoxifen or do not allow for precise timing or dosing of the drugs. For more than two decades, tamoxifen has been administered by oral gavage or injection to CreER^T2-loxP gene-modified mouse models to spatiotemporally control gene expression, with the numbers of such inducible models steadily increasing in recent years. Animal-friendly procedures for accurately administering tamoxifen or other water-insoluble drugs would, therefore, have an important impact on animal welfare. On the basis of a previously published micropipette feeding protocol, we developed palatable formulations to encourage voluntary consumption of tamoxifen. We evaluated the acceptance of the new formulations by mice during training and treatment and assessed the efficacy of tamoxifen-mediated induction of CreER^T2-loxP-dependent reporter genes. Both sweetened milk and syrup-based formulations encouraged mice to consume tamoxifen voluntarily, but only sweetened milk formulations were statistically noninferior to oral gavage or intraperitoneal injections in inducing CreER^T2-mediated gene expression. Serum concentrations of tamoxifen metabolites, quantified using an in-house-developed cell assay, confirmed the lower efficacy of syrup- as compared to sweetened milk-based formulations. We found dosing with a micropipette to be more accurate than oral gavage or injection, with the added advantage that the method requires little training for the experimenter. The new palatable solutions encourage voluntary consumption of tamoxifen without loss of efficacy compared to oral gavage or injections and thus represent a refined administration method.

Fig. 1. Sweetened oil emulsions and training improve voluntary drug consumption.

a, Adult male and female mice were trained for 3 days to consume 80 μl of formulations from a disposable plastic micropipette tip and on the fourth day were offered the same formulation containing TAM. b, Feeding time was recorded and considered voluntary if the mouse, held only slightly by the proximal tail section, drank the offered volume in less than 60 s. c, Any formulation not consumed after 60 s was offered again after gently restraining the mice by the scruff, and the extra time required to drink the remaining solution was recorded. d, The fraction of mice that voluntarily consumed (<60 s) oil (n = 21), sweetened MOE (n = 52) or SOE (n = 33) during 3 days of training (green bars) or TAM-containing formulations on the fourth day. Yellow bar, OIL-TAM (n = 12); blue bar, MOE-TAM (n = 17); orange bar, SOE-TAM (n = 17). The oil used for the formulations in d was sterilized by heat treatment. e, Total consumption time for three training days, recorded from the moment the MOE or SOE formulation was offered (n = 13 per group). Indicated is the time before restraining (white zone) as described in b and after restraining (red zone) as described in c. The oil used for the emulsions in e was sterilized by filtration. Minimum and maximum values, interquartile range and median are depicted as Tukey box plots with individual data points shown as gray circles. Source data

Fig. 2. Efficacy of TAM-induced gene expression.

Comparison of CreER^T2-dependent HY-TCR reporter expression in thymocytes of adult female mice, 40 h after treatment with a single dose of 80 mg/kg TAM. Mice were either treated with TAM dissolved in oil via OG and IP injections or pipette-fed with oil-in-water emulsions made with sweetened milk (MOE) or syrup (SOE). a, The percentage of HY-TCR-expressing thymocytes was determined via flow cytometry as described in Methods. Shown are the means ± s.d. per treatment group (n = 13 per group). b, Noninferiority graph, depicting the 95% confidence interval (CI) of the difference between the mean of the OG-oil group and SOE group or MOE group of the results shown in a. The 95% CIs were computed using Dunnett’s confidence interval formula following the ANOVA. The noninferiority margin ‘d’ was set at 0.18 (1.5 times s.d. of the control OG-oil group). c, TAM metabolite concentrations in the serum of mice described in a were assessed in vitro using MEFs from R26-CreER^T2-Ai14 mice. d, The fold change in body weights relative to the oil group following daily administration of the indicated formulations for 4 or 6 days. Group sizes: OG-oil, n = 9; SOE, n = 14; MOE, n = 21. Depicted are means ± s.d. ***P = 0.0006 according to one-way ANOVA followed by post-hoc Dunnett’s multiple comparison test. e, The percentage of HY-TCR-expressing thymocytes induced after IP injection of OIL-TAM or pipette feeding with MOE-TAM (n = 7). MDA, micropipette-guided drug administration. f, Noninferiority graph, depicting the 95% CI of the difference between the means of the IP oil group and MOE group of the results shown in e. The controls shown in a and c were HY-TCR knock-in mice trained (3 days) with MOE but not treated with TAM (n = 6). Source data

Fig. 3. Repeated administration of TAM using MOE.

a–d, Adult male and female HY-switch mice were trained for 3 days before MOE-TAM treatment. Consumption time for the last training, first (#1) and second (#2) TAM treatment was recorded and displayed as described in Fig. 1e. For subsequent MOE-TAM treatments, mice were immediately restrained (red dot plots). Treatment with 80 mg/kg (a and c) and 40 mg/kg (b and d) TAM in MOE. c and d show the consumption time for individual mice (M1 to M6) for data shown in a and b, respectively. Mice were either offered the formulations without being restrained for the first 60 s (gray or blue dot plots) or immediately restrained (red dot plots). For a and b minimum and maximum values, interquartile range and median are depicted as Tukey box plots with individual data points shown as scatter dot plots. Source data

Fig. 4. Multiple administration of TAM in MOE.

a, Induction of tdTomato expression was assessed in adult R26-CreER^T2-Ai14-tdTomato mice after training (green arrows) and TAM treatments as indicated. Mice were either offered MOE-TAM (20 mg/kg) once a day for 5 days (blue) or only once on day −5 (red) or day −1 (yellow) (n = 6 per group). Mice that did not receive TAM on the indicated days were offered MOE vehicle only (short gray arrows). On day 0, the mice were euthanized and analyzed using flow cytometry (fluorescence-activated cell sorting, FACS) and immunohistochemistry (IHC). b, The percentages of thymocytes and splenocytes expressing tdTomato were determined by flow cytometric analysis. Tukey box plots depict fold change compared to the fraction of positive cells detected in mice treated only once on day −1 and include minimum and maximum values, interquartile range and median with individual data points overlayed as scatter dot plots. Source data