On 2017 Oct 22, George Kunos commented:

We read with interest the above article (Varga B, 2017), comparing the pharmacological properties of a series of cannabinoid receptor 1 (CB1R) blockers, including first generation brain-penetrant compounds and more recently introduced peripherally restricted compounds in various tests, including their anti-obesity effects in a mouse diet-induced obesity (DIO) model. The complete lack of effect of the peripheral CB1R inverse agonist JD-5037 in the DIO mouse model was of particular interest to us, as we have earlier documented its high anti-obesity efficacy in DIO mice (Tam J, 2012, Cinar R, 2014) and in a mouse model of Prader-Willi syndrome (Knani I, 2016), as well as its anti-diabetic efficacy in a rat model of type 2 diabetes (Jourdan T, 2013, Jourdan T, 2014). In these as well as in more recent studies (Tam J, 2017, Hinden L, 2018, Udi S, 2017), JD-5037 was dissolved in 4% DMSO/1% Tween-80 in PBS (vehicle #1) for administration by oral gavage (see Jourdan T, 2013), whereas Varga et al. used a 5% Tween-80 solution (vehicle #2). Because in our hands the inclusion of DMSO was critical for keeping this highly lipophilic compound in solution, we compared the oral bioavailability and peripheral target engagement of JD-5037 dissolved in these 2 vehicles. Using vehicle #1, the peak plasma concentration of JD-5037 measured by LC-MS/MS 1 h after oral administration to lean, male C57Bl6/J mice was 1076 ± 208 ng/mL (1840 nM), similar to values we published earlier (Tam J, 2012). In contrast, using vehicle #2, the plasma level of JD-5037 was 0.38 ± 0.02 ng/mL (0.67 nM), more than 1,000-fold lower and barely detectable. JD-5037 is 99.6% protein-bound in plasma (Tam J, 2012), so its calculated free concentration was 7.4 nM using vehicle #1 versus 2.7 pM using vehicle #2, the latter value being 2 orders of magnitude below the binding Kd of JD-5037 for CB1R (0.4 nM), which predicts no significant CB1R occupancy. We further verified this by using the upper GI motility test as a measure of peripheral CB1R occupancy (Tam J, 2012). In vehicle-treated mice, an oral charcoal bolus traveled 54 ± 3% of the length of the small intestine in 30 minutes, whereas In mice treated with a maximally effective dose of the CB1R agonist ACEA, the charcoal bolus traveled only 25 ± 3%, indicating a 54% inhibition. The inhibitory effect of ACEA was completely blocked by pretreatment with a single dose of 3 mg/kg JD-5037 in vehicle #1 (53 ± 3%), in agreement with our published data (Tam J, 2012). In contrast, the same dose of JD-5037 administered in vehicle #2 was completely without effect, the distance traveled by the charcoal bolus (29 ± 2%) being the same as with ACEA alone. Thus, the negative findings of Varga et al. can be attributed to lack of absorption and a consequent lack of peripheral CB1R occupancy by JD-5037, due to the use of an inappropriate vehicle. Such pitfalls are avoidable by verifying bioavailability and target engagement, which is a basic requirement when testing the in vivo efficacy of novel compounds.

George Kunos, Joseph Tam^1, Resat Cinar, Tony Jourdan; National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD, USA

¹ Current address: School of Pharmacy, The Hebrew University, Jerusalem, Israel

On 2017 Oct 22, George Kunos commented:

We read with interest the above article (Varga B, 2017), comparing the pharmacological properties of a series of cannabinoid receptor 1 (CB1R) blockers, including first generation brain-penetrant compounds and more recently introduced peripherally restricted compounds in various tests, including their anti-obesity effects in a mouse diet-induced obesity (DIO) model. The complete lack of effect of the peripheral CB1R inverse agonist JD-5037 in the DIO mouse model was of particular interest to us, as we have earlier documented its high anti-obesity efficacy in DIO mice (Tam J, 2012, Cinar R, 2014) and in a mouse model of Prader-Willi syndrome (Knani I, 2016), as well as its anti-diabetic efficacy in a rat model of type 2 diabetes (Jourdan T, 2013, Jourdan T, 2014). In these as well as in more recent studies (Tam J, 2017, Hinden L, 2018, Udi S, 2017), JD-5037 was dissolved in 4% DMSO/1% Tween-80 in PBS (vehicle #1) for administration by oral gavage (see Jourdan T, 2013), whereas Varga et al. used a 5% Tween-80 solution (vehicle #2). Because in our hands the inclusion of DMSO was critical for keeping this highly lipophilic compound in solution, we compared the oral bioavailability and peripheral target engagement of JD-5037 dissolved in these 2 vehicles. Using vehicle #1, the peak plasma concentration of JD-5037 measured by LC-MS/MS 1 h after oral administration to lean, male C57Bl6/J mice was 1076 ± 208 ng/mL (1840 nM), similar to values we published earlier (Tam J, 2012). In contrast, using vehicle #2, the plasma level of JD-5037 was 0.38 ± 0.02 ng/mL (0.67 nM), more than 1,000-fold lower and barely detectable. JD-5037 is 99.6% protein-bound in plasma (Tam J, 2012), so its calculated free concentration was 7.4 nM using vehicle #1 versus 2.7 pM using vehicle #2, the latter value being 2 orders of magnitude below the binding Kd of JD-5037 for CB1R (0.4 nM), which predicts no significant CB1R occupancy. We further verified this by using the upper GI motility test as a measure of peripheral CB1R occupancy (Tam J, 2012). In vehicle-treated mice, an oral charcoal bolus traveled 54 ± 3% of the length of the small intestine in 30 minutes, whereas In mice treated with a maximally effective dose of the CB1R agonist ACEA, the charcoal bolus traveled only 25 ± 3%, indicating a 54% inhibition. The inhibitory effect of ACEA was completely blocked by pretreatment with a single dose of 3 mg/kg JD-5037 in vehicle #1 (53 ± 3%), in agreement with our published data (Tam J, 2012). In contrast, the same dose of JD-5037 administered in vehicle #2 was completely without effect, the distance traveled by the charcoal bolus (29 ± 2%) being the same as with ACEA alone. Thus, the negative findings of Varga et al. can be attributed to lack of absorption and a consequent lack of peripheral CB1R occupancy by JD-5037, due to the use of an inappropriate vehicle. Such pitfalls are avoidable by verifying bioavailability and target engagement, which is a basic requirement when testing the in vivo efficacy of novel compounds.

George Kunos, Joseph Tam^1, Resat Cinar, Tony Jourdan; National Institute on Alcohol Abuse and Alcoholism, National Institutes of Health, Bethesda, MD, USA

¹ Current address: School of Pharmacy, The Hebrew University, Jerusalem, Israel