Application of bootstrap f2 to dissolution data from biorelevant media and evidence of the conservative nature of bootstrap f2.

f2 with or without bootstrapping is the most common method to compare in vitro dissolution profiles, but methods to compare dissolution profiles have become less harmonized. The objective was to compare outcomes from bootstrap f2 and f2 (i.e. not-bootstrapped f2) using a large set of in vitro dissolution data. Non-parametric bootstrapping was performed on the 104 profile comparisons that did not meet the percent coefficient of variation (CV%) criteria to use average dissolution data. Bootstrap f2 was taken as the lower 90 % confidence interval of bootstrapped samples. There was concordance between bootstrap f2 and f2 in 92 of the 104 comparisons (88 %). There were no false positives. However, 12 % were false negative. Inspection of these discordance pairs suggests that bootstrap f2 serves as a conservative approach to assess profile similarity, particularly when a value of 50 is being used as decision criteria.

Prediction of food effect on in vitro drug dissolution into biorelevant media: Contributions of solubility enhancement and relatively low colloid diffusivity.

A prior model that showed good predictability for fed-state and fasted-state biorelevant media was extended to predict the degree to which fed-state biorelevant media (i.e. FeSSGF and FeSSIF-V2) enhanced drug dissolution over fasted-state biorelevant media (i.e. FaSSGF and FaSSIF-V2): ϕfood=vfast1/6vfed1/6·(ff,fedDD2/3+fm,fedDD-M,fed2/3)[Dt,fed](ff,fastDD2/3+fm,fastDD-M,fast2/3)[Dt,fast] where ϕfood is the degree by which fed-state biorelevant media enhanced in vitro drug dissolution over fasted state biorelevant media, ff,fast and ff,fed are the fraction of free drug in fasted and fed biorelevant media, fm,fast and fm,fed are the fraction of drug in fasted and fed mixed micelles (or fat globules in FeSSGF), DD is the free drug diffusivity, DD-M, fast and DD-M, fed are the fasted and fed mixed micelle (or fat globule) diffusivity, and [Dt,fast] and [Dt,fed] are the total drug solubilities in fasted and fed-state biorelevant media, respectively. Solubility, particle size measurement, and intrinsic dissolution studies were performed for model BCS Class II drugs griseofulvin, ketoconazole, and ibuprofen each in FaSSGF, FeSSGF, FaSSIF-V2, and FeSSIF-V2 to compare observed versus predicted dissolution enhancement in fed-state over fasted-state biorelevant media. Relative to solubilization, in vitro dissolution was many fold lower in fed media over fasted media, indicating the compromising role of micellar and fat-globule diffusivity in attenuating dissolution rate based on solubility enhancement alone. Results of ϕfood agreed with what was observed and were also corroborated by reported food effects in vivo for griseofulvin, ketoconazole, and ibuprofen. The understanding of attenuation of in vitro dissolution in fed versus fasted biorelevant media may contribute towards predicting in vivo food effects, including lack of in vivo food effect for some poorly water soluble drugs.

Prediction of in vitro drug dissolution into fasted-state biorelevant media: Contributions of solubility enhancement and relatively low colloid diffusivity.

A previously developed model showed good predictability for in vitro drug dissolution into FeSSGF and FeSSIF-V2, where biorelevant media-mediated enhanced drug dissolution by only about one-tenth as much as it enhanced solubility, due to drug-loaded fat globules and mixed micelles exhibiting far slower diffusivity than free drug. The present objective was to quantitatively assess the contributions of fasted biorelevant media-mediated solubility and diffusivity on enhanced drug dissolution in FaSSGF and FaSSIF-V2. Griseofulvin, ketoconazole, and ibuprofen were subjected to dissolution into FaSSGF and FaSSIF-V2, as well as their corresponding "surfactant-free" media. Solubility and dynamic light-scattering studies (DLS) of drug in FaSSGF and FaSSIF-V2 were conducted. Results showed that FaSSGF and FaSSIF-V2 mixed micelles were large and slowly diffusing (diffusivities about 2 × 10-8 cm2/s and 12×10-8 cm2/s, respectively), compared to free drug (about 7 × 10-6 cm2/s), although were smaller and more rapidly diffusing compared to drug-bound fat globules and mixed micelles from FeSSGF and FeSSIF-V2 (about 1 × 10-9 cm2/s and 7 × 10-8 cm2/s, respectively). Dissolution enhancement in FaSSGF and FaSSIF-V2 was the same as solubility enhancement, which was minimal. FaSSGF and FaSSIF-V2 practically did not enhance drug dissolution, since solubility was at best minimally increased, but also since micelles were relatively slowly diffusing relative to free drug diffusivity.

Sources of dissolution variability into biorelevant media.

Identifying sources of dissolution variation is often challenging. Biorelevant dissolution media are compositionally complex, can require multiple steps to fabricate, and are utilized across differing laboratories. The objective was to determine the contributions of location, operator, media fabrication method, day, and tablet to total dissolution variability into biorelevant media. Thirty min dissolution data from a prior report were subjected to analysis-of-variable (ANOVA). Results showed that total dissolution variance was highest for medium dissolution (versus high or low dissolution). Overall, across both ibuprofen and ketoconazole, the rank-order importance of sources of variation were location > operator (nested in sequence) > day > fabrication method > tablet. Location and day were the factors that most frequently were significant sources of variability. However, location was generally much more impactful than day, and location frequently contributed to the majority of total variance. Day never contributed as the majority factor and was generally minor. Fabrication method was occasionally significant, although never a majority contributor. Overall, of all factors, location contributed most to dissolution variance here, using biorelevant media, as observed in other studies not using biorelevant media. This impact of location with and without biorelevant media supports our prior findings that biorelevant media generally showed favorable repeatability and reproducibility.

Prediction of In Vitro Drug Dissolution into Fed-state Biorelevant Media: Contributions of Solubility Enhancement and Relatively Low Colloid Diffusivity.

A model was previously derived to predict in vitro dissolution of drug into surfactant solution and showed good predictability for pharmaceutical surfactants, where surfactant-mediated enhanced drug dissolution was several fold less than enhanced solubility (about 3-fold or less) due to drug-loaded micelles exhibiting slower diffusivity than free drug. The present objective was to quantitatively assess the contributions of biorelevant media-mediated solubility and diffusivity on enhanced drug dissolution in FeSSGF and FeSSIF-V2. Three poorly water soluble drugs were subjected to dissolution into FeSSGF and FeSSIF-V2, as well as their corresponding "surfactant-free" media. Solubility and laser diffraction analysis of drug in FeSSGF and dynamic light-scattering studies (DLS) of drug in FeSSIF-V2 were conducted. Results showed drug-saturated FeSSGF globules and FeSSIF-V2 mixed micelles were large and slow diffusing (diffusivities of about 1×10-9 and 7×10-8 cm2/s, respectively), compared to free drug (about 7×10-6 cm2/s) and drug-bound micelles from pharmaceutical surfactants (about 0.5-1×10-6 cm2/s). Of the three drugs, griseofulvin exhibited the greatest biorelevant media-enhanced solubility and dissolution (652-fold and 6.23-fold respectively in FeSSGF, and 190-fold and 12.7-fold respectively in FeSSIF-V2), but slow colloid diffusivity markedly attenuated large solubility benefits, particularly in FeSSGF.

Similarity of dissolution profiles from biorelevant media: Assessment of interday repeatability, interanalyst repeatability, and interlaboratory reproducibility using ibuprofen and ketoconazole tablets.

Biorelevant media are increasingly being employed as dissolution media in drug development, including in smaller volumes than 900ml and under non-sink conditions. The objectives were to assess interday repeatability, interanalyst repeatability, and interlaboratory reproducibility of dissolution profiles from biorelevant media, as well as to assess the impacts of biorelevant media production method and biorelevant medium volume on dissolution profiles. Ibuprofen and ketoconazole tablets were subjected to dissolution testing in 500ml, 300ml, and 40ml of fasted state simulated gastric fluid (FaSSGF), fed state simulated gastric fluid (FeSSGF), fasted state simulated intestinal fluid version 2 (FaSSIF-V2), and fed state simulated intestinal fluid version 2 (FeSSIF-V2). f2 was used to assess repeatability and reproducibility of dissolution profiles. Results indicate favorable interday repeatability (83 of 88 comparisons were similar), favorable interanalyst repeatability (19 of 21 comparisons were similar), and favorable interlaboratory reproducibility (10 of 14 comparisons were similar) of dissolution profiles from biorelevant media, with commercial media showing greater interlaboratory reproducibility than 'from scratch' media. However, biorelevant medium production had low impact on profiles when one analyst conducted all medium preparations and study procedures at one location. Additionally, biorelevant media detected differences when products were not similar. Overall, biorelevant media showed favorable repeatability and reproducibility performance.

Induction of carbonyl reductase 1 (CBR1) expression in human lung tissues and lung cancer cells by the cigarette smoke constituent benzo[a]pyrene.

Carbonyl reductase 1 (CBR1) reduces various xenobiotic carbonyl substrates to corresponding alcohol metabolites. Here we demonstrated that benzo[a]pyrene (B[a]P), a potent pro-carcinogen and predominant polycyclic aromatic hydrocarbon (PAH) compound in cigarette smoke and air pollutants, upregulates CBR1 gene expression in vitro and in vivo, and that a proximal xenobiotic response element (XRE) motif (₋122XRE) mediates the induction effect of B[a]P. First, we observed 46% and 50% increases in CBR1 mRNA and CBR1 protein levels, respectively, in human lung tissue samples from smokers compared to never-smokers. Second, we detected 3.0-fold (p<0.0001) induction of CBR1 mRNA and 1.5-fold (p<0.01) induction of CBR1 protein levels in cells of the human lung cancer cell line A549 incubated with 2.5 µM B[a]P for 24h. Third, results from experiments with CBR1 promoter constructs indicated that a proximal XRE motif ₋122XRE) mediates induction of reporter activity in response to B[a]P. Furthermore, we detected enhanced nuclear translocation of aryl hydrocarbon receptor (AhR) following B[a]P exposure in A549 cells. Finally, we demonstrated increased binding of specific protein complexes to ₋122XRE in nuclear extracts from B[a]P-treated cells and the presence of the AhR/Arnt complex in the specific nuclear protein ₋122XRE complexes.

Expression of the anthracycline-metabolizing enzyme carbonyl reductase 1 in hearts from donors with Down syndrome.

Cancer patients with Down syndrome (DS) are susceptible to developing anthracycline-related cardiotoxicity. The pathogenesis of anthracycline-related cardiotoxicity has been linked to the intracardiac synthesis of alcohol metabolites by carbonyl reductase 1 (CBR1). CBR1 is located in the DS critical region (21q22.12). The expression of CBR1 in hearts from individuals with DS has not been characterized. This study documented CBR1 expression in hearts from donors with DS (n = 4) and donors without DS (n = 15). The DS samples showed 1.8-fold higher CBR1 mRNA levels compared to the non-DS samples (levels in DS samples were 3.3-relative fold, and those in non-DS were 1.8-relative fold; p = 0.012). CBR1 protein levels were 1.9-fold higher in DS samples than in non-DS samples (13.5 ± 7.7 versus 7.2 ± 3.9 nmol/g cytosolic protein, respectively; p = 0.029). CBR1 activity for daunorubicin was 1.7-fold higher in DS samples than in non-DS samples (3.8 ± 0.1 versus 2.3 ± 0.2 nmol daunol/min · mg, respectively; p = 0.050). CBR1 1096G>A (rs9024) affects CBR1 activity, and one heart trisomic for the variant A allele (A/A/A) exhibited low enzymatic activity. These findings suggest that increased CBR1 expression in the hearts of individuals with DS may contribute to the risk of anthracycline-related cardiotoxicity.