A Preclinical Bioavailability Strategy for Decisions on Clinical Drug


A Preclinical Bioavailability Strategy for Decisions on Clinical Drug...

0 downloads 117 Views 2MB Size

Subscriber access provided by - Access paid by the | UCSB Libraries

Article

A Preclinical Bioavailability Strategy for Decisions on Clinical Drug Formulation Development: an In Depth Analysis An Van Den Bergh, Sandy Van Hemelryck, Jan Bevernage, Achiel Van Peer, Marcus Brewster, Claire Mackie, and Erik Mannaert Mol. Pharmaceutics, Just Accepted Manuscript • DOI: 10.1021/acs.molpharmaceut.8b00172 • Publication Date (Web): 25 May 2018 Downloaded from http://pubs.acs.org on May 28, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 36

Molecular Pharmaceutics

1 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

A Preclinical Bioavailability Strategy for Decisions on Clinical Drug Formulation Development: an In Depth Analysis An

Van

den

Bergh1,

Sandy

Van

Hemelryck2,

Jan

Bevernage3,

Achiel Van Peer2, Marcus Brewster3, Claire Mackie4, Erik Mannaert2 1

Drug Metabolism and Pharmacokinetics

2

Clinical Pharmacology and Pharmacometrics,

3

Pharmaceutical Sciences

4

Drug Product Development

Janssen Pharmaceutica, Turnhoutseweg 30, B-2340 Beerse, Belgium

ABSTRACT The aim of the presented retrospective analysis was to verify whether a previously proposed Janssen Biopharmaceutical Classification System (BCS)-like decision tree, based on preclinical bioavailability data of a solution and suspension formulation, would facilitate informed decision making on the clinical formulation development strategy. In addition, the predictive value of (in vitro) selection criteria such as solubility, human permeability and/or a clinical dose number (Do) were evaluated, potentially reducing additional supporting formulation bioavailability studies in animals. The absolute (Fabs,sol) and relative (Frel,

susp/sol)

bioavailability of an oral solution and

suspension, respectively, in rat or dog and the anticipated BCS classification were analyzed for 89 Janssen compounds with 28 of these having Frel,susp/sol and Fabs,sol in both rat and dog at doses around 10 and 5 mg/kg respectively. The bioavailability outcomes in the dog aligned well with a BCS-like classification based upon the solubility of the Active Pharmaceutical Ingredient (API) in biorelevant media, while the alignment was less clear for the bioavailability data in the rat. A retrospective analysis on the clinically tested formulations for a set of 12 Janssen compounds confirmed that the previously proposed animal bioavailability based decision tree facilitated decisions on the oral formulation type, with the dog as the most discriminative species. Furthermore, the analysis showed that based on a Do for a standard human dose of 100 mg in aqueous and/or biorelevant media, a similar formulation type would have been selected compared to the one suggested by the animal data. However, the concept of a Do did not distinguish between solubility enhancing or enabling formulations and does 1 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 2 of 36

2 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

not consider the API permeability and hence produces the risk of slow and potentially incomplete oral absorption of an API with poor intestinal permeability. In cases where clinical dose estimations are available early in development, the preclinical bioavailability studies and dose number calculations, used to guide formulation selection, may be performed at more relevant doses instead of the proposed standard human dose. It should be noted however, that unlike in late development, there is uncertainty on the clinical dose estimated in the early clinical phases because that dose is usually only based on in vitro and/or in vivo animal pharmacology models, or early clinical biomarker information. Therefore, formulation strategies may be adjusted based on emerging data supporting clinical doses. In summary, combined early information of in vitro assessed API solubility and permeability, preclinical suspension/solution bioavailability data in relation to the intravenous clearance and metabolic pathways of the API can strengthen formulation decisions. However, these data may not always fully distinguish between conventional (e.g. to be taken with food), enhancing and enabling formulations. Therefore, to avoid over-investment in complex and expensive enabling technologies, it is useful to evaluate a conventional and solubility (and/or permeability) enhancing formulation under fasted and fed condition, as part of a first-inhuman study or in a subsequent early human bioavailability study, for compounds with high Do, a low animal Frel,susp/sol or low Fabs,sol caused by precipitation of the solubilized API.

KEYWORDS: absolute and relative bioavailability, rat and dog, physico-chemical properties, Active Pharmaceutical Ingredient (API), Biopharmaceutical Classification System (BCS), solubility in biorelevant media, decision tree, conventional, enhancing and enabling oral formulation, clinical dose number.

2 ACS Paragon Plus Environment

Page 3 of 36

Molecular Pharmaceutics

3 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

INTRODUCTION The oral route is the most commonly used route for drug administration in patients. However, slow or incomplete dissolution of drug substance and potential precipitation throughout the gastro-intestinal tract of poorly soluble drug may prevent therapeutic drug exposure in the patient following oral administration. As it is generally recognized that the majority of drug candidates in pharmaceutical research and development are poorly soluble and lipophilic1, achieving sufficient oral bioavailability is a prerequisite for therapeutic success but can often be challenging. To face the challenge, adequate formulation selection guidance is needed to ensure the required therapeutic exposure is achieved with a reasonably appropriate resource allocation. Throughout the development of a drug candidate, many decisions have to be made regarding the oral drug delivery system. Over the last decade, numerous in vitro, in silico and in vivo tools are developed to support the development of oral formulations in pharmaceutical industry2. The Innovative Medicines Initiative (IMI) funded Oral Biopharmaceutical Tools (OrBiTo) project is centered around the development of new and further refinement of existing tools, with the aim to improve insight into drug absorption, better guide oral formulation development and predict in vivo performance3. In an unpublished survey held across European Pharmaceutical Industry Association (EFPIA) partners involved in OrBiTo, participants were asked when, how and which in vivo tools are currently being applied during the oral formulation development process. All 12 responders indicated they applied animal models for their formulation development, especially in drug discovery or early development phases, dogs and rats being the most popular species. Second to animal studies, more than half of the responders indicated they often or always apply human studies in formulation development. Overall, the survey clearly indicated that companies heavily rely on animal but also human models for their formulation development. In the present paper, a formulation decision tree based on absolute and relative bioavailability data of solution and suspension formulations respectively, previously proposed by Mackie et al.4,

5

and Brewster et al.6 for implementation at Janssen, was evaluated through a

retrospective analysis of an exhaustive Janssen compound database containing bioavailability studies in animals (rat and dog). In addition, available physico-chemical data of Active Pharmaceutical Ingredients (APIs) (e.g. ionization behavior, in vitro solubility and 3 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 4 of 36

4 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

permeability) were considered and attempts were made to link these to available animal in vivo bioavailability information. Specific cases, for which human bioavailability data were also available, were more extensively investigated to link the outcome of the preclinical work to the final formulation concepts used in human. The results led to refinement of the earlier proposed decision tree for the selection of the clinical formulation development strategy based on animal bioavailability data and to an additional formulation decision tree based on API solubility data. METHODS The work described in this paper was performed in 3 parts, each having its own compound database (with some overlap) and corresponding evaluations, as summarized in Table 1. Part

1:

Physico-chemical

API

characteristics,

BCS-like

classification

and

suspension/solution bioavailability in preclinical species In part 1, the link between animal bioavailability of API in suspension and solution formulations, physico-chemical characteristics of APIs and the previously proposed BCS-like biopharmaceutical decision tree4-6 was evaluated using an extensive database of Janssen compounds. The database consisted of 89 Janssen APIs including their physico-chemical characteristics and the results of absolute and relative bioavailability studies in male Sprague Dawley rats and/or male beagle dogs (Supporting information 1). For 28 of these APIs, relative bioavailability data were available in both rat and dog. The APIs were mostly solubilized in mixtures of polyethylene glycol (PEG) or cyclodextrin and water, or were suspended in methylcellulose in their unmilled crystalline form. Dogs were dosed either in fed or fasted condition (after an overnight fast), whereas rats were dosed always fed (standard dog (Labdiet ® - 5L66) and rodent (Scientific Animal Food and Engineering – VRF-1) food was used). The intravenous dose was usually 2.5 mg/kg in the rat and 1 mg/kg in the dog. The oral doses in the rat and dog were usually 10 mg/kg and 5 mg/kg, respectively. These doses were considered as reasonably representative for human because the corresponding human equivalent doses (HED)7 are 113 mg for the rat and 194 mg for the dog. A dosing volume of 20 mL and 2.5 mL was considered for dogs and rats, respectively. In the relative bioavailability studies, the dose of the solution was equal to the dose of the suspension. The in

4 ACS Paragon Plus Environment

Page 5 of 36

Molecular Pharmaceutics

5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

vivo studies in rat and dog were performed under experimental study conditions approved by an Ethics Committee. Absolute (Fabs,sol) and relative (Frel,susp/sol) oral bioavailability values were calculated using the area under the respective plasma API concentration-time curves:  ℎ   .  ∗   ( )  ,  =  ℎ  (. ) ∗    ( )   , / =

 (. ℎ/)  (. ℎ/)

The apparent first-pass extraction in the liver (E = Clb/Q) was calculated based on the intravenous blood clearance Clb and liver blood flow Q taking into account an experimental or in silico B/P ratio predicted by ADMET Predictor v8.0 (Simulations Plus). A Fabs,sol above 0.7 was used as a surrogate for high in vivo intestinal absorption, whereas a Fabs,sol below 0.7 was considered to reflect either an incomplete absorption, a first pass extraction effect, or a combination of both. First pass extraction in the gut was assumed negligible in the current analysis. An Frel,susp/sol above or below 0.8 was used to differentiate between solubility/dissolution unlimited or solubility/dissolution limited absorption from the suspension. The following BCS-like classification of the APIs was applied based on the in vivo data: BCS I-like: Fabs,sol above 0.7 and the suspension – solution Frel,susp/sol above 0.8 in the animal BCS II-like: Fabs,sol above 0.7 and the suspension – solution Frel,susp/sol below 0.8 in the animal BCS III-like: Fabs,sol below 0.7 and the suspension – solution Frel,susp/sol above 0.8 in the animal BCS IV-like: Fabs,sol below 0.7 and the suspension – solution Frel,susp/sol below 0.8 in the animal In vitro determined logP, pKa (of strongest acid/basic function), permeability (measured in Loot Limit Chaotic Player Kill MultiDrug Resistance protein 1 (LLC-PK-MDR1) or MadinDarby Canine Kidney MultiDrug Resistance protein 1 (MDCK-MDR1) cell lines and converted to human effective intestinal permeability (Peff) using a Janssen correlation), API solubility (thermodynamic solubility measured at 37°C) in aqueous buffers of pH 2, 4, 7.4 and API solubility in fasted state simulated intestinal fluid (FaSSIF8) and fed state simulated intestinal fluid (FeSSIF8) were also added to the database. To facilitate the analysis, APIs were

5 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 6 of 36

6 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

categorized according to their functional groups (APIs were considered neutral when the pKa acid was greater than 9 and when pKa base was lower than 2), the Peff (calculated using an internal conversion file depending on the cell line used), the solubility (ranges per log unit were considered) and the percentage ionized in stomach and duodenum (dependent on species-specific gastro-intestinal pH’s). If experimental values of logP and pKa were not available, an in silico value predicted by ADMET Predictor v8.0 (Simulations Plus) was used. Part 2: Clinical formulation type selection based on suspension/solution bioavailability in preclinical species or dose number In Part 2, it was verified whether the clinical formulation type selection was in agreement with the proposed animal bioavailability based decision tree4-6. In addition, it was evaluated if in vitro data, i.e. Peff, API solubility and/or Do would guide us to a similar formulation type as the decision tree based on animal pharmacokinetic data. Therefore, the link between the clinical formulation type and preclinical bioavailability of API in suspension and solution (Fabs,sol and Frel,susp/sol), Peff, API solubility and Do was evaluated. This analysis was performed on a second database (partially overlapping with the first database) consisting of Janssen compounds 1-12 for which preclinical data and information on clinical formulation type were available (Supporting information 2 and 3). The database included physico-chemical characteristics and results of absolute and relative bioavailability studies in rats and/or dogs, supplemented with information of the formulation type that was developed for clinical trials. Preclinical Frel,susp/sol and Fabs,sol data were available in both rat and dog for 6 compounds, only in rat for 5 compounds and only in dog for 1 compound. In Part 2, the pKa and logP values were calculated using ADMET Predictor v8.0. The calculations of the in vivo, solubility and permeability parameters used in the analysis are explained in the previous paragraph (Part 1). In the current section, a Do was calculated based on the API solubility as described below9:  =

!"# # $  ()  250 ( )* "+ (  )

The Do was calculated using an arbitrarily selected standard human dose of 100 mg and solubility data in simple aqueous buffers (aqueous Do: for neutral and basic compounds and amphoteric compounds with strongest basic function, the API solubility in an aqueous buffer 6 ACS Paragon Plus Environment

Page 7 of 36

Molecular Pharmaceutics

7 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

at pH 6-8 was used, for acidic compounds and amphoteric compounds with strongest acidic function the API solubility in an aqueous buffer at pH 1-2 was used), FaSSIF8 or FeSSIF8 medium (biorelevant Do). For adequate data visualization in Figure 5, a scale of 1 to 20 was applied for the Do, by equalizing Do < 1 to 1 and Do > 20 to 20. When the API solubility was reported as higher than a specific value, that specific value was considered to be the API solubility. For API solubility reported as lower than a specific value, the Do was assumed to be 20. The clinical formulations were classified as conventional, enhancing or enabling formulations. Conventional formulations are formulations containing no additives to modify drug solubility, release/dissolution rate, permeability and/or stability. A few examples of applied manufacturing technologies considered in the conventional formulation type are direct compression, dry/wet granulation and encapsulation. This paper refers to solubility enhancing formulations when API solubility or dissolution enhancement is targeted using conventional manufacturing technologies e.g. salt forms of the API or inclusion of solubility or dissolution enhancing excipients (e.g. surfactants, precipitation inhibitors). Addition of permeation enhancers is considered as permeation enhancing formulations.

The term

solubility enabling formulations is used when solubility/dissolution enhancement is targeted using complex manufacturing technologies. Examples are solid dispersions, nanosuspensions and lipid-based systems. The distinction between solubility and/or permeability enhancing methods and enabling formulation technologies is made because the formulation development cost and timelines are impacted by the complexity of the applied manufacturing technology. In Part 2, when multiple formulations were developed for human use, the most complex formulation type (enabling > enhancing > conventional) was included in the analysis. It must be noted that the clinical formulation type in Part 2 was not proven to result in adequate human exposures, since relative bioavailability data of the solid formulation(s) versus an oral solution in human were not available. Part 3: Clinical formulation type, preclinical and clinical solid/solution bioavailability and dose number To evaluate if the formulation type selected based on Do is adequate to guarantee the desired clinical performance in Part 3 the link between the clinical Do and human exposure for conventional, enhancing or enabling clinical solid formulations was evaluated.

7 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 8 of 36

8 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

This analysis was performed on a database (partially overlapping with the second database) consisting of Janssen compounds 10-22, for which physico-chemical data and human bioavailability of oral solid and solution formulations were available (Supporting information 2 and 4). Clinical formulations were dosed either in fed or fasted conditions10. The dose normalized relative exposures of different clinical and preclinical oral solid formulations versus solutions (AUCsolid/sol) were discussed for 3 compounds in order to provide examples of clinical formulation type selection strategies. Physico-chemical parameters were calculated as described in Part 2. The Do was calculated using the actual clinical dose used in the respective clinical bioavailability studies instead of a standard human dose of 100 mg used in Part 2. The relative exposure of solid versus solution formulations in preclinical and clinical studies (AUCsolid/sol) was calculated by dividing the dose normalized AUCinf or AUClast observed for the solid formulations by one of the solution formulations at the same or similar dose. The clinical formulation types assigned in Part 3 were the ones for which AUCsolid/sol data were determined in the respective preclinical or clinical bioavailability studies. RESULTS Part 1: Preclinical suspension/solution bioavailability, physico-chemical characteristics of the API and BCS classification Physico-chemical properties of APIs The distribution of different physico-chemical properties of the 89 Janssen APIs included in the database are illustrated in Figure 1 (Supporting information 1). The majority of the Janssen APIs in this analysis were weak bases, fairly lipophilic (logP > 2) with a Peff > 2*10-4 cm/sec. The solubility of about half of the APIs in water was not available because the solubility in water was below the detection limit as measured by an early research high performance liquid chromatography method.

Higher solubility values were obtained in

simulated intestinal media that reflected fasted (FaSSIF) and fed (FeSSIF) state intestinal conditions. Preclinical Bioavailability – BCS classification

8 ACS Paragon Plus Environment

Page 9 of 36

Molecular Pharmaceutics

9 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Figure 2 displays the Frel,susp/sol against the Fabs,sol for both dog and rat. The colors of the markers reflect the FeSSIF solubility (Figure 2A,B,C) or FaSSIF solubility (Figure 2D,E,F); the size of the markers reflect the in silico Peff (Figure 2A,D), the in vitro Peff (Figure 2B,E) and the first pass extraction in the liver (Figure 2C,F). Not all APIs had an experimentally measured Peff. Figure 2 shows that in dog the very poorly soluble compounds in FaSSIF (< 0.01 mg/mL) had mostly an Frel,susp/sol of less than 20%. For most APIs with a FeSSIF solubility higher than 0.1 mg/mL, the exposure in the dog after dosing a suspension was comparable to the exposure after dosing of the API in a solution (Frel,susp/sol close to 1). Unlike in the dog, the trend in the rat was less clear: higher relative bioavailability or even complete oral absorption was observed in the rat for very poorly soluble APIs. Differences in Peff in the current database appeared to be less critical as an explanation for the observations in both the rat and the dog (Figure 2A and Figure 2B). APIs with a high first pass effect and high intravenous clearance are mainly situated in the area of a low Fabs,sol (Figure 2C). Similar, but less clear observations as in Figures 2A, B and C were obtained when the APIs were categorized according to the in vitro solubility in FaSSIF (Figures 2D, E and F). Figure 3 shows that the Frel,susp/sol in rat was higher compared to the Frel,susp/sol in the dog for most of the APIs for which both data in rat and dog were available. The prandial state (Figure 3), nor the % ionized in stomach or duodenum (not shown) appeared to be an explanation for the difference between the rat and the dog. Part 2: Clinical formulation selection based on preclinical suspension/solution bioavailability and Do Physico-chemical properties of APIs The physico-chemical characteristics of the compounds used in this part of the analysis are included in Supporting information 2. Out of the 12 compounds considered in Part 2, 10 compounds were weak bases, 1 compound was a weak acid, and 1 was amphoteric. The majority of the APIs were fairly lipophilic (logP > 2) and had an in silico Peff > 2*10-4 cm/sec. Preclinical bioavailability - Clinical formulation type Figure 4 displays the Frel,susp/sol against the Fabs,sol in both dog and rat for the 12 Janssen compounds for which an oral clinical solid or liquid formulation was developed. 9 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 10 of 36

10 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Although limited data (7 APIs) were available in the dog, enabling or enhancing solid or liquid formulations had been used in humans when the Frel,susp/sol was lower than 0.4, whereas a conventional solid formulation in humans was used in human when the Frel,susp/sol in the dog was higher than 1. The dog Frel,susp/sol cut-off for enhancing or enabling formulations lies somewhere between 0.4 and 1. Again, the Frel,susp/sol data in the rat were mostly higher than the Frel,susp/sol in the dog and results in rat seemed less discriminative towards the type of clinical formulation that has been developed for early clinical studies. The rat Frel,susp/sol cut-off for enhancing or enabling formulations in this dataset lies around 0.7-0.8, although exceptions (i.e. compounds 3, 4, 7 and 9) were observed. Compounds 3 and 7 with Frel,susp/sol < 0.8 were formulated as a conventional formulation. The low aqueous and biorelevant Do may have contributed to the formulation type selection for these compounds (Figure 5). Compounds 4 and 9 with Frel,susp/sol ≥ 0.8 were formulated as a solution containing cyclodextrins or PEG400 (considered as enhancing technology), which is often used for early clinical studies to maximize the API solubilization with minimal formulation development efforts. Both compounds had a high aqueous and FaSSIF Do (Figure 5). For compounds with low Frel,susp/sol in the dog, solubility enhancing formulations were mainly selected when Fabs,sol was high (≥ 0.7), while enabling formulations were developed when Fabs,sol was low (< 0.7). This trend was not clear in the rat. Figure 5 displays the Frel,susp/sol against the Do calculated based on a standard human dose of 100 mg and the aqueous, FaSSIF or FeSSIF API solubility (panels A, B and C respectively). In the dog, low Frel,susp/sol values (< 0.8) generally corresponded to APIs with an aqueous and biorelevant Do ≥ 10, while high Frel,susp/sol values (≥ 0.8) corresponded to a biorelevant Do < 10. Compound 12, with low Frel,susp/sol and low FeSSIF Do was an exception. Since dogs were mostly dosed in fasted state at Janssen, the FaSSIF Do may have been more predictive for Frel,susp/sol in the dog than the FeSSIF Do. For the rat, the correlation between Frel,susp/sol and Do was less clear. Conventional formulations were generally developed for compounds having a low Do (aqueous Do < 10 and biorelevant Do < 5), while enhancing or enabling technologies were applied for APIs with a high Do (aqueous Do ≥ 10 and biorelevant Do ≥ 5). Compound 12 was again an exception, as it was formulated using an enabling (solid dispersion beads in capsule) and enhancing (tablet with salt form and surfactant) technology, while FeSSIF Do 10 ACS Paragon Plus Environment

Page 11 of 36

Molecular Pharmaceutics

11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

was < 5. The high FaSSIF Do may explain this formulation type selection, to guarantee adequate clinical performance, irrespective of the prandial state. In addition, the maximum intended clinical dose (800 mg) was much higher than the standard human dose used for the Do calculations and would have resulted in a FeSSIF Do ≥ 5. No correlation was observed between formulation type and the Peff of the investigated APIs (not shown). However, the database mainly contained compounds with a high Peff. Part 3: Clinical formulation type based on Preclinical and clinical solid/solution bioavailability and dose number Physico-chemical properties of APIs The physico-chemical characteristics of the compounds used in this part of the analysis are included in Supporting information 2. Eight compounds of the 13 compounds considered in Part 3 were weak bases, 1 compound was a weak acid, 1 was neutral and 3 compounds were amphoteric. The majority of the APIs were fairly lipophilic (logP > 2) and had an in silico Peff > 2*10-4 cm/sec. Dose number and formulation type Figure 6 displays the bioavailability of clinical solid formulation types relative to an oral solution (marker color) as a function of the clinical Do (marker size) calculated based on the actual tested clinical dose and the aqueous, FaSSIF or FeSSIF API solubility (panels A, B and C respectively). Conventional formulations for compounds with a biorelevant Do < 5 had a human bioavailability close to that of a solution (AUCsolid/sol is close to 1). For compound 17, with a low FaSSIF and FeSSIF Do of 1, an enhancing formulation was developed, resulting in an AUCsolid/sol close to 1. A conventional formulation may have led to similar exposures, based on the low biorelevant Do. However, this formulation type was not tested for the API. Compounds with high biorelevant Do ≥ 5 were generally developed as enhancing and/or enabling formulations, resulting in an AUCsolid/sol close to or higher than 1. For compound 12, with high FaSSIF Do ≥ 5 and low FeSSIF Do < 5 at the tested clinical dose of 100 mg, the solubility enhancing formulation showed lower exposures than a solution in fasted state, while exposures in fed state and exposures of the solubility enabling formulation in both fasted and fed state resulted in an AUCsolid/sol close to 1. Also for compound 19, with high FaSSIF Do ≥ 11 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 12 of 36

12 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

5 and low FeSSIF Do < 5, a conventional formulation resulted in a low AUCsolid/sol in the fasted state (AUCsolid/sol of 0.6) and an AUCsolid/sol close to 1 in the fed state. A conventional formulation containing compound 20, with high FaSSIF and FeSSIF Do ≥ 5, showed a low AUCsolid/sol in the fed state (AUCsolid/sol of 0.2 – 0.4).

12 ACS Paragon Plus Environment

Page 13 of 36

Molecular Pharmaceutics

13 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Case Examples For compounds 19, 20, and 22, relative exposures of different oral solid formulations versus solutions in humans (AUCsolid/sol) are discussed and compared to those of similar formulations in animals, if available, in order to provide examples of clinical formulation type selection strategies. Compound 19 Compound 19 was a poorly soluble, weak amphoteric compound (logP of 1.93; basic pKa of 3.37 and acidic pKa of 11.32). FaSSIF Do (3.5 or 4.7) was higher than FeSSIF Do (2.7 or 3.6) for the clinical doses tested (300 or 400 mg). The relative bioavailability (AUCsolid/sol) of a conventional solid formulation (direct compression tablet) was about 0.6 compared to a solution when administered in the fasted state, whereas administration with food increased the oral absorption of the API as compared to API in a solution containing PEG400 and D-αtocopheryl polyethylene glycol 1000 succinate in fed state (AUCsolid/sol of 1.0 to 1.2) (Figure 7). The AUCsolid/sol for a solid formulation with solubility enhancing technology (addition of surfactant) dosed in fed conditions was 1.2 to 1.3, whereas a solubility enabling technology (a self emulsifying drug delivery system) improved the exposure relative to a solution by 10% to 40% under fasted conditions (AUCsolid/sol of 1.1 to 1.4). Compound 20 Compound 20 was a BCS class IV, weak basic compound (logP of 3.88; basic pKa of 2.77) with a high FaSSIF (≥ 10) and FeSSIF Do (4.8 or 9.6) for the two clinical dose levels tested (200 or 400 mg). Compared to a solution, the relative bioavailability of a conventional solid formulation or a solubility enhancing formulation containing a salt form of the API and/or a surfactant was around 0.1 in the dog. In human, the solubility enhancing formulations showed a similar low bioavailability (AUCsolid/sol of 0.2 to 0.4). Multifold higher AUCsolid/sol values were achieved for various solubility enabling technologies (API on beads, granulolayering, solid dispersion, solid solution) in human (Figure 8). Compound 22 Compound 22 was a BCS class II, weak basic compound (logP of 4.03; basic pKa of 5.14) with an intermediate to high FaSSIF (3.6 or ≥ 10) and FeSSIF Do (2.4 or 9.8) for the two clinical dose levels tested (25 or 100 mg). No conventional solid formulation was studied in 13 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 14 of 36

14 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

humans. Solubility enhancement by using a salt form of the API and a surfactant in a tablet resulted in an AUCsolid/sol about 0.7 in fasted conditions and an AUCsolid/sol close to 1 when dosed in the fed state (Figure 9). In the dog, the AUCsolid/sol of the solubility enhancing technology varied between 0.6 and 1. Solubility enabling technology in the dog resulted in an AUCsolid/sol ranging between 0.7 and 1.5 for solid dispersion beads in capsules and between 1.6 and 4.0 for liquid filled capsules. DISCUSSION In the present paper, a previously proposed formulation decision tree (Mackie et al.4-5; Brewster el al.)6 for the selection of the clinical formulation development strategy at Janssen based on preclinical data was verified. At Janssen, preclinical suspension/solution bioavailability studies are performed early in development, given the versatile applicability of solutions and suspensions across development (including use in toxicology studies), to provide a first insight into the in vivo behavior of the compound, including its pharmacokinetics. The formulation decision tree uses these in vivo data to categorize candidate APIs in BCS-like classes. Based on the categorization, guidance was proposed regarding the complexity of the clinical formulation development, i.e. is the need for development of an enabling formulation likely, or will a conventional formulation be sufficient? The verification of the proposed decision tree via the in depth analysis in this paper consisted of 3 parts. In Part 1, we determined whether preclinical suspension/solution bioavailability data together with the physico-chemical properties of the API are indeed in agreement with the BCS-like classification. In Part 2, we evaluated whether the animal bioavailability data reflect the clinical formulation of choice. Furthermore, we investigated whether Peff, API solubility and/or the Do (based on a standard human dose) would have pointed to the same formulation selection. In Part 3, the human bioavailability of selected solid formulations was evaluated to confirm the adequacy of the proposed formulation selection strategies. As in other small molecule database analyses1, the majority of the Janssen APIs used were weak bases, fairly lipophilic (logP > 2) with rather poor aqueous solubility and a Peff > 2*10-4 cm/sec, which can be considered high11. If there was an acid function in the molecule, the acid pKa of the API was mostly above 8 (Figure 1), hence these APIs were unionized within the physiological gastro-intestinal pH range12,13. Therefore, conclusions from the present retrospective portfolio analysis are limited to bases and neutral APIs. 14 ACS Paragon Plus Environment

Page 15 of 36

Molecular Pharmaceutics

15 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Dog bioavailability data and to a lesser extent rat bioavailability data were in agreement with the BCS-like categorization The

Abstract

figure

depicts

the

BCS-like

categorization

based

on

preclinical

suspension/solution bioavailability data, according to the earlier Janssen proposed decision tree4-6. A preclinical Fabs,sol cut-off of ≥ 0.7 is less conservative than the human absolute bioavailability limit of ≥ 85% required in the BCS class I biowaiver guidances to demonstrate high permeability of APIs14. The Frel,susp/sol cut-off of ≥ 0.8 was defined based on solubility data (Part 1) and clinical formulation type information (Part 2). It should be noted that the investigated API dataset of Parts 1 and 2 mainly included in vitro or in silico high permeable compounds (Figure 1, Supporting information 1, 2). As displayed in Figure 2, the retrospective analysis shows that the dog suspension/solution bioavailability data were in good agreement with the BCS-like categorization. Candidate drugs with low Frel,susp/sol in the dog were mostly molecules with a low FaSSIF and FeSSIF solubility (less than 0.01 mg/mL), whereas candidate APIs with a high Frel,susp/sol were mostly those with high solubility in biorelevant media (above 0.1 mg/mL). Higher Frel,susp/sol and even complete oral absorption was sometimes observed in the rat for APIs with a poor in vitro solubility in biorelevant media of less than 0.01 mg/mL (Figure 2), indicating that the dog is the more discriminating animal species in this retrospective analysis and API dataset. It should be noted that the pH of the stomach in the dog is highly variable and on average more basic than in the rat12,13. Pretreatment of the dog with an intravenous dose of a gastric pH lowering agent, e.g. pentagastrin, is often applied during oral bioavailability studies in dog to standardize and ensure acidity of the dog stomach15 but was not applied for the dog bioavailability studies of our in depth analysis. Considering the large proportion of weak basic compounds in the dataset, lack of pretreatment of the dog with pentagastrin in the current dataset could be a theoretical explanation for the lower Frel,susp/sol of the suspension in the dog compared to the Frel,susp/sol in the rat for the very poorly soluble APIs. However, further exploration of the experimental Frel,susp/sol values did not show any dependence of the percentage ionization of the API in the stomach and intestine (not shown). For the calculation of the percentage ionization a pH value of 3 and 6.2 was used for the fasted dog, 5 and 6.2 for the fed dog and 3.2 and 5 for the fed rat stomach and intestine respectively (based on GastroPlus v9.0 physiological parameters). Furthermore, the lower Frel,susp/sol in the dog could not be explained by a food effect as the trend was observed for APIs dosed under fasted and postprandial state of the dog (Figure 3). Nevertheless, it should be acknowledged that the bile 15 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 16 of 36

16 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

concentration in the intestine is on average higher in the rat compared to the dog due to a permanent biliary drainage in the rat and a food-induced secretion of bile in the dog, which may have partially led to our observations. The low preclinical absolute bioavailability in our dataset was mainly caused by high hepatic first pass extraction of high intravenous clearance APIs. Although not part of the analysis, early knowledge of potential involvement of gut wall metabolism (e.g. CYP3A4 or conjugation) and transporters will strengthen the decision tree, particularly with respect to determining whether an animal absolute bioavailability below 0.7 is related to additional API metabolism in the gut wall16,17 in contrast to incomplete API absorption. In the current analysis, Frel,susp/sol study data were obtained after dosing 5 mg/kg and 10 mg/kg in the dog and rat. Pilot formulation bioavailability and a food effect investigation with a more appropriate lower or higher dose can be included in the first-in-human dose escalation study or a subsequent human bioavailability study. Preclinical bioavailability data, though also API solubility data (or Do), reflect the clinical formulation of choice For the APIs that reached the clinical phase in the investigated dataset (Part 2), the developed clinical formulation(s) aligned well with the bioavailability findings in the preclinical species and the suggested formulation type. Conventional formulations were developed for APIs with high preclinical Frel,susp/sol values whereas solubility enabling or enhancing formulation types were developed for APIs with low preclinical Frel,susp/sol values (Figure 4). In the analyzed data set of Part 2, no highly soluble/poorly permeable BCS class III APIs were present so no information was available about any potential impact of permeation enhancing excipients18,19 in clinical solid formulations. Fabs,sol data in dog seem to allow a distinction between solubility enhancing (Fabs,sol ≥ 0.7) and enabling formulations (Fabs,sol < 0.7). It could be hypothesized that for compounds with a low Frel,susp/sol, a high Fabs,sol is likely to correspond to a high permeability (ie. BCS class II), allowing complete dissolution of the API, even if its solubility is only moderately increased by enhancing formulation technologies. On the other hand, a low Fabs,sol could relate to a lower permeability and therefore, a more drastic increase in solubility of the API would be needed by applying enabling technologies to ensure adequate bioavailability of oral solid formulations, like often required for BCS class IV compounds.

16 ACS Paragon Plus Environment

Page 17 of 36

Molecular Pharmaceutics

17 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

However, the APIs in the dataset of Part 2 were mainly highly permeable (Supporting information 2) and a low Fabs,sol may also be linked to high first pass metabolism. Based on the solubility of the API or Do calculated with a standard human dose of 100 mg, the retrospective analysis in Part 2 obtains a similar formulation type selection as was indicated by the preclinical suspension/solution bioavailability study, suggesting a reduced need for animal bioavailability studies. However, no distinction could be made between solubility enhancing and enabling formulations and the analysis did not contain poorly permeable compounds, for which preclinical suspension/solution data may be valuable. In general, compounds with low Do (aqueous Do < 10, biorelevant Do < 5) correspond to a high Frel,susp/sol in animals (especially in dogs) and these APIs were developed in a conventional formulation for humans. On the other hand, compounds with high Do (aqueous Do ≥ 10, biorelevant Do ≥ 5) showed low animal Frel,susp/sol values and these APIs were formulated using enabling and/or enhancing technologies. Proposed clinical formulation type selection strategies Based on the analysis described in Part 1-3, stepwise approaches based on in vitro assessed API solubility and permeability and on preclinical suspension/solution bioavailability data in relation to the intravenous clearance of the API are proposed to support formulation decisions. Figure 10 shows a strategy starting from the API solubility and calculated Do based on a standard human dose of 100 mg or a more representative clinical dose estimate, if available. The API solubility and Do are first determined in aqueous medium. If the aqueous Do is high, the API solubility can be determined in FaSSIF and/or FeSSIF medium to calculate the biorelevant Do in fasted and/or fed state. If the Do is low, conventional formulations may be adequate. For high biorelevant Do, a preclinical suspension/solution bioavailability study may be performed in the dog as the most discriminative species, to evaluate the need for a solubility enhancing or enabling formulation, as shown in the Abstract figure. For compounds with a high Frel,susp/sol (≥ 0.8) a conventional formulation can be developed when Fabs,sol is high (≥ 0.7) or when Fabs,sol is low due to a high first pass effect. When Fabs,sol is low due to a low Peff

or

precipitation

of

solubilized

API,

permeability

enhancing

or

solubility

enhancing/enabling formulations are proposed respectively. Note that Frel,susp/sol values significantly higher than 1 may indicate that the excipients, particle size and/or gut wall metabolism of the suspension are favourable to bioavailability compared to those of the (precipitated) solution. For compounds with a low Frel,susp/sol, solubility enabling formulations 17 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 18 of 36

18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

should be considered, while solubility enhancing formulations may be appropriate when Fabs,sol is high. If the clinical Do is known or estimated to be low, conventional formulations may be developed in cases where solubility enhancing or enabling technologies are proposed. It should be noted, that therapeutic dose estimates made at the discovery and preclinical research stage can be indefinite when based on in vitro and/or in vivo animal pharmacology models. When the need for enabling formulations cannot be excluded based on in vitro or preclinical data, it may be useful to evaluate early in development if a conventional or an enhancing formulation type can lead to the desired therapeutic exposure profile, to avoid significant investments in complex technologies. Therefore, these relatively simple formulations could be included in one of the arms of a first-in-human (e.g. single ascending dose) study for compounds with high Do, low preclinical Frel,susp/sol, or low Fabs,sol caused by precipitation of the solubilized API in order to decide if enabling formulation development is needed. As an alternative or in addition, a preclinical and/or clinical bioavailability study with various solid formulations may further guide formulation selection. For compounds 20 and 22, with a biorelevant Do ≥ 5 for the highest clinical dose level tested, the latter approach was applied. A conventional and enhancing formulation of compound 20 resulted in a low AUCsolid/sol in dogs. In humans, enabling formulations showed multifold higher AUCsolid/sol values as compared to enhancing formulations. One of these high performing enabling formulations was selected for commercialization to minimize the therapeutic dose, pill burden and cost of goods. For compound 22, no conventional formulations were evaluated in preclinical or clinical studies. In dogs, the AUCsolid/sol of enhancing formulations varied between 0.6 and 1 in fed state. The AUCsolid/sol of enabling formulations ranged between 0.7 and 1.5 for solid dispersion capsules in fed state and between 1.6 and 4.0 for liquid filled capsules. Due to shelf-life related issues, the liquid filled capsules were not evaluated in human. The AUCsolid/sol for the solubility enhancing formulation in human was 0.7 in fasted state and close to 1 in fed state. The solid dispersion capsule formulation resulted in a AUCsolid/sol of 0.9 in fed state. Taking into account manufacturing complexity and/or cost, the enhancing tablet (to be taken with food) was selected for commercialization over the solid dispersion capsule. If Do is high in FaSSIF but low in FeSSIF medium, a human bioavailability study or a pilot study arm in the first-in-human dose escalation study with a conventional formulation coadministered with food may be considered to show sufficient bioavailability as an alternative to an investment in solubility enhancing or enabling formulation technology. For 18 ACS Paragon Plus Environment

Page 19 of 36

Molecular Pharmaceutics

19 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

compound 19, a conventional formulation with FaSSIF Do ≥ 5 and FeSSIF Do < 5 showed an AUCsolid/sol of 0.6 in human in fasted state, while AUCsolid/sol was close to 1 (1.0 – 1.2) in fed state and for enhancing and enabling formulations (1.1 – 1.4). Based on these clinical data, the conventional formulation (to be taken with food) was commercialized, since enhancing or enabling technologies did not provide a significant increase in exposure. Lastly, pharmacokinetic simulations may help to provide insight into whether the early solid formulation would be sufficient to obtain the desired exposure after repeated dosing, whereas physiologically based pharmacokinetic modelling and simulations19-21 may help to further explore and validate the critical biopharmaceutical factors throughout the preclinical and clinical stages. CONCLUSION In summary, preclinical Frel,susp/sol and Fabs,sol data and clinical Do estimates provide complimentary information with respect to the formulation type selection or dosing condition, however may not fully distinguish in all cases between conventional (e.g. to be taken with food), solubility (and permeability) enhancing or enabling formulations. Therefore, it is useful to evaluate the human bioavailability using a conventional and solubility (and/or permeability) enhancing formulation under fasting and fed condition in first-in-human studies for compounds with high Do, low animal Frel,susp/sol or low Fabs,sol caused by precipitation of the solubilized API to avoid overinvestment in complex and expensive enabling technologies. AUTHOR INFORMATION Corresponding Authors: Sandy Van Hemelryck *E-mail: [email protected], Phone: +32/(0)14/60.78.31 Notes: The authors declare no competing financial interest. ACKNOWLEDGMENTS The authors are grateful to Lieve Adriaenssen and Ludo Quirijnen (Drug Metabolism and Pharmacokinetics, Janssen, Belgium), Dominique Swerts (Clinical Pharmacology and Pharmacometrics, Janssen, Belgium) for their assistance in building the drug data base, Geert Pille for generating the necessary solubility data and to Prof. Dr. Peter Langguth

19 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 20 of 36

20 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

(Pharmaceutical Technology and Biopharmaceutics, University of Mainz, Germany) for scientific discussion. This work has been performed, without financial support, as part of three tasks in the Innovative Medicines Initiative Joint Undertaking (http:// www.imi.europa.eu). SUPPORTING INFORMATION 1. Supporting

information

1:

Physico-chemical

characteristics

and

Preclinical

bioavailability of 89 compounds in Part 1 2. Supporting information 2: Physico-chemical characteristics of compounds 1-22 in Part 2 and 3 3. Supporting information 3: Preclinical bioavailability, Do and Clinical formulation type of compounds 1-12 in Part 2 4. Supporting information 4: Preclinical and Clinical bioavailability, Do and Clinical formulation type of compounds 10-22 in Part 3 REFERENCES 1. Margolskee, A.; Darwich, A.S.; Pepin, X.; et al. IMI – oral biopharmaceutics tools project – evaluation of bottom-up PBPK prediction success part 1: Characterisation of the OrBiTo database of compounds. Eur. J. Pharm. Sci. 2017, 96, 598-609. 2. Kuentz, M.; Holm, R.; Elder, D.P. Methodology of oral formulation selection in the pharmaceutical industry. Eur. J. Pharmaceutical Sciences 2016, 87, 136-163. 3. Lennernäs, H.; Aarons, L.; Augustijns, P.; Beato, S.; Bolger, L.; Box, K.; Brewster, M.; Butler, J.; Dressman, J.; Holm, R.; Julia Frank, K.; Kendall, R.; Langguth, P.; et al. Oral biopharmaceutics tools – time for a new initiative – an introduction to the IMI project OrBiTo. Eur. J. Pharm. Sci. 2014, 57, 292-299. 4. Mackie, C., Mortishire-Smith, R., Wuyts, K., Brewster, M. Assessing Solid Dosage Form Feasibility and Solubility/Dissolution Rate Limitations for Drug Candidates – Early Solution-Suspension Comparisons in the Rat. Presented at the American Association of Pharmaceutical Scientists (AAPS) Annual Meeting and Exposition, Atlanta, GA, November 16-20, 2008. 5. Mackie, C.; Austin, N.; Wuyts, K.; DuJardin, M.; Vanhoutte, F.; Stokbroekx, S.; Verreck, G.; Brewster, M. Solution-suspension testing in the rat: avoiding overinvestment in superficially difficult-to-formulate drug candidates. Presented at the 20 ACS Paragon Plus Environment

Page 21 of 36

Molecular Pharmaceutics

21 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract American Association of Pharmaceutical Scientists (AAPS) Annual Meeting and Exposition, Chicago, IL, (include dates if available) 2012. 6. Brewster, M.; Mackie, C.; Van Peer, A. Preclinical in vivo evaluation strategies for the evaluation of new compounds and formulations. In Vivo Preclinical Relative BA Study for Early Selection of Decisions on Drug Product Development Strategy. Presented at the Uppsala EUFEPS 5th World Congress on Drug Absorption, Transport, and Delivery, Uppsala, Sweden, June 24-26, 2013. 7. FDA Guidance for Industry. Estimating the maximum safe starting dose in initial clinical trials for therapeutics in adult healthy volunteers. July 2005. 8. Dressman, J. B., Amidon, G. L., Reppas, C. & Shah, V. P. Dissolution testing as a prognostic tool for oral drug absorption: immediate release dosage forms. Pharm. Res. 1998, 15, 11–22. 9. Oh, D.M.; Curl, R.; Amidon, G. Estimating the fraction dose absorbed from suspensions of poorly soluble compounds in humans: a mathematical model. Pharm. Res. 1993, 10, 264-270. 10. FDA Guidance for Industry. Food-Effect, Bioavailability and Fed Bioequivalence Studies. Dec 2002. 11. Chaparian, E.; Tang, L.; XU, G.; Huang, T.; Jin, L. Cell Based Experimental Models as Tools for the Prediction of Human Intestinal Absorption. Conference proceeding. 15th North American Regional International Society for the Study of Xenobiotics Meeting, 2008. 12. Bergström, C.A.S.; Holm, R.; Jorgensen, S.A.; Andersson, S.B.E.; Artursson, P.; Beato, S.; Borde, A.; Box, A.; Brewster, M.; Dressman, J.; Feng, K.-I.; Halbert, G.; Kostewicz, E.; McAllister, M.; Muenster, U.; Thinnes, J.; Taylor, R.; Mullertz, A. Early pharmaceutical profiling to predict oral drug absorption: current status and unmet needs. Eur. J. Pharm. Sci. 2014, 57, 173-199. 13. Sjögren, E.; Abrahamsson, B.; Augustijns, P.; Becker, D.; Bolger, M.B.; Brewster, M.; Brouwers, J.; Flanagan, T.; Harwood, M.; Heinen, C.; Holm, R.; Juretschke, H.P.; Kubbinga, M.; Lindahl, A.; Lukacova, V.; Münster, U.; Neuhoff, S.; Nguyen, M.A.; Van Peer, A.; Reppas, C.; Hodjegan, A.R.; Tannergren, C.; Weitschies, W.; Wilson, C.; Zane, P.; Lennernäs, H.; Langguth, P. In vivo methods for drug absorption – comparative physiologies, model selection, correlations with in vitro methods (IVIVC), and applications for formulation/API/excipient characterization including food effects. Eur. J. Pharm Sci. 2014, 57, 99-151. 21 ACS Paragon Plus Environment

Molecular Pharmaceutics

Page 22 of 36

22 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

14. Davit, B.M.; Kanfer, I.; Tsang, Y.C. BCS Biowaivers: Similarities and Differences Among EMA, FDA, and WHO Requirements. The AAPS J. 2016, 18, 612-618. 15. Zane, P.; Guo, Z.; MacGerorge, D.; Vicat, P., Ollier, C. Use of pentagastrin dog model to explore food effects on formulations in early drug development. Eur. J. Pharm. Sci. 2014, 57, 207-2013. 16. Thummela, K.E.; Kunzeab, K.L.; Shena, D.D. Enzyme-catalyzed processes of firstpass hepatic and intestinal drug extraction. Advanced Drug Delivery Reviews 1997, 27, 99-127 17. Darwich, A.S.; Neuhoff, S.; Jamei, M.; Rostami-Hodjegan, A. Interplay of Metabolism and Transport in Determining Oral Drug Absorption and Gut Wall Metabolism: A Simulation Assessment Using the “Advanced Dissolution, Absorption, Metabolism (ADAM)” Model. Current Drug Metabolism, 2010, 11, 716-729. 18. Parr, A.; Hidalgo, J.J.; Bode, C.; Brown, W.; Yazdanian, M.; Gonzalez, M.A.; Sagawa, K.; Miller, K.; Jiang, W.; Stippler, E.S. The effect of excipients on the permeability of BCS class III compounds and implications for biowaivers. Pharm. Res. 2016, 33, 167-176. 19. Kubbinga, M.; Nguyen, M.A.; Staubach, P.; Teerenstra, S.; Langguth, P. The influence of chitosan on the oral bioavailability of acyclovir – a comparative bioavailability study in humans. Pharm. Res. 2015, 2241-2249.

22 ACS Paragon Plus Environment

Page 23 of 36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Figure Legends Abstract figure. Proposed formulation decision tree based on animal bioavailability and Do Figure 1. Physicochemical characteristics for the compounds used in Part 1. Distribution of pKa (of strongest acid and strongest basic functions), logP, Peff (.10-4 cm/s), solubility in Fassif (mg/mL) and Fessif (mg/mL). Figure 2. Relative (Frel(susp/sol)) versus absolute (Fabs,sol) bioavailability in the rat and dog for 98 Janssen compounds (Part 1). Markers colored by in vitro solubility in Fessif (A,B,C) or Fassif (D,E,F) (red: 0.1 mg/mL; white: no experimental solubility). The size of the marker reflects the size of the human effective permeability in silico (A,D), the human effective permeability calculated from in vitro apparent permeability (B,E), and the degree of calculated hepatic first pass (C,F). Red, green and yellow shaded areas indicate the criteria for which BCS class I, BCS class II/IV and BCS class III-like classification would respectively be proposed based on the decision tree of Mackie et al [ref]. Figure 3. Relative bioavailability (Frel(susp/sol)) in the rat versus the dog (Part 1). Markers are colored by the prandial state of the dog (pink: fasted; blue: fed). The size of the marker reflects size of the ratio of percentage ionized in stomach in rat to percentage ionized in stomach in dog Figure 4. Relative (Frel,susp/sol) versus absolute bioavailability (Fabs,sol) in the dog and in the rat for 12 Janssen compounds (Part 2). Labels indicate the compound number. Marker color and shape represent clinical formulation type, and solid or liquid, respectively. Red, green and yellow shaded areas indicate the criteria for which solubility enabling/enhancing, conventional and permeability enabling formulations would respectively be proposed based on the decision tree of Mackie et al4-6. Figure 5. Relative (Frel,susp/sol) versus dose number (Do) in the dog and in the rat for 12 Janssen compounds (Part 2). Panels A, B and C include Do in aqueous, FaSSIF and FeSSIF medium. Labels indicate the compound number. Marker color and shape represent clinical formulation type and solid or liquid, respectively. Figure 6. AUCsolid/sol in human for 13 Janssen compounds (Part 3). Panels A, B and C for aqueous, FaSSIF and FeSSIF Do respectively. Marker color, shape and size represent clinical formulation type, prandial state and Do, respectively. Figure 7. Relative bioavailability of various formulation types of compound 19 administered to humans in fasted or fed conditions (Part 3).

ACS Paragon Plus Environment

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Marker color and shape represent clinical formulation type and prandial state, respectively. Figure 8. Relative bioavailability of various formulation types of compound 20 studied in the dog and in humans (Part 3). Marker color and shape represent clinical formulation type and prandial state, respectively. Figure 9. Relative bioavailability of various formulation types of compound 22 studied in the dog and in humans (Part 3). Marker color and shape represent clinical formulation type and prandial state, respectively. Figure 10. Formulation type selection strategy based on Do

ACS Paragon Plus Environment

Page 24 of 36

Page 25 of 36

FIGURE 1

p K a a c id

pKa base

15

25

20

count

count

10 15

10

5 5

0

0 0 -1

1 -2

2 -3

3 -4

4 -5

5 -6

6 -7

7 -8

8 -9

9 -1 0 1 0 -1 1 1 1 -1 2

0 -1

1 -2

2 -3

3 -4

lo g P

4 -5

5 -6

6 -7

7 -8

8 -9

9 -1 0 1 0 -1 1 1 1 -1 2

P e ff

40

30

30

count

count

20 20

10 10

0

0 0 -1

1 -2

2 -3

3 -4

4 -5

5 -6

0 -1

1 -2

F a s s if

2 -3

3 -4

4 -5

5 -6

F e s s if

40

40

30

30

count

count

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Molecular Pharmaceutics

20

10

20

10

0

0 1 -1 0

0 .1 - 1

0 .0 1 - 0 .1

0 .0 0 1 - 0 .0 1

0 .0 0 0 1 - 0 .0 0 1

1 -1 0

ACS Paragon Plus Environment

0 .1 - 1

0 .0 1 - 0 .1

0 .0 0 1 - 0 .0 1

0 .0 0 0 1 - 0 .0 0 1

Molecular Pharmaceutics

Frel,susp/sol

Frel,susp/sol

FIGURE 2

Fabs, sol

Frel,susp/sol

Frel,susp/sol

Fabs, sol

Fabs, sol

Fabs, sol

Frel,susp/sol

Frel,susp/sol

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Page 26 of 36

Fabs, sol

Fabs, sol

ACS Paragon Plus Environment

Page 27 of 36

FIGURE 3

Frel,susp/sol rat

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Molecular Pharmaceutics

Frel,susp/sol dog

ACS Paragon Plus Environment

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

FIGURE 4

ACS Paragon Plus Environment

Page 28 of 36

Page 29 of 36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Molecular Pharmaceutics

FIGURE 5

A

B

C

ACS Paragon Plus Environment

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

FIGURE 6

A

B

C

ACS Paragon Plus Environment

Page 30 of 36

Page 31 of 36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Molecular Pharmaceutics

FIGURE 7

ACS Paragon Plus Environment

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

FIGURE 8

ACS Paragon Plus Environment

Page 32 of 36

Page 33 of 36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

Molecular Pharmaceutics

FIGURE 9

ACS Paragon Plus Environment

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

FIGURE 10

ACS Paragon Plus Environment

Page 34 of 36

Page 35 of 36 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Molecular Pharmaceutics

Tables Table 1. Overview of database and evaluations performed in Part 1, 2 and 3 Part Compounds Parameters 1 89 • Physchem API (logP, pKa, solubility, Peff) • Preclinical Frel,susp/sol and Fabs,sol 2 12 (1-12) • Physchem API (logP, pKa, solubility, Peff) • Preclinical Frel,susp/sol and Fabs,sol • Clinical formulation type • Do 3 13 (10-22) • Physchem API (logP, pKa, solubility, Peff) • Preclinical and Human AUCsolid/sol • Clinical formulation type • Do Peff: human effective permeability; Do: dose

• • • •

Evaluations Link between physchem API and preclinical Frel,susp/sol and Fabs,sol Link between BCS-like classification and preclinical Frel,susp/sol and Fabs,sol Link between clinical formulation type and preclinical Frel,susp/sol and Fabs,sol Link between clinical formulation type and Do

• Link between clinical formulation type, Do and human AUCsolid/sol • Examples of formulation selection strategies

number; Frel,susp/sol: relative bioavailability of

suspension over solution; Fabs,sol: absolute bioavailablity of solution; AUCsolid/sol: area under the plasma concentration-time curve of solid formulation over solution; BCS: Biopharmaceutical Classification System

ACS Paragon Plus Environment

Molecular Pharmaceutics 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56

ABSTRACT FIGURE: For Table of Contents Use Only A Preclinical Bioavailability Strategy for Decisions on Clinical Drug Formulation Development: an In Depth Analysis An Van den Bergh1, Sandy Van Hemelryck2, Jan Bevernage3, Achiel Van Peer2, Marcus Brewster3, Claire Mackie4, Erik Mannaert2

ACS Paragon Plus Environment

Page 36 of 36