Understanding dermal drug disposition using TCAT - a novel PBPK model

Transcription

1 Panellists: Anu Krishnatry Valeriu Damian Jessica Spires Understanding dermal drug disposition using TCAT - a novel PBPK model SOT-BMSS Feb 16 th, 2016

2 What is PBPK? It is a modeling technique not a single model Wikipedia: PBPK modeling is mathematical modeling technique for predicting the absorption, distribution, metabolism and excretion (ADME) of synthetic or natural chemical substances in humans and other animal species. PBPK models strive to be mechanistic by mathematically transcribing anatomical, physiological, physical, and chemical descriptions of the phenomena involved in the complex ADME processes. PBPK is: PBPK is not: PBPK is a modeling technique PBPK is striving to be mechanistic PBPK is attempting to separate physiology from drug related parameters PBPK is applicable to all routes of administration PBPK is increasingly accepted by regulatory PBPK is not a single model Even though a Google image PBPK search reveal only variations of the same basic model PBPK is not just for IV / oral Detailed oral absorption models (e.g. ACAT/ADAM) coupled with simple compartmental PK are also PBPK models PBPK is not just an alternative to allometry PBPK can provide tissue concentration predictions, drug delivery predictions for various dosage routes and much more PBPK is not just for small molecules 2

3 The first documented application of PBPK by FDA was in 1990s for review and approval tretinoin topical cream Peck, C. C. (2010). Quantitative clinical pharmacology is transforming drug regulation. Journal of Pharmacokinetics and Pharmacodynamics, 37(6),

4 Is a simple skin absorption model sufficient? capillaries Complex organ structure and Complex formulation formulation epidermis dermis Dose Flux through skin Dermis Diffusion to systemic circulation Systemic receptor sweat gland Systemic clearance 4

5 Site of action The TCAT model allows us to predict drug concentration in each skin sublayer Anissimov, Y. G., et al. "Mathematical and pharmacokinetic modelling of epidermal and dermal 5 transport processes." Advanced Drug Delivery Reviews 65.2 (2013):

6 ml/min/100g skin Not all skin is the same SC Thickness (µm) VE Thickness (µm) DE Thickness (µm) Leg Abdomen Back Arm Scalp Face microns Leg Abdomen Back Arm Scalp Face microns Leg Abdomen Back Arm Scalp Face microns Blood Flow Rate Face Scalp Arm Back Abdomen Leg 6

7 Cumulative amount In vitro testing - cumulative amount over 24 h Dissolution, supersaturation and precipitation are very important events Compound A: Expected behavior Compound B: Unexpectedly low penetration at high dose Precipitation at higher dose Dissolution, supersaturation and precipitation are very important events Solution 1.0% ~ 0.5% Solution 0.5% Solution 0.45% Solution 0.045% Solution 2.2% 2.2% ~ 0.05% Solution 0.05% Time (hours) Time (hours) 7

8 FDA Classification of Transdermal Formulations clear & homogenous SOLUTION LOTION SUSPENSION Emulsion Liquid formulations solid dispersed in a liquid Transdermal Formulations Semi-solid formulations > 50% water and volatiles Emulsion < 50% water and volatiles GEL CREAM OINTMENT PASTE Solution of colloidal dispersion with a gelling agent < 20-50% dispersed solids AND < 50% of solids are waxes, PEGs, HCs OR > 20% water and volatiles < 20-50% dispersed solids >50% of solids are waxes, PEGs, HCs < 20% water and volatiles > 20-50% dispersed solids 8

9 Vehicle for CP Formulations from a modeling perspective SOLUTION LOTION SUSPENSION CP: Liquid vehicle DP: missing CP: Liquid vehicle DP: (oil) Droplet CP: Liquid vehicle DP: Particle Liquid Continuous phase (CP) Disperse phase (DP) Gel DP DP DP Semisolid Drug particle Oil droplet Drug encapsulated CP: Gel vehicle (slower diffusion) CP: semisolid vehicle (slower diffusion) CP: semisolid vehicle (slower diffusion) CP: semisolid vehicle (slower diffusion) DP: missing DP: droplet DP: droplet/particle DP: particle GEL CREAM OINTMENT PASTE 9

10 Percent of initial weight A simple solution formulation Volatile Fast Evaporation (e.g. alcohol) Fast evaporation excipients disappear quickly (few min) Sample 1 Sample 2 They may leave a super saturated solution Sample 3 Sample 4 Huge concentration gradient can drive drug into skin but can also cause precipitation Time (seconds) Acetone 100% Ethanol Volatile Slower Evaporation (e.g. water) Water would evaporate in about 15 min (perhaps longer in a gel formulation) There will be a residual level of nonvolatile oils that would eventually be absorbed in the skin Non-volatile (e.g. oils) These stages are modeled explicitly 10

11 Concentration in Vehicle A likely concentration profile in vehicle Fast evaporation excipient is completely gone Huge concentration increase as the slow evaporating excipient is almost vanished. Precipitation brings concentration back close to solubility in nonvolatile excipient Dissolution rate equals absorption rate Drug concentration in formulation Drug solubility in slow evaporating excipient Drug solubility in non volatile excipient time Volatile Slower Evaporation (e.g. water) For many drugs the absorption window can be quite small (e.g. 15min) No re-dissolution of precipitated drug

12 Dermis Viable epidermis Stratum corneum Vehicle Transdermal Drug Disposition: Key processes Evaporation Release/ Dissolution Precipitation / Redissolution Partitioning Binding Absorption Binding Metabolism Binding Systemic Uptake 12 Thanks to Siladitya Ray Chaudhuri and Simulations Plus

13 SC perm Modeling dermal formulations vdp Evaporation Disperse phase (DP) vcp Evaporation Coalesce of disperse phase Continuous phase (CP) Drug Diffusion in vcp Drug and vdp absorption into SC Film phase (FP) Drug and vcp absorption into SC Surface area Surface area Surface area Drug, vcp and vdp diffusion and partitioning Stratum Corneum of skin of follicles Hair Lipid outside of hair Hair core outside Rest of skin Drug diffusion and partitioning Viable epidermis Sebum Hair lipid Hair core Systemic Drug absorption Dermis 13

14 SC perm Drug Dependent Input Parameters in SC Diffusion and partition vdp Evaporation Coalesce of disperse phase Disperse phase (DP) Continuous phase (CP) vcp Evaporation Drug Diffusion in vcp Drug and vdp absorption into SC Film phase (FP) Drug and vcp absorption into SC Surface area Surface area Surface area Drug, vcp and vdp diffusion and partitioning Stratum Cornea of skin of follicles Hair Lipid outside of hair Hair core outside Rest of skin Drug diffusion and partitioning Viable epidermis Sebum Hair lipid Hair core Systemic Drug absorption Dermis Wang-Kasting-Nitsche Potts-Guy D K SC SC/ w 10 log(0.05) MW logP Wang, T., Kasting, G. B., & Nitsche, J. M. (2007). A multiphase microscopic diffusion model for stratum corneum permeability. II. Estimation of physicochemical parameters, and application to a large permeability database. Journal of Pharmaceutical Sciences, 96(11), doi: /jps Robinson P SC lip pol logp MW 0. 5 cm h P P MW K D SC/ w SC 1 P K h SC SC SC/ w 14

15 SC perm Drug Dependent Parameters in Epidermis and Dermis Diffusion and partition vdp Evaporation Coalesce of disperse phase Disperse phase (DP) Continuous phase (CP) vcp Evaporation Drug Diffusion in vcp Sebum Rest of skin Systemic Drug and vdp absorption into SC Drug, vcp and vdp diffusion and partitioning Drug diffusion and partitioning Drug absorption Film phase (FP) Surface area Stratum Corneum Viable epidermis Dermis of skin Drug and vcp absorption into SC Surface area of follicles Sebum Hair Lipid outside Hair lipid Surface area of hair Hair core outside Hair core Valiveti log K log P D SBM SBM / w SBM ( m / s) log P MW 0.008P K log P SBM SBM / w 0.28 Viable epidermis and Dermis Kretsos log D K VE / w VE cm f 2 aq non / s ; log MW aq 0.7 K K VE / w VE / SC K SC/ w Bunge-Cleek log D K VE / w VE cm 1 2 / s log MW Robinson K D VE / w log P VE VE 1 P K cm / s h VE VE VE / SC log MW 15

16 SC perm Systemic absorption rate in Dermis Diffusion and partition vdp Evaporation Coalesce of disperse phase Disperse phase (DP) Continuous phase (CP) vcp Evaporation Drug Diffusion in vcp First order SysRate j k sys j C DE j C sys free Rest of skin Drug and vdp absorption into SC Drug, vcp and vdp diffusion and partitioning Drug diffusion and partitioning Film phase (FP) Surface area Stratum Corneum Viable epidermis of skin Drug and vcp absorption into SC Surface area of follicles Hair Lipid outside Surface area of hair Hair core outside Sebum Hair lipid Hair core IBRAHIM 1 k sys j sys Cu VolRate 1 ~ Q Systemic Drug absorption gm cm s Dermis cap cap DE sys VolRate j P SA C j, u Cu P UWL 3 D h w UWL J cm 1 cm s 3 cm J Vs 1 s ml 1 s cm VL 2 3 C C sys sys u gm cm C 1 e C DE DE j, u Pe u j, u 1 L PeL cap mem UWL P P P mem log P 1.64 log P 1.37MW e 2.82 s e e Pe Pe L s J J Vs VL L tot s P L tot L P 1 s L n n P s slit i i L s i L S P W W i S P L s 3 f W L D 1 slit Di Dw rhyd i W 2 s A A slit i 9 16 L s L W W L s 2 JVs 1 s Pes P S Pe L slits JVL 1 L P S slitl i x 3 i ln i i i i i Ibrahim, R., Nitsche, J. M., & Kasting, G. B. (2012). Dermal clearance model for epidermal bioavailability calculations. Journal of Pharmaceutical Sciences, 101(6),

17 Permeability and Diffusion in Stratum Corneum As a function of MW Wang-Kasting-Nitsche model 1 for partially hydrated skin Wang-Kasting-Nitsche model 2 for partially hydrated skin Potts&Guy Wang-Kasting-Nitsche model 1 for fully hydrated skin Wang-Kasting-Nitsche model 2 for fully hydrated skin Permeability cm/sec (logp=2.0) Model 1 Model 2 WKN_fhm1 WKN_fhm2 WKN_phm1 WKN_phm2 PottsGuy Diff Coeff cm2/sec (logp=2.0) minutes hours days MW MW 17

18 Permeability and Diffusion in Stratum Corneum As a function of Kow Wang-Kasting-Nitsche model 1 for partially hydrated skin Wang-Kasting-Nitsche model 2 for partially hydrated skin Potts&Guy 10-3 MW=250, Perm cm/sec 10-8 MW=250, Diff Coeff cm2/sec Wang-Kasting-Nitsche model 1 for fully hydrated skin Wang-Kasting-Nitsche model 2 for fully hydrated skin 10 5 MW=250, KP minutes hours days kow kow kow 18

19 KP tissue/water percent drug in each region of skin Partition - skin capacity to absorb drug Assuming no metabolism, no excipient effects, no protein/melanin binding Partially hydrated SCph SCfh VE DE SB All ph All fh SC VE DE SB logp logp Most drugs can partition into sebum (SB) however sebum volume is low For small logp Dermis would have the capacity to hold most drug Extremely lipophilic compounds would partition preferentially in SC Sebum partitioning is most significant for compounds with logp around 5 19

20 MW Permeability in various skin layers Assuming no metabolism, no excipient effects, no protein/melanin binding Stratum Corneum Viable Epidermis Low <----- Permeability [cm/s] -----> High All Skin Dermis Sebum Stratum corneum is the key diffusion barrier for most compounds (not surprising) For most compounds of interest Sebum permeability is quite high Metabolism, binding and systemic uptake can make the dermis a more significant barrier logp 20

21 Hair Three dimensional finite element skin model Attempting to model the 3D skin complex structure thanks Sumeet Thete Formulation Formulation Epidermis Dermis Sytratum Corneum Blood capillaries Sebum 993/jpg/bloodvessels_large jpg Capillaries near root of hair follicle 21

22 Sample simulation 3D finite element model (2D axial symmetry) MW = 250, logp = 5, no vehicle partitioning, high systemic absorption 22

23 LogP effect MW = 250, no vehicle partitioning, high systemic absorption logp = 5 Less lipophylic More lipophylic logp = 2 logp = 8 23

24 Making the PBPK dermal model available Additional Dosage Routes Modules in GastroPlus Additional Dosage Routes Modules Transdermal Pulmonary Ocular Oral absorption - ACAT model Input parameters PBPK model Output 24

25 In-vitro to Human - Model Based Translation Principle Scaling Approach Which factors to use? Is it predictive? Does it translate? Human Physiology Trial Protocol In vitro Drug related Parameter In-vitro to in-vivo scaling factors PBPK model Human tissue PK predictions Predictions not accurate? Change the in-vitro conditions to be more relevant to the in-vivo situation In-vitro Protocol In-vitro Physiology Human Physiology Trial Protocol Model Based Translation In-vitro data Use same PBPK model for both in-vitro and in-vivo PBPK model Using more fundamental drug parameters (e.g. diffusion, lipid bilayer permeability) Human tissue PK predictions Change model to capture more biological processes and use more fundamental drug parameters Predictions not accurate? 25

26 In vitro physiology Using PBPK model to simulate in-vitro skin penetration TCAT model TCAT model Select Human Abdomen skin physiology Adjust Dermis thickness to match skin sample thickness (~450 microns) Remove sebum (set sebum partition to zero) Set high systemic absorption rate PK parameters Match central compartment to receptor chamber No clearance (all compound accumulates in receptor fluid) Franz static cells Bronaugh flow-through cells Match volume of distribution with receptor chamber volume Set Fu plasma to 100% or according to receiving fluid used Match volume of distribution with receptor output volume Set Fu plasma close to zero to match the sink conditions achieved in receptor chamber Instead of using in-vitro flux in PBPK Validate diffusion and partition parameters using in-vitro data and use them in PBPK 26

27 Dermis Viable epidermis Stratum corneum Vehicle Excipient effects Evaporation Shed / Rub/Wash off DRUG Excipients Partitioning Release/ Dissolution Absorption Diffusion Precipitation / Redissolution Partitioning Binding Modified partitioning Absorption Diffusion Partitioning Penetration enhancement Metabolism Binding Metabolism Systemic Uptake Systemic Uptake 27

28 Ketoprofen Propylene Glycol Penetration of co-solvent with drug in the tissue Slower penetration of drug compared to solvent Saar, B. G., Contreras-Rojas, L. R., Xie, X. S., & Guy, R. H. (2011). Imaging drug delivery to skin with stimulated Raman scattering microscopy. Molecular Pharmaceutics 28

29 Topical Formulation Components How can we represent this in the model Soybean oil Medium Chain Triglycerides Light Mineral Oil Oleoyl Macrogolglycerides Xanthan Gum Glyceryl Behenate Hydroxypropyl Cellulose Microcrystalline Wax Wax, Beeswax Mineral Oil Glycerin (glycerol) Used often High amount Polyethylene Glycol 400 Propylene Carbonate Propylene Glycol Used sometimes High amount Reduce diffusion in vehicle Some may affect partition in VH Significant change in partition in both vehicle VH and Stratum Corneum SC Benzyl Alcohol Stearyl Alcohol Phenoxyethanol Potassium sorbate Sorbic Acid Used often High amount Change solubility and partition in VH SC Used often Med amount Evaporate quickly Remove from dosage volume before modeling Lipid vehicle Solvent Thickener Complex formulation Many components Different functions Used often High amount Use hydrated SC model Lower evaporation Slow diffusion in VH Petrolatum Preservative Humectant Used often High amount Change Vehicle (VH) Solubility, Partition and Evaporation Buffer Penetration enhancer Used often High amount (Transient) Increased diffusion in SC (Propylene Glycol) (Transcutol) Sodium Citrate Used often Low amount Minimal impact on penetration Emulsifier Emollient Solubilizer Polysorbate 80 Used often Low amount Use hydrated SC model Reduce evaporation rate Used often High amount Significant change in vehicle solubility Surfactant Evaporation can trigger Used often on precipitation Low amount Can change coalescence and film formation Glyceryl in emulsions Monostearate II Glycerol Monostearate Sorbitan Oleate Cyclomethicone Isopropyl Myristate Octyldodecanol Oleic Acid Stearic Acid Caprylocaproyl Polyoxylglycerides Diethylene Glycol Monoethyl Ether Diisopropyl Adipate Dimethyl sosorbide 29

30 Novel Skin PBPK model in Action: Clindamycin Modeling Anu Shilpa Krishnatry

31 Clindamycin Modeling with TCAT Module Clindamycin is lincosamide antiobiotic Is usually used to treat infections with anaerobic bacteria Common topical treatment for acne Typical IV administration seems to be as clindamycin phosphate - phosphate is quickly converted to clindamycin in the plasma J Med Assoc Thai, 89(5): ,

32 Available Clinical Data Gatti et al., Antimicrob Agents Chemother, 37(5): , 1993 IV infusion: 600mg of clindamycin phosphate over 25 min PO capsule: 600mg of clindamycin hydrochloride (with 170 ml of water) Healthy males, average age 27 years, average body weight 73 kg Plaisance et al., Antimicrob Agents Chemother, 33(5): , mg of clindamycin phosphate infused over 30 min every 6 h 1200mg of clindamycin phosphate was infused over 1h every 12h Healthy males, average age 29 years, average body weight 73 kg Dalacin T solution: 5 mg single application Duac Gel: 10 mg single application Clindagel: 40 mg QD repeat application Evoclin foam: 40 mg QD repeat application Test if TCAT model can predict exposure differences with different formulations

33 What All information needed for the TCAT model? formulation epidermis dermis capillaries receptor sweat gland 33

34 What All information needed to build a TCAT model? Compound details Log P, pka Solubility at various ph, biorelavant media, solvent Permeability Protein binding and skin tissue binding Blood cell association Clearance Formulation details Type (Solution, suspension, gel, ointment..) Volatile components Volume Dose Study Details Dosing region (face, arms, scalp, back, abdomen, legs) Application surface area Skin covered/not Any Washing step performed Any in vitro flux data 34

35 Building Clindamycin TCAT model Compound details Log P, pka Solubility at various ph, biorelavant media, solvent Permeability Protein binding and skin tissue binding Blood cell association Clearance Predicted / Experimental values were used for model building. Systemic disposition was calibrated against plasma concentration-time profiles after intravenous administration from literature Formulation details Type (Solution, suspension, gel, ointment..) Volatile components, solvent evaporation rate Volume Dose API solubility in Formulation Study Details Dosing region (face, arms, scalp, back, abdomen, legs) Application surface area and Time Skin covered/not Any Washing step performed Any in vitro flux data 35

36 Building Clindamycin TCAT model Diffusivity and partition coefficients calculations: In Stratum Corneum using Wang-Kasting Nitsche equation In Viable epidermis using Krestos equation In Sebum calculated according to the equation derived from Valiveti et al., data Sublayers (10, Uniform Grid) 36 Wang et al., J Pharm Sci, 95(3), 2006; Krestos et al., J Pharm Sci, 93(11), 2004; Valiveti S, Wesley J and Lu GW, International J of Pharmaceutics, 346, 2008.

37 Building Clindamycin TCAT model Optimized only 2 parameters after accounting for all compound, formulation and study specific details: Vehicle /Water partition coefficient and systemic uptake rate 37

38 Building Clindamycin TCAT model PK Characterization using Literature Data Published Human data available Gatti et al, Antimicrob Agents Chemother, 37(5): , 1993 IV infusion: 600mg of clindamycin phosphate over 25 min PO capsule: 600mg of clindamycin hydrochloride (with 170 ml of water) Healthy males, average age 27 years, average body weight 73 kg Plaisance et al, Antimicrob Agents Chemother, 33(5): , mg of clindamycin phosphate infused over 30 min every 6 h 1200mg of clindamycin phosphate was infused over 1h every 12h Healthy males, average age 29 years, average body weight 73 kg Data Captured using Digit It (Simulations Plus) 38

39 Building Clindamycin TCAT model PK Characterization using Literature Data Compartmental PK model fitted to match observed plasma concentration vs. time profile after 600 mg IV infusion for 25 mins of Clindamycin phosphate as in Gatti et al., 600 mg and 1200mg repeat IV infusions from Plaisance et al. Antimicrob Agents Chemother, 37(5), 1993; Antimicrob Agents Chemother, 33(5),

40 Simulating Systemic Exposures from Different Formulations In all cases, clindamycin phosphate is dosed and clindamycin is measured Single dose administration Duac gel Dalacin T solution Multiple dose administrations Evoclin foam Clindagel

41 Simulating Systemic Exposures from Different Formulations Simulated Plasma concentration vs. time profile (solid line) and observed plasma concentrations (open squares with CV% as the bars, N=12)

42 Understanding exposures in various skin compartments 42

43 Dissolved Concentration (ug/ml) Dissolved Concentration (ug/ml) Dissolved concentration Simulating exposures in various skin compartments Simulated concentration of Clindamycin in various skin compartments (VH: Vehicle, SC: Startum Corneum; VE: Viable Epidermis and DE: Dermis) at 0, 1, 4, 8, and 24 hours following application of Dalacin T Solution Dissolved concenteration in Dermis Dalacin T Solution Duac Gel Clindagel Evoclin Foam Dissolved concenteration in Sebum Dalacin T Solution Duac Gel Clindagel Evoclin Foam Simulation Time (hour) Simulation Time (hour) Dissolved concentration of Clindamycin in dermis from different formulations of Clindamycin Dissolved concentration of Clindamycin in Sebum from different formulations of Clindamycin 43

44 Clindamycin gets into sebum? Micro-autoradiography to visualize the localization of drug substance in the skin of mice Micro-autoradiographs of H 3 -Clindamycin in mouse skin after 1 hour (left) or 4 hours (right) of administration. Silver grains identify location of radioactive drug in skin As early as 1 hour after application of CT Gel, clindamycin phosphate was diffusely distributed throughout the dermis with a clear accumulation of drug in and around the hair follicles, and some accumulation around the sebaceous glands. 4 hours after application of CT Gel, significantly more clindamycin phosphate was distributed diffusely throughout the dermis with a clear accumulation near hair follicles, and sebaceous glands 44 Jon Lenn, Stiefel Poster at 2010 American Academy of Dermatology

45 Clindamycin gets into sebum? Clindamycin levels in various body tissues and fluids, Journal of Clinical Pharmacology, July

46 Novel Skin PBPK Model in Action: Clindamycin and Tazarotene Modeling AAPS Annual Meeting, San Diego,

47 Conclusions Clindamycin model was able to appropriately characterize exposure differences with different formulations including a solution, gels, and foam. The TCAT model within GastroPlus is a novel model that will enable scientists to simulate the exposures from different topical formulations. Its applications span from candidate compound and formulation decisions to maximizing the potential for the right amount of drug to be delivered to the target tissues in the body. 47

48 48 Acknowledgements Christopher Ruark, Society of Toxicology Richard Lloyd Tom Wilde Betty Hussey Grace Kang Sumeet Thete (summer student) Collaborators: Viera Lukacova Jessica Spires Siladitya Ray Chaudhuri