`
`
`
`PAC
`Ernestina Luna
`Rebecca Evans
`Leonardo Allain
`Chris John
`Yanning Chen
`Elikem Gbeddy
`Betsy Powlus
`Melissa Drexel
`Lea Janowicz
`
`Biopharm. Chem.
`Kari Lynn
`
`
`
`
`
`
`FORMULATION DEVELOPMENT MEMORANDUM
`
`Conrad Winters, Robert Reed
`
`Saurabh Palkar and Ernestina Luna
`
`31 AUG 2005
`
`
`To:
`
`From:
`
`Date:
`
`Subject: L-000224715 (MK-0431) Preliminary Market Formulation Development
`Report
`
`
`
`Contributors:
`
`Formulation Design
`Saurabh Palkar
`Jim Ney
`Yun Liu
`Laura Artino
`Parminder Sidhu
`Honggang Zhu
`Tom Gandek
`Patricia Hurter
`
`Solids Development
`Jeff Givand
`Robert Meyer
`Brad Holstine
`Ed Smith
`Michelle Kenning
`Larry Rosen
`
`
`
`
`
`
`
`
`
`PBRS
`Lei Wang
`Leigh
`Shultz
`Tim Rhodes
`
`
`
`
`
`MCTA
`Dina Zhang
`
`
`
`Compaction Simulator
`Feng Li
`Steve Galen
`
`Summary:
`
`This report describes the design and development of the preliminary market formulation
`for L-000224715 (MK-0431). First, the properties of the bulk drug significant for
`formulation design are discussed. This is followed by a detailed account of the Phase
`IIB/III formulation design that includes selection of the excipients, development of direct
`compression and roller compaction processes and stability of these formulations.
`
`
`CC: Dept. 854, Suhas Shelukar, Leyna Mulholland, ChrisAnne Santangini, Laman Alani,
`Scott Reynolds, Jim Zega, Dominic Ip, Bob Reed, Dave Storey, Sam Mclintock, David
`Toledo, Doug Mendenhall
`
`1 of 88
`
`Merck Exhibit 2123, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`Table of Contents
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Page
`
`1.0 Introduction
`
`2.0 Significant bulk drug properties for formulation design
`2.1 Processing attributes of the monohydrate form
`
`2.2 Comparison of anhydrous and monohydrate forms
`2.3 Excipient
`
` compatibility
` (anhydrous
`
`API)
`
`
`
`
`3.0 Direct
`
` Development
`
` process
`
`(DC)
`
` compression
`
`3.1 DC process development based on the anhydrous API
`
`3.1.1 Mini-formulation
`
` Development
`
`
`
`3.1.2 Small
`
`
`
`scale
` experiments
`
`
`
`3.1.3 Selection of the excipient grades for the Phase IIB
`formulation
`
`3.1.4 Optimization of disintegrant/lubricant
`levels
`
`3.1.5 Clinical Manufacture to support Phase II trials
`3.1.6 Sticking
`issues
`
`
` during
` compression
`
`
`
`3.1.7 Small Scale Sticking Test Development
`
`
`3.1.8 Remedies to alleviate sticking during compression
`
`3.1.8.1 Use of alternate filler in place of mannitol
`
`3.1.8.2 API
`
` Prelubing
`
`
`
`
`3.1.8.3 Evaluation of glidant, anti-adherent and filler ratio
`3.1.9 Selection of Monohydrate API for Development
`
`3.2 DC process development based on the monohydrate API
`
`3.2.1 Evaluation of a lubricant pair to alleviate sticking
`
`3.2.2 Effect of Precompression during Tableting
`
`
`3.2.3 Effect of API particle size on hardness and sticking
`3.2.4 Re-evaluation
`
`of
` mannitol
`
`
`as
` filler
`
`
`
`
`3.2.5 API
`
`
`stress
` experiments/processing
`
` window
`
`
`3.2.5.1
`Influence of the API PSD changes on the DC
`formulations
`
`
`3.2.5.2 Tool damage during the compression of
`DC/A-Tab batches
`
`3.2.5.3 Segregation evaluation of the DC formulation
`3.2.6 Pre-lubrication
`
`with
` single
`
`
` lubricants
`
`
`
`3.2.6.1
` Magnesium
` stearate
`
`
`
`
`
`
`3.2.6.2
` Sodium
` stearyl
` fumarate
`
`
`
`
`4.0 Roller Compaction (RC) Process Development
`4.1 RC development for the anhydrous API
`
`4.1.1 RC feasibility with mini formulations
`4.1.2 RC small scale (500g) experiments
`
`4.2 Monohydrate
`API
` RC
`
`
`
` development
`
`
`
`
`
`
`
`
`
`
`
`2 of 88
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
` 4
`
`4
`5
`6
` 6
`
` 8
`8
` 8
` 10
`11
`
` 12
`15
` 16
`17
`19
`19
` 20
`21
`24
`27
`27
`28
`29
` 31
` 32
`32
`
`35
`
`36
` 36
` 36
` 38
`
`40
`40
`40
`44
` 46
`
`Merck Exhibit 2123, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`46
`
` 47
` 48
`
`49
` 49
` 50
`50
` 52
` 52
`52
` 55
` 55
` 62
` 64
` 65
`65
` 68
` 68
`68
` 69
`
` 71
`75
`
` 75
`76
`78
` 82
`83
`84
`
`86
`
`87
`
`
`
`
`
`
`
`
`
`
`
`
` Performance
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`6.0 Final
`
`
`
`
`
` selection
`
` formulation/process
`
`6.1 Beta press runs for the lead DC and RC formulations
`
`
`6.2 Evaluation of the effect of Cab-O-Sil on formulation performance
`6.3 Final
` formulation
`
` composition
`
`
`
`
`
`
`6.4 Flow diagram for the final process
`
`
`
`
`
`6.5 Rationale for the final formulation composition and
`
`
` process selection
`
`
`
`
`
`
`4.2.1 Use of Dicalcium phosphate powder and dry binders
`to improve tablet hardness
`
`4.2.2 Processability
`
`at
`
` 5-kg
` scale
`
`4.2.3 Re-evaluation
`
` mannitol
`of
`
`as
`
`
` filler
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`5.0 Stability
`
`
`
`5.1 Degradation
`
`
`
`
` chemistry
`
`5.2 Stress
`
`
` screening
`
`
`
`
` conditions
`for
` rapid
`5.2.1 Storage conditions for accelerated stability studies
`5.2.2 Formaldehyde
`
`stress
` experiment
`
`
`
`5.3 Formulation
`
`
` composition
`
`
`
`5.3.1 API selection: anhydrous vs. monohydrate
`
`5.3.2 Excipient
` selection
`
`
`
`
`
`5.3.2.1 Filler
` evaluation
`
`
`
`5.3.2.2 Lubricant
`
` selection
`
`
`5.3.2.3 Coating
` selection
`
`
`
`5.3.3 Residual
` formaldehyde
`
`
`
`5.4 Manufacturing process RC vs. DC
`5.4.1 Blue
`
`
`color
` investigation
`
`5.5 Long
`
`
`term
` stability
`
`
`
`
`5.5.1 Microenvironment pH and API sensitivity
`5.5.2 RH
`
` sensitivity
`
`
`
`
`5.6 Dissolution
`5.7 Summary
`
`
`
`
`7.0 References
`
`Appendix A: Structure of L-000224715 and Major Degradation Pathways
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`3 of 88
`
`Merck Exhibit 2123, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Introduction:
`
`
`
`
`1.0
`
`L-000224715 is a DPP-IV (dipeptidyl-peptidase IV) inhibitor for the treatment of
`Type 2 diabetes. L-000224715 was approved as a PCC in January 2002 and selected
`as the lead DPP-IV inhibitor for development by the DPIV EDT. The compound was
`assigned MK #0431 after the efficacy was demonstrated in Phase Ib and Phase II
`trials. The Phase III trial for this program was initiated in June 2004.
`
`Program timeline
`
`PCC Approval
`Phase I FPI
`Biocomparison study
`(Capsule vs. tablet)
`Phase IIB FPI
`PCS initiation
`Formulation/process selection
`FSS/biobatch initiation
`Phase III FPI
`Earliest WMA filing
`
`For the phase I clinical trials of this program HPMC capsules filled with the neat drug
`were used. After Phase I a tablet formulation was developed. A Direct compression
`(DC) process was developed for this formulation and roller compaction (RC) was
`evaluated as a back-up.
`
` This report
` describes the experimental
` work
`(formulation/process development and stability analysis) leading to the preliminary
`market formulation composition and the manufacturing process selected for this
`compound.
`
`2.0
`
`The chemical and physical properties of L-000224715 relevant to formulation design
`are briefly described below. (See references 1 and 2 for complete details of the
`chemical and physical properties of this compound)
`
`1) Structure of the parent compound and major degradation products (See Appendix
`A. L-000224715 and Major Degradation Products).
`2) The final drug product formulation is based on the monohydrate form of the API
`(referred to as L-000224715-010X), the phosphate salt of MK-0431 (referred to as
`L-000224715-006F) has four known crystalline anhydrous polymorphs (denoted as
`Form I, Form II, Form III, and Form IV) and various crystalline, non-
`stoichiometric solvates. Form I has a monotropic relationship to Form II and Form
`IV, where Form I is the most thermodynamically stable, anhydrous crystalline
`phase at all temperatures. Form I and Form III have an enantiotropic relationship
`with a transition temperature of 34°C as determined by solubility of the pure
`phases at various temperatures in water. Form I is the thermodynamically stable
`
`Significant Bulk Drug Properties for Formulation Design:
`
`Jan. 2002
`Jul. 2002
`Nov. 2002
`
`May 2003
`Sep. 2003
`Oct. 2003
`Apr. 2004
`Jun. 2004
`Dec. 2005
`
`4 of 88
`
`Merck Exhibit 2123, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`1. Pressure Effect: There is no form conversion upon compression as
`characterized by XRPD and SS-NMR
`2. Processing Effect (Blending): no formulation conversion or particle size
`breakage when blended in V-shell, Turbula or Bohle blenders
`3. Solvent Effect:
`a) From water solution: amorphous form
`b) From a suspension:
`o Organic Solvents: Monohydrate ------> Solvate/Anhydrate
`o Water: Monohydrate ------> No Form Change
`o IPA/Water (95/5): Monohydrate ------> No Form Change
`
`crystalline phase at temperatures above 34°C, and Form III is the
`thermodynamically stable phase at temperatures below 34°C. All anhydrous and
`solvated crystalline phases can be converted to the crystalline monohydrate upon
`slurring in water or solvents with a high water activity.
`3) The equilibrium solubility of the monohydrate form in water was found to be
`68.85 mg/g at 24.5C
`4) The available stability data indicate that the monohydrate form is stable when
`stored at 30ºC/65% RH for 9 months and 40ºC/75% RH conditions for 6 months.
`
`
`2.1 Processing attributes of the monohydrate form:
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`5 of 88
`
`Merck Exhibit 2123, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Comparison of anhydrous and monohydrate forms
`
`
`
`
`2.2
`
`
`
`
`Morpholog
`y/Shape
`
`NB# 66839-113
`(monohydrate)
`
`L-224715-006F024
`(Anhydrous)
`
`Mean
`
`198
`143
`
`
`
`D10
`
`
`
`77
`52
`
`
`
`
`D95
`
`
`422
`329
`
`Mean
`
`83
`77
`
`
`
`
`0.28
`0.55
`
`49%
`
`1.96
`
`32
`
`
`D95
`
`224
`206
`
`
`
`D10
`
`
`
`19
`18
`
`
`0.55
`0.81
`
` 32%
`
` 1.47
`
` 19
`
`
`
`
`
`
`
`Parti
`cle
`Size
`( m)
`
`So
`n
`
`0s
`30
`s
`Density
`3]
`[g/cm
`Loose
`Tapped
`Carr s
`Index
`Hausner
`Ratio
`Flodex
`(mm)
`
`6 of 88
`
` Excipient compatibility (anhydrous API)
`
`
`2.3
`
`To screen excipients for Phase II formulation development, the stability of a series of
`probe formulations was investigated. The details of this work can be found in
`reference 5, and only main observations are presented here.
`
`Probe formulations were prepared using a two-level fractional factorial design based
`on the following six variables:
`1) Dry vs. wet (using IPA as granulation solvent, water was not used due to the high
`aqueous solubility of the bulk drug) as process type
`2) 1, 5 and 10% as drug load (through out this report drug load refers to the wt% of the
`free base equivalent)
`3) HPC vs. PVP as the binder type
`4) Avicel vs. mannitol as the filler type (lactose was not evaluated to avoid Maillard
`reaction with the bulk drug)
`5) Croscarmellose sodium vs. crospovidone as the disintegrant type
`6) Magnesium stearate vs. sodium stearyl fumarate as the lubricant type
`
`
`Merck Exhibit 2123, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`
`The impact of citric acid as an acidulent was included in the study. The chemical
`stability of the probe formulations was analyzed (open and closed dish with and
`without desiccant) at 2, 4 and 8-wks. The results from the analysis of this
`experimental design were as follows:
`
`1) Dry processing and 10%API produced the most stable formulations under both
`open dish and closed-dish conditions without desiccant. There were relatively high
`levels of degradation at 60C/amb, closed-dish and 40C/75%RH, open-dish with
`deamination, hydrolysis, oxidation and formaldehyde adduct formation. No
`degradation was detected in any formulations (except 1%API formulations with the
`wet processing) under closed-dish conditions with desiccant, indicating that acceptable
`chemical stability of the formulations can be achieved by proper selection of
`packaging conditions.
`
`2) Open-dish conditions caused higher levels of degradation, especially at
`40C/75%RH with high level of hydrolysis of the drug. The decrease in %API was
`the main contributor to the increase in the rate of degradation. Wet processing, the
`interaction between the wet process and %API, and the interaction of the wet process
`and Avicel also had significant effects on the rate of degradation of the drug at
`40C/75%RH, open dish condition.
`
`3) Severe degradation was observed in the formulation containing citric acid.
`Therefore, citric acid was eliminated from further consideration as an acidulent.
`
`4) The Pearlitol based formulations showed better chemical stability and retention of
`tablet strength than the Mannogem based formulations under closed-dish conditions
`presumably because of lower sorbitol levels. Hence, Pearlitol was chosen for Phase II
`formulation development.
`
`5) Crospovidone is highly hygroscopic and caused blister formation and hardness
`reduction of the tablets, hence croscarmellose was selected over crospovidone as the
`super disintegrant for the Phase II formulation.
`
`In summary the excipient compatibility study revealed:
`a) A dry process (DC or RC) should be used since the wet granulation with IPA
`resulted in significant degradation. Wet granulation of the anhydrous API with water
`was also found to be unstable later in the program.
`b) Avicel and Pearlitol should be used as fillers
`c) Corscarmellose sodium should be used as a disintegrant.
`d) Magnesium stearate or sodium stearyl fumarate could be used as a lubricant
`(combination of these two lubricants not evaluated in this study)
`
`
`
`
`
` 7
`
` of 88
`
`Merck Exhibit 2123, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`
`3.0 Direct Compression Process Development
`
`3.1 DC process development based on anhydrous API
`
`Since the bulk drug exhibited particle characteristics (flow, density, compactibility)
`that were amenable to direct compression, this process was evaluated first.
`Experiments conducted at different scales to evaluate the effects of drug morphology,
`excipient type and composition on tablet physical properties and content uniformity
`are described below. The roller compaction process was developed as a back-up and
`is described in Section 4.0.
`
`3.1.1 Mini-formulation development
`
`Mini-formulations (80 g batch size) were evaluated to determine the impact of Avicel
`(PH102)/Mannitol (Pearlitol 100) ratio (high 3:1, mid 1:1, low 1:3), drug load, and
`drug lot (lots 16 and 17 since they exhibited different bulk densities, on tablet
`properties (3% croscarmellose sodium and 2% Magnesium stearate used for all
`formulations). The formulations were blended in a one-quart V-blender for 10 min
`after which magnesium stearate was added and lubed for an additional 5 min. The
`formulations were then compressed on the F-press into 100 mg tablets at three
`different forces using 8/32 standard concave round punch.
`
`The hardness data for these experiments is presented in figures 3-1 and 3-2 (note Av
`in figure legends stands for the ratio of Avicel to mannitol). These figures indicate
`that: 1) increasing the amount of Avicel gives higher hardness; 2) hardness decreases
`with increasing drug loading, such that at 40% drug load (lot 17) a 3:1 Avicel to
`mannitol ratio was necessary to obtain sufficient hardness (around 4 kP); and 3) lot 17
`with higher bulk density gave better hardness. These results along with probe stability
`data (for dry formulations Avicel was found to be slightly better than mannitol, Ref.
`#5) and excipient properties (high amounts of Avicel might lead to moisture
`absorption and mechanical failure of tablets on storage) were used to select excipient
`amounts for larger scale beta press runs described in the next section.
`
`
`
`8 of 88
`
`Merck Exhibit 2123, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`
`
`
`
`12
`
`10
`
`8
`
`6
`
`4
`
`2
`
`0
`
`Drug L16 Low Bulk Density
`
`d11 High Av 10% drug
`d12 Low Av 10% drug
`d13 High Av 40% drug
` d19 mid Av, 25% drug
`Linear (d11 High Av 10% drug)
`Linear (d12 Low Av 10% drug)
`Linear ( d19 mid Av, 25% drug)
`Linear (d13 High Av 40% drug)
`
`3
`
`6
`
`9
`C ompression Force ( kN )
`
`12
`
`15
`
`
`Figure 3-1: Tablet hardness as a function of compression force (drug
`lot F016)
`
`
`
`Drug L17 High Bulk Density
`
`3
`
`6
`
`9
`Compression Force (kN)
`
`d15 High Av, 10% drug
`d16 Low Av, 10% drug
`d17 High Av, 40% drug
`d18 Low Av, 40% drug
`d20 mid Av, 25% drug loading
`Linear (d15 High Av, 10% drug)
`Linear (d16 Low Av, 10% drug)
`Linear (d20 mid Av, 25% drug loading)
`Linear (d17 High Av, 40% drug)
`12
`15
`Linear (d18 Low Av, 40% drug)
`
`12
`
`10
`
`8
`
`6
`
`4
`
`2
`
`0
`
`
`
`
`
`Hardness (kP)
`
`Figure 3-2: Tablet hardness as a function of compression force (drug lot F017)
`
`
`
` 9
`
` of 88
`
`Merck Exhibit 2123, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
` Small scale experiments
`
`
`3.1.2
`
`Three small-scale (500g) direct compression experiments were conducted to
`investigate the content uniformity of tablets sampled during a longer run on a rotary
`press. Parameters varied were drug lot, drug loading and mannitol/Avicel ratio. The
`amount of Avicel in these experiments was kept constant at 32%. The experimental
`design is shown in Table 3-1.
`
`Table 3-1: Small scale direct compression experimental design
`
`Formulation# Drug Loading Drug Lot # Mannitol/Avicel Ratio
`D25
`10%
`17
`1.58:1 (32% Avicel)
`D26
`25%
`17
`1:1 (32% Avicel)
`D27
`10%
`16
`1.58:1 (32% Avicel)
`
`
`
`Compression was performed on a 16-station Manesty Beta Press (only 2 stations were
`set-up). Samples were acquired every 5 minutes for content uniformity. The duration
`of each run was 25min. Hardness, disintegration and CU data for these runs is shown
`below.
`
`
`
`Beta Press (large scale - 25min run)
`
`D25
`D26
`D27
`
`02468
`
`Hardness (kP)
`
`0
`
`2
`
`10
`8
`6
`4
`Compression Force (kN)
`
`12
`
`14
`
`
`Figure 3-3: Tablet hardness as a function of compression force
`(small scale DC experiments)
`
`The friability of all formulations was found to be excellent.
`
`
`10 of 88
`
`Merck Exhibit 2123, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`D25
`(10%
`Drug Lot
`17)
`
`D26
`(25%
`Drug
`Lot17)
`
`D27
`(10%
`Drug Lot
`16)
`
`Linear
`(D26
`(25%
`Drug
`Lot17))
`Linear
`(D25
`(10%
`Drug Lot
`17))
`Linear
`(D27
`(10%
`Drug Lot
`16))
`
`
`Table 3-2: Friability of DC blends
`
`
`Formulation
`D25 (10% Drug Lot 17)
`D26 (25% Drug Lot 17)
`D27 (10% Drug Lot 16)
`
` %
`
`
`
`
`
`Friability
`0.04
`0.07
`0.03
`
`Content Uniformity for Beta Press Run
`(D25-D27)
`
`102
`
`100
`
`98
`
`96
`
`94
`
`92
`
`90
`
`0
`
`5
`
`15
`10
`Time of Sample (min)
`
`20
`
`25
`
`
`
`
`% Normalized Claim
`
`
`Figure 3-4: Content uniformity for the small scale beta press runs.
`
`CU results for these three batches (Figure 3-4) showed a slight upward trend in %
`claim as a function of time but the RSD was low at each time point.
`
`These results indicated that a) Segregation might be an issue with this formulation b) a
`mannitol to Avicel ratio of 1:1 was found to be adequate to get necessary tablet
`hardness. This ratio of the two fillers was used for all phase II formulations c) drug lot
`16 showed better hardness which was contrary to the effect observed at the smaller
`scale.
`
`3.1.3 Selection of the excipient grades for the Phase IIB formulation
`
`
`Since the small scale DC experiments indicated a slight upward trend in % label claim
`as a function of time, Avicel and mannitol grades were selected so as to match their
`particle size distribution (PSD) and density to that of the bulk drug, while retaining
`
`11 of 88
`
`Merck Exhibit 2123, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`adequate flow characteristics. Particle characteristics of Avicel PH 101/102, Pearlitol
`SD100/200 and Rouquette mannitol 35/60 were determined. The results are presented
`below. Based on these results Avicel 101 and Pearlitol SD 100 were selected for the
`manufacture of the biocomparison batch (comparing phase I capsules to tablets) since
`their size distribution and density was found to be closer to those of the bulk drug.
`
`
`
`
`Table 3-3: Particle Characteristics of Avicel/Mannitol/L-000224715
`
`Avicel PH 102
`Property
`Son Mean D10
`
`D95
`Particl
`
`
`
`e Size
`142
`136
`
`0s
`30s
`
`
`
`
`
`314
`295
`
`46
`46
`
`
`0.33
`0.42
`
`21%
`
`17
`
`1.27
`
`Avicel PH 101
`Mean
`D10
`
`D95
`
`91
`87
`
`
`
`
`
`219
`206
`
`L-000224715-006F024
`Mean
`D10
`
`D95
`
`83
`77
`
`
`
`224
`206
`
`
`
`19
`18
`
`0.55
`0.81
`
`
`32%
`19
`
`1.47
`
`
`
`30
`30
`
`
`0.33
`0.44
`
`25%
`
`26
`
`1.32
`
`Density
`[g/cm
`3]
`Loose
`Tapped
`
`Carr s Index
`
`Flodex
`Hausner
`Ratio
`
`Table 3-4: Particle Characteristics of Mannitol by Roquette
`
`Property
`S
`PS
`n
`m
`s
`
`0
`30
`
`Pearlitol SD 100
` D95
`Mean D10
`
`
`
`116
`104
`
`68
`62
`
`203
`182
`
`96
`84
`
`Pearlitol SD 200
` Mean D10
` D95
`
`
`
`166
`142
`
`293
`250
`
`Mannitol 35
` Mean D10
`
`D95
`
`
`42
`39
`
`Density
`3]
`[g/cm
`Loose
`Tapped
`Carr s
`Index
`
`Flodex
`Hausner
`Ratio
`
`3.1.4 Optimization of disintegrant and lubricant levels
`
`
`
`0.45
`0.56
`
`20%
`
`11
`
`1.24
`
`
`
`0.46
`0.54
`
`15%
`
`4
`
`1.17
`
` A
`
` series experiments (100 g scale) were designed to optimize the levels of the
`disintegrant and lubricant for the tablet formulation. In the previous development
`work, Croscarmellose sodium and Magnesium stearate were held constant at 3% and
`
`12 of 88
`
`Mannitol 60
` Mean D10
` D95
`
`
`
`102
`89
`
`31
`22
`
`261
`236
`
`115
`106
`
`
`
`8
`6
`
`
`
`0.53
`0.80
`
`34%
`
`34
`
`1.51
`
`
`
`0.62
`0.84
`
`26%
`
`28
`
`1.35
`
`Merck Exhibit 2123, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`
`2% respectively. This study investigated disintegrant concentrations of 1% and 3%
`and lubricant concentrations of 1.5% and 3%. The responses investigated were
`hardness, friability, dissolution, and tablet surface appearance (sticking observation).
`The two-level factorial experimental design is shown in Table 3-5. Formulation D26
`was used as a midpoint for this study.
`
`Table 3-5: Experimental design for disintegrant/lubricant study
`Formulation # Drug Loading Mann/Avicel Croscarm Na Mg Stearate
`D28
`25%
`1:1
`1.0%
`1.5%
`D29
`25%
`1:1
`3.0%
`1.5%
`D30
`25%
`1:1
`1.0%
`3.0%
`D31
`25%
`1:1
`3.0%
`3.0%
`
`25%
`
`1:1
`
`3.0%
`
`2.0%
`
`D26
`
`Components were blended in a PK (V-shell) blender for 10 min. Lubricant was added
`and the formulation was mixed for an additional 5 minutes. Compression was
`performed on a 16 station Manesty Beta Press. Compacts of 100 mg were produced
`using 8/32 round, standard concave tooling. The upper punch face included the
`lettering MRBA 84 . These tools were used as a worst case scenario, because the
`past experience in tablet compression has indicated that they typically display the
`worst sticking performance. Compressions were performed over an increasing
`compaction force to achieve a hardness profile for each formulation.
`
`Hardness Results
`
`Figure 3-5 shows the hardness data obtained from the above study. This data suggests
`that there is no significant difference in hardness for the formulations containing 1.5%
`and 2.0% Mg stearate (Formulations D28, D29, and D26). However, increasing
`lubricant concentration to 3.0% causes a significant hardness reduction at compression
`forces of interest (>7kN).
`
`
`13 of 88
`
`Merck Exhibit 2123, Page 13
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`Hardness Data for D28-D31
`
`
`
`D28
`(1%CCNa,
`1.5%MgSt)
`D29
`(3%CCNa,
`1.5%MgSt)
`D30
`(1%CCNa,
`3%MgSt)
`D31
`(3%CCNa,
`3%MgSt)
`D26
`(3%CCNa,
`2%MgSt)
`
`0
`
`10
`5
`Compression Force (kN)
`
`15
`
`01234567
`
`
`
`Harndess (kP)
`
`
`Figure 3-5: Content uniformity for the small scale beta press runs.
`
`Disintegration Results
`
`Figure 3-6 shows the disintegration data obtained from the disintegrant/lubricant
`optimization study. The data obtained from formulations D28-D31 showed that the
`amount of Mg stearate added to the formulation did not have a significant impact on
`tablet disintegration. The disintegration time was only controlled by the amount of
`croscarmellose sodium.
`
`
`D28
`(1%Cros,
`1.5%Mg st)
`D29
`(3%Cros,
`1.5%Mg st)
`D30
`(1%Cros,
`3%Mg st)
`D31
`(3%Cros,
`3%Mg st)
`D26
`(3%Cros,
`2%Mg st)
`
`Disintegration Time D28-D31
`
`4.0
`3.5
`3.0
`2.5
`2.0
`1.5
`1.0
`0.5
`0.0
`
`Disintegration Time
`
`(min)
`
`0
`
`10
`5
`Compression Force (kN)
`
`15
`
`
`Figure 3-6: Disintegrant/Lubricant Optimization Study disintegration results
`
`
`14 of 88
`
`Merck Exhibit 2123, Page 14
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`Friability Results
`
`Table 3-6 shows the friability results obtained from the finished tablets. Friabilities
`for all formulations were very low and well within the accepted range (<0.5%;100
`rev).
`
`Table 3-6: Disintegrant/Lubricant Optimization Study friability results
`
`
`
`
`% Friability (400
`rev.)
`
`
`
`
`
`
`0.249
`0.147
`0.299
`0.248
`------
`
`
`
`
`
`
`
`Form. # % Friability (100 rev.)
`D28
` 0.100
`D29
` 0.033
`D30
` 0.100
`D31
` 0.083
`D26
` 0.069
`
`15 of 88
`
`
`Thus it was observed that a) tablet hardness dropped at 3% Magnesium stearate level
`b) only levels of croscarmellose sodium impacted disintegration and about 2%
`croscarmellose sodium was necessary to get acceptable disintegration times. Based on
`these results 2% Magnesium stearate and 2% croscarmellose sodium were selected as
`the optimum levels for the Phase II formulation.
`
`3.1.5 Clinical Manufacture to support Phase II trials
`
`Due to the tight timelines of this program the scale-up of the DC process was
`conducted as part of the clinical manufacture to support Phase IIB trials. The
`following tables lists the various potencies made, their composition and the process
`train used.
`
`Table 3-7: Compositions of Phase II clinical batches
`wt%
`Ingredient
`
`100
`50
`50
`25
`10
`7.5
`5
`Tablet potency (mg)
`6.2 (5) 9.3 (7.5) 12.4 (10) 31 (25) 31 (25) 15.5 (12.5) 31 (25)
`L-000224715-006F*
`44.9
`43.35
`41.8
`32.5
`32.5
`40.25
`32.5
`Avicel PH 101
`44.9
`43.35
`41.8
`32.5
`32.5
`40.25
`32.5
`Mannitol SD 100
`2
`2
`2
`2
`2
`2
`2
`Croscarmellose sodium
`2
`2
`2
`2
`2
`2
`2
`Magnesium stearate
`100
`100
`100
`100
`200
`400
`400
`Core tablet wt. (mg)
`5
`5
`5
`5
`4
`4
`4
`Opadry White
`* L-000224715-006F is the anhydrous phosphate salt. Quantities in the bracket
`indicate weight % of the free base.
`
`
`
`
`
`
`
`Merck Exhibit 2123, Page 15
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`Table 3-8: Process trains for Phase II clinical batches.
`Theo. # of
`batch size
`PK blender
`tablets
`(kg)
`shell size
`
`
`
`38,000
`3.8
`16 qt
`35,000
`3.5
`16 qt
`125,000
`12.6
`56L
`
`
`
`38,000
`3.8
`16 qt
`140,000
`14.1
`56 L
`
`
`
`22,500
`2.25
`8 qt
`
`
`
`38,000
`3.8
`16 qt
`120,000
`12.1
`56 L
`140,000
`14.1
`56 L
`
`
`
`5,000
`1.005
`4 qt
`12,500
`2.5
`8 qt
`30,000
`12.1
`56 L
`100,000
`40.2
`141L
`
`
`
`30,000
`12.1
`56 L
`35,000
`14.1
`56 L
`
`
`
`
`
`
`100,000
`40.1
`142 L
`
`Batch #
` 5 mg
`001E001
`002E001
`003E001
` 7.5 mg
`001F001
`002F001
` 10 mg
`001D001
` 25 mg
`002A001
`003A001
`004A001
` 50 mg
`001B001*
`002B001
`003B001
`004B001
` 100 mg
`001C001
`002C001
`
` placebo
`001R001
`
`Image
`
`100
`100
`100
`
`100
`100
`
`100
`
`100
`100
`100
`
`200
`200
`400
`400
`
`400
`400
`
`
`400
`
`
`
`pan coater type
`(pan size )
`
`O Hara (19)
`O Hara (19)
`O Hara (24)
`
`O Hara (19)
`O Hara (24)
`
`vector (3.75 lit)
`
`O Hara (19)
`O Hara (24)
`O Hara (24)
`
`hi-coater
`vector (3.75 lit)
`O Hara (24)
`O Hara (30)
`
`O Hara (24)
`O Hara (24)
`
`
`O Hara (30)
`
`(36)
`
`
`
`
`
`
`
`O Hara
`
`vector (3.75 lit)
`
`Press type
`(# of tools)
`
`Beta (8)
`Hata (9)
`Beta (16)
`
`Beta (8)
`Beta (16)
`
`Beta (8)
`
`Beta (8)
`Hata (18)
`Beta (16)
`
`Beta (2)
`Beta (4)
`Beta (8)
`Beta (16)
`
`Beta (8)
`Beta (8)
`
`
`Beta (16)
`Courtoy
`(30)
`
`Beta (4)
`
`202.5
`
`2.5
`
`600 L (Bohle)
`
`8 qt
`
`16 of 88
`
`505,050
`400
`002R001
`
`
`
`12,500
`200
`001Q001
`* Only part of the batch was coated
`
`The results for the in-process testing of tablets (hardness, CU, composite assay) can be
`found in the release documents for these batches. The performance of all of these
`formulations was found to be satisfactory. The beta press was run at a low speed of
`45-60 rpm for these batches to get higher hardness and to minimize sticking during
`compression.
`
`3.1.6 Sticking issues during compression
`
`Sticking and hazing was observed during tablet compression at small scale. Hence
`tablets presses were run at slow speeds (around 45-50 rpm) and with plain tooling to
`prepare Phase IIB clinical supplies in order to minimize sticking during compression.
`Fig 3-7 shows the hazing and sticking observed on the tablet punches even at such
`slow speeds (and plain tools).
`
`
`Merck Exhibit 2123, Page 16
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`
`
`
`
`
`Figure 3-7: Sticking/hazing seen on the upper
`punch for the 25% drug load phase IIB
`formulation (0431FCT002C001) at 45 rpm
`speeds.
`
`
`In order to scale-up this process, higher speeds (around 90 rpm on the beta press to
`match linear velocities on the production scale presses) and embossed tooling were
`evaluated.
`
`As described in the next section, in general, less sticking was observed with roller
`compacted formulations than with DC blends, most likely due to the larger particle
`size of the compacted granules. However sticking during compression was severe
`even with roller compacted formulations when an embossed tool was used at high
`compression speeds. Thus sticking and picking were considered one of the major
`problems for formulation/process development for this product. Numerous remedies
`were evaluated to mitigate this problem. Before getting into the details of these
`solutions, first a small scale sticking test is described that was developed by this team
`to compare the sticking propensity of various formulations.
`
`3.1.7 Small scale sticking test development
`
`To evaluate the sticking propensity of various formulations a quick small scale
`sticking test was found to be necessary. A variety of such methods have been
`developed by members of Dept. 854 in the past. These tests are based on compression
`of 5-10 tablets on a Carver press followed by the analysis of the punch surfaces using:
`a) visual inspection b) image analysis after light microscopy c) Near IR analysis d)
`HPLC assay of the residue on the punch surface. All of these methods were found to
`have varying degrees of success with various formulations. Although more
`quantitative, these methods were found to be time consuming and labor intensive. As
`a result, the MK-0431 development team developed a new fast semi-quantitative
`method to evaluate sticking based on the Korsch press. A brief description of this
`method is as follows: About 20 g of formulations were compressed on a Korsch press
`at low speeds (around 25 rpm) and around 8-9 kN compression force using two tools
`(one plain and one embossed, both 8/32 normal round concave). The punch surfaces
`
`17 of 88
`
`Merck Exhibit 2123, Page 17
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`
`were photographed using a light microscope after compression and then rated by the
`analysts using the following scale:
`
`1) No hazing or sticking
`
`
`
`
`
`
`
`
`
`2) Light hazing in cap
`
`
`
`3) Moderate hazing
`
`
`
`
`18 of 88
`
`Merck Exhibit 2123, Page 18
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`

`

`
`4) Severe Hazing and sticking
`
`
`
`
`
`5) Extreme hazing and sticking
`
`
`
`
`
`
`All of the formulations made for this program were rank ordered using this small scale
`sticking method.
`
`3.1.8 Remedies to alleviate sticking during compression
`
`3.1.8.1 Use of alternate filler in place of mannitol
`
`To reduce sticking, a 33%DL dicalcium phosphate based formulation was prepared
`and tested for sticking tendency using compression on a Carver press (this study was
`conducted before the small scale sticking test method described above was developed).
`The NIR data in Figure 3-8 shows that dicalcium phosphate indeed exhibited less
`sticking than Avicel or Avicel/mannitol based formulations. The t