`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`__________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`__________________
`
`MYLAN PHARMACEUTICALS INC.,
`Petitioner,
`
`v.
`
`MERCK SHARP & DOHME CORP.,
`Patent Owner.
`__________________
`
`Case IPR2020-00040
`U.S. Patent 7,326,708
`__________________
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`
`
`
`Merck Exhibit 2116, Page 1
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`I, Robert M. Wenslow, Ph.D., hereby declare as follows:
`
`I.
`
`INTRODUCTION
`
`1.
`
`I am a named inventor of subject matter claimed in U.S. Patent No.
`
`7,326,708 (“the ’708 patent”). I understand that Merck Sharp & Dohme Corp.
`
`(“Merck”) is the assignee of the ’708 patent.
`
`2.
`
`I understand that Mylan Pharmaceuticals Inc. (“Mylan”) has filed a
`
`petition for inter partes review (“IPR”) challenging the patentability of certain
`
`claims of the ’708 patent. I previously submitted a declaration with Merck’s
`
`preliminary response to Mylan’s petition; my previous declaration is Exhibit 2003
`
`in IPR2020-00040. I incorporate herein by reference the testimony in my
`
`previous declaration, including with respect to my professional background, my
`
`previous roles and responsibilities at Merck, and the work of the DPP-IV project
`
`team to develop sitagliptin into an oral treatment for type 2 diabetes, which Merck
`
`markets today under the tradename Januvia®.
`
`3.
`
`In this declaration, I provide additional facts known to me regarding
`
`the development of Januvia®, in particular the solid-state characterization of
`
`various sitagliptin salts that Merck considered during the development of
`
`Januvia®, the discovery of the crystalline monohydrate of the 1:1 DHP salt of
`
`sitagliptin, and Merck’s decision to select the monohydrate for further
`
`development (and ultimately as the form of sitagliptin used in Januvia®).
`
`2
`
`Merck Exhibit 2116, Page 2
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`4.
`
`At the time of Januvia®’s development, I led a small team of
`
`scientists in the Physical Measurements group, which was then a part of the
`
`Analytical Research department of Merck Research Laboratories (“MRL”). The
`
`primary responsibility of my team was to perform solid-state characterization of
`
`candidate drugs, including the discovery and validation of new polymorphic crystal
`
`forms. The analytical tests performed by me and my team included x-ray powder
`
`diffraction (“XRPD”), solid-state nuclear magnetic resonance (“ssNMR”)
`
`spectroscopy, differential scanning calorimetry (“DSC”), and thermogravimetric
`
`analysis (“TGA”). These tests were performed in accordance with standard
`
`protocols that were established and followed by Merck scientists.
`
`5.
`
`During the development of Januvia®, I regularly discussed the
`
`development of sitagliptin with DPP-IV project team members from other
`
`departments—including Process Research, Chemical Engineering Research &
`
`Development (“CERD”), and Pharmaceutical Research & Development
`
`(“PR&D”)—and was involved in key decisions, including the decision to develop
`
`the crystalline monohydrate form of the 1:1 DHP salt. I also supervised Drs.
`
`Russell Ferlita and Alex Chen, who are also named co-inventors of the ’708 patent,
`
`as well as Yaling Wang, all of whom were members of the Physical Measurements
`
`group that worked on Januvia®. As such, I have first-hand knowledge of the data
`
`generated during Januvia®’s development.
`
`3
`
`Merck Exhibit 2116, Page 3
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`II.
`
`SOLID STATE CHARACTERIZATION OF SITAGLIPTIN SALTS
`
`6.
`
`In the early phases of Merck’s development of Januvia®, scientists on
`
`the DPP-IV project team, including Vicky Vydra and Dr. Karl Hansen, synthesized
`
`various salts of sitagliptin that were considered for further development. In
`
`addition to the 1:1 DHP salt of sitagliptin, other salts that were considered during
`
`development included sitagliptin salts formed with tartaric acid, benzenesulfonic
`
`acid, and hydrochloric acid.
`
`7.
`
`The Physical Measurements group was responsible for the solid-state
`
`characterization of these salts through, inter alia, DSC and TGA, which are
`
`standard tests that provide information on the thermal properties of solid materials,
`
`in particular weight loss, melting or evaporation points, thermal decomposition,
`
`and other phase transitions as a function of temperature.
`
`8.
`
`At the time of Januvia®’s development, the Physical Measurements
`
`group under my supervision followed a set of standard protocols to perform DSC
`
`and TGA. DSC analyses were performed on a TA Instruments DSC 2910 or an
`
`equivalent instrument. Between 2 and 6 mg of the solid material was placed into
`
`an open pan. The pan was then crimped and placed at the sample position in the
`
`instrument’s calorimeter cell. An empty pan was placed at the instrument’s
`
`reference position. The calorimeter cell was then closed and a flow of nitrogen
`
`was passed through the cell. The heating program was set to heat the sample at a
`
`4
`
`Merck Exhibit 2116, Page 4
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`heating rate of 10°C/min to a temperature of approximately 250°C. The heating
`
`program was then started. When the run was completed, the data were analyzed
`
`using the DSC analysis program contained in the system software. The melting
`
`endotherm was integrated between baseline temperature points that were above
`
`and below the temperature range over which the endotherm was observed.
`
`9.
`
`TGA were performed on a Perkin Elmer model TGA 7 or equivalent
`
`using a heating rate of 10°C/min to a maximum temperature of approximately
`
`250°C and samples were placed under nitrogen flow. After automatically taring
`
`the balance, 5 to 20 mg of the solid material was added to a platinum pan, the
`
`furnace raised, and the heating program started. Weight and temperature data were
`
`collected automatically by the instrument. Analysis of the results was carried out
`
`by selecting the Delta Y function within the instrument software and choosing the
`
`temperatures between which to calculate the weight loss. Weight losses were
`
`reported up to the onset of decomposition or evaporation.
`
`10.
`
`In the following sections, I provide a summary of the DSC and TGA
`
`data obtained by the Physical Measurements group for the phosphoric, tartaric,
`
`benzenesulfonic, and hydrochloric acid salts of sitagliptin that were considered
`
`during Januvia®’s development—specifically, the crystalline anhydrates (Forms I,
`
`II, and III) of the 1:1 DHP salt and its crystalline monohydrate form (which
`
`unexpectedly appeared after over a year spent developing the anhydrates), the
`
`5
`
`Merck Exhibit 2116, Page 5
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`crystalline L-tartaric acid salt hemihydrate, the crystalline benzenesulfonic acid
`
`salt anhydrate, and the crystalline hydrochloric acid salt monohydrate.1 These data
`
`were collected by me or under my supervision following Merck’s standard
`
`procedures and were reported in the ’708 patent (EX1001), WO 2005/020920
`
`(“WO ’920,” EX2117), and WO 2005/072530 (“WO ’530,” EX2118).
`
`
`
`Phosphoric Acid
`
`Figure 1. DSC of the 1:1 DHP sitagliptin salt, anhydrate Form I.
`
`
`
`
`1 The XRPD, NMR, DSC, and TGA data for the crystalline monohydrate of the 1:1
`
`DHP salt of sitagliptin disclosed the ’708 patent were generated by myself and Drs.
`
`Ferlita and Chen. In accordance with Merck’s policies and procedures, and as a
`
`condition of my employment, I understood at the time that I was under an
`
`obligation to assign (and did assign) this subject matter to Merck.
`
`6
`
`Merck Exhibit 2116, Page 6
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`11. As shown in Figure 1, the DSC curve for the crystalline anhydrate
`
`Form I of the 1:1 DHP salt of sitagliptin exhibited a melting endotherm with an
`
`onset temperature of 215.37°C and a peak temperature of 217.27°C. See WO ’920
`
`(EX2117) fig.4; id. at 17:29–31.
`
`
`
`Figure 2. TGA of the 1:1 DHP sitagliptin salt, anhydrate Form I.
`
`12. As shown in Figure 2, the TGA curve for the crystalline anhydrate
`
`Form I of the 1:1 DHP salt of sitagliptin exhibited a weight loss of 0.1638% from
`
`ambient temperature to about 180°C. See WO ’920 (EX2117) fig.5.
`
`7
`
`Merck Exhibit 2116, Page 7
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`
`
`Figure 3. DSC of the 1:1 DHP sitagliptin salt, anhydrate Form II.
`
`13. As shown in Figure 3, the DSC curve for the crystalline anhydrate
`
`Form II of the 1:1 DHP salt of sitagliptin exhibited a solid-solid transition
`
`exotherm to Form I with an onset temperature of 114.36°C and a peak temperature
`
`of 124.73°C. See WO ’920 (EX2117) fig.9; id. at 18:3–5. An endotherm with an
`
`onset of 213.80°C corresponding to melting/decomposition was also observed.
`
`Figure 4. TGA of the 1:1 DHP sitagliptin salt, anhydrate Form II.
`
`
`
`8
`
`Merck Exhibit 2116, Page 8
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`14. As shown in Figure 4, the TGA curve for the crystalline anhydrate
`
`Form II of the 1:1 DHP salt of sitagliptin exhibited a weight loss of 0.3900% from
`
`ambient temperature to about 180°C. WO ’920 (EX2117) fig.10.
`
`
`
`Figure 5. DSC of the 1:1 DHP sitagliptin salt, anhydrate Form III.
`
`15. As shown in Figure 5, the DSC curve for the crystalline anhydrate
`
`Form III of the 1:1 DHP salt of sitagliptin exhibited a solid-solid transition
`
`endotherm to Form I with an onset temperature of 80.07°C and a peak temperature
`
`of 83.80°C. See WO ’920 (EX2117) fig.14; id. at 18:13–35. A second endotherm
`
`corresponding to melting/decomposition with an onset temperature of 215.94°C
`
`was also observed.
`
`9
`
`Merck Exhibit 2116, Page 9
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`
`
`Figure 6. TGA of the 1:1 DHP sitagliptin, anhydrate Form III.
`
`16. As shown in Figure 6, the TGA curve for the crystalline anhydrate
`
`Form III of the 1:1 DHP salt of sitagliptin exhibited little to no weight loss before
`
`about 200°C. See WO ’920 (EX2117) fig.15.
`
`Figure 7. DSC of the 1:1 DHP sitagliptin salt, crystalline monohydrate.
`
`
`
`10
`
`Merck Exhibit 2116, Page 10
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`17. As shown in Figure 7, the DSC curve for the crystalline monohydrate
`
`of the 1:1 DHP salt of sitagliptin exhibited an endotherm corresponding to loss of
`
`water and solid-solid phase transitions with an onset temperature of 138.09°C and
`
`a peak temperature of 140.66°C. A second endotherm with an onset of 209.63°C
`
`and a peak of 213.23°C reflects the melting/decomposition of the 1:1 DHP salt.
`
`See ’708 patent (EX1001) fig.5; id. at 14:30–47.
`
`
`
`Figure 8. TGA of the 1:1 DHP sitagliptin salt, crystalline monohydrate
`
`18. As shown in Figure 8, the TGA curve for the crystalline monohydrate
`
`of the 1:1 DHP salt of sitagliptin exhibited a 3.3647% weight loss from ambient
`
`temperature to about 250°C, corresponding the loss of the monohydrate water. See
`
`’708 patent (EX1001) fig.4; id. at 14:12–29. Practically no loss of water was
`
`observed through about 100°C.
`
`11
`
`Merck Exhibit 2116, Page 11
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`
`
`Tartaric Acid
`
`19. As shown in Figure 9, the DSC curve collected for the crystalline L-
`
`tartaric acid salt hemihydrate of sitagliptin exhibited a broad endotherm
`
`corresponding to the evolution of water with an onset temperature of 33.70°C and
`
`a peak temperature of 53.38°C. A second endotherm with an onset of about 200°C
`
`and a peak temperature of 203.99°C reflects the melting point/decomposition of
`
`the L-tartaric acid salt. See WO ’530 (EX2118) fig.6; id. at 18:8–12.
`
`
`
`Figure 9. DSC of the crystalline L-tartaric acid salt hemihydrate of sitagliptin.
`
`20. As shown in Figure 10, the TGA curve collected for the crystalline L-
`
`tartaric acid salt hemihydrate of sitagliptin exhibited a 1.3523% weight loss from
`
`ambient temperature to about 198°C. See WO ’530 (EX2118) fig.5; id. at 19:1–3.
`
`12
`
`Merck Exhibit 2116, Page 12
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`Figure 10. TGA of the crystalline L-tartaric acid salt hemihydrate of sitagliptin.
`
`
`
` Benzenesulfonic Acid
`
`21. As shown in Figure 11, the DSC curve for the crystalline
`
`benzenesulfonic acid salt anhydrate of sitagliptin exhibited a sharp melting
`
`endotherm with an onset temperature of 176.26°C and peak temperature of
`
`179.05°C. See WO ’530 (EX2118) fig.9; id. at 18:13–16.
`
`Figure 11. DSC of the crystalline benzenesulfonic acid salt
`anhydrate of sitagliptin.
`
`
`
`13
`
`Merck Exhibit 2116, Page 13
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`22. As shown in Figure 12, the TGA curve collected for the crystalline
`
`benzenesulfonic acid salt anhydrate of sitagliptin exhibited a 0.1475% from about
`
`63°C to about 203°C. See WO ’530 (EX2118) fig.8; id. at 19:4–6.
`
`Figure 12. TGA of the crystalline benzenesulfonic acid salt
`anhydrate of sitagliptin.
`
` Hydrochloric Acid
`
`
`
`23. As shown in Figure 13, the DSC curve for the crystalline hydrochloric
`
`acid salt monohydrate of sitagliptin exhibited a broad endotherm corresponding to
`
`the loss of water with an onset temperature of 60.34°C and a peak temperature of
`
`74.43°C. A second endotherm with an onset of 164.73°C and a peak of 170.72°C
`
`reflects the melting point of the hydrochloric acid salt. See WO ’530 (EX2118)
`
`fig.3; id. at 18:2–7.
`
`14
`
`Merck Exhibit 2116, Page 14
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`Figure 13. DSC of the crystalline hydrochloric acid salt
`monohydrate of sitagliptin.
`
`
`
`24. As shown in Figure 14, the TGA curve for the crystalline hydrochloric
`
`acid salt monohydrate of sitagliptin exhibited a weight loss of 3.1307% from
`
`ambient temperature to 83°C. See WO ’530 (EX2118) fig.2; id. at 18:33–35.
`
`Figure 14. TGA of the crystalline hydrochloric acid salt
`monohydrate of sitagliptin.
`
`
`
`15
`
`Merck Exhibit 2116, Page 15
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`25. The data above show that the crystalline monohydrate of the 1:1 DHP
`
`salt of sitagliptin is the most stable hydrate as compared to the crystalline L-tartaric
`
`acid salt hemihydrate and the crystalline hydrochloric acid salt monohydrate. The
`
`crystalline L-tartaric acid salt hemihydrate and the crystalline hydrochloric acid
`
`salt monohydrate showed complete loss of their water content well below 100°C,
`
`as shown by their respective TGA curves (see supra Figure 10 and Figure 14). In
`
`contrast, the crystalline monohydrate of the 1:1 DHP salt of sitagliptin did not
`
`exhibit an onset of dehydration until after 100°C (see Figure 8).
`
`26. The data above also show that the various crystal forms of the 1:1
`
`DHP salt have the highest melting and/or decomposition temperature as compared
`
`to the other salts of sitagliptin formed from tartaric, benzenesulfonic, and
`
`hydrochloric acid. These data are summarized in Table 1, below. These higher
`
`melting points are indicative of the higher thermal stability of 1:1 DHP salt as
`
`compared to the other salts that were considered during Januvia®’s development.
`
`Table 1. Comparison of onset of melting/decomposition temperatures
`
`Salt/Crystal Form
`1:1 DHP, Monohydrate
`1:1 DHP, Anhydrate Form I
`1:1 DHP, Anhydrate Form II
`1:1 DHP, Anhydrate Form III
`Tartrate, Hemihydrate
`Besylate, Anhydrate
`HCl, Monohydrate
`
`MP (°C)
`209.63
`215.37
`213.80
`215.94
`~200 (<203.99)
`176.26
`164.73
`
`
`
`16
`
`Merck Exhibit 2116, Page 16
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`III. CREATION OF THE CRYSTALLINE MONOHYDRATE OF THE 1:1
`DHP SALT OF SITAGLIPTIN
`
`27. The crystalline monohydrate of the 1:1 DHP salt of sitagliptin was
`
`first created in March 2003. The appearance of the monohydrate—over a year
`
`after the initial synthesis of the 1:1 DHP salt in December 2001 by Vicky Vydra
`
`and subsequent scale-up efforts by Dr. Karl Hansen in January, February, and
`
`March of 2002 and well into the development of Januvia®—was surprising and
`
`unexpected to me and others on the DPP-IV project team.
`
`28. Throughout Januvia®’s development, the Physical Measurements
`
`team under my supervision undertook extensive efforts to identify and characterize
`
`the polymorphs of the 1:1 DHP salt. Just prior to the monohydrate’s synthesis,
`
`three anhydrous forms had been identified and characterized: Form I, Form II, and
`
`Form III. The Form II anhydrate is a metastable form that was obtained by
`
`desolvating (drying) a non-stoichiometric ethanol solvate of the 1:1 DHP salt, the
`
`major product of the crystallization step used to make early clinical batches. Upon
`
`further drying or storage, Form II (which is metastable) slowly converts to a
`
`mixture of Forms I and III, with faster conversion at elevated temperatures. Form I
`
`and Form III are enantiotropic polymorphs that reversibly interconvert; Form I is
`
`more stable above 34°C (the Form I/III transition temperature), while Form III is
`
`more stable below. Additionally, Form I was determined to convert to Form III
`
`under agitation or shear.
`
`17
`
`Merck Exhibit 2116, Page 17
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`29. The phase relationships between the Form I, Form II, and Form III
`
`anhydrates of the 1:1 DHP salt of sitagliptin are shown in Figure 15.
`
`Form II
`
` RT
`
`Form I
`
`
`Agitation
`
`Ttrans = 34 C
`
`Form III
`
`
`
`Figure 15. Phase diagram of Form I, Form II, and Form III.
`
`30. As a consequence of the interconversion between the anhydrous
`
`forms, delivered lots of the 1:1 DHP salt were determined to be a mixture of Forms
`
`I, II, and III depending on the conditions of their storage and processing. Another
`
`consequence of the enantiotropic relationship between Form I and Form III was
`
`that common pharmaceutical processing techniques such as milling, compression,
`
`and other high-shear techniques resulted in conversion from Form I to Form III.2
`
`
`2 These conclusions, their underlying data, and the tests that were performed to
`
`obtain them were summarized in a February 26, 2003 memorandum co-authored
`
`by me to the DPP-IV project team, a true and correct copy of which may be found
`
`in EX2119, and which I incorporate herein by reference. The phase relationships
`
`and interconversion of the anhydrous forms that was derived from the work of the
`
`Physical Measurements team was also disclosed in WO ’920. See EX2117 (WO
`
`’920) at 4:1–7, 4:30–33, 8:1–26.
`
`18
`
`Merck Exhibit 2116, Page 18
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`31. For instance, in one experiment, samples of anhydrous lots were
`
`subjected to compression at 150 or 250 MPa, and the relative amounts of Form I
`
`and III were quantitated through fluorine ssNMR. The results are summarized in
`
`Table 2. See EX2119 at 8.
`
`Table 2. Form I/III content after 150 or 250 MPa compaction.
`
`
`After 250 MPa
`After 150 MPa
`Before Compaction
`Sample % Form I % Form III % Form I % Form III % Form I % Form III
`A
`81.2
`18.8
`52.3
`47.7
`n.d.
`n.d.
`B
`Mixture of Form I/II
`48.1
`51.9
`29.7
`70.3
`C
`14.7
`85.3
`21.3
`78.7
`23.0
`77.0
`
`
`
`32. Similarly, conversion from Form I to Form III in formulated drug
`
`samples was observed by combining samples of mostly pure Form I, Form II, or
`
`Form III with excipients using a composition similar the Phase IIB formulation at
`
`25% drug load, and analyzing their form content before and after compaction using
`
`fluorine ssNMR. See EX2119 at 6.
`
`33. The complex phase relationships between the anhydrous forms and
`
`the shear-sensitive conversion of Form I to Form III posed a problem for the DPP-
`
`IV project team. Because these anhydrous forms readily interconverted under
`
`normal manufacturing and processing conditions, a high degree of control of both
`
`the manufacture of the 1:1 DHP salt and its pharmaceutical processing during
`
`formulation would have been required to control the identity and proportion of the
`
`polymorphic forms in the final drug product. To avoid having to implement these
`
`19
`
`Merck Exhibit 2116, Page 19
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`controls, the DPP-IV project team hypothesized that pure Form I could be obtained
`
`directly—without having to go through the metastable desolvated Form II
`
`anhydrate—by recrystallizing the 1:1 DHP salt in a non-solvating solvent and then
`
`maintained by adequate storage controls.
`
`34.
`
`In January 2003, members of the DPP-IV project team from Physical
`
`Measurements and CERD worked to identify a solvent that would 1) be non-
`
`solvating for the 1:1 DHP salt; 2) have high solubility for the sitagliptin freebase;
`
`3) be miscible with water (for adding phosphoric acid); 4) not generate an
`
`amorphous salt; and 5) not degrade sitagliptin. The results of our work to identify a
`
`non-solvating solvent for the non-1:1 DHP salt are summarized in Table 3.
`
`Table 3. Results of non-solvating solvent experiments.
`
`Solvent
`
`Water*
`Methanol
`Ethanol
`1-Propanol
`2-Propanol
`t-Butanol
`Cyclohexanol
`Ethyl Acetate
`Acetone
`Methyl
`Butyl Ether
`THF
`n-Hexane
`Cyclohexane
`n-Heptane
`Toluene
`Acetonitrile
`DMF
`DMAC
`DMSO
`
`Forms
`Solvate?
`N
`Y
`Y
`Y
`Y
`N
`N
`Y
`Y
`N
`Y
`Y
`N
`N
`N
`N
`Y
`Y
`Y
`Y
`
`Becomes
`Amorph.?
`N
`N
`N
`N
`N
`Y
`Y
`N
`N
`N
`N
`N
`N
`N
`N
`Y
`N
`N
`N
`N
`
`Freebase
`Soluble?
`N
`Y
`Y
`
`
`Y
`
`
`
`Y
`
`Y
`N
`N
`N
`Y
`
`
`
`Y
`
`20
`
`DHP Salt
`Soluble?
`Y
`
`N
`
`
`
`
`
`
`N
`
`N
`N
`N
`N
`Gels
`
`
`
`Y
`
`Degrades
`Product?
`N
`
`N
`
`
`
`
`
`
`Y
`
`N
`
`
`
`
`
`
`
`
`
`Miscible w/
`H3PO4?
`Y
`
`Y
`
`
`
`
`
`
`Y
`
`Y
`
`
`
`
`
`
`
`Y
`
`Merck Exhibit 2116, Page 20
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`Solvent
`
`MTBE
`Methylene Chloride
`Diethoxymethane
`Methyl Benzoate
`Isoamyl Alcohol
`*As of February 2003.
`
`
`Forms
`Solvate?
`N
`Y
`Y
`N
`N
`
`Becomes
`Amorph.?
`N
`N
`N
`
`N
`
`Freebase
`Soluble?
`Y
`Y
`
`
`Y
`
`DHP Salt
`Soluble?
`N
`N
`
`Sparingly
`N
`
`Degrades
`Product?
`
`
`
`
`N
`
`Miscible w/
`H3PO4?
`N
`N
`
`
`Y
`
`35. As indicated in Table 3, crystallization in isoamyl alcohol (“IAA”)
`
`resulted in a non-solvated crystal of the 1:1 DHP salt, which based on XRPD was
`
`anhydrous Form I. However, the morphology of the Form I crystals was poor
`
`(needles or needle-like). To improve the morphology of the Form I crystal, the
`
`solvent was adjusted to a mixture of 95/5% IAA/water. The resulting particles had
`
`a more favorable rod morphology. But in a surprising and completely unexpected
`
`turn of events, crystallization of the 1:1 DHP salt in 95/5% IAA/water also resulted
`
`in the creation of a new form: the monohydrate.
`
`36. The creation of the monohydrate was a highly surprising development
`
`to the DPP-IV project team as a whole—and to me and the Physical Measurements
`
`group in particular—as the determination of the Form I/III transition temperature
`
`had been performed by determining the solubility of the anhydrous forms in
`
`water—an experiment that could no longer be repeated due to conversion of the
`
`anhydrous forms to the monohydrate. See EX2119 at 10; see also id. at 11 (water
`
`slurry experiments conducted before the appearance of the monohydrate).
`
`Additionally, over a year of development on the 1:1 DHP salt had taken place,
`
`21
`
`Merck Exhibit 2116, Page 21
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`including extensive experiments with the 1:1 DHP salt in both water and aqueous
`
`solvent mixtures, as well as significant efforts to identify all of its polymorphs,
`
`hydrates, and solvates. See EX2119 at 1–5, 10–11. Despite this, the DPP-IV
`
`project team had not observed a monohydrate until its surprising and unexpected
`
`creation very late in the development cycle.
`
`IV. SELECTION OF THE CRYSTALLINE MONOHYDRATE FOR
`FURTHER DEVELOPMENT
`
`37. The creation and identification of the monohydrate took place
`
`approximately two months before the DPP-IV project team was scheduled to
`
`produce clinical supplies for Phase IIB clinical studies in human subjects. This
`
`presented a challenge to the project team with respect to which crystal form of
`
`sitagliptin to take forward in the development cycle: the previous anhydrous
`
`mixture of Forms I, II, and III; or the new crystalline monohydrate.
`
`38.
`
`In order to provide data to guide this decision, scientists in both the
`
`Physical Measurements group and PR&D worked closely together to characterize
`
`the monohydrate’s physical properties and behavior in formulations. Ultimately,
`
`and serendipitously, the monohydrate’s physical stability, chemical stability, and
`
`its improved pharmaceutical processability (notably, reduced stickiness) led the
`
`DPP-IV project team to select the monohydrate over the anhydrous forms for
`
`further development. These data were summarized in an April 9, 2003
`
`memorandum co-authored by me and other scientists from Physical Measurements,
`
`22
`
`Merck Exhibit 2116, Page 22
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`PR&D, and CERD, as well as a presentation, “Parallel Development of Multiple
`
`Crystal Forms for a New Drug Candidate: Selection of the Final Form via
`
`Integrated Chemical and Pharmaceutical Process Evaluation,” which was delivered
`
`at the November 2003 Polymorphism and Crystallization Forum held in
`
`Philadelphia, Pennsylvania, true and correct copies of which may be found in
`
`EX2120 and EX2122.
`
`
`
`Physical Stability
`
`39. The monohydrate was shown to be stable against dehydration under
`
`ordinary conditions through dynamic vapor sorption experiments conducted at
`
`25°C and 40°C. See EX2122 at 31 (“Monohydrate NEVER loses water at 25°C
`
`and 40°C > 0.8% RH”); see supra Figure 7 and Figure 8 (DSC and TGA
`
`confirming the stability of the monohydrate). Nevertheless, experiments
`
`conducted by Physical Measurements determined that the monohydrate could be
`
`dehydrated by heating the crystal to temperatures between 40°C and 130°C and
`
`placing it under nitrogen flow. This dehydrated phase exhibited a similar XRPD
`
`pattern but was distinguishable by its lack of weight loss from TGA. The
`
`transition was found to be reversible by exposing the dehydrated monohydrate to
`
`ambient conditions. Continued heating of the dehydrated monohydrate resulted in
`
`the formation of another anhydrate, Form IV, which converted back to the
`
`monohydrate at 98% relative humidity and room temperature. Continued heating
`
`23
`
`Merck Exhibit 2116, Page 23
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`of Form IV resulted in Form I, which was found to convert back to the
`
`monohydrate by slurrying in water. The phase relationships for the monohydrate
`
`system are summarized in Figure 16. Notably, following the appearance of the
`
`crystalline monohydrate, slurrying the 1:1 DHP salt in water resulting in the
`
`formation of the monohydrate, a process that previously only yielded anhydrous
`
`forms. See EX2122 at 33; see also id. at 27 (“All anhydrous forms, when put in
`
`water, now turn over to the monohydrate (previously, Form I).”).
`
`RT
`
`Monohydrate
`
`40–130 °C
` N2 flow
`
`Dehydrated
`monohydrate
`
`60–130 °C
`
`Form IV
`(anhydrous)
`
`130 ~ 180 °C
`
`Form I
`
`RH=98%, RT
`
`Water slurry with
`or without
`monohydrate seed
`
`Figure 16. Phase diagram of the 1:1 DHP monohydrate system.
`
`
`
`40. This data indicated that, unlike the anhydrous forms, it was possible to
`
`maintain a single polymorphic form of the monohydrate at ambient conditions.
`
`This was an extremely positive result that overcame a critical challenge in the
`
`development of Januvia®—adequate control of crystal polymorphism and the
`
`overall properties of the drug substance—and one that was not predictable even
`
`after the creation of the monohydrate.
`
`24
`
`Merck Exhibit 2116, Page 24
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`41. Additionally, unlike the anhydrous forms, the monohydrate did not
`
`convert to another crystal form under shear or pressure. As shown in Figure 17,
`
`compression of an anhydrous lot (“L-224715-006F024” or “Lot 24”) at 200 MPa
`
`resulted in a material with different XPRD pattern, while compression of the new
`
`crystalline monohydrate (“NB# 66839-113”) did not. See Ex. 2120 at 5. This
`
`property was another unexpected benefit of the monohydrate over the previous
`
`anhydrous forms.
`
`
`
`
`
`Figure 17. Compression effect on anhydrous lot (top) and monohydrate (bottom).
`
`25
`
` L-224715-006F024
` Lot 24 Compressed at 200 Mpa
`
`1500
`
`1000
`
`Intensity
`
`500
`
`10
`
`20
`Diffraction Angle
`
`30
`
`40
`
` NB# 66839_113
` NB# 66839_113 Compressed at 200 Mpa
`
`2000
`
`
`
`1500
`
`1000
`
`500
`
`10
`
`20
`
`30
`
`40
`
`Merck Exhibit 2116, Page 25
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
` Chemical Stability
`
`42. The monohydrate was also unexpectedly found to have improved
`
`chemical stability as compared to the anhydrous forms. In probe stability studies
`
`conducted at 40°C and 75% relative humidity, formulations using the monohydrate
`
`were found to generate fewer impurities than comparable formulations using
`
`anhydrous forms. See EX2122 at 40. Additionally, in formaldehyde stress tests,
`
`the bulk monohydrate was shown to resist discoloration as compared to the bulk
`
`anhydrous forms. See EX2122 at 39.
`
`
`
`Pharmaceutical Processing
`
`43. Both the bulk monohydrate API and formulations of the monohydrate
`
`were found to also have unexpectedly reduced sticking compared to the bulk
`
`anhydrous API and comparable formulations. Using a Carver press with a punch
`
`surface embossed with “MSD,” a sample of the monohydrate (“L-224715-66839-
`
`113”) was determined to produce less sticking compared a sample of anhydrous
`
`forms (“Lot 24”), with only about 150 µg of monohydrate remaining on the punch
`
`surface, as compared to over 300 µg of the anhydrous forms. See EX2120 at 5.
`
`44. Comparable formulations of the monohydrate and anhydrous forms
`
`were additionally subjected to a 5 minute compression run using a Korsch tablet
`
`press applying approximately 9 kN of force. See EX2122 at 38. As shown in
`
`Figure 18, a smaller amount of the monohydrate formulation was left on the
`
`26
`
`Merck Exhibit 2116, Page 26
`Mylan Pharmaceuticals Inc. v. Merck Sharp & Dohme Corp.
`IPR2020-00040
`
`
`
`SECOND DECLARATION OF ROBERT M. WENSLOW, PH.D.
`
`surface of the tablet punch. Id. This result was a further surprising and unexpected
`
`property of the monohydrate that led to it selection over the anhydrous forms for
`
`further development.
`
`
`
`
`
`Figure 18. Tablet-punch sticking of anhydrous formulation (left) and
`monohydrate formulation (right).
`
`*
`
`*
`
`*
`
`I hereby declare that all statements made herein of my own knowledge are
`
`true and that all statements made on information and belie