`
`
`
`
`
`
`
`
`
`Filed on behalf of: CSL Behring GmbH and CSL Behring LLC
`
`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`
`__________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`__________________
`
`CSL BEHRING GMBH and CSL BEHRING LLC,
`Petitioners,
`
`v.
`
`
`SHIRE VIROPHARMA INC.,
`Patent Owner.
`
`__________________
`
`
`U.S. Patent No. 9,616,111
`
`__________________
`
`
`
`DECLARATION OF DR. CHRISTOPHER J. ROBERTS
`
`CSL EXHIBIT 1015
`CSL v. Shire
`
`Page 1 of 50
`
`

`

`
`
`I.  
`
`II.  
`
`TABLE OF CONTENTS
`
`Page(s)
`
`INTRODUCTION .......................................................................................... 1  
`
`QUALIFICATIONS ....................................................................................... 1  
`
`III.   MATERIALS CONSIDERED ....................................................................... 3  
`
`IV.   SUMMARY OF OPINIONS .......................................................................... 3  
`
`V.  
`
`BACKGROUND AND STATE OF THE ART ............................................. 5  
`
`A.  
`
`B.  
`
`C.  
`
`Introduction to Protein Formulation ..................................................... 5  
`
`Subcutaneous Protein Formulations ................................................... 10  
`
`C1-INH ............................................................................................... 21  
`
`1.  
`
`2.  
`
`Chemical and Physical Properties ............................................ 21  
`
`The Literature Disclosed High Concentration
`Formulations for IM Administration, and SC
`Administration of Both Low and High Concentration
`Formulations ............................................................................ 24  
`
`VI.   The ’111 Patent ............................................................................................. 29  
`
`VII.   LEVEL OF ORDINARY SKILL IN THE ART .......................................... 33  
`
`VIII.   A POSA WOULD HAVE BEEN MOTIVATED TO INCREASE
`THE CONCENTRATION OF PRIOR SC C1-INH
`FORMULATIONS ....................................................................................... 33  
`
`IX.   A POSA WOULD HAVE HAD A REASONABLE EXPECTATION
`OF SUCCESS IN FORMULATING C1-INH FOR SC
`ADMINISTRATION AT A CONCENTRATION OF AT LEAST
`400U/mL OR 500U/mL ................................................................................ 37  
`
`X.  
`
`CONCLUSION ............................................................................................ 43  
`
`
`
`i
`
`
`
`Page 2 of 50
`
`

`

`I.  
`
`INTRODUCTION
`1.  
`
`I have been retained by Finnegan, Henderson, Farabow, Garrett &
`
`Dunner, LLP, on behalf of CSL Behring GmbH and CSL Behring LLC
`
`(collectively “CSL”) to provide my opinions in this proceeding based on my
`
`qualifications as a formulation scientist, chemical engineer, and expert in stability
`
`and physicochemical properties of therapeutic proteins.
`
`2.  
`
`I have been engaged at my customary hourly consulting rate of
`
`$350.00 per hour. My compensation is not contingent on the outcome of this
`
`proceeding.
`
`II.   QUALIFICATIONS
`3.  
`I hold the rank of Professor in the Department of Chemical and
`
`Biomolecular Engineering at the University of Delaware (UD), and I am a faculty
`
`member in the Center for Molecular and Engineering Thermodynamics, and the
`
`Chemistry-Biology Interface Program at UD. I have courtesy positions as a Guest
`
`Researcher at the National Institute of Standards and Technology, and as a Visiting
`
`Faculty member in the School of Chemical Engineering and Analytical Sciences at
`
`the University of Manchester.
`
`4.  
`
`At the University of Delaware, I serve as Director of the Center for
`
`Biomanufacturing Science and Technology, and Director of the Biomolecular
`
`1
`
`Page 3 of 50
`
`

`

`Interaction Technologies Center. Prior to my academic positions, I spent three
`
`years as a formulation scientist at Pfizer Global Research and Development.
`
`5.  
`
`Prior to my time in the pharmaceutical industry, I received my Ph.D.
`
`in Chemical Engineering from Princeton University in 1999, and my Bachelor’s
`
`degree in Chemical Engineering from the University of Delaware in 1994.
`
`6.  
`
`I serve on
`
`the Editorial Advisory Board of
`
`the Journal of
`
`Pharmaceutical Sciences, and the External Advisory Board for an EU-funded
`
`industry-university research consortium (acronym, PIPPI, for Protein-Excipient,
`
`Protein-Protein Interactions consortium). I am also an active member of various
`
`professional organizations in the field of protein formulation, including the
`
`American Association of Pharmaceutical Scientists (AAPS) and the American
`
`Chemical Society.
`
`7.  
`
`I am the recipient of multiple awards and plenary / keynote
`
`lectureships at academic, industrial, and regulatory institutions, and I have given
`
`over 120 invited lectures. I have authored or co-authored over 75 publications in
`
`peer-reviewed journals, as well as five book chapters, and one book, in the fields of
`
`protein stability, protein interactions, and physicochemical properties of protein
`
`solutions.
`
`8.  
`
`I have studied the quantitative prediction, design, and control of the
`
`physical stability of proteins and their solution properties for over 17 years. My
`
`2
`
`Page 4 of 50
`
`

`

`research combines experiments, modeling, and molecular
`
`theory across
`
`biochemical engineering, applied biophysics, and pharmaceutical sciences to
`
`address a range of topics that include: protein misfolding and aggregation; protein-
`
`protein and protein-excipient
`
`interactions; proteins at
`
`interfaces; high-
`
`concentration behavior and stability; and protein engineering to create improved
`
`product properties.
`
`9.  
`
`A copy of my curriculum vitae is provided as Exhibit 1018.
`
`III.   MATERIALS CONSIDERED
`10.  
`In preparing this declaration, I have relied on my extensive experience
`
`in protein formulation sciences, and specifically my experience in developing
`
`formulations of protein drugs for parenteral administration through my time
`
`working in the pharmaceutical industry and my extensive industry collaborations
`
`and technical consulting roles since joining the University of Delaware. I have
`
`also considered the materials listed in Appendix A.
`
`IV.   SUMMARY OF OPINIONS
`11.  
`I have been asked to provide an opinion on the state of the art in
`
`formulating protein drugs, and specifically formulations for subcutaneous (“sc”)
`
`administration of protein drugs, as of March 2013. At the time, the state of the art
`
`in formulating protein drugs was well-developed and approaches for dealing with
`
`3
`
`Page 5 of 50
`
`

`

`issues of stability, aggregation, high viscosity, and phase behavior were well
`
`known.
`
`12.  
`
`I have also been asked to review information that was publicly
`
`available as of March 2013 regarding the C1 esterase inhibitor protein (termed
`
`“C1-INH” hereafter). At the time, C1-INH was well-characterized, and I am aware
`
`of no reports of any solubility, viscosity, or practical limitations for developing a
`
`concentrated C1-INH formulation for sc administration.
`
`13.  
`
`It is my opinion that in light of this well-developed field and well-
`
`characterized protein, a person of ordinary skill in the art (“POSA”), in March
`
`2013, would have had a reasonable expectation of success in formulating C1-INH
`
`at a concentration of at least 400U/mL, including at least 500U/mL, for sc
`
`administration based on routine experimentation that was typical of formulation
`
`development activities in the field.
`
`14.  
`
`I also specifically considered Dr. Schranz’s assertions, which were not
`
`supported with any citations, that it is difficult or impossible to formulate protein
`
`drugs like Cinryze® for sc administration, and that C1-INH has certain
`
`characteristics that posed unique challenges in developing a high concentration sc
`
`formulation, specifically with respect to physical and chemical instability, as well
`
`as high solution viscosity. The reasons Dr. Schranz cited for these perceived
`
`difficulties included the large size of the protein and its high level of glycosylation.
`
`4
`
`Page 6 of 50
`
`

`

`As explained below, Dr. Schranz’s assertions are incorrect both from a general
`
`perspective of protein formulation science and protein physical chemistry, and
`
`because they do not reflect the state of the art with respect to C1-INH as of March
`
`2013.
`
`V.   BACKGROUND AND STATE OF THE ART
`A.  
`Introduction to Protein Formulation
`15.   Protein drugs are almost invariably administered parenterally (i.e.,
`
`non-orally) due to their susceptibility to chemical degradation by the proteases of
`
`the digestive system (Ex. 1060 [Sola, p. 1229], 13), issues with how quickly they
`
`are processed in vivo, and because there is very little intracellular transport of
`
`protein molecules in the gut.
`
` The most common modes of parenteral
`
`administration of protein drugs are: intravenous (iv), intramuscular (im), and
`
`subcutaneous (sc). Ex. 1006 [Gatlin pp. 417-418], 29-30.
`
`16.  
`
`Intravenous administration allows for greater formulation volumes
`
`than im and sc, which enables the use of lower protein concentrations to achieve a
`
`given dose. Intramuscular and sc administration require smaller volumes, usually
`
`on the order of a few milliliters, necessitating higher protein concentrations to
`
`achieve a given dose. Ex. 1006 [Gatlin, pp. 417-418], 29-30; see also Ex. 1019
`
`[Connolly, p. 69], 6; Ex. 1044 [Shire 2009, p. 709], 5. Because the requirements
`
`for developing an sc or im formulation are almost identical from the perspective of
`
`5
`
`Page 7 of 50
`
`

`

`protein aggregation/stability, interaction with key excipients and the product
`
`container, and the requirements for the viscosity of the injectable solution, a POSA
`
`would have understood that if one had a stable im formulation at a sufficient
`
`concentration, then that formulation could be used for sc administration, as also
`
`evidenced by the common formulations for sc and im formulations of marketed
`
`high-concentration products such as Subcuvia® and Subgam® (see Table 1).
`
`Table 1. Illustrative subset of marketed protein therapeutics as of 2013.
`
`“LYO” denotes a lyophilized product; “LIQ” denotes a liquid product; “Mab”
`
`denotes monoclonal antibody.
`
`Drug  name  
`
`MW  
`(kDa)  
`
`Protein  type,  
`liquid  v.  lyo  
`
`pH,  buffer*   Protein  
`conc.  
`
`Synagis  
`(Palivizumab)1  
`
`Herceptin  
`(Trastuzumab)2  
`Herceptin  SC3  
`
`148   MAb,  LYO  
`
`145.5   MAb,  LYO  
`
`Not  listed,  
`47mM  
`Histidine  
`6    
`
`145.5   MAb,  LIQ  
`
`~  6  
`
`100  
`mg/mL  
`
`21  
`mg/mL  
`140  
`mg/mL  
`
`Vivaglobin4  
`
`Not  
`listed  
`
`Immunoglobulins  
`from  plasma,  
`LIQ  
`
`6.4-­7.2  
`
`160  
`mg/mL  
`
`Other  excipients   SC,  IM,  
`or  IV  
`delivery  
`IM  
`(0.5  to  1  
`mL)  
`IV  
`
`3mM  glycine,  5.6  
`%  mannitol  
`
`Trehalose,  
`histidine,  PS20  
`Trehalose,  
`histidine,  PS20,  
`methionine,  
`vorhyaluronidase    
`alfa  
`  
`Glycine,  NaCl  
`
`Approval  
`yr  US  
`
`1998  
`
`1998  
`
`2013**  
`
`2006  
`
`2002,  
`2011  
`
`2003,  
`2015  
`
`2008  
`
`SC  (5mL  
`per  site)  
`
`SC  
`(max  vol  
`=  15mL  
`per  site)  
`SC  (0.4  
`to  0.8  
`mL)  
`
`SC  (1.2  
`mL)  
`
`SC  (1  
`mL),  IV  
`
`Humira  
`(Adalimumab)5  
`
`148   MAb,  LIQ  
`
`149   MAb,  LYO  
`
`Xolair  
`(Omalizumab)6  
`  
`Procrit  (Epoetin  
`alpha)7  
`
`5.2,  
`sodium  
`phosphate  
`/  citrate  
`Not  listed,  
`histidine  
`
`50  
`mg/mL  
`
`125  
`mg/mL  
`
`6.16mg/mL  
`NaCl,  1.2%  
`mannitol,  0.1%  
`PS80  
`Sucrose,  PS20  
`
`30.4   Glycoprotein,  
`LIQ  
`
`Not  listed,  
`sodium  
`
`<  0.4  
`mg/mL  
`
`NaCl,  sodium  
`phosphate,  
`
`6
`
`Page 8 of 50
`
`

`

`Neupogen  
`(Filgrastim)8  
`
`18.8   G-­CSF,  LIQ  
`
`Enbrel  
`(Etanercept)9  
`
`~150   TNFR-­Fc  fusion,  
`LIQ  
`
`Enbrel  
`(Etanercept)9  
`
`~150   TNFR-­Fc  fusion,  
`LYO  
`
`citrate  
`4,  sodium  
`acetate  
`
`6.1-­6.5,  
`sodium  
`phosphate  
`7.1-­7.7,  
`not  listed  
`
`Actimmmune  
`(Interferon  
`gamma-­1b)10  
`Intron  A  
`(interferon  
`alpha-­2b)11  
`Subcuvia12  
`
`~  33  
`
`19.3  
`
`Interferon  
`gamma  dimer,  
`LIQ  
`Interferon,  LYO  
`
`>100  
`
`Not  listed,  
`sodium  
`succinate  
`Not  listed,  
`sodium  
`phosphate  
`Not  listed  
`
`Subgam13  
`
`>100  
`
`Not  listed  
`
`Immunoglobulin  
`mixture  (IgG1,  
`IgG2,  IgG3,  
`IgG4),  LIQ  
`Immunoglobulin  
`mixture  (IgG1,  
`IgG2,  IgG3,  
`IgG4),  LIQ  
`Immunoglobulin  
`mixture  (IgG1,  
`IgG2,  IgG3,  
`IgG4),  LIQ  
`Immunoglobulin  
`mixture  (IgG1,  
`IgG2,  IgG3,  
`IgG4),  LIQ  
`Zinc,  m-­cresol,  
`3.6  
`Human  insulin,  
`glycerol    
`mg/mL  
`LIQ  
`*for LYO products, pH and buffer type/concentration is post reconstitution
`
`Hizentra14  
`
`>100  
`
`Gammanorm15  
`
`>100  
`
`4.6  –  5.2  
`
`200  
`mg/mL  
`
`Proline,  PS80  
`
`Not  listed  
`
`165  
`mg/mL  
`
`NaCl,  sodium  
`acetate,  PS80  
`
`Humulin  R16,17  
`
`6  
`
`7  –  7.8  
`
`0.3  
`mg/mL  
`
`50  
`mg/mL  
`
`25  
`mg/mL  
`
`0.2  
`mg/mL  
`
`Albumin  
`Sorbitol,  PS80  
`
`SC  
`(<1.5mL),  
`IV  
`SC  
`(<1mL)  
`
`1  %  sucrose,  
`100mM  NaCl,  
`25mM  Arg-­HCl  
`Mannitol,  
`sucrose,  
`tromethamine  
`Mannitol,  PS20,     SC  
`(<1mL)  
`
`SC  
`(<1mL)  
`
`<  0.2  
`mg/mL  
`
`NaCl,  EDTA,  m-­
`cresol,  PS80  
`
`160  
`mg/mL  
`
`Glycine,  NaCl  
`
`160  
`mg/mL  
`
`Glycine,  sodium  
`acetate,  NaCl,  
`PS80  
`
`1991  
`
`1998  
`
`1998  
`
`1990  
`
`1995  
`
`2003  
`
`2004  
`
`2010  
`
`SC  (1  
`mL),  IV,  
`IM  
`SC  (10  
`mL/hr),  
`IM    
`
`SC  (10  
`mL/hr),  
`IM  
`
`SC  (15  
`mL/hr)  
`
`SC  (10  
`mL/hr)  
`
`2008  
`
`SC    
`
`1982  
`
`** the product is administered with an enzyme that reduces the pain upon injecting a higher
`
`volume by SC; the product was approved in 3Q2013, but publications and a patent were
`
`available prior to that which enabled the marketed product
`1  https://www.accessdata.fda.gov/drugsatfda_docs/label/2002/palimed102302LB.pdf
`2  https://www.accessdata.fda.gov/drugsatfda_docs/label/2010/103792s5250lbl.pdf
`
`3A  http://www.medsafe.govt.nz/consumers/cmi/h/herceptinsc.pdf  
`
`3B  https://www.thieme-­connect.com/products/ejournals/html/10.1055/s-­0032-­1321831  
`
`7
`
`Page 9 of 50
`
`

`

`3C  https://www.google.com/patents/US20110044977  
`
`4  https://www.fda.gov/downloads/Biolog...ionatedPlasmaProducts/ucm070585.pdf  
`
`5  https://www.accessdata.fda.gov/drugsatfda_docs/label/2011/125057s0276lbl.pdf  
`
`6  https://www.accessdata.fda.gov/drugsatfda_docs/label/2016/103976s5225lbl.pdf  
`
`7  https://www.accessdata.fda.gov/drugsatfda_docs/label/2008/103234s5196PI.pdf  
`
`8  https://www.accessdata.fda.gov/drugsatfda_docs/label/1998/filgamg040298lb.pdf  
`
`9  https://www.accessdata.fda.gov/drugsatfda_docs/label/2015/103795s5548lbl.pdf  
`
`10  https://www.accessdata.fda.gov/drugsatfda_docs/label/2015/103836s5182lbl.pdf  
`
`11  https://www.accessdata.fda.gov/drugsatfda_docs/label/2015/103132s5191lbl.pdf  
`
`12  https://www.medicines.org.uk/emc/medicine/30223  
`
`13https://www.medicines.org.uk/emc/medicine/14826/SPC/Subgam,+Human+normal+immunogl
`
`obulin+solution/  
`
`14https://www.fda.gov/downloads/BiologicsBloodVaccines/BloodBloodProducts/ApprovedProduc
`
`ts/LicensedProductsBLAs/FractionatedPlasmaProducts/UCM203150.pdf  
`
`15  https://www.medicines.org.uk/emc/medicine/23491  
`16 https://www.accessdata.fda.gov/drugsatfda_docs/label/2011/018780s120lbl.pdf  
`17 Multiple insulin analogues were marketed prior to 2013 under very similar if not
`
`the same product conditions as Humulin;
`
`https://www.accessdata.fda.gov/drugsatfda_docs/label/2013/020563s115lbl.pdf;
`
`https://www.accessdata.fda.gov/drugsatfda_docs/label/2015/020986s082lbl.pdf;;  
`
`https://www.accessdata.fda.gov/drugsatfda_docs/label/2008/021629s015lbl.pdf;;  
`
`https://www.accessdata.fda.gov/drugsatfda_docs/label/2009/021081s034lbl.pdf;;  
`
`https://www.accessdata.fda.gov/drugsatfda_docs/label/2012/021536s037lbl.pdf  
`
`8
`
`Page 10 of 50
`
`

`

`17.   A primary goal of protein formulation is to mitigate the physical and
`
`chemical instabilities of proteins. Ex. 1066 [Wang, pp. 2-3], 7-8. As a result, all
`
`protein drugs require some formulation screening to assure sufficient solubility,
`
`limited chemical and physical instabilities, and acceptable practical properties,
`
`such as solution viscosity. By March 2013, however, the factors affecting protein
`
`stability were well-established, and well-developed techniques for stabilizing
`
`protein solutions and reducing undesirable properties, such as high viscosity, were
`
`well documented for developing protein drug formulations. See, e.g., Ex. 1066
`
`[Wang, Tables 1, 3], 13, 19; Ex. 1019 [Connolly, p. 69], 6; Ex. 1044 [Shire 2009,
`
`pp. 708-09], 4-5; Ex. 1067 [Yadav, pp. 1974, 1982], 6, 14; Ex. 1060 [Sola, Table
`
`1, p. 1225], 8-9.
`
`18.   For instance, the literature included numerous examples of excipients
`
`and additives, such as sugars, polyols, amino acids, amines, salts, polymers, and
`
`surfactants that could be included in a formulation to reduce or prevent protein
`
`aggregation and improve stability. Ex. 1066 [Wang, p. 20, Table 3], 25, 19-22.
`
`Similarly, it had been well-established that appropriate buffers, pH, and
`
`excipients/additives could be selected to reduce solution viscosity. Ex. 1019
`
`[Connolly, Figure 2], 11; Ex. 1067 [Yadav, Figures 3, 5, 7], 10-11, 14. By
`
`following such guidance, appropriate formulation conditions could be selected as a
`
`matter of routine experimentation that was typical of formulation development.
`
`9
`
`Page 11 of 50
`
`

`

`19.   Common buffers (weak acids and bases) used
`
`in parenteral
`
`formulations of protein drugs include: acetic acid, ascorbic acid, benzoic acid,
`
`citric acid, diethanolamine, glutamic acid, glycine, phosphoric acid, succinic acid,
`
`triethanolamine, and tromethamine. Ex. 1006 [Gatlin, Table 17.3], 20. As part of
`
`identifying a desired pH for a formulation, an appropriate buffer is selected to
`
`maintain that pH. Ex. 1006 [Gatlin, p. 408], 20. While the choice of buffer, buffer
`
`concentration, and solution pH can affect injection pain by contributing to the
`
`tonicity of the final product (Ex. 1006 [Gatlin, p. 409], 21), buffers typically have
`
`little impact on the therapeutic efficacy of the protein drug itself.
`
`B.  
`Subcutaneous Protein Formulations
`20.   As discussed above, sc formulations are typically prepared at lower
`
`volumes than iv formulations. A primary goal in developing an sc formulation is to
`
`reduce the volume and number of injections to maximize therapeutic efficacy and
`
`patient convenience and compliance. Ex. 1006 [Gatlin, p. 405], 17. Lower
`
`injection volumes reduce patient discomfort and leakage of the formulation from
`
`the site of administration. Ex. 1006 [Gatlin, p. 417], 29; see also generally Ex.
`
`1061 [’432 patent, 1:29-39], 9.
`
`21.   Although target volumes of less than 2mL are typically cited for sc
`
`administration (e.g., Ex. 1006 [Gatlin, p. 417], 29; Ex. 1019 [Connolly, p. 69], 6),
`
`larger volumes (e.g., 5mL) can also be injected. Ex. 1036 [Clinician’s Pocket
`
`10
`
`Page 12 of 50
`
`

`

`Reference, p. 277], 302. For example, in March 2012, Roche announced that its
`
`Herceptin® product would be administered as a 5mL sc injection in combination
`
`with a hyaluronidase enzyme.1 Ex. 1056 [Roche press release, p. 2], 2. In
`
`addition, even larger volumes (e.g., 3-20mL) had been administered by “push
`
`method” injections (e.g., Ex. 1024 [Jolles, p. 8]), and subcutaneous infusions had
`
`also been used to administer larger volumes. Ex. 1005 [Jiang 2010, p. 324], 9; Ex.
`
`1047 [Martinez-Saguer abstract], 3; Ex. 1064 [Vivaglobin® label, p. 2], 2.
`
`However, although subcutaneous administration of larger volumes is known, as
`
`discussed above, a primary goal in developing a sc formulation is to reduce the
`
`injection volume.
`
`22.   The desire to deliver lower volumes via sc administration typically
`
`requires the development of higher-concentration formulations to maintain an
`
`appropriate dose of the protein drug. Ex. 1044 [Shire 2009, p. 709], 5.
`
`Formulation scientists typically measure concentration in terms of mass-per-
`
`volume (e.g., mg/mL). This is in part due to requirements for manufacturing
`
`specifications, and in part due to the fact that physical properties and behaviors of
`
`
`1 Hyaluronidases increase tissue permeability by catalyzing the reversible
`
`degradation of hyaluronan, thereby enhancing the dispersion and delivery of
`
`subcutaneously-administered drugs.
`
`11
`
`Page 13 of 50
`
`

`

`protein solutions such as solubility, viscosity, and aggregation are more easily
`
`generalized between different proteins when the data are reported on a mass-per-
`
`volume basis instead of molar basis. The appropriate concentration range for a
`
`given protein drug will depend on the individual molecule: all other things being
`
`equal, larger molecules typically require a larger mass-per-volume to achieve a
`
`sufficient number of molecules to reach a target efficacy, while smaller molecules
`
`typically require a smaller mass-per-volume to reach the same target for the
`
`number of molecules per dose, as illustrated schematically in Figure 1.
`
`23.   Factors that affect the ability to increase the concentration of a protein
`
`formulation include protein solubility, its tendency to aggregate, and the solution
`
`viscosity; all of which depend on the chosen formulation for a given protein. Ex.
`
`1066 [Wang, pp. 2-3], 7-8; Ex. 1044 [Shire 2009, pp. 709, 712-13], 5, 8-9; Ex.
`
`1060 [Sola, Table 1, p. 1225], 8-9. In terms of these factors, the following basic
`
`concepts were known to apply to protein formulation in March 2013:
`
`a.   Solubility: the term solubility typically refers to the maximum
`
`concentration of protein that can be dissolved in a given liquid, at a given
`
`solution condition (e.g.,
`
`temperature, pH,
`
`ionic strength), before
`
`macroscopic phase separation occurs. The process of phase separation
`
`represents a form of reversible protein aggregation, which is to be
`
`distinguished from irreversible aggregate formation that is often simply
`
`12
`
`Page 14 of 50
`
`

`

`referred to as “aggregation” in the pharmaceutical stability literature (see
`
`below). Ex. 1057 [Weiss, p. 1246], 7. If a protein has a particular
`
`solubility threshold, it will be difficult to concentrate the formulation past
`
`that
`
`threshold without significant changes
`
`in solution pH, salt
`
`concentration, or the chemical identity of the salt ions. However,
`
`glycosylation had been shown to increase the solubility of many proteins,
`
`including Shire’s alpha-galactosidase A (Replagal®) product. Ex. 1060
`
`[Sola, p. 1231], 15.
`
`b.   Tendency to aggregate: as concentration increases, the possibility of
`
`physical interactions between protein molecules increases, leading to an
`
`increased possibility of protein aggregation, all other factors being equal.
`
`FIGURE 1. Schematic illustration of the difference in mass-per-
`
`volume concentrations (LEFT: low concentration, RIGHT: high
`
`
`
`13
`
`Page 15 of 50
`
`

`

`concentration) for the same number of protein molecules (gray
`
`circles) per unit volume (i.e., the same molar concentration).
`
`Irreversible protein aggregation is mediated by changes in three-
`
`dimensional conformation of the individual protein chains – i.e., so-
`
`called conformational stability. See generally Ex. 1057 [Weiss, p. 1257],
`
`18. At a given concentration (mass of protein per unit volume), the
`
`likelihood of aggregation is expected to be higher for smaller proteins
`
`than
`
`larger proteins, all other factors being equal, because
`
`the
`
`hydrophobic and hydrophilic driving forces for folding are roughly
`
`proportional to the molecular weight and amount of molecular surface
`
`area that is sequestered from contact with water. Ex. 1033 [Myers, p.
`
`2141, 4143], 4, 6. That is, from the perspective of the size or molecular
`
`weight of a protein, larger proteins tend to be less prone to unfolding and,
`
`therefore, less prone to aggregation. As discussed above, the literature
`
`included numerous examples of excipients and additives, such as sugars,
`
`polyols, amino acids, organic and inorganic salts more generally,
`
`polymers, and surfactants that could be included in a formulation to
`
`reduce or prevent protein aggregation on short or long time scales needed
`
`for administration and storage of protein therapeutics. Ex. 1066 [Wang,
`
`p. 20, Table 3], 25, 19-22. In addition, several reports had demonstrated
`
`14
`
`Page 16 of 50
`
`

`

`that protein glycosylation reduces, and in some cases prevents protein
`
`aggregation. Ex. 1060 [Sola, p. 1231], 15. Finally, lyophilization
`
`(freeze-drying) and spray-drying were common methods to “encapsulate”
`
`proteins in pharmaceutically benign solid matrices that can arrest
`
`aggregation during long-term storage. The solid is then easily and
`
`rapidly dissolved in sterile water prior to administration to the patient
`
`(e.g., via sc, im, or iv dosing). In those cases, the solution-state stability
`
`of the protein is mostly irrelevant, except in so far as the protein is
`
`sufficiently stable on the time scale of hours (i.e., during dissolution and
`
`the time for injection), or at most days if it is a multi-use product.
`
`c.   Viscosity: the possibility of increased physical interactions between
`
`protein molecules at increased concentration can also lead to increased
`
`solution viscosity. 50 millipascal-seconds (mPa×s)2 was considered a
`
`reasonable value for the maximum acceptable viscosity for adequate
`
`syringeability of sc and im formulations as of March of 2013, although
`
`there was no general requirement in that regard. Ex. 1061 [’432 patent,
`
`
`2 One millipascal-seconds (mPa×s) is equal to a centipoise (cP). The related
`
`quantity, kinematic viscosity (cs), is simply the dynamic viscosity divided by the
`
`density of the liquid.
`
`15
`
`Page 17 of 50
`
`

`

`2:56-60], 9. Globular proteins (i.e., proteins having a roughly spherical
`
`three-dimensional shape) exhibit qualitatively similar concentration-
`
`viscosity curves:
`
`
`
`
`
`
`
`
`
`FIGURE 2. Viscosity of different forms of hemoglobin (MW ~ 64 kDa) as a
`
`function of protein concentration at 25°C and approximately neutral pH. Adapted
`
`from Ex. 1049 [Monkos 1994, Figure 1], 4.
`
`16
`
`Page 18 of 50
`
`

`

`
`
`
`
`
`
`
`
`FIGURE 3. Viscosity of ovalbumin (MW ~ 45 kDa) as a function of protein
`
`concentration at 25°C (black circles) and 20 °C (blue squares). Adapted from Ex.
`
`1052 [Monkos 2000, Figure 1], 7.
`
`Such proteins typically exhibit viscosities well under 50mPa×s, even at
`
`higher protein concentrations of 200-300mg/mL. See id. For instance,
`
`BSA, which has a molecular weight of 66.5 kDa, exhibits the following
`
`viscosity-concentration profiles at various pH values:
`
`17
`
`Page 19 of 50
`
`

`

`Ex. 1067 [Yadav, Figure 5], 11.
`
`
`
`As can be seen, BSA exhibits a viscosity between 5mPa×s and 35mPa×s at
`
`concentrations between 250mg/mL and 300mg/mL and at pH values
`
`between 4.0 and 7.0. Similar behavior was known for other globular
`
`proteins as a function of typical formulation variables such as solution
`
`pH. Ex. 1061 [’432 patent, 2:46-52], 9.
`
`In contrast, monoclonal antibodies, which are non-globular, Y-shaped
`
`proteins with much more extended molecular structures, are more prone
`
`to exhibit higher viscosities at lower concentrations than globular
`
`proteins. See, e.g., Ex. 1019 [Connolly, p. 69], 6; Ex. 1044 [Shire 2009,
`
`p. 709], 5. For instance, as shown in the profiles below, some
`
`monoclonal
`
`antibodies
`
`exhibit viscosities
`
`above 50mPa×s
`
`at
`
`18
`
`Page 20 of 50
`
`

`

`concentrations somewhat below 150mg/mL. See Ex. 1044 [Shire 2009,
`
`Figure 1, 709], 5.
`
`
`
`However, as discussed above there were a number of formulation
`
`approaches, including the use of appropriate buffers and excipients, and
`
`the choice of solution pH, that had been shown to greatly decrease
`
`solution viscosity of
`
`formulated proteins, allowing
`
`for higher
`
`concentrations to be easily accessed simply by adjusting solution
`
`conditions within commonly accepted limits for formulated products.
`
`Id.; see also Ex. 1019 [Connolly, Figure 2], 11; Ex. 1067 [Yadev, Figures
`
`3, 5, 7], 10-11, 14. Indeed, U.S. Patent No. 6,875,432 demonstrated that
`
`solution viscosity could be reduced by adjusting pH within accepted
`
`ranges for pharmaceutical products. Ex. 1061 [’432 patent, 2:49-52], 9.
`
`19
`
`Page 21 of 50
`
`

`

`24.  
`
`It was also well known in March 2013 that glycosylation is expected
`
`to increase the overall stability of proteins. Ex. 1060 [Sola, pp. 1223, 1225], 7, 9.
`
`A number of studies had shown that glycosylation can lead to enhanced stability
`
`and therapeutic efficacies for protein pharmaceuticals. Ex. 1060 [Sola, p. 1225], 9.
`
`In terms of chemical stability, glycosylation had been shown to prevent the
`
`formation of disulfide and non-disulfide linkages, and also to reduce protein
`
`oxidation. Ex. 1060 [Sola, p. 1230], 14. And in terms of physical stability, as
`
`discussed above, glycosylation had been shown to increase protein solubility, and
`
`to reduce or prevent protein aggregation. Ex. 1060 [Sola, p. 1231], 15. In
`
`addition, glycosylation had been shown to prevent pH denaturation, improve
`
`thermal stability, and increase the long-term stability of proteins. Ex. 1060 [Sola,
`
`p. 1225, 1231, 1234], 9, 15, 18; see also generally Ex. 1081 [Latypov].
`
`25.   As can be seen from the foregoing discussion, the state of the art in
`
`March 2013 relating to formulating protein drugs for sc administration was well-
`
`developed and reasonably established. In fact, a number of protein drugs having a
`
`wide range of molecular weights and concentrations had been formulated for sc
`
`administration, and a number of marketed drugs utilized the same formulation for
`
`sc and im products, as well as for use in iv administration. See Table 1, supra.
`
`20
`
`Page 22 of 50
`
`

`

`C.   C1-INH
`1.  
`Chemical and Physical Properties
`26.   C1-INH is a serine protease inhibitor (“serpin”) that is used to treat
`
`hereditary angioedema (“HAE”). Ex. 1028 [Over, p. 241], 3. C1-INH is a
`
`medium-sized protein, having a molecular weight of ~76kDa by neutron scattering
`
`(Ex. 1054 [Perkins, p. 751], 5), ~100kDa by analytical ultracentrifugation (Ex.
`
`1053 [Nilsson, P. 275], 9]), and ~105kDa by sodium dodecyl sulfate
`
`polyacrylamide gel electrophoresis (SDS-PAGE) (Ex. 1042 [Harrison, p. 5001], 5).
`
`27.   The C1-INH protein has two domains: (1) a C-terminal domain,
`
`which is a typical serpin domain; and (2) a non-serpin N-terminal domain. Ex.
`
`1028 [Over, p. 241], 3. A crystal structure of the C-terminal serpin domain was
`
`published in 2007, which revealed typical globular protein folding. Ex. 1028
`
`[Over, p. 241, Figure 17.1 (citing n.14)], 3-4, 14.
`
`28.   C1-INH is considered a heavily glycosylated protein, with 13
`
`glycosylation sites. Ex. 1028 [Over, p. 241], 3. The carbohydrates are unevenly
`
`distributed over the molecule: three are associated with the C-terminal serpin
`
`domain, and 10 are associated with the N-terminal domain. Id.
`
`29.  
`
`I am not aware of any references reporting a solubility threshold for
`
`C1-INH as of March 2013. The lack of any reported solubility limitation is not
`
`surprising, since the high levels of glycosylation would have been expected to
`
`21
`
`Page 23 of 50
`
`

`

`improve C1-INH’s solubility, as discussed above. Thus, a POSA would not have
`
`expected solubility to be a limiting factor for developing a concentrated sc
`
`formulation of C1-INH, given that examples already existed in the literature where
`
`elevated concentrations (e.g., 333U/mL) had already been administered. Ex. 1004
`
`[Schranz poster], Figure 1. And this is consistent with the studies provided in the
`
`’111 patent, which “demonstrated that there is not a solubility limit to preparing
`
`C1-INH at concentrations up to 500 U/ml.” Ex. 1000 [’111 patent, 10:35-36], 11.
`
`30.  
`
`I am likewise unaware of any reports prior to March 2013 of C1-INH
`
`having a particularly high tendency to aggregate, such that it would be difficult if
`
`not impossible to achieve concentrated formulations for sc administration. While
`
`all plasma-purified proteins have some tendency to aggregate, the size (i.e.,
`
`molecular weight) of a protein has little correlation with its tendency to aggregate.
`
`For example, some of the most aggregation-prone therapeutic proteins are
`
`relatively small (e.g, ~5 – 20 kDa for insulin and G-CSF), while large proteins
`
`(~150 kDa) such as monoclonal antibodies (which are also glycoproteins) are
`
`much more stable. In addition, as discussed above, C1-INH’s high glycosylation
`
`would have been expected to reduce or prevent protein aggregation. Thus, a POSA
`
`would not have expected aggregation to be a limiting factor f