IN THE UNITED STATES PATENT AND TRADEMARK OFFICE
`
`Trial Number: IPR2017-00805
`
`
`
`
`
`
`
`In the Inter Partes Review of:
`
`U.S. Patent No. 7,371,379
`Filed:
`June 20, 2003
`Issued:
`May 13, 2008
`Inventor(s): Sharon Baughman, Steven Shak
`Assignee: Genentech, Inc.
`Title:
`Dosages for Treatment with Anti-
`
`
`ErbB2 Antibodies
`_________________________________________________________________
`Mail Stop Inter Partes Review
`
`Commissioner for Patents
`P.O. Box 1450
`Alexandria, VA 22313-1450
`
`
`
`
`DECLARATION OF GEORGE GRASS, Ph.D.
`
`
`
`
`
`Genentech 2039
`Hospira v. Genentech
`IPR2017-00805
`
`

`

`I.
`II.
`III.
`IV.
`V.
`
`TABLE OF CONTENTS
`Introduction ...................................................................................................... 1
`Qualifications ................................................................................................... 1
`Summary of Opinions ...................................................................................... 3
`Person of Ordinary Skill in the Art .................................................................. 6
`Dr. Jusko’s Calculations and Predictions Are Based on Incorrect
`Assumptions that Contradict the Prior Art ...................................................... 7
`A Skilled Artisan Would Not Model Trastuzumab as if It Had Dose-
`Independent, Linear Kinetics ....................................................................... 7
`The Prior Art Taught that Trastuzumab Had Dose-Dependent, Non-Linear
`Pharmacokinetics ....................................................................................... 14
`Dr. Jusko’s Computations and Predictions Are Inconsistent with the
`Teaching in the Prior Art ............................................................................ 16
`VI. A Skilled Artisan Would Not Rely on Dr. Jusko’s Erroneous Assumptions to
`Predict the Efficacy of Three-Week Dosing ................................................. 18
`Maintaining Adequate Serum Trough Concentrations Is Important in
`Predicting the Efficacy of a Dose Regimen for Trastuzumab ................... 18
`A Skilled Artisan Would Conclude that Dr. Jusko’s Calculations Would
`Likely Overestimate Trough Serum Concentrations ................................. 20
`VII. The Prior Art Did Not Provide Data Sufficient to Predict the Efficacy of a
`Three-Week Dosing Interval for Trastuzumab .............................................. 24
`A Skilled Artisan Would Not Have Relied on the Prior Art’s Half-Life
`Data to Predict the Efficacy of a Three-Week Dosing Interval ................. 24
`The Prior Art Did Not Provide Enough Data from Which a Skilled Artisan
`Could Construct an Accurate Pharmacokinetic Prediction of the Efficacy
`of a Three-Week Dosing Interval of Trastuzumab .................................... 28
`VIII. A Skilled Artisan Would Not Have Been Motivated to Try Three-Week
`Dosing Based on the Teachings of the 1998 Herceptin® Label, Baselga ’96,
`and Pegram ’98 .............................................................................................. 33
`
`- i -
`
`

`

`
`
`
`
`
`
`The Limited Pharmacokinetic Data Available Would Not Have Motivated
`a Skilled Artisan to Try Three-Week Dosing ............................................ 33 
`The Prior Art Teaches that Shed Antigen and Tumor Burden Affect the
`Pharmacokinetics of Trastuzumab ............................................................. 34 
`
`- ii -
`
`

`

`
`
`I.
`1.
`
`Introduction
`I have been asked to review and respond to the opinions set forth in the
`
`January 20, 2017 Declaration of William Jusko, PhD.
`
`II. Qualifications
`2.
`I am President of G2 Research, Inc., a company I founded in August 2001 to
`
`provide consulting services to pharmaceutical and biotechnology companies in a
`
`variety of areas. Among other things, I have performed pharmacokinetic modeling
`
`to evaluate clinical regimens for antibodies and small molecules. I have also
`
`developed computer simulation software and models to predict drug
`
`pharmacokinetics.
`
`3.
`
`I obtained a Ph.D. in Pharmaceutics at the University of Wisconsin, Madison
`
`in 1985. My Ph.D. thesis was entitled “Mechanisms of Corneal Drug Penetration.”
`
`As a result of this research, I was the co-recipient of the 1989 Ebert Prize, awarded
`
`by the American Pharmacists Association Academy of Pharmaceutical Research
`
`and Sciences, for a series of manuscripts published in the Journal of
`
`Pharmaceutical Sciences entitled “Mechanisms of Corneal Drug Penetration.” I
`
`obtained a M.S. degree in Pharmaceutics at the University of Wisconsin, Madison
`
`in 1983. I obtained a Pharm. D. degree from the University of Nebraska in 1980,
`
`and was formerly licensed to practice pharmacy in the state of Nebraska.
`
`- 1 -
`
`

`

`
`
`4.
`
`I have spent more than 30 years working in the pharmaceutical industry.
`
`From 1985 to 1991, I worked as a Research Scientist at Syntex Research in Palo
`
`Alto, where I was responsible for formulation development and research in oral
`
`drug absorption, including methods to orally deliver peptides. Since 1991, I have
`
`been a pharmaceutical industry consultant. In 1991, I started my own company,
`
`Precision Instrument Design Inc., and, in 1997, another company, NaviCyte, Inc.
`
`In 1999, NaviCyte, Inc. was acquired by Trega Biosciences, and I served as Chief
`
`Technology Officer at Trega Biosciences, Inc. until 2001. In 2001, I founded G2
`
`Research, Inc., and also founded RaptorGraphics, Inc., a computer graphics and
`
`simulation business. From 2005 to 2007, I was Vice President of Product
`
`Development and Chief Technology Officer for PDxRx, Inc., a specialty-focused
`
`pediatrics company. From 2007 to 2010, I was Senior Vice President of Research
`
`and Development for Sorbent Therapeutics, Inc., a company developing novel
`
`polymer therapeutics for sodium fluid removal. From January 2016 until May
`
`2017, I was Senior Vice President of non-clinical development and founder for
`
`NeuroVia, Inc., a company developing a novel compound for childhood cerebral
`
`adrenoleukodystrophy.
`
`5.
`
`I am the author or co-author of more than 30 published scientific articles,
`
`primarily in the areas of models to predict drug pharmacokinetics, corneal
`
`permeability and drug transport, and intestinal transport and drug absorption. I
`
`- 2 -
`
`

`

`
`
`have authored book chapters related to drug delivery and have been an invited
`
`speaker at multiple scientific meetings, including meetings of the American
`
`Association of Pharmaceutical Scientists. I have been a peer reviewer for a
`
`number of journals such as Pharmaceutical Research and Journal of
`
`Pharmaceutical Sciences. I have presented technical information and development
`
`plans to the U.S. Food and Drug Administration (FDA). I am an inventor or co-
`
`inventor on eight U.S. patents and four additional U.S. patent applications and
`
`several foreign patents. My curriculum vitae is attached as Appendix A.
`
`III. Summary of Opinions
`6.
`The 1998 Herceptin® Label (“the Label”), Baselga ’96, and Pegram ’98 do
`
`not contain sufficient data with respect to trastuzumab to reasonably predict that a
`
`three-week dosing regimen, as recited in the claims of U.S. Pat. No. 7,371,379
`
`(“the ’379 patent”), would maintain therapeutically effective trough
`
`concentrations. Prediction of a dosing regimen for a drug such as trastuzumab
`
`would require additional data that was not reported in the prior art, such as serum
`
`concentration levels from initial doses through steady-state, or require conducting a
`
`“washout study” where serum concentration data is collected over a period of at
`
`least several half-lives.
`
`7.
`
`The prior art upon which Petitioner relies reported administering
`
`trastuzumab to patients as a 250 mg loading dose followed by weekly doses of
`
`- 3 -
`
`

`

`
`
`100 mg (Baselga ’96 and Pegram ’98) or as a 4 mg/kg loading dose followed by
`
`weekly doses of 2 mg/kg (the Label). Those doses were used in Phase II and Phase
`
`III studies and became the weekly dosing regimen subsequently approved by the
`
`FDA. The selection of this weekly dose regimen was based on much of the same
`
`Phase I data upon which Petitioner’s pharmacokinetics expert, Dr. Jusko, relies in
`
`his analysis.
`
`8.
`
`The Label, Baselga ’96, and Pegram ’98 also reported a number of different
`
`half-lives for trastuzumab, including 1.7 days, 5.8 days, and 12 days in the Label,
`
`1.8, 8.3, and 9.1 days in Baselga ’96, and 2.9, 4.0, 9.2, and 11.0 days in Pegram
`
`’98. The cited prior art further taught that trastuzumab had dose-dependent, i.e.,
`
`nonlinear kinetics—which a skilled artisan would have known to cause changes in
`
`serum concentration and therefore clinical effect. As the scope and extent of such
`
`non-linear effects could not be predicted from the limited data reported in the prior
`
`art, a skilled artisan would not have had a reasonable expectation of success in
`
`predicting that a three-week dosing regimen as claimed would be safe and
`
`effective. Specifically, none of this prior art would have suggested to a skilled
`
`artisan that three-week dosing would reliably achieve and maintain therapeutically
`
`effective trough concentrations.
`
`9.
`
`The prior art reported that 64% of patients had shed antigen present and that
`
`shed antigen had a direct effect on the half-life of trastuzumab. Given the high
`
`- 4 -
`
`

`

`
`
`prevalence of shed antigen and its documented ability to lower trastuzumab’s half-
`
`life, a skilled artisan would not have been motivated to try every-three-week
`
`dosing of trastuzumab as claimed in the ’379 patent.
`
`10. A skilled artisan would have not have been motivated to try the three-week
`
`dosing regimen of the claims of the ’379 patent based on the pharmacokinetic
`
`information reported in the Label, Baselga ’96, and Pegram ’98.
`
`11.
`
`I disagree with Dr. Jusko’s opinion that the calculations set forth in his
`
`declaration would have given a skilled artisan confidence that trastuzumab could
`
`safely and effectively treat breast cancer when administered every three weeks
`
`according to the regimens claimed in the ’379 patent.
`
`12. Dr. Jusko’s calculations with respect to trastuzumab are based upon
`
`assumptions that are unsupported by and indeed contrary to the explicit teachings
`
`of the prior art upon which Petitioner relies:
`
`First, Dr. Jusko’s calculations assume that trastuzumab exhibits linear
`kinetics, i.e. a single half-life across all dose amounts. (See, e.g., Ex. 1003,
`Jusko Decl. ¶¶60, 69.) This assumption is directly contrary to the explicit
`teaching in the prior art that trastuzumab exhibited dose-dependent, non-
`linear kinetics.
`
`Second, Dr. Jusko’s calculations assume that the half-life of trastuzumab is
`12 days, across all doses and dosing intervals; this single value for half-life,
`taken from the Label, is used throughout the entirety of his analysis. (See,
`e.g., Ex. 1003, Jusko Decl. ¶49.) Although the Label reported a 12-day half-
`
`- 5 -
`
`

`

`
`
`life for a 500 mg dose administered weekly in a Phase I study, it also
`reported a 5.8-day half-life when trastuzumab was administered weekly in a
`Phase III clinical trial and a wide range of half-lives for trastuzumab,
`ranging from 1.7 days with a 10 mg dose to 12 days with a 500 mg dose,
`when trastuzumab was administered weekly in Phase I studies. Dr. Jusko’s
`decision to use the half-life from a small Phase I study (which happened to
`be the longest reported half-life available) as the basis for all of his
`calculations is an unsupportable assumption that leads to an unreliable and
`incorrect analysis.
`
`13. A skilled artisan would know that calculations based on Dr. Jusko’s series of
`
`unsupported assumptions would not reliably predict the minimum trough serum
`
`concentrations of trastuzumab over a three-week dosing interval. Importantly,
`
`such calculations would likely overestimate trough serum concentrations and thus
`
`incorrectly make it appear as if therapeutically effective trough serum
`
`concentrations (the parameter noted as critical to efficacy in the prior art) would be
`
`maintained during a three-week dosing interval.
`
`IV. Person of Ordinary Skill in the Art
`14.
`I understand that the Petitioner and Dr. Jusko defined a person of ordinary
`
`skill in the art as a team including:
`
`(1) a clinical or medical oncologist specializing in breast cancer with several
`years of experience with breast cancer research or clinical trials, and
`
`- 6 -
`
`

`

`
`
`(2) a person with a Ph.D. in pharmaceutical sciences or a closely related
`field with an emphasis in pharmacokinetics with three years of relevant
`experience in protein based drug kinetics.
`
`(Ex. 1003, Jusko Decl. ¶15.)
`15. For the purpose of this declaration, I have applied this definition of a skilled
`
`artisan, and my opinions are offered from the perspective of a skilled artisan as that
`
`hypothetical person would have understood matters on August 27, 1999.
`
`V. Dr. Jusko’s Calculations and Predictions Are Based on Incorrect
`Assumptions that Contradict the Prior Art
`16. Dr. Jusko’s calculations incorrectly assume that trastuzumab has linear
`
`pharmacokinetics and exhibits a consistent 12-day half-life across all doses and
`
`dosing intervals. (See, e.g., Ex. 1003, Jusko Decl. ¶¶ 33-34, 60, 69.) As explained
`
`more fully below, these assumptions are contrary to the prior art’s teaching that
`
`trastuzumab has dose-dependent, non-linear pharmacokinetics across the doses that
`
`were tested, including the doses and plasma concentrations contemplated by Dr.
`
`Jusko. (See, e.g., Ex. 1008, Label at 1; Ex. 1013, Baselga ’96 at 10.)
`
` A Skilled Artisan Would Not Model Trastuzumab as if It Had
`Dose-Independent, Linear Kinetics
`17. To develop a dose regimen for any drug, including trastuzumab, a person of
`
`ordinary skill in the art would need to understand the pharmacokinetics of the drug
`
`at issue. Pharmacokinetics is the study of the absorption, distribution, metabolism,
`
`and elimination of drugs in the body over time, i.e., where the drug goes and the
`
`- 7 -
`
`

`

`
`
`rates at which it gets there. The kinetics of a drug affect its concentration at the
`
`site of drug action, the potential for toxic response, the onset and duration of
`
`effects, the dosing amounts and frequency of dosing, and the rates of metabolism
`
`(if any) and elimination.
`
`18. Of particular relevance here, the study of pharmacokinetics is a useful part
`
`of the understanding of the relationship between the concentration of drug in
`
`plasma and therapeutic effects. If the relationship between the blood plasma
`
`concentration and clinical safety and efficacy is well-understood, then it becomes
`
`possible to draw inferences about safety and efficacy based on anticipated plasma
`
`concentrations. (See, e.g., Ex. 1022 Vol. 1, Rowland ’95 at 16.)
`
`19.
`
`In the case of drugs with linear pharmacokinetics, plasma concentration
`
`levels change proportionally to dose amount and dose interval, such that the period
`
`of time for the drug serum concentration to decrease by half—i.e., the drug’s half-
`
`life—remains constant over time, regardless of the concentration of the drug in the
`
`serum (or plasma). (See Ex. 2006, Sarfaraz Niazi, Chapter 7: Pharmacokinetic
`
`Principles, in TEXTBOOK OF BIOPHARMACEUTICS AND CLINICAL
`
`PHARMACOKINETICS 141 (1979) at 143, 179; Ex. 1022 Vol. 3, Rowland ’95 at 108-
`
`09.) Such drugs are said to have “dose-independent” pharmacokinetics.
`
`- 8 -
`
`

`

`
`
`20. A drug’s half-life can be determined by the rate at which the drug is
`
`eliminated from the plasma. For linear (first order)1 kinetics a constant fraction of
`
`drug in the body is eliminated per unit of time. The drug concentration halves
`
`predictably according to fixed time intervals. Pharmacokineticists often refer to
`
`the rate at which a drug is eliminated in terms of the “elimination rate constant,”
`
`which is the fractional rate of drug removal from the plasma. (Ex. 1022, Vol. 1,
`
`Rowland ’95 at 34-35.) The elimination rate constant is defined by the rate of
`
`elimination divided by the amount of the drug in the body, such that the
`
`elimination rate constant will only remain constant (i.e., remain as a single,
`
`unchanged value) as long as the drug is eliminated from the body proportionally to
`
`the amount of drug that is in the body. (See id.)
`
`
`1 The term “first order” comes from chemistry, where it has classically been used
`
`to describe reaction kinetics. When doubling the concentration of reagents also
`
`doubles the reaction rate, the increase in rate is by a factor of 2 (2 to the first
`
`power, or 21). That “first power” gives rise to the term “first order.”
`
`- 9 -
`
`

`

`
`
`21. For drugs with linear pharmacokinetics, there is a linear relationship
`
`between rate of elimination and concentration.2 Half-life is constant and
`
`independent of concentration.
`
`22.
`
`In contrast to drugs with dose-independent pharmacokinetics, dose-
`
`dependent drugs exhibit non-linear pharmacokinetics, meaning that plasma
`
`concentrations do not change proportionally with dose or interval. (See Ex. 2006,
`
`Niazi at 179; Ex. 1022 Vol. 3, Rowland ’95 at 108-09.) Half-life can change with
`
`changes in concentration. (Ex. 2008, Johan Gabrielsson & Daniel Weiner, Chapter
`
`3: Pharmacokinetic Concepts, in PHARMACOKINETIC AND PHARMACODYNAMIC
`
`DATA ANALYSIS 58 (2d ed. 1997) at 123.)
`
`
`2 More specifically, for a drug with linear pharmacokinetics, the rate at which the
`
`drug is eliminated from the body will always change in proportion to the amount of
`
`drug in the plasma such that the elimination constant will remain a constant value.
`
`In contrast, for a drug with non-linear kinetics, the rate at which the drug is
`
`eliminated from the body will not change in proportion to the amount of drug in
`
`the body such that the elimination rate constant and therefore half-life will, in fact,
`
`change.
`
`- 10 -
`
`

`

`
`
`
`23. The graph above shows differences in kinetics that can exist between dose-
`
`independent and dose-dependent drugs. Dose-independent drugs exhibit linear
`
`kinetics, meaning that the elimination rate (and thus half-life) remains constant
`
`regardless of serum concentration, as illustrated by the straight dotted lines in the
`
`graph above. In contrast, for dose-dependent drugs, the elimination rate and half-
`
`life vary depending upon concentration of drug in plasma. (Ex. 1022 Vol. 3,
`
`Rowland ’95 at 108-09; Ex. 2006, Niazi at 143.) For example, as drug
`
`concentration decreases over time, half-life decreases, which means that
`
`elimination increases. This is illustrated by the solid lines with slopes that change
`
`depending on the dose amount—high, medium, and low, as shown in the graph
`
`above—and/or the change in elimination (or half-life) as drug concentration
`
`- 11 -
`
`

`

`
`
`decreases over time. As will be discussed more fully below, the prior art taught
`
`that trastuzumab had dose-dependent, non-linear kinetics. See infra Section V(B).
`
`24.
`
`In August 1999, a skilled artisan would know that for a dose-independent
`
`drug, the half-life will remain constant over any dose interval, whether the dose
`
`interval is 7, 14, or 21 days (this is the assumption used by Dr. Jusko). A skilled
`
`artisan would also know that for a dose-dependent drug, the half-life and
`
`elimination rate will not remain constant across different doses and dosing
`
`intervals, or even within a single dose interval, but can vary as drug concentration
`
`changes. (See Ex. 2008, Gabrielsson at 123.) However, the actual rates of
`
`elimination for a dose-dependent drug would be unknown and unpredictable
`
`without sufficient data, such as by conducting a “washout study” where serum
`
`concentration data is collected over a period of at least 5 to 7 half-lives after a
`
`single administration of the drug or by measuring serum concentration levels from
`
`initial doses through steady-state at a given dose amount. There is no prior art
`
`reference for trastuzumab that describes such data.
`
`25. Thus, for a drug with non-linear kinetics like trastuzumab, a
`
`pharmacokineticist could not assume that a 250 mg dose would be eliminated from
`
`the body at the same rate as a 500 mg dose. This also means that a half-life
`
`measurement for a particular dose amount over one period of time could not be
`
`confidently applied to a different dose amount or dose interval. In other words,
`
`- 12 -
`
`

`

`
`
`based on the cited prior art, a skilled artisan would appreciate that the measured
`
`half-life of a 500 mg dose of trastuzumab over a one-week period could not be
`
`applied to a two-or-three-week dosing interval. This is a direct result of the non-
`
`linearity of trastuzumab, and the fact that its half-life is dependent on the
`
`concentration of the drug in the serum. As illustrated in the graph above, applying
`
`a constant value for half-life over a three-week period, based on the one-week data
`
`reported in the prior art, to a dose-dependent drug like trastuzumab could
`
`overestimate trough serum concentration levels by failing to account for the non-
`
`linear increase in elimination and corresponding decrease in the half-life that
`
`would be expected to occur as serum concentration declines.
`
`26. Take, for example, the anticancer agent indisulam, which is known to have
`
`non-linear kinetics. (Ex. 2052, Anthe S. Zandvliet et al., Saturable Binding of
`
`Indisulam to Plasma Proteins and Distribution to Human Erythrocytes, 34 DRUG
`
`METABOLISM & DISPOSITION 1041 (2006) at 1041.) The graph below is adapted
`
`from Zandvliet 2006 and shows a plasma concentration vs. time profile of
`
`indisulam. If a skilled artisan were to predict the serum concentration of indisulam
`
`after 175 hours using only the data collected from the first 50 hours and assumed
`
`the half-life would remain constant (i.e., assuming linear kinetics as depicted by
`
`the red line), the predicted serum concentration would be an overestimate of the
`
`measured serum concentration by at least an order of magnitude:
`
`- 13 -
`
`

`

`
`
`
`
`(See id. at 1045, Fig. 6 (edited for clarity, red straight and red dotted lines added).)
`
`While not all drugs with non-linear kinetics will have deviations of this magnitude,
`
`it is not possible to predict the extent of such a deviation without sufficient data.
`
`27. As a general matter, drugs with linear kinetics can be accurately described
`
`with relatively limited data. A pharmacokineticist could predict plasma
`
`concentrations for different doses and different dose intervals by applying a
`
`constant half-life. In contrast, for drugs with non-linear, dose-dependent kinetics a
`
`pharmacokineticist could not reliably predict plasma concentrations for different
`
`doses and dose intervals by applying a constant half-life, but would require more
`
`comprehensive information to define the deviations due to non-linear kinetics.
`
`
`
`The Prior Art Taught that Trastuzumab Had Dose-Dependent,
`Non-Linear Pharmacokinetics
`28. The prior art cited in the Petition explicitly teaches that trastuzumab
`
`demonstrated dose-dependent pharmacokinetics. For example, the Label reports
`
`that “[s]hort duration intravenous infusions of 10 to 500 mg once weekly
`
`- 14 -
`
`

`

`
`
`demonstrated dose-dependent pharmacokinetics.” (Ex. 1008, Label at 1.) The data
`
`presented in the Label is consistent with this conclusion because it teaches that
`
`“[m]ean half-life increased and clearance decreased with increasing dose level”
`
`and that “[t]he half-life averaged 1.7 days and 12 days at the 10 and 500 mg dose
`
`levels, respectively.” (Id.)
`
`29. Thus, when reporting different half-lives for the 10 and 500 mg doses, the
`
`Label specifically stated that trastuzumab demonstrated dose-dependent
`
`pharmacokinetics. (Id.) Baselga ’96 similarly reports that Phase I clinical trials
`
`conducted at University of California, Los Angeles, and Memorial Sloan-Kettering
`
`Cancer Center demonstrated that trastuzumab “had documented dose-dependent
`
`kinetics.” (Ex. 1013, Baselga ’96 at 10.)
`
`30. A skilled artisan would understand the term “dose-dependent
`
`pharmacokinetics” as used in Baselga ’96 and the Label to indicate that
`
`trastuzumab had demonstrated non-linear kinetics. (See, e.g., Ex. 2006, Niazi at
`
`181.) I am not aware of any prior art information that would have caused a skilled
`
`pharmacokineticist to doubt that trastuzumab had dose-dependent
`
`pharmacokinetics.
`
`31. Baselga ’96’s reference to the use of a one-compartment model (Ex. 1013,
`
`Baselga ’96 at 10; cf. Ex. 1003, ¶33) does not support Dr. Jusko’s conclusion that
`
`trastuzumab would have linear kinetics over a three-week dosing interval.
`
`- 15 -
`
`

`

`
`
`32. The question of whether a drug has non-linear kinetics over a given dosing
`
`schedule is independent of the question of whether a one or two compartment
`
`model best fits the data. Under certain circumstances, a drug with non-linear
`
`kinetics could be described with a one-compartment model. Similarly, drugs with
`
`linear kinetics are often described with a two-compartment model.
`
`33. A skilled artisan reading Baselga ’96 in its entirety would understand that a
`
`one-compartment linear model may have been descriptive of the data collected
`
`over a weekly dosing interval of a single dose amount and schedule where only
`
`two data points are determined (peak and trough) before another dose was
`
`administered. A skilled artisan would not assume that the same linear, one-
`
`compartment model could be used to reliably predict what would happen over a
`
`longer dosing interval or a different dose amount or if more comprehensive data
`
`were obtained. (See also infra ¶¶60-64.)
`
` Dr. Jusko’s Computations and Predictions Are Inconsistent with
`the Teaching in the Prior Art
`34. Contrary to the teaching of the prior art, Dr. Jusko modeled a three-week
`
`dosing regimen based upon the assumption that trastuzumab demonstrated linear
`
`pharmacokinetics. (Ex. 1003, Jusko Decl. ¶¶46-66, 69-71.) These assumptions are
`
`reflected in the equations Dr. Jusko applies at ¶¶46 to 66 of his declaration. All of
`
`these equations assume that the half-life, elimination rate constant, and volume of
`
`distribution of trastuzumab remain constant as the dose amount changes and as the
`
`- 16 -
`
`

`

`
`
`concentration of the drug in the body decreases. For example, to determine the
`
`elimination rate constant used in his equations, Dr. Jusko assumed a half-life of 12
`
`days. (See, e.g., id. ¶49.) Dr. Jusko obtained this 12-day half-life from the 500 mg
`
`trial reported in the Label, where the Label states that trastuzumab demonstrated
`
`dose-dependent kinetics. (Ex. 1008, Label at 1.) Dr. Jusko then used the
`
`elimination rate constant calculated using the 12-day half-life to determine the
`
`serum trough concentration levels after single and repeated doses. (See, e.g., Ex.
`
`1003, Jusko Decl. ¶¶ 50, 54.) A skilled artisan would recognize that the
`
`assumptions which underlie these calculations, i.e., a single 12-day half-life and
`
`constant elimination rate, are contrary to the prior art teaching that trastuzumab has
`
`dose-dependent kinetics.
`
`35. As explained above, drugs with linear kinetics can be easily described by a
`
`single unchanging half-life. (Supra ¶¶20-26.) That is not the case for drugs
`
`displaying non-linear kinetics, such as trastuzumab. A drug with non-linear
`
`kinetics will have one value for elimination rate constant at one concentration of
`
`the drug, and a different value for elimination rate constant at a different
`
`concentration of the drug. Since concentration of drug is continually changing
`
`over the duration of the time between doses, the elimination rate constant (and
`
`therefore half-life) are also continually changing. (See Ex. 2008, Gabrielsson at
`
`123; see also Ex. 1022 at 3:108-09.)
`
`- 17 -
`
`

`

`
`
`36.
`
`In the face of the limited data of the prior art and direct evidence of non-
`
`linear kinetics, a skilled artisan would not predict a dose regimen based on the
`
`assumption that the drug has elimination described by a constant singular half-life.
`
`The half-life derived from the weekly administration of a particular dose of a drug
`
`with documented dose-dependent kinetics has limited predictive value for
`
`estimating the half-life of a longer dosing interval for such a drug. (See Ex. 2008,
`
`Gabrielsson at 123 (“The half-life is a function of plasma concentration for the
`
`nonlinear system.”).);
`
`VI. A Skilled Artisan Would Not Rely on Dr. Jusko’s Erroneous
`Assumptions to Predict the Efficacy of Three-Week Dosing
` Maintaining Adequate Serum Trough Concentrations Is
`Important in Predicting the Efficacy of a Dose Regimen for
`Trastuzumab
`37. For most drugs with repeat dosing—including trastuzumab—the goal is to
`
`achieve plasma concentration levels within a therapeutic window, i.e., above a
`
`certain minimum or “trough” level where efficacy is maintained and below a
`
`ceiling or “peak” where unacceptable toxicity may occur. For a drug that has
`
`already been shown to be safe and effective at a particular dose amount and dose
`
`interval, a pharmacokineticist would generally seek to ensure that any alternative
`
`dose amount and interval would yield serum concentration data within the range
`
`previously known to be safe and clinically effective. Without an adequate
`
`- 18 -
`
`

`

`
`
`understanding of the pharmacokinetics of a drug, it is impossible to make such a
`
`determination.
`
`38. Serum drug concentration reaches a “peak” or maximum concentration
`
`shortly after administration. As the drug is eliminated from the body over time, it
`
`reaches a “trough” or minimum concentration just prior to the next dose. A dosing
`
`regimen for trastuzumab must be designed such that the fluctuations in drug
`
`concentration within a dosing interval do not result in a trough concentration that
`
`fall below the minimum effective concentration. Dose amount and dose interval
`
`affect both peak and trough levels. For example, increasing the dose and extending
`
`the dose interval can result in significantly higher peak levels and significantly
`
`lower trough levels as compared to the same overall amount of drug given in lower
`
`doses more frequently. In fact, administering lower doses more frequently is a
`
`common method to minimize the variation between peak and trough levels.
`
`39. At the priority date in August 1999, a skilled artisan considering an
`
`alternative dosing regimen for trastuzumab would know that efficacy was
`
`associated with maintaining adequate steady-state trough serum levels, i.e.,
`
`maintaining a “therapeutic trough concentration.” A skilled artisan would thus
`
`want to ensure that any alternative dosing regimen maintained therapeutic trough
`
`concentrations throughout the course of treatment. A dosing regimen with a
`
`- 19 -
`
`

`

`
`
`potential to result in concentrations below therapeutic trough levels would not be
`
`considered acceptable as it would risk leaving patients without therapeutic benefit.
`
` A Skilled Artisan Would Conclude that Dr. Jusko’s Calculations
`Would Likely Overestimate Trough Serum Concentrations
`40. A skilled artisan would conclude that Dr. Jusko’s erroneous assumptions
`
`would likely overestimate trough concentration levels of trastuzumab. In
`
`particular, a skilled artisan would know that using a 12-day half-life from a study
`
`of weekly administration of a dose-dependent drug such as trastuzumab to predict
`
`a three-week dosing regimen would very likely overestimate serum trough
`
`concentrations. Such an error would have serious implications for accurately
`
`predicting the efficacy of a breast cancer treatment regimen when maintaining a
`
`therapeutic serum trough concentration is needed for efficacy.
`
`41. A skilled artisan would understand from the Label that the half-life of
`
`trastuzumab will decrease as the concentration of trastuzumab in the blood
`
`decreases. The Label expressly teaches that “[m]ean half-life increases and
`
`clearance decreased with increasing dose level.” (Ex. 1008, Label at 1.) A skilled
`
`artisan would understand this to mean that as the concentration of trastuzumab in
`
`the bloodstream increased, the half-life increased. A skilled artisan would
`
`understand the corollary to also be true: as the concentration of trastuzumab in the
`
`bloodstream decreased, the half-life decreased.
`
`- 20 -
`
`

`

`
`
`42. The prior art relied upon by the Petitioner reports shorter half-lives for
`
`smaller doses of trastuzumab when measu