`
`
`
`UNITED STATES PATENT AND TRADEMARK OFFICE
`________________
`
`BEFORE THE PATENT TRIAL AND APPEAL BOARD
`________________
`
`TARO PHARMACEUTICALS U.S.A., INC.,
`Petitioner,
`
`v.
`
`APOTEX TECHNOLOGIES, INC.,
`Patent Owner.
`________________
`
`Case IPR2017-01446
`U.S. Patent No. 7,049,328 B2
`
`Title: USE FOR DEFERIPRONE
`________________
`
`
`SECOND DECLARATION OF DUDLEY J. PENNELL, M.D.,
`IN SUPPORT OF PATENT OWNER’S RESPONSE
`
`
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`I.
`
`II.
`
`Table of Contents
`
`INTRODUCTION, BACKGROUND, AND QUALIFICATIONS ............... 1
`
`LEGAL STANDARDS ................................................................................... 1
`
`III. BACKGROUND OF THE TECHNOLOGY .................................................. 1
`
`A.
`
`
`There Were No Reliable Methods to Quantify and Correlate Cardiac
`Iron Content with Cardiac Disease at the Time of the Invention ......... 1
`
`IV. THE ’328 PATENT ......................................................................................... 5
`
`V.
`
`CLAIM CONSTRUCTION ............................................................................ 5
`
`VI. HOFFBRAND 1998, OLIVIERI ABSTRACT 1995, AND OLIVIERI 1995
`
`FAIL TO EXPLICITLY OR INHERENTLY DISCLOSE A PATIENT
`HAVING AN IRON OVERLOAD CONDITION OF THE HEART ............ 6
`
`
`A. Hoffbrand 1998 does not disclose treating a patient having an iron
`overload condition of the heart .............................................................. 7
`
`
`B. Olivieri 1995 does not disclose treating a patient having an iron
`overload condition of the heart .............................................................. 9
`
`
`C. Olivieri Abstract 1995 does not disclose treating a patient having an
`iron overload condition of the heart ....................................................10
`
`
`VII. GROUNDS 6-10 FAIL TO ESTABLISH THAT CLAIMS 1, 2, 4-17, AND
`
`19 ARE OBVIOUS OVER THE PRIMARY REFERENCES IN VIEW OF
`THE KNOWLEDGE OF A POSA ................................................................12
`
`
`
`
`ii
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`I, Dudley J. Pennell, M.D., declare as follows:
`
`I.
`
`INTRODUCTION, BACKGROUND, AND QUALIFICATIONS
`
`1. My compensation, background, and qualifications have remained the
`
`same since the submission of my Declaration in Support of Patent Owner’s
`
`Preliminary Response (Ex. 2003). (Ex. 2003 at 1-3.)
`
`II. LEGAL STANDARDS
`
`2. My understanding of the legal standards for anticipation and
`
`obviousness are the same as set forth in my Original Declaration. (Ex. 2003 at 3-
`
`5.)
`
`III. BACKGROUND OF THE TECHNOLOGY
`
`A. There Were No Reliable Methods to Quantify and Correlate
`Cardiac Iron Content with Cardiac Disease at the Time of the
`Invention
`
`3.
`
`As discussed in my Declaration (Ex. 2003), there was a misguided
`
`belief that a liver iron concentration (“LIC”) >80 µmol of iron per gram of liver
`
`(“µmol/g”), wet weight (or >15 mg/g, dry weight), was associated with an
`
`increased risk of heart disease in blood transfusion dependent patients. (Ex. 2003
`
`at 7-8; see e.g., Ex. 1012 at 921; see also Ex. 2011 at 419; Ex. 1017 at 565.)
`
`4.
`
`In addition to my own experiments showing that LIC is not reflective
`
`of, and in no way predicts cardiac iron content (Ex. 2007), a study by Berdoukas
`
`1
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`(Ex. 2022)1 directly challenged and disproved Olivieri and Brittenham’s proposed
`
`association between cardiac iron levels and LIC. (Ex. 2022 at 685-686.) The
`
`study included 58 transfusion-dependent patients stratified into three groups:
`
`- Group I included patients with LIC <7 mg/g dry weight;
`
`- Group II included patients with LIC between 7 mg/g dry weight and 15 mg/g
`
`dry weight; and
`
`- Group III included patients with LIC >15 mg/g dry weight.
`
`(Ex. 2022 at 685, Table 1.)
`
`5.
`
`All groups were evaluated for cardiac dysfunction with multiple
`
`uptake gated acquisition (“MUGA”) scans. (Ex. 2022 at 685.) The empirical
`
`results showed that thalassemia (“TM”) patients having LIC >15 mg/g, dry weight,
`
`and even as high as 62 mg/g, dry weight, exhibited the same cardiac function as
`
`TM patients having LIC of 6.8 mg/g, dry weight. (Ex. 2022 at 685, Table 1.)
`
`Further, the study demonstrated that there was no statistical difference in the
`
`cardiac function of TM patients averaging an LIC of 32.4 mg/g, 10.4 mg/g, and 4.2
`
`mg/g, dry weight. (Ex. 2022 at 685-686, Table 1.) Thus, a POSA can only
`
`conclude that a LIC >15 mg/g, dry weight, cannot be associated with iron overload
`
`in the heart, let alone indicative of cardiac disease. See e.g., Ex. 2006 at 2973; Ex.
`
`
`1 V. Berdoukas et al. “Lack of correlation between iron overload cardiac
`dysfunction and needle liver biopsy iron concentration,”
`Haematologica 90(5):685-6 (2005) (Ex. 2022).
`
`2
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`20272 at 100; Ex. 20283 at 272; Ex. 20294 at 2; Ex. 20305 at 214. Berdoukas also
`
`highlights that in “15 patients with a LIC in the range in which one might expect
`
`serious problems from iron overload, only one had an abnormal resting LVEF and
`
`only two had a significant reduction in their ΔLVEF.” (Ex. 2022 at 686.)
`
`6.
`
`In my Declaration, I also commented that serum ferritin (“SF”) levels
`
`are not indicative of cardiac iron content. (Ex. 2003 at ¶¶ 26, 51.) To be clear, it
`
`has been well known since 1976 that SF levels are not reflective of parenchymal
`
`iron content (i.e., the iron content of organs such as the heart and liver). (Ex. 20316
`
`at 46; Ex. 20327 at 333-4.)
`
`7.
`
`In my Declaration, I further commented on the state of magnetic
`
`resonance imaging (“MRI”) technology, and specifically that, as of the 1990s, MRI
`
`
`2 A. Aessopos et al. “The heart in transfusion dependent homozygous thalassaemia
`today – prediction, prevention and management,” Eur. J. Haematology 80:93-106
`(2008) (Ex. 2027).
`3 N. Koonrungsesomboon et al. “Early detection of cardiac involvement in
`thalassemia: From bench to bedside perspective,” World J. Cardiol. 5(8):270-279
`(2013) (Ex. 2028).
`4 A. Aessopos et al. “Prevention of Cardiomyopathy in Transfusion-Dependent
`Homozygous Thalassaemia Today and the Role of Cardiac Magnetic Resonance
`Imaging,” Advance in Hematology Article ID 964897:1-4 (2009) (Ex. 2029).
`5 S. Lekawanvijit et al. “Iron overload thalassemic cardiomyopathy: Iron status
`assessment and mechanisms of mechanical and electrical disturbance due to iron
`toxicity,” Can. J. Cardiol. 25(4):213-218 (2009) (Ex. 2030).
`6 DL Johnston et al. “Assessment of tissue iron overload by nuclear magnetic
`resonance imaging,” Am. J. Med. 87(1):40-7 (1989) (Ex. 2031).
`7 WH Crosby “Editorial: Serum ferritin fails to indicate hemochromatosis--nothing
`gold can stay,” N. Engl. J. Med. 294(6):333-4 (1976) (Ex. 2032).
`
`3
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`T2 could not quantify cardiac iron loading. (Ex. 2003 at 8; Ex. 2015 at 2177.) Of
`
`note, Dr. A. Victor Hoffbrand, the first named author of Ex. 1007, reiterated my
`
`opinion that “MRI T2 failed because of lack of sensitivity, motion artifacts and
`
`poor signal-to-background noise ratios.” (Ex. 2036 at 2140.)8 Dr. Hoffbrand
`
`further agreed that “Myocardial T2* provides the earliest, and probably the most
`
`sensitive and reproducible test yet available of cardiac iron and consequent cardiac
`
`damage.” (Ex. 2036 at 2141.)
`
`8.
`
`In addition to the inaccuracy of MRI T2 to quantify cardiac iron
`
`loading, MRI T2 relaxation data cannot independently diagnose a patient for
`
`cardiac disease. As shown in a 1998 study by Mavrogeni et al.9 (“Mavrogeni”),
`
`focusing on MRI T2 relaxation times and iron overload in β-thalassemia,
`
`evaluation of cardiac disease required examination by a cardiologist and
`
`classification according to the criteria defined by the New York Heart Association
`
`(“NYHA”), where NYHA II-III is the classification for heart failure. (Ex. 2033 at
`
`8.) Viewed in isolation, the MRI T2 data reported by Mavrogeni was inconclusive
`
`of cardiac iron loading or cardiac disease. Specifically, the MRI T2 data in
`
`Mavrogeni showed no correlation between MRI T2 relaxation times and cardiac
`
`
`8 A.V. Hoffbrand “Diagnosing myocardial iron overload,” EUR. HEART J. 22:2140-
`2141 (2001) (Ex. 2036).
`9 SI Mavrogeni et al. “T2 relaxation time study of iron overload in b-thalassemia,”
`MAGMA 6(1):7-12 (1998) (Ex. 2033).
`
`4
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`disease, as 12 of 30 patients having an average T2 relaxation time of 29.9 msec ±
`
`5.7 were diagnosed with heart failure (NYHA II-III) and 3 out of 24 patients
`
`having an average T2 relaxation time of 33.4 msec ± 6.8 were diagnosed with heart
`
`failure (NYHA II-III). (Ex. 2033 at 9, Table 2.) Thus, a POSA would have
`
`understood, at least as of 1998, that there is no correlation between MRI T2
`
`relaxation time and cardiac disease, and that there was no accepted threshold MRI
`
`T2 relaxation time that could indicate cardiac disease.
`
`IV. THE ’328 PATENT
`
`9. My understanding of the ’328 patent and the claims therein remain the
`
`same as set forth in my original declaration in support of Patent Owner’s
`
`Preliminary Response.
`
`V. CLAIM CONSTRUCTION
`
`10.
`
`In forming the opinions set forth herein, I have considered the
`
`preamble of independent claims 1-2 and 4-10 to be limiting. I have also
`
`considered the disputed terms to be material limitations of the claims. However,
`
`my opinion does not change if the disputed terms are not considered material
`
`limitations of the claims. Further, it is my opinion that the words of each of the
`
`challenged claims would be readily understood by a POSA.
`
`11.
`
`I also understand that the PTAB has reviewed the ’328 patent
`
`specification and did not find a definition for the phrase “iron overload condition
`
`5
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`of the heart.” (Paper No. 15 at 3.) I note that the ’328 patent refers at least twice
`
`to an “iron overload condition of the heart” as “iron induced cardiac disease.” (See
`
`Ex. 1001 at col. 10, ll. 18-19, ll. 29-31.) In my opinion, a POSA would understand
`
`that a patient with “an iron overload condition of the heart,” to mean a patient with
`
`levels of cardiac iron that would elicit a heart condition diagnosable by common
`
`methods such as LVEF, radionuclide ventriculography (e.g., MUGA scan),
`
`echocardiogram, electrocardiogram (e.g., Holter monitoring), and the like. Thus, a
`
`POSA would understand that “an iron overload condition of the heart” represents a
`
`diagnosable cardiac condition that is due to iron accumulation in the heart that
`
`could range from something as mild as arrhythmia or NYHA I-II angina to more
`
`severe cardiac disease such as NYHA III-IV heart failure.
`
`VI. HOFFBRAND 1998, OLIVIERI ABSTRACT 1995, AND OLIVIERI
`1995 FAIL TO EXPLICITLY OR INHERENTLY DISCLOSE A
`PATIENT HAVING AN IRON OVERLOAD CONDITION OF THE
`HEART
`
`12.
`
`I understand that the instituted grounds of invalidity are:
`
`Reference(s)
`
`Basis
`
`Claims Challenged
`
`Hoffbrand 1998
`Olivieri Abstract 1995
`Olivieri 1995
`Hoffbrand 1998
`Olivieri Abstract 1995
`Olivieri 1995
`
`§ 102(b)
`§ 102(b)
`§ 102(b)
`§ 103(a)
`§ 103(a)
`§ 103(a)
`
`1, 2, 4-11, 13-17, 19
`1, 2, 4-11, 13-17, 19
`1, 2, 4-11, 13-17, 19
`1, 2, 4-17, 19
`1, 2, 4-17, 19
`1, 2, 4-17, 19
`
`13. For all of the reasons discussed below, it is my opinion that the
`
`6
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`Primary References do not disclose, either expressly or inherently, administering
`
`deferiprone to a transfusion dependent patient having or experiencing an iron
`
`overload condition of the heart as required by independent claims 1, 2, and 6-9
`
`(and dependent claims therefrom).
`
`A. Hoffbrand 1998 does not disclose treating a patient having an iron
`overload condition of the heart
`
`14. Dr. Mehta and Taro assert that Hoffbrand 1998 anticipates claims 1, 2,
`
`and 6-9 because it discloses “that ten patients had a liver iron content above
`
`15.0 mg/g dry weight, which falls in the range of iron content that has been
`
`associated with cardiac disease due to iron overload.” (Ex. 1002 at ¶ 74; Pet. at 35-
`
`36.) Dr. Mehta’s opinion is incorrect.
`
`15. As discussed above, it has been clinically shown that there is no
`
`correlation between LIC and an iron overload condition in the heart. Specifically,
`
`Berdoukas found that, LIC, assessed by needle biopsy, is not predictive of cardiac
`
`dysfunction assessed by MUGA scan,” and that the historical guidepost of LIC
`
`>15.0 mg/g, dry weight, had no correlation with abnormal cardiac function. (Ex.
`
`2022 at 685-686, Table 1, Figure 1.)
`
`16. Of particular note, Berdoukas specifically referenced and dispelled the
`
`7
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`conclusions of Brittenham et al. (Ex. 2034)10 concerning a greater risk for TM
`
`patients having LIC >15 mg/g relied upon by Hoffbrand, Taro, and Dr. Mehta.
`
`(See Ex. 2022 at 686 n. 3; see also Ex. 1007 at 298 (“levels above 15 mg/g, levels
`
`at which liver and cardiac damage are likely to occur.”) n. 21.) Indeed, Berdoukas
`
`makes clear that the focused study “[did] not support the threshold values proposed
`
`by Olivieri and Brittenham11 [(i.e. that LIC >15 mg/g dry weight was associated
`
`with cardiac disease)].” (Ex. 2022 at 686.)
`
`17. Thus, in my opinion there is no scientific basis to conclude that any
`
`TM patient in the Hoffbrand 1998 study in fact had an iron overload condition of
`
`the heart, let alone cardiac disease. Thus, Taro’s reliance that “Hoffbrand 1998
`
`discloses that ten patients had a liver iron content above 15.0 mg/g dry weight, due
`
`to iron overload,” cannot indicate, let alone even suggest, iron overload in the heart
`
`or cardiac disease or dysfunction. (Pet. at 35–36 (citing Ex. 1007, 297 and Ex.
`
`1002 ¶¶ 74–75).)
`
`18. Moreover, Hoffbrand admits that LIC >15 mg/g, dry weight, are
`
`merely “levels at which liver and cardiac damage are likely to occur,” tacitly
`
`acknowledging the lack of scientific support for this assumption. (Ex. 1007 at 297
`
`
`10 Brittenham GM et al. “Efficacy of deferoxamine in preventing complications of
`iron overload in patients with thalassemia major,” N. Engl. J. Med. 331:567-73,
`(1994) (Ex. 2034).
`11 Id.
`
`8
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`(emphasis added), (n. 21 referencing Ex. 2034.) However, even Hoffbrand’s
`
`characterization that “cardiac damage [i]s likely to occur” is dispelled by
`
`Berdoukas. Specifically, Berdoukas reports that in “15 patients with a LIC in the
`
`range in which one might expect serious problems from iron overload, only one
`
`had an abnormal resting LVEF and only two had a significant reduction in their
`
`ΔLVEF.” (Ex. 2022 at 686.) A POSA would have understood that a majority of
`
`TM patients with LIC >15mg/g, dry weight, would not have abnormal heart
`
`conditions. Thus, a POSA would have understood that Hoffbrand’s description
`
`that cardiac damage was “likely to occur” at LIC >15 mg/g, dry weight, in no way
`
`inherently teaches that cardiac damage was “necessarily present” in these TM
`
`patients.
`
`19. Accordingly, I maintain and further support my opinion that a POSA
`
`would not view Hoffbrand 1998 as teaching the treatment of blood transfusion-
`
`dependent patients who have an iron overload condition of the heart.
`
`B. Olivieri 1995 does not disclose treating a patient having an iron
`overload condition of the heart
`
`20. For the reasons stated above concerning Hoffbrand, Olivieri 1995
`
`cannot anticipate claims 1, 2, and 6-9 based on the erroneous belief that LIC >80
`
`µmol per gram of liver tissue, wet weight, is a threshold indicative of cardiac
`
`disease.
`
`21. Olivieri 1995 reports that “[w]e used hepatic iron concentration under
`
`9
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`80 µmol per gram . . . as the criteria for effective chelation therapy to facilitate the
`
`comparison between deferiprone and deferoxamine. These values may not
`
`represent the optimal goals for the treatment of iron overload. They are derived
`
`from prospective trials lasting more than a decade,3,4 . . . .” (Ex. 1012 at 921.)
`
`Olivieri 1995 references Brittenham’s 1994 publication (Ex. 2034) that proposed
`
`LIC >80 µmol per gram of liver tissue, wet weight, as associated with cardiac
`
`dysfunction, which Berdoukas proved as incorrect.
`
`22. Moreover, a POSA was well aware that SF levels do not reflect
`
`parenchymal iron content (i.e. the iron content of organs such as the heart and
`
`liver). (Ex. 2031 at 46; Ex. 2032 at 333-4.)
`
`23. Furthermore, Olivieri 1995 does not disclose any additional clinical
`
`methods of assessing a patient’s heart condition due to iron overload. Thus, for the
`
`same reasons discussed above concerning Hoffbrand 1998, it is my opinion that
`
`Olivieri 1995 did not explicitly or inherently treat any TM patients who had an iron
`
`overload condition of the heart, let alone cardiac disease.
`
`C. Olivieri Abstract 1995 does not disclose treating a patient having
`an iron overload condition of the heart
`
`24.
`
`I respectfully disagree with Dr. Mehta’s opinions that MRI T2
`
`relaxation time (“TRT”) is an indicator of the extent of iron overload in the heart
`
`and that measurements below a threshold value of 32 msec are indicative of
`
`cardiac disease due to cardiac iron overload. (Ex. 1002 at ¶ 75; Pet. at 38.)
`
`10
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`25. As discussed above, TRT was not a reliable tool to quantify cardiac
`
`iron levels, and cannot indicate whether a patient has a heart condition due to iron
`
`overload. (See Ex. 2015 at 2177; Ex. 2036 at 2140.) The unreliability of TRT
`
`measurements is reflected by the large standard deviation in the data presented in
`
`Olivieri Abstract 1995. For example, the relative standard deviation of TRT data
`
`in Olivieri Abstract 1995 is 23.9 msec ± 27% and 32.4 msec ± 29%. The large
`
`variability in the data would have led a POSA to conclude that TRT was much too
`
`imprecise to make reliable conclusions on the cardiac iron load and cardiac health
`
`of a TM patient. Indeed, a POSA at the time avoided the use of TRT because of its
`
`unreliability. (Ex. 2031 at 46 (“T2 relaxation times were not used in this study,
`
`since the calculation of these values is imprecise.”).)
`
`26. Further, as discussed in Section III(A)(5), subsequent studies proved
`
`that TRT in TM patients showed no correlation between TRT values and cardiac
`
`disease, and completely disprove Dr. Mehta’s assumption that there was a
`
`threshold TRT value indicative of heart failure. (See Ex. 2033 at 8-9.) Thus, a
`
`POSA would have understood that a TRT value under 32 msec does not explicitly
`
`or inherently indicate a heart condition due to cardiac iron overload. Moreover, a
`
`POSA would have recognized that Olivieri 1995 Abstract fails to disclose any
`
`other clinical methods (i.e. MUGA scan, LVEF, Holter monitoring, etc.) that
`
`would explicitly or inherently disclose whether patients in the study were afflicted
`
`11
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`with a heart condition due to iron overload. It is my opinion that a POSA would
`
`not have been able to make any conclusions regarding the heart conditions of
`
`patients as disclosed in Olivieri Abstract 1995.
`
`27. Moreover, I note that the TRT result disclosed in Olivieri Abstract
`
`1995 was never published in any form that would have been subjected to critical
`
`peer review. Whilst we don’t know whether Dr. Olivieri and co-authors attempted
`
`to publish the TRT cardiac data, the fact that it was not published in a peer review
`
`journal suggests that either she and her co-authors or peer reviewers found it to not
`
`be of adequate scientific quality for publication in a peer reviewed scientific
`
`journal. The resulting peer-reviewed scientific publications (Exs. 1012, 2011,
`
`2012, and 2013) concerning the administration of deferiprone to this same cohort
`
`of patients is wholly silent as to cardiac iron loading in these patients, let alone its
`
`purported measurement by TRT.
`
`28. Thus, it is my opinion that a POSA would not view Olivieri 1995
`
`Abstract as teaching the treatment of transfusion-dependent patients who have an
`
`iron overload condition of the heart.
`
`VII. GROUNDS 6-10 FAIL TO ESTABLISH THAT CLAIMS 1, 2, 4-17,
`AND 19 ARE OBVIOUS OVER THE PRIMARY REFERENCES IN
`VIEW OF THE KNOWLEDGE OF A POSA
`
`29.
`
`I maintain my previous opinions that Taro has failed to show that
`
`claims 1, 2, 4-17, and 19 are obvious. I further, supplement my previous opinion
`
`12
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`with the discussion herein.
`
`30. As shown above, my opinions do not merely criticize the accuracy of
`
`the data presented in Hoffbrand 1998, Olivieri 1995, and the Olivieri 1995
`
`Abstract. Instead, my opinions rely on empirical evidence that disproves the
`
`association of LIC and TRT with iron overload in the heart and iron-induced
`
`cardiac disease. Thus, a POSA would have understood that the TM patients treated
`
`with deferiprone in Hoffbrand 1998, Olivieri 1995, and Olivieri Abstract 1995 did
`
`not explicitly or inherently have iron overload in the heart or iron-induced cardiac
`
`disease based on the lack of correlation between LIC, SF, and TRT values with
`
`these conditions.
`
`
`
`
`
`13
`
`Apotex Tech.
`Ex. 2026
`
`
`
`
`
`I declare under penalty of perjury that, to the extent of my knowledge and belief,
`
`the foregoing is true and correct. I further declare that all statements made herein
`
`of my own knowledge are true and that all statements made on information and
`
`belief are believed to be true; and further that these statements were made with the
`
`understanding that knowing and willful false statements and the like so made are
`
`punishable by fine or imprisonment, or both, under section 1001 of title 18 of the
`
`United States Code.
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`___________________________
`
`
`
`
`
`
`Date: February 22, 2018 Dudley J. Pennell, M.D.
`
`
`
`
`
`14
`
`Apotex Tech.
`Ex. 2026
`
`