ATTORNEY DOCKET NO.: 056291-5004-01
`
`PATENT
`
`IN THE UNITED STATES PATENT AND TRADEMARK OFFICE
`
`Confirmation No.
`
`2093
`
`) ) ) )
`
`In re PATENT APPLICATION of:
`
`EVANS et al.
`
`Application No.:
`
`10/872,784
`
`) Group Art Unit:
`
`1617
`
`Examiner: Hui, San-Ming R
`
`) ) )
`
`))
`
`Filed:
`
`June 22, 2004
`
`FOR:
`
`FORIVIULATION
`
`DECLARATION UNDER 35 U.S.C. § 1.132
`OF PAUL RICHARD GELLERT
`
`PAUL RICHARD GELLERT of Astrazeneca, Alderley Park, Macclesfield, Cheshire, UK
`
`declares:
`
`l.
`
`I graduated from the University of Oxford in Chemistry in 1984. I undertook postgraduate
`
`research with Professor Brian Howard in the Physical Chemistry Laboratory at the
`
`University of Oxford leading to the award of a D.Phil in 1988. From February 1988 until
`
`the present I have been employed by AstraZeneca, (formerly Zeneca and ICI) initially as a
`
`Senior Research Scientist and subsequently as a Team Leader/Manager, Principal
`
`Scientist and, since 2004, a Senior Principal Scientist.
`
`2.
`
`I have worked in the formulation and drug delivery area throughout my career with
`
`Astrazeneca, where my research and development Work has covered a range of
`
`formulation types including sustained released injections, including fulvestrant.
`
`3. During the course of my study of the subject application (hereinafter “the Evans
`
`Application”) and the underlying data, Ihave become aware of several transcription or
`
`other errors between certain disclosures of the subject application and the underlying
`
`laboratory notebook data. One purpose of this Declaration is to point out the existence
`
`DB1/'62042606.l
`
`|nnoPharma Exhibit 1020.0001
`
`

`

`and nature of these errors and to report further testing that has been carried out under my
`
`guidance to obtain additional data (paragraphs 4-10 below and Attachments A—D). A
`
`further purpose of this Declaration is to set out and document the manner in which an
`
`experienced forrnulator would likely have approached the task of developing a sustained
`
`release injectable formulation suitable for human use for a steroidal compound such as
`
`fulvestrant in about early 2000, which I understand is when the priority applications
`
`supporting the Evans Application were filed (paragraphs 11 ~ 25 below and Attachment
`
`E). Citations to literature and patent references in this Declaration will be in the format
`
`Lead Author (Date), and the full citations are given in the Table of References at the end
`
`of this Declaration. A copy of each cited reference (or cited portions of the longer
`
`references) is included in Attachment F under the Tab number noted in the Table of
`
`References.
`
`4.
`
`In Table 2 of the Evans Application, the solubility of fulvestrant in Castor oil appears to
`
`have been transcribed incorrectly from the original source, the laboratory notebook. The
`
`value in the latter is 245 mg/ml and not 20 mg/ml. In other experiments to determine the
`
`solubility of fulvestrant in castor oil and also in benzyl benzoate, some variability was
`
`observed.
`
`5.
`
`In Table 3 of the Evans Application, the given solubility values were generated at 4°C
`
`and not at 25°C as is stated in the title of Table 3. For fulvestrant formulations, it is
`
`preferable that the fulvestrant remains completely in solution at both 4°C and 25°C. The
`
`4°C temperature corresponds to the storage temperature (2°C to 8°C in the FDA approved
`
`label for Faslodex), and the 25°C temperature corresponds to the administration
`
`temperature (ambient temperature). In addition, the specified solubility values on this
`
`Table 3 are mean values calculated from analysis of replicate samples from one or more
`
`trials. The individual values are shown in handwriting in the amended version of Table 3
`
`in Attachment A. In addition, it appears that the mean values for the last three
`
`compositions have been incorrectly calculated. The corrected mean Values, together with
`
`the correction of the temperature from “25“C” to read “4°C”, are also shown in
`
`handwriting in the amended version of Table 3 in Attachment A.
`
`DB 13620426061
`
`ta)
`
`|nnoPharma Exhibit 1020.0002
`
`

`

`6.
`
`I have evaluated the transcription and other errors against the original application
`
`disclosures and conclude that these do not change the ultimate conclusions made from the
`
`data as originally reported. The addition of 15% wfv benzyl benzoate to compositions
`
`having total alcohol concentrations in castor oil of 10%, 15%, 20% and 30% w/v
`
`unexpectedly provides a positive effect on fulvestrant solubility, significantly increasing
`
`the solubility of fulvestrant in the compositions despite fulvestrant having a lower
`
`solubility in benzyl benzoate than in either alcohol or castor oil.
`
`7. An additional set of experiments has been conducted at 25 °C under my guidance to
`
`obtain consistent data with reduced variability from a single set of rigorously controlled
`
`solubility experiments and to demonstrate that the unexpected increase of solubility of
`
`fulvestrant by adding benzyl benzoate into compositions containing ethanol, benzyl
`
`alcohol and castor oil, is present across the broader range of composition encompassed by
`
`the claims being presented With this Declaration. The solubility of fulvestrant in benzyl
`
`benzoate and in castor oil was also measured in the same set of experiments using the
`
`same batch ofbenzyl benzoate and the same batch of Castor oil as were used to make up
`
`the compositions. The Experimental Test Procedure is described in Attachment B.
`
`8. The results from these solubility experiments are shown in the table in Attachment C.
`
`These results show that the solubility of fulvestrant in castor oil alone (21.4 mg/ml) is
`
`significantly greater than the solubility of fulvestrant in benzyl benzoate alone (3.8
`
`mg/ml) and demonstrate the unexpected increase in fulvestrant solubility on the addition
`
`of 10, 15 and 25% w/V benzyl benzoate, in place of an equivalent amount of castor oil, to
`
`compositions having total alcohol concentrations in Castor oil of 10%, 15%, 20%, 25%
`
`and 30% W/v.
`
`9. Thus, the results that were obtained from experiments conducted under rigorously
`
`controlled conditions and with an expanded range of compositions, as shown in
`
`Attachment C; confirm the ultimate conclusions drawn from the results shown in Table 3
`
`of the original application disclosure, namely that the addition of 10% to 25% wfv benzyl
`
`DB1.»‘62042606.1
`
`3
`
`|nnoPharma Exhibit 1020.0003
`
`

`

`benzoate to compositions having total alcohol concentrations in castor oil of between
`
`10% to 30% wfv unexpectedly provides a positive effect on fulvestrant solubility,
`
`significantly increasing the solubility of fulvestrant in the compositions despite
`
`fulvestrant having a lower solubility in benzyl benzoate than in either alcohol or castor
`
`oil.
`
`10. During the course of my study of the Evans Application and the underlying source
`
`materials it was drawn to my attention that some of the composition data given for
`
`Delestrogen and Delalutin somehow had been shifted one column to the right. Thus, for
`
`Delestrogen, the 78% and 58% figures shown under the BzBz column should have been
`
`under the OIL column; the 20% and 40% figures shown under the BZOH column should
`
`have been under the BzBz column; and the 2% figures shown under EtOH should have
`
`been under the BZOH column. Similarly for Delalutin, the “up to 2%” shown under the
`
`EtOH column should have been under the BzOH column. This table reports that the
`
`source of this data was J .Pharm.Sci (1964) 53(8) 891, which is Riftkin (1964) elsewhere
`
`referred to in this Declaration, and I have also verified the corrected data from the entries
`
`for Delalutin and Delestrogen in PDR (1973). A copy of Table 1 from the Evans
`
`Application is reproduced as Attachment D, on which these corrections have been made
`
`in handwriting, and I have additionally more correctly noted that Delalutin is 17-hydroxy
`
`progesterone caproate, and that the “COMP” designation for Delalutin should be “BMS”
`
`G3ristol-Myers Squibb). Attachment D also includes a one page explanation of the
`
`corrections to this Table 1.
`
`l 1. In about early 2000, a person responsible for developing a sustained release inj ectable
`
`formulation suitable for administration to humans for a new steroidal compound such as
`
`fulvestrant, would have had specialized training and experience in developing
`
`pharmaceutical formulations and methods for their administration. In developing such a
`
`formulation for fulvestrant, the objective would have been to formulate an intramuscular
`
`(LVI) injection that would provide for the satisfactory sustained release of fulvestrant over
`
`a period of at least two weeks and preferably over a period of at least four weeks to
`
`reduce the frequency of administration, and would have a target fulvestrant content of at
`
`DB 1162042606.}
`
`4
`
`|nnoPharma Exhibit 1020.0004
`
`

`

`least 45 mg;’mL so as to provide a fulvestrant dose of at least 250 mg in a single 5-6 mL
`
`injection. From my personal experience and knowledge of the literature at about that
`
`time, I believe that such an experienced formulator would likely have approached the task
`
`of developing a fonnulation for fulvestrant in about the following manner.
`
`12. Given the foregoing objective, the experienced formulator would have appreciated that
`
`the traditional administration options to explore were intramuscular (IM) injection of a
`
`sustained release aqueous or oil suspension or an oil-based solution (depot) containing at
`
`least 250 mg of fulvestrant in a volume of vehicle that is tolerable for injection, i.e., no
`
`more than 5 or 6 mL.
`
`13. Because of the extremely low solubility of fulvestrant in water, a reasonable starting point
`
`would have been to investigate intramuscular injection of an aqueous or oil suspension of
`
`fulvestrant. However, the formulator would have found that injection of an aqueous
`
`suspension of fulvestrant resulted in extensive local tissue irritation at the injection site as
`
`well as a poor release profile, such as reported in paragraph [0042] of the Evans
`
`Application. Since suspensions thus were not an acceptable option for fulvestrant, the
`
`experienced formulator would have moved on to further explore whether 250 mg of
`
`fulvestrant could be solubilised in no more than 5-6 mL of an oil-based vehicle, z'.e., to
`
`achieve the target fulvestrant concentration of at least 45 mg/mL.
`
`14. In the preformulation phase, the experienced formulator would have conducted a
`
`literature review or otherwise would have become familiar with commercially marketed
`
`inj ectable formulations, particularly injectable sustained release formulations of steroids
`
`or other relatively insoluble compounds such as those listed in Table 1 of the Evans
`
`Application, with the objective of identifying potential oil vehicles, co—solvents and other
`
`excipients that already had been found to be tolerated and/ior to have passed through
`
`regulatory review, and which might be candidates for further consideration and testing for
`
`the fulvestrant formulation. This review also would have provided guidance with respect
`
`to concentration levels of such co~solvents and other excipients that generally had been
`
`found acceptable in sustained release oil-based intramuscular injections administered to
`
`DB1/620426061
`
`5
`
`|nnoPharma Exhibit 1020.0005
`
`

`

`humans. This objective is confirmed, for example, in Nema (1997) at page 166:
`
`Generally, a knowledge of which excipients have been deemed
`safe by the FDA or are already present in a marketed product
`provides increased assurance to the formulator that these excipients
`will probably be safe for their new drug product. Regulatory
`bodies may View an excipient previously approved in an injectable
`dosage form favorably, and will frequently require less safety data.
`
`The purpose of this Nema paper was thus “to present the various excipients that have
`
`been included in the formulation of injectable products marketed in the USA.“ Similar
`
`objectives were intended to be served by the compilations of commercial formulations in
`
`Strickley I (1999), Strickley II (2000) and Strickley HI (2000):
`
`This compilation will also be useful for those interested in
`knowing what additives are currently used in inj ectable products
`and at what concentrations they are administered in practice. This
`compilation only focuses on marketed formulations and does 11ot
`delve into the subject of preclinical or drug discovery formulations
`associated with early-stages pharmacokinetics or proof—of-concept
`pharmacodynamics, where the formulation scientist is not bound
`by regulatory constraints.
`
`(Stricklcy I (1999) at 324).
`
`Powell (1998) similarly states at page 238 with respect to its compilation of commercially
`
`used excipients:
`
`Thus, the formulation scientist is often faced with a dilemma --
`
`which excipients are truly available for use (based on what has
`been used previously), and which are not? And at what
`concentrations, and by what route?
`Herein are listed the excipients found in most of the approved
`and marketed parenteral formulations, given systematically by
`excipient name. In this format it is easy to detennine what
`concentrations were used, the route of administration, the main
`
`rationale for addition of that excipient, the drug that was
`formulated, the manufacturer, brand name, etc.
`
`15. From the literature review, the formulator would have noted reference to a number of
`
`intramuscular injectable sustained release oil-based steroidal formulations that had been
`
`“ Nema (199?) does caution, however, that there is no guarantee that the new drug product will be safe as excipients
`are combined with other additives and/for with a new drug, creating unforeseen potentiation or synergistic toxic
`effects.
`
`DBl;’62042606.l
`
`6
`
`|nnoPharma Exhibit 1020.0006
`
`

`

`commercially marketed:
`
`0
`
`Strickley I (1999), Table V II:
`
`Haloperidol Dccanoatc;’Haldol dccanoatc (50-100 mgimL in sesame oil, benzyl
`
`alcohol 12%);
`
`Testosterone Enanthate;'Delatestryl (200 mg;"mL in sesame oil, chlorobutanol 5
`
`mgfmL);
`
`I PDR (1973) at pages 1277-1278
`
`Proluton/progesterone (50 mg/mL in sesame oil, 150 mg/ml benzyl benzoate, 5
`
`mgfml benzyl alcohol, 1 mgfml propylparaben);
`
`o
`
`PDR (1973) at pages 1349-1354
`
`Deladumone!Test0steronc Enanthate & Estradiol Valerate (90 & 4 mg/‘mL in
`
`sesame oil, 0.5% chlorobutanol);
`
`Deladumone OBfTestosterone Enanthate & Estradiol Valerate (180 & 8 mg/mL in
`
`sesame oil, 2% benzyl alcohol);
`
`Delalutin/hydroxyprogesterone caproate (250 mg/mL in 52% Castor oil, 46%
`
`benzyl benzoate, 2% benzyl alcohol);
`
`Delestrogen/estradiol valcrate (20 mg/mL in 78% Castor oil, 20% benzyl benzoate,
`
`2% benzyl alcohol and 40 mg/mL in 58% Castor oil, 40% benzyl benzoate, 2%
`
`benzyl alcohol);
`
`Delatestryl/Testosterone Enanthate (200 mg/mL in sesame oil, 0.5%
`
`chlorobutanol);
`
`Delaluteval 2X:’hydroxyprogesterone caproate & estradiol valerate (250 mg/mL &
`
`5 mg/mL in Castor oil, 45% benzyl benzoate, 1.6% benzyl alcohol);
`
`0
`
`PDR (1973) at pages 1391-1392
`
`Prolixin Enanthate/Fluphenazinelinanthate (25 mgfmL in sesame oil, 1.5% benzyl
`
`alcohol);
`
`0 Wang (1980):
`
`Depo-Testosterone/testosterone cypionate (100 mg.-i’mL in 87.4% cottonseed oil;
`
`0.1 mL benzyl benzoate, 9.45 mg benzyl alcohol as a preservative);
`
`0 Mackey (1995):
`
`Testoviron Depotftcstostcrone enanthatc (250 mgz‘mL in castor oil and benzyl
`
`DB1z'62()42(106.l
`
`7
`
`|nnoPharma Exhibit 1020.000?
`
`

`

`benzoate);
`
`as well as a number of other commercialized oil based long—acting IM injectable
`
`formulations reported on Table 1 of the Evans Application.
`
`16. As a further part of the prefonnulation phase, the experienced formulator would have
`
`conducted a preformulation solubility screen, separately measuring the solubility of
`
`fulvestrant in a range of pure solvents, including the potential oil and co—solvent
`
`candidates that had been identified in the above literature review as being suitable for
`
`inclusion in intramuscular injection formulations. See, for example, Gupta (1999),
`
`Chapter 17 at page 402, under the heading “Formulation Development”:
`
`The activities necessary to develop a parenteral product can be
`placed into the following three broad areas: preformulation,
`formulation, and scale-up. While there are alternative development
`perspectives, all development ultimately needs to accomplish the
`same activities. Preformulation includes the characteristics of the
`
`bulk drug plus initial screening for excipient compatibility with the
`drug.
`
`“Prcformulation studies” are said to “provide fundamental data and experience necessary
`
`to develop formulations for a specific compound” including, as item 8.1 in the outline of
`
`areas of specific interest, a determination of “solubility” in “selected solvents” (at 403).
`
`“Significant formulation activities begin with initial preformulation data and knowledge
`
`of the specific route of administration” (at 405), which “formulation activities include the
`
`identification and selection of a suitable vehicle (aqueous, nonaqueous or co—solvent
`
`system) ...” (at 404). It is further noted that “injection volume is one of the most
`
`important considerations in the formulation development of a commercial product” (at
`
`405). \Vl1en carrying out such a preformulation solubility screen with fulvestrant, the
`
`formulator would have found that fulvestrant had extremely low solubility in water, low
`
`solubility in most oils (but highest in castor oil), low solubility in benzyl benzoate, and
`
`the highest solubility in ethanol and benzyl alcohol, such as reported in Table 2 of the
`
`Evans Application.
`
`17. With the information on prior commercialized formulations and the fulvestrant solubility
`
`data from the prefonmulation screen (such as reported in Table 2 of the Evans
`
`DB1/'62U4?.606.3
`
`3
`
`|nnoPharma Exhibit 1020.0008
`
`

`

`Application), the experienced formulator would have selected castor oil as the oil vehicle
`
`because of the higher solubility of fulvestrant in castor oil relative to the other oils tested.
`
`Nevertheless, he would have appreciated that the target fulvestrant concentration of at
`
`least 45 mgx"mL could not be achieved with castor oil alone, and that a co—solvent would
`
`be required.
`
`18. A number of the commercialized fonnulations that would have been identified in the
`
`literature review (including the castor oil-based formulations) have a sub stantial benzyl
`
`benzoate component, which may be present as a co-solvent. See, for example, Delalutin
`
`noted in paragraph 15 above, which is reported in PDR (1973) and noted in Table I of the
`
`Evans Application, and is one of the fonnulations discussed in Riftkin (1964), “Castor
`
`Oil as a Vehicle for Parenteral Administration of Steroid Hormones” (see Riffkin n. 6).
`
`Delalutin is 250 mg/mL l7—hydroxyprogesterone caproate dissolved in 52% castor oil,
`
`46% benzyl benzoate and 2% benzyl alcohol. However, Riffldn Table 11 reports that the
`
`solubility of 17-hydoxyprogesterone caproate in castor oil alone is only 55.6 mg/'mL, but
`
`the solutility of 17-hydroxyprogesterone caproate in benzyl benzoate is substantially
`
`higher, being at least 250 mg/mL (see example 4 of Huber H} S ‘520) and Attachment E
`
`discussed below). Even if not needed as a cosolvent, Riffkin (1964) notes that “the
`
`addition ofbenzyl alcohol or benzyl benzoate to castor oil resulted in a lower and more
`
`favorable viscosity, making it easier to inject” (paragraph bridging pages 893-894).
`
`19. However, the skilled formulator would have appreciated from the fulvestrant solubility
`
`data generated in the preformulation screen that fulvestrant had very different solubility
`
`characteristics relative to the steroids of previous commercial formulations. Attachment E
`
`is a compilation showing the chemical structures and relative solubilities in castor oil and
`
`sesame oil of the compounds named in Rifflcin (1964) Table II compared to the structure
`
`and the solubility of fulvestrant in these oils. It can be seen that the solubility of
`
`fulvestrant in castor oil and in sesame oil (20 mg/mL and 0.58 mg,fmL, respectively, from
`
`Table 2 of the Evans Application) is appreciably lower than the solubility of the other
`
`steroids in these oils (taken from Table II of Riffldn (1964)). The second page of
`
`Attachment E tabulates the concentration in benzyl benzoate of five named steroids, taken
`
`DBlv£-2042606.]
`
`9
`
`|nnoPharma Exhibit 1020.0009
`
`

`

`from Examples 1-5 of Huber (US ‘520), ranging from 200 to 400 mg/c’ml.2 By
`
`comparison, the solubility of fulvestrant in benzyl benzoate is reported in Table 2 of the
`
`Evans Application as being only 6.15 mg/mL, and only 3.8 mg/imL as determined in the
`
`recently conducted tests reported in Attachment C.
`
`20. The experienced formulator thus would have expected that benzyl benzoate would not act
`
`as a co-solvent for fulvestrant in castor oil because the solubility of fulvestrant in benzyl
`
`benzoate was significantly lower than its solubility in castor oil. The addition of benzyl
`
`benzoate to castor oil, for whatever reason, would have been expected to decrease, rather
`
`than increase, the solubility of fulvestrant in the resulting castor oil/benzyl benzoate
`
`mixture. This is confirmed in Table 4 of the Evans Application, which reports a
`
`fulvestrant solubility of only 12.6 mg/mL in the castor oil Vehicle containing only 15%
`
`benzyl benzoate, compared to the 20 mg/ml. solubility of fulvestrant in castor oil alone as
`
`reported in Table 2.3
`
`21. Based on the solubility data determined in the preformulation screen (such as reported in
`
`Table 2 of the Evans Application), ethanol and/or benzyl alcohol would have been seen as
`
`the best co—solVent candidates for raising the fulvestrant solubility to the 45 mg/mL target
`
`in the castor oil Vehicle, and would also function to lower the viscosity of the resulting
`
`formulation and make it easier to inject. Consistent with this solubility data, Dukes (US
`
`‘8l4) added 40% W/v benzyl alcohol in order to dissolve 50 mg/mL fulvestrant in the
`
`castor oil-based formulation used in the experimental rat studies of his Example 3. It thus
`
`would have been apparent that 40% W/V benzyl alcohol could fimction as a co-solvent in
`
`castor oil to achieve the target fulvestrant concentration. Nevertheless, the skilled
`
`formulator would have been concerned with using such a high alcohol content in
`
`intramuscular inj ectable formulations for administration to a human.
`
`2 Data taken from the Examples of Huber (US ‘520); these are concentrations used in the examples and not
`necessarily the actual maximum solubility of each steroid in benzyl benzoate, which may be higher. Huber was a co-
`author on Riftkin (1964).
`1 It should be noted that in the further tests that were recently conducted under my guidance (paragraphs 7-9 above
`and Attachments B and C hereto), the solubility of fulvestrant in castor oil alone was again tested and found to be
`21.4 mg.=’mL, and the solubility of fulvestrant in benzyl benzoate alone was again tested and found to be only 3.8
`mgfmL, which further confirms that benzyl benzoate would not be expected to act as a cosolvcnt for fulvestrant in
`castor oil.
`
`DB1:s2o42sum
`
`10
`
`|nnoPharma Exhibit 1020.001O
`
`

`

`22. First of all, the experienced formulator would want to minimize the amount of co-
`
`solvents and exeipients in any injeetable formulation. For example, as stated in Gupta
`
`(1999), Chapter 17, “Formulation and Administration Techniques to Minimize Injection
`
`Pain and Tissue Damage Associated with Parental Products” at page 414:
`
`Cosolvents are commonly used to enhance drug solubility and
`stability. Cosolvents may include ethanol, propylene glycol,
`polyethylene glycols, and glycerine. These components have
`intrinsic effects on biologic tissue and can alter the properties of
`other excipients, thus influencing the tissue damage or pain caused
`by a product. There is a dearth of literature on the pain caused by
`cosolvents, but there is also a growing body of knowledge on the
`tissue damage that they can cause. It is not certain that tissue
`damage is always directly correlated with the injection pain, but
`minimization of both pain on injection and potential for tissue
`damage should be included in the product development plan.
`
`See also Gupta (1999), Chapter I 1, titled Cosolvent Use in lnjectable Formulations, page
`
`217:
`
`Ideally, it is best to select and use solvents that would maximize
`the solubility of the compound. Maximizing the solubility of a
`compound in a paiticular eosolvent system would result in lower
`total levels of the non-aqueous solvent(s) being administered to the
`patient, thereby lowering the chance for potential side effects.
`
`This objective would have applied to aqueous and oil-based systems alike, in that the
`
`precedent of commercialized formulations identified in the literature review would have
`
`confirmed that fixed oils, such as castor oil, have long been commercially used and
`
`accepted as the major component of oil-based sustained release intramuscular inj ectable
`
`steroidal formulations. On the other hand, co-solvents such as ethanol or benzyl alcohol
`
`have generally been used only in far lesser concentrations, as discussed in the following
`
`paragraph.
`
`23. Thus, use of such a high content of either benzyl alcohol or ethanol would have been
`
`contrary to precedent as shown from the review of commercialized oil-based
`
`intramuscular injectahle sustained release formulations. The literature review as of early
`
`2000 would have shown that any benzyl alcohol in such formulations was almost always
`
`DB1;’62042606.l
`
`1 1
`
`|nnoPharma Exhibit 1020.0011
`
`

`

`present as a preseivative in a concentration of about 2% or less, occasionally at a
`
`concentration of up to 5%, but only rarely at higher concentrations. With respect to
`
`benzyl alcohol see, for example:
`
`0 Gupta (1999), Chapter 11 at page 229 stating that benzyl alcohol “is typically used in
`
`concentrations of up to 2 percent as a preservative and up to 5 present as a solvent,”
`
`and then discussing reported toxicities.
`
`0 Nerna (1997), Table V at page 168, reporting that benzyl alcohol was present as an
`
`antimicrobial preservative in 74 injectable fonnulations (not limited to oil—based IM
`
`formulations) at concentrations of from 0.75-5% (note that benzyl alcohol is not
`
`included at all in Nema Table I, “Solvents and Co—solvents”;
`
`0
`
`Powell (1998), the benzyl alcohol listing at pages 244—246, particularly those
`
`indicated as being used in IM formulations;
`
`0
`
`Strickley I (1999) at page 329 notes the inclusion of 2% benzyl alcohol in an IM
`
`lorazapam formulation in a propylene glycol vehicle, but does not include benzyl
`
`alcohol at all in Table VI listing “Cosolvents Used in Parenteral Formulations;”
`
`0 Lopatin (1972) noting in Table 3 at page 727 opposite Benzyl alcohol, “Toxic. Used
`
`in concentration of 11ot over 3%. Has irritant action in concentration of 5%,”
`
`0 Cornelius (US ‘863), col. 1, lines 30-35 stating, “It is known that the solubility of
`
`steroids in vegetable or animal oils can be increased by the addition of excipients such
`
`as benzyl alcohol and benzyl benzoate. An objection to the use of such excipients, and
`
`specifically benzyl alcohol in somewhat higher concentrations, is that these agents
`
`may irritate the tissues.”
`
`The literature review as of early 2000 also would have shown that, with few exceptions,
`
`ethanol was not included in such fonnulations in excess of about l0‘/o. See, for example:
`
`0 Gupta (1999), Chapter 11 at page 225 noting that ethanol has been used at levels up
`
`to 50 percent, but these levels typically are associated with pain on injection;
`
`0 Strickleyl (1999), Table VI, “List of Cosolvents Used in Parenteral Fonnulations”
`
`more specifically lists the ethanol content in IM formulations for specifically
`
`identified drugs, which concentrations range only from 2.5 to 10%; an Il\/I/IV
`
`lorazapam formulation in a propylene glycol Vehicle is noted at page 329 as having
`
`13% alcohol, but is not included with the IM formulations in Table VI;
`
`DB1f62042606.1
`
`12
`
`|nnoPharma Exhibit 1020.0012
`
`

`

`0 Nema (1997), Table I, “Solvents and Co—solvents" at page 167, lists ethanol as being
`
`in 24 fonnulations with a concentration range of 06-80% (for Prograt); note that this
`
`is misleading, however, since Prograf is a concentrate for intravenous infusion only,
`
`and is to be diluted 250 to 1000 times before administration;
`
`0 Powell (1998), lists “alcohol” at page 242 and “ethyl alcohol” at page 255, wherein
`
`the ethanol concentration for [M formulations ranges from 0.61 -10%.
`
`24. Thus, even though Dukes (US ‘8l4) had demonstrated that the target 45 mg/mL
`
`fulvestrant concentration could be achieved by adding 40% benzyl alcohol to the castor
`
`oil vehicle, the precedent of commercialized IM oil—based systems would have motivated
`
`the experienced fonnulator to substantially reduce the benzyl alcohol content of the
`
`formulation intended for human use, and this commercial precedent would have made
`
`him very reluctant to replace benzyl alcohol with the substantial amount of ethanol that
`
`would be needed to maintain the target fulvestrant concentration. Benzyl benzoate clearly
`
`would not be considered to solve this dilemma, but rather would be expected to have a
`
`negative effect on fulvestrant solubility since fulvestrant was even less soluble in benzyl
`
`benzoate than in castor oil, that is, one would have expected that adding benzyl benzoate
`
`would require still more alcohol to maintain the target fulvestrant concentration.4
`
`25. Under these circumstances, the discovery by Evans et al., that the addition of benzyl
`
`benzoate to the castor oil/alcohol mixture actually increases the solubility of fulvestrant
`
`such that more fulvestrant could be dissolved in a given volume of formulation, was
`
`unexpected and truly surprising. This positive benzyl benzoate effect on fulvestrant
`
`solubility in the resulting formulation is shown in Table 3 of the specification (and is not
`
`changed by the above-noted corrections), and is confirmed and demonstrated over a
`
`broader range of formulation composition by the additional set of experiments conducted
`
`under my guidance and discussed in paragraphs 7-9 above, the results of which are
`
`reported in Attachments C.
`
`4 It should be noted that even apart from this solubility issue, there would have been no motivation to add benzyl
`benzoate for Viscosity reduction since the significant quantity of alcohol would serve the dual function of acting as a
`co—solvent as Well as reducing the injection viscosity and making it easier to inject, whereas the benzyl benzoate
`would be expected to have a negative effect on the fulvestrant solubility.
`
`DB1f62042606.l
`
`13
`
`|nnoPharma Exhibit 10200013
`
`

`

`The undersigned declares fuither that all statements made herein of his own knowledge are true
`
`and that all statements made on information and belief are believed to be true; and fiirther that
`
`these statements were made with the knowledge that wilful false statements and the like so made
`
`are punished by fine or imprisonment, or both, under Section 1001 of Title 18 of the United
`
`States Code and that such wilful false statements may jeopardise the validity of the application or
`
`any patent issuing thereon.
`
`R Saw
`
`Date;
`
`-
`
`DBl;‘620426G6.1
`
`14
`
`|nnoPharma Exhibit 1020.0014
`
`

`

`TABLE OF REFERENCES
`
`Author/Inventor
`
`Reference Citation/Patent
`
`Cornelius (US ‘863) US Patent 4,212,863
`
`Dukes (EP ‘014)
`
`EP 0 346 014 Al (corresponds to US Patent 5,183,814)
`
`Dukes (US ‘814)
`
`US Patent 5,183,814 (corresponds to EP 0 346 013 A1)
`
`Gupta (1999)
`
`P.K. Gupta and G.A. Brazeau (eds). Injectable Drug
`Development: Techniques to Reduce Pain and Irritation. Chapters 11 &
`17 lntelpharm Press, Denver, Colorado (1999)
`
`Huber (US ‘520)
`
`US Patent 3,164,520
`
`Lopatin (1972)
`
`Maekey (1995)
`
`Nema (1997)
`
`PDR (1973)
`
`Powell (1998)
`
`Riffkin (1964)
`
`P.V. Lopatin, V. P. Safonov, T. P. Litvinova and L. M. Yakimenko. Use
`of nonaqueous solvents to prepare injection solutions. Pharm. Chem. J.
`6:724-733 (1972)
`
`M.A. Maekey, A..l. Conway and DJ. Handelsman. Tolerability of
`intramuscular injections of testosterone ester in oil vehicle. Hum.
`' Reprod. 10: 862-865 (1995)
`
`S- Nema, R.J. Washkuhn, and R.J. Brendel. Excipients and their
`use in injectable products. PDA J. Pharm. Sci. Technol. 511166-71
`{ (1997)
`Physicians ’DeskRef2rence (27th edition). 1277-1278, 1350-1354, 139l—
`1392 Medical Economics Company, Oradell, NJ (1973)
`
`g M. F. Powell, T. Nguyen, and L. Baloian. Compendium of excipients for
`parenteral formulations. PDA J. Pharm. Sci. Technol. 52:23 8-311 (1998)
`
`C. Riffkin, R. Huber and C.H. Keysser. Castor oil as a vehicle for
`parenteral adrninstation of steroid hormones. J.Pharm.Sci. 53: 891-5
`(1964)
`
`Strickleyl (1999)
`
`R. G. Strickley. Parenteral formulations of small molecule therapeutics
`marketed in the United States (1999) —Part I. PDA J.