PATENT
`.:‘XT'”I‘ORNEY DOCKET NC); {}S€'i291-«’50i)4-E11
`
`IN THE UNITED STATES P.s’§.’I.‘ENT JKND 'I‘R;~‘kI)EMARI€. UFFICE
`
`In re PRTENT APPLlC.§-*s'f£{}N sf:
`
`} Cenfiniwetitan ?."si£:2.V
`
`2393
`
`EVAKS at :11.
`10fS‘72,'3'34
`Applicatian Nu:
`Filed:
`June 22}, EH04
`FQR:
`FGRMEELATIEJN
`
`I461’?
`i Czmup Art Unit:
`; Examiner: Hui, San—Ming R
`g
`
`§)ECLARATI()?‘E UNBER 35 U.(S.C.§ § 1.132
`OF PAUL RICE-IARI} GELLERT
`
`PAUL RICHARD (BELLERLT cf Sxstraéiantzca, Adderley Parkfi Macciastieid, Cheshire, UK
`
`declares:
`
`1.
`
`i graduated fnzzrn that Uzziversity €11?" Clxferci in Ciwenzistry in 193%. 1 uxxécrtouic pafitgtraéuate
`
`research with Frnfesgor Brian Howard in the Phyisicai Chemistry Laburatery at the
`
`University ::’:t‘lZZ1::,¥*'::1rri ieaaiing tea the: awarci afa I.).E’hii in $988‘ ‘F‘mm Fatbruary 1983 umié
`
`the present I haw been employed by Astra.?,e11r:£:a,(fe3rzner1y Zenaua and ICE} initially as a
`
`Scsniasr Reeaearcla Scimtist and suhsequeaxtly as a Team Leanivarflvtanaggézr, Principal
`
`Scientist and. aincs:-: 2004:, :3 Senim: Prittcipal Seieiltist.
`
`2.
`
`I have warked in 111% fiinrmuiatmn amii cimg deeriivery araa thrcaughaut my tzarear with
`
`AstraZer:eCa§ xvlwre my research anti dcwelapment mark. has €“:€3~’&-’£.’I‘€2€1
`
`a range :31’
`
`furmulatinn types including sustaineci rcieaazed injectiens, inclmtii-;:1g fuivestram.
`
`3. During the wursivsz tjfmy atudy cf‘ the: subject a;}p1icatir3n(here:iné1t“t€:r "the: Evans
`
`.r1.pp1iCati<3n"‘} and the underlying data, E have become aware emf’ rzervczragl transcriptian cur
`
`{rather ermrs Eaetwezan certain disclosures :31” the subject appiicati-mi and the undarlying
`
`IEi§J{}}’3{frI'_}FI3{)tt°E1'3€}f.1k
`
`tiara, fine purgfrmie ::Jffi"2is. Desciaratienn is m ptaim nut the *3)i§S{£’:I‘£L‘-£21‘
`
`HE I.’f13E.M2§:{!E:. 1
`
`Astrazeneca Ex. 2135 p. 1
`Mylan Pharms. Inc. V. Astrazeneca AB IPR2016-01325
`
`

`
`and mature eat’ these setters amt te report furttter testing that has been earrieti out untier my
`
`guiéarzee tu tahtain atiditétma! data (paragraphs; 4» if} beiew anti Attachments A-Q). A
`
`further perpese et‘ this Deeiaratiers is tax
`
`em and titactxment the manner in which an
`
`experienced fertnulator emttid tikety have appmaeheti the task :1? tttwetopittg 2: smttttineti
`
`teleasae ittgeetabie fetttmlatien sttttabte fut htmtatz use {er a etetettial s::c3m;:}Qt1.ttt;t such as
`
`tulveettattt in zttmut early 2300, which i ttttdersttmtfi is Whtttt the ptierity‘ appiicatitms
`
`snppertittg the Elvants Appiteation were fi1ed(parag,tapt1.~.t 11 ~ 25 beiew and Attachtneet
`
`El). Citetiens te literature and patent referettces in this Deeiatatien wiii be in the fomzat
`
`Lead Auttmr {Date}, and the full citatiem-3 are gittett in the Table of References; at the end
`
`at this Deelaratien. A copy at" each eited reference (at etted ptartimts at‘ the Eonger
`
`references} ts; included ‘tn Attachment F‘ under the Ttth {1t1t‘t‘13:}ti‘}‘ note-ct in the Table at‘
`
`Re fetettces,
`
`In Tame 2 set the Evans Applieatitm, the ttelnbiiity est‘ fitlvesttant in caster tail appears to
`
`ha we been tranemihed ineerreetty frtmz the etiginai tteutee, the iaberatery tteteheek. Ttte
`
`value in thtt Iattet is 24.5 mgimt and net 2%} mgftttl. In ether ettperitnestts to determitte the
`
`stfltztaiiity ref futveetrant in caster sail and also in bettztyi benaeeate, same variabiiity was,
`
`observed.
`
`In Table 3 of the Evans Application, the given eelubility values were generated at 4"€Z
`
`ttnti tam at 25"C as is stated in the title of "t‘at::1e 3. For fttlvesttant fommlatiene, it is
`
`preferabte that the fillvestrant remains eemptetely in soiutimt at both 4"C and 25°C.. The
`
`4"(Z temperature eerrespends te the sterage temperature {2“C to 83C in the FDA approved
`
`Label fer Fasleéex}, and the 256C» temperature etzrtteepende to the admiztisttatitm
`
`temperature (ambient temperature}. In aciditiett, the stpetzifiectt selttlaility values mt this
`
`Tahie 3 are mezattt vatttes eaieutated from ettttlysie t:«t’t'epIit:ttte samples fttzttt (me in‘ mere
`
`trials. The itttiiwsitittat values are etttwxm in handwriting in the amemitetzi ttersitm {sf Tattle 3
`
`in Attachment A. In atltiitiett, it appears that the mean vaittee for the East three
`
`eettzpeeitiene have been ineetreetijs; eaimtlated. the etarreeted mean values, together with
`
`the etzrreetien of the temperature {mm “25'“C“ be read “x‘-i"‘C"1 are ttise shown in
`
`tteandttstiting in the axtteettted versitm of "fahle 3 in Attatztnfnent. 2%..
`
`€33 I :’E'13:t'I-43-f1t}f1 L i
`
`ta?
`
`Astrazeneca Ex. 2135 p. 2
`
`

`
`1 havé mrainaxaci :11:-.: transcriptimn anti rather rsrrmrs agaémfit ms: miginai appiieatinxx
`
`dimlnsures and cmmiucie that theme do am changt: the uitimatxs: cnnciizsions made fmm £316.-.
`
`data as: miginally repmted. This: adciitiiazx {taf 15% WW bzengyi henzcsate m r:csmposi1;im2$
`
`hzwitzg tmsfl akrolmi uzcmcentratiens in caster 013 {sf iEZr‘?x‘E;, if:%. 28% 31111 3{}% xvhs
`
`umzxpatczttrdljy prtzvitiass 3 peaiiive effeci an fixlwsstmw; miubiiéty, significantly increasing
`
`the rmiubility :3?’ fulvestrant in the compositimns despite fhlvestzrani; having a Imwcr
`
`snalixbiiity in betzzyl banzoate than in either alcnhztzk :31“ casmr {‘.‘.vii.
`
`An ariditinnal 3:21: {sf experiments has been ccrziciuctmi at 3.5%‘. under my guidance to
`
`mam cangiszazni data with reduced variability fffllfl a sing}:-;: set :3?“ rigamusigr tzcsmrcsiifld
`
`sszaiuizxfiity fixgzerimenta and :9 citzmenatrata {hat the unexpecteei izzcreasts crf soiubility csf
`
`fuivtmtrant by adding b£3:‘12,jfl benzoate into <:{3mp<:xs&tions ccmtainiag mhanxai, bcnzyl
`
`almhmi 3:11:23 caster -ail, is present acrcsss the Ezmacicr range: :.:>f’c:>:np<}3it§nn emrampaszsed by
`
`the vziiaimfi being presenied with {his ilaciaration. T136 :~:0i1.:biLIit1i xi}? fxiivmtrant in hanzyi
`
`h::::1:e:nate and in casmr mi was also measumci in the same 3% cf ezuzparimcms using the
`
`smmt: bait}: azfibxztrzzyl bmnmate and?‘ the same batch Kiif casmr mil as were used ta make up
`
`me mm positiaomg. The Experimental Tea: Pracatdure: is diifiéffifibtifi in Attachmem B~
`
`”.I”hr.-3 results fmm these: sniubiiity s:::::perimm1ts are smczwn in the tabh‘: in Attachment C.
`
`These results shew that the salubiiity cf fhlvestrant in caster ail aians: (21.4 mgfml’) is
`
`significantiy greater than the sczxlubiiity cf fuivc-zfitrant in bmsszyl btznzzcrate aicma (3.8
`
`mgfzxzi) and ciem<:m:~'strate the uziexpecied increase in fuhn;-strant smiubility an the adéitinn
`
`:3? If}? 15 mad 25% was benzyl banzmte, in glans xzsfzm equivalent ammmt mi‘ caster ail? ta
`
`:3-i}II’3§}£}S$iZ§(3nS$ having mtai alcohol ccancxantraiinns in castor oil :3"? I{}%, 15%, 28%, 25%
`
`and 3{}"’-Exéz Wm
`
`Thug, the r:;:s.:1¥::s that ware: mbtaimzrd fmm exp::rim::nm czctrnziuxztmi under rigxzumusiy
`
`rzsrmiroiltzci czsnfiitimns anti with an expanded I“£i£3g§i’*<}fi3{33fI1§1i3$iiiC>fiS$, as Sh(.3W'i‘I in
`
`Attaahxnetxt if, confinn the ultimate canciusinns drawn mam the resuits; -shuwn in Tabk: 3
`
`£;;ft11::: Uriginai appiicatian dimin:)$u.re:, namaiy that the aci::1iti::::1 Qf H}‘.’x‘E~; in .2:3% wfv benzyl
`
`{Z313 1:{:3€1~$2§:€}fa.
`
`E
`
`Astrazeneca Ex. 2135 p. 3
`
`

`
`bemsaate tar; izonmasitiuns l1:;wingii}tal almlwl C()11CflEl'£i‘a’fliJE'IS in caamr ail €24‘ bewseen
`
`10% to 30% wfv uslerxpectedly provides a pasitive effect 011 fulvcstrant scxlubility,
`
`significantly incraaaiszg the sulubility ef fhlveaztraxxt in the cumpositimxs deapita
`
`ftxlvestrant having :21 lczswer solubility in bemxyl bemaate than in eithe.-‘:r aiicahal €31” castzyr
`
`ail.
`
`10.
`
`{luring aha cfirurse: wf my study atzsf {he Em.-Ina Agzplicatiim and the underlyring Source
`
`materials it was drawn ta» my attanticjm that same cf” the compesiticyn clam given fur
`
`Ilelestmgen Eiflii Delalutin swmehzrrw llad been sllified we C€3l!.1f?£“£l1
`
`tn the right. Thus, fitsr
`
`Daiesimgen, the 73% and 5§.l% figures. ahowri under the B233: ::s::~lu:1'm shauid have been
`
`under the OIL ceiumn; me 20% and 4€}“r’£: figurea shawn under that 330%! mlumn shnuld
`
`have been under tin: E32232 Cfllumii; and the 2% figures alarms: uraciitr l€tC}H shcnzicl have:
`
`been umiar the B:e;C)I:I eealumnl Similearly fur Delalutin, the “up tn 2%" filmwn under til»;
`
`EtOH mlumxz slmuld have been under the E22101-1 wlumn. This table regmrts that the
`
`source nf this data was J.Phan1i1.Sci (1964) 53(8) 391, which is: Rifflcm (W54) elsewhere:
`
`referred III! in this llleclaratitrm, anzl I have aim verified the c::nrre3ct€:d data fmm the entries
`
`fur lfmlalutln and Qelestragen in FDR (lԤ3'?3). fix cup}? sf Table: l fmm the Evans
`
`Applicaticxn is; repruduced as Attaulximrnt I), cm which theas: carreetitztnat haw: best: made
`
`in harxdwxitlng, arid K have additiimaily mart: czzarectly ncyteci {hat Elalalutin is I7-hydrtrxxy
`
`pmgestemne rsaprmste, ané tlxzzt the “C€)MP"’ ciesignatian far D‘t§l.£lllJti1i’£ Eihiiuld be “Bl\«'[S“
`
`{Bristal—Myer:s Squibb). attachment 1) 31:50 includes a {ma page explanatitan af‘ the
`
`c0£1*ectic::1s :0 this ‘Fable 1.
`
`ll.
`
`In abczsuz early 2906, a perssan refipanaihle Fm‘ daveluping E1 sustaitieri release in_§:::ctal:J¥<:
`
`fatmnulation sauitable fer aciminisiratian UL} humans far a new szeruidal campnund such as
`
`fuivemz-ant, waulcl haw hat? specialized training and experience in davelapiug
`
`phamlasaeutical fuzmulaticna and I1’Et3fl'1t"J(LlS Fur their ailministratimi, In develaping slush 3
`
`formulaziazm fair: fulvestrant, the cvbjectiwi: wcxulsi have been ta fcrmulatta an lf1tfE1TI'£t1S£2L1lf:*1f
`
`UM) injectiarr that would prmside for the satisfactary sustaimtd releaxc ssflfhlvearirant over
`
`:1 parted at‘ at la.-as: zwa weeks and preferably sweat a pericci ofzatt Ieassiz fcyur weeks is
`
`rss:i11::e: the fr:-'::;ue11c.::,«‘ crf administratiatx. and wszmld have a target fitlvetsttrant mutant :11’ at
`
`$1123 I :€;?.«:m:?;0z:«.a
`
`.4
`
`Astrazeneca Ex. 2135 p. 4
`
`

`
`12_
`
`13.
`
`least 45 mg,§mL so 213 to provide a fizlxsestrant dose of at lea31:25O mg in a single 5-6 mL
`
`injection. From my personal experience and knowledge of the iiterarure at about that
`
`time, I believe that such an experienced forrnnlaror would likely have approached the task:
`
`of developing a formulation for fuliresrrnnt in about the following manner.
`
`Given the foregoing objective, the experienced forinolator would have appreciated that
`
`the traditional administration options to explore were inrramuocuiar (Hvl) injection of a
`
`inisrained reieaee aqueous or oil suspension or an oiI—based solution (depot) containing at
`
`least 258 mg of folvestrant in a volnmelofvehiele that is tolerable for injection, 123., no
`
`more than 5 or 5 mL.
`
`Because of the extremely low solubility of fulvesrrant in water, 3 reasonable stariing point
`
`would have been to investigate intramuscular injection of an aqueous; or oil suspension of
`
`fuivestrant. However, the formulator woniciihave found that injection of an aqueous
`
`suizpension of fulvestrant resulted in extensive local tismie irritation at the injection site an
`
`well as a poor reiease profiie} 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
`
`fulveotrant could he solubiiised in no more than 56 mi. of an oiifioaserzi vehicle, 21:2,, to
`
`achieve the target fnlvestrant concentration of at least 45 mgirnh.
`
`14.
`
`In the preformniation phase, the experienced form“-ulator would have conducted a
`
`literature review or otherwise wonkzi have become familiar with eominereiaiiy marketed
`
`injeetable forrnulaticms, gaarticulariy inj eerable sustained reiease formulations of eteroids
`
`or other relatively insoluble compounds noon on rhoee listed in Table I of the Evans
`
`Application, with the objective of identifying potential oil Vehieleii, emsolvents and other
`
`oxeipients that alreaciy had been found to be tolerated andfor to have passed through
`
`regulatory review, and which might be eanciiontes for fi.irLhe1' eonsirierarion and testing for
`
`the fulvestrant formulation. This review also would have provided guirianee with respeer
`
`to concentration ieveis of such eo~so1vent$ anti other excipients that generally had been
`
`fonnri acceptable in sizsraineci release oil~bas;e<:1 intramuscular injections; arlminisrerezi to
`
`DB-i;(i20-'1i2€>L3€>.i
`
`Us
`
`Astrazeneca Ex. 2135 p. 5
`
`

`
`humans. Thia objective is ::m1finne<i,, for example, in Nema (199?) at page 166:
`
`Generaiiy, 3 knawlaadge efwhich excipients have been deemed
`safe by the FDA or are mrraady present in as marketeci pmduct
`prcsvides increased assurance: t0 the ibnnulator that these excipients
`will probably be safe for their new drug product.
`Reguiamry
`bodies may View an cxcipicnt prsvieusly apprnved in an injectable
`desage form fa\+’iZII'fib1"_§a’.,, and will frequently require kiss safety data.
`
`The purpose of this Nema paper was thus “to pmsent the various excipients that have
`
`been inciuded in the furtnulation of injectabie pmducts marketed in the USA?” Similar
`
`objectives ware intended to be Served by the compilations mi’ commercial farmulations in
`
`Strickhazy I (1999), Strickiey H (2000) and Strick1eyIII(20UD):
`
`in
`This ccsmpilation will also be useful for those il1IBI‘f3ST.€(3
`krmwing What additives are currsntly used in injectable products
`and at what concentraticms they are acizninistared in gzracticze. This
`cempilatimn only focuses on marketed fnnnulatiens and does not
`fieive into the subject of preciinical or drug discovery formuiaticms
`associated with Barby-stages pharmamkineticg (gr prcr<:af«£1f«cmncept
`phannacodynamics, where {hit fonnulation scientieai is not bound
`by reguiatcmy constraimsz.
`
`(Strickiey I (1999) at 324).
`
`Powefl (1998) gimiiarly states at page 233 with respect tn its compilation mfmtnmercialiy
`
`used excipients:
`
`Thus, the fonnulation scientist is ofien facfid with a ditemma -~
`
`which excipicnts arr: truly avaiiabie fur use (based on whaz has
`been usseci previously), and which are mm‘? And at what
`c{:enc:ent1*ations, and by what route‘?
`Herein are listed the excipients found in most nf the approved
`and marketed par-antrzrai fcmnulations, given syst-amaticaiiy by
`cxcipiem name. In this format it is easy in deteimine what
`concentratians were used, the route of administraticm, the main
`ratianale for addition uf that excipient, the drug {hat was
`formulatcct, the manufamturar, brand name; etc.
`
`15. From the literature review, the formulator would have ms.-ted reference to a number of
`
`éntraxnuscular injectable sustaimzd release oémzased steroidal fc::;n:nu1ation3 thai had been
`
`‘ Nszmxa (E99?) dares caution lxowever, that there is; as guaramee that the new drug preduci will be safe as excipicnts
`arr: ccfsmbintci wiih anther addiiivcs and.-‘<:vr with a new cirug, r:r<:at§ng unfsrcsecn pcmzniiaticen m synergistic toxic
`cfihctx,
`
`DB i.v'6'2{J~$26£36. 1
`
`55
`
`Astrazeneca Ex. 2135 p. 6
`
`

`
`mmmercially marketed:
`Q
`
`Strickiey 1 (1999), Tabb: ‘V11:
`
`Halcvperidcel Deczanoateffialdol decazaoate (50-1130 mgfmi, in scaame 0211, 13E‘:I‘IZjy’1
`
`31301101 1.2%);
`
`"1“estostc:'one Exzanthahzfflalavszstryl (2130 mgfnm in sesame 011, chime-butanol 5
`
`1r1gx’mL);
`
`FDR (19?3) at pages 1277-12?‘ 8
`
`Proiutmrzfprogesterone (Sf) mgfmi, in srctsamc 05.1, 150 mgfxnl benzyi benzoate, 5
`
`mgiml benzyl alcohoi, 1 mg/ml propyiparabenj;
`
`FDR (1973) at pages 13494354
`
`Deladuxhoneffestosterone Enanthate 8:. Estradiol Va1erat«::(9D 8; 4 mgfm}; in
`
`gesmne: 011, 8.5% £:hi0r(:rbutar101);
`
`Drsladumomz Qflffestastaronc Enanthate &: Estradiul Vaisraie (181) 8: 8 mgfmL in
`
`sesame Q11, 2% benzyl alcohol);
`
`1)e1a1utinfhyciroxypmgesterone caproate (250 mgfmli. in 52% cagmr 011, 46%
`
`benzyi benzcuatet, 2% benzyl alcohol};
`
`Defestrogezmfestraciioi valrzrate: (20 mg/mL in 78% castor oil, 20% he-enzyl bcnzcate,
`
`2% hanzyi aimhol and 40 mgfrnl. in 58% zzasmr oil, 40% benzyl henzoate, 2%
`
`benzyl a1cnho1);,
`
`DelatestryiffF3$StQ$1ZE1‘0I1€i: Enanthate (290 mg/ml, in sesame 011, 0.5%
`
`ch1c:n:*{:sbntano1};
`
`Delaluteval 2Xfhydroxypmgesterone capmate: & estradicol vmerate (250 mgfmll &
`
`S mgfmL in caster Q11; 45% benzyl benzmate, 1.6% benzyl aimhoi);
`
`FDR (19?3) at pages 1391-1392
`
`Pmlixin Enanthatefmuphenazinefinanthate (25 mg;"mL in sesame oih 1.5% benzyl
`
`alcahol};
`
`Wang (1980):
`
`Depo-Testosteroneftestosterone cypim1atr.=: (100 mg;"mL in 87/14% cottonseed (311,
`
`0.1 mL bcnzyl benzoate, 9.45 mg benzyl a1cohx::1 as a preservative);
`
`Mackey (19%):
`
`Tr:smviz'oz2 Depotficatosmrntxc rmantimattz (250 mgfmfi in casmr ml and b<.°.'i‘I’.?'.}{1
`
`DB1;€:—2{)-i2(:«L3&.1
`
`7
`
`Astrazeneca Ex. 2135 p. 7
`
`

`
`b{i‘1'3Ié'.£}2i[ti:);
`
`as wail as a nmnbzizr of other cun1xnc:rciali:v:a:1 mi} based Emngactitzg {M ixxjactabie
`
`fmnnulatimas 1‘r;.*{:}<;sz‘:<:ci an Tabie 1 mi" {he Fivaraa Appiicatinxz.
`
`15. A3 a further p.ari::;f"t%1e prefarmulatiann {ah
`
`33, {hit cxperimcarci fsnrrmziator wmsld have
`
`cendmzteci a prefmnnuiatian sniuiniiity screen, separately measuring the zmmbiliiy <:-sf
`
`fulvestram in 23 range cf pure scfiwztxlm, "Em: uding the pmcmial 0&1 and cm-scelvent
`
`camiidatczfi that had been identified in the ahxwe literature review as being suitable far
`
`inclusion in intraxnuscular injtacztérm fmmu atiazms. See, for axarnpie, (Emma (1999),
`
`Chapter 1’? at page 462? umtier the hmding "‘F::mm1iation §j)av::i:::1:1m:3..*nt“t
`
`The activities necessary 1:: ciexre op :32: parenteral preduci can be
`plactzsd inm ihvai: fniinwitag threat: bread areas: prefnnnulazinn.
`fianixuiatinsn. and scale-up. Wizile: 1} era: are: altematfixm development
`perspectives, afi dtwcinpmaant um amly rmeds in aztmmpiish the
`zaaxne activities. Prefmrmulatimx inc. mi»:-:3 {he characmrizazics cf the
`
`bulk drug pins initial screening ‘far mzcipiasnt cmmpatibility with the
`drug.
`
`“Pmfor1nu1ati:"3n studies” are said tn *‘;:>r+:wid+;=: fnnaamentai dam and cxparicmzre necszgsary
`
`tr: téemizlmp fmrnulatinmz far as sgxxiritifm rsrrzmpimmrl" inciuxiing, as; itrsm 8.1 in the aniline :3?’
`
`areas nf specific iratratrest, 3 <:§et:ermi:1mii3:': ::>f“$:::iu¥::i}ity” in “schizctrsxi s::1v:ants” {at 483}.
`
`“Significant fizmnisfatimza activi:ie$ begin wiih initiai pfef{)F3‘I3£31£fl‘.ii3I1 ciata and knowiecige
`
`at" £312 specific; :'<:aui£: tnfadminisirafixzzxf’ {at 405}, which “f::m'nul23tiLm activities inciudfl the
`
`idemificatinn and sciactiran Q? 2: suitable vehicle {acguatmm mnaqumus or m~soIw:nt
`
`system} .3’ {at -’«1("}4). It is. furihfir fluted that “inji:i;tion Valium: is mm :31’ the; mom
`
`important zzmmicieratimm: in the fnrmniatimn dseveiogzment (fa {gamma-rciai pmduct” {at
`
`405), Wheat} zrzzrrying {mt such 2 prefarstzulatimx s:::Eubi,1it)g screw: with fuixrestrant, ihe
`
`fcsnnuiazor wouid have fmmd that fuivestram had tzxmamstiy ims.»-* soiubiliry in water, law
`
`soiubiiiiy in most mils (but highest in caaiur oii}, low 1-miubility in hmxzyl benzoate, and
`
`the highest sanmhiiity in ethaxml and br::t13};! aicnhm, such as !’:°:p(Zéi"tf;t::i in Table: 2 0f the
`
`Evans Appiicmimn.
`
`1”},
`
`With thzi: §nf::r:*mm§z.m cm prmr crizemmarazzéaiifizgz-ti farmniaticms and {he fuwesiram so1ubi§iI§;
`
`flats: fr:;m1 the pr:.:§':,:.:m1;.zIai:i:::-n screen {.~.i'.i.:£:l’i as rcperted in Table 3 safthc Evaszs
`
`[313 t §E12‘£3~‘2l?.f:«(3{: . §
`
`Astrazeneca Ex. 2135 p. 8
`
`

`
`Application), the experienced fermulater would liaise selected easier oil as the nail vehicle
`
`because of the liiglter seltibility of fulvestrant in caster oil relative to the other exile tested.
`
`lwlevertheleaa, he would have appreciated that the target fulvesttant eeneentratien of at
`
`least/-15 nigfml; could net be achieved with caster oil alone, and that a eeeelvent would
`
`be required.
`
`18.
`
`A number ef the commercialized formulations that would have been identified in the
`
`literature review (including the easier oil—l:>ased fennulatiens) have a substantial benayl
`
`benzteate component, which may be present as a eeeelvent. See, for example, Delalutin
`
`noted in paragraph 15 above‘ which is reported in FDR (1973) and noted in Table I of the
`
`Evans; Application, and is (me of the formulations discussed in Riffkin (196243, “Caster
`
`Oil aa a Vehicle fer Parenteral Administration ef ‘Steroid Hormones” (see Rifflcin ti. 6}.
`
`Delaltitin is 250 mg.«*’mL 17~hydre-itypregeaterene eapteate dieselved in 52% eastor oil,
`
`46% benzyl benaeate and 2% benzyl alcohol. However, Riffkin Table ii reports that the
`
`selubility of I7-hydexyéprogestetene eapmate in caster oil alone is only 55.6 mg!mL, but
`
`the seltttilitjy ef l?~hyclr0xypr0gestetene eapmate in benzyl benaeate is Substatttially
`
`higher, being at leaet 250 mgfml. {see example 4 of Huber (US ‘$26) and Attachment E
`
`diseuased below). Even if net needed as a eeselvent, Riffkin (1964) notes that “the
`
`addition ef benzyl alcohol or beitzyl benzeate to caster oil resulted in 3 lower and more
`
`favorable viseesity, making it easier to irijeet” (paragraph bridging pages 893-894}.
`
`1*}.
`
`Hewevet, the skilled fennulater would have appreciated from the fulveetrant solubility
`
`data generated in the prefemiulatiori aereeii that fulixestrant had very different solubility
`
`ehatacteristiea relative tea the steroids ef previous commercial fertttulatimis. Attachment E
`
`is a eempilatiott showing the ehemieal stmetures and relative seltibilities in caster oil and
`
`sesame ell ef the eetnpounds named in Riftltin (l964}Tab1e It eemgiared te the structure
`
`and the aelubiiity of fillVf3SlZI‘aI1l in these oils. It can be seen that the solubility ef
`
`fulvestrant in easier oil and in sesame tail (2!) mgx‘mL and $3.58 mg!mL, respectively, from
`
`Table 2 of the Evaiis Applieatimi} is appreciably lower than the solubility of the ether
`
`stereids in theee mils (taken from Table II of Riffkin (1964)). The second page of
`
`Attachment E talzvttlatez; the eeneentratietn in beazyl benzzeate of five named stereids, taken
`
`DE l;fi2€tsl2r€-Gt’:-.l
`
`9
`
`Astrazeneca Ex. 2135 p. 9
`
`

`
`fmm Examples L5 ttf Huber (US ‘S233, ranging from 200 in 40{} mgfmlf” By
`
`cemparisc:-n, the solubility of fixlvestrant in hertzyl benzeate is reported in Table 2 efthe
`
`Evans Applieatien as being ettly 6.15 mgfmh, and only 3.8 mgfml, 38 determined in the
`
`recently cenduetetl tests reported in Attachment C.
`
`20.
`
`The experiemzed thnnulator thus would have expected that benzyl item-toate would not act
`
`as a emscxlvent fer fulvestrzmt in caster all because the solubility of fulvestrant in benzyl
`
`benzoate was significatttly lewer than its selnbility in castor oil. The addition of benzyl
`
`benzoate to caster oil, fer whatever reaecm, would have been expected to decrease, razfizer
`
`than increase, the solubility at‘ fultrestrant in the resulting Castor oilfbenzyl benzoate
`
`mixture. 'l"h.is is mnfltmed in Table 4 of the Evans Application? which reports at
`
`ftzlvestrant solubility of only 12.6 mgfmL in the master all vehicle containing :2.-ttly 15%
`
`benzyl benzoate, compared to the 20 mg/ml, solubility of fulvesttartt in C3311)!‘ oil alone at
`
`reported in Table 2.3
`
`21.
`
`Based can the solubility clata determineci in the prefermulation eereen (suelt as reported in
`
`Table 2 of the Evans Application}, ethanol antilfor benzyl alcohol weuld have been seen as.
`
`the best <:e—s-oltzent candidates fer raising the fulvestrartt solubility to the 45 mgmlt, target
`
`in the caster oil vehicle, and wmtld also funzzticm to lower the viscosity of the resulting
`
`fennulatien and make it easier to inject. Consistent with this solubility tlata, Dukes (U53
`
`‘S143; added 4fl"Z’/<3 WW benzyl aicahei in order to tlisselve SO mgfml; ftzlvestrant in the
`
`caster oil-based fettmxlation used in the experimental rat; studies of his Example 3. It thus
`
`would have been apparent that 40% WW benzyl alcohol coultzl function as a t:ca-«solvent in
`
`caster oil tr: achieve the target fulvesttant concentration. Nevertheless, the skilled
`
`fermulater would have been cencemed with using such a high alcattel content in
`
`intramuscular ittjeetable fermulatiens fat administratien tn) 2: lmman.
`
`3 Data taken fmm the Examples Uf Huber {US ‘$20); these are csrmrsezttraticcns used in the examples and net
`necessarily the actual maximum selubility of t:3£:'l”t steroid in lncnzyl benzmate, which may be higher. Huber was a cm
`gtuilmr en Iiiftkin ( 119134}.
`" It should be noted that in the furtltet tests that were recently tremlucted under my gtxiclattee (patagtaplm '?-9 above
`tmtl Attaeltmetttts 13 and C2 txettztojlt the seltttvllity tat‘ fulvesttzmt in easttor eil alarm’: was again tested and fczsumfi te be
`214 mgjmL, and the selubility effulwstrant in benzyl benzoate alone was again tested and tlmmil tat: be Gill}? 3.8
`tngsml... wltich fitrtlter cam tinm that hettzyl benzeate would net be extnectecl to act :12: 2: cczstcrlxtent for fitlvestrant in
`master all,
`
`DE l i{32i}426f}t3.l
`
`ll}
`
`Astrazeneca Ex. 2135 p. 10
`
`

`
`2?” First :2? 3115, the experienced fizmnulamr wmuid warm in zmnisnism the amount craf ctr
`
`scaivents ané excipients in any injeczahie fmmulatitm. Fm cxampla, as stateé in Gupta
`
`{K999}, (lhzqzter 31?, “F»::nm1}ati<m and Acizniniszraiinn Teshniques in Minimize In_§e:::t§c:-n
`
`Pain and Tissue Damage Assuciateci with Pzxremai ‘.Pmdn::t:~;:” at page 41%:
`
`Cm;::xhrr.=::1ts am cnmmcnly useci {:3 ctziaance drug asamubiiiiy and
`stability. Casaivenis may it1<:ia..u:ir: ethanmi, gnrapylcnc giycnl,
`pfilwthyiczmc giycoig, and giymmine. These c<3rnp13w:::1ishaw:
`intrinsic effects on bioisngic tisasuaa and can alter th<~;: prmperties of
`m:}1:3r mtzczipiems, thus influessncing the tissxxea damage: or pain caused
`by 21 product. There is; 3 (Earth 52%‘ iiteratum {rm {hit pain L‘::€m$&fi by
`i.?.C35f.’JI’\:"tZ:1?‘!tS, but there is also 3 grmwing bmiy of knowiedge am the
`tiéasue daznage that {hey can cause. It is rm: certain that Iisimrs
`damage is always directiy carmiaiexi with the irzjcctian pain, but
`minimizaiion of bath again an iflfiisiiflfl and gacmzntiai far tissm:
`tianmgra xlzmuid be inciudemi in the prmxiuct ciemzslmpment pian.
`
`See aim ffiiupia {I999}, Chapter 3 1, titled Cmscatveazt Use in injeiztabhet Fnnnulatiens, page
`
`21?:
`
`Iciraaiijg, it is beat is Sfflii-‘SJ! and mat: :-::::s}v+::rm-2 that wssuid maximize
`this scalability of the cnmpnund. Mmrcimizing the 3oiubiILity mfg:
`cxampcmnci in 3 pariixmlar ccmsmivent zwstam wmzid ms-suit in lower
`mm} levels cf the non»-aquecus. 3x:>Iv:=:m{s} hating administered tar the
`patient, thereby lowering Ihe chance for pntmiial saide effects,
`
`Tixis objcciive woulfi have apphed to aqueous and ni1»ba3:3d syésterns alike, in that ‘ihe
`
`precedent of mmmerxzialized formutatiom-:. identified in the literature review wcmki haw:
`
`mnfirrnrsd that fixeci miis, such as casts}: mi}, have hang been cmnumcrcialty used and
`
`E1£‘.:i'34:;'pIt3i‘{3 as the nzajcar cczmgmzzent mi? {:sii—bast:c§ susiaimci mirazase intramuscniar injectahle
`
`stcraieial fhnnuiazions. On the rather hané} cxaqnalvcarxts such as eiharmi <31‘ banzyl aicohol
`
`have gencraliy been used {mi}; in far teaser cczncentraiiiz-115, as discusseci in the fcsilmxving
`
`paragraph.
`
`23, Thus, 133$ tgfsuch a high cement cf either buzrmyl alcizhmi :33‘ s::tham3§ wmzici have Eaten
`
`sanitary 1:: prszcedexit as shmwn fmm the mvizzw of commeruiaiimré c;i§-basmi
`
`imramuscuiar ingectabie Sustained release fmnnulatitnras. The Iiteraturaz review as of early
`
`2{1{}€,1 W1:-uh?! llama 3h<3wn that any bengyi aizsuhm in smtln fezammlatémm was aimogt always
`
`:3:41:a¢2m2n1:.:,.z
`
`1 1
`
`Astrazeneca Ex. 2135 p. 11
`
`

`
`presem as a preservative in a concentration of about 2% er less, O€:{:Z:‘lSi{)I12lll}’ at 3
`
`concmltraticsn of up to 5%, but only rarely at higher canczexitratiinnsi. With respect to
`
`lrienzyl alcehol see, fer example:
`«I
`
`Gupta (1999), Chapter ll at page 229 stating that benzyl alcohol “is typically used in
`
`concentralicms of up to 2 percent as a preservative and up is 5 present as a solvent“
`
`and then discussing reported toxicities.
`
`Nema (1997), Table. ‘V’ at page 168., reporting max benzyl alcohol was present as an
`
`amimicrcibial prszscrvativc in 7'4 injectable foirinulations (rim: limited to Gil-l3EiS(":til IM
`
`formulations) ax miiceniratians inf from 0.735% {note that lmnzyl alCOl‘l€)l
`
`is net
`
`included at all in Name: Table l, “S»::alx:<:i1ts; and (3o»3:2lVei1ts"’;
`
`Powell (1998), the bemtyl alcohol lisiing at pages 244446, particularly those
`
`indicated as being used in mi fertnulatinns;
`
`Strickley I (1999) at page 3283 notes the imzlusimi of 2% benzyl alccahtil in an IM
`
`lmzazapam fnmmlatiizm in a pmpylr.-tm: glycol vehicle, but (11163 not include bmizyl
`
`almzshol at all in Table ‘VI listing ‘“C0s<::l“vents Used in Parenteral P’<::rmulaiions;"
`
`Lopatin (1972) rioting in Table 3 at page 727 opposite: Benzyl alcohol, “T<:i:»s;ic. Used
`
`in concemratinn inf not ever 3%. Has irritant aciion in mncentration of S%;’""
`
`Ctzarnelius {US ‘$63), col. 1, lines 30-35 stating, “It is known that the scilnbility of
`
`steroids in vegetable or animal oils can be increased by the addition of zitxcipients such
`
`as ltzenzyl alcohol and banzyl benzoate. An‘ -cibjvatztion tn the use of such excipirmts, and
`
`spacifically’ bemzyl almlml in somewhat higher concentrations, is that these agents
`
`may irritate the ti-;~:sues.”
`
`The literature review as of early 2000 alga wmild haw shown that, with few excepiiona,
`
`ethanml was not included in such fcrmnulations in excess of about l()%. See, for example:
`
`Gupta {I999}, Chapter I I at page 225 nesting that ethanol has been used at levels up
`
`to 50‘p%rc;e:1t, but these lcvttls typically are assosiated with pain on injection;
`
`I
`
`Strickley I (1999), Table V}, “List sf Cosnlvents Uaed in Parenteral Formulations”
`
`more specifically lists the ethanol content in IM fizsrmulalions fur spacifically
`
`identified drugs, which concentrations range only from 2.5 to 10%; an {MFKV
`
`lmazapam, formulatimi in a propylenrz glycol vrzhicla is noted at page 329 as having
`
`lS% alccolml, but is 11m; included with the 1M lhriniilatitins in Table ‘W;
`
`[38 lé62fl4.’2tf:fZl6 .l
`
`12
`
`Astrazeneca Ex. 2135 p. 12
`
`

`
`an mama (39‘§'?)‘ Tabie I, “Scs§w:nts anti £i2<:»~:sn:rim~“c:m:s“ at nag: 3 $71 iists etlzancsi as being
`
`in 24 fnnnulatimzg with a cnncentratinn range of{},{3-~é§i}‘E‘ri3 (ff)? ?mgra:t}; ncm‘: {hat this;
`
`is misieadiazg, hawawer, since Pmgmf is :1 €?£>H::?a?:1£:*(I£-L? fer izztrawezmus infusimn mfiy,
`
`aim is in be dilated 258 {cs §0{}{} {imam b»*2:f::srr: administraiimx;
`
`an
`
`Fmwxzli {E993}, lists: “‘“2.1::s::»h<::I”" at page 242 3:15} "‘t:i:I1yI aicohczci” at page 255, wherein
`
`the etiman-at mxxcesitration for IM fiormulatiom ra:1ge:;: £’r=x::2m f}.Ei1«1{}%.
`
`24.
`
`Thus, even thtmgh Qukes (US ‘$14) had :}mnm1str2m:d that the target 45 mg,z’:mL
`
`Fulvitstrant concemratian ccmici be achieved by ackiing 49% benzyk alcohxzl to the caster
`
`c;miIv£::hi.:.:1z:, {he prececéent of commerci:-:x1i2ev;i ¥M c:»i1«ba:‘~;¢:d systems wmfid have mmcivamd
`
`ttm expericnmd famzulatc:-r in suhséamiaiiy reduce ‘rim bensayi aI::s::Imi Cflfltfilit aafthc
`
`fiarrnutmimn imexxnieri f€}I’h11i‘1‘i3I1 use, and {hi3 <:z3mmr.=:r::ia§ preuezient wsrmié haw: maée
`
`him wry rciuizzrant to replace benzyi aiczishul with {ha smhaztainiial armtaum sf mhalml that
`
`wmxid be needed to maintain the tzargxzi fniwsstrani. cranmatraiiezz. Benzyi benxoatc cleariy
`
`wouid am he Eiflilfiidflffid I:-:1 £{2d’S.*‘i3 this diicrnnam but rather would be expected m have 3
`
`negatiwz ::f"f7ect an fulvestrant sambility Sifiiiii fuivastrani wag even Ease: sniuble in benzyi
`
`benanate than in caster ail, thai is, me: wmuid ham expected that aéding benzyl bxztazcaate
`
`wmxkzi requirc sztiil mare alcuhui in maintain 2113 target fuhzegtraxit +::(:«ncentra2:im11.‘*
`
`25.
`
`Undar rhea:-2:’: circumstancas, the digcavexjr by Evan:-1. an‘ (23,: that the addition of benzyl
`
`btmzmam to the castmr czilfalcohcl mixture actuaiiy i:1crc:as»35: thrs: golubility cf fulvesitranr
`
`such that Imam fuivestrant cauid be dissaived in 3 given valume of forxnulation, was
`
`uxlcxpccted and tnity surprising. This pcxsitive benzrtyl benzitoatc effczcz cm fuivestrant
`
`zmlubiiity in the: resuiting f<:armu§atE(m is Shawn in "habits: 3 :31’ the spacificatian {ané is net
`
`charzgeci by the abmxemmed =‘3{)1‘l‘!K3£?§if}¥1S}3 and ii; cmxfirrned anti siemmsstrated over 3
`
`brmader range cf fermuiatian cnmpositiazn by the adciiticmal set {}ff5Xp£:3‘i1‘flE:IltS conducted
`
`under my guidance anti discussed in paragraphs ‘L9 abews, ihe E‘{f:.'§l1IICS mfwhich are
`
`repurmd in Attachments (3.
`
`“ It zailmuid 13$ am:::<§ t%1ats:w::::1 agmn Fm-m this scziubiiésy £55133, imzrrgs *s.mx1Ici haw: i3c:v:'.-1': ma maztimtirm tcs add hsanzyi
`‘magnate fiat vgiscmsity rxzéumtiun sigma the signiiicant quaxmtity £3? 2:1::e::1mi wcmlxi mrrvxr the dual fuxzciian ofacting as :3
`:::1~::<;xkw:nt as wait as rccincéng tin: §:33'<:<:ii<m vi$r::meity zmci amzking it azmicr me injcct. whereas fin: b-::*11zj~;lbca:mai:{:
`wvzmld km fltxptiitkrd in hszwxx :2: rmgzaiire ::ffZ:»:.: rim 211:: iislwscsirmzt szxlubiiity,