`
`
`
`
`
`
`PHIGENIX
`PHIGENIX
`Exhibit 1015
`Exhibit 1 0 1 5
`
`

`

`From (613) _991—2760_
`
`0rd_er'I# 05421267DP04474253
`
`Mon Apr 18 23:24:30 2005
`
`Page 2 of 17
`
`1.. Lafleur. Y. Fmdct.
`vating proteins-con-
`:lder cztrcinorrta 601]
`l.
`aslan. Pseudnmouas
`iody against ovarian
`arian cancer cells in
`:i. USA 84 (I987)
`
`thlingham. RJl.
`Chemoimmunotoxin
`HT-29) xenografted
`t 3562—3567.
`:, R.C. Young. LW.
`:Clay. S. Howell. E.
`l. I. Pastan. Clinical
`aortas cxotoxitt
`im-
`vithiovan'an cancer,
`
`I.
`in", S. Ktntwnr,
`that
`is active on
`:ed of the Fv region
`truncated form of
`land. Sci. USA 90
`
`1, K. Covalcuic, l).
`I vitro and in viva
`tmunotoxin contain—
`er Res. 55 “995)
`
`hire, R. Aoki. M.
`of gclonin immuno—
`N—l
`tn cndomefrinl
`unto: cells in yitro.
`
`argcting — a new
`, Pharmuc. ’icr. 64
`
`If large solid tumors
`:ted against
`tumor
`\ 90 (1993) 8996—
`
`L. Watkins. T.S.
`'ction in mice by
`tctor to tumor vas-
`
`Advanced Drug Delivery ReVicws 31 (1998) 89-104
`
`
`
`advanced
`drug delivery
`reviews
`
`
`
` Targeted delivery of chemotherapeutics: tumor-activated prodrug therapy
`
`Ravi VJ. Chari’k
`
`ImmmmGrm, nun, I48 Sidney Street, Cambridge, MA 02139. USA
`
`
`
`Abstract
`
`‘
`
`The potential of targeted delivery of chemotherapeutic drugs for the treatment of cancer has not yet been realized owing to
`the difficulty of delivering therapeutic concentrations to the target site. While in vivo studies in animal tumor models have
`produced very encouraging results, clinical studies with antibody—drug conjugates have been less successful. This paper will
`review the current status of the targeted delivery approach and analyze some of the reasons for the lack of success so far.
`Starting with a historical perspective,
`this review will end .with a description of newer, more potent and specific
`antibody—drug conjugates, which behave like tumor-activated prodrugs that may yet fulfil the promise of the targeted
`delivery approach for the treatment of cancer. © 1998 Elsevier Science B.V.
`
`Keywords: Antibody—drug conjugates; Cancer; Prodrugs; Anti-cancer agents
`
`Contents
`
`
`
`1. Introduction ............................................................................................................................................................................
`2. ’l‘umor-activated prodrug (TAP) therapy ..........................................................
`
`3. Tumor-specific agents .....................................................................................
`4. Early antibody—drug conjugates ...............................................................................................................................................
`4TI. Design of conjugates ...................................................................................
`In vin‘o evaluation.
`4.2.
`
`4.3.
`in vivo cfficacy..
`
`4.4. Clinical evaluation
`
`5. Factors influencing the effectiveness of antibody—drug conjugates
`..........
`
`6. Antibody conjugates willt more potent drugs ...........................
`
`6.1. Design of conjugates ................................................
`
`.
`..
`6.2. in vitm evaluation.................................................
`6.3. Anti-tumor efficacy in vivu ...............................................................................................................................................
`7. Conclusion ...... . ......................................................................................................................................................................
`Acknowledgements ......................................................................................................................................................................
`References ..................................................................................................................................................................................
`
`89
`90
`91
`92
`93
`94
`95
`96
`96
`97
`97
`99
`loo
`[01
`l0'2
`l 03
`
`1. Intruduction
`
`Cancer chemotherapy today relies on the expecta—
`
`‘Tet; V+l rm 4971113; Fax: +1 617 4975406; email:
`raViuchari@immunogen.cemaiLeompuservecom
`
`
`
`
`
`
`
`
`0l69—409X/98/51900 © 1998 Elsevier Science B.V. All rights reserved.
`PII 30169-409X(97)0l)095-l
`
`tion that anti-cancer drugs will preferentially kill
`rapidly proliferating tumor cells rather than normal
`cells. Typically, cancer patients with disseminated
`disease present themselves with approximately 10”
`tumor cells, and it
`is well establislted'that at least
`99% of these cells have to be killed (Le. a two-log or
`
`--
`'
`PAGE 2/17 " RCVD AT 4/18I2005 8:28:24 PM [Paclt‘lc Dayllght Tlme] " SVR:SVCSO1I0 ‘ DNI
`
`or
`6034 " CEI
`
`"3)991-2150
`
`" DURATION mm-ss 218-10
`
`PHIGENIX
`
`Exhibit 1015-01
`
`

`

`Order . # 05421267DP04474253
`
`Mon Apr 18 23:24:30 2005
`
`Page 3 of 17
`
`R.V.J. Clittri / Advanced Drug Delivery Reviewr 3] (I993) 89— 704
`
` From (613) 91—2760,
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`radioisotope or a small cytotoxic drug. This chapter
`will review the progress made in the area of targeted
`delivery of small chemotherapeutic drugs, discuss
`some of the shortcomings of the earlier approaches,
`and provide potential. solutions that may help restore
`the promise of the targeted delivery approach for
`cancer therapy.
`
`2. Tumor-activated prodrug (TAP) therapy
`
`The basic premise of the targeted delivery ap-
`proach is that conjugation of drug to a tumor-specific
`molecule renders the drug inactive until
`it reaches
`the target site. Once at the tumor site. the conjugated
`drug binds to the surface of tumor cells and is further
`processed (internalized, released from the carrier
`
`molecule) to restore its original potency. Thus, drug
`conjugates can he considered as
`tumor-activated
`
`prodrugs (TAPS). While most conventional prodtugs
`are converted to active drugs by mechanisms such as
`chemical or enzymatic hydrolysis.
`restoring the
`activity of TAPS should ideally be dependent on
`interaction with antigens or receptors specifically
`found on the surface of tumor cells. Historically,
`conventional prodrugs have been designed to over-
`come a physiological barrier, such as poor oral
`bioavailability or rapid metabolism 13,4]. Often, the
`oral delivery of a drug is
`improved by merely
`converting it into a water soluble prodrug. Conven-
`tional pnodrugs are designed with the expectation
`that improving the pharmacokinetic properties of a
`parent drug will result in increased levels in circula-
`tion and thus greater levels at the target site. In TAP
`therapy. the specific affinity of the tumor-associated
`antigen or receptor for the targeting component of
`the drug conjugate results, in addition, in a greater
`uptake and retention of the TAP at
`the targeted
`tumor site and,
`therefore,
`in increased selectivity.
`This is followed by liberation of the active drug
`resulting in high local concentration at the target site.
`Ideally, TAPS would be stable during circulation-
`such that no conversion of the prodrug to the active
`form occurs outside the targeted tumor. Also, TAPs
`would not bind to non-target tissues and thus will be
`non—toxic while in circulation in vivo.
`
`The principle of drug conjugates as tumor-acti—
`vated prodrugs
`is
`illustrated with antibody—drug
`conjugates as an example in Fig. 2. In an ideal
`
`greater cell kill) to achieve a complete remission.
`Continued treatment during remission is required to
`achieve complete
`eradication of
`the
`tumor. A
`schematic
`representation of
`the drug treatment
`categories for cancer as originally described by Frei
`[.11 is shown in Fig. 1. Unfortunately, clinically used
`anti—cancer drugs have limited selectivity for the
`tumor. The levels of drug required to kill sufficient
`number of tutuor cells to achieve and maintain a
`
`state of complete remission in patients causes signifi-
`cant
`toxicity towards actively proliferating non-
`' malignant cells, such as normal cells of the gastroin-
`testinal tract and bone marrow. Thus, a continuing
`challenge in cancer treatment
`is
`to develop new
`cytotoxic agents with greater selectivity for
`the
`tumor. To achieve this goal,
`it is first necessary to
`identify inherent differences between normal and
`
`cancer cells that can be potentially exploited.
`The discovery that. tumor cells expressed specific
`dcwrminants on their cell surface that were not found
`on normal cells suggested that this distinction could
`form the basis for the selective targetingof tumors.
`The advent of monoclonal antibody technology [2]
`led to the development of a myriad of monoclonal
`antibodies, each with its own binding specificity for
`novel tumor-specific antigens. The logical outcome
`was to exploit the binding specificity of the antibody
`to deliver a cytotoxic agent selectively to the tumor
`site with the hope of delivering a ' high,
`lethal
`concentration of drug to the target cells. The cytotox-
`ic agent could be in the form of a protein toxin, a
`
`Complete remission
`
`
`
`
`
`
`
`
`
`“ Treatment
`lnductlontreatment \durlng remissim
`
`8
`
`Eradicatlon of
`neoplastic cells
`
`T?
`Immune Response
`
`Time ( from start of treatment)
`
`Fig. I. Chemotherapy of cancer (reproduced with permission from
`Frei [1]).
`"
`
`
`
`
`
`
`
`(Log,0)
`CancerCells
`
`
`PAGE 3/17 ' RCVD AT 4/18/2005 8:28:24 PM [Paclflc Dayllght Tlme] " SVR:SVCSOl/0 ‘ D
`
`
`
`PHIGENIX
`
`Exhibit 1015-02
`
`

`

`From (513) 991—2760 ,
`
`Order .# 05421267DP04474253
`
`Mon Apr 18 23:24:30 2005
`
`Page 4 of 17
`
`-
`
`.
`
`It.V..l. Chari / Advanced Drug Delivery Reviews 31 (I998) 89- 104
`
`91
`
`1g. This chapter
`area of targeted
`dings, discuss
`lier approaches,
`nay help restore
`'y approach for
`
`') therapy
`
`3d delivery ap—
`a tumor-specific
`until it reaches
`
`., the conjugated
`lls and is further
`70111
`the. carrier
`
`ncy. Thus, drug
`tumor-activated
`
`ational prodrugs
`hanisms such as
`
`restoring the
`,
`2 dependent on
`tors specifically
`lls. Historically,
`:signcd to over-
`h as poor oral
`[3,4]. Often, the
`ved by merely
`fodrug. Conven-
`the expectation
`properties of a
`evels in circula-
`
`‘gct‘site. In TAP
`armor-associated
`g component of
`on, _in a greater
`at
`the targeted
`ased selectivity.
`the active drug
`at the target site.
`tring circulation
`rug to the active
`nor. "Also, TAPS
`and thus will be
`'0.
`
`s as tumor-acti-
`
`antibody—drug
`2.
`In an ideal
`
`1000
`
`900
`
`800
`
`>.
`2 700
`E 600
`2
`m 500
`E
`< 400
`ii
`m 300
`
`200
`100
`
`°__
`
`
`
`Drug
`
`Conjugate Drug
`
`__
`___
`Conjugate
`
`TARGET CELL
`
`NON-TARGET CELL
`
`Fig. 2. Relative potency of free drug and antibody—drug conju-
`gates (tumor—activated prodrugs. TAPS) towards target and non«
`target cells.
`
`situation, for an antigen-negative cell which does not
`bind conjugate, conversion of the free drug into :1
`TAP by conjugation to an antibody results in in-
`activation of the drug. For an antigen-positive cell,
`binding to the TAP is foll0wed by internalization and
`release of the free drug in its fully active form. Thus
`the free drug and TAP have equal potency for the
`target cell. The therapeutic window is determined by
`the difference in cytotoxicity of the TAP for the
`target versus the non~target cell. As we will see later,
`the effectiveness of TAP therapy depends on several
`factors-including choice of the targeting molecule,
`the potency of the drug and the nature of the release
`mechanism for conversion of the prodrug into the
`active drug.
`
`3. Tumor-specific agents
`
`The success of the targeted drug delivery approach
`for the treatment of cancer relies to a great extent on
`the tumor-specificity of the targeting agent. The cell
`
`surface molecrrle which the targeting agent recog-
`nizes can be a tumor-specific. antigen (typically a
`glycoprotein, carbohydrate or oncoprotein), a growth
`factor receptor, or a receptor for a hormone. Ideally,
`the cell surface molecule would have the following
`properties:
`(a) well defined molecule expressed ex-
`clusively on tumor
`tissue,
`(b) not expressed on
`normal
`tissues (c) binds to the targeting molecule
`with high affinity,
`((1) expressed homogeneously on
`all target tumor cells, (e) present on the tumors of all
`patients with the same type of cancer (f) not shed
`into the serum of patients.
`In the early phase of the targeted therapy ap-
`preach, monoclonal antibodies were heralded as
`ideal
`targeting agents that hound exclusively to
`antigens expressed on tumor cells. However, the use
`of more sensitive analytical
`techniques
`such as
`immunofluoresccnce
`and
`imtnunohistochcmical
`
`staining revealed that most antibodies bound to
`tumor-associated antigens that were only preferen-
`tially expressed on the surface of tumor cells.
`In
`most cases,
`the antibodies also bound to varying
`extent.
`to antigens found on a limited number of
`normal tissues. In fact, the target antigens for most
`antibodies [5,6! developed against solid tumors were
`selected mainly on the basis of the higher expression
`of the antigen on tumors in comparison with normal
`tissues. The only truly tumor-specific antigens appear
`to be those found in hematopoietic tumors, such as
`idiotypes present on the surface of B-cell tumors {7],
`and the T-cell receptor expressed in T—cell leukemia
`and lymphOma. Although the cross-reactivity of
`antibodies with normal tissues is a matter of concern,
`
`the benefit potentially gained from the improvement
`in the therapeutic window of cytotoxic drugs by
`conjugation to antibodies often outweigh the toxicity
`concerns. Of course, selection of antibodies with an
`
`In
`acceptable cross-reactivity profile is important.
`addition,
`thorough pre—clinical toxicology studies in
`animals that bear the same antigenic determinants
`and show similar crossvreactivity patterns to that
`found with human tissues is critical.
`Monoclonal antibodies are also attractive as target—
`
`ing agents because of their high binding affinity for
`their respective antigens. This should allow for the
`localization and retention of high concentrations of
`drug at the tumor site. In addition, the long circula-
`tion time of antibodies also allows for a greater
`probability that the. drug will reach the tumor site.
`
` PAGE 4/17 " RCVD AT 4/18/2005 8:28:24 PM [Paclflc Dayllght Tlme] " SVR:SVCSO1
`
`PHIGENIX
`
`Exhibit 1015-03
`
`

`

`From (613) 991—2760.»
`:-
`
`Order.# 05421267DP04474253
`
`Mon Apr 18 23:24:30 2005
`
`Page 5 of 17
`
`
`
`
` 92 .
`
`R.V..l. Cliari / Advanced Drug Delivery Reviews .31 {1998’} 89—104
`
` poorly vascularized. The use of smaller antibody
`
`The therapeutic potential of conjugates of cytotox-
`ic drugs with monoclonal antibodies derived front
`murine hybridomas is dampened by the development
`of a predictable anti—globulin immune response in
`humans. The generation of a human anti-mouse
`antibody (HAMA) response leads to rapid neutraliza-
`tion and clearance of the itnmunoconjugate from the
`blood stream,
`thus limiting its therapeutic utility.
`Recent advances in recombinant DNA technology,
`and knowledge of antibody gene structure have been
`applied to the engineering of rodent antibodies to
`make them less immunogenic. A- ‘humanized’ anti-
`body is constructed by transferring the murine
`complementarity determining regions (CDRs) on to
`an appropriate human framework region. Since
`CDRs form the antigen combining site, a humanized
`or CDR-grafted antibody preserves the murine an-
`tigen specificity, but because most of the antibody
`structure is human,
`it is likely to be less immuno—
`genic in patients than the parent mouse antibody.
`Recent clinical studies [8] with humanized antibodies
`
`fragments instead of whole lgG molecules may be
`advantageous in sonte cases. Three comparative
`studies [ll—13] of the tumor localization of radio-
`labeled intact IgGs and smaller fragments [Fv, Fab'.
`F(ab')2] in mice have shown that smaller fragments
`penetrate the tumor faster (maximum tumor penetra-
`tion of Fv is at 0.5 h) than intact IgG. which showed
`an equivalent degree of tumor penetration only at 48
`h post-injection. However,
`the smaller fragments
`displayed faster clearance and delivered lower over—
`all tumor doses than the intact IgG, suggesting that
`conjugates with intact lgG molecules may be prefer-
`able for the specific application.
`Although monoclonal antibodies probably provide
`the greatest binding selectivity for cancer cells, other
`targeting agents that preferentially bind to tumor cell
`surface markers may provide distinct advantages
`such as smaller size. rapid internalization and non-
`intnnmogcnicity. For example, the epidermal growth
`factor receptor (EGFR) gene is amplified in a high
`proportion of human squamous carcinoma cell lines _
`[l4]. Levels of EGFR are three to sixty—four—fold
`higher in several tumor types, such as lung. breast
`and head and neck, as compared to that found on
`normal keratinocytes [15]. Human EGF is a single
`polypeptide of 53 amino acids and is
`specially
`attractive as a targeting agent because of its small
`size (M,=6201), high binding affinity for its re-
`ceptor (apparent K‘1 = 2*4 X ill—:0 M) and its rapid
`internaligation upon binding to the receptor [16]. In
`addition'human EGF will not be immunogenic. An
`example of another polypeptide that can also be lused
`to target EGFR-exprcssing carcinomas is transform-
`ing growth factor alpha (TGl"'a) []7|. Other exam-
`ples of low molecular weight targeting agents that
`appear
`to bind preferentially to tumors
`include
`melanocyte
`stimulating hormone (MSH) against
`melanomas [18], thyroid stimulating hormone (TSH )
`and thyrotropin against
`thyroid cancers [19], and
`interleukin 2 (1L2) for T—cell leukemias [’20].
`
`:
`
`'
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`in 46 patients have demonstratedthat unlike human-
`
`mouse chimeric antibodies, COR-grafted antibodies
`
`were found not
`to induce a primary immune re-
`
`sponse, even after several courses of treatment.
`
`However, humanization by CDR-grafting often re-
`
`sults in an antibody with a lower binding affinity
`
`than the parent murine antibody. Further amino acid
`
`substitutions in the framework region are usually
`
`required to maintain the correct conformation of the
`
`CDRs. Even with this improvement, COR-grafted
`
`antibodies with somewhat
`lower affinity than the
`
`parent. antibodies are often produced. A newer
`
`technique called variable domain resurfacing 19]
`
`takes advantage of the generally accepted view that
`
`the antigenicity of proteins is determined solely by
`
`surface epitopes. In this approach, the binding affini-
`
`ty is maintained by retaining the CDRs and the core
`
`of the murine variable region framework. Only the
`
`surface residues in the tnurine variable region are
`
`replacedby' those front a human variable region.
`
`This technique was applied to two murine antibodies
`
`and,
`in both cases, affinity was fully preserved
`
`[10].
`
`Although monoclonal antibodies have been the
`
`most commonly used targeting agent
`for chemo—
`
`therapeutics,
`the pharmacodynamics of these large
`Early antibody—ding conjugates were comprised
`
`immunoglobulin molecules may impede their ability
`of a monoclonal antibody covalently linked to sever-
`
`al molecules'of a clinically used anti-cancer drug.
`to access or penetrate solid tumors which are often
`
`
`4. Early antibody—drug conjugates
`
`
`
` PAGE 5/17 . RCVD AT 4I18l2005 8:28:24 PM [Paclt'lc Daylight Tlme] ‘ svnfiycgggq;3N1
`
`PHIGENIX
`
`Exhibit 1015-04
`
`

`

`From (613) 991—2760 .
`we
`
`0 rder‘ ,# 05421267DP044742 5 3
`
`Mon Apr 18 23:24:30 2005
`
`Page 6 of 17
`
`RM]. C/mri / Advanced Drug Delivery Reviews 3] ( I 998'} 6‘9— /04
`
`93
`
` l l
`
`smaller antibody
`molecules may be
`nee comparative
`lization of radio—
`
`gments [Fv. Fab’,
`unaller fragments
`m tumor penetra-
`G, which showed
`
`nation only at 48
`mallet
`fragments
`Iered lower over-
`
`}, suggesting that
`:5 may be prefer-
`
`probably provide
`aancer cells, other
`tirtd to tutnor cell
`
`:tinct advantages
`lizalion and non
`
`epidermal growth
`tplified in a high
`cinoma cell lines
`
`:o sixty-four—fold
`h as lung, breast
`to that found on
`
`EGF is a single
`and is
`specially
`lust: of its small
`
`finity 'for its re-
`M) and its rapid
`receptor [16}. in
`.nmunogenic. An
`can also be used
`nas is transform-
`
`.7}. Other exam-
`:ting agents that
`tumors
`include
`
`against
`(MSH)
`' hormone (’I‘SH)
`tncers
`[l9], and
`mias [20].
`
`is
`
`were comprised
`' linked to sever-
`
`uni—cancer drug.
`
`The linker connecting the antibody and the drug was
`either non-cleavable or cleavable upon entry into the
`cell. Representative examples of types of linkers and
`drugs that were used for the generation of drug
`conjugates are shown in Fig. 3.
`
`4.1. Design of conjugates
`
`in the early development phase of antibody-drug
`conjugates. it was believed that the tumor specificity
`of anti-cancer drugs could be improved merely by
`linking these drugs directly to antibodies via amide
`bonds. Thus, several anti-cancer drugs [21—26] such
`as methotrexate. daunorubicin,
`the Virrca alkaloids,
`mitomycin C,
`idarubicin’ and N-acetyl melphalan
`were linked via non-cleavable bonds to a multitude
`
`of murine monoclonal antibodies [27—29]. Typically,
`an average of 4—8 molecules of drug werelinked per
`antibody molecule.
`in most cases.
`the conjugates
`lacked cytotoxic potency and were less potent than
`the unconjugated drugs. Attempts were then made to
`increase their potency by linking a larger number of
`drug, molecules either directly to the antibody or
`through the intermediaey of a macromolecular car-
`rier,
`such as poly-L-lysine, poly-I,«glutainic acid,
`dextran or human serum albumin [30,31]. However,
`these conjugates displayed impaired binding affinity
`in vitro, and unfavorable pharmacokinetics in ani-.
`mals deterring further development.
`Since the full potency of the drug could not be
`observed when such non—cleavable
`linkers were
`
`used. the emphasis shifted towards the development
`of linkers that wbuld be cleaved inside the cell to
`release active drug. An approach, that was explored
`by Trouet et al. and Umemoto et at. [32.33]. consists
`of linking daunorubicin or methotrexate to an anti-
`body via a tetrapeptide spacer arm under the premise
`that free‘drug can be released from the conjugate by
`the action of lysosomal peptidases. One of the first
`cleavable linkers to be widely employed was an
`acid-labile linker based on cis-aconitic acid,
`that
`takes advantage of the acidic environment of the
`endosomes
`that might be encountered by these
`conjugates after receptor mediated endocytosis. Shen
`and Ryse'r' used acid labile linkers for the preparation
`of conjugates of daunorubicin with macromolecular
`carriers [34]. Subsequently, Yang and Reisfeld [351
`and Dillman et ali [36] used a similar approach to
`
`LlNKEFl
`
`Non-Cleavable
`
`Amide
`O
`a) RtNJk
`H
`
`R'
`
`Succinimide
`
`0
`
`H
`
`a) R‘OJWN‘R'
`
`0
`
`Cleavable
`
`DRUGS
`
`Methotrexate [21]
`Daunorubicin (22]
`Vinca Alkaloids [23]
`Mitomycin C {24]
`Idarubicin [25]
`N-Ac-Melphalan [26] .
`
`Peptldase—Labile
`Fl-Ala-Leu-Ata—Ala-R'
`
`Other tri- and tetra- peptides
`
`Doxorubicin {32]
`Methotrexate [33]
`
`Acld Lablle
`
`a) Cis-Aconitic
`
`o
`
`OH
`
`b) Hydrazide
`
`RI
`
`/
`Fl/
`
`N‘N’
`*5
`
`Re
`
`c) Thiocarbamoyl
`S
`NJLNH o
`HoH
`R.N,
`0
`H
`
`Doxorubicin [34)
`Daunorubicin {35,36}
`
`Vlnblastine [37]
`Doxorubicin {38]
`
`Daunorubicin [39]
`
`Fig. 3. Linkers and drugs used in early antibody—drug conjugates.
`
`conjugate daunorubicin to an anti-melanoma anti—
`body and an anti-T cell antibody, respectively. A
`more popular method of conjugating drugs to anti—
`bodies uses acid—labile hydrazone bonds. The hy—
`drazone linkage has been used to conjugate vin-
`blastine and doxoruhiein to antibodies {37,38l. An
`thiocarbamoyl
`acid-labile
`linker
`for
`coupling
`
`
`
`PHIGENIX
`
`Exhibit 1015-05
`
`

`

`From (613) 99172760 ..
`
`0 rder‘ .# 05421267DP044742 5 3
`
`Mon Apr 18 23:24:30 2005
`
`Page 7 of 17
`
`
`
`
`
`
`
`
`
`
`
`
`
`R.V.J. Chari I Advanced Drug Delivery Reviews 31 (I998) 89-104
`
`daunorubicin to a polymeric carrier has also been
`described [39]. Structures of representative examples
`of acid—cleavable conjugates are provided in Fig. 4.
`
`4.2,
`
`In vitro evaluation
`
`Evaluation of antibody—drug conjugates for bind-
`ing to antigen-positive cells indicated that.
`in most
`cases,
`the affinity of the antibody was either fully
`preserved or only slightly diminished upon conjuga-
`tion of up to eight molecules of drug. In many cases.
`the results of in vitro cytotoxicity evaluation of the
`conjugates have not been reported. In the remaining
`cases, evaluation of in vitro cytotoxicity was general-
`ly performed using 3H-thymidine incorporation as a
`measure of cell growth inhibition. Antibody conju—
`
`gates of methotrexate, the Vina: alkaloids or doxont-
`bicin had [C50 values in the range of 10"b M to 10-7
`M [40—42]. The efficiency and extent of cell kill in
`vitro is difficult
`to gauge because the cytotoxicity
`assays are limited in their capability of detecting low
`survival numbers (lower limit of detection is about
`5%). Thus, these assays do not provide a measure of
`the steepness of the killing curve or the number of
`logs of cells killed. Often, conjugation appeared to
`diminish the cytotoxicity of the drug as compared to
`the parent unconjugated drug. For example, conju-
`gates of doxorubicin with the antibodies BR96 and
`BR64. directed against. the Lewis Y antigen found on
`several
`tumors, were eight-fold and live—fold less
`active respectively in vitro than free doxombicin
`[42.43]. Similarly. free doxorubicin is fivc-fold more
`
`0
`
`o
`
`iW
`
`l1
`
`"
`
`"l/N
`
`
`
`0 u
`
`se
`
`0
`
`H36
`
`0
`
`c
`
`
`
`OH H"
`
`°
`on
`
`|
`
`0
`
`o
`
`2
`
`Ah
`
`on
`
`on HN
`
`H
`N
`"\n’ V
`5
`
`n
`
`0
`
`R = Polymeric backbone
`
`a
`
`Fig. 4. Stnictures of representative acid-labile ding conjugates: hydrazone-linked doxorubicin conjugate, 1
`daunorubicin conjugate. 2 [35.36], daunombicin conjugate containing a thiocarbamoyl link. 3 [39].
`
`
`
`[38,42], cis-aconitate linked
`
`mess
`034 ' C
`
`PAGE 7/17 ‘ RCVD AT 4/18/2005 8:28:24 PM [Paclflc Daylight Tlme] SVR.SVCSO1I
`
`PHIGENIX
`
`Exhibit 1015-06
`
`

`

`From (613) 991—2760 .
`
`,
`
`0 rder ,# 054212670 P0447425 3
`
`Mon Apr 18 23:24:30 2005
`
`Page 8 of 17
`
`
`
`RM]. Char} I Advanced Drug Delivery Reviews 31 (I993) 89— 104
`
`95
`
`:aloids. or doxoru—
`.r 10'“ M to to"
`eat of cell kill in
`
`the cytotoxicity
`:
`y of detecting low
`ietection is about
`vide a measure of
`
`or the number of
`ation appeared to
`tg as compared to
`' example, conju—
`bodies BR96 and
`
`' antigen found on
`.nd fivefold less
`free doxorubicin
`. is live-fold more
`
`iK.
`
`i;
`:5.
`
`ii;
`i;
`j
`.'
`l:
`
`
`
`.,..-.._..___...,._.._._._.
`
`, cis-aconitate linked
`
`potent than doxorubicin conjugated to the antibody
`SEN‘7 directed against small cell
`lung cancer [44].
`Although conjugation of antibodies
`to drugs
`is
`purported to increase their specificity of cell kill, in
`vitro cytotoxicity data with antibody conjugates of
`anti-cancer drugs have been somewhat disappointing.
`‘ In conjugates where antibody and doxorubicin were
`linked via acid-labile hydrazone bonds, antigen-
`negative cells were only five to ten-fold less sensitive
`to conjugate (24 h exposure) than antigen-positive
`cells, suggesting that this type of link is quite labile
`[42-44]. Even when stored at optimal pH and
`temperature, significant hydrolysis of the hydrazonc
`link to release active drug has been reported I45].
`Conjugates of daunorubicin and doxorubicin linked
`via an acid labile cis-aconitate linker were also
`cytotoxic in the 10"6 to 10 7 M range, but showed a
`greater degree of antigen-specificity Suggesting that
`this link was more stable than the hydrazone link.
`[35].
`
`4.3.
`
`In. vivo efficacy
`
`New anticancer agents are gcncrally assessed for
`in vivo anti-tumor efficacy in animal tumor models
`to determine their potential for further evaluation in
`patients. Antibody—drug conjugates have been no
`exception and several of them have been evaluated
`for anti-tumor activity in human tumor xcnograft
`models in immunodeficient mice. Most of these
`conjugates showed greater
`therapeutic efficacy in
`vivo than the corresponding unconjugated drugs or
`isotype-matched
`but
`non-binding
`antibody—drug
`conjugates. in somc cases, biodisl‘ribution studies in
`tumor-bearing mice indicated that the antibody~drug
`conjugate was indcul localized at the tumor site. For
`example, when lymphoma-bearing mice were treated
`with a methotrcxatc conjugate of the anti-EL4 anti—
`body, after 3 11 as much as 15% of the administered
`dose was localized per gram of .tumor
`[46].
`In
`another experiment, comparative studies
`in mice
`with free doxorubicin or a conjugate of doxorubicin
`with the anti-melanoma antibody 9.2.27 showed that
`there was a preferential accumulation of cloxorubicin
`at
`the tumor when delivered as ‘a conjugate as
`compared to'that of free doxorubicin [47]. After'48
`11, there was a 45-fold greater amount (if conjugated
`doxorubicin at the tumor as compared to free dom-
`rubicin. Experiments in tumor—bearing mice indi-
`
`cated that greater tumor localization translated into
`better anti-tumor efficacy, with 9.2.27-Doxombicin
`being significantly more efficacious than free dom-
`rubicio. Another doxorubicin conjugate (BR96-Dox)
`with the human-mouse chimeric monoclonal anti-
`
`body BR96 has been shown to completely Cure
`athymic mice that had been implanted subcutaneous-
`ly with human lung or colon tumor xenografts [42].
`The anti-tumor effect observed is especially impres-
`sive as the tumors were well—established prior to
`commencement of therapy, and the treated animals
`were tumor-free for up to 200 days. However,
`the
`dose of conjugate necessary to achieve 100% cures
`was quite large (2.1 to 2.7 g/kg conjugate, equiva-
`lent to 60 mg/kg doxorubicin). Doaorubicin conju-
`gates of other monoclonal antibodies have also been
`shown to induce tumor regressions or cause signifi—
`cant delay in tumor growth in mice, while unconju—
`gated doxorubicin or non-binding antibody-doxoru-
`bicin only showed a slight therapeutic effect in these
`models {48,491.
`Several conjugates of the Vinca alkaloids linked
`via acid-labile hydrazonc bonds to different mono-
`clonal antibodies have been tested for anti-tumor
`
`tumor models. The best efficacy
`efficacy in animal
`was observed with a conjugate of the vinblastinc
`derivative DAVLBl-IYD with the murine monoclonal
`
`antibody K5114 which reacts with human adeno-
`carcinomas. Significant suppression of tumor growth
`in a well-established subcutaneous lung xcnograft
`model
`in nude mice was demonstrated [50]. Al-
`though complete tumor regression was not observer],
`the conjugate showed much greater efficacy than -
`both the unconjugated vinblastine derivative or the.
`non-binding antibody conjugate. The same conjugate
`was also tested in a human ovarian carcinoma
`
`survival mode] in athymic mice and showed a five-
`fold increase in survival when compared with un-
`
`the
`In this model,
`conjugated DAVLBHYD [51].
`also
`non-binding antibodyvvinblastine
`conjugate
`showed some therapeutic effect causing a 25-fold
`increase in life span as compared to frcc drug. 'l‘his
`non-specific anti-tumor effect might be explained by
`the fact. that the non-binding conjugate lasts longer in
`circulation than free vinblastine,
`thus acting as a
`‘slow-release’ formulation offtlie drug.
`in vivo tumor efficacy has also been demonstrated
`with conjugates of monoclonal
`antibodies with
`milomycin C [24] and idaiubicin [26].
`
`
`
`PAGE 8/17 ' RCVD AT 4/18/2005 8:28:24 PM [Paclflc Dayllght Tlme] ‘ SVR:SVCSD1IO ' DNI
`
`PHIGENIX
`
`Exhibit 1015-07
`
`

`

`From (613) 991—2760 ..
`
`‘
`
`Order .# 0 54212 67DP0447425 3
`
`Mon Apr 18 23:24:30 2005
`
`Page 9 of 17
`
`
`
`R.V.J. (.‘liari I Advanced Drug Delivery Reviews 3] (19.98) 8.9- I04
`
`4.4. Clinical evaluation
`
`The pre-Iclinical studies with antibody conjugates
`of anti—cancer drugs demonstrated that the conjugates
`generally displayed greater anti—tumor efficacy than
`the respective free drugs, suggesting that
`targeted
`delivery of drugs was indeed improving the tumor
`localization and selectivity of . the dl'ngS. These
`encouraging results led to the further evaluation of
`some of these agents.
`in humans. A conjugate of .
`methotrexate and the monoclonal antibody, KSl/4.
`was evaluated in two different Phase I human
`
`trials in patients with non-smallcell lung
`clinical
`cancer. A majority of patients elicited a human
`anti-mouse antibody (HAMA)
`response
`[52,53].
`Although, immunopemxidase staining of carcinoma
`samples provided evidence of post-treatment locali-
`zation of the conjugate,
`there was no therapeutic
`effect and clinical responses were not observed in
`either study. The same KSl/4 antibody was also
`linked to the vinblastine derivative DAVLBl-IYD.
`
`trials in patients with
`and evaluated in clinical
`adenocarcinomas. Again. a majority of the patients
`elicited a HAMA response resulting in rapid clear-
`ance of the conjugate. and no clinical restionses were
`observed in these studies [54,55]. Further human
`studies with these immunoeoniugales appear to have
`been discontinued. Recently, a phase I human clini~
`cal
`trial with BR96-Dox (chimeric monoclonal anti-
`body BR96 linked to the anti-cancer drug doxorubi—
`cin) has been completed in patients with colon
`cancer
`[56]. Preliminary indications are that
`this
`conjugate was also not
`therapeutically effective.
`Thus,
`in clinical
`trials conducted so far, early
`antibody