`
`
`
`Wi—
`
`
`
`Reseorch ,
`
`
`
`
`
`
`R. CANQE-SRRESEARC‘H“'
`EjAfME‘h-liégA'Niqksisfl’QlC';,ATiONé
`
`
`
`
`
`HEALTH SCIENCES LIBRARY
`University of Wisconsin
`
`SEP 2 I 1995
`
`1305 Linden Dr'
`
`Madison, WI 539,86
`
`‘SepTember15, 1995
`
`golume 55 ' Number 18
`P39494204
`lssN 0008—5472 - CNREA 8
`
`
`
`IMMUNOGEN 2177 pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2177 pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`[CANCER lll‘SltAIK‘H 55. 4117‘IAJIN4. September 15,
`
`l‘ILlSl
`
`Enhancement of the Selectivity and Antitumor Efficacy of a CC-1065 Analogue
`
`through Immunoconjugate Formation
`
`Ravi V. J. Chari,l Kristine A. Jackel, Lizabeth A. Bourret, Susan M. Derr, B. Mitra Tadayoni, Kristin M. Mattocks,
`sudhir A. Shah.2 Changnian Lin, Walter A. Bltittler, and Victor S. Goldrnacher
`lumirr'mGr'n. In... Carri/Midge,
`l1ItL\.\U(‘/lll.i‘t'il.\ (NHL) 42 )9
`
`MATERIALS AND METHODS
`
`tDCl-(lerivatives 1a and lb) and SAMB 3.
`Bis-indolyl-seco-Clil
`Swo»(‘Bl was prepared as described previously (20) and convened to l')(.‘l
`derivatives la and lb by methods that will be described separately." SAME 3
`is rr new bit'rrnctional cross—linking agent
`that enables the introduction of
`sterically hindered sultlrydryl groups on an antibody, Details of the synthesis
`of SAMB will be described elsewhere.1
`mAbs. Murine mAhs anti-B4 (anti-CD19: Ref. 21) and N901 (antieCD56:
`Ref. 22) were purified by methods described previously for anti-134 (23). The
`humanized versions of autieB4 and N901, called h—arrti-Bcl and hN901, were
`generated at
`lmmunoGen by trsing variable domain resurfacing (24) and
`grafting of complementarity determining regions,S respectively, on the frame-
`work ol' a human lgGl antibody molecule.
`Conjugation of DCl-derivalive 1b with mAbs. To introduce sultlrydryl
`groups, antibodies (10 mg/url) were first
`incubated at 30°C with a 14-fold
`molar excess of SAMB 3 in 0.1 M potassium phosphate buffer (pH 7.0).
`containing 2 mM HD'l'A, for 30 min. followed by treatment Willi hydroxyla-
`mine (50 mM) for 15 rrrirr at 30°C. The modified antibody (typically containing
`between 5 and 0 thiol groups/molecule of antibody, as determined by an
`Ellman’s assay; Ref. 25) was separated from low molecular weight comporrnds
`by gel filtration through a Sephadex 625 column equilibrated in 5 mM sodium
`acetate buffer (pH 4.7). containing 50 min NaCl and 1 mm EDTA. The
`modified antibody was then diluted to 1 mg/ml. and the pH was adjusted to 7.5
`by the addition of 0.1 M sodium phosphate buffer (pll 75). containing 20%
`dimetliylacetamide and 1 mM ED’tA and then incubated immediately at 37°C
`with [)t'fl derivative lb (1.5 equivalerrts/sultliydryl group. from a stock solu-
`tion of 10 mg/ml in dimcthylrrcetamide) for 80 min. Any unrcacted thiol groups
`on the antibody were then converted to thioethers by treatment with N»
`ethylmalciniidc (0.075 mM) for 15 min at 37°C. The conjugate was purified by
`mixing with Porapak Q (Millipore Corp. Milford. MA) chrorrratography resin
`(0.4 ml settled resin/mg aritihody) for 2 h at room temperature to remove
`unconjugated drug and subsequent dialysis at ambient temperaturc for 24 11
`into PBS containing 20% propylene glycol. The content of free drug in the
`antibody-DC] conjugate preparations was determined as follows: (a) low
`molecular weight compounds were separated from conjugate by passing the
`conjugate solutions through a Sepd’ak C-lS cartridge (Waters Chromatogra-
`phy, Milford, MA); and (Ir) recovered small molecular weight compounds
`were tlreu analyzed and quantitated by lll’LC on a reverse-phase Cm column
`(BAKERBOND; l. T. Baker, lnc.. l’hillipsburg, NJ). The conjugates contained
`«1.25% free drug (lower detection limit of the assay). Conjugates containing
`an average of 4 5 drug molecules linked/antibody molecule were prepared by
`this method as calculated from absorbauce measurements by using the extinc—
`tion coefficrents 52w "m
`2.24 X 105 M"cm" for antibody. and
`61mm“ 7 4.17 X 10' MT'CmTl. and em "I“ = 4.17 X 104 Mi'cm" for drug.
`Conjugate solutions ((1.5 rug/ml) were quickly frozen in liquid nitrogen and
`then stored at —20°C.
`Cell Lines. Human cell lines Namalwa (Burkitt‘s lymphoma, ATCC CR1.
`1432). MOLT-4 (acute lyrnphoblastic leukemia. ATCC CRL 1582), NALM-6
`(non—B iron-T acute lymphoblastic leukemia; Ref. 26), and SW2 (human small
`cell
`lung carcinoma; Ref. 27) were maintained as exponentially growing
`cultures in RPMI 1040 medium supplemented with 10% heat-inactivated
`bovine calf serum and 2 mM L-glutamine.
`Binding Studies. Relative affinities of antibodies and their DCI conjugates
`for their cognate antigens were determined in competition binding assays on
`isolated plasma n‘rembrancs (28). The proportion of inactive aritieBr—l (lgG
`
`
`
`" R. V. .1. Char
`'nthc s and antiturnor activity or distrlfrrlc containing bcnzarnnelatcd
`analogues ot' (‘C-ltloS. manuscript in preparation.
`‘ M, Rnguska and 13. Guild. unpublished data.
`
`
`
`
`
`
`
`ABSTRACT
`
`Bis-indolyl-(seca1-1,2,9.9a—tetraliydrocyclopropa[rlbenziaIindol-4-0ne
`compounds are synthetic analogues of (TC-1065 that are highly cytotoxic
`toward a broad spectrum of tumor cell lilies. One of these compounds.
`called DC], was conjugated to antibodies via novel cleavable (lisullide
`linkers. Conjugates of DCl with murine mAbs anti-B4 and N901 directed
`against tumor-associated antigens CD19 and CD56. respectively, proved
`to be extremely potent and antigen selective in killing target cells in
`culture. DCl conjugates with humanized versions of anti~ll4 and N901
`antibodies were also constructed and demonstrated to he as cytotoxic and
`selective as the respective murine antibody conjugates. The ariti-B4-DC1
`cnnjugate showed antitumor efficacy in an aggressive metastatic lnrman
`B-ccll lymphoma survival model
`in SCID mice and completely cured
`animals bearing large tumors. Anti-B4-DC1 was considerably more effec-
`tive in this tumor model than doxoruhicin. cyclophosphamide, ctoposide,
`or viricristine at their maximum tolerated doses.
`
`lNTt‘LODUCTION
`
`A continuing challenge in designing new anticancer chemothera—
`peutic drugs is to develop agents possessing superior antiturnor effi—
`cacy combined with reduced toxic side effects. One widely used
`approach toward this goal has been to target conventional anticancer
`drugs to tumor sites by conjugating these drugs to antibodies that have
`selective affinity [or tumor cells (1714). In general, these antibody-
`drug conjugates proved to be only modestly cytotoxic for cells in vitro
`with IFS“ values ranging from 1077 M to 1075 M. and most of them
`were significantly less potent than the respective drugs in nonconju-
`gated form.
`We (15) and others (5) have reported recently about novel airti-
`body-drug conjugates that had greatly increased potency. These cone
`jtlgates were at least 100-fold rrrore Cytotoxic to Cultured cells than
`imrnrruocoirjugates of clinically used anticancer drugs such as doxo—
`ruhicin. melphalan, mitomycin C. cisplatin. bleomycin, and Vim-(1
`alkaloids. Furthermore,
`in contrast to the unconjugated drugs. these
`tinniunoconjugates killed cells in an antigeneselective manner.
`in an extension of this work, we report here the preparation and
`testing of immunoconjugales comprising a new class of cytotoxic
`drugs that have a broad spectrum of activity and are highly Cytotoxic
`for a wide variety of tumor cell
`types. These compounds are his—
`irrtlolyharem-CB]3 derivatives, which. like adozelesin (16) are syn—
`lhetic analogues of CC—1065. The mode of action of these cytotoxic
`drugs is DNA alkylation at adenine bases after binding with high
`affinity to the minor groove of the DNA in AT-rich regions (17—19).
`
`Received 4/6/95; accepted 7/I0/05
`The costs of publication of this article were defrayed in part by the payment at page
`chills“. This article must llicrcforc be hereby marked iii/rit'rtitr'nrwn in accordance with
`18 USC. Section 1734 solely to indrcate this lact.
`
`1To whom requests for reprints should be addressed. at lrumunchn. Inc. [2s Sidney
`
`
`Stree1.(arirbri 5 MA (1213‘) 4 i9.
`;l‘rc\cnl rrddr
`Oncogene
`cicncc. St) Rogers Street. L‘ruubridgc. MA 021112.
`The abbrevralions used are:
`(‘81.
`l.2.9.9a-tctrrthydrocyclopropa[r'jlicnz[u]indol-
`
`‘1‘006;
`L'l’l.
`-
`iydro 7 mctbylcvclopropa[i'lpyrrololllulindolr—H5H)eonc;
`
`SAME r\l‘-.
`_r| 4—/\L(2’eact>1yltlriopropiouyl)-N—rnelhylr-lerrnirnobulyrate; MST.
`median ~urw
`tune: MTD. nirrxrmuru tolerated dose: sum-C151.
`lrclrlorurnetlrylrir
`hydmxyl.Z-drliydro-SHebcnz-indolc.
`
`407‘)
`
`L
`
`This material may be protected by Copyright law (Title 17 US. Code)
`
`IMMUNOGEN 2177 pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2177 pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`AN‘IIH'MOR ACTIVITY OF DCt
`
`IMMUNOC‘ONJUGATES
`
`incapable of binding to CD197expressing cells) in samples of native anti-B4
`antibody and anti-B4-DCI was determined by the method described elsewhere
`(29). Samples (0.6 ml) of anti-B4 (l .0 nM) or antleB4~DC1 (0.8 nM) in Eagle‘s
`MEM for suspension cultures supplemented with 2.5% pooled human serum of
`AB type and 10 mM HEPES buffer (pH 72) were incubated with CD197
`expressing Namalwa cells or CD19—negative MOLT»4 cells in several succes-
`sive incubations (4 X 107 fresh cells/incubation) until
`the samples were
`depleted of all
`the conjugate that was capable of binding to the cells. The
`concentration of the remaining immunoconjugate was determined by an
`ELISA (31)) using anti-E4 antibody and anti—B4-DC1 for generating the
`standard curves.
`In Vim; Cytotoxicity Assays. The method for the determination of sur-
`viving fractions of cell cultures by back extrapolation of the exponential
`growth curve is similar to that described previously (Ref. 31 and references
`therein). Cell cultures that had been exposed to a drug or a conjugate for a
`specified period of time were washed once with prewarmed medium and
`resuspended in fresh growth medium at a density of about 273 X 105 cells/ml.
`The cells were then cultured and counted daily by usmg a Coulter Counter. The
`cultures were diluted with fresh medium to 3 X 103 cells/ml when they reached
`a density above 6 X 10‘ cells/ml. The daily increase in the cell density was
`used to calculate an increase of the number of cells in the culture assuming no
`dilution. An estimate of the number of surviving cells was nrade by extrapo-
`lation of exponential growth curves of treated cultures (in semilogarithmic
`plots) to the end of the treatment period. The surviving fractions were calcu-
`lated as the extrapolated number of surviving cells in treated culturenrumber of
`cells in the control culture ratio. One unique feature of DC1-containing
`compounds and their antibody conjugates is that
`the process of cell death
`induced by these agents is slow, and the cells undergo up to four divisions
`before dying (data not shown). Therefore, to obtain accurate surviving frace
`tions of cells. the proliferation of cell cultures treated with DC1 derivatives or
`their immunoeonjugates was monitored for at least eight doubling times.
`In Vivo Toxicity Studies. Systemic toxicity of DC1 immunoconjugates
`was evaluated in female CD-1 mice. Animals received either a single i.v.
`(through tail vein) bolus injection (acute toxicity study, eight animals/dose) or
`daily iv bolus injections for 5 consecutive days (subacute toxicity study, ten
`animals/dose) of a conjugate in PBS containing 20% propylene glycol and
`human serum albumin (1 mg/ml). Toxicity was assessed by observing animals
`for weight
`loss and survival. The maximum tolerated dose under subacute
`dosing conditions was also determined in female CB-17 SCID mice obtained
`from the Department of Radiation Oncology, Massachusetts General Hospital,
`Boston. MA.
`Antitumor Activity of Conjugates. The therapeutic efficacy of immuno—
`conjugates with DC] was evaluated in 7—week-old female SCID mice bearing
`a human lymphoma (Namalwa) xenograft (32). Animals were injected iv.
`(through tail vein) with 4 X 10“ Namalwa cells. It has been shown previously
`that tumors metastasize to ovaries,
`to the leptomeningcs of the brain, bone
`marrow of the vertebra, the dura matter of the spinal cord, and to muscles
`surrounding the spinal column. Control animals received no treatment because
`we observed previously that treatment with PBS did not alter the survival of
`the tumorebearing mice. The median survival
`time of untreated mice was
`23—24 days. Treatment with irnrnunoconjugatcs or uneonjugated drugs was
`initiated 7-days posttumor inoculation unless indicated otherwise. The extent
`ofcell kill (log cell kill) in viva was estimated as described previously (32, 33).
`Mice that were alive on day 120 of the studies were sacrificed and examined
`histopathologically for signs of residual tumor by Dr. Roderick '1‘. Bronson
`(School of Veterinary Medicine, Tufts University, Boston, MA).
`Determination of Serum Concentrations of anti-B4—DC1. Serum cone
`centrations of anti-B4—DCI
`in SCID mice bearing Namalwa tumors were
`determined leh post i.v, administration of the conjugate at a dose of 5 rug/kg
`by an ELISA method by using CD I 9-expressing Namalwa plasma membrane—
`coated plates to bind the conjugate and Enuelabeled anti-murine lgG to detect
`the conjugate as described elsewhere (28).
`
`RESULTS
`
`We modeled DC1 after adozelesin 2, which is a synthetic analogue
`of the natural product CC—1065 (Fig. 1). Adozelesin has been found to
`be as cytotoxic as CC—lDoS but without having the delayed toxicity in
`4080
`
`animals of CC~1065 (16). The alkylating portion of adozelesin Con.
`sists of the CPI unit. which we replaced in DC] with seen-CB1 (20)
`Bis~indolyleCBI compounds are at
`least as cytotoxic as the COrre
`sponding bis-indolyl-CPI compounds (34. 35) and, importantly, FBI
`and its Openvr'lrrg prodrug form, called seen-CB1, are more stable iii
`aqueous solutions and easier to synthesize than CPI. seco-CBI readily
`converts into the active form, CB1, under physiological conditmns
`with a half-life of about 3.5 h in human plasma at 37°C (data m
`shown). We, therefore, incorporated seen—CB1 into DC1. Finally, [0
`enable linkage of DC1 to an antibody via a disulfide bond, We
`replaced the terminal unsubstituted benzofuran unit of adozelesin by
`an indole ring bearing a reactive pyridyldithiopropionamido substit.
`uent at the C-5 position,
`The anti—B4 antibody was modified with the cross-linking reagcm
`SAMB 3 and then treated with hydroxylamine to deprotect the thin]
`groups. These antibody-bound thiol groups reacted readily with DC1
`derivative 1b via a disulfide bond exchange to yield an antibody-DC1
`conjugate (Fig. 1). Immunoconjugates containing. On the average, 4—5
`DC1 molecules/anti-B4 antibody molecule were prepared and are
`referred to as antivB4—DC1 in this report. Binding studies on CD19—
`expressing cell membranes demonstrated that such conjuga‘es pre—
`served nearly all of the binding activity of native anti—B4 antibody
`(Fig. 2). The apparent affinity was only 2737fold lower, which, in
`principle, can be explained by either a complete inactivation of
`50~75% of the antibody molecules during their modification and
`conjugation to DC1, or by a general 273efold decrease in the avidity,
`We, therefore, determined the percentage of immunoreactive conju-
`gate in the preparation. A sample of anti—B4—DC1 was incubated
`successively with two batches of CD19-expressing Namalvta cells.
`After each incubation, the remaining conjugate in solution was deter-
`mined by an ELISA (Table 1). AntieB4-DC1 behaved similarly to
`anti-B4 antibody, and after two incubations, <6.3% of the initial
`amount of conjugate was present in solution (the detection lirrit). We
`concluded that at least 93% of the conjugate had bound to CD19-
`positive Namalwa cells. The binding was antigen specific because
`similar irrcubations with the CDl9—negative cell litre MOLT-4 did not
`significantly lower the initial concentrations of conjugate in the solu-
`tion.
`
`Anti-B4—DC] proved to be highly potent, and killed cultured CD19—
`expressing Namalwa cells after an exposure of as short as 15 min (Fig.
`3) with an ICDU value of 3 X 10’ '0 M (0.21 ng/ml) of DC1 (equivalent
`to 7.5 X 10"1 M conjugate). A 15—min exposure to a higher concen-
`tration of the conjugated DC1 of4 X 10’8 M (equivalent to l X 10-5
`M conjugate) eradicated >99.999% cells (the detection limit of tht‘
`assay). The effectiveness of cell kill further increased with a prolon-
`gation of the exposure time to 24 or 72 h, with ICSO values Of
`6 X 10’” M and 5 X 10’“ M, respectively (Fig. 3). Remarkably.
`concentrations as low as 4 X 10"U M (24—h exposure) and 1.2 *1 10'9
`M (72-h exposure) were sufficient to kill S logs of cells, In marde
`contrast, at all exposure times measured, the conjugate was essentially
`nontoxic for antigen-negative MOLT-4 cells with IC50 values
`>4 X 10 8 M DC1 (equivalent
`to l X 10’“ M conjugate). Tire
`effectiveness of anti-B4-DC1 was not limited to the target cell linC
`Namalwa because it was effective in killing NALM-6. Smother
`CD19-positive cell line with an Icy, value of 8 x 10*“ M D91
`(2 X IOTll M conjugate) after a 24—h exposure (Fig. 3). T1115
`cytotoxic effect was abolished in the presence of nonconjugaled
`antieB4 antibody (Fig. 3), again demonstrating antigen specificity
`of the cytotoxic effect.
`In contrast
`to anti-U4-DC1 conjugate,
`uneonjugated DC1-derivative 1a was equally cytotoxic for Nflmfl‘
`lwa and MOLT—4 cells (ICSH of2 X 10—” M after a 24-h canSleri
`data not shown).
`The effectiveness of [)Cl when incorporated into immun-rCOnlu‘
`
`
`
`IMMUNOGEN 2177 pg. 3
`Phigenix v. Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2177 pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`/\N'll'|'UMOtt ACTIVITY OF DC]
`
`IMMUNOCONJUGATES
`
`AbwwSH
`
`°
`a
`at“:
`::1~10(3(c112)31ucocuscocua
`o
`
`3
`
`1_
`
`2. NH20H
`
`1 b
`
`Ab‘A’VV‘SH
`
`s —
`
`/N
`
`H
`
`0
`
`N
`
`
`
`Ab
`
`'4,‘
`
`/Cl
`
`\
`
`
`
`jxflH
`HmNHCOCHZCHZSSWAb
`
`
`(80 ug/kg/day ofDCI , equivalent to 5 mg/kg/day of conjugate), or the
`isotypeanalchcd, nonbinding conjugate N901—DC1 (5 mg/kg/day of
`conjugate), or a mixture consisting of anti—B4 antibody (5 mg/kg/day)
`and unconjugated DCl derivative la (80 og/kg/day). The control
`group of mice that was left untreated had a MST of 23 days. and all
`animals were dead by day 20 (Fig. 4A). Mice treated with anti,
`B47DC| showed significantly prolonged survivals with a 27—fold
`longer MST of 62 days. The nonbinding conjugate N901—DC1. or the
`mixture of anti—B4 antibody and drug la, had only modest therapeutic
`effects. prolonging the MST to 42 and 30 days. respectively. The
`increased therapeutic effect of the nonbinding conjugate of DCI
`compared to the nonconjugated drug may be due to the slow release
`of the drug from the conjugate over an extended period of time.
`Inoculation of SCID mice with as few as 4 X 10} Namalwa cells
`killed all mice within 42 (lays with an MST of 39 days (32). In the
`experiment shown in Fig. 4A, the tumor burden was estimated to he
`7 X l0' Namalwa cells at the beginning of treatment (based on their
`in viva doubling time of 1.7 days; Ref. 32). This direct comparison
`shows that anti—B4—DCI with an MST of ()2 days eradicales more than
`4 logs of tumor cells. The tumor cell kill
`is at
`least 6 legs when
`calculated by the method of Corbett and Valeriote (33). Anti-B47DC1
`was similarly effective when administered every other day or every
`third day (in both cases, MST values were between 77 and 81 days;
`data not shown).
`To better understand the meaning of the in viva therapeutic effects
`of anti—B4—DC1 in the xenograft tumor model, we treated mice hear
`ing the same tumors with chemotherapeutic drugs that are known to
`)e clinically effective against lymphoma (Fig. 43). Drugs were used
`at their previously determined MTDs (causing transient body weight
`oss of 20%). Higher doses resulted in systemic toxicity manifested by
`Jody weightless of >20% and early death (within 3 days after the last
`injection) of several animals. Optimal
`treatment schedules. which
`resulted in the maximal therapeutic effect, were determined by vary—
`ing doses, schedules (daily, every other day. every third day, or every
`"hourt day), and the total numbel of injections (between a single
`injection and 11p to tlnee injections) were evaluated in this survival
`mode] for each drug. The most effective treatment schedules at MTD
`or cyclophosphamide (100 mg/kg/day X 3 every 2 days. i.p.), vin-
`cristine (1 mg/kg/day X 3 every 2 days, iv), etoposide ([S ing/kg/
`day X 3 every 2 days. i.v.), or doxorubicin (3 mg,’kg/day X 3 every
`4 days i..v) had only a modest prolongation of survival of the
`urnorbeating mice with MST values of 44 37 32, and 28 days
`respectively,111 comparison to unheated mice that died with an MST
`of 23 days. Anti—B4—DC1 (80 ug DCl/kg/day X 5 every day i.v.) with
`an MST of 62 days was, therefore, significantly more efficacious than
`any of these chemotherapeutic agents.
`Animals treated on day 7 postinoculalion of tumor cells had a large
`tumor burden of 7 X 107 cells and were not completely cured. We
`then wished to determine if an earlier initiation of anti—B4—DC1
`therapy at the time of a lower tumor burden could achieve complete
`cures of mice. Treatment of animals with anti—B4—DC] (80 og/kg/
`day X 5 of DCl) that were again inoculated iv. with 4 X 106
`Namalwa cells was. therefore, begun on day 2, 3. or 5 postinoculation,
`representing tumor burdens of 9 X 10“. 1.4 X 107, and 3 X 107 cells.
`respectively (Fig 4C) More than one——half of the animals (6 of 8 and
`Set 8) that we1e treated2 or 3 days after inoculation of tumor were
`alive on day I20 and showed no signs of toxicity 01 forum growth.
`The mice were then sacrificed on day 120 and were examined his—
`topathologically for signs of residual tumors. No evidence of tumors
`SCID mice bea1ing a disseminated aggressive human lymphoma
`was found, and the animals were considered cured. 1n the group of
`Ken0g1’alt(32)that was lethal 1o1 nontrealcd mice in 22 20 days. Mice
`Were i.v.
`inoculated with 4 X 10“ human Namalwa cells, and the
`eight mice that were treated 011 day 5 after tumor inoculation.
`two
`animals were cured, and the remaining six animals had a much
`1“mm was allowed to establish for 7 days before therapy. Animals
`prolonged MST of 73 days.
`WW then
`treated on 5
`consecutive days with antiiB4—DC1
`4081
`
`
`
`
`
`
`
`:1. structures of adozclcsin Z and hIs-indolyl-a‘ccu-CBI drugs 1:1 and lb. 8.
`Fig, l.
`preparation of antihody-DCI conjugates.
`
`gates is not limited to the anti7B4 antibody A conjugate prepared with
`the murme anti—CD56 antibody N901 directed against small cell lung
`cancer cells (36) was found to be equally potent
`in an antigen—
`Selective manner toward its target cells (Table 2). lmportantly, for
`future clinical development, we also demonstrated that DC1 conju—
`
`gates ofhumanized versions ofantiiB4 and N901 antibodies were as
`
`potent and specific as their respective murinc antibody conjugates
`Table 2).
`T0 dc ermine the dose range of antiiB4iDC1 for efficacy studies in
`human tumor xenograft models in SCID mice, we first evaluated its
`Systemic toxicity. Under acute (a single i.v. dose) and subacute (one
`olus (lose/day for 5 days) conditions in CD—1 mice, LDSU values for
`anti~B4—DC1 were 225 Mg DCl/kg (equivalent to 14 mg/kg of anti-
`34‘13Cl Conjugate) and 465 ug DCl/kg cumulative dose over 5 days
`(equivalent to 29 mg/kg of anti—B4—DC1 conjugate). respectively. The
`MTD for anti—B4—DC1 in both SCID mice and CD—l mice was found
`‘0 be 80 Mg DCl/kg/day X 5. The dose-limiting toxicities of anti-
`3‘l-DCl were hematological effects. such as the early depression of
`Circulating leukocytes,
`followed by recovery, and lesions in renal
`ll‘lbules and lungs. The renal lesions were resolving on day 29 after
`lrst injkalOl] of conjugate.
`
`
`
`Theantitumorefficacyofmurineanti—84—DC]wasevaluatedin
`
`l l
`
`I
`
`l
`
`|
`
`‘
`
`‘
`
`IMMUNOGEN 2177 pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`l L
`
`IMMUNOGEN 2177 pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`ANTI TUMOR ACTIVITY (1)7 UPI
`
`iViMUNOCONJUGATES
`
`reasons for the higher antitumor efficacy of drug conjugates in X5
`nograft models compared to human trials may be the mucl: highs
`doses used in the animal studies; and the better delivery of the
`targeting antibodies to the xenograft tu1no1 sites in animals [105011
`tumor localization of injected iv dose (1, 49) in animal.
`“91.1111
`0 l~0. 000270 in humans (50)]
`
` i
`
`Recently, novel immunoconjugates incorporating more potent Cy
`totoxic agents (at least 100—fold more cytotoxic than CDmentional
`anticancer drugs) have been reported. Maytansinoids (15,-1), Cali.
`cheamicins (5) and morpholinodoxorubiein (8) were utilized in the“
`conjugates. Immunoconiugates containing DC] are comparable in
`their 1'11
`111110 potency to those containing maytansinoids (15) and
`appear to be somewhat more potent than calicheamicin cmtjugates
`(1C50 values of 3 X 10 J” M after a 15 min exposure vet.1113 a X 109
`M after a 7—min exposure respectively). Morpl1olinodoxo1uhici1-1'rm—
`munoconjugates described thus far (8) appeared not to be active
`Survival of SCID mice bearing a human lymphoma xenogmfi
`(Namalwa cells; Ref. 32) was used to examine the antitumor activity
`of antiB4—DC1. We chose this model for several reasons:
`((1)1113
`cancer is aggressive, highly metastatic, and quickly disseminates
`throughout the body; (b) the tumor model mimics the clinical situation
`because most cancer patients die of disseminated disease; (c) [he
`survival of patients is the most direct and widely used CIiterion for
`therapy evaluation: and (d) survival models allow the studv of the
`effects of therapy on minimal residual tumor burden
`In a direct comparison, antiB4DC1 showed a greater therapeutic
`effectin this model than did the anticancer drugs doxorubich1. cyclo-
`phosphamide etoposidc, and vincristine when used at their MTD. in
`an experiment, where animals were treated on day 7 after tumor
`inoculation, the median inCIease 111 life span in animals treated with
`
`i i
`
`i
`
`1
`
`i
`
`100
`
`so
`
`60
`
`40
`
`20
`
`>
`'3—
`.sg
`E':
`118
`3115
`as
`m
`
`0
`
`1mm
`IO‘” 10"“ 10-9
`10-8
`Concentration, M
`
`Fig. 2. Competition binding. Assays measure the competitron of binding of biotin‘
`labeled anti‘B4 by anti-B4»DC1 conjugate (O) or by antr—Brl antibody (C). Solutions
`Conturnrng biotinilabeled anti-B4 antibody (02 nM) and various concentrations of corn—
`peting anti-B4-DC‘] or anlirB4 antibody were incubated (18 it at ambient temperature)
`with CD19—expressing Namalwa cell membranes immobilized in 96»\vcii plates. Positive
`controls lacked competing antibody, whereas negative controls lacked biotin-labeled
`antibody. Plate—associated biotin-labeled anti-B4 was then quantified with Eu‘ ' riabeied
`streptavidin and plotted as 11 percentage of the fluorescence of the positive control sample
`versus the concentration of competing reagent.
`
`Table 1 Removal of inrnrnnnrenctiw 111111011111! [111111 samples qunli-B—tt milibodv 11ml
`(1111iVB4-DC] conjugate
`Antibody or
`Fraction of immunoreactive antibody
`
`
` Cell line conjugate remaining in the supernatant”
`Namalwa (CD19-positive)
`anti-B4
`41.0113”
`Namalwa (LiDl‘Lpositive)
`anti-B4»DC1
`<0.063b
`MOLT-4 (CD19-negalive)
`antiiB4
`0.92
`
`MOLT-4 (CD19—negative) 0.99 anti~B4—DC1
`
`" Antibody remaining in the supernatant after 2 cycles of treatment with cells as a
`fraction ofthc initial concentration The initial concentrations ol‘zmlirBéi and anti-B4-DC1
`were i and 0.8 nM respectiv.cly
`I Lower limits oitleteclionIn these experiments as determined by the standard curves.
`
`DISCUSSION
`
`
`
`1 W 1 MAP
`101
`10.2
`J
`
`1
`ac1..
`E 10-1
`I
`D 102
`.2
`5
`10-3
`E
`>
`g 1014
`._
`in
`E
`10.;
`
`_
`
`a
`
`Bis—indolyl-secn-CBI belongs to a group of synthetic analogues of
`CC—1065, such as adozelesin and carzelesin, which possess the high
`cytotoxic activity of CC-1065, but lack its delayed systemic toxicity.
`These drugs are DNA alkylators that bind tightly to the minor groove
`of DNA and then react covalently with adenine bases. To confer
`cellular specificity to biseiudoiyl-seco-CBI while maintaining its po—
`tency, we prepared several
`immunoconjugates with a bis—indolyl~
`seeo-CBI derivative called DC1. Interestingly, when target cells were
`exposed to these conjugates, the cytotoxic effects were only observed
`after several cell divisions after the treatment. This response is similar
`to that reported for cells exposed to CC—1065 analogues (37), sug—
`gesting to us that the conjugated and nonconjugated DC1 kill cells in
`an analogous manner.
`Antibody—DC1 conjugates proved to be extremely potent (ICSU in
`the range of 10— m M DC1) and antigen selective in killing tumor cells
`in 11111-0 and were able to eradicate greater than five logs of cells from
`cultures of antigenipositive cells. High potency may be an important
`feature in achieving efficacious levels at tumor sites in 111110. In cancer
`therapy,
`it has been well established that at least a 2—iog or greater
`reduction in the number of tumor cells is necessary to achieve a
`clinical remission and even a greater cell kill is required to prevent
`relapse (38). In general, previously reported conjugates of antibodies
`Target cells Nontarget cell». Specificity fact“r
`
`with conventional anticancer drugs had ICSO values of 1077-10 5 M
`Conjugate
`'l‘argel
`Nonlargcl
`A
`B
`131A ’_
`(17 4, 14, 39—43) and were effective in treatment of a number of
`Anti-B4-DCI
`Namalwa 1101.141
`so x111 1'
`>4 xro‘f
`><00
`xenograft tumors in mice. In contrast, for those few drug immuno—
`irrArrlirBh’irDCl Namalwa MOLT-4
`1.5 xmrt“
`>4 x1011
`>260
`11901—11171
`swz
`Namalwa
`1.2 x10"“
`2.11 ><10'b
`3110
`
`conjugates that have been evaluated in human clinical trials, the peak
`hN901»DCl
`swz
`Namalwa 96x10‘”
`39 >110"
`«10/
`5
`circulating levels achieved were 111
`the range of their 1'11 vitro [€51J
`'Ceils were exposed to the Conjugates for 24 I1 and the surviving flilCii“; is of cell
`values (between 104‘ and 10—7 M; Refs. 44—48), and little therapeutic
`were then determined by track extrapolating the e\poncn1ial gronlh eurvcs (seeMaterifl'S
`.nnl
`'Vlettl1.ods") ICE”, inhibiting concentratron ol drug that kills SOC/r ot Cells.
`effect has been observed thus far (1, 44—48). Some of the possible
`4082
`
`\
`
`‘
`
`B
`
`107
`
`1o
`10-4
`.
`10”
`‘
`A
`mm Hwy—WWWr
`10-1“
`to9
`10.1
`111-7
`o
`10"0
`10°
`101
`Concentration, M
`Fig. 3 In 111110 cytotoxicity and specificity of antiB4DC] A 111 1111111 cyli1loxiciii'9‘
`antiB4DC1 tor CD19~positive Namalwa cells (0, A, and I) or CD19negativeMOLT‘i
`
`
`
`cells (0L., and D) Cells were exposed to various concentrations of hairB4-DC1 for 15
`min (0 and O), 2411(A and A) or 72 h (I anti
`) [he surviving fractions 01 cells We“?
`then determined by the growth back-extrapolation assay as ilesrribed clsewher (31) The
`211111784110 used in 11115 experiment had 4.0 DC1 molecules linked/antibody molecule.
`and the results are plotted in terms of DC1 concentration. B 111 111111 Cytotoxicity Hf
`anti-i34—DC1 for CD19positive i\Ai.M (1 cells after a 24-h exposure was measured in iii”
`absence (0) or in the presence of excess (l X 10" yr) antiB4 antihodi
`(O). The
`surviving iractions of cells were determined by the growth backextrapolation assay (3])‘
`
`Table 2 I11 1111171 cytotoxicity and specificity of antibody-DC] conjngri'cs
`Cell line
`[Cam M"
`
`IMMUNOGEN 2177 pg. 5
`Phigenix v. Immunogen
`|PR2014-00676
`
`IMMUNOGEN 2177 pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`ANTlTUMOR ACTIVITY OF DCl
`
`IMMUNOCONJUGATES
`
`100'
`
`50
`
`D
`
`0
`_r O O l
`
`50
`
`_; O O I
`
`
`
`PercentSurvival
`
`20
`
`40
`
`60
`
`80
`
`100
`
`120
`
`60
`
`80
`
`100
`
`120
`
`
`
`
`
`0
`
`20
`
`4O
`
`60
`
`BO
`
`100
`
`120
`
`Days After Tumor inoculation
`
`Fig. 4. In viva antitumor efficacy and specificity of anti-B4-DC1 and of conventional chemotherapeutic drugs. A, antitumor efficacy and specificity of untieBLieDCl. SCID mice (10
`
`animals/group) were i.v. inoculated with 4 X 10" Namalwa cells. Animals were left untreated (control group, I) or were treated iv. on day 7 after tumor inoculation with a mixture
`ofanti—B-i (5 mg/kg/day X 5 every day) and unconjugatcd DC] (80 ug/kg/day X 5 every day) (I), or with the isotypeemalchcd but nonbinding conjugate, N001-DC1 (O) at a DCl
`
`
`
`dose of EU uglkg/day x 5 every day, or with anti-B4—DCI (O) at a DCl dose of 80 ng/kg/day X 5 every day. B. comparison of the therapeutic efficacy of anti~BIl7DCl with that of
`conventional anticancer drugs. SCID mice (8710 animals/group) were i.v. inoculated with 4 X if)6 Namalwa cells. Animals were left untreated (control group
`), or treated on day
`7after tumor inoculation with one of the following agents: anti»l54-DCl (C) at a DUI dose of 80 ug’kg/day X 5 every day, i.v.. duxorubicin (A), 3 mg’kg’day X 3 every 4 days, i.v.;
`etoposidr (O),
`5 mg/kg/day X 3 every 2 days, i.v ; vincrislinc (I), l mgkg/day X 3 every 2 days, i.v.; or cyclophospliamide (A), 100 mg/kg/day X 3 every 2 days. i.p. C, antitumor
`
`
`efficacy of anti-B4-DC1 in SCID mice bearing varying tumor burdens. Mice (8710 animals/group) were i.v. inoculated with 4 X IO“ Namalwa cells. Animals were left untre