Volume 2, Number 1, 1989
`ISSN: 0892-7049
`
`IMMUNOGEN 2327, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`~~b11~0PDHAY, IMMUNOCONJUGATES, AND
`v
`RMACEUTICALS
`olume 2, Number I 1989
`Mary Ann Liebcn, In~., Publishers
`
`Pre-Clinical and Clinical Studies with N-Acetyl
`Melphalan Immunoconjugates and Tumor
`Necrosis Factor a
`
`GEOFFREY A. PIETERSZ, MARK J. SMYTH, JOE J. TJANDRA,
`and IAN F.C. McKENZIE
`
`Research Centre for Cancer and Tra11.1plantation, Department of Pathology, The University of Melbourne,
`Parkville, Victoria, 3052 Australia
`
`ABSTRACT
`
`t
`
`. Melphalan {MEL), a potent alkylating agent has been modified to a non-
`ox~c derivative, N-Acetyl Melphalan {N-AcMEL) which can be conjugated to
`monoclonal antibodies {MoAbs ) and used for immunotherapy. On average , 20
`rnol~cules of N-AcMEL can be attached to each antibody molecule and this
`con~ugate h as greater activity than free N-AcMEL. The N-AcMEL conjugates
`sat~sfy the basic requirements of specificity and potency for immunoconjugate
`use in vit r o.
`In vivo in mice, tumors growing in ascites fluid, and 20% of
`~mall subcutaneous tumors can be totally eradicated by N- AcMEL conjugates. For
`iarger tumors, a major problem exists in achieving access to the tumor by the
`mmunoconj ugate . To increase immunoconjugate localization we have used
`recombinant human tumor necrosis factor-a {r-TNF-U) to incite an inflammatory
`response in the tumor . Under these circumstances where N-AcMEL conjugates or r-
`TNF-a each
`·
`h
`·
`'
`·
`g~ve a proportion {20 %) of cures, t e cornb~nat~on of N-AcMEL-ant~body
`anct r-TNF-a leads to partial or complete regression of the majority of tumors.
`Furthe
`·
`dd
`r ~mprovement was found when recombinant mouse interferon-y {IFN-y) was
`a
`ect to the regimen
`on the basis of our preliminary preclinical data on N-
`AcMEL-MoAb
`.
`.
`,
`conJugates , a Phase I c l~ni ca l trial was done in patients with
`act
`vancect col
`.
`AcM
`orectal cancer us1ng MoAbs to colon carcinoma conjugated with N-
`EL . Thus far eight patients have been examined and in excess of lgrn MoAb
`~sect conjugated with 35mg of N-AcMEL. No patient showed any side effects of
`~:portance as 20mg/m2 is the maximum tolerated dose {MTD) for MEL in its free
`tr;m~ Only 6 of the 8 patients had evaluabl e disease {t otal of 9 courses of
`al a ment as one patient had two courses ) . Subjective improvement was noted in
`most all of th
`·
`res
`e pat1ents examined b u t only 29% {2/7) of the treatments s howed
`Ponse by strict objective criteria.
`
`INTRODUCTION
`
`Interest in the use of MoAbs to target cytotoxic drugs has grown
`enormousl
`.
`.
`Y ln recent years from the original concept of "mag~c bullets " of Paul
`Eh
`,
`{ 4 ~ 1 ~?h earlier this century {1,2). Since the early work of Mathe{3) and Ghose
`anct w~t? ~ethotrexate and chlorambucil , a number of other drugs {5 ), toxins {6)
`radlolsotopes {7) have been used as warheads for the "magic bullet". These
`
`47
`
`IMMUNOGEN 2327, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`Ehrlich earlier this century (1 , 2). Since the early work of Mathe(3) and Ghose
`(4) with methotrexate and chlorambucil, a number of other drugs (5) , toxins (6)
`and radioisotopes (7) have been used as warheads for the "magic bullet ". These
`conjugates are specifically toxic to antigen positive cell lines in vitro.
`In
`vivo efficacy of the immunoconjugates in animal models is varied, but the
`general pattern emerging is that ascites tumors growing in the peritoneum can be
`cured, while relatively poor responses are observed with solid subcutaneous
`tumors (8 , 9).
`Studies using radiolabeled MoAbs show that only a small f raction of the
`injected dose (<2 0 % in animal models and 0.01 % in humans) localized to the
`tumor, altho ugh a higher dose of radiolabeled MoAb fragments localized (10,11,
`12). Therefore, the lack of efficient permeation of immunoconjugate into the
`tumor tissue could be the cause of the less satisfactory response found with
`solid tumors. To increase blood flow through the tumor, and therefore antigen
`binding by immunoconjugate , we have used vasoactive agents with the
`immunoconjugate and an increase in efficacy was noted ( 13). However,
`immunoconjugates made using F(ab') 2 fragments were only marginally more effective
`than the corresponding conjugates made using intact antibody (14) .
`It is known
`that the permeability of the endothelium to immunoglobulin is enhanced during
`bacterial infection , produced by factors released in the inflammatory process
`(15); would such an event (increa se in capillary permeability) at the tumor site
`increase the amount of immunoconjugate that enters the tumor? TNF, a monokine
`~ith a myriad of biological activities has the capability of causing an
`lnflammatory response in tumors (16) and this paper reports the results of
`preclinical studies on the combination therapy of N- AcMEL- MoAb immunoconjugate
`with r-TNF -a and clinical studies in colon cancer using the immunoconjugate.
`
`MATERIALS AND METHODS
`
`(C57BL/6xBALB/c)F1 (B6CF 1) mice were produced in the Department of
`Patho l ogy at The University of Melbourne.
`
`l'umor cells
`E3, a clonal variant of the thymoma ITT(1 ) 75NS was used (17). The cell
`line was maintained in vitro in Dulbecco's modified Eagles medium (DMEM),
`s uppl emented with 10 % heat-inactivated newborn calf serum (Flow Laboratories ,
`Sydney, Australia), 2mM glutamine (Commonwealth Serum Laboratories (CSL) ,
`Melbourne Australia), 100~g streptomycin/ml (Glaxo Laboratories, Melbourne) and
`100IU penicillin/ml (CSL ) .
`In vivo E3 was maintained by serial passage in
`ascitic fluid in B6CF1 mice . For in vivo experiments the cells from the ascitic
`fluid were washed and centrifuged (1,500rpm for 5 min.) twice in DMEM and PBS ,
`re s us pended in Phosphate buffered saline (PBS ) , and inj ect e d s ubcutaneou s ly
`(s . c. ) int.o B6CF 1 mice .
`
`Monoclonal Antibodies
`Several MoAbs were used in this study (Table 1). A resetting assay (21)
`or an immunoperoxidase assay (2 2 ) was used to determine antibody activity of
`MoAbs and immunoconjugates . The choice of MoAb for use in drug conjugation was
`made by immunoperoxidase staining on frozen (for 250-30.6, I-1 & JGT MoAbs)
`and/or formalin - fixed (for I-1 and JGT MoAbs) tissue sections of colon cancer
`tissue of individual patients. The sections were then assessed by light
`microscopy to estimate the percentage o f colon carcinoma ce lls s tained with each
`of the antibodies and t h e result expre ssed on the scale 0 to 4 according to
`whether nil (0), up to 25 % (1), 25 - 50 % (2), 51 - 75 % (3) or >75%
`(4) of carcinoma
`ce ll s s tained
`(The intensity and the distribution of stain in the cancer cells
`and the staining of extracellular material were not taken into account) (23).
`The MoAbs se l ec ted for use in drug conjugation for an individual patient had a
`s taining score of 3 or 4. Therapy given to patient 1, the second time and
`patient s 3 and 6 had at least two MoAbs in the final drug-MoAb conjuga~e
`preparation, given in equivalent prop o rtions.
`
`48
`
`IMMUNOGEN 2327, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`1
`TABLE
`in Experimental Studies . 1
`Details of MoAbs used
`
`Antibody
`
`Reference
`
`Immunoglobulin
`Subclass
`
`Spec if icity
`
`anti-Ly-2 .1
`250 - 30.6
`I - 1
`JGT
`
`18
`19
`20
`20
`
`IgG2a
`IgG2b
`IgGl
`IgGl
`
`murine Ly-2.1
`Colon cancer
`Carcinoembryonic antigen (CEA )
`CEA
`
`All antibodies were purif i ed using Protein-A Sepharose.
`
`Preparation of N AcMEL-MoAb Coniugates
`N- AcMEL was co upled to various MoAbs by a previously described procedure
`(24). For clinical studies , the final preparation after drug conjugation was
`batch te s ted for pyrogens and sterility (Pharmacology Department, University of
`Melbourne; and Sigma Pharmaceuticals , Clayton , Victoria, Australia) .
`
`Drug Activity
`Cytotoxicity of the free drug and immunoconjugates wa s determined by
`measuring the inhibition of DNA synthesis using [ 3H)-thymidine as previously
`described (24 ).
`
`In Viyo Experiments
`E3
`(Ly- 2 .1 +ve ) tumor cells were injected subcutaneously into the abdominal
`wa ll o f B6CF1 (Ly-2.1 -ve , Ly-2 .2+ve ) mice and allowed to d e velop into a palpable
`lump (0. 1 -0 .6cm2) depending on the experiment before commencing treatment.
`Conjugates were admini stered intravenously and the tumor size measured daily
`with a caliper square , measuring along two p erpendicular axes of the tumor, and
`the data recorded as mean tumor size (product of two diameters ± standard error
`o f the mean) . Experimental groups of 10-20 mice, all of the same sex and age ,
`were used in each experiment . The details of the preclinical studies of
`antibodies used in the clinica l studi es are provided elsewhere (25,26).
`
`Recombinant Human-TNF - a
`!r-IFN -y l
`
`Cr-TNF - al
`
`and Recombinant Mouse-IFN y
`
`Both human r-TNF-a (6xl0 7 U/mg, purity >99%, ~l .Ong endotoxin/mg ) and mouse
`r -I FN-y (l-2xl0 7 U/mg) were generously s upplied by Boehringer Ingelheim; these
`were diluted in PBS and s tored in single dose aliquots.
`
`Biodistribution
`B6CF1 mice bearing subcutaneou s E3 tumors (0. 2 - 0.Scm2) were used to compare
`th e di s tribution of 125 r - anti - Ly-2.1 in the presenc e or absence of therapeutic
`levels of r-TNF-a (S~g) . Groups of 4 mice were sacrificed 24hrs after the
`injection of labeled anti-Ly-2.1 a nd the biodistribut ion of 125!-anti-Ly-2.1 was
`determined by counting the radioactivity of blood, heart, splee n, liver, kidneys
`and tumo r. The distribution of isotope is reported as the % injected dose/gm
`tissue. All mice received l.Sxl0 6 cpm of 125 r activity ( 60 ~g protein) by tail
`vein injection 2 4 hours after the intraperitoneal (i.p.) administration of r (cid:173)
`TNF-a.
`
`Clinical Studies
`Patients
`Seven patients with disseminated colorectal carcinoma and one
`p a tient who previously had a curative colonic resection (Duke C) were entered
`into the study . All (except patient 6) had previous resection of the primary
`colorecta l carcinoma and all (except patient 8) presented with progressive
`metastatic disease: Two patients had pulmonary and hepatic metastases and five
`patients h ad extensive h epatic metastases. Two patients previously had failed
`intens ive chemotherapy . Table 2 summarises the characteristic and clinical
`feature s of the patients . The ages ranged from 38 years to 62 y ears. All
`
`49
`
`IMMUNOGEN 2327, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`patients (except patient 8 ) had measurable d i sease (hepatic metastases), a
`performance status (ECOG) of at least 3 , and had not received anti - cancer
`therapy for at least 4 weeks and no such therapy was administered during
`evaluation (at l east 9 weeks after therapy ) . Thi s Phase I study was approved by
`the Medical Research Board of the Roya l Melbourne Hospital and written informed
`consent was obtained from every patient.
`
`TABLE
`2
`Characteristics and Clinical Features of Patients
`
`Patient No.
`
`Age
`
`Disease
`
`Performance
`Status (ECOG )
`
`Previous
`Treatment
`
`1
`2
`
`3
`4
`
`5
`6
`7
`
`8
`
`59
`57
`
`61
`58
`
`62
`46
`62
`
`38
`
`Hepatic metastases
`Hepatic and Pulmonary
`meta stases
`Hepatic metastases
`Hepatic and Pulmonary
`metastases
`Hepatic metastases
`Hepatic metastases
`Hepatic metastases
`and unresected primary
`colon carcinoma
`
`2
`2
`
`2
`2
`
`3
`3
`3
`
`1
`
`a HAl
`
`Hepatic Artery Infusion
`
`HAi a with Cisplatinum
`
`HAl with Cisplatinum
`
`Administration of N- AcMEL-MoAb Conjugates
`Using the Seldi nger technique , an external arterial catheter was
`introduced percutaneously into the left axil lary or high brac hial artery . The
`catheter was placed into the common hepatic artery under fluoroscopic guidance.
`The N-AcMEL-MoAb conjugates were administered by oxyrnetric pump through the
`hepatic artery cannula in lOOml of normal sa line over 2 h o urs per day for 2
`days. All patients had 3 doses of the irnrnunoconjugates (t= O,
`t=24hrs , t=48hrs)
`In between infusion of the irnrnunoconjugates, the patency of the hepatic artery
`catheter wa s accomplished by heparinised saline (5000 I . U. aqueous heparin in 1
`litre normal sa line at the rate of 50ml/hour ) . The arterial catheter was
`removed at the e nd of the 2 - day infusion. A dose escalation protoco l of 5mg/m2,
`10mg/m2, 15mg/m2, 20mg/m2 N-AcMEL conjugated to MoAb was used. One patient
`r ece ived 5mg/m2, and two month s later, 10mg/m2,
`two received 10mg/m2,
`t wo
`r ece ived 15mg/m2 and three r eceived 20mg/m2 N- AcMEL conjugated to MoAbs . The
`s tudy wa s closed at the 20mg/m2 d ose of N-AcMEL conj ugated because of the
`f inanci a l cos t involved.
`All patients had dexamet hasone 8mg admi ni stered intravenous ly just before
`ea ch infusion of i rnmunoconjugates and oral prednisolone lOmg dai ly for 7 days
`after completion o f therapy to diminish hypersensitivity phenomena .
`
`~aluation of Responses
`Patients were monitored clinically including noting changes in
`temperature , pulse rate, blood pressure and respiratory function during and
`after the infusion . Blood studies ( f ull blood count, serum tests of renal and
`liver function and human a nti -mouse antibody r espon se ) were performed before,
`during and wee kly for 9 weeks after the therapy to assess potential toxicity.
`The tumor re s ponses were evaluated by weekly physical examination s , weekly blood
`tests (liver function test , carcinoernbryonic antigen l eve l) and computed
`tomography (CT ) sca ns of the abdomen , performed before therapy , 1 month and 2
`months after therapy. Complete response (CR ) is defined as a reduction of at
`lea st 50% in the sum o f the products of the two greatest diameters of measured
`lesions ; minor response (MR) is a reduction o f greater than 25 % but less than
`5 0 % in the s ize of measurable l es i o n s in the absence of progression or
`occurrence of new lesions elsewhere; stable disease (SD ) is no objective change
`of all meas urable tumors; and progressive disease (PD) is the appearance of new
`lesions or increase in size of one or more measurable lesions by at least 25 %.
`
`Human IgM and IgG antibodies to the murine MoAbs were
`Serological Tests
`measured by a n indirect enzyme linked irnmunosorbent assay (EL ISA ) . Ninety- six
`
`50
`
`IMMUNOGEN 2327, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`well flexible polyvinyl ch l oride plates (Costar , Cambridge , MA ) were coated with
`lO J.l.g/ml of purified 30 . 6 ,
`I - 1 or JGT MoAb in a carbonate buffer , pH 9. 6 ,
`(50J.l.l
`per well ) and no n- spec i fic binding blocked wi th 1 % bovine serum albumi n/PBS pH
`7 . 6 for lhr at 3 7°C. Doubl ing dilutio n s of patien ts ' sera a nd poo l ed normal
`human serum (Red Cross Bl ood Bank, Melbourne , Austra l ia )
`i n PBS/ 0 .05% Tween 20
`to a final di l ution of 1/256 were performed. After washing the coated and
`blocked plate with phosphate buffered saline (PBS ) /0.05 % Tween 20 , di l uted serum
`was added to the coated wells and left in overnight incubation at 4°C.
`Pl ates
`(50J.l.l/well ) were then washed with PBS/0.05% Tween 2 0 and then reacted with a
`1:400 dilution (50 J.l.l/ well ) of p h osphatase labe l ed affinity purif i ed goat anti (cid:173)
`(Ki r kegaard & Perry , Mary land) at 37°C for 3 hours . The
`human IgM and IgG
`plates were t he n was hed exten sively wi th PBS /0 . 05 % Tween 20 , and 5 0J.l.l/wel l of
`alkaline p hosph atase s u bstrate added . The co l o ur reaction was read wit h a
`Dynatech ELISA p l ate reader at a wave l e ngth o f 405 nm. Resu lt s we r e expressed as
`the absorbance value of patient serum compared with poo l ed normal huma n serum
`both at a dilution of 1: 256 .
`
`RESULTS
`
`Preclinical Studies
`Me l phalan was coupled to MoAbs using
`to MoAb
`Conjugation of N-AcMEL
`the non-toxic derivative N-AcMEL . The active ester of N- AcMEL was reacted with
`the monoclo n a l antibody as described previou sly
`(2 4)
`(Fig. 1 ). About 15 - 20
`residues of MEL was coupled with good protein recovery and retention of
`satisfactory antibody activity .
`
`Cl
`
`Cl
`
`~N -o-~ j
`r
`-
`Melphelen
`
`0
`II
`/ C- OH
`CH£!CH
`
`Aceti c
`
`Anhydri de
`
`0
`II
`C-O H
`
`.
`/
`-+- CH
`l
`"""
`II
`;
`NH-C- CH3
`N- AcME L
`
`0
`
`"""
`
`NH 2
`
`i
`;
`
`0
`II
`/ C- NH -Ab
`.
`-+- CH
`l
`i
`
`0
`"""
`II
`NH - C- CH3
`
`N- AcMEL-Mo Ab
`Conjugete
`
`lc er bo di imid e
`
`N- Hy dr ox ysuc cin (cid:173)
`imid e
`
`0
`
`0~
`
`i / ~-o-N:J
`-+- CH
`0
`-7
`0
`i
`"""
`II
`i
`NH - C- CH3
`
`FIGURE 1
`Coupling of N-AcMEL
`to Mono clonal Antibody.
`
`The activity of free and bound drug
`In Vi tr o Ac tiv i ty o f N-AcMEL-Mo Ab
`wa s determined by measuring the inhibition of DNA synthe s is of target cell
`line s . N-AcMEL is considerably less toxic than MEL (30-100 fold) and when
`c onjugated to MoAb the resulting immunoconjugate is more toxic than N- AcMEL
`(24 ) . These conjugates react only with antigen positive cells: Figure 2 s hows
`the cytotoxicity of N-AcMEL- anti - Ly-2 . 1 conjugate on E3 (Ly-2.1 +ve ) and BW5147
`(Ly- 2.1 - ve ) cells .
`It was noted that conjugation of N- AcMEL to MoAb did not
`affect the alkylating activity of ME L (demonstrated by the Epstein method (27 ))
`
`51
`
`IMMUNOGEN 2327, pg. 6
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`and the conjugates were stable at 4°C overnight but rapidly lose activity at
`room temperature overnight.
`
`100
`
`80
`
`w z
`i5
`:E
`>
`X z
`1-;" 0
`:r i= 60
`!2 oCt
`u. a:
`0 0
`0..
`z a:
`0 0
`i= 0
`iii ~
`i:
`~
`<fl.
`
`40
`
`20
`
`0
`
`-5.30 1
`
`-5.000
`
`-4 .301
`
`-4.000
`
`LOG CONCENTRATION NAcMEL (M)
`
`2
`FIGURE
`Inhibition of DNA Synthesis {measured as % inhibition of 3H-thymidine uptake) b~
`Ly- 2+ Antigen Positive {E3) and Ly-2- Antigen Negative {BW5147) Cells. Cells
`were incubated with drug or immunoconjugate for 30 mins, wa s h ed , and after 24
`hrs pulsed with [ 3H] - thymidine for 6 hrs , harvested and counted. N- AcMEL {[] a~ct
`N- AcMEL- anti - Ly-2 . 1 (.) on E3 Ly-2.1+ cells; N- AcMEL {~) and N-AcMEL-anti-Ly-2 . 1
`{00) on BW5147 Ly- 2.1- cells.
`
`Combined Effect of R-TNF-a and N- AcMEL-anti-Ly - 2.1 on the E3
`Thymoma.
`N-AcMEL- anti - Ly-2.1 conjugates have been s hown to be a r e asonably
`effective anti-tumor agent against the murine thymoma E3 {24) . R- TNF - a is toxiQ
`and therefore an initial experiment was performed to determine the most
`effective and non-toxic dose regimen; a total dose of 24~g N- AcMEL conjugated to
`Ly-2.1 wa s chosen for this study. R-TNF - a was admin i stered in 3 doses; a total
`of 3 (3xl), 15 (3x5) and 30 (3xl0) ~g was used i n conjunction with the N- AcMEL(cid:173)
`anti - Ly- 2.1 conjugate. The tumor growth and toxicity {by histopathological
`examination of major organs) was monitored {Figure 3).
`Several important points were apparent. At a total dose of 30~g r - TNF - a, 20% of
`the mice died of toxicity and ,in 10 % the tumor was totally eradicated. At lower
`doses of r - TNF - a (3~g) there were no cures and on day 20 the tumor was 66% of
`the con trol mean tumor size. The combination of N-AcMEL-anti - Ly- 2.1 with
`different doses of r-TNF - a gave a synergistic effect where 30, 50 and 40 % of
`tumors regressed, with doses of 3 , 15 and 30~g r-TNF-a respectively .
`It wa s
`also apparent that MEL and anti-Ly-2.1 was more effective when TNF was also
`administered. A similar experiment was performed to ascertain the effect of two
`different doses of N-AcMEL-anti-Ly-2.1 conjugate dose together with a single
`therapeutic dose of r - TNF - a (5 - lO~g) . Two groups of mice were given {3x4~g) and
`( 3x8 ~g) N-AcMEL- anti- Ly- 2.1, when no tumors regres sed , but on day 20 the tumors
`were 45 % a nd 26% of the control tumor size (Fig. 4). Where (5x5~g) r-TNF-a was
`also added to the treatment regime 25 % and 35% of the tumors completely
`regressed with the two different doses of the immunoconjugate. Furthermore, on
`day 20, the tumors were 17% and 12% of the tumors found in control mice treated
`with PBS. R-IFN-y is capable of enhancing the activity of r-TNF-a and therefore
`
`52
`
`IMMUNOGEN 2327, pg. 7
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`2 :·
`
`20
`
`1 0
`
`.,.- 1 G
`::
`!,!. 1 4
`l!j
`v;
`a:
`0
`
`1 2
`
`~ 1 0
`z 00
`:3
`" 0 6
`
`1110
`
`1110
`
`\110
`
`04
`
`02
`
`00
`
`22
`
`2 0
`
`1 0
`
`.,.- 1 6
`"' !,!. 1.4
`w
`N v; 1 2
`a:
`0
`~ 1 0
`....
`z 00
`:3
`"' OG
`
`04
`
`02
`
`00
`
`0
`
`10
`
`1 2
`
`14
`
`1 6
`
`10
`
`20
`
`22
`
`24
`
`26
`
`28
`
`30
`
`2110
`
`3110
`
`51 10
`
`·11 10
`
`10
`1G
`14
`1 2
`10
`0
`OA YS AFTER TUMOR INOCULATI ON
`
`20
`
`22
`
`2 4
`
`2G
`
`28
`
`30
`
`3
`FIGURE
`Inh ibition of tumor growth by r - TNF - a and N- AcMEL- anti - Ly- 2 .1 . Growth of the E3
`(i . v .) or TNF (i.p.)
`tumor in B6CF1 mice treated on day s 6 , 9 ,1 4 wi th N- AcMEL- MoAb
`on days 7 , 9 and 1 3. For c l arity two graphs of t h e same experiment is shown .
`a ) PBS (~); r - TNF-a, 3 ~g (4) ; 1 5 ~g (Ill, 30 ~g ( ~ ); a mi xture of MEL and
`anti-Ly- 2 . 1 wit h
`( []) or without 30 ~g r-TNF -a (0 ) ; and
`(+ ) ; 15~g r(cid:173)
`b) N- AcMEL-anti - Ly-2.1
`( 24 ~g )w ithout r- TNF- a
`(x ) ; 5~g r - TNF - a
`TNF-a (•) and 30 ~g r-TNF - a
`(83 ) . Error bars represent standard error of th e
`mean and the numbers represent t h e n umber of complete regressions .
`
`(28 ) . There was only a marginal
`
`this was also i ncluded i n the treatment r egime
`increase in t herapeutic efficacy.
`From the above results and those published elsewhere (24 ) it can be seen
`that N-AcMEL- anti-Ly- 2. 1 i s reasonably effective in eradicating s mall tumors
`(<0.2cm2). Therefore the effect of r-TNF - a on the efficacy of N-AcMEl- anti-Ly-
`2 . 1 o n larger tumo rs (0 . 4- 0.6cm2l was i nvestigated . A total dose of 26~g
`on day
`N- AcMEL-an ti - Ly- 2. 1 and 45 ~g r-TNF -a was c h osen for this st u d y
`(Fig . 5 ) .
`20 the groups of mice that had rece i ved only r - TNF - a or N-Ac MEL-anti-Ly-2 .1 had
`tumors 50% of the size of that of t h e PBS treated gro up wh ile no regressions
`were seen. By contrast , o n day 22 , the g r oups of mice that received both
`treatments had tumor s that were 15% of the size of that of t he PBS treated group
`and 10 % of the tumors were totally eradicated.
`As the results show that r-TNF- a does enhance the anti - tumor e ff ect of N(cid:173)
`AcMEL-anti - Ly- 2.1 a possible mechanism for s ynergy wa s sought .
`
`The
`and Immunoconjugate.
`Mechanism of the Synergy Between r-TNF- a
`effect of r - TNF - a o n the distr i b u tion of radi olabe l ed anti- Ly-2 .1 was examined
`to a sce rtain whether an increased up take of MoAb resu l ted d u e to concurrent
`\
`
`53
`
`IMMUNOGEN 2327, pg. 8
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`2.2
`
`2.0
`
`1.8
`
`1.6
`
`N~ 1.4
`::E
`£
`w 1.2
`N
`iii
`a: 1.0
`0
`::E
`::l 0.8
`t-
`:z
`<t 0.6
`w
`::E
`
`0.4
`
`0.2
`
`0.0
`
`1/10
`
`1110
`
`5!20
`
`3110
`
`7120
`
`0
`
`2
`
`4
`
`6
`
`18
`16
`14
`12
`10
`8
`DAYS AFTER TUMOR INOCULATION
`
`20
`
`2 2
`
`24
`
`26
`
`28
`
`30
`
`4
`FIGURE
`Inhibition of E3 thymomas by r-TNF-a, r-IFN-y and N-AcMEL-anti-Ly-2 . 1 conjugates.
`N-AcMEL-MoAB (i.v) wa s given on days 6,8,10; TNF and/or yiFN (Ig) on days
`7,10,13 , 1 6 .
`PBS (x); r-TNF-a (t); r-TNF-a and r-IFN - y (+); N-AcMEL-anti-Ly-2.1
`; N-AcMEL-anti-Ly- 2.1 ( 12).lg) with T-TNF-aand y-IFN-1 (M
`(1 2).lg)
`(0) and 24).lg
`•
`and TNF-a (j; and N-AcMEl-anti-Ly-2.1
`(24).lg) and r -TNF-a (£). The error bars
`denote the s tandard error of the mean a nd the numbers are the number of complete
`regres s i o ns/ tot a l number of mice.
`120
`
`100
`
`w
`!:::! 80
`(/)
`a:
`0
`::E
`::l 60
`t-
`...J
`0 a:
`t-:z
`0 u
`~
`
`40
`
`20
`
`1/10
`
`0
`
`2
`
`4
`
`6
`
`8
`
`10
`
`12
`
`14
`
`16
`
`18
`
`20
`
`22
`
`24
`
`26
`
`28
`
`30
`
`DAYS AFTER TUMOR INOCULATION
`
`5
`FIGURE
`Ef f ect of r-TNF-a and N-AcMEL-anti-Ly-2.1 on E3 tumor (0.4-0.6cm2). Mice
`rece ive d N-Ac MEL-MoAB
`(IV) on days 8 , 10 ,14,1 5 ;
`(TNF)
`(I P) on days 9,11,14,16 and
`2 1 . The tumor gro wth c urve s as the % control (PBS ) size for each group is shown
`-PBS (t); N-Ac MEL-anti - Ly - 2.1 ~; a mixture of N-AcMEL-anti-Ly-2.1 and r -TNF -a
`(4) ; a nd r-TNF-a C). The s tandard error for each group did not exceed 5 % for
`eac h po int . The number repre sent s the total number of complete regressions/
`tota l numbe r of mi c e.
`
`IMMUNOGEN 2327, pg. 9
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`20
`
`::E
`
`(!) > 1->
`
`t=
`(.)
`<t
`0 10
`15
`<t
`a:
`;f!.
`
`0
`
`BL
`
`TU
`
`L1
`
`TISSUE
`
`SP
`
`1-E
`
`Kl
`
`6
`FIGURE
`Effect of r -TNF-a on the biodistribution of 125 1 - anti-Ly- 2 .1. The percentage
`injected radioactivity dose/gm is shown for the various tissues- blood (BL );
`tumor (TU); liver (LI ); spleen (SP); heart (HE ) and kidney (KI) without filii and
`with ~ the administration of r-TNF - U
`
`inflammation. Administration of r-TNF-a 24hrs earlier than 125 1 - anti - Ly-2.1 to
`E3 tumor bearing mice, a nd examin ation of the tissues for radiolabel 24hrs later
`showed an improvement in tumor localization. Administration of r - TNF-a
`increased the % dose/gm in the tumor from 14.87±0.90% to 1 9 . 30±0.68 % (Fig. 6)
`and an increase of the tumor/blood ratio from 0.810.03 to 1.63±0.02. R- TNF-a
`also has a direct cytotoxic or cytostatic effect on some cell lines (29 ) which
`is not necessarily related to the presence of high or low numbers of receptors
`scatchard a nalys i s of E3 cells binding 1251-r-TNF-a
`on the surface (30 ) .
`indicated the presence of 5000±20 0 TNF recept ors /c e ll with a Ko = 1.2 6±0 .0 5nm
`(not shown ) .
`I n vitro cytotoxicity of r - TNF-~r-IFN-y and the mixture on E3
`cells was examined using a cell viability assay, a nd 3H-thymidine uptake assay
`(Fig . 7). Both r -TNF -U and r -IFN - y decreased the viability >70 % at doses of
`lOOU/ml , but no s ynergistic effect was apparent in vitro.
`
`Clinical Studies
`
`The results of
`Immunohistochemical Testing of Biopsy Specimens
`immunoperoxidase sta ining, perfo rmed o n sections of biopsy specimens of primary
`colon cancer , using a panel of MoAbs, are summarized in Table 3 . There was a
`wide range of staining grades with different MoAbs although all MoAbs
`demonstrated some degree of stain ing. To overcome the problem of tumor cell
`heterogeneity and to ensure maxima l
`immunoreactivity, a combination of at least
`two MoAbs was used in 5 /9 treatments -
`the selection of MoAb was individualized
`for each patient; except patient 5 a nd 6, a lthough the colon tumor was reactive
`with both I -1 a nd JGT MoAbs
`(staining grade 3 ), they were treated with 250-30.6
`MoAb because of a shortage of I - 1 a nd JGT MoAbs .
`Such a decision was empirical
`as there was no fresh co l o n cancer tissue available for immunohistochemical
`testing , but based on our previous study , 250-30.6 MoAb reacted with >90 % colon
`cancer tissue (1 9 ) .
`
`55
`
`IMMUNOGEN 2327, pg. 10
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`80
`
`60
`
`40
`
`20
`
`> ...
`::J
`iii
`~ >
`..J
`..J w
`()
`u.
`0
`z
`0
`i=
`()
`:::>
`0 w
`a:
`:.!! 0
`
`10
`
`1 00
`
`1000
`
`DOSAGE (Uimf)
`
`7
`FIGURE
`Cytotoxicity of r - TNF - a , r-IFN-y on E3 cel l s . Cells were cu l tured with r -TNF - a
`II '
`r - IFN-y 0 or a mixture of both Et for 72hr at 37°C and cell viabi l ity
`determined by uptake of Trypan blue .
`
`Immunop e roxi dase
`
`3
`TABLE
`s t aining of Prima ry Colon Cance r with Different
`Monoclonal Ant i bodies
`
`Patient
`
`Staining Score of MoAbsa
`250-30.6b
`I -1
`JGT
`
`1
`2
`3
`4
`5
`6
`7
`8
`
`3
`
`3
`3
`
`4
`4
`4
`4
`3
`3
`4
`4
`
`3
`2
`4
`4
`3
`3
`4
`4
`
`b
`
`Staining s core expressed as 0-4 according to whether nil , up to 25 %, 25 - 50% ,
`51-75% or more than 75% of carcinoma ce l ls stain ed, othe rw ise not t ested.
`250-30.6 MoAb tested on fresh colon cancer tiss ue.
`I - 1 and JGT tested on
`formalin fixed or fresh colon cancer tissue .
`
`Tox ic i ty Table 4 summarizes the results of the patients treated with hepatic
`artery infusion of N-AcME L- MoAb conjugates .
`I n ge ne r a l, the therapy was well
`tolerated with no disturbance in hematological , gastrointestina l, renal or
`cardiac parame ters ; the externa l arterial cathete r was a l so well tolerated with
`n o complications. Patient 1 h ad 2 courses of N- AcMEL- MoAb con j u gates two months
`apart , despite the development of a high titre of human a n ti - mouse antibody
`a~ter the first course: he deve l oped moderate ly seve r e back pain , fever (39°C ),
`r~gors , urticaria and bronchospasm duri ng the second course of treatment ,
`de s pite prior administration of dexamethasone . He was treated with
`antihistamine and an additional dose of dexamethasone with rapid control of the
`reactions .
`In three other patients (patients 6 , 7 & 8 ), fever of 38°C- 39 . 0°C
`
`56
`
`IMMUNOGEN 2327, pg. 11
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`immunoconjugates;
`was noted during the second or third day of infusion o f
`febrile reactions we r e more common i n patients given higer doses of
`immunoconjugates.
`Patient 1 also described symptoms (fever, polyarticular arthralgia)
`suggest ive of se rum sickness , deve l oping 2 weeks after cessation of the second
`course of therapy. No other patient s had delaye d s ymptoms that could be
`assoc iated with se rum s i ckness .
`Seven o f the eight patients mo unted a positive
`HAMA response, and the peak response usually occurred within 3 weeks after
`therapy. Patient 7 described worsening o f the pre-existent diarrhoea since the
`third week after treatment ; ho wever, he a l so h ad unresected advanc ed primary
`colon cance r caus ing subacute bowel obst ruction which might account for this .
`
`TABLE
`Results of Hepatic Artery Infusion
`
`4
`(HAI) of N-AcMEL-MoAb Conjugates
`
`Patient
`Number
`
`Dose
`N-AcMEL:MoAb
`
`Response
`Response
`by CT scana Duration
`(mont hs)
`
`Serum CEA level (ng/ml) b
`Before
`4 wks after
`Treatment
`treatment
`
`3
`
`2
`
`7
`
`3
`
`5
`
`1
`
`2
`
`3
`
`4
`
`5
`
`6
`
`7
`
`8
`
`5mg/m2 :120mg/m2
`(I - 1)
`10mg/m2:16 0mg/m2
`(I-1 + JGT)
`10mg/m2 :250mg/m2
`(I -1 )
`10mg/m2 :980 mg /m2
`(I-1 + JGT)
`15mgfm2 : 340mg/m2
`(30.6 + I - 1 + JGT)
`15mgfm2:380mg/m2
`(30.6 )
`20mgfm2:440mg/m2
`(30.6)
`20mg/m2 :1000mg/m2
`(30.6 + I-1 + JGT)
`20mg.m2:820mg/m2
`(I-1 + J GT)
`
`MR
`
`SD
`
`MR
`
`SD
`
`SD
`
`PD
`
`PD
`
`Died before
`CT scan
`
`3920
`
`1160
`
`26
`
`460
`
`2 70
`
`5425
`
`62
`
`325
`
`<1
`
`920
`
`10 50
`
`15
`
`405
`
`35
`
`4300
`
`144
`
`42
`
`<1
`
`a MR = minimal response, SD = stable disease, PD
`b Upper limit of normal is 5ng/ml
`
`progressive disease
`
`Response Table 4 s ummari ses the tumor respon ses as evaluated by CT scan.
`Minor responses we r e seen in 2 patients (patients 1 & 2 ) a nd a ll patients except
`patien t 6 are alive with disease at the time of wri ting (mean duration since
`the rapy = s i x months, mean duration sin ce diagnosis of hepatic metastases =
`twelve months ) .
`Patient 1 had subtotal co l ectomy 24 months previously for synchronous
`carcinoma of the transverse and sigmoid co l on . At the time of laparotomy, it
`was n oted that he had multiple large metasta ses in the liver . The hepatic
`metastases were too exten sive for s urgica l
`r esection a nd he was treated with 2
`courses o f hepatic artery infusion of Cisp l atinum, with no response . He
`continued to deteriorate over the fol l o wing 10 months , with anorexia , nausea ,
`lethargy and severe hepatic pain. He had resigned from his job as the
`headmaster of a high schoo l for terminal care when he was entered into the
`study. The liver function test was grossly deranged with elevated transamina se
`(AST = 136IU/l) a nd CEA ( 3920 ng/ml) levels. The abdomi na l CT scan revealed
`multiple large metastases in both l obes occupying about 60 % of t he liver.
`Within 3 weeks after treatment with N-Ac MEL-MoAb con jugates , he made a dramatic
`improvement in hi s con stit u t i on, regained his appetite and had good r e lief of
`his hepatic pain. He felt so we ll t hat he r et urned to work . There was an
`improvement in hi s liver function test (AST- 60 IU/1 ), a fa ll in CEA l evel (CEA
`= 920ng/ml) and an improvement in abdominal CT scan with a reduction in the size
`of the liver and the metastases (2 5% ) and , of the remaining metastases , over
`half the size of the lesions were ca l cified. A repeat abdominal CT scan
`
`57
`
`IMMUNOGEN 2327, pg. 12
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`peformed 5 months after treatment showed that the hepatic metastases had not
`progressed since treatment. Patient 1 is still alive , with disease , 12 months
`after therapy.
`Patient 2 had right hemicolectomy for carcinoma of caecum originally ,
`followed by subtotal colectomy for a metachronous carcinoma of the transverse
`colon 2 years later . At the second presentation he was found to have multiple
`hepatic and pulmonary metastases and had symptoms referable to the hepatic
`metastases which were too extensive for surgical resection; he was therefore
`trea