'Vol. 4 no. 2, 1989
`
`CONTENTS
`
`Analysis of the human immune repertoire using human monoclonal antibodies ..............
`R. J. Cote
`
`Individually specified drug immunoconjugates in cancer. treatment ..................................
`R. K. Oldham, M. Lewis, D. W. Orr, S—K. Liao, J.R. Ogden, W.H. Hubbard, R. Birch
`
`Distribution of CAT125 in placental tissues ..........................................................................
`LC. Fuith, E. MI'i/Ier-Ho/zner, C. 'Marth, E. Perkmann, A. Zeimet, G. Daxenbichler
`
`Clinical behaviour of prostatic specific antigen and prostatic acid phosphatase: a compa-
`rative study ...........................................................................................................................
`J. Morote FI’ob/es, A. Huibal More/l, J.A. De Torres Mateos, A. Soler Rose/lo
`
`Correlation of an estrogen receptor-related phosphoprotein with histopathological featu-
`res in breast cancer .............................................................................................................
`
`Ft. M. Tomasino, E. Daniele, R. Nuara, V. Morel/o, M. Salvato, A.M. Florena
`
`Preoperative and logitudinal serum levels of CA 125 and CA 15.3 in patients with breast
`cancer ...................................................................................................................................
`Y. 7'. Omar, A.E. Behbehani, N. AI-Naqeeb, M. M. Mota wy, M. 0. Foudeh, A.H. Awwad, M. Y.
`Nasralla, J.J. Szymendera
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`Serum levels of soluble interleukin-2 receptors and their relation to CD4ICD8 ratio in
`patients with advanced solid tumor (Letter to the Editor) ...................................................
`
`
`
`G. Tancini, P. Lissoni, S. Barni, Fl. Rescaldani, F. Rove/Ii, A. Sormani, S. Crispino,
`T. Saporito, E. Tisi, C. Benenti
`
`
`
`CA 72.4 serum levels in 298 patients with non-malignant diseases (Letter to the Editor)..
`125
`E. Sagredo, A. Soriano, L. Pombo, M.C. Roiz, M. T. Allende, A. Ruibal
`
`
` 127
`Announcements ...................................................................................................................
`
`
`
`
`
` THE INTERNATIONAL JOURNAL
`
`
`
`OF BIOLOGICAL MARKERS
`
`
`
`59
`
`65
`
`78
`
`81
`
`88
`
`95
`
`Diagnosis of bone and liver metastases in breast cancer comparing tumor markers and
`imaging techniques ..............................................................................................................
`P. Zanco, G. Rota, V. Sportiello, N. Borsato, G. Fer/in
`
`103
`
`A brief guide to the practice of radioimmunoscintigraphy and radioimmunotherapy in
`cancer
`
`Report of the European Association of Nuclear Medicine Task Group ..............................
`K.E. Britton, G.L. Buraggi, Fl. Bares, A. Bischof-Dela/oye, U. Buell, D. Emrich,
`M. Granowska
`'
`
`106
`
`Tissue polypeptide antigen (TPA) in bronchoalveolar lavage (Letter to the Editor) ..........
`R. Blasco, E. M. Barbero, R. MufiozCa/vo, F. Gomez de Terreros
`
`119
`
`Do mast cells and their products play a role in AIDS-related angiogenesis and immune
`dysfunction? (Letter to the Editor) ........................................................................................
`R. Manconi
`
`, 121
`
`123
`
`IMMUNOGEN 2305, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 1
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`© Wichtig Editore 1989
`The International Journal of Biological Markers / Vol. 4 no. 2, pp. 65-77
`M
`
`
`
`Individually specified drug immunoconjugates in
`cancer treatment
`‘
`
`R.K. OLDHAM, M. LEWIS, D. W. ORR, S-K. LIAO, J.R. OGDEN, W.H. HUBBARD, R. BIRCH
`
`Williamson Medical Center Franklin, TN—USA
`
`
`
`ABSTRACT: Forty-three patients with disseminated refractory malignancies each received an
`individually-specified combination ofeither Adriamycin (24 patients) or mitomycin—C ( I9 patients)
`conjugated murine monoclonal antibodies. Tumors were typed using a panel ofantibodies with both
`immunohistochemistry andflow cytometry. Cocktails ofup to six antibodies were selected based on
`binding greater than 80% of the malignant cells in the biopsy specimen. These monoclonal antibody
`cocktails were drug conjugated and administered intravenously.
`Seventeen out of twenty-four patients had reactions to the administration of Adriamycin im-
`munoconjugates, but these were tolerable in all but two patients. Fever, chills, pruritis and skin rash
`were by far the most common transitory reactions. All were well controlled with premedication. In
`several patients it was demonstrated that there was limited antigenic drift among various biopsies
`within the same patient over time. Up to 1 gram of Adriamycin and up to 5 grams of monoclonal
`antibody were administered. The limiting factor appeared to be a variable dissociation of active
`Adriamycin from the antibody, which unpredictably caused hemopoietic depression.
`Similar findings were noted in 19 patients with mitomycin-C conjugates. Thrombocytopenia at a
`60mg dose of mitomycin-C in this schedule Was dose limiting. Preliminary serological evidence
`suggests that the development of an IgM antibody 'which is specific against the mouse monoclonal
`antibody has the specificity and sensitivity to predict clinical reactions. These antibodies were
`quantitatively less in mitomycin—C patients.
`Selectedpatients were re-treated. One patient with chronic lymphocytic leukemia had re-treatment
`on three occasions and demonstrated regression ofperipheral lymph nodes. Two patients with breast
`carcinoma had definite improvement
`in ulcerating skin lesions and two patients with tongue
`carcinoma had shrinkage of their lesions. No responses were seen with mitomycin-C conjugates but
`binding was noted to tumors and colon with likely drug induced colitis seen after colon binding.
`This study demonstrates the feasibility and illustrates technical considerations in preparing drug
`immunoconjugate cocktails for patients with refractory malignancies.
`Cocktailformulation and antibody delivery was accomplished. The major technical hurdle appears to
`be the selection of effective conjugation methods that can be used to optimally bind drugs to
`monoclonal antibodiesfor targeted cancer therapy. (Int J Biol Markers 1989; 4: 65-77)
`
`KEY WORDS: Monoclonal antibodies, Drug immunocon/‘ugate cocktails, Targeted cancer therapy
`
`INTRODUCTION
`
`Since Kohler and Milstein ( 1) provided the semi-
`nal technology, there has been an explosion in the use
`of monoclonal antibodies in patients with malig-
`nancies. This paper describes our series of patient
`treated with combination monoclonal antibodies,
`specifically tailored for individual patients, combined
`with Adriamycin or mitomycin-C.
`The hypothesis that a combination of monoclonal
`
`antibodies would be necessary to cover virtually all
`cancer cells in a variety of sites and that each patient
`would require an individually specified immunocon-
`jugate dominated this research. Single monoclonal
`antibodies localize in areas of malignancy and to
`individual malignant cells (2, 3). However, it is well
`known that cancer cells have a variety of antigens
`which are not cancer specific. Antigens can vary
`within patients in clusters of tumor cells both by
`location and over time (microheterogeneity). Tumor
`
`65
`
`IMMUNOGEN 2305, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 2
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`
`
`Individually specified drug immunocanjugates in cancer treatment
`
`
`
`antigens may also vary during phases of tumor cell
`maturation. In addition, we have typed tumors from
`more than 150 patients and quantitative differences
`are the rule. No two have demonstrated precisely the
`same typing pattern (macroheterogeneity). Thus, an
`attempt has been made to identify combinations of
`antibodies which could potentially recognize up to
`100% of malignant cells within a variety of primary
`and metastatic sites. This was done by making a large
`number of monoclonal antibodies against freshly dis-
`persed cells, xenograft cells or cell lines recently de-
`rived from biopsies of tumors and then typing the
`individual patient’s tumor biopsies with these anti-
`bodies. Cocktails were specified to bind greater than
`80% of the cells within the malignancy. To that end,
`preparation of as many as six antibodies were admini—
`stered to patients following drug conjugation.
`These antibodies were usually greater than 95%
`pure, maintained immunoreactivity after conjugation
`and were tested for safety in a variety of systems prior
`to administration to patients. This paper demon-
`strates
`the feasibility of treating patients with
`mixtures of monoclonal
`immunoeonjugates and
`addresses technical considerations involved in the
`
`process. Observations on side effects, the retreatment
`of patients subsequently with similar or identical anti-
`bodies, the biological effects and tumor localization
`of the antibodies and the efficacy were recorded.
`
`TABLE 1 - DISEASE CATEGORIES
`
`MATERIALS AND METHODS
`
`Patient Selection
`
`females
`twenty-three
`and
`Twenty males
`participated in this trial. The most common cancers
`were breast (14), lung (3), colorectal (5), ovary (3),
`renal (2), and prostate (3).
`A variety of other cancer types were included as
`shown in Table 1. Selected characteristics of the
`
`treated patients are shown in Table 2.
`This clinical trial was carried out in Williamson
`
`Medical Center, after approval by the Investigational
`Review Board. Patients were referred primarily by
`oncologists after failure of standard modalities. Each
`patient was initially seen by a medical oncologist
`who reviewed the history and medical records, confirm—
`ed the failure of standard therapeutic options and,
`informed the patient of the experimental nature of the
`study. Each patient understood the strategy involved
`in this therapy and the other experimental therapeutic
`options available. After a determination of a suita—
`bility for the study and informed consent,
`tissue
`samples were obtained by biopsy. All typing was done
`on frozen tissue, either directly or on tissue which had
`been expanded by a xenograft in nude mice or by
`tissue culture propagation. Antibody selection was by
`immunoperoxidase and flow cytometry as described
`
`Tumor
`type
`
`Breast CA
`Carcinosarcoma
`Cholangiosarcoma
`Colo-rectum CA
`Hepatoma
`Islet Cell CA
`Leiomyosarcoma
`Lung CA
`Lymphoma (CLL)
`Ovarian
`Parotid
`Prostate CA
`Renal Cell CA
`Schwannoma
`Squamous Cell CA
`(tongue, mouth, penis)
`
`
`
`No. patients
`treated with
`Adriamycin
`immunoeonj ugates
`
`No. patients
`treated with
`Mitomycin—C
`immunoconjugates
`Total
`
`
`10“
`0
`l
`2
`0
`1
`1
`2
`l
`1
`1
`1
`1
`0
`3
`
`
`24
`
`4‘I
`l
`0
`3
`l
`0
`1
`1
`0
`2
`0
`2
`1
`I
`2
`
`14"l
`1
`l
`5
`1
`1
`2
`3
`1
`3
`1
`3
`2
`1
`4
`
`19 43
`
`
`
`’ Note that one patient with breast infiltrating ductal carcinoma first received Adriamycin immunoconjugate and six months later Mitomycin—C
`immunoconjugatc therapy. Thus, the total number ofdifferent patients with breast carcinoma is 13 instead of 14.
`
`66
`
`IMMUNOGEN 2305, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 3
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`
`
`Oldham et a]
`
`in detail elsewhere (4, 5). A minimal period of 45 days
`was necessary for tissue typing and preparation of
`sufficient quantities of immunoconjugate for treat-
`ment. A typical regimen consisted of three days to
`type with a panel of monoclonal antibodies, four
`weeks for production of sufficient quantities of the
`individual antibodies followed by conjugation of
`drug and extensive safety testing over a final three
`weeks. Thus, within three months, patients were seen
`and treated with an individually tailored combination
`of drug conjugate antibodies.
`Immunoconjugate preparations dissolved in
`normal saline were given over a period of one to five
`hours on a Monday, Wednesday and Friday. The
`total amount of immunoconjugate was given over a
`,two to three week period. An initial test dose based on
`5-10 mg of drug bound to antibody was given. The
`dose was then quickly escalated depending on the
`phase of the study. Toward the end of the investiga-
`tion, antibody amounts were escalated to try to give
`as much as one gram of Adriamycin and 60 mg of
`
`mitomycin C conjugated to 3-5 grams of antibody
`over a period of 2-3 weeks. A registered nurse was
`always available during administration and patients
`were pre-medicated with acetaminophen and diphen-
`hydramine for fevers, meperidine for rigors and
`epinephrine (in four patients) for significant allergic
`reactions.
`
`Antibody Selection and Preparation
`
`Immunization of mice and preparation of
`hybridomas are described elsewhere (4, 5). Over 100
`antibodies were available for tissue typing and we
`selected 28 for the standard panel. Seven of these were
`acquired elsewhere and 21 were produced in the bio-
`therapeutics’ laboratory. Five of these originated
`from immunization with breast cancers, eleven from
`melanomas, three from adenocarcinomas of the kid-
`ney, two from an islet cell carcinoma of the pancreas,
`and seven from colon carcinomas (4).
`The majority of the antibodies were IgGl with the
`
`TABLE 2 - MONOCLONAL ANTIBODIES USED IN THE PRESENT STUDYa
`
`Isotype
`
`Immunogen
`
`Antigen
`structure
`
`
`
`MoAb
`
`BA-Br—l
`HA—Br-Z
`BA-Br-3
`
`BT-Br—4
`BA-Br-S
`
`BT-Br—6
`BT-Co-lc
`
`BT-Co-Z
`BT-Co-3
`BT-Co-4
`BT-Co-Sc
`BT-Co-6
`BT-Me—3
`BT~Me—4
`
`BT-Me-5
`
`BT-Me-7
`BT-Me-8
`BA-Me-lO
`BA-Me—l l
`
`IgGl
`IgGl
`lgGl
`
`IgGl
`IgGl
`
`IgGI
`IgG3
`
`IgG3
`IgG3
`IgG3
`lgG3
`IgG3
`IgGl
`IgGl
`
`IgGZa
`
`IgGl
`IgGl
`IgGZ“
`IgGl
`
`Membrane extract of breast mrcinoma tissue
`Dispersed cells from breast carcinoma tissue
`Membrane extract of breast carcinoma
`cell line CAMA-l
`Dispersed cells from breast carcinoma tissue
`Membrane extract of breast carcinoma tissue
`
`L
`
`>
`
`.
`
`NDb
`ND
`> 300kD
`glycoprotein
`ND
`220kD — 400kD
`glycoprotein
`ND
`Dispersed cells from breast carcinoma tissue
`29kD + 31kD
`Dispersed cells from colon carcinoma grown
`protein
`as xenografts in nude mice
`ND
`”
`ND
`”
`ND
`”
`29kD + 31kD
`"
`ND
`7.
`ND
`Dispersed cells from melanoma tissue
`95kD—150kD
`Melanoma cell line
`glycoprotein
`CaCL 78-1
`p97—like (97kD)
`Melanoma cell line
`glycoprotein
`CaCL 78-1
`llOkD protein
`Melanoma cell line BUR
`llOkD + 40kD
`Melanoma cell line BUR
`280kD + > 400kD
`Melanoma cell line
`i
`” ”
`
`
`
`a All antibodies are referenced as to source in reference 5
`b Not yet defined, although attempts were made to determine the molecular mass of antigen involved
`L’ Based on epitopc blocking and indirect immunoprecipitation experiments. BT-Co~1 and BT—Co—S recognized different epitopes residing on the same or
`similar molecules.
`
`
`67
`
`IMMUNOGEN 2305, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 4
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`
`
`
`
`Individually specified drug immunoconjugates in cancer treatment
`
`exception of two Ing’s (melanoma) and five IgG3’s
`(colon carcinomas). Table 3 illustrates the character-
`istics of 19 antibodies from the panel used in this
`clinical study.
`The initial selection of antibodies was done by
`immunohistochemical phenotyping (6-9). This im-
`munoperoxidase based system is
`fully described
`elsewhere (5). The selection of antibodies was based
`on a grading system of l to 4+ which includedjudge-
`ments on the intensity of staining as well as the distri-
`bution of staining and the specific characteristics of
`the staining. The variation in grading between ob-
`servers was less than 10%. The selection by the same
`observer was reproducible over 90% of the time.
`Staining patterns varied from homogeneous staining
`of membrane and cytoplasm to patchwork staining of
`given tumor areas leaving adjacent tumor areas virtu-
`ally unstained, to scattered reactivity of tumor cells in
`a “sea” of non—reactive tumor cells (5).
`Selections of antibodies were made which encom-
`
`passed considerations of intensity, distribution, and
`patterns of staining. Positive controls included anti-
`HLA; negative controls included nonspecific random
`
`mouse immunoglobulins. All results were scored in-
`dependently by two scientists.
`Frozen tissue specimens (from fresh or cryopre-
`served primary or metastatic tumors) were assessed
`_ for binding of 19 murine monoclonal antibodies from
`our panel. One-micron thick sections of frozen tissue
`embedded in OCT compound (polyvinyl alcohol,
`benzalkonium chloride, polyethylene glycol, d.H20;
`Miles Labs, Elkhart, Indiana), were cut, placed on
`gelatin-coated slides and evaluated by means of the
`avidin-biotin—peroxidase complex technique (Vecta-
`stain ABC kit, Vector Laboratories, Burlingame,
`CA). Enzymatically or mechanically, disaggregated
`tumor cells (when available) were also subjected to
`flow-cytometric analysis to determine cell percentage
`and intensity of antibody binding.
`The use of the HLE-l antibody, which binds to
`leucocytes, permitted the correction for any white
`blood cell contamination. Propidium iodide exclu—
`sion staining excluded non-viable cells. Simultaneous
`antibody exposures‘provided comparisons of various
`antibody combinations to define possible additive or
`interfering interactions.
`
`IMMUNOHISTOCHEMICAL RESULTS OF METASTATIC MELANOMA LESIONS REMOVED AT VARIOUS TIMES FROM
`TABLE 3 -
`PATIENT BUR
`
`Occipital
`LN met
`01/23/86
`
`Tissue culture
`cells from
`LN met
`01/23/86
`
`Mediastinal
`LN met
`05/05/86
`
`Femoral
`, Supraclavicular
`Neck
`LN met
`LN met
`LN met
`04/27/87
`06/15/87
`06/15/87
`
`
`Brain
`met
`09/28/87
`
`ANTI-
`MELANOMA
`BT-M 6-7
`BT-M e-B
`BT-M e-3
`BT-Me-4
`ET-Me-S
`
`ANTI-BREAST CA
`BA-Br-l
`BA-Hr-3
`BA-Br-4
`BA-Br-S
`BA—Br-6
`R-ll
`R-l3
`
`ANTI-COLON CA
`CO-Co-l
`BT»Co-2
`BT—Co-J
`BT-Co-4
`BT—Co-S
`BRACo—6
`BRACo-7
`BR-Co-8
`BR'Co-9
`
`ANTI—RENAL CA
`BT-Ne-3
`
`Abbal++++
`
`illlrll
`
`lllllllll
`
`+
`
`N#3}l+l++
`
`bat-h+++ll|l
`lllllllll
`
`
`#N ll
`
`Ill.l|l|
`EEEEIIIII
`
`1+
`
`4+
`4+
`
`l +
`—
`
`—
`—
`
`4+
`4+
`3+
`
`lllllllll
`
`l
`
`4+
`4+
`
`4+
`1+
`
`1 +
`3+
`
`4+
`4+
`4+
`
`lllllllll
`
`l
`
`NE = Not available because tissue section was washed of)“
`NT = Not tested
`
`
`68
`
`4--
`4—7
`
`4”
`‘ 2 W
`
`
`
`4--
`4”
`
`4”
`2 W
`
`
`
`1 +
`3+
`
`4+
`4+
`4+
`
`lllllllll
`
`l
`
`2+
`2+
`
`2+
`1 +
`NE
`
`l|l||l|||
`
`l
`
`IMMUNOGEN 2305, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 5
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`——:—-——————_
`
`Old/1am et a]
`
`
`Out of the 43 patients (one patient received one
`course of Adriamycin immunoconjugate and another
`course of mitomycin-C immunoconjugate) included
`in the subject study, flow cytometry analysis was
`performed on tumor biopsy cells from 23 of the pa-
`tients (53%). The agreement between positives on
`immunohistochemistry
`and
`positives
`on
`flow
`cytometry was approximately 90%. Interaction of
`different antibodies was determined on flow cytome—
`try by measurement of the above-described para-
`meters on cells exposed to two or more of the anti-
`bodies simultaneously, or sequentially.
`Comparison of the percent positive and peak
`mean channel
`(PMC) obtained in these mixing
`studies, with those obtained for the single antibody,
`enabled us to determine the degree to which the anti-
`bodies interacted, either additively 0r subtractively.
`These studies demonstrate that greater than 80% of
`the tumor cells were able to be coated with antibodies.
`
`Adriamycin was tightly associated with antibody,
`but the exact nature of the linkage is unknown (10). A
`' significant percentage of Adriamycin (30-40%) was
`covalently linked; the remaining drug was tightly but
`non-covalently associated with antibody. These pre-
`parations were stable for at least 6 months in phos-
`phate buffered saline and would tolerate gel filtration
`under mild conditions; stringent competition for
`binding or denaturation (incubation with Amberlite
`XAD-2 or SDS-PAGE respectively), however, could
`remove the Adriamycin which was not covalently
`linked to antibody. General safety tests revealed that
`non-specific toxicities could occur if immunocon~
`jugate levels were high enough. As reported here,
`early preparations using these methods demonstrated
`little clinical toxicity while later preparations proved
`less stable in serum and caused mild to moderate
`
`Preparation and Testing of Adriamycin or
`Mitomycin-C Immunoconjugates
`
`After selection of appropriate monoclonal anti-
`bodies, production was scaled up in ascites, and these
`antibodies were purified in gram amounts and chemi-
`cally conjugated with Adriamycin or mitomycin-C
`(10). Immunoconjugates of Adriamycin and murine
`antibody using a cis-aconitate linker were prepared
`by a modification of the method of Shen and Riyser
`(11, 12) and are fully described elsewhere (10).
`Clinical grade Adriamycin containing lactose
`(Adria Labs) was converted to a carboxylic acid inter-
`mediate by reacting cis-aconitic anhydride with the
`free Adriamycin base. This compound was in turn
`converted to the active ester by its reaction with
`1- ethyl —3— (3 -dimethylamino- propyl) carbodiimide
`(EDC) and N~hydroxysuccinimide (NHS). The ac-
`tivated Adriamycin solution was mixed with purified
`antibody and stirred overnight at room temperature
`in the dark. Free Adriamycin was removed by tan-
`gential
`flow ultrafiltration using MWCO 30,000
`membranes.
`
`Adriamycin toxicity.
`Concurrent with this change in toxicity and sta-
`bility was a formulation change by Adria Labs involv-
`ing the addition of methylparaben. While not en—
`cumbering the chemical reactions involved in the
`conjugation, methylparaben may have affected the
`strength of the non-covalent association between
`antibody and Adriamycin and thus decreased the
`stability of the complex in serum leading to the release
`of more free Adriamycin.
`The chemotherapeutic agent, mitomycin-C, was
`conjugated via an N-hydroxysuccinimide (NHS) ac-
`tive ester intermediate to amine groups (principally
`provided by lysine, arginine, aspartic acid, glutamic
`acid, and NHz-terminal amino acids) displayed by the
`antibody protein. Mitomycin-C monoclonal anti-
`body molar ratios reproducibly range from 4 to 15,
`but differed amongst various antibodies. The conju-
`gation procedure, modified from Kato et a1. (13)
`entails the addition of a glutaric acid spacer arm to
`mitomycin-C, followed by its conversion to an active
`ester by treatment with NHS and dicyclohexylcar-
`bodiimide (DCC). The resulting compound is crys-
`tallized and stored dessicated at —20° centigrade.
`Conjugation was accomplished by dissolving 86 mg
`_of this active ester in dimethylformamide (DMF),
`which is added to one gram of monoclonal antibody
`Adriamycin-to-antibody molar ratios were cal—
`and borate buffer at pH 8.5 so that the DMF con-
`culated from protein concentrations estimated by the
`centration was 10%. Free mitomycin-C was removed
`Bradford Coomassie Blue binding assay and
`Adriamycin concentrations estimated from a stan-
`by tangential flow ultrafiltration (Minitan) and the
`conjugate mixed with mannitol and stored in the dark
`dard curve of absorbence at 495 nm. Ge] permeation
`at 10° centigrade.
`HPLC with a Superose 6 column (Pharmacia) and
`SDS-PAGE were performed to
`examine
`the
`High yield conjugation was confirmed by high
`performance liquid chromatography (Waters HPLC
`homogeneity of antibody. Immunoconjugates were
`with two Protein Pak 300 SW columns in series)
`sterilized by filtration and stored in the dark at 4°
`demonstrating superimposable peaks of antibody
`centi grade until used.
`
`69
`
`IMMUNOGEN 2305, pg. 6
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 6
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`é I
`
`ndividually specified drug immunoconjugates in cancer treatment
`____—_——-——————————-——’——'—'—
`
`Patient biopsies were tested for the presence (in
`viva targeting) of murine antibody by immunoperoxi-
`dase histochemistry or by flow cytometry using goat
`anti-mouse antibodies as the developing agent (4, 5).
`
`RESULTS .
`
`Table 1 summarizes the tumor categories. For this
`report, 43 patients treated with immunoconjugates
`were analyzed. All antibodies were against membrane
`determinants of tumor cells except for one patient
`with chronic lymphocytic leukemia who was treated
`with a cocktail of anti-idiotypic immunoconjugates.
`No significant changes were seen in total serum
`complement, immunoglobulins or lymphocyte sub-
`sets (data not shown). Five patients received less than
`500 mg of antibody, four patients had between 500
`and 1000 mg and thirty-four patients greater than 1.5
`grams of antibody.
`The antibodies used in this study are summarized
`in Table 2.
`
`Heterogeneity and Selection
`
`and mitomycin-C with less than 2% free drug. Pre-
`servation of mitomycin-C anti-tumor activity was
`documented by exposing the active ester to water and
`determining equivalents in cell-killing capacity be—
`tween free drug and the hydrolyzed active ester.
`The mitomycin-C and Adriamycin immunocon-
`jugates underwent endotoxin analysis (chromagenic
`LAL test), sterility testing (trypticase soy broth
`(aerobes, fastidious), thioglycolate broth (anaerobes,
`aerobes), Sabourad’s dextrose agar (fungi), general
`safety testing (in guinea pigs and mice), and, when
`cells were available, in vitro cytotoxicity testing 3H-
`thymidine or 7SSe-selenomethionine uptake using the
`patients own tumor as target and non—tumor cells)
`and in vivo anti—tumor testing (nude mouse xenograft)
`prior to administration.
`Analysis by flow cytometry indicated that these
`immunoconjugates retained immunoreactivity after
`conjugation. Conjugates prepared in this manner ex—
`hibited antibody specific in vitro cell-killing pro-
`perties. Animal studies indicated 5—15-fold lower non—
`specific lethal toxicity from the conjugate when com-
`pared with free drug. Examination of the bone-mar-
`row of animals treated with DXR conjugates also
`revealed a decreased toxicity compared to free drug.
`These conjugates also failed to produce soft tissue
`necrosis when injected intradermally at drug levels,
`which in an unconjugated state cause severe damage.
`Flow cytometric analysis of Adriamycin and
`MMC conjugates indicated that the breadth of tumor
`cell coverage remains similar with only a small de-
`crease in the intensity of the binding following con-
`jugation (10). The Adriamycin conjugates were very
`stable in vitro but the deconjugation half-time of the
`Br—l-MMC immunoconjugate was found to be one
`day at 37 ° centigrade and 21 days at 10° centigrade. In
`studies of the immunoconjugated Me-4, conjugation
`preserves in vitro cytotoxicity of the immunocon-
`jugate compared to free mitomycin-C, while confer-
`ring specificity. In vivo xenograft studies with Me-7
`and Br-l have demonstrated both comparable anti-
`tumor activity and diminished cytotoxicity of the
`immunoconjugate to free drug. These preclinical re-
`sults with mitomycin—C conjugates are fully described
`elsewhere ( 14).
`
`In Table 3, data from a patient with melanoma
`can be seen. Me-3 was positive in the initial lymph
`node and stayed positive in the subsequent tissue
`culture derived cells but was negative in all other
`specimens. Other antibodies, such as Me-7 and Me-8
`generated from immunization with the first tumor
`biopsy were present in all specimens. Br-6 was not
`present in the original specimen but was in all sub-
`sequent samples though not in the tissue culture line
`derived from the original lymph node. Thus, of the
`five antibodies specifically derived from melanoma
`tumors, only Me-4, Me-7 and Me-8 reacted consist—
`ently with all the tissues and Me-3 and Me—S had
`variable patterns. This stresses the importance of
`having a panel of antibodies and illustrates micro-
`heterogeneity in multiple biopsies over time from a
`single patient.
`,
`In phenotyping the tumor of a patient who had a
`melanoma removed from the right lung (12/86) and,
`from the left lung (7/87) who also had tissue culture
`cells grown from an earlier lymph node biopsy (2/85),
`the melanoma lymph node typed with Co-l or Co-4
`while both lung samples were negative. The left lung
`Testing of serum samples for human anti-mouse
`immunoglobulin was done with a particle concen-
`melanoma metastasis typed strongly with Me-ll but
`the right lung tumor was negative. A weak reaction
`tration fluorescent immunoassay (15, 16). Multiple
`with Br-6 and Br-7 was present in tumor in the left
`samples were tested to detect the development and
`effect of antiglobulins.
`lung but not in the right lung. This patient confirms
`W—
`
`Clinical Monitoring
`
`70
`
`IMMUNOGEN 2305, pg. 7
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 7
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`————_——_——————_’____
`
`Oldham et al
`
`
`
`the variability of antigenic expression on tumor cells
`within an individual over a period of time. At least
`two variables are involved, time and site of meta-
`stasis. Among the ten antibodies generated with mela-
`noma immunization, all three samples were positive
`with six of the antibodies (data not shown).
`-
`Ten patients with breast carcinoma were typed
`with 8 monoclonal antibodies and the variability of
`tissue typing of breast carcinomas is apparent with
`monoclonal antibody Br—l typing positive for 9 out of
`10 the breast carcinomas, but never with 100% of the
`cells in any tumor. Five of the antibodies typed only
`one ofthe breast tumors (data not shown). In order to
`insure typing and saturation, a panel of monoclonal
`antibodies was needed.
`
`The histological evaluation for selection of anti-
`bodies was complemented by flow cytometry of
`viable tumor cell suspensions reacted with the mono-
`clonal antibodies. In Table 4, two examples are shown
`demonstrating both an additive effect and a negative
`effect of mixing other antibodies. Perhaps hindrance
`related to the closeness of the epitopes or modulation
`of the epitopes contributed to the negative interac-
`tions.
`
`In 53% of the patients, final selection was based
`on results of both immunohistochemical and flow
`
`cytometry testing.
`
`Localization of Antibody and Saturation of Cancer
`Cells After Infusion of Monoclonal Antibodies
`
`Tables 5 and 6 Show localization of antibodies
`after infusion of the immunoconjugate into the pa-
`tients. Biopsies of carcinoma and normal tissues were
`studied at varying times after infusion of immunocon-
`j ugates before and after (in vitro) addition of cocktails
`to demonstrate saturation of the antibody mixture on
`carcinoma cells. Tube thoracentesis permitted inter-
`mittent sampling of malignant cells in the pleura at
`
`varying intervals. Localization of the antibody cock—
`tail was consistently seen (19/23; 83%).
`Saturation was seen in 9 instances. Two patients
`studied simultaneously in effusion and solid tumors,
`demonstrated higher saturation in the effusion com-
`pared to the soft tissue biopsies.
`Localization was demonstrated on cancer cells as
`
`early as 4 days after starting treatment and as long as
`10 days after the last infusion.
`Saturation was seen at 1.65 gram antibody in one
`case and in 5 cases at over 2 grams cumulative dose
`antibody. This suggests high doses of antibody (over
`2 grams) are necessary to saturate the various
`epitopes on carcinoma cells. This may represent one
`of the limitations of small doses of unconjugated
`mouse antibody as previously used by others. It may
`partially explain the previous lack of efficacy of
`mouse antibody conjugates in destroying human
`tumors. It strengthens the concept of using large
`, doses of antibody to localize toxic substances to can-’
`cer cells where antibody is used as a carrier.
`Measurements of the serum concentration of in-
`
`fused antibodies in 13 patients (data not shown) de-
`monstrated variable disappearance of the mono-
`clonal antibody from serum by 24 hours after infu-
`SlOI‘l.
`
`Antibody persisted in the serum for several days in
`selected patients. Precise plasma half lives were not
`calculated.
`
`Toxicities
`
`In general, the most frequent toxicities were rash,
`fever, and chills (Tab. 7). The fever and chills were
`easily controlled, and no temperature went above
`102°/ F. A rash with or without pruritis was seen in 19
`patients. Abdominal pain and arthralgia was seen
`only in one patient. In those patients given repeated
`infusions,
`the type of
`reaction mimicked that
`
`TABLE 4 - FLOW CYTOMETRIC ANALYSIS OF MONOCLONAL ANTIBODY BINDING TO TUMOR CELLS
`
`Test
`,
`0/o Cells
`Peak mean
`Type of
`
` target Antibody positive channel interaction
`
`
`
`
`
`
`Ovarian
`carcinoma
`
`97
`Br-l
`104
`Br—3
`Co-6
`95
`Br-l + Br-3 + Co-6
`126
`
`
`additive
`
`Carcinoma—
`sarcoma
`cells
`
`Br-l
`Br-3
`Br«5
`Br-I + Br-3 + Br-S
`
`121
`87
`53
`10
`
`subtractive
`
`71
`
`IMMUNOGEN 2305, pg. 8
`Phigenix v. Immunogen
`IPR2014-00676
`
`IMMUNOGEN 2305, pg. 8
`Phigenix v. Immunogen
`IPR2014-00676
`
`

`

`g I
`
`ndividually specified drug immunoconjugates in cancer treatment
`“Pf/—
`
`TABLE 5 -
`
`IN VITRO AND IN V1 V0 REACTIVITY OF MOABS WITH TUMOR CELLS IN SURGICALLY REMOVED LESIONS DURING
`AND AFTER ADRIAMYCIN-MOAB IMMUNOCONJUGATE THERAPY
`WM
`Immunoperoxidase
`reactivity of MoAb
`
`cocktail scorea
`Comments
`
`Patient
`(tumor type)
`
`MoAb
`cocktail
`
`Type of
`specimen
`
`.
`Saturation
`
`No
`
`No
`
`Cumulative
`dose of
`MoAb (mg)
`before
`biopsy
`
`
`Cumulative
`days post
`initial
`treatment
`before
`biopsy
`
`Number of
`days post
`last
`treatment
`before
`biopsy
`
`In vitm
`
`In vivo
`
`Localization
`
`WEN
`(breast)
`
`Br-l
`Br-J
`Br-4
`
`Pleural
`effusion
`Pleural
`effusion
`Pleural
`eiIusion
`
`1 +
`4+
`4
`4
`75
`0
`3 +
`2
`IS
`150
`50