`
`.
`
`~
`
`PLA
`
`APRIL 1, 1991
`VOLUME88
`NUMBER7
`
`NATURAL
`
`.,.
`sc,cNcrs
`
`Proceedings
`
`OF THE
`
`.
`
`National Acadeiily
`of Sciences
`
`OF THE UNITED STATES OF AMERICA
`
`PFIZER EX. 1628
`Page 1
`
`
`
`Proceedings
`OF THE
`National Academy
`of Sciences
`OF THE UNITED STATES OF AMERICA
`
`Officers
`of the
`Academy
`
`Editorial Board
`of the
`Proceedings
`
`FRANK PRESS, President
`JAMES D. EBERT, Vice President
`PETER H. RAVEN, Home Secretary
`JAMES B. WYNGAARDEN, Foreign Secretary
`ELKAN R. BLOUT, Treasurer
`
`ROBERT H. ABELES
`GORDON A. BAYM
`MICHAEL J. CHAMBERLIN
`MARY-DELL CHILTON
`
`LAWRENCE BoGORAD, Chairman
`RONALD L. GRAHAM
`GORDON G. HAMMES
`ERIC R. KANDEL
`PHILIP W. MAJERUS
`HERBERT A. SIMON
`
`MAXINE F. SINGER
`HAROLD V ARMUS
`THOMAS A. WALDMANN
`SHERMAN M. WEISSMAN
`
`Managing Editor: FRANCES R. ZWANZIG
`Senior Associate Editor: GARY T. COCKS
`Associate Editor: CAY BuTLER
`Associate Editor: JOHN M. MALLOY
`Associate Editor: MARILYN ] . MASON
`Associate Editor: JANET L. MORGAN
`Associate Editor: T. PEARSON
`Associate Editor: DOROTHY P. SMITH
`Associate Editor: COLENE RUCH WALDEN
`Assistant Managing Editor: JOANNE D'AMICO
`
`Senior Production Editor: BARBARA A. BACON
`Production Editors: DEBORAH G. CHLEBOVE, EILEEN P. DELANEY, JAMIE M. FEAR,
`BILL FOGLE, SCOTT C. HERMAN, KATHLEEN RUBY,
`ANNE M. SUNDERMANN, DoN C. TIPPMAN
`Proofreader: MARY E. McLAUGHLIN
`Administrative Assistants: DELORES BANKS, BRENDA L. McCoy
`Manuscript Coordinators: PATRICIA A. GoDLEY, JACQUELINE PERRY
`Circulation: JULIA LITTLE, CYNDY MATHEWS, VIRGINIA TREADWAY
`
`Editorial correspondence: PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES, 2101 Constitution Avenue, Wash(cid:173)
`ington, DC 20418.
`
`Business correspondence: Circulation Office of the PROCEEDINGS, National Academy of Sciences, 2101 Constitution
`Avenue, Washington, DC 20418.
`Information for Contributors: Seep. i (of this issue} and pp. i-viii of issue number 1, January 1, 1991.
`Copyright: The National Academy of Sciences has copyrighted this journal as a collective work and does not own copyright
`for individual articles. Requests for permission to reproduce parts of individual articles or for reprints of individual articles
`should be addressed to the authors. Microforms of complete volumes are available to regular subscribers only and may be
`obtained from University Microfilms, Xerox Corporation, Ann Arbor, Ml48103. This journal is printed on acid-free paper
`effective with volume 84, issue 1.
`Subscriptions: All correspondence concerning subscriptions should be addressed to the Circulation Office of the PROCEED(cid:173)
`INGS. Subscriptions are entered on a calendar year basis only. For 1991, subscription rates are as follows-in the United States:
`student/postdoctoral, $80; personal, $210; institutional, $380; elsewhere by surface mail: student/postdoctoral, $175;
`personal, $305; institutional, $475; elsewhere by Air Cargo at a surcharge of$102. Information regarding other air mail postage
`rates is available from the Circulation Office. Subscribers are requested to notify the Circulation Office of the PROCEEDINGS
`6 weeks in advance of any change of address; also the local postmaster. The Academy is not responsible for nonreceipt of
`issues because of an improper address unless a change of address is on file. The notice of address change should list both the
`old and new addresses. Claims for replacement copies will not be honored more than 60 days after the issue date for domestic
`subscribers and not more than 90 days after the issue date for foreign subscribers.
`Back Issues: Volumes 83-87, January 1986 and thereafter, are available from the Circulation Office of the PROCEEDINGS.
`The price of a single issue is $20.00.
`PRINTED IN THE USA
`S•cond dass postal(• pllid til Washington. DC. and at additional mailing offic.s.
`PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA (JSSN-0027-8424) is
`published semimonthly by THE NATIONAL ACADEMY OF SCIENCES, 2101 Constitution Avenue, Washington, DC 204/8.
`© 1991 by THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED STATES OF AMERICA .
`POSTMASTER: Send address changes to: PROCEEDINGS OF THE NATIONAL ACADEMY OF SCIENCES OF THE UNITED
`STATES OF AMERICA, 2101 Constitution Ave., Washington , DC 20418.
`
`PFIZER EX. 1628
`Page 2
`
`
`
`Proc. Nat/. Acad. Sci. USA
`Vol. 88, pp. 2869-2873 , April 1991
`Immunology
`
`Humanized antibodies for antiviral therapy
`(herpes simplex virus I computer modeling)
`
`MAN SuNG Co*t, MARGUERITE DESCHAMPS *, RICHARD J . WHITLEY*, AND CARY QuEEN *
`
`*Protein Design Labs, Inc. , 2375 Garcia Avenue, Mountain View, CA 94043 ; and *Department of Pediatrics, University of Alabama, Birmingham, AL 35294
`
`Communicated by Stanley Falkow, January 14, 1991
`
`Antibody therapy holds great pr~mise for the
`ABSTRACT
`treatment of cancer, autoimmune disorders, and viral infec(cid:173)
`tions. Murine monoclonal antibodies are relatively easy to
`produce but are severely restricted for therapeutic use by their
`lmmunogenicity in humans. Production of human monoclonal
`antibodies has been problematic. Humanized antibodies can be
`generated by introducing the six hypervariable regions from
`the heavy and light chains of a murine antibody into a human
`framework sequence and combining it with human constant
`regions. We humanized, with the aid of computer modeling,
`two murine monoclonal antibodies against herpes simplex virus
`gB and gD glycoproteins. The binding, virus neutralization,
`and cell protection results all indicate that both humanized
`antibodies have retained the binding activities and the biolog(cid:173)
`ical properties of the murine monoclonal antibodies.
`
`It was first shown in 1891 that the antibodies induced during
`a viral infection can neutralize the inciting virus (1). For
`certain acute viral infections such as rabies, hyperimmune
`serum from infected patients has been a traditional therapy
`(2). More recently, the development of monoclonal antibody
`technology has allowed generation of specific antibodies
`against various viral antigens (3). Several reports have ap(cid:173)
`peared showing that monoclonal antibodies can protect
`against various viral di seases in animal models {4-9). The use
`of monoclonal antibodies thus provides a new approach to
`antiviral therapy .
`The production of murine monoclonal antibodies is rela(cid:173)
`tively straightforward, but problems in the production of
`human monoclonal antibodies have persisted (10). In addition,
`the resulting human antibodies are frequently not of the
`appropriate isotype or do not possess the desired specificity.
`On the other hand, because xenogeneic antibodies are highly
`immunogenic in humans , the potential use of murine mono(cid:173)
`clonal antibodies for human therapy is limited, especially
`when repeated administration is necessary. The immune re(cid:173)
`sponse against a murine monoclonal antibody may potentially
`be reduced by transforming it into a chimeric antibody. Such
`antibodies combine the variable binding domain of a mouse
`antibody with human antibody constant domains (11, 12).
`However, in a study to evaluate the immunogenicity of chi(cid:173)
`meric antibodies, it was found that the anti-variable domain
`response was not attenuated in the chimeric antibody, dem(cid:173)
`onstrating that foreign variable frameworks can be sufficient to
`lead to a strong anti-antibody response (13). Therefore, for
`therapeutic purposes it may be necessary to fully humanize a
`murine monoclonal antibody by re shaping both the variable
`and the constant domains to make them human-like.
`Winter and colleagues (14) first successfully humanized
`both chains of a rat antibody, directed against human lym(cid:173)
`phocytes, by introducing the six hypervariable regions from
`the rat heavy- and light-chain variable regions into human
`variable region framework sequences. Recently, a human-
`
`ized antibody that binds to the human interleukin 2 receptor
`(p55) has also been reported (15). However, generation of
`other fully humanized antibodies has proved unexpectedly
`difficult, because significant loss of binding affinity generally
`resulted from simple grafting of hypervariable regions , prob(cid:173)
`ably due to di stortion of the complementarity-determining
`region (CDR) conformation by the human framework .
`Herpes simplex virus (HSV) infections range from asymp(cid:173)
`tomatic to life threatening (16). More than 50 HSV polypep(cid:173)
`tides have been identified in HSV -infected cells, including
`seven major cell-surface glycoproteins (17). The specific
`biologic functions of these glycoproteins are not well defined ,
`although gB and gD have been shown to be associated with
`cell fusion activity (18, 19). gB and gD express both type(cid:173)
`specific and type-common antigenic determinants. Many of
`the antibodies against gB and gD have shown high neutral(cid:173)
`izing activities in vitro and in vivo (20-24). Oakes and Lausch
`(20) demonstrated that monoclonal antibodies against gB and
`gE suppress replication of HSV-1 in trigeminal ganglia. Dix
`et a!. (21) showed that anti-gC and -gD antibodies protect
`mice against acute virus-induced neurological disease. Whit(cid:173)
`ley and colleagues (22-24) produced a panel of murine
`monoclonal antibodies against HSV-1 and showed that sev(cid:173)
`eral of the antibodies protected mice against encephalitis and
`death following ocular inoculation with the virus. Clone Fd79
`(anti-gB) prevented encephalitis even when immunization
`was delayed until 48 hr postinfection. Fd79 and Fdl38-80
`(anti-gD) significantly reduced the severity of epithelial kera(cid:173)
`titis and lowered the frequency of persistent vi rill infection in
`an outbred mouse model , suggesting potential therapeutic
`uses in humans. Because murine monoclonal antibodies are
`limited by their immunogenicity for human therapy , we chose
`to humanize these two antibodies. In this article, we describe
`the construction of humanized antibodies for Fd79 and
`Fdl38-80. These humanized antibodies retain the binding
`affinities and biological properties of the murine antibodies.
`
`MATERIALS AND METHODS
`Reagents. Vero cells were obtained from American Type
`Culture Collection (CCL 81) and maintained in minimum essen(cid:173)
`tial medium with 10% fetal bovine serum and nonessential amino
`acids. HSV-1 [~305 mutant (F strain)] (25) was a gift of Ed
`Mocarski (Stanford University). All enzymes were obtained
`from New England Biolabs and all chemicals were from Sigma
`unless otherwise specified. Staphylococcal protein A-Sepharose
`CL-4B was from Pharmacia. 1251 was from Amersham. Immu(cid:173)
`nostaining reagents were ordered from Tago.
`Synthesis of Variable Domain Genes. The construction of
`variable domain genes for the humanized antibody heavy
`chain and light chain generally follows ref. 15. The nucleotide
`sequences were selected to encode the protein sequences of
`the humanized heavy and light chains, including signal pep(cid:173)
`tides, generally utilizing codons found in the mouse se-
`
`The publication costs of this article were defrayed in part by page charge
`payment. This article must therefore be hereby marked "advertisement"
`in accordance with 18 U .S .C . §1734 solely to indicate this fact.
`
`Abbreviations: HSV , herpes simplex virus ; CDR, complementarity-
`determining region ; pfu , plaque-forming unit(s).
`·
`t ro whom reprint requests should be addressed.
`
`2869
`
`PFIZER EX. 1628
`Page 3
`
`
`
`2870
`
`Immunology: Co et a/.
`
`Proc. Nat/. Acad. Sci. USA 88 (1991)
`
`quence. Several degenerate codons were changed to create
`restriction sites or to remove undesirable ones. For each
`variable domain gene , two pairs of overlapping oligonucle(cid:173)
`otides on alternating strand s were synthesized (380B DNA
`synthesizer; Applied Biosystems), which encompassed the
`entire coding sequences as well as the signal peptide and the
`splice donor signal. Each oligonucleotide was 110-140 bases
`long with a 15-base overlap. Double-stranded DNA frag(cid:173)
`ments were synthesized with Klenow polymerase, digested
`with restriction enzymes, ligated to the pUC18 vector, and
`sequenced. The two fragments with the correct sequences
`were then ligated into the Xba I sites of expression vectors
`similar to those described in ref. 15.
`Expression and Purification of Humanized Antibodies. For
`each humanized antibody constructed , the heavy-chain and
`light-chain plasmids were linearized at the BamHI sites and
`transfected into Sp2/ 0 mouse myeloma cells by electropora(cid:173)
`tion. Cells were selected for gpt expression. Clones were
`screened by assaying human antibody production in the
`culture supernatant by ELISA.
`
`Antibodies from the best-producing clones were purifiect
`by passing tissue culture supernatant over a column or
`staphylococcal protein A-Sepharose CL-4B . The bound an.
`tibodies were eluted with 0.2 M glycine·HCI (pH 3.0) and
`neutralized with 1 M Tris·HCI (pH 8.0) . The buffer was
`exchanged into phosphate-buffered saline (PBS) by passina
`over a PD10 column (Pharmacia).
`Fluorocytometric Analysis. Vero cells were jnfected with
`HSV-1 at 3 plaq~e~forming units (pfu) per cell overnight.
`Cells were tryps1mzed at 0.5 mg/ ml for 1 min, washed
`extensively with PBS , and resuspended in FACS buffer
`(PBS/ 2% fetal calf serum/ 0.1% azide) at = 5 x 106 cells per
`mi. One hundred microliters of cell suspension was trans(cid:173)
`ferred to a polystyrene tube and incubated with 100 ng of
`purified antibody on ice for 30 min . The cells were washed
`with F ACS buffer and incubated with fluorescein isothiocy.
`anate-labeled goat anti-human antibody on ice for another 30
`min. The cells were washed again and resuspended in
`PBS/ 1% paraformaldehyde. Cells were analyzed on a FAC(cid:173)
`Scan (Becton Dickinson).
`
`A
`
`c
`
`2 0
`1 5
`10
`5
`1
`EMI LVESGGGLVKPGASLKL
`E V Q L L E S G G G L V Q P G G S L R L
`
`20
`1 5
`10
`5
`1
`QVQLQQSDAELVKPGASV K I
`Q V Q L V Q S G A E V K K P G S S V K V
`
`7 5
`7 0
`SREDAKNTL Y
`S RNDSKNTL Y
`
`75
`70
`65
`60
`S E K F K GIK A T L T A D K sA s T A y
`. EKFKGKATITADESTNT AY
`l A
`
`c
`8 2 a
`95
`9 0
`85
`b
`80
`c
`8 2 a
`9 5
`90
`85
`b
`8 0
`M H L N S L T S E D S A V Y F C A R iGRl
`L Q M S S L K S E D T A L Y Y C L R ~
`M E L S S L R S E D T A V Y F C A R ~
`L Q M N S L Q A E D T A L Y Y C L R~
`
`c
`1 00 a
`b
`d
`1 OS
`11 0
`DD S R ER N G FA Yv i Ww G Q G T LV T V S
`S R E R N G F A
`. G Q G T L V T V S
`
`11 o
`1 os
`1 oo a b c d
`k
`Y Y A D Y G F F D Vv i Ww G T G T T V I V
`y y A D Y G F F D
`. G Q G T L V T V
`
`.
`I
`
`11
`
`
`
`.
`I
`
`113
`
`s s
`s s
`
`B
`
`1
`D
`E
`
`20
`15
`10
`5
`V L T Q S P A S L A V S L G 0 R A T
`V M T Q S P A T L S V S P G E R A T
`
`113
`A
`s
`
`D
`
`1
`D
`D
`
`20
`15
`10
`5
`V M T Q S H K F M S T S V G D R V S
`Q M T Q S P S T L S A S V G D R V T
`
`40
`3 5
`30
`25
`T CciKK AS Q D V GSA V VviWw H Q Q K S
`T
`. A S Q D V G S A V
`. H Q 0 K P
`
`45
`40
`Q Q K P G Q P P K L L
`Q Q K P G Q S P R L L
`
`45
`G Q S P K L L
`GKAPKLL
`
`75
`70
`65
`60
`G V PARFSGSGFGTDFT L NI H
`Gl PARFSGSGSGTEFTLT IS
`
`80
`75
`70
`65
`RFTGSGSGTDFTLT I TNV O S
`RFT G SGSGTEFTLTI SSLO P
`
`85
`80
`PVEEEDT V T YY
`R L E S E D F A V Y Y
`
`90
`
`1 OS
`107
`107
`100
`105
`GTRLELK
`K
`TFGGGT K LE
`GTKVE V K
`K
`T F G Q G T R V E
`FIG. 1. Amino acid sequences of the heavy chain (A ) and the light chain (B) of the murine and humanized Fd79 antibodies and the hea~~
`chain (C) and light chain (D ) of the murine and humanized Fd138-80 antibodies . The sequences of the murine antibodies as deduced from t k
`eDNA (upper lines) are shown aligned with the humanized ·antibody sequences (lower line.s). The humanized Fd79 and Fd138-80 framewo~
`sequences are derived from Pom and Eu antibodies, respectively. Residues are numbered according to the Kabat system (30). The three CD s
`in each chain are boxed . Residues in the Pom and Eu framework that have been replaced with murine sequences or consensus human sequence
`are underlined .
`
`PFIZER EX. 1628
`Page 4
`
`
`
`Immunology : Co et a/.
`
`Proc. Nat/. Acad. Sci. USA 88 (1991)
`
`2871
`
`Measurements. Binding affinities of the mouse and
`antibodies were determined by competitive bind(cid:173)
`, Vero cells infected with HSV-1 as described
`used as a source of gB and gD antigens. Increas(cid:173)
`of competitor antibody (mouse or humanized)
`to 1.5 ng of radioiodinated tracer mouse antibody
`1 Ci = 37 GBq) and incubated with 4 x 105
`cells in 0.2 ml of binding buffer (PBS/ 2% fetal
`1% azide) for 1 hr at 4°C. Cells were washed and
`, and their radioactivities were measured . The con(cid:173)
`of bound and free tracer antibody were calcu(cid:173)
`binding affinities were calculated according to the
`of Berzofsky and Berkower (26).
`Neutrali:zation Assay. Neutralizing activity of the roo(cid:173)
`humanized antibodies was assayed by a plaque
`method. Briefly, serial dilutions of antibodies were
`with 100 pfu of virus and incubated at 37°C for 1 hr. The
`were then inoculated onto six-well plates with conflu(cid:173)
`cells and adsorbed at 37oC for 1 hr. Cells were
`with 1% agarose in complete medium and incubated
`Plaques were stained with neutral red. The anti(cid:173)
`.............. ntr•>t.·r'" was recorded for 90% plaque reduction.
`Protection Assay. Twenty-four-well plates of con(cid:173)
`cells were inoculated with virus at 0.1 pfu per cell
`to adsorb for 2 hr at 37oC before adding 1 ml of
`sin medium (10, 1, or 0.1 !J.g/ ml). At the end of 4
`culture medium with antibodies was removed and
`were washed and dried by placing overnight in a 37°C
`. To detect viral antigens, each well was incubated
`!J.I of mouse Fd79 antibody at 0.5 1-Lg/ ml for 1 hr at
`twice, and incubated with 200 !J.I of peroxidase(cid:173)
`goat anti-mouse immunoglobulin (1:300 dilution)
`at 37°C. The plates were washed and then developed
`substrate 3-amino-9-ethylcarbazole for 15 min at
`temperature. The reaction was stopped by rinsing with
`and air drying.
`Analysis. Sequence analyses and homology
`were performed with the MicroGenie sequence
`software (Beckman). The molecular model of the
`domains was constructed with the ENCAD program
`examined with the MIDAS program (28) on an Iris
`graphics workstation (Silicon Graphics, Mountain
`,CA).
`
`RESULTS
`of Heavy-Chain and Light-Chain eDNA. cDNAs for
`"""'"'v"'-"''~·n and light-chain variable domain genes were
`by using anchored polymerase chain reactions (29)
`3' primers that hybridized to the constant regions and 5'
`that hybridized to the dG tails (details to be published
`. The heavy-chain variable domain gene of Fd79
`mouse heavy-chain subgroup IIIB , and the light
`ongs to K-chain subgroup III . The heavy chain and
`· ofFd138-80 belong to the heavy-chain subgroup II
`· subgroup V, respectively . The translated amino
`sequences of the two antibodies are shown in Fig. 1.
`Modeling of Humanized Antibodies. To retain
`affinity in the humanized antibodies, the general
`_..,u,,rp~ of Queen eta/. (15) were followed . First, a human
`variable region with maximal homology to the
`antibody is selected to provide the framework se(cid:173)
`for humanization of the mouse antibody. Normally
`chain and light chain from the same human
`are chosen so as to reduce the possibility of incom(cid:173)
`in the assembly of the two chains. Based on a
`homology search against the NBRF protein se(cid:173)
`ti
`data base, the antibody Porn was chosen to provide
`ramework sequences for humanization of Fd79.
`
`The computer program ENCA D (27) was used to construct
`a model of the Fd79 variable region. Inspection of the refined
`model of murine Fd79 revealed two amino acid residues in the
`framework that have significant contacts with the CDR
`residues (Table 1). Lysine in light chain position 49 has
`contacts with three amino acids in CDR2 of the light chain
`(L50Y , L53N , L55E ; see Table 1 for explanation of coding
`system) and two amino acids in CDR3 of the heavy chain
`(H99D, H100Y). Leucine in heavy-chain position 93 shows
`interactions with an amino acid in CDR2 of the heavy chain
`(H35S) and an amino acid in CDR3 of the heavy chain
`(H100cF) . Hence, L49K and H93L were retained in the
`construction of humanized Fd79, as their replacement with
`human Porn framework residues would be likely to introduce
`distortions into the CDRs. Also, seven other residues in the
`Porn framework (five in the light chain and two in the heavy
`chain) were substituted with consensus human residues
`(identical to the murine Fd79 sequence in six of the choices)
`because of their rare occurrence in other human antibodies.
`The elimination of unusual amino acids in the framework may
`further reduce immunogenicity. The murine Fd79 sequences
`and the corresponding humanized sequences are shown in
`Fig. 1 A and B. Substituted residues in the Porn framework
`are underlined.
`Similarly , the murine heavy-chain and light-chain se(cid:173)
`quences of Fd138-80 were compared to the NBRF protein
`sequence data base, and the human antibody Eu was selected
`to provide the framework sequence for humanized Fd138-80 .
`Inspection of a computer-generated model of Fd138-80 re(cid:173)
`vealed six amino acid residues in the framework that show
`important contacts with CDR residues. The residues and
`their contacting counterparts are listed in Table 1; these
`murine residues were retained in the construction of human(cid:173)
`ized Fd138-80. Two other residues (L87F and H37M) show
`significant contacts with L98F, which is immediately adja(cid:173)
`cent to CDR3, so these two mouse residues were also
`retained . Eight amino acids in the Eu framework (two in the
`light chain and six in the heavy chain) were substituted with
`the murine residues (which are also consistent with the
`human consensus residues) because of their rare occurrence
`in other human antibodies. The murine Fd138-80 sequences
`and the corresponding humanized sequences are shown in
`Fig. 1 C and D. Substituted residues in the Eu framework are
`underlined.
`Properties of Humanized Antibodies. The humanized Fd79
`and Fd138-80 antibodies were characterized by comparisons
`with the murine and chimeric antibodies. Both humanized
`antibodies bind to Vero cells infected with HSV-1 in a
`fluorocytometric analysis in a manner similar to the chimeric
`
`Fd79
`
`Fd138-80
`
`Table 1. Residues in the framework sequence showing contacts
`with residues in the CDRs
`Amino
`Contacting CDR residues
`acid
`Residue
`L50Y , L53N , L55E, H99D, HlOOY
`K
`L49
`H35S , H100cF
`L
`H93
`L34V , L89Q
`H
`L36
`y
`H32H , H341
`H27
`y
`H32H , H53R
`H30
`H63F
`F
`H48
`H63F
`K
`H66
`H67
`H63F
`A
`The amino acid residues are numbered according to the Kabat
`system (30) : the first letter (H or L) stands for the heavy chain or light
`chain , the following number is the residue number, and the last letter
`is the amino acid single-letter code . The CDRs are defined according
`to Kabat. Light chain: CDR1 , residues 24-34; CDR2, residues 50-56;
`CDR3, residues 89-97. Heavy chain: CDR1 , residues 31-35; CDR2 ,
`residues 50-65 ; CDR3, residues 95-102.
`
`PFIZER EX. 1628
`Page 5
`
`
`
`2872
`
`Immunology: Co et a/.
`
`Proc. Nat/. A cad. Sci. USA 88 (1 991)
`
`A
`
`B
`
`FIG. 2. Fluorocytometry of HSV-1-infected Vero cells stained with Fd79 (A) and Fdl38-80 (B ) antibodies . · · ·, lsotype-matched control
`antibody;·····, humanized antibody ; - -, chimeric antibody.
`
`antibodies (Fig. 2), showing that they recognize their respec(cid:173)
`tive viral antigens. Chimeric antibodies (unpublished data)
`rather than the original mouse antibodies were used for this
`analysis so the same second-step staining reagent could be
`used . To more quantitatively assess the binding activity,
`radioiodinated murine antibodies were bound to virally in(cid:173)
`fected cells and Scatchard analysis was performed . The
`affinities of the humanized antibodies were determined by
`competition with the iodinated antibodies. The measure(cid:173)
`ments indicate that there is no significant loss of binding
`affinities in the humanized antibodies. Specifically , there is
`an =2-fold decrease in affinity in humanized Fd79 compared
`to the murine Fd79 (Ka. 5.3 x 107 M- 1 vs 1.1 x 108 M- 1). The
`affinity of humanized Fd138-80 is comparable to that of the
`murine antibody (K3 , 4.8 x 107 M- 1 vs 5.2 x 107 M- 1) . These
`results suggest the general usefulness of computer modeling
`in the design of humanized antibodies.
`Murine Fd79 and Fd138-80 have been shown to neutralize
`HSV-1 in vitro without complement (22), so the neutralizing
`activities of the humanized antibodies were compared with
`the mouse antibodies. Serial dilutions of equal quantities of
`murine and humanized antibodies were incubated with virus
`for 1 hr before inoculation onto Vero cells. After 4 days , cells
`were stained with neutral red to visualize plaques. Results
`from these plaque-reduction assays indicated that both hu(cid:173)
`manized Fd79 and Fd138-80 neutralize virus as efficiently as
`their murine counterparts (Fig. 3). Both humanized and
`murine Fd79 cause a 90% reduction of plaques at an antibody
`concentration of 7 nM (1 ~-tg/ml). Similarly , humanized and
`murine Fd138-80 were able to cause a 90% plaque reduction
`at equivalent levels.
`The antibodies were also investigated for their ability to
`protect cells from viral spread in tissue culture. Vero cells
`were inoculated with virus at 0.1 pfu per cell and allowed to
`adsorb for 2 hr at 37oC before addition of 10, 1, or 0.1 1-Lg per
`ml of antibody. After 4 days, antibodies were removed and
`cells were stained with mouse Fd79 antibody for detection of
`viral antigens on infected cells. Results indicated that human(cid:173)
`ized Fd79 at 1 ~-tg/ml (Fig. 4A) and murine Fd79 (data not
`shown) protected culture cells from viral spread. Cells stained
`with anti-gB antibodies were negative, except isolated single
`cells, which were infected with virus before introduction of
`protective antibodies. However, neither humanized (Fig. 48)
`nor murine (data not shown) Fd138-80 was able to protect cells
`against viral spread, despite their ability to neutralize virus
`before inoculation. Fig. 48 shows that total cell lysis and
`staining with anti-gB antibodies were observed even in the
`presence of humanized Fd138-80 (10 ~-tg/ml) . Both gB and gD
`are thought to be associated with cell fusion and virus infec(cid:173)
`tivity (18, 19). However, it is possible that Fd79 blocks both
`the infectivity and cell fu sion functions of gB , while Fd138-80
`does not block the fu sion epitope of gD, so virus can still
`spread cell to cell. This is not surprising, as it has been reported
`
`that polyclonal antibody to glycoprotein D did not prevent the
`spread of virus from cell to cell in culture (31).
`
`DISCUSSION
`The binding, neutralization, and protection results all indi(cid:173)
`cate that the humanized Fd79 and Fd138-80 antibodies have
`retained the binding activities and the biological properties of
`the murine monoclonal antibodies. The use of murine mono(cid:173)
`clonal antibodies for therapy is hindered by the generation in
`humans of an immune response to the mouse antibodies (32).
`The potential advantages of a humanized antibody are (i) the
`
`120.--- - - - - - - - -- -- -- -- ,
`A
`
`•
`
`1.00
`
`c
`0
`-~
`.!:!
`~ :;
`<ll z
`~ 0
`
`80
`
`60
`
`40
`
`20
`
`~·
`
`0 ~---~-------~----~
`.1
`1
`10
`100
`1000
`Antibody Concentration (nM)
`
`12 0 . - - - - - - - - - - - - - - - --
`8
`
`--,
`
`100
`
`c
`0
`~
`.!:!
`
`~ al
`
`z
`~ 0
`
`80
`
`60
`
`40
`
`20
`
`0 ~~~~~---~---~---~
`.1
`1
`10
`100
`Antibody Concentration (nM)
`
`FtG. 3. Neutralization of HSV-1 by Fd79 (A) and Fdl38-80 (B)
`antibodies. e, Mouse ; o, humanized .
`
`PFIZER EX. 1628
`Page 6
`
`
`
`Immunology: Co et a/.
`
`Proc. Nat/. Acad. Sci. USA 88 (1991)
`
`2873
`
`· enhance effector functions such as antibody-dependent cel(cid:173)
`lular cytotoxicity, leading to greater therapeutic efficiency in
`human patients. The actual efficacy of the antibodies in
`human patients must be evaluated by clinical trials.
`
`We thank Michael Levitt and Phil Payne for helpful discussions in
`designing the humanized antibodies and Barry Selick for the expres(cid:173)
`sion vectors.
`
`1. Babes, V. & Cerchez, T. (1891) Clin. Bucuresci 2, 133.
`2. Baltazard , M., Bahnanyan, M., Ghodssi, M. , Sabeti, A., Gaj(cid:173)
`dusek, C. & Rouzbehi , E. (1955) Bull. WHO 13, 747- 772.
`3. Kohler, G. & Milstein , C. (1975) Nature (London) 256, 495-
`497 .
`4. Schmaljohn, A. L., Johnson, E. D. , Dalrymple , J. M. & Cole,
`G. A. (1982) Nature (London) 297, 70-72.
`5. Balachandran, N., Sacchetti , S. & Rawls, W. E. (1982) Infect.
`lmmun . 37, 1132-1137.
`6. Rector, J. T., Lausch, A. R.N . & Oakes, J. E. (1982) Infect.
`lmmun . 38, 168- 174.
`7. Mathews, J. H. & Roehrig, J . T. (1982) 1. lmmunol. 129,
`2763-2767 .
`8. Letchworth , G. I. & Appleton, J. A. (1983) Infect. lmmun. 39,
`208-212.
`9. Kumel , G., Kaerner, H. C., Levine , M. , Schroder, C. H. &
`Glorioso, J. C. (1985) 1. Virol. 56, 930-937.
`10. James , K. & Bell , G. T . (1987) 1. 1mmunol. Methods 100,5-40.
`11. Morrison, S. L., Johnson, M. J., Herzenberg, L.A. & Oi ,
`V. T. (1984) Proc. Nat/. Acad. Sci. USA 81, 6851-6855.
`12. Boulianne , G. L. , Hozumi , N. &Shulman,M. J. (1984)Nature
`(London) 312, 643-646.
`13 . Bruggemann, M., Winter, G., Waldmann , H. & Neuberger, J.
`(1989) 1. E.xp. Med. 170, 2153-2157.
`14. Reichmann , L. , Clark, M., Waldmann , G. & Winter, G. (1988)
`Nature (London) 332, 323-327.
`15. Queen , C., Schneider, W. P., Selick, H. E., Payne, P. W.,
`Landolfi , N. F., Duncan , J. E., Avdalovic, N. M., Levitt, M.,
`Junghans , R. P. & Waldmann, T . A. (1989) Proc. Nat/. Acad.
`Sci. USA 86, 10029-10033.
`16. Corey, L. & Spear, P. G. (1986) N. Engl. 1. Med. 314,749-757.
`17. Spear, P. G. (1985) in The Herpesvirus, ed. Roizman , B .
`(Plenum, New York) , Vol. 3, pp. 315-356.
`18. Cai , W., Gu , B. & Pearson , S. (1988) J . Viro/. 62, 2596-2604.
`19. Fuller, A. 0 . & Spear, P. G. (1987) Proc. Nat/. Acad. Sci. USA
`84, 5454-5458.
`20. Oakes, J. E. & Lausch, R. N. (1984) 1 . Virol. 51, 656-661.
`21. Dix , R. D. , Pereira, L. & Baringer, J. R. (1981) lnfect. lmmun .
`34, 192-199.
`22. Koga, J., Chatterjee, S. & Whitley, R. J. (1986) Virology 151,
`385-389.
`23. Metcalf, J. F., Koga, J. , Chatterjee, S. & Whitley, R. J. (1987)
`Curr. Eye Res. 6, 173-177.
`24. Metcalf, J. F., Chatterjee, S., Koga, J. & Whitley, R. J. (1988)
`lnterviro/ogy 29, 39-49.
`25. Post, L. E., Mackem , S. & Roizman, B. (1981) Cell 24,
`555-565.
`26. Berzofsky, J. A. & Berkower, I. J. (1984) in Fundamental
`Immunology, ed. Paul , W. E. (Raven , New York), pp. 595-
`644 .
`27. Levitt, M. (1983) 1. Mol. Bioi. 168, 595-617.
`28. Ferrin , T. E., Huang , C. C., Jarvis , L. E. & Langridge, R.
`(1988) 1. Mol. Graphics 6, 13-17.
`29. Loh , E. Y., Elliot, S., Cwirla, S., Lanier, L. L. & Davis ,
`M. M. (1989) Science 243, 217-220.
`30. Kabat, E. A., Wu , T. T., Reid-Miller, M., Perry , H. M. &
`Gottesman , K. S. (1987) Sequences of Proteins of Immunolog(cid:173)
`ical Interest (Natl. lnst. Health , Bethesda, MD) .
`31. Hoggan, M. D., Roizman, B. & Turner, T . B. (1960) 1. lmmu(cid:173)
`nol. 84, 152- 159.
`32. Schroff, R. W., Foon, K. A., Beatty , S. M., Oldham , R. K. &
`Morgan , A. C. (1985) Cancer Res. 45, 879-885.
`33. Hale, G. , Dyer, M. J. S. ,Clark, M. R., Phillips , J. M., Marcus,
`R., Reichmann , L., Winter, G. & Waldman , H. (1988) Lancet
`ii, 1394-1399.
`34. Corey, L. , Nahmias , A. J., Guinan , M. E., Benedetti, J. K.,
`Critchlow, C. W. & Holmes, C. K. (1982) N. Engl. 1. Med. 306,
`1313-1319.
`
`Immunostaining of Vero cell monolayers infected with
`4.
`in the presence of humanized Fd79 antibodies (1 J.Lg/ ml) (A)
`........... ,,.,u Fd138-80 antibodies (10 J.Lg/ ml) (B).
`
`of, or significantly reduced, immune response allowing
`treatment; and (ii) an increased serum half-life,
`the required dose as well as extending the effective
`remains to be evaluated in clinical trials whether
`antibodies will induce an anti-isotypic response
`likely, an anti-idiotypic response, and whether
`antibodies will be superior to chimeric antibodies.
`...... , .. u.oo;;u antibody (CAMPATH-lH) used to treat two
`with non-Hodgkin lymphoma was able to induce
`with no anti-globulin response (33).
`- · .. 4V''"J of humanized antibodies with specificity
`gB and gD should provide an opportunity for
`the in vivo potency and immunogenicity of human(cid:173)
`ies in treating viral diseases. The recognition by
`and Fd138-80 of type-common epitopes of gB and gD
`expands the therapeutic potential to HSV-2 as well as
`The use of a combination of two or more humanized
`in therapy could be important to reduce the
`nt of antibody-resistant strains. Combination ther(cid:173