`
`*#$##**#*#**#$$t
`D
`0569 SCI 522816X 08/
`LOWELL H ERICSSCN
`6296 89TH
`E
`MERCER.
`IS
`
`PFIZER EX. 1026
`
`Page 1
`
`PFIZER EX. 1026
`Page 1
`
`

`

`AMERICAN
`Assocv.TION FOR THE
`ADVANCEMENT OF
`ScIBNCB
`
`SciENCE ISSN 0036-8075
`
`9 JANUARY 1987
`VOLUME 235
`NUM.BER +785
`
`•OdhMIM- 141
`•d i i ik - 144
`
`139 This Weck in Science
`
`Fraud in Science:
`
`Misuse of the Freedom of Information Act: J. R. WILLIAMS • Ri.idenberg's
`Patents: J. L. HUMMER • Quality of Biomedical Literature: R. G. MARTIN •
`Quantitative Risk Aspects of the "Woburn Case": M. T. SMITH • Moonlight and
`Circadian Rhythms: R. M. SINCLAIR; C. A. CzBISLBR AND J. S. ALLAN •
`Research Practices: W. W. STEWART AND N. FEDER
`
`Science Budget: More of the Same • R&D and the Deficit
`Peer Rcview-'oops--Merit Review in for Some Changes at NSF
`Cancer M.D.'s Clash over Interleukin Therapy
`Landsat Commercialization Stumbles Again
`Hazardous Waste: Where to Put It?
`EEC Research Program in Jeopardy
`Court Rejeets Rifkin in Biotech Cases
`Math Papers Called lnaccc:ssible
`Briefing: Comings and Goings
`
`Oncogenes Give Breast Cancer Prognosis
`Materials Scientists Seek a Unified Voice
`Diabetics Should Lose Weight, Avoid Diet Fads • High-Carb Diets Questioned
`Delving into Faults and Earthquake Behavior: How to Stop a Quake by Jogging
`• Earthquakes Arc Giving Little Warning • Coastal Ups and Downs Point to a
`Big Quake • Cutting the Gordian Knot of the San Andreas
`
`Community Diversity: Relative Roles of Local and Regional Processes:
`R. E. RICKLEFS
`Band-Gap Engineering: From Physics and Materials to New Semiconductor
`Devices: F. CAPASSO
`Human Breast Cancer: Correlation of Relapse and Survival with Amplification of
`the HER-2/neu Oncogene: D. J. SI.AMON, G. M. CLARK, S. G. WONG,
`w. J. LEVIN, A. ULLRICH, W. L. McGUIRE
`The Atomic Strucrure of Mengo Virus at 3.0 A Resolution: M. Luo,
`G. VRJBND, G. KAMER, I. MINOR, E. ARNOLD, M. G. ROSSMANN, u. BOEGE,
`D. G. SCRABA, G. M. DUKE, A. C. PALMENBERG
`
`SCIENCE II publlllled "*ly on Fr1dey, ••cei>I Ille 11e1 -
`wttll en H1nl " - In Fetwue,y
`In - . -
`ol Science, 1333 H SIJNC, NW, WNlllngton, DC 20005. S.C,.
`by the American ANoclellon lor the -
`oncl-dass postage (publication No. 484-460) paid at Washlnglon. DC. and at an additional entry. Now combined with The
`Scientific~~ Copyright C> 1987 by the American Association tor the Advancement ol Science. The title SCI·
`ENCE la a registered trlldemlsl< o1 the MAS. ~ individual - ip and oulJocription (51 -
`): 165. Do(cid:173)
`mestic lnstitutlonal 51Jb9cription (51 Issues): $98. Fo,eign poslage ..tra.: C..- $32. Oll>e< (IIU!face md) $27. lir-sur(cid:173)
`faco via -.tam $65. First ciaos, airmail. - -· and sulonl ,_ on rlqUNI. Single ooplN $2.50 ($3 by moi);
`bad< Issues S4 ($4.50 by mail): Biotechnology issue. SS.50 (S6 by mu): ctassroom ratn on ~
`: Gulde to Blotoch·
`notogy Products and Instruments $16 ($17 by mai~. Ctw,ge a l - : . - s - .. gMng old and new -
`and seven-<ligit account numbaf. Authorization to pl-.x,py matarieJ lo< lntemeJ °' po<sOnal -
`unoe, cltcumstancea
`not fall ng within Ille flir uoe provisions ol the Copyright Ad is granted by MAS to 1barin and Oll>e< ,-a regiswed
`with the Copyright Cleerance Center (CCC) T...-ia! Reporting Servioo. provided that lhe -
`fee ol S1 per copy
`p4ua $0.10 per page la paid directly to CCC, 21 Congresa Street. Salem. MassachUMlts 01970. The ldontiticalion code
`lo< Science la 0036-8075183 $1 + .10. - : Send Form 3579 to Science, 1333 H 5tJwt, NW, WUhlngton, DC
`spec:iallnd lndoxn .
`The American Aasoclation ""the Advancement o1 Science wu lcundod in 1848 an d~ 1n 1s1,. 11s ot,fects
`are to lur1het' the won< ol sdentista. to facilitate cooperation among them. to looter !dentific -
`and ""'l)()f\1li>.
`to lmpn,Ye the elloctlveness of oc:ience In the promotion ol human -
`.... and to lncfeue public undemanding and ap(cid:173)
`preciation of the i~ and promise ol the methods ol oc:ience In human progress.
`
`20005. Science is Indexed In lhe -s Guide to Periodical l.itrntun, and In -
`
`SCIENCE, VOL. 235
`
`WiW&ibhhhih- 151
`
`153
`154
`155
`156
`158
`159
`
`•RliiiUiWtWt- 160
`
`161
`163
`165
`
`167
`
`172
`
`177
`
`182
`
`•
`
`•
`
`136
`
`PFIZER EX. 1026
`Page 2
`
`

`

`Lava flow after eruption of Mauna Loa Volcano, March 1984. Lava
`COVBll
`flows advanced nearly 20 kilometers in about 5 days toward the city of Hilo,
`largest city on the island of Hawaii. The flow slowed and stopped about 8
`kilometers from nearest buildings on city's outskirts. All eruptive activity ceased
`by early April 1984. Sec page 196. [Scott Lopez, National Park Service]
`
`Outward-Dipping Ring-Fault Structure at Rabaul Caldera as Shown by
`Earthquake Locations: J. Molli AND C. McKEE
`Sensory Tuning of Lateral Linc Receptors in Antarctic Fish to the Movements of
`Planktonic Prey: J. C. MoNTGOMEllY AND J. A. MACDONALD
`Disruption of the Mauna Loa Magma System by the 1868 Hawaiian Earthquake:
`Geochemical Evidence: R. I. TILLING, J. M. RHODES, J. W. SPAllKS,
`J. P. LocICWOOD, P. W. LIPMAN
`Oxygen Supersaturation in the Ocean: Biological Versus Physical Contributions:
`H. CllAJG AND T. HAYWAllD
`Identification of a Juvenile Hormone-Like Compound in a Crustacean:
`H. l..AUFEll, D. BORST, F. c. BAJCEll, c. CAllllAsco, M. SINKUS, c. C. REtrrEll,
`L. w. TSAI, D. A. ScHOOLBY
`Evolution of Male Pheromones in Moths: Reproductive Isolation Through Sexual
`Selection?: P. L. PHELAN AND T. C. BAJCEll
`Direct Aaivation of Mammalian Atrial Muscarinic Porassium Channels by GTP
`Regulatory Protein Gk: A. YATANI, J. CoDINA, A. M. BllOWN, L. BIRNBAUMEll
`Mctalloregulatory DNA-Binding Protein Encoded by the merR Gene: Isolation
`and Characterization: T. O'HALI.ollAN AND C. WALSH
`Nerve Growth Factor Treatment After Brain Injury Prevents Neuronal Death:
`L. F. l<B.OMEll
`Lung Cancer and Indoor Air Pollution in Xuan Wei, China: J. L. MuMFOllD,
`X. Z. HE, R. S. CHAPMAN, S. R. CAo, D. B. HAlllUs, X. M. LI, Y. L. XIAN,
`W. Z. JIANG, C. W. Xu, J. C. CHUANG, W. E. WILSON, M. CooKB
`Preferred Microtubulcs for Vesicle Transpon in Lobster Axons: R. H. MILLE&,
`R. J. LAsEK, M. J. KATZ
`
`The Darwinian Heritage, rer>ie'tPed by A. DESMOND • Confronting Nature,
`Y. GINGllAS • Are Australian Ecosystems Different?, G. H . OIUANs • Under the
`Qoud and Justice Downwind, R. A. DMNE • Books Received
`
`Portable Macroscope • Diacylglycerol Kinase • Polyacrylarnide Gels for
`Electrophoresis • FORTRAN Language Subroutines • PVC Tubing • Miniature
`Centrifuge • Tcfton-Coated Slides • Literature
`
`193
`
`195
`
`196
`
`199
`
`202
`
`205
`
`207
`
`211
`
`214
`
`217
`
`220
`
`224
`
`MUiiihh4**1MiGUMt• 229
`
`...... al Dlrecton
`Gerard Piel
`11111mg , ,_ ,,
`
`°"""'1an = Bogorad
`,,,_.,,_
`
`lllollaE. Wodnall
`
`__
`
`E -E. Balley
`David BaltimoNt
`William F. Bririunan
`Phllip E. Con-,,e
`Jooeph L Goldlteln
`James D. Idol, Jr.
`~ Knopo4I
`Seymour lJpoo!
`W-Musoy
`~E. Nollon
`Devld V. Ragone
`David M. Ral4>
`Vera C. Rullin
`Lally L Smarr
`Solomon H. Snyder
`Aol>el1M. So1ow
`Jameo D. Watoon
`
`-.iMcC. Adams
`-
`w. Bef1ine<
`Floyd e. Bloom
`Maly E. Clutter
`Milcnd s. Oresselhaus
`Donald N. Langenbe<g
`Dorothy Neikln
`Linda s. Wilson
`WIiiiam T. Golden
`T,...,,.,
`Wollam D. C.,oy
`Execut/veOflc:,er
`
`-o l~ng
`Edltora
`John-
`Oeil AI-Awqati
`JameoP. Alioon
`DonLAndlnon
`E-H.lllackbum
`Floyd E. Bloom
`Chat1N A. Canlor
`JamNH. Clarl<
`Bruce F. Eldndge
`StanleyF(cid:173)
`Theodo!11 H. Gebde
`Roger I. M. Glas
`6'opllln p. Gol
`-B. ~
`
`Co,wy s. Goodman
`RlchardM . . (cid:173)
`Glorla Heppner
`Eric:F. (cid:173)
`Kornd B. Ktauel<opf
`I. -
`l.elvnar,
`KIit L MOQloby
`JooephB. Martin
`John C. McGII
`
`Allon----Ollon
`
`Gordon H. Oriana
`John$. -
`
`Yeshayau Packer
`Jean Paul Revel
`James E. Rolhman
`Thomas C. Sc:hoUing
`Ronald H. Schw111Z
`Si,,pl1an M. Schwartz
`Otto T. Solbrig
`Robert T. N. ljlan
`VlrginlaTMlble
`Geerat J. verme,
`MartJn G. Weigert
`Harold Weintraub
`Irving L Weissman
`George M. Whitesides
`OwenN. Wotte
`Wl liam B. Wood
`
`9 JANUAllY 1987
`
`TABLE OF CONTENTS
`
`137
`
`PFIZER EX. 1026
`Page 3
`
`

`

`Human Breast Cancer: Correlation of
`Relapse and Survival with Amplification
`of the HER-2/neu Oncogene
`D ENNIS J. SIAMON, * GARY M. CLARK, STEVEN G. WoNG, W ENDY J. L EVI N
`AxEL ULLRICH, WILLIAM L. McGUIRE
`
`The HER-2/neu oncogene is a member of the erbB-like
`oncogene family, and is related to, but distinct from, the
`epidermal growth factor receptor. This gene has been
`shown to be amplified in hwnan breast cancer cell lines.
`In the current srudy, alterations of the gene in 189
`primary hwnan breast cancers were investigated. HER-2/
`neu was found to be amplified from 2- to greater than 20-
`fold in 30% of the rumors. Correlation of gene amplifica(cid:173)
`tion with several disease parameters was evaluated. Am(cid:173)
`plification of the HER-2/neu gene was a significant pre(cid:173)
`dictor of both overall survival and time to _rel~pse in
`patients with breast cancer. It retained its sigoiticance
`even when adjustments were made for other known
`prognostic factors. Moreover, HER-2/neu amplification
`had greater prognostic value than most currently used
`prognostic factors, including hormonal-receptor status,
`in lymph node-positive disease. These data indicate that
`this gene may play a role in the biologic behavior and/or
`pathogenesis of human breast cancer.
`
`T HE EVIDENCE LINKING P ROTO-ONCOGENES TO TH E INDUC·
`
`tio11 or maintenance of human malignancies is largely cir(cid:173)
`cumstantial, but has become increasingly compelling. This
`circumstantial evidence is derived from studies of animal models,
`tumor cell lines and aetual human rumors. Data from ani mal models
`and cell lines include: (i) sequence homology between human proto(cid:173)
`oncogene and the viral oncogene of tran forming retroviruses that
`are known to be tumorigenic in some species (1, 2); (ii) transfection
`rudies showing the tran forming potential of prom-oncogenes in
`lH 3T3 ceUs and primary embryo fibroblasts (3- 5); and (iii) the
`central role of certain proto-oncogenes in cumorigenesi by chronic
`transforming retroviruses uch as avian leukosis viru (6). Data from
`human rumors include: (i) increa ed expression of pecific proto(cid:173)
`in ome human malignancies (7, 8); (ii) localization of
`oncogene
`proto-oncogenes at or near the site of pecific tumor-as ociaced
`chromo omal tran locations (9); and (iii) amplification of proto(cid:173)
`oncogenes in some human rumors (10, 11 ).
`Additional data linking prom-oncogenes to cell growth is their
`ex pre ion in respon e to cer:tain proliferation ignals (12, 13) and
`their expression during embryonic development (14, 15). More
`direct evidence comes from the fact that, of the 20 known proto(cid:173)
`oncogencs, three are related to a growth factor or a growth factor
`include c-sis which is homologous to the
`receptor. The e gene
`
`transforming gene of the imian sarcoma viru and i the J3 chain of
`platelet-derived growth factor (PDGF) (16, 17) ; c-fois, which i
`homologous to the transforming gene of the feline sarcoma virus
`and is do cly related to the macrophage colony-stimulating factor
`receptor (CSF-l R) (18); and c-erbB, which encodes the EGF
`receptor (EGFR) and is highly homologous to the transforming
`gene of the avian erythroblasto is virus (19 ). The two receptor·
`related proto-oncogenes, c-fins and c-erbB, arc members of the
`tyrosine-specific protein kinase family to which many proto-onco(cid:173)
`gencs belong.
`Recently, a novel tran forming gene was identified as a result of
`transfection rudics with D A from chemically induced rat neu(cid:173)
`roglioblastomas (20). Thi gene, called neu, was shown to be related
`to, but di tinct from, the c-erbB proto-oncogene (21 ). By means of
`v-erbB and human EGFR as probes to screen human genomic and
`complementary DNA (cD A) libraries, two other group indepen(cid:173)
`dently isolated human erbB- related genes that they called HER-2
`(22) and c-erbB-2 (23). Subsequent sequence analy i and chromo(cid:173)
`somal mapping rudies revealed all three gene (nm, c-erbB-2, and
`HE R-2) to be the same (22, 24, 25). A fourth group, also using v(cid:173)
`erbB as a probe, identified the same gene in a mammary carcinoma
`cell line, MAC 117, where it was found to be amplified five- to ten(cid:173)
`fold (26).
`This gene, which we will call HER-2/nei,, encodes a new member
`of the tyrosine kinase family; and is do cly related to, but distinct
`from, the EGFR gene (22). HER-2/neu differ from EGFR in that it
`is found on band q2 1 of chromo omc 17 (22, 24, 25), as compared
`to band pll- pl3 of chromosome 7, where the EGFR gene is
`located (27) . Also, the HER-2/11e11 gene generates a messenger
`RNA (mRNA) of 4.8 kb (22), which differs from the 5.8- and IO(cid:173)
`kb transcripts for the EGFR gene (28). Finally, the protein encoded
`by the HER-2/tteu gene i 185,000 daltons (21 ), as compared to the
`170,000-dalton protein encoded by the EGFR gene. Conver ely, on
`the basi of equence data, HER-2/nm i more closely related to the
`EGFR gene than to ocher members of the tyrosine kinase famil y
`(22). Like the EGFR protein, HER-2/neu has an extracellular
`domain, a transmembrane domain that includes two cysteine-rich
`repeat clusters, and an intracellular kinase domain (21 ), indicating
`
`. G. Wong, and W. / . Levin ffl: in the Division of Hcmorolog)'·
`lamon,
`D. ).
`Oncology, Department or Medicine and Jo ns.son Comprchcnsi\'c Cancer Center,
`U LA School of Medicine, Los Angeles, CA 90024. G. M. Clork and W. L. McGuire
`arc in the Divisio n o f Oncology, Depann,cnt of Medicine, nh·ersit)' of Texas He,lth
`Science Center at San Antonio, San Anton io, T X 78284. A. UUrich is in the
`Department of Molecular Biology, Gcncntcch, Inc., South an Francisco, CA 94080.
`
`*'l"o whom co rrespondence should be: adcl=d.
`
`9 JANUARY 1987
`
`ARTlCLES 177
`
`PFIZER EX. 1026
`Page 4
`
`

`

`that it too i likely to be a ccUular recepror for an as yet unidentified
`ligand.
`Al> a result of the published data showing amplification ofHER-
`2/neu in a human mammary carcinoma cell line, and as pan of an
`ongoing survey in our laborarory of prom-oncogene abnormalities
`in human rumors, we evaluated alterations of the HER-2/,ieu gene
`in a large series of human primary breast cancers. Our results show
`that amplification of this gene occurs relatively frequently in breast
`cancer, and that it is associated with disease relapse and overall
`patient survival.
`Factor that are known to be imponant in the prognosis of breast
`malignancies in individual patients include: size of the primary
`rumor, stage of disease at diagnosis, hormonal receptor status, and
`number of axillary lymph nodes involved with disease (positive
`nodes) (29). The current study, which was conducted in two parts,
`involved the evaluation of tissue from 189 separate breast malignan(cid:173)
`cies that were pan of a breast cancer srudy ongoing at the Univer ity
`of Texas, San Antonio. This cohon of rumors was of interest
`because considerable information was available on the majority of
`the specimen including size of the primary rumor, estrogen recep·
`tor starus, progesterone receptor starus, age of patient, disease stage,
`and status of the axillary lymph nodes.
`In the initial survey, tissue from 103 prima.ry breast cancers was
`evaluated for alterations in the HER-2/neu gene. DNA from
`individual rumors was prepared as described (30), digested with
`Eco Rl, and subjected to Southern blot analysis with a 32P-labeled
`HER-2/nttt-l probe, which is known to detect a 13-kb hybridizing
`band in human D A (22). Examples of rumors from the initial
`survey are shown in Fig. l. Of the 103 samples examined, 19 (18%)
`showed evidence of HER -2/,ieu gene amplification. The degree of
`amplification in individual cases was determined by dilution analysis
`(Fig. 2A), as well as sofi: laser densitometry scanning. To determine
`that the amount of DNA loaded in each lane was equivalent, all
`filters were washed and rehybriclized with a 32P-labeled arginase
`gene probe (31 ). This probe identifies a 15-kb hybridizing band on
`Eco RI-digested human D A, and was selected as a control
`because ir more appropriately assesses the relative amount and
`
`Table 1. Association between HER·2Jneu amplification and disease parame-
`ters in I 03 breast rumors.
`
`Nwnbcr of cwnors
`
`Factor*
`
`Single
`copy
`
`2 to 5
`copies
`
`5 to 20
`copies
`
`> 20
`copies
`
`Pt
`
`Total
`
`ER+
`ER-
`PgR+
`PgR-
`
`s2
`2-5
`>5
`Unknown
`
`s50
`> 50
`Unknown
`
`0
`1-3
`>3
`Unknown
`
`53
`31
`42
`42
`
`13
`34
`17
`20
`
`21
`52
`11
`
`30
`20
`17
`17
`
`I
`4
`2
`3
`
`Humwnal naptur srlll'US
`2
`9
`2
`I
`2
`6
`s
`I
`Tumor si:u (unrimam)
`l
`1
`l
`5
`I
`2
`0
`3
`A,ge at diagnosis (yuin)
`I
`I
`2
`4
`2
`7
`0
`2
`0
`Number of positive lymph >U!da
`I
`0
`3
`I
`I
`0
`2
`4
`2
`I
`l
`3
`
`0
`1
`2
`2
`
`0.99
`
`0.85
`
`0.82
`
`0.83
`
`0.11
`
`65
`38
`52
`51
`
`15
`41
`22
`25
`
`25
`65
`13
`
`34
`22
`25
`22
`
`•Rcccpror st>.rus was analyzed as described (39). ER, c:srrogcn receptor: + and - «:fer
`co the presence or absence of ., 3 finol o( =cptor per inilligrun of protein. PgR,
`progc:stcronc receptor: + and - refers to the pr=ncc or abot:ncc of ;;,,S frnol of receptor
`per milli~ of protein.
`t Sotistical analyses for concb.tion of HER· 21""' amplifi ·
`cation with disease parameters we«: performed by the x' tc:st. P V2lucs were computed
`after combining the cases with S to 20 and > 20 copic:s.
`
`transfer of high molecular weight species than a probe hybridizing
`with low molecular weight pecics, which transfer more readily on
`Southern blotting. All lanes were hown to contain equivalent
`amounts of high molerular weight DNA (Fig. 2B). lnclividual
`rumors were assigned t0 groups containing a single copy, 2 to 5
`copies, 5 to 20 copies, and greater than 20 copies of the HER-2/,im
`gene (Fig. 1). Assignment of rumors to the various group was done
`
`Fig. 1. Analysis of alterations of the HER-2/neu
`gene in human breast cancer. hown arc 79 of the
`189 breast rumors used in this anal)•sis. Tumors
`with a single copy of HER-2/,m,: 3, 4, 10 to 15,
`20, 23 co 25, 27 to 29, 31, 38, 42 10 46, 48, 49,
`52, 55, 61, 65, 66, 71, 72, and 74. Tumors with
`two to 6ve copies of HER-2/,ieu: I, 2, 5, 7, 9, 16,
`17, 19, 21, 22, 32, 35, 36, 47, so, 54, 56 to 58,
`60, 62, 70, and 75 to 77. Tumors with 5 to 20
`copies of HER-2/neii: 6, 8, 26, 34, 37, 39 to 41,
`51, 53, 63, 64, 67, 69, 73, and 79. Tumors with
`more than 20 copies of HER-2/mu: 18, 30, 33,
`59, 68, and 78. Examples of rumors 77 10 79 have
`rearrangements in the HER-2/ner, gene. DNA
`W3S extracted from tissues and digested with Eco
`Rl as described (30). A total of 12 µg of Eco RI(cid:173)
`digested DNA was loaded onto 0.8% agarose
`gels, separated by electrophoresis, and rransferrcd
`onto nylon filter papers (Biodync) (30). All filters
`were baked in a vacuum oven for 3 hours at 80°C,
`prchybridizcd in 5 x
`(standard saline citrate)
`containing 50% formamide, 10% dcxtran sulfate,
`0.1 % SD , dcnarured salmon sperm D A ( 1 m~
`ml), and 4 X Dcnhardts solution for 12 hours,
`then hybridized in the same solution containing
`32P-labc:led nick-translated HER-2 probe (2/)
`specific activity of I x 108 cpm per microgram of
`DNA; 2 x 106 cpm/ml. Hybridization occurred
`at 42°C for 48 hours, foUowcd by washing of
`filters under the foUowing conditions in succcs·
`
`1 2 3 4 5
`
`• 6 7 8 9
`
`· 10 11 12 13 14
`
`15 16 17 18 19 20
`
`21 22 23 24 2526 2728 29
`1
`
`4142 43 44 45 46 47 48 49 50
`
`51 52 53 54 55 56 ST 58 58 60 61 62 63 64 65 66
`
`67 68 81 70 71 72 73 74 75 76
`
`n 1a 7'9
`
`kb
`•12
`
`- 4
`•31
`
`•18
`
`sion: 2 X SSC for 20 minutes at room tempera·
`rure; cwo washes of 30 minutes each in 2 x SSC,
`0.1% SOS at 65° ; one wash of 30 minures in
`
`, 0.1% D at 65°C. Filters were then
`O.Sx
`exposed 10 XAR-5 x-ray 6Jm (Kodak) for autora·
`diography.
`
`SCIENCE, VOL. 23S
`
`PFIZER EX. 1026
`Page 5
`
`

`

`2. (A) Example of dilutional analysis ro assess degree of HER-2/nn.
`pie amplification. Lanes a, g, k, and p were loaded with 12 µg of Eco RI(cid:173)
`digested breast rumor DNA. Lane a is DNA from rumor 31 (Fig. I), which
`rq,rcscnts a rumor with a single copy of the HER/2-nn. gene. Lane g is
`DNA &om rumor 33, which represents a rumor with > 20 copies of the
`HER-2/neu gene. Lanes bro fare serial dilutions (1:100, 1:20, 1: 10, 1:5,
`and I: 2, respectively) of the D A sample in lane g. Lane k is D A from
`nunor 35 (Fig. l ), which represents a tumor containing two to five copies of
`me HER-2/neu gene. Lanes b to j arc serial dilutions ( 1: 10, 1: 5, and 1:2,
`respectively) _of the D A sample in l~e k. Lane p is _D A from rumor 34
`(fig. 1), which represents a tumor with 5 to 20 copies of the HER-2/neu
`P· Lanes I ro o arc serial dilutions ( l: 20, l : 10, l: 5, and 1: 2, respectivc(cid:173)
`ly)__~f the D A sample in lane p. The filter was prepared and hybridized with
`a 31P-labclcd HER-2 probe as in Fig. I. (B) Example of arginasc probe
`hybridization to demonstrate that equivalent amounts of tumor D A were
`loaded inro each lane. Rehybridization of filter containing lanes 30 to 40
`(Fig. I). The filter was first stripped of label by washing in a buffer made up
`of 50% formamidc, 3x S , and 0.1% SDS at 65°C for 20 minutes,
`following by three su cessive washes of 5 minutes each in 0 .1 x SSC at room
`temperature. Filters were exposed overnight on XAR-5 film (Kodak) to
`ensure removal of all radioactive probe, then rchybriclizcd as in Fig. I with a
`np.tabelcd human arginasc gene probe (31).
`
`in a blinded fashion, in that they were made without knowledge of
`disease parameters. Analysis of the data for association between gene
`amplification and a number of clisease parameters was then per(cid:173)
`formed.
`Of 103 rumors evaluated in the initial survey, there was essentially
`no correlation between gene amplification and estrogen receptor
`starus, progesterone receptor starus, size of rumors, or age at
`diagnosis (Table l ). However, when analysis was performed for
`association between HER-2/neu amplification and number of posi(cid:173)
`tive lymph nodes, a trend was noted. This analysis showed that 4/34
`(11%) of patients with no involved nodes, 2/20 (10%) with l to 3
`involved nodes, and 8/25 (32%) with >3 involved nodes had gene
`amplification (P = 0.11 ). lfthese data were examined by comparing
`0 to 3 positive nodes versus > 3 positive nodes, the correlation with
`gene amplification became more significant (P < 0.05). Thus, there
`was a significant increase in incidence of HER-2/neu gene amplifica(cid:173)
`tion in patients with > 3 axillary lymph nodes involved with clisease.
`A multivariate regression analy is to correlate HER-2/neu amplifica(cid:173)
`tion with various di ease parameters identified the number of
`positive nodes as the only
`ignificant factor, either alone or in
`combination, to correlate with amplification.
`This initial srudy indicated that it might be possible to discrimi(cid:173)
`nate among node-po itive patients on the basis ofHER-2/neu gene
`
`A
`a
`
`B
`30
`
`b
`
`c
`
`d
`
`e
`
`g
`
`h
`
`m n
`
`o
`
`p
`
`31
`
`32
`
`33
`
`34
`
`35
`
`36
`
`37
`
`38
`
`39
`
`40
`
`amplification. It is well known that the number of positive nodes is
`the best progno tic factor for clisease recurrence and survival in
`patients with breast cancer (29). Given the correlation between
`number of nodes positive and HER-2/neu amplification, one might
`preclict that amplification of this gene might also have some
`prognostic value.
`o long-term follow-up data, however, were
`available on the l 03 patients analyzed in the initial study. For this
`reason a second study was conducted on 100 breast cancer sample
`from patients with positive axillary lymph nodes. All of the informa(cid:173)
`tion available for the first group of I 03 patients was available for
`these patients. ln adclition, relapse and urvival information was
`available, since these cases had a meclian foUow-up of 46 month
`( range 24 to 86 months). Of these l 00 samples, 86 yielded ufficient
`D A for rudy. Amplification of the HER-2/ntt, gene was mea(cid:173)
`sured as in the initial survey, and examples of tumors from this srudy
`are shown (Fig. 1). Amplification was found in 34/86 (40%) of
`these patients. For this larger sample of node-positive patients,
`several statistically significant or nearly significant relationship were
`observed. In agreement with the preliminary urvey, there was an
`association between number of involved lymph nodes and HER-2/
`,1eu amplification (Table 2). In adclition, the presence of gene
`amplification was correlated with estrogen receptor starus and size
`of primary rumor (Table 2). Together, these two surveys yielded
`data on 189 patients and the association of HER-2/neu amplifica(cid:173)
`tion with various clisease parameters in the combined group is
`shown in Table 3.
`While these correlations were of interest, the strong relationship
`
`1.0
`
`~ 0.8
`~
`~ 0.6
`a.
`J 0.4
`i
`; 0.2
`
`0
`
`Amptllted (n : 34)
`
`A
`
`C
`
`Not amplified (n:52)
`
`Amptllted (II = 11 )
`>5 copies
`
`1.0
`:E 0.8
`:g
`D 0.6
`e a.
`
`;;; 0.4 i 0.2
`
`Amp lllled (/1 =34)
`
`Amplilied (n :::11 )
`>5 copies
`
`B
`0 + - - - - - - - -- ~~·--,--,
`12
`36
`60
`72
`84
`48
`0
`24
`Time (mon l hs)
`
`D
`
`12
`
`24
`
`36
`
`60
`
`72
`
`84
`
`ARTICLES 179
`
`Fig. 3. Actuarial curve for relapse in (A) node(cid:173)
`positive patients with no amplification versus
`node-positive patients with any amplification (> 2
`copies) of HER-2/neu and (C) node-positive ~a(cid:173)
`ticnts with no amplification versus node-posiave
`patients with greater than 5 copies ofHER-2/nn..
`Actuarial curve for overall survival in (B) node(cid:173)
`po itivc patients with no amplification versus
`node-positive patients with any amplification (> 2
`copies) of HER-2/,,,u and (D) node-positive pa(cid:173)
`tients with no amplification versus node-positive
`patients with greater than 5 copies ofHER-2hieu.
`Actuarial curves for both relapse and overall sur·
`viva! were computed by the method of Kaplan
`and Meier (44) and compared by the log rank test
`(42-44).
`
`9 JANUARY 198
`
`PFIZER EX. 1026
`Page 6
`
`

`

`between HER-2/11t11 amplification and nodal starus (P = 0.002)
`indicated that information on amplification of this gene may
`correlate with disease behavior; that is, recurrences and survival. To
`test this, univariate survival analyses were performed in which
`amplification was compared to relapse and survival in this patient
`group. A total of 35 patients had a recurrence of the disease, and 29
`had died at the time of the analyses. Median times to relapse and
`death were 62 months and 69 months, respectively. The median
`follow-up time for patients still alive was 47 months, ranging from
`24 to 86 months. A total of 71 of the 86 patients (83%) received
`some form of therapy after mastectomy: adjuvant systemic therapy
`alone, 47%; adjuvant ystemic therapy plus local radiation, 19%;
`and local radiation alone, 17%. A strong and highly statistically
`significant correlation was found between i:he degree of gene
`amplification and both time to disease relapse (P = < 0.0001 ) and
`survival (P = 0.0011 ) (Table 4). Moreover, when compared in
`univariate analy es to other parameters, amplification ofHER-2/neu
`was found to be superior to all other prognostic factors, with the
`exception of the number of positive nodes (which it equaled) in
`predicting time to relapse and overall survival in human breast
`cancer (Table 4). The association between HER-2/ne,i amplification
`and relapse and survival can be illustrated graphically in actuarial
`survival curves (Fig. 3, A co D). While there was a somewhat
`shortened time to relapse and shorter overall survival in patients
`having any amplification of the HER-2/neu gene in their rumors
`(Fig. 3, A and B), the greatest differences were found when
`comparing patients with > 5 copies of the gene to those without
`amplification (single copy) (Fig. 3, C and D). Patients with greater
`than five copies of HER-2/neu had even shorter disease-free survival
`times (P = 0.015) and overall survival times (P = 0.06) when
`compared to patients with no amplification. The phenomenon of
`greater gene copy number correlating with a worse prognosis has
`also been seen in evaluations of N-myc gene amplification in human
`neuroblastoma (32).
`To determine if an1plification of HER-2/neu was independent of
`other known prognostic factors in predicting disease behavior,
`multivariate survival analyses were performed on the 86 node(cid:173)
`positive cases. Amplification of the gene continued to be a strong
`progno tic factor, providing additional and independent predictive
`information on both time to relapse and overall survival in these
`
`51 52 53 54
`
`55 56 57 58 59 60
`
`61
`
`62
`
`63 64 65 66
`
`Fig. 4. Example of rehybridization of filter with human EGFR probe. Filters
`were stripped as in Fig. 2B, and hybridized with 32P-labelcd human EGFR
`probe (28), as in Fig. l. Shown are the lower molecular weight bands
`hybridized with 32P-labeled EGFR probe in filter-containing lanes 51 to 66
`(Fig. l ). The bands from top to bottom arc 2.8, 2.2, and 1.8 kb, respectively.
`Lane 52 is an example of a nimor showing marked amplification (>50
`copies) of the EGFR gene.
`
`patients, even when other prognostic factors were taken into
`account (Table 4 ).
`Rearrangement of the HER-2/ne11 gene was rare. Of the total 189
`rumors evaluated, three showed evidence of rearrangement, and in
`two of the three cases, the rearrangement was identical (Fig. I, cases
`77 to 79). Also, two of the rearranged HER-2/neu loci were
`amplified (Fig. 1, cases 78 and 79). The incidence of HER-2/neu
`rearrangement as determined by Eco RI digestion was too small ro
`attempt statistical correlations.
`To determine whether the phenomenon of amplification ofHER-
`2/neu in breast cancer extended to related growth factor receptors,
`all filters were analyzed with the EGFR probe (Fig. 4). Amplifica(cid:173)
`tion of the EGFR gene was found in 4/189 (2%) of the cases, and
`rearrangement of the EGFR gene was found i.n one of those four
`cases. The incidence of EGFR amplification and rearrangement was
`too small to attempt statistical correlation. Comparison of HER-2/
`neu amplification (531189 or 28%) with that of the EGFR gene
`reveals the incidence of the former to be 14 tinies greater than that
`of the latter indicating that the phenomenon of gene amplification
`is not a general one for a related tyrosine kinase-specific receptor in
`human breast cancer. Moreover, studies examining alterations of
`two other tyrosine kinase-specific proto-oncogenes 1Wl and Jes, in
`breast cancer did not show amplification of these genes (33).
`Alterations of non-tyrosine kinase-related prow-oncogenes in these
`
`Table 3. Association between HER-2/neu amplification and disease parame(cid:173)
`ters in combined survey (189 paticncs).
`
`Table 2. AssQciation between HER·2/11eu amplification and disease parame·
`tcrs in 86 breast tumors from node-positive patients.
`
`Faaor*
`
`Single
`copy
`
`2 to 5
`copies
`
`5 to 20
`copies
`
`>20
`copies
`
`Total
`
`pt
`
`Faaor•
`
`inglc
`copy
`
`2 to 5
`copies
`
`5 to 20
`copies
`
`>20
`copies
`
`Total
`
`pt
`
`ER+
`ER-
`PgR+
`PgR-
`
`S 2
`2-5
`>5
`
`S50
`>50
`
`l- 3
`>3
`
`38
`14
`31
`21
`
`18
`28
`6
`
`16
`36
`
`31
`21
`
`l
`I
`
`H omuma/ rueptur rtatus
`21
`5
`2
`4
`18
`4
`5
`5
`Tumqr size (centimetm)
`8
`3
`12
`2
`3
`4
`Age at diagnosis (Jean)
`12
`6
`l
`3
`11
`1
`Number of positiw: lymph ,wdes
`7
`5
`0
`16
`2
`4
`
`0
`1
`1
`
`65
`21
`54
`32
`
`29
`43
`14
`
`35
`51
`
`43
`43
`
`0.05
`
`0.14
`
`0.09
`
`0.06
`
`0.06
`
`t Starutical analyses for correlation of
`•ER and PgR arc as described in Table I.
`HER·21neu amplification with various disease ~cters w<r<: performed by the x'
`ten. P values were computed after combining e 5 to 20 and > 20 cases, since th=
`were so few samples in the > 20 group.
`
`ER+
`ER-
`PgR+
`PgR-
`
`s2
`2-5
`>5
`Unknown
`
`S50
`>50
`Unknown
`
`0
`1- 3
`>3
`Unknown
`
`91
`45
`73
`63
`
`31
`62
`23
`20
`
`37
`88
`11
`
`30
`51
`38
`17
`
`2
`5
`3
`4
`
`H ornw,u,I receptur rtatus
`23
`14
`6
`3
`20
`JO
`6
`10
`T,mwr siu (cmtimam)
`9
`4
`13
`7
`4
`6
`0
`3
`Age at diag,,otis (Jean)
`2
`13
`8
`13
`10
`5
`0
`0
`2
`N11mber of positive lympi, notks
`0
`3
`l
`7
`6
`l
`18
`8
`4
`3
`I
`l
`
`0
`2
`3
`2
`
`0.05
`
`0.06
`
`0.19
`
`0.11
`
`0.002
`
`130
`59
`106
`83
`
`44
`84
`36
`25
`
`60
`116
`13
`
`34
`65
`68
`22
`
`Statistical analyses for corrtlation of
`·ER and PgR arc as described in Table I.
`HER-2/neu amplification \\ith ,•arious distasc par.tm<tcrs were: performed by the x'
`'"'· P valucs were computed after combining the cases with 5 to 20 and > 20 copics.
`
`180
`
`SCIENCE, VOL. 235
`
`PFIZER EX. 1026
`Page 7
`
`

`

`4. Univariate and multivariate analyses comparing disease-free survival (relapse) and overall survival to prognostic f