`
`Human Breast Cancer: Correlation of
`Relapse and Survival with Amplification
`of the HER-2/neu Oncogene
`
`DENNIs J. SLAMON,* GARY M. CLARK, STEVEN G. WONG, WENDYJ. LEVIN,
`AXEL ULLRICH, WILLIAM L. MCGUIRE
`
`The HER-2/nex oncogene is a member of the erbB—like
`oncogene family, and is related to, but distinct from, the
`idermal growth factor receptor. This gene has been
`shown to be amplified in human breast cancer cell lines.
`In the current study, alterations of the gene in 189
`primary human breast cancers were investigated. HER-2/
`neu was found to be amplified from 2- to
`greater than 20-
`fold in 30% ofthe tumors. Correlation ofgene amplifica-
`tion with several disease parameters was evaluated.
`Am-
`plification of the HER-2/nex gene was a significant pre-
`dictor of both overall survival and time to relapse in
`patients with breast cancer. It retained its significance
`even when adjustments were made for other known
`prognostic factors. Moreover, HER-2/nex amplification
`ad greater prognostic value than most currently used
`prognostic factors, including hormonal-receptorstatus,
`in
`lymph node~positive disease. These data indicate that
`this gene mayplay a role in the biologic behavior and/or
`pathogenesis of human breast cancer.
`
`HE EVIDENCE LINKING PROTO-ONCOGENESTO THE INDUC-
`
`tion or maintenance of human malignancies is largely cir-
`cumstantial, but has become increasingly compelling. This
`circumstantial evidence is derived from studies of animal models,
`tumorcell lines, and actual human tumors. Data from animal models
`andcell lines include:(i) sequence homology between human proto-
`oncogenes and the viral oncogenes of transforming retroviruses that
`are known to be tumorigenic in some species (1, 2); (ii) transfection
`studies showing the transforming potential of proto-oncogenes in
`NIH 3T3cells and primary embryo fibroblasts (3-5); and(iii) the
`central role of certain proto-oncogenes in tumorigenesis by chronic
`transforming retroviruses such as avian leukosis virus (6). Data from
`human tumors include: (i) increased expression of specific proto-
`oncogenes in some human malignancies (7, 8); (ii) localization of
`proto-oncogenes at or near the site of specific, tumor-associated
`chromosomal translocations (9); and (iti) amplification of proto-
`oncogenes in some human tumors (10, 11).
`Additional data linking proto-oncogenes to cell growth is their
`expression in response to certain proliferation signals (12, 13) and
`their expression during embryonic development (14, 15). More
`direct evidence comes from thefact that, of the 20 known proto-
`oncogenes, three are related to a growth factor or a growth factor
`receptor. These genes include c-sis, which is homologous to the
`9 JANUARY1987
`
`transforming gene of the simian sarcoma virus andis the 6 chain of
`platelet-derived growth factor (PDGF) (16, 17); c-fins, which is
`homologous to the transforming gene ofthe feline sarcoma virus
`and is closely related to the macrophage colony-stimulating factor
`receptor (CSF-1R)
`(18); and c-erbB, which encodes the EGF
`receptor (EGFR) and is highly homologous to the transforming
`gene of the avian erythroblastosis virus (19). The two receptor-
`related proto-oncogenes, c-fims and c-erbB, are members of the
`tyrosine-specific protein kinase family to which many proto-onco-
`genes belong.
`Recently, a novel transforming gene was identified as a result of
`transfection studies with DNA from chemically induced rat neu-
`roglioblastomas (20). This gene, called ne, was shown to be related
`to, bur distinct from, the c-er6B proto-oncogene (21). By means of
`v-erbB and human EGFRas probes to screen human genomic and
`complementary DNA (cDNA)libraries, two other groups indepen-
`dently isolated human er#B—related genes that they called HER-2
`(22) and c-erbB-2 (23). Subsequent sequence analysis and chromo-
`somal mapping studies revealed all three genes (mex, c-erbB-2, and
`HER-2)to be the same (22, 24, 25). A fourth group,also using v-
`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-
`fold (26).
`This gene, which wewill call HER-2/neu, encodes a new member
`of the tyrosine kinase family; and is closely related to, but distinct
`from, the EGFR gene (22). HER-2/nex differs from EGFRin thatit
`is found on band q2] of chromosome17 (22, 24, 25), as compared
`to band pll—p13 of chromosome 7, where the EGFR geneis
`located (27). Also, the HER-2/nex gene generates a messenger
`RNA (mRNA)of4.8 kb (22), which differs from the 5.8- and 10-
`kb transcripts for the EGFR gene (28). Finally, the protein encoded
`by the HER-2/neu gene is 185,000 daltons (21), as compared to the
`170,000-dalton protein encoded by the EGFR gene. Conversely, on
`the basis of sequence data, HER-2/neu is moreclosely related to the
`EGER gene than to other members of the tyrosine kinase family
`(22). Like the EGFR protein, HER-2/nen has an extracellular
`domain, a transmembrane domain that includes two cysteine-rich
`repeat clusters, and an intracellular kinase domain (21), indicating
`
`and W.J. Levin are in the Division of Hematology-
`D. J. Slamon, S. G. Wong,
`Medicine and Jonsson Com
`ive Cancer Center,
`Oncology, Department
`UCLA
`] of Medicine, Los Angeles, CA 90024. G.
`M. Clark and W. L. McGuire
`are in the Division of Oncology,
`of Medicine, University ofTexas Health
`Science Center at San Antonio,
`San Antonio, TX 78284. A.
`Ulrich is in the
`Department of Molecular Biology, Genentech, Inc., South San Francisco, CA 94080.
`
`*To whom correspondence should be addressed.
`
`ARTICLES 177
`
`(cid:43)(cid:82)(cid:86)(cid:83)(cid:76)(cid:85)(cid:68)(cid:3)(cid:89)(cid:17)(cid:3)(cid:42)(cid:72)(cid:81)(cid:72)(cid:81)(cid:87)(cid:72)(cid:70)(cid:75)(cid:3)
`Hospira v. Genentech
`(cid:44)(cid:51)(cid:53)(cid:21)(cid:19)(cid:20)(cid:26)(cid:16)(cid:19)(cid:19)(cid:27)(cid:19)(cid:24)(cid:3)
`IPR2017-00805
`(cid:42)(cid:72)(cid:81)(cid:72)(cid:81)(cid:87)(cid:72)(cid:70)(cid:75)(cid:3)(cid:40)(cid:91)(cid:75)(cid:76)(cid:69)(cid:76)(cid:87)(cid:3)(cid:21)(cid:19)(cid:23)(cid:22)
`Genentech Exhibit 2043
`
`

`

`Factor*
`
`Single
`copy
`
`ER+
`ER-
`PgR+
`PgR-
`
`<2
`2-5
`>5
`Unknown
`
`53
`31
`42
`42
`
`13
`34
`17
`20
`
`>20
`copies
`
`yi,
`
`Pt
`
`1
`4
`2
`3
`
`0
`1
`2
`2
`
`0.99
`
`0.85
`
`0.82
`
`65
`38
`52
`51
`
`15
`41
`22
`25
`
`Table 1. Association between HER-2/nex amplification and disease parame-
`that it too is likely to be a cellular receptor for an as yet unidentified
`ters in 103 breast tumors.
`ligand.
`
`As a result of the published data showing amplification of HER-
`Number of tumors
`2/neu in a human mammary carcinomacell line, and as part of an
`2105
`5t020
`ongoing survey in our laboratory of proto-oncogene abnormalities
`copies
`copies
`in human tumors, we evaluated alterations of the HER-2/neu gene
`in a large series of human primary breast cancers. Our results show
`Hormonal mee status
`2
`that amplification of this gene occurs relatively frequently in breast
`1
`2
`cancer, and thatit is associated with disease relapse and overall
`2
`6
`patient survival.
`1
`5
`Factors that are known to be importantin the prognosis of breast
`Tumor size (centimeters)
`malignancies in individual patients include: size of the primary
`1
`1
`tumor, stage of disease at diagnosis, hormonal receptor status, and
`1
`5
`number of axillary lymph nodes involved with disease (positive
`1
`2
`0
`3
`nodes) (29). The current study, which was conducted in twoparts,
`involved the evaluation oftissue from 189 separate breast malignan-
`Age at diagnosis (years)
`
`=50 25=0.8321 1 2 1
`
`
`
`
`cies that were part ofa breast cancer study ongoing at the University
`2
`7
`>50
`52
`4
`65
`of Texas, San Antonio. This cohort of tumors was of interest
`Unknown
`ll
`0
`2
`0
`13
`because considerable information was available on the majority of
`Number ofpositive lymph nodes
`the specimens including size of the primary tumor, estrogen recep-
`0 34=0.1130 0 3 1
`
`
`
`
`
`tor status, progesteronereceptorstatus, age ofpatient, disease stage,
`0
`1
`1-3
`20
`1
`22
`and status of the axillary lymph nodes.
`>3
`17
`2
`4
`2
`25
`Unknown
`17
`1
`3
`1
`22
`In theinitial survey, tissue from 103 primary breast cancers was
`evaluated for alterations in the HER-2/nex gene. DNA from
`as described
`ER,
`+ and — refer
`(oepnen:oras >3 fmol eich ‘Perniligramofproscia. PgR,
`individual tumors was prepared as described (30), digested with
`i,t tinngtad + and— referstothe
`Pacetekbanetvaaas
`Eco RI, and subjected to Southern blot analysis with a **P-labeled
`oemaaan— forcorrelationofHER-2/newamplifi-
`HER-2/neu-1 probe, which is known to detect a 13-kb hybridizing
`band in human DNA (22). Examples of tumors from the initial
`survey are shown in Fig. 1. Ofthe 103 samples examined, 19 (18%)
`showed evidence of HER-2/nex gene amplification. The degree of
`amplification in individual cases was determined by dilution analysis
`(Fig. 2A), as well as soft laser densitometry scanning. To determine
`that the amount of DNA loaded in each lane was equivalent, all
`filters were washed and rehybridized with a >*P-labeled arginase
`gene probe (31). This probe identifies a 15-kb hybridizing band on
`Eco RI-digested human DNA, and was selected as a control
`because it more appropriately assesses the relative amount and
`
`transfer of high molecular weight species than a probe hybridizing
`with low molecular weight species, which transfer more readily on
`Southern blotting. All
`lanes were shown to contain equivalent
`amounts of high molecular weight DNA (Fig. 2B). Individual
`tumors were assigned to groups containing a single copy, 2 to 5
`copies, 5 to 20 copies, and greater than 20 copies ofthe HER-2/neu
`gene (Fig. 1). Assignment oftumors to the various groups was done
`
`6:7 .8::0..*
`
`
`
`41 42 43 44 45 46 47 48 49 50
`
`ee
`
`wn 1213 4
`15 16 17 18 19 20 - 21 22 23 24 2526 27 28 29
`1.2.3.4. 5.
`Fig. 1. Analysis of alterations of the HER-2/neu
`ragglt + !
`Her
`“a a Haw
`ey a
`in human breast cancer. Shown are 79 ofthe==jee
`89 breast tumors used in this
`. Tumors
`e
`with a single copy ofHER-2/neu: 3, 4, 10 to 15,
`20, 23 to 25,ZFto 29, 31, 38, 42 to 46, 48, 49,
`52, 55, 61, 65, 66, 71, 72,and 74. Tumors with
`two tofive
`copies of HER-2/neu: 1, 2, 5, 7,9, 16,
`17, 19, 21, 22, 32, 35, 36, 47, 50, 54, 56 to 58,
`60, 62, 70, and 75 to 77. Tumors with 5 to 20
`ies of HER-2/neu: 6, 8, 26, 34, 37, 39 to 41,
`
`51, 53, 63, 64, 67, 69, 73, and 79. Tumors with
`more than20 copies of HER-2/nex: 18, 30, 33, =
`59, 68, and 78. Examples oftumors 77 to 79have
`Pan
`ts in the HER-2/new gene. DNA
`was extracted from tissues and
`with Eco
`RJ as described (30). A total of 12 yg ofEco RI-
`digested DNA was loaded onto 0.8% agarose
`and transferred
`gels, separated by
`onto nylon filter papers (Biodyne) (30). Alll filters
`67 68 68 70 71 72 73 74 75 76
`78
`
`were in a vacuum ovenfor3 hours at 80°C,
`Tg
`oe KD
`ad
`prehybridized in 5x SSC (standard saline citrate)
`containing 50% formamide, 10% dextran
`0.1% SDS, denatured salmon s
`DNA(1 mg/
`ml), and 4x Denhardts solution for 12 hours,
`then hybridized in the same solution eae
`=Plabeled nick-translated ee probe Lapl
`DNA; 2 x 10°
`onere
`specific activity aexoH microgram
`at 42°C for 48 aiaaee
`filters under the following conditions in succes-
`178
`
`51 ae SS eS ee ee €2 63 6&4 6566
`
`‘
`sion: 2X SSC for 20 minutes at room
`ture; two washes of 30 minutes each in 2x SSC,
`0.1% SDS at 65°C; one wash of 30 minutes in
`
`Birer or:
`'
`OB w:s:
`-18
`
`0.5x SSC, 0.1% SDSat 65°C.Filters were then
`exposed to XAR-5 x-ray film (Kodak) for autora-
`diography.
`
`SCIENCE, VOL. 235
`
`

`

`Fig. 2. (A) Example of dilutional analysis to assess
`of HER-2/neu
`gene amplification. Lanes a, g, k, and p were loaded
`12 yg ofEco RI-
`with
`digested breast tumor D.DNAfacade DNA fim tuene AL (Fig. 1), which
`ts a tumor with a single copy of the HER/2-new gene. Lane g is
`DNA from tumor 33, which
`a tumor with >20 copies of the
`HER-2/neu gene. Lanes b to f are serial dilutions (1:100, 1:20, 1:10, 1:5,
`and 1:2,
`) of the DNA sample in lane g. Lane k is DNA from
`tumor 35 (Fig. 1), which represents a tumor containing two to five copies of
`See Lanes h to j are serial dilutions (1:10, 1:5, and 1:2,
`respectively) of
`DNAsample in lane k. Lane p isDNA fromtumor 34
`(Fig. 1), which represents a tumor with 5 to 20
`copies of the HER-2/neu
`Lanes | to o are serial dilutions (1:20, 1:10, 1:5, ae2,respective
`F)ofthe DNAsample in lane p. The filterwas
`with
`a *?P-labeled HER‘?probe as in Fig, 1. (B)
`pro!
`byteidicadon to demonstrate thaterySoamounts oftumor DNAwere
`loaded into cach lane. Rehybridization of filter containing lanes 30 to 40
`(Fig. 1). The filter was first stri
`of label by
`in a buffer made up
`of 50% formamide, 3x SSC, and 0.1% SDS at 65°C for 20 minutes,
`following by three successive washes of5 minutes each in 0.1 SSC at room
`castrc removalofall radioactieeprobe,terchybridized as in og 1 with a
`t
`rature. Filters were
`tt on XAR-5 film (Kodak) to
`32P.labeled human arginase gene probe (31).
`
`in a blinded fashion, in that they were made without knowledge of
`disease
`. Analysis ofthe data for association between gene
`amplification and a number of disease parameters was then per-
`formed.
`Of103 tumors evaluated in theinitial survey, there was essentially
`no correlation between gene amplification and estrogen receptor
`status, progesterone receptor status, size of tumors, or age at
`diagnosis (Table 1). However, when analysis was performed for
`association between HER-2/nex amplification and number of posi-
`tive lymph nodes,a trend was noted. This analysis showed that 4/34
`(11%)of patients with no involved nodes, 2/20 (10%) with 1 to 3
`involved nodes, and 8/25 (32%) with >3 involved nodes had gene
`amplification (P = 0.11). Ifthese 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/new gene amplifica-
`tion in patients with >3 axillary lymph nodes involved with disease.
`A multivariate regression analysis to correlate HER-2/nex amplifica-
`tion with various disease parameters identified the number of
`positive nodes as the only significant factor, either alone or in
`combination, to correlate with amplification.
`This initial study indicated that it might be possible to discrimi-
`nate among node-positive patients on the basis of HER-2/neu gene
`
`
`
`amplification. It is well known that the numberofpositive nodes is
`the best prognostic factor for disease recurrence and survival in
`patients with breast cancer (29). Given the correlation between
`number of nodes positive and HER-2/neu amplification, one might
`predict
`that amplification of this gene might also have some
`prognostic value. No long-term follow-up data, however, were
`available on the 103 patients analyzed in the initial study. For this
`reason, a second study was conducted on 100 breast cancer samples
`from patients with positive axillary lymph nodes.All ofthe informa-
`tion available for the first group of 103 patients was available for
`these patients. In addition, relapse and survival information was
`available, since these cases had a median follow-up of 46 months
`(range 24 to 86 months). Ofthese 100 samples, 86 yielded sufficient
`DNAfor study. Amplification of the HER-2/nex gene was mea-
`sured as in the initial survey, and examples oftumors from this study
`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 relationships were
`observed. In agreement with the preliminary survey, there was an
`association between number of involved lymph nodes and HER-2/
`neu amplification (Table 2). In addition,
`the presence of gene
`amplification was correlated with estrogen receptor status and size
`of primary tumor (Table 2). Together, these two surveys yielded
`data on 189 patients and the association of HER-2/nex amplifica-
`tion with various disease parameters in the combined group is
`shown in Table 3.
`While these correlations were of interest, the strong relationship
`
`1.0
`
`
`Not amplified (n =52)
`
`
`Amplified (n =11)
`
`>5 copies
`
`
`
`
`Not amplified (n =52)
`
`
`Amplified (7 =11)
`>5 copies
`
`
`
`12
`
`24
`
`36
`
`48
`
`60
`
`72
`
`84
`
`ARTICLES
`
`179
`
`0
`
`0
`
`12
`
`24
`
`36
`
`48
`
`60
`
`=
`0
`72
`84
`Time (months)
`
`Disease-free
`
`Overall
`
`Fig. 3. Actuarial curve for relapse in (A) node-
`positive patientswith no
`tion versus
`node-positiveHERDinewith any amplification (>2
`copies) of
`2/neu and (C) node-positivepa-
`tients with no ampli
`versus node-positive
`patients with greater than 5
`ofHER-2/neu.
`Actuarial curve for overall survival in (B) node-
`a ents with no
`versus
`pre
`amplification (>2
`copie)sofHER-2/neu and (D)
`“positivepa-
`positive
`patients with any am
`tients with no amplification versus node-positive
`patients with greater than 5 copies ofHER-2/neu.
`Actuarial curves for both relapse and overall sur-
`vival were computed by the method of Kaplan
`and Meier (44) and compared by the log rank test
`(42-44).
`
`9 JANUARY 1987
`
`
`
`survivalprobability ° o
`
`2>
`
`2cS
`
`0
`
`~ °
`
`2@
`
`
`
`survivalprobability ocso>oa
`
`°iy
`
`Not amplified (n =52)
`
`
`
`Not amplified (n=52)
`
`Amplitied (n =34)
`
`
`Amplified (n=34)
`
`
`
`

`

`51
`
`52 53
`
`54
`
`55
`
`56
`
`57 58
`
`59
`
`60
`
`between HER-2/nex amplification and nodal status (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 ofthe disease, and 29
`had died at the time of the analyses. Median times to relapse and
`Fig. 4. Example of rehybridization offilter with human EGFRprobe.Filters
`death were 62 months and 69 months, respectively. The median
`were stripped
`as in Fig. 2B, and hybridized with *?P-labeled human EGFR
`follow-up time for patients still alive was 47 months, ranging from
`(28), as in Fig. 1. Shown are the lower molecular weight bands
`24 to 86 months. A total of 71 of the 86 patients (83%) received
`with **P-labeled EGFR
`in filter.
`ining
`lanes 51 to 66
`idized
`
`
`some form of therapy after mastectomy: adjuvant systemic therapy (Fig. 1). The bands fromtopto bottom are 2.8, 2.2, and 1.8 kb, respectively.
`Lane 52 is an exampleofa tumor showing marked amplification (>50
`alone, 47%; adjuvant systemic therapy plus local radiation, 19%;
`copies) of the EGFR gene.
`and local radiation alone, 17%. A strong and highly statistically
`significant correlation was found between the 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 analyses to other parameters, amplification of HER-2/nex
`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/nex amplification
`and relapse and survival can be illustrated graphically in actuarial
`survival curves (Fig. 3, A to 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 tumors
`(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/new 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 numbercorrelating with a worse prognosis has
`also been seen in evaluations of N-myc gene amplification in human
`neuroblastomas (32).
`To determineif amplification of HER-2/neu was independent of
`other known prognostic factors in predicting disease behavior,
`multivariate survival analyses were performed on the 86 node-
`positive cases. Amplification of the gene continued to be a strong
`prognostic factor, providing additional and independentpredictive
`information on both time to relapse and overall survival in these
`
`61
`
`62
`
`63
`
`64
`
`65
`
`66
`
`
`PoghiaeeaeseS”
`
`patients, even when other prognostic factors were taken into
`account (Table 4).
`Rearrangement ofthe HER-2/neu gene was rare. Ofthe total 189
`tumors evaluated, three showed evidence of rearrangement, and in
`two ofthe three cases, the rearrangement was identical (Fig. 1, 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 to
`attempt statistical correlations.
`To determine whether the phenomenonofamplification ofHER-
`2/neu in breast cancer extended to related growth factor receptors,
`all filters were analyzed with the EGFR probe (Fig. 4). Amplifica-
`tion of the EGFR gene was found in 4/189 (2%) ofthe cases, and
`rearrangement of the EGFR gene was found in one of those four
`cases. The incidence of EGFR amplification and rearrangement was
`too small to attempt statistical correlation. Comparison of HER-2/
`new amplification (53/189 or 28%) with that of the EGFR gene
`reveals the incidence of the former to be 14 times greater than that
`ofthelatter, indicating that the phenomenon of gene amplification
`is not a general one for a related tyrosine kinase—specific receptorin
`human breast cancer. Moreover, studies examining alterations of
`two other tyrosine kinase—specific proto-oncogenes, abl and fés, in
`breast cancer did not show amplification of these genes (33).
`Alterations ofnon—tyrosine kinase—related proto-oncogenes in these
`
`Table 3. Association between HER-2/neu amplification and disease parame-
`ters in combined surveys (189 patients).
`
`
`Table 2. Association between HER-2/nex amplification and disease parame-
`ters in 86 breast tumors from node-positive patients.
`
`Factor*
`
`Single
`copy
`
`Factor*
`
`Single
`copy
`
`ER+
`ER-
`PgR+
`PgR-
`
`s2
`2-5
`>5
`
`350
`>50
`
`38
`14
`31
`21
`
`18
`28
`6
`
`16
`36
`
`>20
`5to20
`2to5
`copies
`copies
`copies
`Hormonal receptor status
`21
`5
`2
`4
`18
`4
`5
`5
`Tumor size (centimeters)
`8
`3
`12
`2
`3
`4
`
`1
`1
`1
`1
`
`0
`l
`l
`
`Age at diagnosis (years)
`12
`6
`ll
`3
`
`1
`1
`
`Total
`
`Pt
`
`65
`21
`54
`32
`
`29
`43
`14
`
`35
`51
`
`0.05
`
`0.14
`
`0.09
`
`0.06
`
`43
`0
`5
`7
`31
`1-3
`>3
`21
`16
`4
`2
`43
`
`
`0.06
`
`ER+
`ER-
`
`PgR+
`PgR-
`
`=2
`2-5
`>5
`Unknown
`
`350
`>50
`Unknown
`
`91
`45
`
`73
`63
`
`31
`62
`23
`20
`
`37
`88
`ll
`
`>20
`5 to 20
`2to5
`z
`4
`copies
`copies
`copies
`Hormonal receptor status
`23
`14
`3
`6
`
`2
`5
`
`10
`20
`10
`6
`Tumor size (centimeters)
`9
`+
`13
`7
`4
`6
`0
`3
`Age at diagnosis (years)
`13
`8
`13
`10
`0
`2
`
`3
`4
`
`0
`2
`3
`2
`
`2
`5
`0
`
`Total
`
`130
`59
`
`106
`83
`
`44
`84
`36
`25
`
`60
`116
`13
`
`Pt
`
`0.05
`
`0.06
`
`0.19
`
`0.11
`
`*ER and PgR are as described in Table 1.
`tStatistical analyses for correlation of
`eeeee
`*ER and PgR are as described in Table 1.
`Statistical analyses for correlation of
`+ HER-2/neu amplification with various disease
`were
`by the x?
`
`test. P values were after combiningthe5 to 20 and >20 cases, since there
`were so few samples in the >20 group.
`test. P values were computed after combining the cases with 5 to 20 and >20 copies.
`180
`
`SCIENCE, VOL. 235
`
`Number ofpositive lymph nodes
`34
`0
`3
`1
`30
`0
`65
`7
`6
`1
`51
`1-3
`68
`18
`8
`4
`38
`>3
`Unknown
`17
`1
`3
`1
`22
`
`
`0.002
`
`

`

`Table 4. Univariate and multivariate analyses comparing disease-free survival (relapse) and overall survival to prognostic factors in node-positive patients.eeeeeeeeeeeeeeeeeenereeereeceeecS
`Multivariate*
`
`Univariate (P)
`
`Factor
`
`Survival
`
`Relapse
`
`Survival
`
`Relapse
`
`Number ofpositive nodes
`0.0001
`HER-2/neu
`0.0011
`Log (PgR)
`0.05
`Tumorsize
`0.06
`Log (ER)
`0.15
`Age
`0.22
`GaaatResultsortshownas Pircprosion oocficient & SIE).
`regression model was used to evaluate the predictive power ofvarious combinations andinteractions ofprognostic factors in a multivariate manner
`
`0.0003
`0.02
`
`(0.0938 + 0.0256)
`(0.0872 + 0.0388)
`
`0.001 (0.0849 + 0.0266)
`0.001 (0.1378 + 0.0425)
`
`0.03
`
`(-0.5158 + 0.2414)
`
`0.0002
`<0.0001
`0.05
`0.06
`0.10
`0.61
`
`REFERENCES AND NOTES
`
`tumors have been examined. In a survey of 121 primary breast
`receptors in the biology of breast cancer is well established (29, 39,
`malignancies, amplification of the c-mye gene was found in 38
`40). It is easy to speculate that a gene encoding a putative growth
`(32%) (34). Attempts to correlate c-myc gene amplification with
`factor receptor, when expressed in inappropriate amounts, may give
`stage ofdisease, hormonal receptorstatus, histopathologic grade, or
`a growth advantage to the cells expressing it. Alternatively, alter-
`axillary node metastases showed no association. There wasastatisti-
`ation in the gene product itself may lead toacritical change in the
`cally significant association between c-myc amplification and age at
`receptor protein. A single point mutation in the transmembrane
`diagnosis >50years in a group of95 ofthese patients (34). Data on
`domain of the protein encoded by the rat new oncogene appears to
`relapse and survival were not presented in this study; however, there
`be all that is necessary for the gene to gain transforming ability (41).
`was no correlation between c-myc amplification and nodal status to
`Whether this or a similar alteration is found in the amplified HER-
`indicate an association with disease behavior.
`2/neu gene in human breast cancer will require sequence analysis of
`Theexact role of various proto-oncogenes in the pathogenesis of
`the homologous region in the amplified human gene. In addition,
`human malignancies remains unclear. Oneline ofevidence implicat-
`studies evaluating the expression of this gene at the RNA and/or
`ing abnormalities of these genes in human disease is association of
`protein level will prove important in determining if HER-2/nex
`their amplification with tumor progression in specific cancers. The
`amplification results in an expected increased gene expression. The
`N-myc gene is frequently amplified in human neuroblastomas and
`question of amplification of HER-2/neu in metastatic as compared
`neuroblastoma cell lines (35, 36). Studies on the N-myc proto-
`to primary lesions in a given patient is important. The current study
`oncogene were the first to show a direct association between
`utilized only primary breast tumors for analyses. It would be of
`abnormalities in a proto-oncogene andclinical behavior of a human
`interest to determine if HER-2/neu copy numberis altered as the
`tumor. N-myc amplification and expression correlate both with stage
`tumor metastasizes. A recent study evaluating N-myc copy number
`of disease and overall survival in patients with neuroblastoma (10,
`in human small cell carcinoma of the lung showed no difference
`32, 37). Moreover the greater the N-myc gene copy number, the
`between primary and metastatic lesions (11).
`worse the patient prognosis for all stages of the disease (32). Taken
`Theinitial survey from the current study showed that 15% of
`together, these data indicate a role for the N-myc gene in the
`breast cancer patients with stage I disease (node-negative) have
`pathogenesis of neuroblastoma (32).
`HER-2/neu amplification. Unfortunately, no long-term follow-up
`Neuroblastomais a relatively rare disease with an incidence ofone
`data were available for these patients. This stage I setting may be an
`per 125,000 children. Carcinoma of the breast, however,
`is a
`additional group in which HER-2/new measurements will have an
`common malignancy affecting one ofevery 13 womenin the United
`impact in predicting biologic behavior of the tumor, andas a result,
`States. There are 119,000 new cases per year, and approximately
`in design of treatment strategy. Finally, if the HER-2/neu gene
`40,000 women will die of the disease in 1986 (38). Current
`product functions as a growth factor receptorthatplaysa role in the
`treatment decisions for individual patients are frequently based on
`pathogenesis of breast cancer,
`identification of its ligand and
`specific prognostic parameters. The major prognostic factors for
`developmentof specific antagonists could have important therapeu-
`breast cancer include presence’ or absence of tumorin the axillary
`tic implications.
`nodes, size of the primary tumor, and presence or absence of
`hormonal receptors (29). The current study indicates that amplifica-
`tion of the HER-2/nex geneis a significant predictor of both overall
`survival and time to relapse in node-positive patients with breast
`cancer. Amplification of the gene retains its prognostic significance
`in multivariate analysis, even when adjustments are made for other
`known prognostic factors. Moreover, amplification of HER-2/neu
`has greater prognostic value than most currently used prognostic
`factors,
`including progesterone and estrogen receptors, and is
`equivalent to and independent of the best known prognosticator—
`number ofpositive lymph nodes. Finally, the degree of HER-2/neu
`amplification appears to have an effect on survival, with greater copy
`number being associated with a worse prognosis (Fig. 3, C and D).
`A similar phenomenonhas been observed for N-myc gene amplifica-
`tion in human neuroblastoma (32).
`The potential role of HER-2/neu in the pathogenesis of breast
`cancer is unknown. Like N-myc, the correlation of HER-2/neu
`amplification with disease progression indicates it may be an
`important gene in the disease process. The role of other cell
`
`. Bishop, Annu. Rev. Biochem. 52, 301 (1983).
`ne
`us, Annu. Rev. Genet. 18, 553 (1984).
`
`-Land,L.EefsBarada]RA.VeniNature(London) 304,596(1983).
`M.Bishop,iid 316, 160 (1985).
`“S.
`ibid. 290, 475 (1981).
`“N56 1984
`(1984).
`. D. Erisman e¢3Mek Cell.Biol. 5, 1969 (1985).
`. M. Croce and G. Klein, Ses. Am. 252, 54 (March 1985).
`M.Brodeur et al., Science 224, 1121 (1984).
`., ibid. 233, 461 (1986).
`>Call833,387, 1983),
`and
`, Nature (London) 311, 433 (1984).
`> sbid. 369,‘640 (1982).
`J. Slamon and M. j. Cline, Proc. Nati.Acad. Sci. USA. 81, 7141 (1984).
`. D. Waterfield et al., Nature (London) 304, 35 (1983).
`F. Doolittle, M.M.W. Hunkapiller, L. E. Hood, Science 221, 275 (1983).
`. J. Sherr et al., Cell 41, 665 (1985).
`etal., Nature (London)307, 521 (1984).
`:
`. Shih, L. Padhy, M. Murray,
`Weinberg, ibid, 290, 261 (1981).
`21. A. L. Schechter et al., ibid. 3 2,03(1984).
`22. L. Coussens et al., Science 230, 1132 (1985).-
`23. K. Semba, N. Kamata, K. Toyoshima, T. Yamamoto, Proc. Natl. Acad. Sci. U.S.A.
`82, 6497 (1985).
`
`$0.90UTwaSzmmm Ses) no
`SnoremPoEEO™: & 8S
`ele
`
`9 JANUARY1987
`
`ARTICLES 181
`
`

`

`=; | E
`
`PZ
`posPS
`PRPZoOS
`cea
`
`Segaloff, Science 189, 726
`
`).
`
`, Cell 45, 649 (1986).
`
`1978
`cin!
`1972).
`Rev. 43, 45 (1975).
`ier, J. Am. Stat. Assoc. 53, 457 (1958).
`U.S. Public Health Service grants CA 36827 and CA 30195 and a
`triton Biosciences, Inc. We thank D. Keith, L. Gordon, and W.Aft for
`technical assistance and Dr. S. Cedarbaum for providing the human arginase
`
`AgrOryny>
`
`[i
`
`FB
`
`g
`
`Fsi
`SSRRRS
`Ht et al., Science 229, 976 (1985).
`et al., Mol. Cell. Biol. 6, 955 (1985).
`A. Aaronson, Science 229, 974 (1985).
`weBT
`M.
`Cytogenet. Cell. Genet. 37, 71 (1984).
`f
`lature (London)
`309, 418 (1984).
`2
`mEaf”g
`. Engl. J. Med. 309, 1343 (1983).
`Manual,
`itsch, J. Sambrook, Molecular
`ing: A
`5 ¢
`fi
`Laboratory, Cold Spring Harbor,
`NY, 1982), pp. 282~285.
`ikes, W. W. Grody, R. M.
`S. B. Cedarbaum,Biochem. Biophys. Res.
`.. in press
`oe oO
`RSS ZO
`. Seeger et al., N. Engl. J. Med. 313, 1111 (1985).
`°
`—
`Yokota et al., Science
`231, 261 (1986).
`. Escot ef al., Proc. Nati. Acad. Sci. U.S.A. 83, 4834 (1986).
`. Schwab et al., Nature (London) 305, 245 (1983).
`
` Researeh ‘Wontocie
`
`The Atomic Structure of Mengo Virus
`at 3.0 A Resolution
`
`MING Luo, GERRIT VRIEND*, GrEeG KAMER, IwoNA Minor, EDwaRD ARNOLD,
`MICHAEL G. ROSSMANN, ULRIKE BOEGE, DoUGLAS G. SCRABA, GREG M. DUKE,
`ANN C. PALMENBERG
`
`The structure of Mengo virus, a representative member of
`the cardio picornaviruses, is substantially different from the
`structures of rhino- and polioviruses. The structure of
`Mengo virus was solved with the use of human rhinovirus
`14 as an 8 A resolution structural approximation. Phase
`informati:ion was then extendedto 3 A resolution by use of
`
`without the use of the isomorphous replacement tech-
`nique. Although the organization of the major capsid
`proteins VP1, VP2, and VP3 ofMengo virus is essentially
`the same as in rhino- and polioviruses, large insertions
`
`and deletions, mostly in VP1, radically alter the surface
`features. In particular, the putative receptor binding
`“canyon” ofhuman rhinovirus 14 becomes a deep “pit” in
`Mengo virus because of polypeptide insertions in VP]
`that fill part of the canyon. The minor capsid peptide,
`VP4,
`is completely internal in Mengo virus, but its
`association with the other capsid proteins is substantially
`different from that in rhino- or poliovirus. However, its
`carboxyl terminus is located at a position similar to that in
`human rhinovirus 14 and
`poli