CANCER RESEARCH57. 536()-5368. December 1. 19971
`
`Role of 06—Methylguanine-DNAMethyltransferase in the Resistance of Pancreatic
`Tumors to DNA Alkylating Agents
`
`Demetrius M. Kokkinakis,' Mansoor M. Ahmed, Ruby Delgado, Mushdaq M. Fruitwala, Mohammed Mohiuddin, and
`Jorge Albores-Saavedra
`Departmeizis of Neurolog@ (D. M. K.J and Pathology (R. D.. J. A.S./. The Universi@ of Texas Southwestern Medical Center at Dallas, Dallas Texas 75235-9036, and Department
`of Radiation Medicine. The Uni@'ersitvof Kentucky. Lsxing:on. Kentucky 40536 fM. A.. M. F.. M. M.J
`
`ABSTRACT
`
`Pancreatic adenocarcinomas rarely respond to radiation or chemother
`apy, indicating that a large percentage of these tumors possess complex
`mechanisms of resistance. The failure of alkylating agents, such as
`carmustine
`I1@bis(2-chloroethyl)-1-altrosourea;
`BCNUJ,
`lomustine
`Ll-(2-chloroethyl)-3-cyclohexyl-l-nitrosourea; CCNUI, and streptozoto
`cm, to yield consistent
`therapeutic results further
`suggests that one of
`these mechanisms may be the high expression of O'-methylguanine-DNA
`methyltransferase (MGMT). All 12 human pancreatic ductal adenocarci
`nomas
`assayed
`for MGMT
`activity
`showed
`unusually
`high levels,
`implying
`that these malignancies are efficient in repairing genotoxic 06-alkylgua
`nine lesions induced by methylating (streptozotocin) and 2-chloroethylat
`ing (BCNU and CCNU) chemotherapeutic genotoxic agents. Immunohis
`tochemical analysis of an additional 15 pancreatic tumors showed that
`high levels of MGMT protein reside in the nucleus and the cytoplasm of
`malignant
`cells. Both
`nuclear
`and
`cytoplasmic
`staining
`were
`absent
`in
`hyperplastic duct epithelium, but staining was invariably present in mod
`erate to highly dysplastic
`foci and especially strong in invasive components
`of the tumor. With the exception of lymphocytes that were MGMT
`positive,
`acinar,
`ductal,
`and
`islet
`cells did not
`stain
`for MGMT
`in histo
`logically normal pancreata. These data indicate that MGMT activity is
`up-regulated in dysplastic epithelium, and its expression increases during
`tumor progression, reaching the highest levels in the invasive components
`of the tumor. Resistance of pancreatic tumor cells to alkylating agents was
`verified
`with
`four
`pancreatic
`tumor
`cell
`lines.
`CAPAN-2,
`CFPAC-1,
`PANC-1, and MIAPaCa-2, having MGMT levels of 1800, 987, 700, and
`880 fmol/mg protein,
`respectively, were resistant
`to BCNU, but
`their
`resistance
`declined
`sharply
`following
`pretreatment
`with the MGMT in
`hibitor O'-benzylguanine (O'-BG). On the other hand, PANC-1 and MI
`APaCa-2 could not be eradicated with N-methylnitrosourea (MNU) at
`concentrations as high as 2 mM,even when pretreated with O@-BG.These
`two lines were shown to be modified genetically in microsatellite sequences
`by MNU and are believed to have a defective mismatch repair system,
`which may explain their resistance to methylating agents. Failure of
`pancreatic tumors to respond to nitrosoureas is related to high levels of
`MGMT expression and in some cases to genomic instability. However,
`these tumors can be sensitized to chloroethylating drugs and eradicated
`following the elimination of MGMT activity by O@-BGor homologous
`MGMT inhibitors.
`
`INTRODUCTION
`
`In spite of its low incidence of only 0.01% (26,300 new cases
`diagnosed in 1996), pancreatic cancer remains the fifth leading cause
`of death from cancer in North America (1 , 2). A cure is only possible
`by surgical
`removal of the tumor, but in practice, even this is rarely
`achieved because most tumors are no longer resectable at the time of
`diagnosis (3, 4). Fifteen to 20% of patients with carcinoma of the head
`of the pancreas can be resected, with about 14—33%of them surviving
`more than 5 years (5). Theoretically,
`the rate of survival could be
`
`of high-risk groups.
`surveillance
`improved further by intensifying
`of the etiology of pancreatic cancer
`However,
`lack of understanding
`(5) obscuresidentificationof suchgroupsandfurtherimpedesthe
`development of preventive measures. A rising incidence of pancreatic
`cancer, at a rate of 1% per year since 1937, combined with our failure
`to prevent, diagnose, or treat pancreatic cancer successfully, makes a
`powerful
`case for
`further
`understanding
`of the biology
`of this disease
`as related
`to the development
`of more potent
`drugs
`for
`its treatment
`(6—10).
`is usually a
`the time of diagnosis,
`seen at
`cancer,
`Pancreatic
`disseminated disease (1 1—13),and therefore, surgery alone is unlikely
`to increase survival
`rates in the absence of standard and well-tested
`adjuvant
`therapies. Present and future strategies for treatment
`should
`include the addition of adjuvant modalities or the development of new
`agents with significantly
`greater antitumor
`activities. Thus far,
`the
`results obtained with cytotoxic chemotherapy
`have been disappoint
`ing. More
`than 40 Phase
`II studies
`of many
`new agents
`and combi
`nations
`have been reported
`in the last 15 years
`(14,
`15). Among
`the
`more vigorously tested drugs are streptozotocin
`in combination with
`mitomycin;
`cisplatin in combination with Ara-C; 5-fluorouracil
`in
`various combinations with CCNU2; and methyl-CCNU with strepto
`zotocin, mitomycin C, doxorubicin,
`cisplatin, methotrexate,
`cyclo
`phosphamide,
`and vincristine (16—22).Treatment with 5-fluorouracil,
`doxorubicin,
`and mitomycin has shown a response varying between
`10 and 34%, with a median survival of 10 months. All other treat
`ments elicit a response lower than 20%,
`thus failing the standard for
`further
`testing. A newer drug, gemcitabine
`(2',2'-difluorodeoxy
`cytidine), with advanced cytotoxic properties
`in the laboratory, has
`induced a partial
`response in 5 of 44 patients ( 11%), with a median
`survival of 5.6 months and with 32% of the patients alive at 1 year
`(7, 23). The failure of a large number of chemotherapeutic
`agents
`against
`pancreatic
`tumors
`is consistent
`with the presence
`of a multi
`drug-resistant
`phenotype
`(not necessarily conferred by the p21 gly
`coprotein)
`in a large fraction of these tumors, but also it may be due
`to the relatively
`poor
`systemic
`access
`of several
`drugs
`to this organ
`(24, 25).
`to
`tumors
`of pancreatic
`is the resistance
`interest
`Of particular
`nitrosoureas
`and other genotoxic agents. This could
`chloroethylating
`be auributed
`to effective
`repair of critical cytotoxic DNA lesions
`induced by such agents. One of the most supported mechanisms of
`tumor
`resistance
`to DNA-damaging
`agents is the overexpression
`of
`the repair protein MGMT (26—35). This protein rapidly reverses
`formation of adducts at
`the @6position of guanine and averts the
`formation
`of
`lethal DNA cross-links
`induced
`by chloroethylating
`nitrosoureas
`(36—38).MGMT is the major mechanism of resistance of
`tumor cells to bifunctional nitrosoureas, and its depletion with suicide
`inhibitors,
`such
`as 06-BG,
`has been
`shown
`repeatedly
`to sensitize
`
`Received 4/25/97; accepted I0/3/97.
`The costs of publication of this article were defrayed in part by the payment of page
`charges. This article must therefore be hereby marked advertisement in accordance with
`18 U.S.C. Section 1734 solely to indicate this fact.
`To whom requests for reprints should be addressed, at Department of Neurology, UT
`Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas, TX 75235-
`9036. Phone:
`(214) 648-6314; Fax:(214) 648-7992.
`
`@
`
`used
`are:
`CCNU,
`l-(2-chloroethyl)-3-cyclohexyl-l-nitrosourea
`2 The
`abbreviations
`intraductal papillary mucinous
`(lomustine);
`IPMC,
`carcinoma; MGMT, O@-methylgua
`nine-DNA methyltransferase; MeG, methylguanine; O°-BG,O°-benzylguanine; BCNU,
`l,3-bis(2-chloroethyl)-I-nitrosourea
`(carmustine); MNU, N-methylnitrosourea; MMR,
`mismatch repair; MIN, microsatellite
`instability; UTSWMC. University of Texas South
`western Medical Center.
`
`5360
`
`
`
`cancerres.aacrjournals.org Downloaded from
`
`on March 13, 2017. © 1997 American Association for Cancer Research.
`
`NOVARTIS EXHIBIT 2058
`Par v. Novartis, IPR 2016-01479
`Page 1 of 9
`
`

`

`MGMT IN PANCREATIC TUMORS
`
`tumor cells to BCNU (39—42).In this communication, we examined
`the capacity of pancreatic tumors to express MGMT and the effect of
`such
`expression
`on
`the
`resistance
`of pancreatic
`tumors
`to DNA
`methylation and chloroethylation
`by MNU and BCNU,
`respectively.
`
`MATERIALS AND METHODS
`
`selected
`randomly
`of
`frozen samples
`of Tissue. Twelve
`Procurement
`tissue were obtained from the University of Kentucky Hos
`pancreatic
`tumor
`pital. Eight additional samples were obtained fresh, following Whipple or total
`pancreatectomies
`performed at UTSWMC. Seemingly normal
`tissue resected
`from sites remote from the tumor was available for four of the University of
`Kentucky Hospital tumors. Normal pancreatic parenchyma was also collected
`from all tumors obtained at UTSWMC. In addition, seemingly normal ductal
`epithelium and hyperplastic
`epithelium were obtained from the main pancre
`atic duct of pancreata removed at the UTSWMC site. Part of each sample was
`shock frozen no later than 0.5 h after resection by immersing it in liquid
`nitrogen, where it was stored until used, whereas another part was fixed in
`formalin and processed for histological
`examination.
`Preparation of Tissue and Ceil Extracts. Frozen tissues were cut in thin
`sections with a razor and immediately immersed in 50 mMcold (0°C)Tris-HC1
`(pH 8.0) containing I mMEDTA, I mistf3-mercaptoethanoland 0.2% soybean
`trypsin inhibitor (homogenizing buffer). Sections containing necrotic or hem
`orrhagic regions were cut off under a magnifying scope and discarded; the
`remainder
`was homogenized
`on ice using
`an Ultraturrax
`(Omni, Waterbury,
`CT) at 15,000 rpm and subsequently sonicated on ice using a Branson Soni
`cator (Branson, Danbury, CT). Homogenates were centrifuged at 10,000 rpm
`for 10 mmnat 4°C,and clear supernatants were frozen in liquid nitrogen until
`assayed. Specimens from the main duct were cleaned from the surrounding
`tissue, and the duct was cut in a section of approximately 1 cm and inverted
`using a stainless steel wire. The surface of the duct was scraped with a scalpel,
`and the material removed was suspended in homogenizing buffer. Cell extracts
`were prepared by lysing cells in homogenizing buffer, thawing to room
`temperature, refreezing in liquid nitrogen (three times), and subsequently
`sonicating
`at 70% maximum output
`for 5 s (three times). Cell
`lysates were
`centrifuged at 15,000 X g, and supernatants were kept frozen in liquid nitrogen
`until use.
`Biochemical Determination of MGMT. MGMTactivitywas determined,
`as described
`previously
`(43), by high-performance
`liquid chromatography
`analysis of
`the
`[3H-CH3]O@-MeGcontent
`in acid hydrolysates of
`[3H-CH3]MNU methylated calf thymus DNA following incubation of the
`DNA with tissue or cell homogenates. In this assay, the detection limit for
`MGMT is dependent on the specific activity of the 06-MeG in the DNA
`substrate.
`The
`substrate
`used
`had a specific
`activity
`of 16 Ci/mmol,
`and the
`detection limit for MGMT was approximately 5 fmol/mg protein. Protein
`determination
`was made
`using
`the Bio-Rad
`assay
`and corrected
`for
`the pres
`ence of trypsin inhibitor
`in the samples.
`in terms of
`Statistical Analysis. Comparisons
`among groups of tumors
`MGMT activity were done by one-way ANOVA using Instat (Graph Pad, San
`Diego, CA).
`for MGMT. Fifteen tumors were examined
`In Situ Immunohistostaimng
`immunohistologically
`using the anti-MGMT monoclonal
`antibody MT23.2,
`which was developed
`originally
`by Dr. Thomas Brent
`(St. Jude Children's
`Research Hospital)
`and Darell Bigner
`(Duke University)
`and donated gener
`ously by Dr. Brent. All immunostaining was performed at room temperature on
`a BioTek Solutions TechMate 1000 automated immunostainer (Ventana
`BioTek Systems, Tucson, AZ). Buffers, blocking solutions, secondary anti
`bodies,
`avidin-biotin
`complex
`reagents,
`chromogen,
`and hematoxylin
`counter
`stain were used as supplied in the ChemMate
`secondary detection kit (Ventana
`BioTek Systems). Paraffin sections were cut at 3-sm intervals on a rotary
`microtome, mounted on positively charged glass slides (POP100 capillary gap
`slides, Ventana BioTek Systems), and air dried overnight. Sections were then
`deparaffinized
`in xylene and ethanol and placed in 200 ml of heat-induced
`epitope retrieval buffer (pH 6.8; Ventana BioTek Systems). The buffer was
`heated to the boiling point, after which 50 ml of deionized water were added.
`The buffer was again heated to the boiling point for 5 mm, and then the slides
`were cooled in buffer for 20 mm, after which they were rinsed thoroughly in
`deionized water followed by buffer. Sections were incubated in unlabeled
`5361
`
`at
`antibody
`primary
`for 5—10 mm either with
`solution
`antibody
`blocking
`various dilutions ranging from 1:200 to 1:800 in buffer or with buffer alone, as
`a negative reagent control. After being washed in buffer, sections were incu
`bated for 25 mm with biotinylated
`polyvalent
`secondary
`antibody
`solution
`(containing
`goat
`antibodies
`to rabbit. mouse,
`and rat
`immunoglobulin).
`After
`another buffer wash, sections were incubated with three changes, 2.5 mm each.
`of 3% hydrogen peroxide to inhibit endogenous
`tissue peroxidase
`activity and
`washed
`again
`in buffer.
`Sections
`were
`incubated
`for 25 mm with
`freshly
`prepared horseradish
`peroxidase-conjugated
`avidin-biotin
`complex and then
`washed in buffer. The sections were incubated with three changes, 5 mm each,
`of a freshly prepared mixture of diaminobenzidine
`and H,O,
`in buffer,
`fol.
`lowed by washing in buffer and then water. Sections were counterstained with
`hematoxylin,
`dehydrated in a graded series of ethanol and xylene, and cover
`slipped.
`Slides were viewed
`by light microscopy.
`Positive
`reactions
`to diami
`nobenzidine
`were
`identified
`as dark
`brown
`reaction
`product.
`Sections
`were
`photographed
`on a Nikon Optiphot microscope
`(Nikon Instruments, Melville,
`NY).
`lines. CAPAN-2,
`Cell Growth and Cytotoxicity Assays. Four cell
`CFPAC-1, MIAPaCa-2, and PANC-l,
`isolated originally from pancreatic
`ductal adenocarcinomas, were used in this study. All cell lines were maintained
`in Eagle's MEM (Life Technologies,
`Inc.)
`supplemented
`with lysine.
`valine,
`and leucine (100 @.LMeach) and with 10% dialyzed fetal bovine serum. The
`medium was also supplemented
`with nonessential
`amino
`acids
`(I : 100 dilution
`of
`stock from Life Technologies,
`Inc.),
`1 m@i sodium pyruvate,
`sodium
`bicarbonate, 6 .LMa-hydroxycobalamin, 100 MMfolic acid, 0.2 mg/mI gen
`tamicin,
`and
`100 @ML-methionine.
`The
`cytotoxicity
`of BCNU on cultured
`cells was measured as follows: cells growing at an exponential
`rate (10—20%
`confluency)
`in 100-mm Petri dishes were incubated at 37°C,either in medium
`alone or in medium containing 20 p@M06-BG for 2 h. O°-BGwas synthesized
`and donated to us by Dr. Robert Moschel
`(National Cancer
`Institute-Frederick
`Cancer
`Research
`and Development
`Center,
`Frederick, MD).
`Subsequently,
`dishes were washed with PBS and incubated
`with freshly
`prepared
`BCNU in
`PBS
`(pH 7.4)
`at concentrations
`ranging
`between
`0 and 500
`j.tM for an
`additional
`45 mm. Afterward.
`the BCNU was washed
`off with PBS,
`and the
`cells were incubated with medium containing
`5 j.@MO@'-BGfor 4 days. The
`cells were then trypsinized
`and counted
`using a Coulter Counter
`(Coulter
`Electronics,
`Hialeah,
`FL). The
`experiment
`was
`repeated
`using MNU at con
`centrations
`ranging
`between
`0 and 2000
`@M.The concentration
`of BCNU or
`MNU that halved the growth rate of the tumor cells (IC@)) was determined
`from plots of drug concentration
`versus
`the percentage
`of change
`in cell
`numbers
`as
`compared
`with
`untreated,
`nonconfluent
`controls
`4 days
`after
`treatment. Conditions were optimized
`for
`the cells used in this
`study by
`exposure to BCNU and MNU in PBS, because the presence of serum or even
`nutrients yielded nonconsistent data. Also, incubations with nitrosoureas were
`done on attached cells and were limited to only 45 mm, because three of four
`of the cultured pancreatic
`cells began to detach I h after
`the medium was
`replaced with PBS. Cells were
`not exposed
`to 06-BG during
`treatment
`with
`BCNU or MNU, because nitrosoureas alkylate the MGMT inhibitor at rates
`sufficient
`to cause
`a reduction
`of
`their
`effective
`concentration.3
`Determination
`of Mitosis and Apoptosis.
`Trypsinized
`cells derived from
`a 24—48-h treatment with BCNU or MNU were washed with PBS three times
`and spread on slides with a cytospin centrifuge
`at 500 rpm for 1 mm. The
`mitotic
`index
`(mitosis
`per
`1000 cells) was determined
`from a set of
`slides
`stained with H&E. Apoptosis
`was determined
`in situ on another
`set of slides
`using
`terminal
`deoxynucleotidyl
`transferase-mediated
`nick
`end
`labeling
`(Oncor, Gaithersburg, MD). Apoptosis was distinguished from necrosis by
`morphological
`characteristics,
`i.e., condensed
`nuclei
`and nuclear
`fragmenta
`tion.
`lines, CFPAC-l, MIAPaCa-2, and
`MIN. Three pancreatic tumor cell
`PANC-l, were negative for MIN during the course of I year of culturing. To
`evaluate possible loss of fidelity in DNA replication following treatment with
`MNU, cell lines were exposed to MNU alone (500 p@Mto 2 mM) for 45 mm or
`to the same concentrations
`of MNU after
`they were treated with 20 @MO°-BG
`for 2 h. 06-BG-treated
`cells were subsequently
`cultured in medium supple
`mented with S @LMO@-BGfor 4 days and after that in regular medium. Controls
`(no 06-BG or MNU) were also cultured in parallel. Analysis of MIN was
`
`3 D. M.
`
`Kokkinakis,
`
`unpublished
`
`observations.
`
`Downloaded from
`
`cancerres.aacrjournals.org
`
`on March 13, 2017. © 1997 American Association for Cancer Research.
`
`NOVARTIS EXHIBIT 2058
`Par v. Novartis, IPR 2016-01479
`Page 2 of 9
`
`

`

`MGMT IN PANCREATICTUMORS
`
`performed on DNA extracted from cultures treated with MNU alone, MNU
`plus O'@-BG [following
`several
`(>10)
`cycles of division],
`and also from
`controls. The following six loci were screened for MIN: D2SJ23, D2SJ36,
`D3SJ067, D5J07, D6S87, and DJ8S34. Primers for the amplification
`of these
`loci were
`purchased
`from Research
`Genetics,
`Inc.
`(Huntsville,
`AL).
`PCR
`amplification was performed
`using 32P-end-labeled
`sense primers. The PCR
`product was denatured and electrophoresed on urea polyacrylamide (8%) gels.
`After electrophoresis, the gels were dried and visualized by autoradiography.
`Instability was determined as an alteration in banding patterns in MNU or
`MNU plus BG-treated
`cells as compared
`to matching
`nontreated
`(control)
`cells. After
`treatment,
`cells were washed with PBS and cultured in medium
`alone (MNU-treated) or medium containing S @MO@-BG(MNU/O@-BG
`treated) until they were confluent.
`
`RESULTS
`
`The median age at diagnosis of patients
`Patient Characteristics.
`in the pancreas was 57 years (range,
`with ductal adenocarcinomas
`24—68years). In addition to ductal adenocarcinomas,
`one patient had
`a tumor in the ampulla, one had a mucinous cystic adenocarcinoma,
`and two had a borderline mucinous
`cystic tumor.
`In addition,
`four
`cases of intraductal papillary mucinous carcinoma were identified.
`MGMTActivity. A totalof 20 pancreaticneoplasms,12segments
`of seemingly normal pancreatic parenchyma,
`and 6 samples of the
`seemingly normal main pancreatic duct epithelium were assayed for
`MGMT activity (Table 1). MGMT activity in pancreatic neoplasms
`was highly variable. Higher activity was observed in adenocarcinomas
`(mean, 459 fmol/mg
`protein;
`range,
`160—875 fmol/mg
`protein;
`n = 12), followed by other less common tumors such as mucinous
`cystic tumors and IPMCs. The difference in MGMT levels between
`IPMCs and mucinous cystic tumors (cases 14—20),on one hand, and
`grade 3 ductal adenocarcinomas
`(cases 1—10),on the other, was
`statistically significant
`(P = 0.0006). No correlation was found be
`tween the age or sex of the patient and MGMT levels in the tumor.
`Unlike MGMT activity in tumors, MGMT activity in both normal
`parenchyma
`(mean, 41 fmol/mg protein;
`range, 29—66fmol/mg pro
`tein; n = 12) and the seemingly normal main duct (mean, 35 fmollmg
`protein;
`range, 32—46fmollmg protein; n = 6) was low in those
`patients
`from whom histologically
`normal
`tissue was available. The
`results suggest a correlation between tumor grade and level of MGMT
`in both adenocarcinomas
`and IPMCs.
`Immunohistochemistry.
`Fifteen ductal carcinomas examined im
`munohistochemicaliy
`showed intense nuclear and cytoplasmic
`stain
`
`ing with the MT23.2 MGMT monoclonal antibody. Highly dyspiastic
`ductal cells stained in all 12 tumors, as shown in Fig. 1A and B,
`whereas
`hyperplastic
`ductal
`epitheiium showed
`no reactivity
`or
`stained weakly (Fig. 1, C and D). Invasive carcinomas usually showed
`more intense staining than did nomnvasive
`components of the same
`tumor (Fig. 1E). Cells staining for MGMT were only focally positive
`
`in
`the
`more
`differentiated
`areas
`of
`intraductal
`papillary
`carcinomas
`(Fig. 1, F and G) and varied in MGMT content as it was assessed from
`the intensity of staining. A more uniform and intense staining was
`observed
`in less-differentiated
`ductal
`carcinomas
`(Fig. 1H). Histolog
`ically normal pancreata did not usually stain. An exception was found
`in 2 specimens
`(of 12), in which islets and acinar cells trapped in the
`tumor showed variable nuclear staining (data not shown). Likewise,
`normal
`lymphocytes
`present
`in the tumor
`stroma and in adjacent
`lymph nodes also showed positive reactivity (data not shown). No
`reactivity was observed when the antibody was omitted (Fig. 11). It
`must be noted that all tissue sections were processed simultaneously
`and subjected to the same treatment conditions. Failure to stain does
`not preclude the presence of MGMT, which may be present at an
`undetectable
`level. This is supported further by the biochemical de
`tection of MGMT in normal ductal epithelium. On the basis of strong
`positive staining of Daoy tumor xenografts
`(MGMT, 320 fmoL/mg
`protein) but not of U-138 xenografts containing less than 160 fmol/mg
`protein, we conclude that
`the cutoff point
`for positive detection of
`MGMT by our method lies between these two values.
`Cell Lines. Four pancreatic tumor cell lines tested were all positive
`for MGMT activity. Levels of MGMT measured from 790 to 1800
`fmol/mg protein (Fig. 2), a range that was significantly higher than the
`mean found in pancreatic
`adenocarcinomas.
`In addition, Daoy, a
`human medulloblastoma
`line, and U-263, a human glioblastoma
`line,
`which are well characterized
`in terms of MGMT expression
`and
`resistance to alkylating agents in our laboratory, were used for com
`parative purposes. Rates of growth for these lines measured as dou
`blmg time in our medium were 22 h for Daoy; 28 h for CFPAC-1,
`PANC-l,
`and U-263; 36 h for CAPAN-2; and 48 h for MIAPaCa-2.
`The rates were not affected by the inclusion of 5 ,.LM06-BG in the
`medium for 4 days.
`MGMTInactivationby O@-BG.As shownin Fig. 2, CAPAN-2
`expressed
`1800
`fmol/mg
`protein MGMT
`activity.
`CFPAC-l,
`PANC-1, and MIAPaCa-2,
`expressed approximately one-half
`to one
`third of this activity. Daoy and U-263 expressed 380 and 160 fmol/mg
`
`Table 1 MGMT activity in pancreas tumors
`
`Tumor
`
`307
`875
`
`MIN@
`MJN@
`MIN@611
`
`333
`
`Ki-ras
`
`3MIN
`
`MIN@350
`
`433
`
`3
`
`33
`
`DAC
`DAC
`DAC4
`
`DAC3
`
`cystic
`cystic
`
`Age (yr)SexType
`Patient
`of surgery(months)LocationHistological
`typeâ€(cid:157)StageGradealterationsâ€(cid:157)mg/protein)1
`62MWhipple6HeadDAC33—4422
`58FWhipple8HeadDAC33—5113
`61MWhipple6HeadDAC43Ki-ras5274
`65
`44
`68
`62F
`24
`MWhippleWhipple>52
`62F
`9
`21110
`64FWhipple>6HeadDAC32—7701
`166FWhipple14BodyDAC22—44512
`62MGastrojejunostomy4HeadDAC41—1601
`349FWhipple>39AmpullaryAdenocarcinoma42—35914
`63FDistal
`pancreatectomy1TailMucinus
`borderlineI2—13515
`41FDistal
`pancreatectomy>5TallMucinus
`borderline11—14516
`47MPancreatectomy>7HeadDAC-Chronic
`pancreatitisI1—9517
`68MWhipple>24HeadIPMC11—21118
`54FWhipple>18HeadIPMC12—18519
`52MWhipple>84HeadIPMCI1—13020
`63FWhipple>12HeadIPMC12—240a
`
`MMM
`
`Whipple
`Whipple
`Whipple
`Whipple4
`
`5
`6
`7
`4608
`
`10
`14
`12Head
`
`Head
`Head
`HeadDAC
`
`14Tail HeadDAC
`
`DAC,
`
`ductal
`not done.adenocarcinoma;IPMC,
`
`b
`
`intraductal papillarymucinouscarcinoma.
`
`5362
`
`Downloaded from
`
`cancerres.aacrjournals.org
`
`on March 13, 2017. © 1997 American Association for Cancer Research.
`
`NOVARTIS EXHIBIT 2058
`Par v. Novartis, IPR 2016-01479
`Page 3 of 9
`
`

`

`@
`
`
`
`@@
`
`@1
`
`MGMT IN PANCREATIC TUMORS
`
`@..
`
`f
`
`@8
`
`c
`
`:
`
`‘,-
`
`@1
`
`-‘.--..
`
`‘1
`
`@.
`
`“4
`,@
`
`‘
`
`ti.
`
`@.‘
`
`4
`
`,“. ,‘,,
`
`:•
`
`
`
`@@
`
`@
`
`,
`
`
`
`@‘:‘:@ . @:
`
`- ‘@;t
`
`. -
`
`..‘‘....
`
`..
`
`_a@@
`
`--
`
`@,_
`
`
`
`
`
`@@@
`
`@
`
`@@
`
`
`
`
`
`@@@
`
`..;
`
`.,....
`
`.‘.,.
`
`.
`
`‘.
`
`‘@- ..
`
`-
`
`-
`
`.
`
`-
`
`@@%@S1l@
`
`.,@
`
`‘S.'
`
`•;:.
`
`.
`
`@: @.
`
`•
`
`.
`
`@$.
`
`.
`
`.,
`
`.
`
`
`
`cancerres.aacrjournals.org Downloaded from
`
`on March 13, 2017. © 1997 American Association for Cancer Research.
`
`5363
`
`NOVARTIS EXHIBIT 2058
`Par v. Novartis, IPR 2016-01479
`Page 4 of 9
`
`

`

`MGMT IN PANCREATIC TUMORS
`
`CAPAN-2
`s@
`*@ MIAPaCa-2
`mumCFPAC-1
`=
`PANC-1
`— DAOY
`— 11-139
`
`j
`
`2100
`
`I 800
`
`I 500
`
`I 200
`
`0)
`
`00
`
`.
`0)
`
`-@
`0
`
`@
`
`(data not
`(Table 2) and MIAPaCa-2
`in PANC-l
`inhibited mitosis
`shown) more effectively when cultures were treated with 06-BG.
`Eradication
`of
`these two lines
`(Table 2; Fig. 3) was also more
`effective
`when
`06-BG was used
`to suppress MGMT.
`Overall,
`pan
`creatic tumor cell
`lines were markedly more resistant
`to BCNU than
`was the MGMT-efficient Daoy, which had a moderate MGMT activ
`ity. Elimination
`of MGMT
`resulted
`in a similar
`sensitization
`of both
`pancreatic
`and nonpancreatic
`cell
`lines
`to BCNU, which
`substantiates
`the role of MGMT as a common and obviously major mechanism of
`resistance of a variety of tumor cell
`lines to the toxic effect of this
`drug.
`Toxicity to MNU in the Absence and Presence of 06-BG. Of the
`three pancreatic tumor
`lines tested, only CFPAC-l was sensitive to
`DNA methylation by MNU, with an IC50 of approximately
`360 pM
`(Fig. 4). The IC50 was reduced to 120 p.M when MGMT activity was
`suppressed with 06-BG. Complete
`eradication
`of the culture was
`observed following treatment with 06-BG and 1500 p.MMNU or with
`2000 @.LMMNU alone. Daoy was also resistant
`to MNU and was
`sensitized by MGMT elimination, as determined by a reduction of the
`IC50 from 500 to 150 p.MMNU in the presence of 06-BG. Complete
`eradication of this cell line required exposure to 1 mi@tMNU. Unlike
`Daoy and CFPAC-l, MIAPaCa-2
`and PANC-1 were resistant
`to
`MNU, and such resistance was only marginally affected by depletion
`of MGMT activity. Because cell numbers (shown in Fig. 4) depended
`not only on cell death, but also on inhibition of cell proliferation via
`DNA synthesis blocks, we tested the relative contributions
`of these
`two factors in the observed reduction of cell numbers in MIAPaCa-2
`and PANC-l
`cultures treated with MNU. MNU concentrations
`up to
`1000 @LMinhibited mitosis
`in PANC-l
`and MIAPaCa-2
`cultures at
`48 h posttreatment
`but had no effect on the frequency of cell death.
`Furthermore,
`inhibition of DNA synthesis
`and cell death were not
`affected by use of 06-BG. Complete eradication of these two cultures
`could not be achieved at the concentrations
`of MNU used, regardless
`of the presence of the MGMT inhibitor. This suggests that inhibition
`of cell cycle by MNU (Table 2) is reversible
`and that
`the apparent
`resistance
`of PANC-l
`and MIAPaCa-2
`to MNU is due to mechanisms
`other than the ability to repair 06-MeG. Due to continued inhibition
`of MGMT activity by 06-BG during culture,
`it is assumed that
`in
`these two cultures, cell cycle and DNA synthesis are operational while
`06-MeG lesions
`in genomic DNA are still present.
`In contrast
`to
`
`900
`
`600
`
`300
`
`0
`
`>I
`
`—
`
`0 I
`
`- 0
`
`0
`
`10
`
`20
`
`06-BG (j.tM)
`
`cell
`Fig. 2. Depletion of MGMT activity by O@-BGin four pancreatic adenocarcinoma
`lines. O@-BG dissolved in medium was added to attached proliferating cells. Cells were
`harvested 2 h after initiation of treatment and washed extensively with PBS, and the
`remaining MGMT activity was determined using the biochemical
`assay.
`
`in all of the lines, was com
`respectively. MGMT activity,
`protein,
`pletely eliminated with the addition of 10 /.LM06-BG to the culture
`medium for 4 h or 20 p.M of 06-BG for 2 h. Because
`06-BG
`concentrations
`in culture medium did not change
`significantly
`with
`time, frequent changes of 06-BG-containing medium were not nec
`essary.
`A depleted
`MGMT
`state
`(< 15 fmol/mg
`protein)
`was main
`tamed by adding 5 @.aMof 06-BG in the medium.
`Toxicity of BCNU in the Absence and Presence ofO@-BG. Three
`pancreatic
`cell
`lines
`tested
`for
`toxicity
`to BCNU were
`found
`to be
`highly
`resistant
`compared
`with Daoy
`(Fig. 3).
`In these
`lines, mitotic
`activity was
`somewhat
`inhibited
`up to 48 h from treatment
`without
`affecting
`survival
`following
`a 45-mm incubation
`with 60 LM BCNU
`(Table 2), a concentration that is known to eradicate MGMT-deficient
`cell cultures (43). Severe inhibition of mitosis, however, was evident
`at BCNU concentrations
`as low as 40 p.M (Table 2), when CFPAC-l
`was pretreated with 06-BG before exposure to BCNU.
`In addition,
`treatment of CFPAC-l with 06-BG effectively reduced the IC50 for
`BCNU from 200 to less than 40 p.M (Fig. 3). Similarly, BCNU
`
`120
`80
`
`40
`
`20
`
`10
`
`5
`
`I
`
`B
`
`0
`
`40
`
`80
`
`120
`
`160
`
`0
`
`100 200 300 400 500 600
`
`0
`
`100200300400500600
`
`0
`
`40
`
`80
`
`120
`
`160
`
`Concentration of BCNU (j.tM)
`
`(C), and MIAPaCa-2 (D). Exponentially growing cells
`(B), PANC-l
`Fig. 3. Potentiation of BCNU toxicity by O@-BGin Daoy (A) and in three pancreatic tumor cell lines. CFPAC-l
`were treated with 20 @.LM06-BG in full medium for 2 h. Cells were subsequently washed with PBS and treated with varying concentrations of BCNU in PBS for 45 mm at 3TC. Cultures
`were washed again with PBS and supplemented with full medium containing 5 @LMO@-BG. The medium was changed every 48 h for 4 days (before any of the cultures
`reached
`confluency). Finally, cells were trypsinized and counted, and survival was determined as a percentage of the control
`(not treated with BCNU). A parallel experiment was run in which
`O@-BGwas omitted. Relative survival is shown for cells treated with BCNU alone (•)or with BCNU plus O@-BG(0).
`5364
`
`Downloaded from
`
`cancerres.aacrjournals.org
`
`on March 13, 2017. © 1997 American Association for Cancer Research.
`
`NOVARTIS EXHIBIT 2058
`Par v. Novartis, IPR 2016-01479
`Page 5 of 9
`
`

`

`linesCell
`
`Table 2 Effect of alkvlating drugson division and survival ofpancreatic
`
`tumor cell
`
`indexâ€(cid:157)(% of control)Cell
`
`
`
`deathâ€(cid:157)(% ofcontrol)—
`
`MGMT IN PANCREATIC TUMORS
`
`lineTreatmentConcentrationMitotic
`06-BGCFPAC-1BCNU20
`
`06-BG—
`06-BG
`
`+
`
`1200PANC-l.BCNU20
`
`1100MNU100
`
`1200MNU100
`
`40
`60
`80
`10088
`
`200
`500
`1000
`200095
`
`40
`60
`80
`100100
`
`200
`500
`1000
`200095
`
`59
`56
`52
`54
`
`75
`58
`56
`51
`
`100
`92
`70
`64
`
`92
`95
`81
`71
`
`55
`36
`20
`16
`12100
`
`90
`63
`55
`55
`48100
`
`95
`92
`83
`20
`11100
`
`82
`84
`80
`75
`75100
`
`06-BG+
`
`100
`100
`100
`150150
`
`100
`400
`650
`800100
`
`100
`150
`150
`200200
`
`100
`100
`100
`150100
`
`500
`1000
`1100
`
`100
`900
`1100
`
`450
`950
`1000
`
`100
`100
`100
`200
`
`a Mitotic
`
`index,
`
`percentage
`
`of decrease
`
`of
`
`the number
`
`of mitoses
`
`per
`
`1000
`
`cells
`
`counted
`
`48 h after
`
`treatment
`
`as compared
`
`to untreated;
`
`average
`
`of
`
`five
`
`determinations.
`
`b Apoptotic
`
`bodies,
`
`number
`
`of cells
`
`staining
`
`with Apop-Tag
`
`48 h after
`
`treatment
`
`as compared
`
`to untreated;
`
`average
`
`of
`
`three
`
`determinations
`
`rounded
`
`to the nearest
`
`increment
`
`of 50.
`
`MIAPaCa-2 and PANC-l, CFPAC-1 was both inhibited in terms of
`growth and killed by MNU (Table 2). In this line, killing but not
`mitotic inhibition was stimulated by O@-BG.
`MIN of Pancreatic
`Tumor Cells. The failure of MNU to kill
`PANC-l
`and MIAPaCa-2
`and the apparent
`inability of 06-BG to
`sensitize these cultures to MNU could be related to defective MMR in
`these lines. Because MMR defects are frequently associated with the
`hypermutable phenotype and MIN, we tested these lines for the latter
`characteristic. Detection
`of
`instability was based on comparisons
`between MNU-treated and untreated tumor lines, respectively, assum
`ing that the treatment
`induces mutations
`reflected by an alteration of
`microsatellite
`sequences. MNU treatment of both MIAPaCa-2
`and
`PANC-l
`resulted in microsatellite
`alterations at several
`loci,
`indicat
`ing the presence of an ineffective MMR system (Fig. 5, Table 3). On
`the other hand, CFPAC-l, which was susceptible to MNU treatment,
`especially when pretreated with 06-BG, did not show MIN. This is in
`
`of MNU to kill cells
`the failure
`that
`with the assumption
`accordance
`is due to a defective MMR in MIAPaCa-2 and PANC-i.
`
`DISCUSSION
`
`to
`Pancreatic tumors express various levels of MGMT that appear
`correlate
`with the degree
`of malignancy.
`MGMT
`activity
`is low in
`seemingly normal ductal epithelium and pancreas in general, but it is
`present at high levels in invasive ductal adenocarcinomas
`and,
`to a
`lesser extent,
`in intraductal
`tumors. Although the absence of reactivity
`with MGMT antibodies
`does not necessarily
`indicate
`a MGMT
`deficient phenotype,
`the weak or negative
`immunostaining
`of the
`normal pancreas supports the assumption that expression of MGMT is
`up-regulated
`during the carcinogenic
`process
`in th