lENT
`
`9
`
`ANTIFOLATE DRUGS IN
`CANCER THERAPY
`
`Edited by
`ANN L. JACKMAN
`The Cancer Research Campaign Centre
`for Cancer Therapeutics,
`The Institute of Cancer Research,
`Sutton, Surrey, UK
`
`HUMANA PRESS
`TOTOWA, New JERSEY
`
`Sandoz Inc.
`Exhibit 1012-0001
`
`JOINT 1012-0001
`
`

`
`© 1999 Humana Press Inc.
`999 Riverview Drive, Suite 208
`Totowa, New Jersey 07512
`
`All rights reserved.
`
`No part of this book may be reproduced, stored ia a retrieval system, or transmitted in any form or by any means,
`electronic, mechanical, photocopying, microfilming, recording, or otherwise without written permission from the
`Publisher.
`
`All authored papers, comments, opinions, conclusions, or recommendations are those of the author(s), and do not
`necessarily reflect the views of the publisher.
`
`For additional copies, pricing for buIk purchases, and/or information about other Humana titles, contact Humana at the
`above address or at any ofthe follow ng numbers Te 973-256-1699" Fax’ 973-~56-8341 ’ E-mail" humana@humanapr corn
`
`This publication is printed on acid-free paper. ~3
`ANSI Z39.48-1984 (American Standards Institute) Permanence of Paper for Printed Library Materials.
`
`Cover illustration:
`
`Cover design by Pairicia F. Cleary.
`
`Photocopy Authorization Policy:
`Authorization to photocopy items for internal or personal use, or the internal or personal use of specific clients, is granted
`by Humana Press Inc., provided that the base fee of US $5.00 per copy, plus US $00.25 per page, is paid directly to the
`Copyright Clearance Center at 222 Rosewood Drive, Danvers, MA 01923. For those organizations that have been
`granted a photocopy license from the CCC, a separate system of payment has been arranged and is acceptable to Humana
`Press Inc. The fee code for users of the Transactional Reporting Service is: [0-89603-460-7/97 $5.00 + $00.25].
`
`Printed in the United States ofAmerica. I0 9 8 7 6 5 4 3 2 1
`
`t
`
`t
`
`r
`
`Sandoz Inc.
`Exhibit 1012-0002
`
`JOINT 1012-0002
`
`

`
`)yle et al.
`
`Effects of
`inhibitor
`
`). Phase I
`
`I trial of
`nuous in-
`
`12 Preclinical and Clinical Evaluation
`
`of the Glycinamide Ribonucleotide
`Formyltransferase Inhibitors
`Lometrexol and LY309887
`
`Laurane G. Menddsohn, John F. Worzalla
`andJackie M. Walling
`
`CONTENTS
`
`INTRODUCTION
`
`INHIBITION OF GAR~ AND POLYGLUTAMATION BY
`
`FOLYLPOLYGLUTAMATE SYNTHETASE
`
`FOLATE-RECEPTOR SELECTIVITY OF ANTIFOLATES
`
`CELLULAR PHARMACOLOOY
`IN VIVO ANTITUMOR ACTIVITY OF LY309887
`THERAPEUTIC INDEX DETERMINATiONs
`
`THE EFFECT OF DIETARY FOLATE ON DISPOSITION OF LOMETREXOL
`AND LY309887 IN LIVER
`THE EFFECT OF LFD AND DIETARY I~OLATE SUPPLEMENTATION ON
`THE EFFICACY AND TOXICITY OF LOMETREXOL AND LY309887
`HUMAN FOLATE STATUS
`CLINICAL EVALUATION OF LOMETREXOL AND LY309887
`PHASE I STUDIES WITHOUT FOLIC ACID SUPPLEMENTATION
`PHASE I STUDIES WITH FOLIC ACID SUPPLEMENTATION
`PHASE I STUDY OF LOMETREXOL WITH FOLINIC ACID
`HOW DOES FOLIC ACID MODULATE TOXICITY?
`ANTITUMOR ACTIVITY IN PHASE I
`PHARMACOKINETICS
`CLINICAL DEVELOPMENT OF LY309887
`CONCLUSION
`
`-
`
`From: Anticancer Drug Development Guide." Antifolate Drugs in Cancer Therapy
`Edited by: A.L. Jackman © Humana Press Inc., Totowa, NJ
`
`261
`
`Sandoz Inc.
`Exhibit 1012-0003
`
`JOINT 1012-0003
`
`

`
`Mendelsohn, Worzalla, and Walling~
`
`c~
`
`1. INTRODUCTION
`
`The importance of the purine de novo pathway in providing DNA precursors for can-
`cer cell growth led to the hypothesis that novel antifolate inhibitors of glycinamide ri-
`bonucleotide formyltransferase (GARFT), the first folate-dependent enzyme in this
`pathway, might have utility in the treatment of cancer. In 1987, clinical investigations
`were initiated with lometrexol (6R-dideazatetrabydrofolic acid, 6R-DDATHF), a novel
`"tight-binding" inhibitor of GARFT with potent antitumor activity in a number of
`murine and human xenograft solid tumors. Unexpected observations of delayed cumu-
`lative toxicity in phase I clinical trials prompted extensive preclinical investigations of
`the dynamics of folate status on the efficacy and toxicity of GARFT inhibitors and other
`antifolates (_1). In addition, structure-activity studies have led to the identification of a
`second generation GARFT inhibitor, LY309887 (2’,5’-thienyl-dideazatetrahydrofolic
`acid), which is more potent than lometrexol and has greater antitumor efficacy in vivo
`(2). Biochemical and pharmacological differences between LY309887 and lometrexol
`with respect to potency to imhibit GARFT, differential transport and storage in liver, and
`polyglutamatiou suggest that LY309887 may have greater antitumor efficacy and more
`manageable toxicity in the clinic than lometrexol. A routine model of the delayed cu-
`mulative toxicity seen with lometrexol has been refined and characterized to provide
`greater understanding of the pharmacokinetics and pha~macodynamics of these events.
`In concert with recently published nutritional data on the folate status of humans and
`more sophisticated methods of assessing and modulating antifolate toxicities through vi-
`tamin supplementation, antifolate therapy may be poised to enter a new phase of clini-
`ca~ success. In this report, we describe LY309887, a GARFT inhibitor with unique
`biochemical and pharmacological properties that has antitumor activity against a broad
`panel of human xenograft tumors, and greater potency than lometrexol both as an in-
`hibitor of GARFT and as an inhibitor of tumor growth in vivo.’ An overview of the phase
`I clinical results with lometrexol and the design of the phase I clinical trial with
`LY309887 will be presented.
`
`2. INHIBITION OF GARFT
`AND POLYGLUTAMATION
`BY FOLYLPOLYGLUTAMATE SYNTHETASE
`
`The natural forms of folic acid and "classical" inhibitors of folate-dependent enzymes
`are polyglutamated intracellularly by the enzyme folylpolyglutamate synthetase. Poly-
`glutamylation enhances both intracellular retention and affinity of folates and gntifolates
`for many of the folate-utilizing enzymes (3). Table 1 summarizes the inhibition of
`GARFT by lometrexol, compound LY254155 (6R,S-2’,5’-thienyl-DDATHF, a di-
`astereoisomeric mix of LY309887 and LY309886, respectively) and theft- polygluta-
`mates. Monoglutamated LY254155 was approx 30-fold more potent than lometrexol.
`Polyglutamation enhanced inhibition by both compounds: lometrexol-triglutamate
`(LY235337) was approx 4.5-fold more potent than parent compound; a 10-fold increase
`in inhibition was seen with the triglutamated thiophene (LY314209). These data demon-
`strate that the thiophene was inherently more potent as a GARFT inhibitor and that in
`the polyglutamated state it achieved picomolar affinity for GARFT.
`The kinetic constants (Kin, Vmax, and first-order rate constant [Vmax!KmJ)’for activa-
`
`tior
`Tat
`eve
`util
`
`stYa
`obt~
`a fi~
`FP(
`cub
`HPI
`all ~
`vert
`in u
`serv
`at tr
`
`late
`ular
`cycl
`non
`
`Sandoz Inc.
`Exhibit 1012-0004
`
`JOINT 1012-0004
`
`

`
`Walling
`
`~or can-
`aide ri-
`in this
`gations
`a novel
`~ber of
`cumn-
`ions of
`d other
`an of a
`lrofolic
`in vivo
`etrexol
`er, and
`d more
`,ed cu-
`provide
`events.
`ns and
`agh vi-
`f clini-
`unique
`¯ broad
`an in-
`’,phase
`1 with
`
`zymes
`Poly-
`."olates
`ion of
`. adi-
`’gluta-
`xexol.
`amate
`erease
`:mon-
`fiat in
`
`.ctiva-
`
`Chapter 12 / Lometrexol and LY309887
`
`263
`
`Table 1
`Inhibition of hGARFT
`by Lometrexol, 254155, and Polyglutamates
`
`Compound No.
`
`Compound Name
`
`hGARFT Ki (nM)
`
`LY249543
`LY235540
`LY235337
`LY266978
`LY235542
`LY254155
`LY314565
`LY314209
`
`lometrexol
`diglu
`triglu
`tetraglu
`pentaglu
`thienyl-DDATHF
`diglu
`triglu
`
`59.7 (n = 2)
`15.4 (n = 2)
`13.3 (n = 2)
`7.1 _+ 2.2 (n = 4)
`5.3 (n = 2)
`2.1 -+ 0.2 (n = 5)
`1.2 (n = 1)
`0.25 (n = 2)
`
`The potency of antifolate analogs to inhibit monofunctional human GARFT
`was assessed spectrophometrically using the Morrison equa~on, which is appro-
`priate for determining the affinity of "tight-binding" compounds that produce sto-
`ichiometiic inhibition (2,4).
`
`Table 2
`Kinetic Constants for Activation of GARFT Inhibitors by FPGS
`
`Compound
`
`Km (IxM)
`
`Vmax (t~moles/h/mg)
`
`k "(Vm/Km)
`
`lometrexol
`LY309887
`MTA
`
`16.4 _+ 1.0
`6.5 _ 1.1
`1.9 + 0.5
`
`977 ----- 128
`-686 + 116
`. 725 +95
`
`60
`43
`381
`
`tion of lometrexol and LY309887, determined using hog liver FPGS are summarized in
`Table 2 (5). Lometrexol and LY309887 had similar Km values as FPGS substrates. How-
`ever, lometrexol had a significantly higher Vma~. The relative efficiencies of substrate
`utilization by an enzyme can be determined by comparing first-order rate constants, k’
`(Vrnax]Km). The data suggest that desPite equal Km values, lometrexol was a better sub-
`strate, which would be more extensively polyglutamated in vivo. For comparison, data
`obtained with the multitargeted antifolate inhibitor, LY231514 (MTA), is shown. With
`a first-order rate constant of 381, it clearly had the greatest affinity and efficacy as an
`FPGS substrate. In other experiments, polyglutamated products formed during a 24-h in-
`cubation of lometrexol, LY309887 or MTA with FPGS were separated by quantitative
`HPLC. At low substrate concentrations, i.e., below the Km (1 IxM), pob, glutamation of
`all substrates was more extensive and a higher percentage of the total product was con-
`vened to tetra- and pentaglutamated forms than at high substrate conqentrations (20 IxM)
`in which over 70% of each antifolate was present as the triglutamateanalog. These ob-
`servations are consistent with the known substrate inhibition of FPGS that occurs in vivo
`at high intracellular folate concentrations (6).
`An important inference from these data is that folate-deficient patients may accumu-
`late and retain greater amounts of highly polyglutamated "classical antifolates," partic-
`ularly in liver, a known folate depot, than patients who are folate-replete. Continuous
`cycling of stored antifolate through the enterohepatic pathway may explain the phe-
`nomenon of delayed and cumulative toxicity in cancer patients oo 10metrexol and in
`
`Sandoz Inc.
`Exhibit 1012-0005
`
`JOINT 1012-0005
`
`

`
`264
`
`Mendelsohn, Worzalla, and Walli.p__g
`
`Table 3
`Affinity of Folic Acid and Antifolates
`for Isoforms of the Folate Receptor
`
`Compound
`
`folic acid
`methotrexate
`lometrexol
`S-DDATHF
`raltitrexed
`LY309887
`LY309886
`LY231514
`5-CH:~-THF(6S)"
`
`Human KB cell
`( FR e~ )
`
`Ki (~M _+ SEM)
`Human Liver
`( FRf3 )
`
`Ratio
`Kit~/Kic¢
`
`0.07 _+ 0.11
`30.0 _+ 1(3.6
`0.29 -_+ 0.05
`0.20 -+ 0.07
`29.1 + 7.44
`1.74 -+ 0.18
`1.04 +- 0.31
`99.7 -+ 22.9
`1.00
`
`0.23 _+ 0.03
`108.00 _+ 4.4
`1.44 + 0.13
`0.82 _+ 0.01
`285.00 _+ 76.9
`18.2 _+ 4.4
`4.48 +-_ 0.36
`482.00 + 141
`55.0
`
`3.29
`3.6
`4.97
`4.10
`9.79
`10.5
`4.31
`4.83
`55.0
`
`Kic~ values were determined in equilibrium binding assays using 125I-folic
`acid and membranes prepared from human (KB) nasopharyngeal carcinoma cells
`or human liver. Values are the mean from three or more determinations.
`~FRe~, KB cell; FRI3, human placenta (9).
`
`lometrexol-treated mice on a low folate diet. Compound LY309887 may have a reduced
`potential for delayed toxicity in humans because it is less extensively polyglutamated.
`Furthermore, since it is a more potent GARFT inhibitor than lometrexol and requires less
`polyglutamation to achieve tight-binding inhibition, !ower-doses can be administered
`without compromising potency. Thus, in folate:replete patients having normal liver fo-
`late concentrations, accumulation, polyglutamation and retention of LY309887 in liver
`would be expec’ted to be lower than observed with lometrexol. In vivo experiments to
`test this hypothesis were carried out in mice and results are summarized in subheading
`7(5,7).
`
`3. FOLATE-RECEPTOR SELECTIVITY OF ANTIFOLATES
`
`Activation of classical antifolates through formation of polyglutamates and conse-
`quent deposition in liver may only partially account for the extended bioavailability and
`risk of delayed cumulative toxicity of classical antifolates. Folate-transport mechanisms,
`including the reduced folate carrier and folate receptor (FR) isoform-selectivity are im-
`portant features of antifolates that may also modulate clinical efficacy and toxicity of
`these compounds. It is noteworthy that FRc~ is highly overexpressed in many epithelial
`forms of cancer including ovarian and head and neck cancer (8). Dissociation constants
`of folic acid and several antifolates for the human FRot and FRI3 isoforms are presented
`in Table 3: The smaller the Ki, the more tightly the compound binds to the FR. Inspec-
`tion of the ratio of affinities of these antifolates for the isoforms reveals that all of the
`compounds have a lower affinity for the liver (FR[3) isoform relative to their FRc~ affin-
`ity. Such selectivity may enhance uptake of the natural folates, e.g., 5-CH3-THF into
`normal tissues or tumors over uptake into liver, particularly at physiological serum fo-
`late concentrations (10-40 nM). The affinity of lometrexol for both FR isoforms was
`very high, and the poor selectivity ratio of 4.97 suggests that lometrexol would readily
`
`c
`
`c
`
`f
`
`[
`
`d
`
`Sandoz Inc.
`Exhibit 1012-0006
`
`JOINT 1012-0006
`
`

`
`d Walling
`
`Chapter 12 / Lometrexol and LY309887
`
`265
`
`accumulate in human liver as well as tumors expressing FR. The affinity of LY309887
`for the liver receptor was 12.6-fold lower than lometrexol’s affinity and the selectivity
`of LY309887 for the ¢~ over the [3 isoform was 10.5, suggesting that this new compound
`will not accumulate in liver as much as !ometrexol. In addition to transport by FR, both
`lometrexol and LY309887 are taken up by the reduced folate carrier. This carrier, how-
`ever, has a high capacity but a tow affinity (>micromolar) for reduced folates. In vivo
`experiments in mice using radiolabeled lometrexol or LY309887 have confirmed these
`suggestions and these data are presented in Subheading 7.
`The clinical significance of these observations is relevant to the risk of delayed, cu-
`mulative toxicity. Murine studies have demonstrated that lometrexol is taken up and
`stored in the liver (10). It is thought that the clinical observations of delayed and cumu-
`lative toxicity may result from gradual release of stored lometrexol over time resulting
`in an extended "y-phase plasma half life (11). The affinity of LY309887 for liver folate
`receptors suggest that it is less likely to be transported into the liver, and, because it is
`less efficiently polyglutamated than lometrexol, it should not be retained in the liver as
`well. The greater potency of LY309887 to inhibit GARFT suggests that lower doses of
`this compound may be used. Collectively, these properties suggest that LY309887 may
`have a reduced risk of delayed toxicity in humans.
`
`4. CELLULAR PHARMACOLOGY
`
`LY309887 showed potent in vitro cytotoxicity (ICso of 2.9 ruk!) against CCRF-CEM
`leukemia cells and was approx fivefold more cytotoxic than lometrexol (ICs0 of 15.2
`~ (2). The cytotoxicity of both compounds was reversed by hypoxanthine, but not by
`thymidine. The cell cycle effects of LY309887 were evaluated Using the human
`leukemia cell line CCRF-CEM. Exposure of cells to 29 nM of this inhibitor for 24-96 h
`resulted ina slowed progression of cells through S-phase and an increase in the number
`of S-phase ceils (12). Measurement of deoxynucleotide pools demonstrated marked de-
`creases in dATP pools in these cells within 6 h of exposure (13).
`
`5. IN VIVO ANTITUMOR ACTIVITY OF LY309887
`
`reduced
`amated.
`ires less
`fistered
`iver lo-
`in liver
`~ents to
`reading
`
`The dose-response curves of LY309887 and lometrexol against the murine C3H
`corlge-
`mammary tumor are shown in Fig. 1. LY309887 demonstrated antitumor activity over a
`ity and
`broad dose range in this model with 94% tumor inhibition seen at 1 mg/kg/dose and up
`misms,
`to 99% inhibition seen at 100 mg/kg/dose. LY309887 was a more potent antitumor agent
`¯ re im-
`especially since lometrexol was dosed more frequently (q2d × 5) compared to
`city of
`LY309887 (dosed q3d × 4). Thus LY309887 showed 99% inhibition of tumor growth
`ithelial
`from 3 to 100 mg/kg/dose with acceptable toxicity (lethality <20% at all efficacious
`astants
`doses), whereas lometrexol showed 97% inhibition of tumor growtlrat 400 mg/kg/dose.
`:sented
`No lethality was noted at any of the doses of lometrexol including the 400 mg/kg dose,
`nspec-
`the highest dose tested for !ometrexol in this study.
`of the
`The antitumor activity of LY309887 (or LY254155, the mixture of its "6-R" and "6-
`: affin-
`S" stereoisomers) and lometrexo! (or the mixture of its "6-R" and "6-S" stereoisomers,
`~F into
`LY237147) in a number of preclinica! models is summarized in Table 4. Maximum in-
`hibition was obtained at doses producing 25% or less lethality. There were no significant
`differences in the maximum inhibition between the LY254155 isomers and lometrexol
`
`IS was
`eadily
`
`Sandoz Inc.
`Exhibit 1012-0007
`
`JOINT 1012-0007
`
`

`
`266
`
`Mendelsohn, Worzalla, and X~allin_~
`
`Table 4
`Antitumor Activity of GARFT Inhibitors
`
`Tk iopheneb
`
`DDA THF~
`
`¯
`Tumor
`
`CX-1 colon
`GC3 colon
`HC1 colon
`VRC5 colon
`LX-1 lung
`MX-1 mammary
`BXPC3 pancreatic
`PANC1 pancreatic
`
`6 R
`6 R,S
`diastereomer diastereomers
`LY309887
`LY254155
`
`95% (30)
`97% (30)
`86% (30)
`
`52% (10)
`
`92% (30)
`
`64% (40)
`96% (40)
`98% (40)
`96% (5)
`98% (10)
`61% (40)
`87% (40)
`85% (40)
`
`6R
`diastereomer
`LY249543
`(lometrexol)
`
`94% (200)
`99% (50)
`
`81% (I00)
`
`67% (300)~
`
`6 R,S
`diastereomers
`LY237147
`
`73% (100)
`98% (100)
`94% (200)
`99% (100)
`
`76% (100)
`59% (50)
`
`’~Compound dosed ip at the indicated dose (mg/kg) and schedule; e.g., for the CX-1 colon xenograft.
`6R,S-diastereomer was dosed ip q3d × 4 at 40 mg/kg.
`~Dose schedule: q3d × 4.
`LDose schedule: q2d × 5.
`
`Antitumor Activity of LY309887 vs.
`Lometrexol Against C3H Mammary Tumor
`Regular Diet -- 10 Days After First Dose
`100- : ÷++ . . . .... - ~ ,/:~. ,~,,~, .... 100
`
`activity
`
`.80 ~
`
`w
`
`rn
`
`c(
`re
`al
`W
`
`m
`th,
`,60-- ~
`
`LY309887 Dosed I.P, "
`q3dx4
`
`d ~..~
`
`~
`40 ~
`Lometrexol Dosed I,P.
`q2d x 5 ~
`
`~
`
`~
`: 40.
`~
`--
`
`20-
`
`25% Le~all~
`/
`
`0.0o 0.01 0.1 1 10 100 500
`Drug Dosage (mg/kg)
`
`~ig. ]. Andtu~or Activity of L~309887 vs. Lome/~exo] Against C3~ ~a~m~ Tumor.
`
`isomers. Against the LX-1 lung, VRC5, HC1, and GC3 colon xenografts, both the
`LY309887 and lometrexol produced over 80% tumor inhibition. Against the MX-1
`mammm-y and CX-1 colon, both the thiophene isomers and lometrexol produced be-
`tween 60% and 80% inhibition. In these studies, the thiopfiene analogs were dosed q3d
`× 4 while IometrexoI was dosed q2d X 5.
`The ability of LY254155 to prevent tumor regrowth of the human colon xenograft
`
`xe
`
`co
`
`ap,
`hu
`sht
`of
`He
`tail
`EE
`lov
`ina
`pec
`
`Sandoz Inc.
`Exhibit 1012-0008
`
`JOINT 1012-0008
`
`

`
`:[ Wallin~
`
`Chapter 12 / Lometrexol and LY309887
`
`267
`
`Regrowth Study LY254155 (Thiophene Mixture of
`"6-R" & "6-S" Isomers) Against HC1 Human
`Colon Xenograft - Dosed I.P. q3d x 4
`
`10000=
`
`g lOOO,
`
`100=
`
`..r.o"~3 ontrols
`
`,A" 2.5 mg/kg
`
`5 rng/kg,. #
`
`10"
`14
`
`21 28 35 42 49 56 63 70 77 84
`Days Since Therapy Initiated
`Fig. 2. Regrowth Study LY254155 (Thiophene Mixture of "6-R" & "6-S" Isomers) Against HC 1 Hu-
`man Colon Xenograft.
`
`was investigated (Fig. 2). A single course of LY254155 (mixture of"6-R" and "6-S" iso-
`mers of the thiophene) was given starting 14 d after inoculation with the HC1 human
`colon xenograft. At the lowest doses, tumor growth was slowed; at higher doses tumor
`regrowth was prevented for at least 42 d. The highest dose of 40 mg/kg in this study was
`also very effective, but two of the seven mice died (>25% lethality), and thus this data
`was not included. At the 10 mg/kg dose, three of six mice were tumor free, and at the 20
`mg/kg dose~ four of six mice were tumor free 39 d after’completion of a single course of
`therapy.
`Similar results were obtained using GC3 xenograft tumors. Thus for two colon
`xenograft models, the LY254155 prevented growth of tumor for periods of 4-7 wk, and
`in one of these tumor models, the HC1 colon tumor, a single course of the thiophene re-
`suited in a majority of mice being tumor free more than 5 wk after the therapy had been
`completed.
`
`6. THERAPEUTIC INDEX DETERMINATIONS
`
`The activities of LY309887 and lometrexol were compared using calculations of ther-
`apeutic index ranges for the two drugs. For this comparison, both drugs were tested in
`human xenograft tumor models (dosing schedule: q3d × 4). Ideally these comparisons
`should be made using tumor models in which both drugs were higlily active. The ratio
`of LDio divided by the ED9o would give therapeutic indices relevant to clinical safety.
`However, in some tumor models, a maximal tumor growth inhibition of 70% was ob-
`tained. Therefore, calculations of therapeutic index were based on the ratio of LDlo to
`ED7o (2). Calculations of therapeutic indices using ED9o values would give slightly
`lower values. The LDlo for these studies was assessed in nontumor-bearing mice to elim-
`inate uncertainty regarding cause of death, i.e., from tumor burden vs drug toxicity, es-
`pecially since lethality resulting from antifolate toxicity was delayed for many days:
`
`)/T/ers
`1_47
`
`100)
`lOO)
`200)
`100)
`
`lOO)
`~o)
`
`.enografl,
`
`}th the
`MX-1
`ed be-
`.’d q3d
`
`.ograft
`
`Sandoz Inc.
`Exhibit 1012-0009
`
`JOINT 1012-0009
`
`

`
`268
`
`Mendelsohn, Worzalla, and Walling
`
`c
`
`Table 5
`Estimates of Therapeutic Indexa Ranges for LY309887
`and Lometrexol Based on Ratio of LDi0/EDT0
`
`LY309887 Lometrexol (LY249543)
`
`PANC- 1 pancreatic
`LX-1 lung
`MX-1 mammaxy
`HC1 colon
`VRC5 colon
`
`10--30
`3-10
`1-3
`3-10
`3-10
`
`3-10
`1-3
`1-3
`1-3
`3
`
`~Therapeutic index range:
`(e.g., for LY309887 against PANC-1 = LDIo!EDvo = (30-100 mg/kg)/(3
`mg/kg) = 10-30
`All values for LDm and EDvo based on q3d X 4 ip dosing with regular diet.
`LDlo values determined in C3H female mice.
`ED7o values determined in CDI Nu/Nu mice.
`
`Lethality was determined in C3H mice which were used for testing murine tumor mod-
`els. However, the human xenograft models utilized CD1 Nu/Nu mice. Thus the thera-
`peutic index calculations use efficacy data from this strain of mouse.
`Compound LY309887 was more potent than lometrexol in producing antitumor ac-
`tivity in several-rm0"~dels, and it also was more potent in producing lethality in mice. In
`the PANC-1 pancreatic tumor and the LX-1 human lung xenografts, LY309887 was
`about 100 times more potent in its antitumor activity compa[ed to lometrexol when both
`were dosed q3d × 4. The LDlo for LY309887 was betWeen 30 and 100 mg/kg, whereas
`for lometrexol, an LDm between 1000 and 3000 mg/kg was obtained. Thus, LY309887
`was approx 30 times more potent for producing lethality .in mice compared to lome-
`trexol. Estimates of therapeutic indices were calculated as shown in Table 5. In PANC-
`1 pancreatic, LX-1 lung, HC1 colon, and VRC5 colon xenograft tumor models,
`preliminary estimates of therapeutic index showed an approx threefold greater thera-
`peutic index for LY309887 compared to lometrexol. In MX-1 mammary tumors, both
`compounds were less active, and there was no difference in therapeutic index in this tu-
`mor model.
`
`7. THE EFFECT OF DIETARY FOLATE
`ON DISPOSITION OF LOMETREXOL AND LY309887 IN LIVER
`
`In phase I clinical trials with lometrexol, unexpected delayed and cumulative toxicity
`in patients was encountered (1). The observed toxicities, mucositis and myelosupres-
`sion, are classically associated with antifolate therapy. However, the duration of toxic-
`ity following a single-drug dose and the cumulative nature of the toxicity were novel
`observations of antifolate toxicity. It was hypothesized that cancer patients may be
`marginal in their folic acid stores. To this end, a folate-deficient murine model was es-
`tablished to characterize biochemical and pharmacological effects of low dietary folat~
`(LFD) on the efficacy and delayed toxicity of antifolate inhibitors.
`In mice receiving a LFD for 2 wk, Schmitz et al. showed that folate pools were re-
`duced in plasma, liver, intestine, and tumors (14). In addition, significant changes were
`noted in the density of FR in tumors and liver and in the isoforms expressed in tumors
`
`Fi
`
`()
`
`ol
`
`L
`ui
`
`P~
`
`8
`C
`
`3(
`th
`
`Sandoz Inc.
`Exhibit 1012-0010
`
`JOINT 1012-0010
`
`

`
`t Walling
`
`Chapter 12 / Lometrexol and LY309887
`
`269
`
`-13-- Lometrexol Standard Diet
`
`--d-- Lometrexol Low Folate Diet
`
`"-G’- LY309887 Standard Diet
`
`~ LY309887 Low Folate Diet
`
`2500 -
`
`2000.
`
`500
`
`O,
`
`0
`
`20
`
`40
`
`60
`
`80 t00
`
`120 140 160 t80
`
`Time(hours)
`
`Fig. 3. Total Disposition of [14C]-lometrexol and [14C]-LY309887 in Murine Liver.
`
`(15). Increases in tumor and liver FPGS activity were also noted (16). Whole-body au-
`toradiography studies in mice on standard diet or LFD receiving either [14C]-lometrexol
`or [14C]-LY’309887 demonstrated significantly higher accumulation of drug in livers of
`mice on LFD (]0, 7). The accumulation and polyglutamation state of lometrexol and
`LY309887 in livers of mice on SD and LFD were compared by dosing mice iv with eq-
`uitoxic doses (LDso doses in LFD mice) of radiolabeled parent compound. The total ac-
`cumulation of each drug in liver.was determined over a 7-d period (Fig. 3).
`Polyglutamates were also determined by reversed-phase HPLC (7).
`Regardless of diet, more lometrexol accumulated in murine liver than LY309887.
`Furthermore, animals on LFD accumulated roughly 6-10-fold more drag than animals
`on SD and clearance of drug from livers of LFD mice was slower than clearance from
`SD mice. Polyglutamation profiles showed that on SD, the most common metabolite af-
`ter 24-h for both drugs was pentaglutamate (80-90%). In mice on LFD receiving
`LY309887, penta- and hexa-glutamyl-metabolites were still the predominant forms; 70
`and 30%, respectively. In contrast, lometrexol was polyglutamated mgre extensively: an
`additional percentage, approx 15%, was recovered as the septa- and octa-glutamyl forms
`(17).
`
`8. THE EFFECT OF LFD AND DIETARY FOLATE SUPPLEMENTATION
`ON THE EFFICACY AND TOXICITY OF LOMETREXOL AND LY309887
`
`In mice on a LFD for 2 wk, the toxicity of lometrexol and LY309887 ine?eased
`300-1000-fold. Antitumor activity cannot be assessed because of the lethality of
`these agents. Oral supplementation with folic acid (0.6-600 mg/kg) restores sensitivity
`
`?r mod-
`e thera-
`
`mor ac-
`nice. In
`87 was
`en both
`vhereas
`309887
`) lome-
`PANC-
`nodels,
`: thera-
`:s, both
`this tu-
`
`oxicity
`supres-
`." toxic-
`’, novel
`nay be
`vas es-
`r folate
`
`ere re-
`’,s were
`tumors
`
`Sandoz Inc.
`Exhibit 1012-0011
`
`JOINT 1012-0011
`
`

`
`270
`
`Mendelsohn, Worzalla, and Walling,
`
`Ch
`
`to the antitumor activity of both GARFT inhibitors. High doses of folic acid (>600
`mg/kg) eliminated both toxicity and antitumor activity.
`The therapeutic indexes (LDlo/ED90) of LY309887 and lometrexol were determined
`over a range of supplemental folic acid doses in two antitumor models, the human
`xenograft GC3 colon and the routine mammary tumor C3H (Table 6). The data show
`that increasing folic acid supplementation doses from 0.0 to 6-15 mg/kg/d resulted in an
`enhanced therapeutic index ranging from 6 to 60 for LY309887 in both models. Fur,
`thermore, the ability to delay regrowth of the GC3 tumors over a broad dose range was
`only seen at the higher doses of folate supplementation. A small but modest increase in
`lometrexol’s therapeutic index to 2-5 was also observed. Higher doses of folic acid sup-
`plementation resulted in less robust increases in the therapeutic index.
`
`9. HUMAN FOLATE STATUS
`
`The folate status of cancer patients has not been systematically evaluated. However,
`early studies reported decreased serum folio acid activity in patients with metastatic can-
`cer (18-20). Other investigators have demonstrated decreased urinary clearance of a
`folic-acid load (21,22). Saleh et al. demonstrated that patients with metastatic disease in-
`corporated more folic acid into their reduced folate pools, had decreased catabolism of
`folate and more rapid clearance of serum folate than controls even in the presence of
`maintained serum 5-CH3-THF concentrations (22). They concluded that patients were
`folate deficient and that there was an increased demand for folate in patients with ma-
`lignant disease. In these patients, variability in the metabolism, pharmacokinetics, and
`toxicity of classical antifolates compared to humans with normal folate status would not
`be unexpected. Furthermore, dietary supplementation with folic acid may "normalize"
`the dose response.for achieving antitumor activity and reduce, toxicity to normal tissues
`by restoring folate pools in tissues having low folate requirements, without meeting the
`high folate demands of rapidly dividing tumor cells.
`The biochemical pathways that utilize folate cofactors also require adequate amounts
`of vitamins B I2 and B6. Thus, the status of all three vitamins in patients may significantly
`influence the severity of toxicity observed during chemotherapy. R. Allen and his col-
`leagues have established that measuring specific amino acid metabolites, especially ho-
`mocysteine, N-methyl glycine and others, from these metabolic pathways provides a
`more sensitive and reliable assessment of patient vitamin status (23). These surrogate in-
`dicators of functional folate status are more indicative of deficiencies and more respon-
`sive to dietary supplementation.
`
`10. CLINICAL EVALUATION OF LOMETREXOL AND LY309887
`
`Cancer chemotherapy was born in 1948 with the discovery that the antifolate an-
`timetabolite aminopterin, an inhibitor of dihydrofolate reductase (DHFR), induced re-
`missions in patients with acute lymphoblastic leukemia (24). Over the last 50 yr,
`intensive structure-activity studies directed at the folate pathway led initially to the iden-
`tification of methotrexate (25), also an inhibitor of DHFR, and 5 fluorouracil (5FU) an
`inhibitor of the enzyme thymidylate synthase (TS) (26). More recently a number of pure
`
`TS
`den
`dict
`witl
`met
`CB:
`has
`of c
`eral
`vek
`mul
`pm-i
`ter c
`bios
`aste
`kno
`basi
`testi
`tire
`sch~
`pha.,
`beer
`tion
`ducl
`lom~
`scrit
`
`Sandoz Inc.
`Exhibit 1012-0012
`
`JOINT 1012-0012
`
`

`
`~Tallin~
`
`(>600
`
`mined
`human
`t show
`dinah
`s. Fur-
`ge was
`:ase in
`,d sup-
`
`¢¢ever,
`ic can-
`e of a
`ase in-
`ism of
`nce of
`~ were
`:h ma-
`:s, and
`fld not
`talize"
`issues
`ng the
`
`lounts
`cantly
`is col-
`ly ho-
`ides a
`ate in-
`’.spon-
`
`tO an-
`ed re-
`30 yr,
`iden-
`U) an
`f pure
`
`Chapter 12 / Lometrexol and LY309887
`
`271
`
`Table 6
`Effect of Increasing Folic Acid Supplementation on the
`Therapeutic Index
`
`Tkerapeutic Index
`
`Folate Supplement
`mg~g/d
`
`Z~lmor
`
`GC3 colon
`
`C3H mammary
`
`n.d.: not determined.
`
`LY309887
`
`0
`3 (2,4)
`25 (25,25)
`60
`n.d.
`0
`1.4 (1.3,1.5)
`9 (6, 12)
`6
`n.d.
`
`LD l o/EDgo
`lometrexol
`
`0
`0.6
`1.5
`n.d.
`2
`0
`2
`5
`n.d.
`1
`
`none
`0.6
`6.0
`15.0
`60.0
`none
`0.6
`6
`15
`60
`
`TS inhibitors have been studied, the lead compound being CB3717. Despite initial evi-
`dence of activity (27,28), this compound failed in the clinic largely because of unpre-
`dictable renal toxicity, which was thought to result from crystallization of the compound
`within the kidney (27). The replacement of the l’-amino and 10 propargyl groups to a
`methyl group and the replacement of the l’,4’-phenyleneto 2’,5’-thienyl bioisostere on
`CB3717 gave a more potent and soluble TS inhibitor, raltitrexed (ZD1694). Raltitrexed
`has recently been shown in randomized comparative trials to be active in the treatment
`of colorectal cancer. These studies have led to the granting of a product license in sev-
`eral countries including the U.K. Other antifolate antimetabolites are currently under de-
`velopment including the 5FU prodrugs, capecitabine (29) and UFT (30), and the
`multitargeted antifolate (MTA, LY231514) (31-33). However, no specific inhibitors of
`purine biosynthesis are yet utilized in routine clinical practice. The first molecule to en-
`ter clinical trials that was directed at inhibition of an enzyme essential for de novo purine
`biosynthesis, glycinamide ribonucleotide transformylase (GARFT), was the 6R di-
`astereomer of dideazatetrahydrofolic acid (DDATHF), an analog of tetrahydrofolate,
`known as lometrexol (34,35). Lometrexol was selected for clinical development on the
`basis of its preclinical activity in vitro and in vivo (36,37). Hematologic and gastroin-
`testinal toxicity was dose limiting in both mice and dogs. The dog was the most sensi-
`tive of the an