~9-1353.00/0
`Copyri~t © by The American Society for Pharmacology and Expe~mental The~peuti¢~
`All rights of reproduction in any form reserved.
`~OL/~JLAR PHA~ACOI~6~, "4&459--471 (1996).
`
`Carrier- and Receptor-Mediated Transport of Folate
`Antagonists Targeting Folate-Dependent Enzymes: Correlates
`of Molecular-Structure and Biological Activity
`
`(3. ROBBIN WESTERHOF, JAN H. SCHORNAGEL, IETJE KATHMANN, ANN L JACKMAN, ANDRE ROSOWSKY, RONALD
`FORSCH, JOHN B. HYNES, F. THOMAS BOYLE, GODEFRIDUS J. PETERS, HERBERT M. PINEDO, and GERRIT JANSEN
`
`Department of Oncology, University Hospital Vdje Universiteit, Amsterdam, 1007 MB The Netherlands (G.R.W., LK., G.J.P., H.M.P., G.J.),
`Department of Internal Medicine, Netherlands Cancer Institute, Amsterdam, 1066 CX The Netherlands (J.H.S.), Drug Development Section,
`Institute of Cancer Research, Sutton, Surrey, 5N2 5NG United Kingdom (A.L.J.), Dana-Farber Cancer Institute and Department of Biological
`Chemistry and Molecular Pharmacology, Hamard Medical School, Boston, Massachusetts 02215 (A.R., R.A.F.), Department of Pharmaceutical
`Sciences, Medical Univetsily of South Carolina, Charleston, South Carolina 29425 (J.B.H.), and Z
`"cals, Macclesfleld,
`Cheshire, SKIO 4TG United Kingdom (F.T.B.)
`
`Received January 30, 1995; Accepted June 14, 1995
`
`SUMMARY
`The transport propertias and growthAnhibitocy poten’dal of 37 classic
`
`(mFBP), or both. The ;,,becellular targets of these drugs were dihy-
`
`ge:~h-~fbib~ potential as a function of cellular RFC/mFBP ex-
`pression, and the protective effect of either FA or leucovodn against
`
`d~k~.~’~k: a~d (CB3717), ZD1694, 5,~k= add (U~HQ),
`
`.were we~ tolerated. Gmwlt~Wt~lk~ studies iderfdlied a sedes of
`
`turas) or mFBP (CB3717, 2-NH=-ZD1694, or 5,8-dk:ieazabo~ic
`
`port pathways (e.g., DOATHF, ZD1694, BW1843U89, or LY231514).
`
`substrate affinity for all analogues except CB3717, 2-NI-I~-ZD1694,
`
`For more than 40 years, the folate antagonist MTX has had
`an established role in cancer chemotherapy, both as a single
`
`This study was supported by Dutch Cancer Society Grant IKA 89-34. G. J.
`is a recipient of a fellowship from the Royal Netherlands Academy of Arts and
`Sciences.
`
`agent and in combination regimens (I). Knowledge of the
`factors that contribute to the preclinical and clinical activity
`of MTX, i.e., membrane transport, intracellular retention,
`and inhibition of the target enzyme DHFR, has provided a
`solid basis for the design and synthesis of novel antifolates
`that are either transported more efficiently, have a prolonged
`
`ABBREVIATIONS: RFC, reduced folate/methotrexate carder;, mFBP, membrane-associated folate binding protein; FCS, fetal calf serum; HBSS, HEPES
`balanced salt solution; LV, L-leucovodn (L-5-formyttetrahydrofotate); 5-CH3-THF, 5-metf~elrahy~-ofolate; DHFR, d~drofolate reductase; FPGS,
`folytpol~lutamate synthetmm; GARTF, glycinamide dbonucleotide transforrnytase; TS, thymidylate synthase; FA, folic acid; PteGlu, pteroyl glutamate
`(FA); Ivrrx, methotrexate; 2-dMTX, 2-desamino-MTX; 2-CHa-MTX, 2-desamino-2-meft~M3X; AMT, aminopte~; 2-dAIvrr, 2-desamlno-AMT; 2-CI~-
`AMT, 2-desamino-2-methyI-AM’r; 10-EdAM, 10-etttyl-10-d~_ ==u~minoptedn; PT523, N=-(4-amlno4-d~xypteroy~-(hemlpflltBIo~-omilJ3ine;
`DDATHF, 5,10-dideaza-5,6,7,8-tetrahydrofolic acid; 5-d(i)H4PteGlu, 5-desza-5,6,7,8-tetrahydroisofotic acid; N~-CH3-5-d(i)H4PteGlu, Ne-rnet~-5-
`deaza-5,6,7,8-tetrahydroisofolic acid; 5-dPteHCysA, N=-(5-de~v~.nteroyl)-L-homocystetc acid; 5-dPteAPBA, N=-(5-de~_ ,~.nteroyl)-oL-2-andno-4-phos-
`phonobutanoic acid; 5-dPteOrn, N=-(5-de~v-,nteroy~-omithine; 5-dH,PteHCysA, N=-(5-deaza-5,S,7,8-tetrahydropteroyl)-L~ acid;
`5-dI-~PteAPBA, N=-(5-deaza-5,6,7,8-telmhydropteroyl)-DL-2-arnino-4~p/no~c~:x~Jtanol¢ acid; 5-dH~PteOm, N=-(5-deaza-5,6,7,8-tetrahydropteroyl)-L-
`orni~ine; CB3717, N~ °-propargyl-5,8-dideazafolic acid; IC1-198,583, 2-desamino-2-me~lW~°-propargyl-5,8-dideezafolic acid; 4-H-IC1-198,583, 4-de-
`oxy-IC1-198,583; 4-OCH3-1C1-198,583, 4-rnethoxy-ICl-198,583; Glu--*Va14CI-198,583, valine-ICl-198,583; Glu-,Sub-ICl-198,583, 2-amino-suberate-
`IC1-198,583; 7-CH3-1C1-1 98,583, 7-methyl-ICl-198,583; ZD1694, N-[5(N-(3~4-dihydr~~2-methy~-4-~x~quinaz~~in-6-y~-me~~n~~)amin~)2-thany~)]-L-g~utamic
`acid; 2-NH~-ZD1694, 2-amino-ZD1694; BW1843U89, (S)-2-{5-(((1~3~methy~1-~x~-~?/qt~‘~m-9-y~)me~j~)an~)-1-~x~-24s~ind~n~
`gluladc acid; LY231514, N-(4-(2-(2-amin~-4~7-dihydr~4-~x~-~H-pyrr~k)~2~3-~]pydmidin-5-y~)e~.~)banz~y~]-L-g~u~amic acid; IAHQ, 5,8-dideezeisofollc
`acid; 2-d-IAHQ, 2-desamino-lAHQ; 2-CH~-dlAHQ, 2-desamino-2-methyl-lAHQ; 5-d(i)PteGlu, 5-deazaisofolic acid; hP-CHa-5-d(i)PteGlu, Ne-methyl-5-
`deazaisofolic acid; hP-CHO-5-d(i)PteGlu, Ne-formyl-5-deazaisofolic acid; AG337, 3,4-dihydro-2-amirm-6-methyl-4-oxo-5-(4-pyrtdytlNo)qulnazol~;
`AG377, 2,4-diamino-6-[N-(4-(pher~jlsu ~ino]q uinazoline.
`
`Sandoz Inc.
`Exhibit 1022-0001
`
`

`
`!
`
`intracellular retention, or are more potent inhibitere of the
`target enzyme (2-13). Other than DHFR, key enzymes in
`folate metabolism, such as TS, GARTF, and FPGS, have been
`recognized as potential targets for folate-based chemothera-
`peutic agents (14-19).
`Traditionally, the biological activity of novel antifolates is
`evaluated with in vitro model systems (usn~lly leukemia cell
`lines) that express the RFC as the major transport route for
`natural reduced folate cofactors (e.g., LV and 5-CHs-TI-IF)
`and classical antifolate compounds (6, 20-22). Although the
`role of the RFC in antifolate transport is undisputed, an
`increasing number of reports suggest that at least one other
`folate trsn~port protein, an mFBP, may have an additional
`role in antifolate uptake (23-28X A functional role of mFBP
`in folate uptake has been demonstrated in a number of in
`vitro (28-32) and in vivo (33) studies in which mFBP-medi-
`ated transport of natural reduced folate cofactors supported
`cellular growth at nanomolar extracollular concentrations.
`Other reports have demonstrated mFBP-mediated transport
`of antifolates, particularly ff the mFBP exhibits a high bind-
`ing affinity for these drugs (23, 34-36). Because the mFBPs
`appear to be expressed in a range of normal and neoplastic
`cells and tissues (37-42), further analysis of the potential
`role of this protein in antifolate drug uptake is warranted.
`The RFC and mFBP are’ structurally and functionally un-
`related proteins. The RFC is an integral membrane protein
`with a molecular weight ranging from 43-48 kDa in rodent
`cells (43, 44) to 80-120 kDa in human cells (42, 45, 46). The
`mFBP is a 38-40-kDa protein that is linked to the plasma
`membrane via a glycosylphosphatidylinositol anchor (47, 48).
`The mechanism of RFC-mediated folate and antifolate up-
`take has been the subject of extensive studies, which have
`demonstrated the uptake process is both temperature and
`energy dependent and able to be inhibited by structurally
`unrelated anions (6, 49-52). These observations support the
`hypothesis that an anion exchange mechanism may be in-
`volved in carrier-mediated antifolate uptake (6). At least two
`mechanisms have been described for mFBP-mediated folate
`uptake: one by the classic receptor-mediated endocytosis
`pathway (53) and the other, most extensively studied in
`monkey kidney MA-104 cells, via a novel mechanism called
`potocytosis. Potocytosis (27) comprises a series of events that
`involve clustering of folate-bound mFBP molecules in mem-
`brane invaginations (caveolae), followed by transient closure
`of caveolae from the extracollular medium and acidification
`of their lumen, after which the folate molecule is dissociated
`from the mFBP and translocated across the plasma mem-
`brane via a specific carrier protein (27, 54). Whether the
`latter protein is the RFC is not clear.
`The klnet~c properties of the RFC and mFBP for transport
`also differ significantly. The RFC exhibits a high affinity (Kin,
`1-10/m) for natural reduced folate cofactors and antifolates
`such as MTX (6), which can be internalized at a maximal rate
`of 10-20 molecules per minute per transport molecule (20). A
`characteristic feature of the RFC is its poor affinity for FA
`(Kin, 200-400 ~), which is in contrast to that of mFBP.
`mFBP has a high affinity (Kd, 1-10 n~) for FA and reduced
`folate cofactors, which may favor uptake at a physiological
`folate concentration of 5-50 nM (26). On the other hand, the
`recycling rate of mFBP is much slower than that for the RFC:
`from -30 rnin in MA104 cells (27, 55) to 5 hr in L1210 cells
`(56, 57).
`
`We report on a comprehensive study of the functional as-
`pects of both the RFC and mFBP in antifolate drug transport,
`which allowed us to directly compare the efficiency of these
`transport proteins. As models, three CCRF-CEM human leu-
`kemia cell lines were used, which are characterized by nor-
`real, upregulated, or defective RFC transport (22, 58). For
`comparison, three (variant) L1210 murine leukemia cells
`were included that express the RFC (50), coexpress the
`mFBP and the RFC (59), or lack functional RFC activity
`while retaining mFBP (60). In the present study, we used a
`series of folate-based inhibitors of DHFR (MTX and AMT)
`(61), GARTF (DDATHF and 5-d(i)H4PteGlu) (15, 62), FPGS
`(glutamate side chain-modified 5-deazafolic, 5-deazatetrahy-
`drofolic acid structures) (63), and TS (CB3717, ZD1694, and
`IAHQ) (64-66). These compounds and their analogues with
`structural alterations in the pteridine/quinazoline ring (sub-
`stitutions at C-2, C-7, N-3, and C-4), p-smlnobenzoate ring
`(thiophene for benzoyl), the C~-N1° bridge, and the glutamate
`side chain (replacement by 2-aminosuberate, valine, and or-
`nithine) were used in the study. All of these compounds were
`evaluated for changes in the substrate specificity for the RFC
`and mFBP; the growth-inhibitory effects against cell lines
`with different expression levels of the RFC, mFBP, or both;
`and the protective effect of LV, FA, or both.
`The structure-activity relationships described in the
`present study for antifolate transport via the RFC, mFBP, or
`both may be of predictive value in assessing the preclinical
`and clinical activity of antifolate drugs and can also be used
`for the rational future design of new antifolates.
`
`Experimental Procedures
`
`Materials. RPMI 1640, with and without FA, and FCS, dialyasd
`and nondialysed, were obtained from GIBCO (Grand Island, NY). FA
`and DL-5-CHs-THF were purchased from Sigma Chemical Co. (St.
`Louis, MO). LV (L-stereoisomer) was a gift from Lederle (Pearl River,
`NY). [3’,5’,7,9-SH]FA (35 Ci/mmo]), [3’,5’,7-SH]MTX (20 Ci/mmol),
`and [3,4-SH]glutamic acid (56.6 Ci/mmol) were obtained from
`Moravek Riochemicals (Brea, CA). [SH]FA and [SH]MTX were puri-
`fied as described previously (20, 59, 60). All other chemicals were of
`the highest purity available.
`Drugs. The chemical structures of the antifoiste compounds de-
`scribed in the study are depicted in Fig. 1. The inhibitory potential of
`these compounds against their target enzyme, as well as their sub-
`strate ~fl~nity for FPGS, are described in Tables 1 (DHFR, GARTF,
`and FPGS inhibitors) and 2 (TS inhibitors). It should be noted that
`these data were taken from the indicated references and that the
`experimental conditions were not identical. Nevertheless, these data
`were included to explain possible growth-inhibitory effects on the
`basis of poor enzyme inhibition or substrate affinity for FPGS. Fur-
`ther details regarding chemical synthesis of the drugs are given in
`references as indicated in Tables 1 and 2.
`DHFR inhlbitors. MTX was a gi~ from Pharmachemie (Haar-
`lem, The Netherlands). AMT was purchased from Sigma Chemical
`Co. 2-dMTX, 2-CHs-MTX, 2-dAMT, and 2-CHs-AMT were synthe-
`sized as previously described (61). 10-EdAM was a ~ from CIBA-
`GEIGY (Basel, Switzerland). PT523 (67, 68) was a gift from Dr. W. T.
`McCulloch (Sparta Pharmaceuticals, Research T~gle Park, NC).
`GARTF Inhlbitors. The synthesis of 5-d(i)I-I~PteGlu and Ng-CHs-
`5-d(i)H~PteGlu has been described previously (62). DDATHF was a
`gift from the late Dr. G. B. Grindey (Lilly Research Labs., Indianap-
`otis, IN).
`FPGS i~tbitors. The 5-deazafolate analogues 5-dPteHCysA,
`5-dPteAPBA, 5-dPteOrn, 5-dH~PteHCysA, 5-dH~PteAPBA, and
`5-dH~PteOrn were synthesized by Rosowsky et al. (63).
`
`Sandoz Inc.
`Exhibit 1022-0002
`
`

`
`Structure-Activity Relationshilm of Antifolate Transport 461
`
`!
`
`o
`
`DHF~RGARTF F~PGS
`
`TS
`
`Rg. 1. Molecular structure of folate-based inhibltors of DHFR, GARTF, FPGS, and TS.
`
`TS ;-hibitors. CB3717 (64, 69) and ICI-198,583 (70) served as
`basic structures for the following ~nalogues: 3-deaza-ICI-198,583,
`4-H-ICI-198,583, 4-OCHa-ICI-198,583, Glu--~Val-ICI-198,583 (71,
`72), 7-CHa-ICI-198,583 (73, 74), Glu-~Sub-ICI-198,583 (72), ZD1694
`(65), and 2-NH2-ZD1694 (75, 76). The isofolic analogues IAHQ, 2-des-
`Amino-IAHQ, and 2-des*mino-2-methyl-IAHQ and the 5-deazaisofo-
`lates 5-d(i)PteGlu, N~-CHa-5-d(i)PteGin, and l~-CHO-5d(i) eGlu
`were synthesized as described previously (62, 66, 77, 78). LY231514
`(79) was a gift from the late Dr. G. B. Grindey. BW1843U89 (80) was
`
`provided by R. Ferone (Burroughs Wellcome Co., Research Triangle
`Park, NC). AG337 (81, 82) and AG377 (83) were gitts from Agouron
`Pharmaceuticals, Inc. (San Diego, CA).
`Cell culture. The cell lines used in the present study include
`RFC-expreseing CCRF-CEM human leukemic lymphoblaste (20-22);
`a variant of this line, CEM/MTX, lacking functional RFC (58); and
`another variant (CEM-7A), characterized by a 30-fold overexpression
`of the RFC compared with CCRF-CEM coils (22). In addition, the
`following murine cell lines were used: wild-type RFC expressing
`
`Sandoz Inc.
`Exhibit 1022-0003
`
`

`
`Charactedstics of folate-based inhibitors of DHFR, GARTF, and FPGS with respect to enzyme inhibition and sublimate activity for
`
`Compound Main target FPGS substrete activity~ Ref.
`DHFR TS GARTF
`
`! TABLE I
`
`MTX
`2-dMTX
`2-CH3-MTX
`AMT
`2-dAMT
`2-CH3-AMT
`10-EdAM
`PT523
`DDATHF
`5-d(i)H4PteGlu
`N~-CH3-5-d(i)H4PteGlu
`
`5-dPteHCysA
`5-dPteAPBA
`5-dPteOm
`5-dH4PteHCysA
`5-dH4PteAPBA
`5odH4PteOm
`
`DHFR
`
`GARTF
`
`FPGS
`
`0.024
`6
`>20
`0.025
`19
`>50
`2.3 pM
`0.052
`>100
`>1000
`>1000
`
`0.4
`4.7
`8.3
`12
`220
`
`29
`
`>10
`
`>100
`> 100
`>100
`
`55
`300
`
`2100
`4200
`
`0.12
`5.3
`26
`
`0.19
`0.047
`3.7
`
`+
`NR
`NR
`+ + +
`NR
`NR
`+++
`NS
`+++
`+ +
`+ + +
`
`570
`9.0
`5.7
`22
`3.2
`0.027
`
`61,110
`61
`61
`61, 87
`61
`12, 61
`111
`68
`112,113
`62
`62
`
`63
`63
`63
`63
`63
`63
`
`"The data in this table are obtained from the indicated references. Please note that Kin determinations of enzyme inhibition and substrate activity determinetion
`for FPGS may not have been carded out under Identical (x)ndlttons with respect to substr~e, Inhibitor, and cofactor concentration. For experimental details, see the
`indicsted references.
`~ FPGS subetrate activity (Kin) is defined as follows: Km 5-25 M,M: (+++); Km 25-100/ZM: (++); Km > 100 ~: (+). NR, not reported; NS, nonsubstrate.
`
`TABLE 2
`Characted~cs of folate-based inhibitom of TS with respect to enzyme Inhibition and substrate activity for FPG~’
`
`Compound Main target FPGS substrate activity~ Ref.
`DHFR TS GARTF
`
`MTX
`CB3717
`IC1-198,583
`3-deaza-ICl-198,583
`4-H-IC1-198,583
`4-OCH3-1C1-198,583
`Glu--,VaI-ICI-198,583
`Glu--*Sub-ICl-198,583
`7-CH3-1C1-198,583
`ZD1694
`2-NH2-ZD1694
`BW1843U89
`LY231514
`IAHQ
`2-dlAHQ
`2-CHa-IAHQ
`5-d(i)PteGlu
`N~-CH3-5-d(i)PteGIu
`NS-CHO-5-d(i)PteGlu
`AG337
`AG377
`
`DHFR
`TS
`
`0.024
`0.91
`5.5
`
`12
`
`0.11
`0.32
`3.1
`0.21
`0.25
`1.95
`
`0.02
`0.046
`6.3
`7.9
`6.0
`0.042
`0.018
`0.028
`0.65
`0.44
`0.00009
`0.44
`1.3
`25
`17
`7.1
`0.48
`28.8
`0.011
`0.033
`
`> 17
`
`85.9
`>100
`> 100
`
`+
`+ +
`+ +
`+
`
`NS
`+ + + +
`
`+ + + +
`+ + + +
`+++
`+ + +
`+ + +
`+
`+ +
`+
`NS
`NS
`
`110
`7O
`
`110
`70,
`69,
`72
`114
`114
`71
`72
`73
`65
`76
`80
`79
`78
`78
`78
`62
`62
`62
`81
`63
`
`¯ The data in this table are depicted from the indicated references. Please note that KI,.~ determinations of enzyme inhibition and substrat.e activity determination
`for FPGS may not have been ~ out under identical conditions with respect to subetr~e, inhibitor, and cofactor concentration. For expenmental details, see the
`indicated references. Moreover, it should be noted that the KI for the compounds listed may not reflect the increased potency of inhibiting TS when polyglutamete
`
`~ FPGS substrate activity (Kin) is defined as follows: Km < 5/~M: (++++); Km 5-25 p.u: (+++); Km 25-100 p~: (++); K,, > 100 p~: (+). NS, nonsubstrate.
`
`L1210 leukemia cells, one variant (L1210-B73) expressing beth the
`RFC and mFBP (59), and another variant (L1210-FBP) in which
`mFBP is the only functional transport protein (60). (This cell line
`was erroneously reported to be derived from CEM/MTX cells, but on
`karyotype and restriction polymorphism analysis it appeared to be a
`subline of L1210-B73 cells lacking functional RFC activity (60a).
`Parental cells (CCRF-CEM and L1210) and CEM/MTX cells were
`
`grown in RPMI-1640 medium cent~ining 2 /~M FA supplemented
`with 10% FCS, 2 mM L-ghit~mine, and 100 units/m] each of penicillin
`and streptomycin. CEM-7A, L1210-FBP, and L1210-B73 cells were
`grown in folate-free RPMI 1640 supplemented with 10% dialysed
`FCS, antibiotics, and ghit~mine as described. LV was added as folate
`source at final concentrations of 0.2, 1.0, and 1.0 nM, respectively
`(unless otherwise indicated). The cell cultures ofL1210, L1210-FBP,
`
`Sandoz Inc.
`Exhibit 1022-0004
`
`

`
`!
`
`and L1210-B73 cells also contained 50 ~M ~-mercaptoethanol. Cells
`were kept at 37° in a humidified atmosphere with 5% CO2.
`DHFR and FPGS levels in cell lines. The DHFR concentration
`of the cell lines used in the present study as assessed by [SH]MTX-
`bindin~ (84) was comparable in CEM, CEM/MTX, and CEM-7A cells
`(1.3 pmol/mg protein). Likewise, DHFR levels in L1210 cells and
`L1210-FBP cells were similar (2.6 pmol/mg protein). The enzyme
`level in L1210-B73 cells was 2.8-fold higher (7 pmol/mg) than in
`L1210 cells. FPGS activity was measured in all cell lines as described
`by Jansen et al. (85) with 250 /~M MTX as a substrate. Enzyme
`activities expressed as picemoles of MTX-[SH]glutamato formed per
`hour per mi|ligram of protein were as follows: 402 (L1210-FBP), 618
`(L1210-B73), 829 (CEM), 992 (CEM-7A), 1168 (L1210), and 1200
`(CEM/MTX) (results not shown).).
`ItFC substrate specificity. CEM-7A cells (3 × l0s) in the loga-
`l~t.hmlc phase of growth were washed with 10 mi HBSS buffer (107
`mM NaC1, 20 mM HEPES, 26.2 mM NaHCOs, 5.3 mM KC1, 1.9 mM
`CaC12, 1.0 mM MgC12, and 7.0 mM D-glucose adjusted to pH 7.4 with
`NaOH), centrifuged, and suspended in 1 ml HBSS-buffer at 37°.
`Influx of [SH]M~X (0.5 Ci/mmol) was measured over a period of 1.5
`rain at 37° at an extracellular concentration of 5 p~M in the absence or
`presence of increasing concentrations of unlabeled folate or antifo-
`late compound. Uptake was terminated by the addition of 9 ml of
`ice-celd HBSS buffer, centrifugation at 800 x g for 5 rain at 4°, and
`a second wash with 10 ml ice-celd buffer. Cell pellets were resus-
`pended in water and analyzed for radioactivity with Ultima-Gold
`scintillation fluid and a scintillation counter (both from Packard,
`Brussels, Belgium) with a counting efficiency for [3H] of approxi-
`mately 55%. The drug concentration required to inhibited [SH]MTX
`influx by 50% of control values (221 pmol/min/107 cells) was taken to
`be a measure of relative affinity for the RFC.
`mFBP substrate specificity. An intact cell binding assay for
`competitive binding of [SH]FA was performed as described previously
`(86). Briefly, L1210-FBP cells were washed twice with ice-cold HBSS
`buffer. One milliliter of L1210-FBP cell suspension (3 × 10s cells)
`was incubated with 100 pmol [OH]FA (specific activity, 0.5 Ci/mmol)
`in the presence or absence of increasing concentrations of unlabeled
`folate or antifolate compound. After 10 rain, the cells were collected
`by centrifugation (for 5 rain at 800 x g at 4°), after which the
`supernatant was removed. Pellets were resuspended in water and
`analyzed for radioactivity as described. Relative aiYmities are de-
`fined as the inverse molar ratio of cempeund required to displace
`50% of [aH]FA from mFBP on L1210-FBP cells. The relative affinity
`of mFBP for FA is set at 1.
`Growth-lnhlbition studies. All cell lines were plated at an
`initial density of 7.5 x 104 cells/m] in individual wells of a 24-well
`tissue culture plato. The growth medium for CCRF-CEM, CEM/
`MTX, and L1210-cells was RPM11640 with 10% FCS, supplemented
`as described. CEM-7A cells were grown in folate-fi, ee RPMI 1640
`with 10~ dialyzed FCS and 1.0 nM LV as the sole folate source;
`L1210-FBP and L1210-B73 cells were grown in the same medium
`but with 1 nM LV, 20 nM LV, or 20 nM FA as folate source. Drugs were
`added at the time of plating. After 72 hr of exposure, cell counts and
`viability were determined with a hemocytometer by trypan blue
`exclusion. IC~o values are given as the concentration of drug at which
`the growth is inhibited by 50% compared with controls.
`
`Results
`RFC substrate specificity. The affinity of the RFC for
`the series of antifolate drugs, expressed as the drug concen-
`tration required to inhibit [aH]MTX influx by 50%, is shown
`in Fig. 2. MTX (represented by a black bar and vertical line)
`was used as a reference compound. Compounds on the left
`side of this line were better substrates for the RFC than
`MTX, whereas compounds on the right side were poorer
`substrates. All DHFR inkibitors included in the study ap-
`
`Structure-Activity Relationships of Antlfolate Transport 463
`
`peared to be better substrates by RFC than ~ Replace-
`ment of the 2-Rmino group by 2-methyl in MTX and AMT led
`to an enhanced affinity. For example, 2-CHs-dAMT was a
`10-fold better substrate for RFC than MTY~
`In the series of GARTF inhibitors, only DDATHF is a more
`efficient substrate for RFC than MTX. The group of FPGS
`inhibitors [5-deazatetrahydrofolic acid analogues with the
`glutamate side chain replaced by other charged residues
`(homocysteic acid, 2-~mino-4-phosphonobutanoic acid, and
`ornithine)] are very poor substrates for RFC-mediated trans-
`
`Within the group ofTS inhibitors, the 2-amino-based struc-
`tures CB3717, 2-NH~-ZD1694, and IAHQ are low affinity
`substrates for the RFC. However, their 2-methyl analogues
`(ICI-198,583, ZD1694, and 2-CH3-IAHQ) are excellent sub-
`strates for the RFC. Modification at the 4-oxo position of
`ICI-198,583 did not alter the substrate affinity for the RFC.
`A modification at the N-3 position led to a 2-fold better
`affinity for the RFC. Replacement of the glutamate side chain
`by valine and 2-~mlnosuberate or a 7-CHa substitution re-
`suited in a 4-5-fold and a 3-fold lower affinity for the RFC
`than for ICI-198,583, respectively. The RFC had the highest
`affinity for BW1843U89. LY231514 was a 2-fold better sub-
`strate for the RFC than MT~ The ai~inity of the RFC for
`5-d(i)PteGlu and its N~-substituted analogues was more than
`10-fold lower compared with MTT~ As might be expected, the
`lipophilic antifolates AG337 and AG377 were poor substratos
`for the RFC.
`mFBP substrate specificity. Relative affinities of the
`mFBP for the series of antifolate drugs are presented in Fig.
`3. FA (represented by the black bar and vertical line), for
`which mFBP has a binding ~fl~nlty constant (Kd) of 0.5-1 aM
`(not shown), was used as the reference compound. Compared
`with FA, the mFBP had a lower affinity for all of the DHFR,
`GARTF, and FPGS inhibitors, especially the 4-amlno-based
`compounds, e.g., MTX, AMT, and their 2-des~mlno/2-methyl
`analogues 10-EdAM and PT523. Modifications at sites other
`than the pteridine ring, such as the Ce-N1° bridge and the
`glutamate side chain, were better tolerated with respect to
`efficient binding by mFBP. For eT~mple, mFBP demon-
`strated good binding affinity for DDATHF and 5-dPteHCysA.
`In contrast to 5-dPteHCysA, the tetrahydrofolate forms of
`5-dPteAPBA and 5-dPteOrn were better substrates for
`mFBP than were their nonreduced counterparts.
`From the group of TS inhibitors, four compounds were
`identified for which mFBP has a higher affinity than for FA:
`LY213514, CB3717, IAHQ, and 2-NH2-ZD1694. Unlike ob-
`servations for the RFC, the 2-methyl analogues of the latter
`three compounds are characterized by significantly de-
`creased binding affinity to mFBP. Most dramatic in terms of
`abrogation of binding affinity are modifications at the N-3
`position (3-deaza-ICI-198,583) and structural alterations at
`the 4-oxo position (4-deoxy-ICI-198,583), although the 4-
`methexy analogue of ICI-198,583 retained the same binding
`affinity as the parent compound. Modifications at other sites
`of the molecule (the qulnazoline ring [substitution at the
`7-position], the C~-N~° bridge, the p-aminobenzoyl, and the
`glutamate side chain) did not dramatically influence the
`binding affinity with respect to binding by mFBP. mFBP
`exhibits a moderately high binding ~Winity for BW1843U89.
`Similar to the RFC, the mFBP has a poor affinity for the
`lipophilic drugs AG337 and AG377.
`
`Sandoz Inc.
`Exhibit 1022-0005
`
`

`
`RFC
`
`2-d~
`2-CH~-MTX
`
`2-dAMT
`=-CH,-AM’r
`tO-F=dAM
`
`S-d~P~C~A
`S-dH~UmA
`5-dH4Pt~)m
`
`CB3717
`K:1-198.583
`3-d~z~Cl-198,583
`4-H4C1-198,583
`
`Glu-~Vd-ICI-1S8,583
`Glu~Sub-1~-198,583
`7~H=-IC1-1~8,583
`
`2-NH=-ZD1694
`
`LY231514
`IAHQ
`2-dlAHQ
`
`5-d(i)Pt~Glu
`
`AG337
`AG377
`
`-)
`
`0)
`
`Fig. 2. Concentrations for 50%
`inhibition of RFC-mediated influx
`of [~H]MTX. Values represent the
`concentrations of folate or anti-
`folate compound necessary to
`inhibit RFC-mediated [~H]MTX
`influx in CEM-7A cells by 50% at
`5 pJ~ extracelluler concentration.
`Verffcal line and black bar, con-
`
`centration of unlabeled M’rx
`necessary to inhibit [~H]MTX up-
`take by 50%. For further details,
`see Experimental Procedures.
`Ba/s labeled with a through g,
`actual value exceeds the indi-
`cated value. Residual uptake val-
`ues as percentage of control at
`the indicated drug concentration
`were as follows: a, 57% at 100
`p.M; b, >90% at 100 /~M; C,
`>90% at 100 p~; d, 57% at 100
`p.M; e, >90% at 200 p~; f, >90%
`at 50 ~; g, >90% at 50
`
`I I
`
`;)
`
`0.1
`
`1
`
`10
`
`pM
`
`100 600
`
`Growth-;nk|bition studies. The role of the P,~’C in drug
`transport and growth inhibition was analyzed by using three
`cell lines that differ in RFC expression and MTX transport
`(22) (Tables 3 and 4): CEM-7A cells are characterized by a
`30-fold overexpression of RFC compared with parental
`CCRF-CEM cells, whereas CEM/MTX cells exhibit a MTX
`transport defect. In general, compounds that are good sub-
`strates for the RFC (e.g., MTX and AMT with their 2-deso
`nmino and 2-methyl Rnnlogues ICI-198,583, ZD1694,
`LY231514, BW1843U89, and DDATHF) were potent growth
`inhibitors provided that the compounds were efficiently poly-
`glutamated, are potent inhibitors of their target enzymes, or
`both (Tables 1 and 2). For e~Rmple, drugs like 2-dMTX,
`2-CHa-dAMT, 3-deaza-ICI-198,583, and 4-H-ICI-198,583 ap-
`peared to be efficient substrates for the RFC, but because
`
`they are poor inhibitors of both DHFR and TS, their growth-
`inhibitory activity was low. Nevertheless,’the role of the RFC
`in transport of these compounds and of most of the TS inhib-
`itors is illustrated by the fact that RFC-overproducing
`CEM-7A cells are more sensitive than parental coils, whereas
`MTX transport-defective CEM/MTX cells are cross-resistant
`to these drugs. In this respect, it is of interest to note that
`cross-resistant factors for DDATHF and PT523 were signifi-
`cantly lower than for MTX, AMT, and 10-EdAM. Poor growth
`inhibition associated with inefficient RFC transport was ob-
`served for CB3717, 2-NH2-ZD1694, IAHQ, and the series of
`glutamate side chain-modified FPGS inhibitors. In addition,
`for 5-deazaisofolic acid, 5-deazatetrahydroisofolic acid and
`their N~-substituted analogues, the poor growth-inhibitory
`effect correlates with a low transport efficiency and target
`
`Sandoz Inc.
`Exhibit 1022-0006
`
`

`
`mFBP
`
`Sb’uctwe-Actlvity Relationships of Antlfo~te Transport
`
`0.008
`0.005
`0.005
`0.008
`0.002
`0.011
`0.009
`0.009
`
`2-d~
`2-CH,-MTX
`
`2-dAMT
`2-C~,-AMT
`10-EdAM
`m23
`
`DDATHF
`
`N~-CH=-5-d(i)H4Pte~u
`
`S-dPt~W:~A
`
`CB3717
`IC1-198,583
`3-deaz~lCl-198,583
`4-H4C1-198,583
`
`Glu--Vd-ICl-198,583
`Glu-,Sd~-198.583
`7-CH=4C!o198o583
`
`2-NH=-ZD1
`
`10.019
`
`LY231614
`LAHO
`2-dlAHO
`2-CH,4AHO
`5-dli)PIsGiu
`
`I
`
`AG337
`AG377
`
`|0.028
`|o.o$$
`
`I
`
`Fig. 3. Relative affinity of
`mFBP for folate and antifolate
`compounds. Values indicate
`
`inverse molar ratio of com-
`pound required to displace
`50% of [~H]FA from mFBP in
`L1210-FBP ceils. Relative af-
`finity of FA Is set at I (b/ack bar
`and vertical line).
`
`/
`
`0.00 0.25 0.50 0.75 1.00 1.25 1.50
`
`Relative affinity
`
`enzyme inhibition. The lack of correlation between RFC ex-
`pression and growth-inhibitory effect of the lipophilic drugs
`AG337 and AG377 is indicative of an RFC-independent pro-
`cess of drug uptake.
`The drug sensitivity profiles of human CCRF-CEM coils
`(Tables 3 and 4) and routine L1210 colls (Tables 5 and 6)
`were in large part similar. Nevertheless, absolute IC~o values
`for the individual drugs were most often lower for L1210 colls
`than for CCRF-CEM coils, which may be consistent with a
`3-fold higher level of RFC expression in L1210 cells versus
`CCRF-CEM cells, whereas other transport kinetics proper-
`ties are similar (20). The one excoption is the uptake of
`BW1843U89 (80), which is a 17-fold poorer substrate for the
`routine RFC (K~, 14 p~) compared with human RFC (Ki, 0.8
`
`/~; see Fig. 2). This difference in transport characteristics is
`reflected in a 10-fold lower growth-lnhibitory activity of
`BW1843U89 ag~inRt L1210 versus CCRF-CEM colls (80).
`In addition to binding of antifolates, the results given in
`Tables 7 and 8 support a functional role for mFBP in the
`uptake process of antifolates. In the absence of competing
`folate cofactore in the growth medium, most of the com-
`pounds tested demonstrated a growth-inhibitory effect
`against (RFC-/mFBP+ + +) L1210-FBP cells, provided they
`were efficiently polyglutamated or were potent inhlbitere of
`their target enzyme (Tables 1 and 2). Compounds for which
`the RFC has poor Aff;nlty (e.g., CB3717, IAHQ, and 2-NH=-
`ZD1694) appeared to be good growth inhibitore, consistent
`with the high bindin£ affixdty of mFBP for these compounds.
`
`Sandoz Inc.
`Exhibit 1022-0007
`
`

`
`!
`
`466 Westedmf et al.
`
`TABLE 3
`Growth-inhibitory effects (IC.o values in nM’) Of antifolate
`=)mpounds against human CF.M cells with upregulated and
`downregulated RFC-medlated transport (RFC expression in
`parentheses)
`
`TABLE 5
`Growth-inhibitow effects (IC~ values in ~ of anUfolate
`compounds against routine L1210- (RFC+)b and L1210-B73
`(RFC+/mFBP+++)~ cells grown in different folate
`concentrations in the medium
`
`Cell line (RFC expression)a
`
`L1210
`
`L1210-B73
`
`Compound
`
`MTX
`2-dMTX
`2-CH3-MTX
`AMT
`2-dAMT
`2-CH3-AMT
`10-EdAM
`PT523
`DDATHF
`5-d(i)H4PteGlu
`NS-CH~-5-d(i)H4PteGlu
`5-dPteHCysA
`5-dPteAPBA
`5-dPteOm
`5-dH4Ptel.-ICysA
`5-dH4PteAPBA
`5-dH4PteOm
`
`CCRF.-CEM
`(+)
`
`8.1
`1,870
`837
`2.0
`500
`1,157
`1.3
`1.1
`11
`37,020
`42,670
`>25,000
`>25,000
`>50,000
`>25,000
`6,100
`>50,000
`
`CEM/MTX
`(-)
`1,950
`40,000
`41,300
`1,010
`12,600
`10,000
`533
`33
`80
`36,025
`>100,000
`>50,000
`>50,000
`>50,000
`>50,000
`>50,000
`>50,000
`
`CEM-7A
`(++++)
`1.2
`49
`31
`0.70
`14.6
`42
`0.3
`0.7
`2.2
`164
`1,740
`4,500
`>25,000
`31,500
`13,000
`1,400
`3,200
`
`¯ IC~o values are given as the concentration of drug at which growth is inhibited
`by 50% compared wflh controls. VaJuas ~e the mas~ of at lasst four separate
`experiments (SD < 15%).
`a RFC expression levels: CCRF-CEM, CEM-MTX, and CEM-TA-ceiis, 0.23,
`<0.02, and 8.1 pmoVmg protein, respectiv~y (22).
`
`TABLE 4
`Growth-lnhibitow effect~ (IC~o values in nM=) of antifolate
`compounds against human CEM cells with upregulated and
`downregulated RFC-medtated transport (Rr-C expression in
`pa~)
`
`Cell line (RFC expression)a
`
`Compound
`
`MTX
`CB3717
`IC1-198,583
`3-deaza-ICl-198,583
`4-H-IC1-198,583
`4-OCHa-IC1-198,583
`Glu-*Val-ICl-198,583
`Glu--*Sub-lCI-198,583
`7-CH~-IC1-198,583
`ZD1694
`2-NH=-ZD1694
`BW1843U89
`LY231514
`IAHQ
`2-dlAHQ
`2-CH~-IAHQ
`5-d(i)PteGlu
`Ng-CHa-5-d(i)PteGlu
`Ng-CHO-5-d(i)PteGlu
`AG337
`AG377
`
`CCRF-CEM
`(+)
`
`8.1
`705
`17
`>100,000
`680
`46
`4,400
`910
`243
`3.5
`115
`2.4
`23
`1,090
`690
`41
`17,465