mature blood-brain barrier. This phenomenon has been ob 3. Mckenzie AF, Budd RS, Yang C, et al. Technetium-99m-methylene diphosphonate uptake in the fetal skeletal at 30 weeks gestation. J Nuc! Med 1994;35:I338—1341. served with other substances (11,12). 4. Martin B, Payan JM, Jones JS, Buse MG. Placental localization in abdominal pregnancy using technetium-99m-labeled red blood cells. J Nuc! Med l990;3 I :I 106— CONCLUSION 1109. Technetium-99m d, 1-HMPAO binds to both human and 5. Herbert RJT, Hibbard BM, Sheppard MA. Metabolic behavior and radiation dosimetry of Tc-99m albumin in pregnancy. J Nuc! Med I969; 10:224—232. guinea pig placentas. It is transfered across the placental barrier 6. Newman RA, Gallagher JO, Clements JP, Krakoff IH. Demonstration of Ga-67 into fetal circulation mostly in the liver with increasing gesta localization in human placenta. J Nuc! Med 1978;I9:504—506. tional age. 7. Mahon DF, Subramanian G, McAfee JG. Experimental comparison of radioactive agents for studies of the placenta. J Nuci Med l973;I4:65I—659. 8. Smith CH, Adcock EW III, Teasdale F, Ct al. Placental amino acid uptake: tissue ACKNOWLEDGMENTS preparation, kinetics and preincubation effect. Am J Phvsio! 1973;223:558—664. This research was supported by Kuwait University Council 9. Guerre-Millo M, Challier JC, Rey E, Ct al. Matemofetal transfer of two benzodiaz epines effect of plasma protein binding and placental uptake. Des' Pharmaco! Ther grants MN042 and MLNO18. I982;4:I58—I72. 10. Nakamura K, Tukatani Y, Kubo A, et al. The behaviour of@―TcHMPAO in blood REFERENCES and brain. EuriNuc!Med I989;lS:100—I07. 1. Palestro Ci, Malat J, Richman AH. Incidental diagnosis of pregnancy on bone and I I. Pardridge WM, Mietus L. Kinetics of neutral amino acid transport through the gallium scintigraphy. J Nuc! Med I986;27:370—372. blood-brain barrier of the newborn rabbit. J Neurochem l982;38:955—962. 2. Maguire C, Florence 5, Powe JE, et al. Hepatic uptake of technetium-99m HM-PAO 12. Moos T. Moligard K. Cerebrovascular permeability to azo dyes and plasma proteins in in a fetus. J Nuc! Med I990;3I:237—239. rodents of different ages. Neuropatho! App! Neurobiol I993;19:I20—I27.

Biological Properties of Biotin—ChelateConjugates for Pretargeted Diagnosis and Therapy with the Avidin/B iotin System

David A. Goodwin, Claude F. Meares and Maureen Osen Department ofNuclear Medicine, Veterans Affairs Palo Alto Health Care System, Palo Alto; Department of Radiology, Stanford University, Stanford; and Department of Chemistry, University of Cahfornia, Davis, Cal@fornia

much higher than that in streptavidin controls. No significant target Three-step pretargeting increases target-to-background ratios in uptake was seen with anti@LAkMAb-streptavidin pretargeted for 3 or radioimmunodetection and can potentially decrease harmful radia 20 hr. Kidney uptake -@-20%/g,which was lower than that of tion to normal tissues in radioimmunotherapy.We studied four streptavidin alone. Conclusion: Three biotinyl chelates bind the biotin—chelateconjugates (BCCs) for use in the avidin/biotin pretar diagnosticand therapeuticradiometals111lnand seY (@, t@y geting system. Methods Pharmacokineticsand biodistribution analogy, 90Y)with the required in vivo stability and physiological were studied in normal BALB/c QA'@-negative),normal C3H QA'@ properties for pretargeted diagnosis and therapy. Kidney uptake of positive) and LS174T tumor-bearing BALB/c severe combined im streptavidin was decreased by conjugation to MAb. Failure of anti munodeficient mice. Streptavidin alone and antibody—streptavidin conjugates[monoclonalanthody(MAb)10-3.6anb-1A@'@lgG2a]were @kMAb-streptavidin conjugate to bind BCC after pretargeting may be due to rapid internalization of MAb@@streptavidin@IAkcomplex by used. Indium-i 11- or @‘Y-BCCswere given alone intravenously; they were mixed with streptavidin or MAb-streptavidin conjugate the lymphocyte or to endogenous biotin. Either or both of these and given intravenously; or streptavidin and MAb-streptavidin con would make streptavidin unavailable to subsequent BCCs. jugate were pretargeted, and 2-3, 5 and 21 hr later, BCCs were Key Words radioimmunolmaging; radsosmmunotherapy, yttrium injected intravenously. Samples were taken 2-3 hr after intravenous 90; indium-i 11; pretargeting injection of labeled BCCs. Results Three of the four BCCs were J Nuci Med 1998 391813-1818 rapidly excreted by the kidneys, with <2.5%/g in any organ or tumor at 2—3hr. Gut excretion eliminated biotinyl-(S)-1-p-aminobenzyl ethylenediaminetetraacetic acid (EDTA) for use in pretargeting. Ra diolabeled monoclonal antibodies (MAbs) have shown Ninety percent of BCCs were bound to circulating pretargeted promising clinical results in the diagnosis and therapy of cancer streptavidin at 1-6 hr, and —15% were bound to pretargeted (1—5).Radionuclides such as ‘I‘Inand 90Y have been of streptavidin at 24 hr. Kidney uptakes were: preformed streptavklin particular interest in radioimmunoscintigraphy and radioimmu BCC given intravenously,—80%/g(24 hr);streptavidin pretargeted for 2—3hr, —60%/g; and streptavidin pretargeted for 5—21hr, notherapy (RAIT) due to their nuclear properties (6—9).Among —10%—20%/g.Kidney uptake was dose-dependent: 0.2, 0.67 and the radionuclides for therapy, @°Yis ofparticular interest due to 1.0 nmol of streptavrdin pretargeted for 21 hr showed increasing its superior properties, including pure beta emission and the concentrations (24 hO. Uptake of monoclonal anti@IAk@streptavidin@high-dose yield per nanomole (10). Bifunctional chelating BCC complex into spleen (70% ±10%/g; p < 0.05) and lymph agents (1 1,12) complex these metal ions and attach the chelated @ nodes (10% ±3.5%/v; < 0.01) was higher in lAk@posftiveC3H radionuclide to a protein or MAb (13). These conjugates act as mice than it was in IA -negative control BALB/c mice, and it was carriers of radiometals for tumor targeting and radiotherapy. Chelates that hold radiometals with high stability under physi ReceivedAug.29,1997;revisionacceptedDec.24,1997. ological conditions are essential to avoid excessive radiation For correspondence or reprints contact: David A. Goodwin. MD, Department of Nuclear Medicine, veterans Affairs Palo Alto Health Care System, 3801 Miranda Ave., damage to nontarget cells (14). Palo Alto, CA 94304. Renn and Meares (15) reported the large-scale synthesis of

BIOTIN—CHELATES FOR PRETARGETING ‘Goodwin et al. 1813 the macrocyclic bifunctional chelating agent, 2-p-nitrobenzyl 1,4,7,l 0-tetraazacyclododecane-N,N',N' p‘-tetraacetic acid (nitrobenzyl-DOTA), which holds yttrium with extraordinary stability under physiological conditions (16). Nitrobenzyl DOTA binds both yttrium and indium with superior stability, so it should be possible to use DOTA chelates of the gamma emitting 1IIIn as a tracer for DOTA chelates of the beta emitting 90Y to gain accurate measurements ofradiation dosim 0 H etry. HN@Y@ HN NH @ However, the use of radiolabeled antibodies for radiotherapy g H000cH, [email protected] [email protected]

and radioimmunotargeting is not without disadvantages. The @occ@p-@ooH ,CH,COOH accumulation of an antibody in target tissue requires several

days, so the distribution kinetics of directly labeled antibodies —‘-I bH,CO0H HOOCCH,\—/bH,cOOH are not optimal for accurate imaging or delivery of toxic substances to the tumor tissue for radiotherapy. A relatively low FiGURE1. SChematicstructuresofthesynthesizedbiotin-chelateconju tumor-to-background ratio and high nontarget activity are the gates. major problems in immunoscintigraphy (17), due to to the slow pharmacokinetics of radiolabeled antibodies (18). In addition, physiological properties for pretargeted diagnosis and therapy. because the high levels of circulating radioactive antibody Some abbreviations, the pretargeting terminology and the time conjugates limit the total amount of radioactivity that can be sequence of the pretargeting steps are given in Table I. given safely, low absolute tumor uptake and low radiation dose to the tumor are a problem in RAIT. Several attempts have been MATERIALS AND METHODS made to improve biodistribution of the radiolabel, such as the Synthesis of the BIOtin-Chelate and Monoclonal Antibody use of a second antibody (19), local delivery (20), the use of Streptavidin Conjugates metabolizable linkers between antibody and the chelate (21 ) or As BCCs are prepared for medical use, it is important to confirm the use of MAb fragments that display a faster blood clearance that they bind to avidin. The BCCs were labeled with excess than do whole antibodies (22). 57Co/Co of known molar concentration, and then streptavidin was Further improvements are desirable. The diagnostic or ther added. The complex was purified from excess “Co/Coand apeutic agent should be delivered at a time when the tumor unbound BCC as described (35). In a separate experiment, biotin tissue is already labeled by the tumor-specific antibody. This LC-DOTA was bound to hen egg avidin, and the complex was can be accomplished ifthe label or the therapeutic agent, which mixed with excess 57Co/Co. should show a fast clearance, is captured by an antibody that is Streptavidin was obtained from Boehringer Mannheim or Cal already targeted to a tumor cell. These considerations led to the biochem and was used without further purification. The MAb concept of multistep tumor targeting, based on separate admin 10-3.6 (IgG2a), specific for lAk, was conjugated to streptavidin. istrations of antibody and label (18,23—26). Several molecular pretargeting systems have been proposed, the hapten/antibody TABLE I and the avidin/biotin being the two most thoroughly studied to Two- and Three-Step Pretargeting Terminology date (27). A mathematical model has also been developed to Term Definition determine an optimal drug delivery protocol for multistep pretargeting systems (28). Targetingmolecule Longcirculatingmacromoleculespecific The avidin/biotin system is widely used for in vitro and in fortarget (MAb—streptavidin vivo applications, in immunohistochemistry, enzyme-linked conjugate) immunosorbent assay and molecular biology (29,30). Avidins Effectormolecule Small,rapidlydiffusingand excreted are proteins that bind biotin with high affinity and specificity 111ln-or anY@BCC Pretargetingtime (e.g., Timefrominjectionof targeting (31 ). Theoretically, they can bind up to four molecules of streptavkiinpretargeted molecule(bsMAb, MAb-streptavidin biotin, and the dissociation constant of the avidin/biotin com for21 hr, streptavidin conjungate)untilchase or effector plex is on the order of l0 15M, so binding can be considered was givenintravenously (11ln-or @Y-BGCinthese an irreversible process (32). (pretargeted)21 hr experiments) One practical adaptation of the biotin/avidin pretargeting beforethe next step) system has now been in Phase I/Il clinical imaging trials for Chase* Polyvalentmacromoleculethat several years (33 ). Biotinylated antitumor MAbs are given crosslinksand removestargeting intravenously, followed by avidin chase and then a biotin moleculesfromthe circulationbut not chelate conjugate (BCC) in a three-step protocol. Improved the target Samplingor imagingtime Time from injection of effector until kill I I IIn human tumor images with low blood and liver background or imaging(2—3hr inthese and high contrast have been obtained (26). Encouraging results experiments) have now been reported by these authors using a similar Totaltime Pretargetingtime + chase time fdchase protocol in a Phase I @°Ytherapy trial (34). is used) In this article, we describe the pharmacokinetics and biodis Pretargeting+ effectortime [e.g., tribution of four biotinylated chelates: biotinyl-ABE, biotin-P pretargeting(21hO+ chase (1 hr) + CITC,biotin-P-DOTAand biotinyl-LC-DOTA(Fig. 1). De sample (3 hOafter intravenous111ln tailed synthesis and chemical characterization of these or @Y= 24 hr total@ conjugates have been reported elsewhere (35). Among these conjugates, biotinyl-LC-DOTA, biotin-P-DOTA and biotin-P *bsMAb= blspec@cMAb. CITC bind the diagnostic and therapeutic radiometals 1I1In and Note:Not used inthese experiments.Givenjust before (0.5—1hr)radio 88y with the required in vivo stability and have excellent activeeflectormolecule.

1814 THEJOURNALOFNUCLEARMEDICINE•Vol. 39 •No. 10 •October 1998 lAk is a major histocompatibility complex alloantigen (Class II) 3.5 present on C3H mouse B lymphocytes but not on BALB/c B lymphocytes. We have shown in other experiments with directly 3.0 labeled ‘‘‘In-lO-3.6that the spleen and lymph nodes in C3H mice are the target organs (1 7). BALB/c mice served as the control. The synthetic pathways and detailed chemical characterization 2.5 of the four BCCs studied here have been reported previously (35). @ The MAb—streptavidinconjugates were synthesized by first form 2.0 ing the 2-iminothiolane (Traut's reagent) derivative of anti@IAk @ (IgGl) and of streptavidin. Then the streptavidin conjugate was 1.5 activated with 5,5 ‘-dithio-bis(2-nitrobenzoic acid) and added to 2@IT@anti@IAkin a 3:4 molar ratio. The resulting disulfide S—S 1.0 linkage coupled the two molecules together in a molar ratio of approximately 1:1. 0.5 Radiolabeling Biotin-Chelate Conjugates with Yttrium-88 0.0 or Indium-I II Blood Heart Luap Liver Spies. Kidatys Mssek Bout Ski. Gut Indium- 111 was chosen as an imaging radionuclide, and 88Ywas chosen as an analog of the therapeutic radionuclide @°Y.Yttri U Biotln.P.CITC n-3 0 Bloda.P.DOTA .-3 I Blodn.LC.DOTA n-3 @ um-88 or ‘‘In was dried in 1 ml of metal-free polypropylene FIGURE2. Distributionof three111ln-biotin-cheiateconjugates2 hrafter snap-top tubes in an Eppendorf heat block at set at 95°C.Fifty intravenousinjectionintonormalBALB/cfemalemice. Errorbars = ±1 s.d. microliters of 0.1 M NH4Ac (jH 5) were added and vortexed, and Kidneysare the major route of excretion, with a small amount excreted after 5 mm, 1 p.1 was spotted onto a thin-layer chromatography throughthe gut. (TLC) strip (first lane). Ten microliters of 1 mM BCC (or a volume = 10 nmol) were added, and the solution was mixed and plex given intravenously and the organ concentrations were mea heated at 95°for 30 mm. The solution was cooled, and 1 p.1was sured at 3 hr. In some experiments, the mice were also pretargeted spotted on 10-cm TLC strip in the second lane. Then, 10 nmol of with 0.2—1.0nmol ofstreptavidin intravenously; II11n-labeled BCC nonradioactive metal were added to fill up the chelator. The was injected intravenously 1, 2, 3, 4, 6 and 2 1 hr later; and organs solution was heated again at 95°for 30 mm, and 1 @lwas spotted were sampled at 3 hr. In other experiments, the mice were on the TLC strip in lane 3. The TLC strip was dried and run in pretargeted with MAb—streptavidin conjugate, and labeled BCC MeOH-NH4Ac [10% NH4Ac:absolute MeOH (vol/vol)]. was injected at 3 and 20 hr afterward. Organ and tumor samples An autoradiograph was made on DuPont film, the TLC strip was were usually taken 2—3hr after injection of radiolabeled BCCs. cut in 1-cm sections and counted and the percentage of labeled BCC was calculated. All activity in the first spot should be at the RESULTS origin (free metal), and that in the second spot should run 100% to The in vitro assay of streptavidin binding gave a lower than Rf —2—4(metal-DOTA complex). For injection, 1 @lof 0.1 M theoretical 4: 1 ratio, based on the known streptavidin valence of Ca/Na2-EDTA was added during dilution, to bind any free metal 4. The measured BCC-to-streptavidin ratio was approximately and promote rapid renal excretion. 1.87:1 ±0.3 (35). In the hen avidin experiment, only 0.41 ± Biological Half-Life 0.08 cobalt ions (instead of 4) became bound to the complex. Twenty-five to 100 @tCiII‘In-BCCwere injected intravenously The organ distributions of 11‘In-labeledbiotin-P-CITC, bi for half-life measurements. For whole-body counting, the mice otin-P-DOTA and biotin-LC-DOTA 2 hr after intravenous were placed in a small stoppered plastic beaker (with air holes) injection into normal BALB/c female mice are shown in Figure equidistant between the two 180°opposed arms of a Picker Dual 2 (error bars = ±1 s.d.). The kidneys were the major route of Probe scintillation counter. The empty mouse holder was counted excretion, with a small amount excreted through the gut. Very for background, and the first mouse count was obtained immedi low concentrations (< l%/g) were seen in all organs except ately after intravenous injection, before any renal excretion could kidney and gut, even at the early 2-hr time point. occur. This count was used as the 100% standard zero time value The organ distributions of 11‘In-labeledbiotin-P-CITC, bi for subsequent Tl,2b calculations. Two 1-mm mouse counts from otin-P-DOTA and biotin-ABE 2 hr after intravenous injection each detector were performed and averaged. For later measure into LS147T tumor bearing BALB/c SCID mice at 2—3hr are ments, a minimum of 1000 counts were obtained by increasing the seen in Figure 3. With the exception of biotin-P-ABE, similar counting time. The background counts of the empty mouse holder patterns with very low concentrations (< l%/g) in all organs were subtracted, and the mean net counts per minute were except the kidney were demonstrated. In addition, the tumor calculated. The time elapsed from zero time was noted, the counts showed no accumulation above the background seen in the per minute were decay-corrected back to zero time and the other organs. Of interest was the very high level of biotin ABE percentage of the injected dose remaining in the mouse was in the gut due to the large amount excreted through the biliary calculated. system. Nothing in the structure predicted the unique metabolic route of this derivative. Biodistribution The whole-body retention of II‘Inlabeled biotin-P-CITC in In control experiments, the biodistributions of four I11In-labeled BALB/c female mice injected after pretargeting with streptavi BCCs injected intravenously alone were measured in normal din for 1, 2, 3, 4, 6 and 24 hr is shown in Figure 4. The curves (IA―-negative)BALB/c, LS174T tumor-bearing BALB/c severe were plotted from immediately after injection of II‘In-labeled combined immunodeficient (SCID) and normal (IAk@positive)C3H biotin-P-CITC up to 48 hr after injection. Almost quantitative mice. Streptavidin and anti@IAk(IgG2a) MAb—streptavidinconju binding of BCC to circulating streptavidin was seen up to 6 hr gates were also studied. Indium-i 11- or 88Y-labeled BCCs were of pretargeting, with much less binding after 24 hr of pretar mixed with streptavidin or streptavidin—MAbconjugate, the corn geting.

BIOTIN—CHELATES FOR PRETARGETING •Goodwin et al. 1815 30.0 100

25.0 90 so 20.0 (I) 70 z @ 15.0 60

10.0 50

40 5.0 30 0.0 Btood Ikurt Lungs Unr Sp1@ KIday Tumor Muick Bo@ Skia Gut 20

a Biotin.P.CITCn=3 UBiotin.P.DOTAu=3 •[email protected] n=3 10

FiGURE 3@Organ distribution of three 111ln-biotin-chelete conjugates at 2-3 0 hr in BALB/cSCID mice with LS174Ttumor in the left flank. Very low A B C D E F concentrationsare seen inallorgans and the tumorat the earlytime points, withthe exceptionofbiotin-ABE,a largefractionofwhichisexcretedthrough the liverand gut. Pretargeting Interval FIGURE 5. Kidney uptake in normal BALB/c female mice of streptavidin Kidney uptake of streptavidin in normal BALB/c mice is moed with 111lnblotin-P-CIrC and injected (Column ,@4or pretargeted for 1 (ColumnB),2 (ColumnC),3 (ColumnD),4 (ColumnE)and 21 (ColumnF)hr shown in Figure 5. A molar excess of streptavidin was mixed andsampled3 hrafterintravenousinjectionof 111ln-biotin-P-CITC. with I1‘In-biotin-P-CITC; the complex was injected and kid neys were assayed at 24 hr (Column A); or streptavidin was significantly greater than that expected by chance (p < 0.0001) pretargeted for 1 (Column B), 2 (Column C), 3 (Column D), 4 when compared by analysis of variance. (Column E) and 21 (Column F) hr, and the kidneys were The organ, spleen and peripheral and mediastinal lymph node assayed 3 hr after intravenous injection of an equimolar amount distributions of a bispecific conjugate capable of binding both 11‘In-biotin-P-CITC. A fairly high kidney uptake was seen at the target lymphocytes and BCC are shown in Figure 7. all times but especially at the earlier times, up to 3 hr. However, Indium-I 1l-biotin-P-CITC was mixed with MAb anti@IAk_ a significant amount (> lO%/g) was also seen at 4 and 21 hr streptavidin conjugate and injected intravenously, and the 24-hr pretargeting. organ distribution was obtained in normal C3H (IAk@positive) The kidney uptake of increasin@ amounts of streptavidin and control normal BALB/c (IAk@negative) mice. Figure 7 pretargeted for 21 hr, 3 hr after 1 ‘In-biotin-P-CITC (24 hr shows significantly higher uptake in the spleen and lymph total), is shown in Figure 6. The kidney uptake of streptavidin nodes of C3H mice than control BALB/c mice (@p < 0.05; is dose-dependent, with the variation among group means being @ ±p< 0.01 ; < 0.001). However, there was no significant

25

20

‘I,

z z @ z 15

@ 10 2

5

.-+-- Pretarget 1-6 Houra u-S —•—Pretarget24 Hours u-I 0 0.2 amok u-3 0.667 amok n—2 1 umok n—3 10 20 30 40 50 smoksSTREPTAVIDIN TIME-HOURS FIGURE 6. Three hour kidney uptake of 111ln-biotin-P-CflC in normal FIGURE4. Whole-bodyretentionof 111ln-labeledbiotin-P-Cfl@Cin normal BALB/cfemalemiceafterstreptavidinpretargetedfor2l hr.Kidneyuptake is BALB/cfemalemiceinjectedafterpretargetingwithstreptavklinfor1,2,3,4, dose dependent, withthe variationamong group means being significantly 6 and 24 hr.Curvesare plOttedfromimmediatelyafterinjectionto 48 hrafter greater than that expected by chance when compared by analysis of injection. variance

1816 THEJOURNALOFNUCLEARMEDICINE•Vol. 39 •No. 10 •October 1998 TABLE 2

UCJH u-3 Properties of Biotin-Chelate Conjugates DBALH/cp.3 Chemicalproperties Physiologicalpropertiss

@0.0 Lowmolecularweight(small<10 kDa) Renalexcretionexclusively Hydrophilic GFR(lOOcc/min)* Rapidlydiffusible Extracellularfluiddistribution Net negativecharge No proteinbindingin blood HighspecificactMty(1000 CVmmoI) No intracellularuptake . 30.0 Choiceof radionuclide;bifunctionalchelate Nonimmunogenic 20.0 + N

10.0 •Notmeasured, but WhOle-bOdycle@r@nc@similarto the known GFR agents 1111n-DTPAand 111ln-EDTA. 0.0 Blood Herl Luup urn Sçkt. Kldeys Muielu R@ S@1 Gui PNds MNde, GFR = glomerularfiltrationrate. ORGAN

FIGURE7.Twenty-four-hourorgandistributionof 111ln-biotin-P-CITCmixed In other experiments using a hapten/antibody pretargeting with MAbanti-LA'@-strepta@iidinconjugate in normal C3H(@-positn/e)and system, we have achieved fast in, slow out tumor kinetics in a BALB/c(lA'@-negative)mice.ThereissignificantlyhigheruptakeinC3Hspleen and lymphnodes than incontrolBALB/cmice.*p< 0.05; @p< 0.01; 0p < mouse tumor model that are ideal for therapy (38). The tumor 0.001. biological half-life of a bivalent 88Y JANUS-DOTA hapten, measured over 5 days, was —24 hr. The therapeutic ratio, uptake in the target spleen (0.43% ±0. 15%/g) or lymph nodes obtained from the integrated tumor and blood concentrations (0.35% ±0. 13%/g) when the MAb—streptavidin conjugate was over 5 days, was —‘20:1, compared to 2—3:1 with directly pretargeted for 3 and 2 1 hr. labeled MAb. The total injected dose remaining in the mouse at 24 hr was 5.5%, of which 23% was in the tumor. These result DISCUSSION suggested that it may be possible to deliver tumoricidal radia Pretargeting combines the pharmacokinetics of long-circulat tion doses with @°Yusing pretargeting techniques, without ing targeting molecules needed for high blood and tumor severenormal marrow irradiation. concentrations and rapidly diffusing and excreted effector The avidin/biotin system has several advantages over the molecules necessary for low concentration in normal tissue and hapten/antibody system, including the high affinity (K@ low toxicity. This is accomplished by temporal separation of the 10@15) that exists between avidin and biotin. Avidin and targeting and effector steps (27). The most popular pretargeting streptavidin are commercially available, well-defined structures techniques use either the biotinlavidin system discussed here or that can also be made by genetic engineering techniques. The the hapten/antibody system. However, DNA/DNA (36) and system is flexible allowing the use of several protocols; for prodrug/enzyme systems (37) have also been used, and many example, the use of biotinylated MAb followed by streptavidin, other ligand/receptor systems are theoretically possible. All the chase and 90Y-BCC (39). Despite its immunogenicity, strepta methods depend on a long-circulating conjugate to get high vidin is preferred over avidin for the MAb conjugate due to its target uptake with a diffusible, rapidly excreted effector mole slower blood clearance (40). cule. The chemical and physiological properties ofthree ofthe four In the past, several bifunctional chelating agents were rou BCCs tested are listed in Table 2. The most important are: rapid tinely synthesized and used for conjugation to antibodies in the diffusion into the extracellular fluid, no intracellular uptake in laboratory ofClaude F. Meares have shown promising results in any organ and rapid excretion through the kidneys. High terms of physiological stability. Among these chelates are both specific activity (>1000 Ci/mmol) ofeffector molecules is also EDTA- and diethylenetriaminepentaacetic acid (DTPA)-based necessary for receptor (antigen) targeting at nanomolar concen ligands. Biotinylated chelating agents have now been prepared trations. The speed and ease of radiolabeling highly purified from these precursors, thus exploring the complete set of low molecular weight BCCs at high concentration and/or high bifunctional chelating agents, including the newer DOTA de temperature makes high specific activity readily obtainable with rivatives. In the biotinylated chelates studied here, the chelating no-carrier-added radionuclides such as 99mTc, II‘In,@°Yand site is separated by a spacer from the biotin group and, thus, 68Ga (35). The half-life of the biotin-LC-DOTA in the blood from the avidin-binding site (Fig. 1). was sufficient for targeting, suggesting that it was not cleaved It is interesting that only about two molecules of the chelate by biotinidase in vivo as rapidly as the deferoxamine derivative were able to bind to the tetravalent protein. In addition, the hen of LC-biotin described by Rosebrough (41 ) was cleaved in egg avidin result suggested that the avidin-bound c-amino vitro. caproic linker does not make chelating groups accessible for Failure in these initial mouse experiments of anti@IAkMAb postcomplex radiolabeling. The fact that the binding sites for streptavidin conjugate to bind BCC after pretargeting may have biotin occur in pairs that are fairly close to each other can cause been due to rapid internalization of MAb—streptavidin@IA'@ steric hindrance to the binding of more than two biotinylated complex by the lymphocyte or to endogenous biotin. Either or molecules (26). Thus, avidins are expected to have fewer than both of these factors would make streptavidin unavailable to four biotinylated molecules bound. Too short a distance be subsequently administered BCC. The former explanation is tween avidin and the surface of the biotinylated chelate might more likely because pretargeted streptavidin was still able to also result in unsatisfactory binding of the biotinylated chelate bind BCCs almost quantitatively up to 6 hr (Fig. 4) and, to a to avidin. For sufficient binding capability to avidin, the biotin lesser extent, up to 24 hr after injection. Also, pretargeted site and the chelating site of the biotinylated chelate conjugate streptavidin binding sites were still available in the kidney up to must be separated by a spacer. The observed values we obtained 21 hr after injection, indicating that they had not been blocked are in agreement with this hypothesis and suggested that a by endogenous biotin. Furthermore, using pretargeted carcino varying degree of steric hindrance occurs between the sites (35). embryonic antigen markers known not to internalize, excellent

BI0TIN—CHELATE5 FOR PRETARGETING •Goodwin et al. 1817 localization of ‘‘‘In-BCChas been obtained with the avidin/ I I. Sundberg MW, Meares CF. Goodwin DA, Diamante CT. Chelating agents for the binding of metal ions to macromolecules. Nature l974;250:587—588. biotin system in humans (42). 12. Meares CF. Wensel 1G. Metal chelates as probes ofbiological systems. Acc Chem Res Recent preliminary clinical results of pretargeted @°Yradio 1984; 17:202—209. immunotherapy using the avidin/biotin system have been very 13. Brinkley M. A brief survey of methods for preparing protein conjugates with dyes, haptens and cross-linking agents. Chemistry 1992:3:2—13. promising. Paganelli et al. (34) have reported a Phase I—Il 14. Klein JL, Nguyen TH, Laroque P. et al. Yttrium-90 and iodine-13 1 radioimmuno clinical trial in 11 patients using 9°Y-labeled biotinyl-LC globulin therapy of an experimental human hepatoma. Cancer Res 1989:49:6383— DOTA in a three-step protocol. Each patient received 30—50 6389. 15. Renn 0, Meares CF. Large-scale synthesis of the bifunctional chelating agent mg of biotinylated MAb F023C5 (anti-carcinoembryonic anti 2-(p-nitrobenzyl)-I,4,7,l0-tetraazacyclododecane-N,N',N',N' ,‘-tetraacetic acid, and gen), followed 36 hr later by an avidin chase and streptavidin, the determination of its enantiomeric purity by chiral chromatography. Bioconj Chem and 18—24hr later, a second chase ofbiotinylated human serum 1992;3:563—569. 16. Moi MK, Meares CF. DeNardo Si. The peptide way to macrocyclic bifunctional albumin, followed immediately (10 mm) by 5 mg of @°Y chelating agents: synthesis of 2-(p-nitrobenzyl)-I,4,7,l0-tetraazacyclododecane labeled biotinyl-LC-DOTA. Doses ranging from 1.85 to 5.55 N,N',N―,N―‘-tetraaceticacid and the study of its yttrium (III) complex. J Am Chem Soc 1988;)10:6266—6267. GBq (50—150 mCi) of @°Ybound to 5 mg of biotinyl-LC I7. Goodwin DA, Meares CF, McCall MJ, et al. Chelate conjugates of monoclonal DOTA were administered without any acute toxicity. One antibodies for imaging lymphoid structures in the mouse. J Nuc! Med I985;26:493— patient with brain tumor achieved >50% regression, whereas 502. 18. Goodwin DA. Pharmacokinetics and antibodies. J Nuc! Med I987;28: 1358—1362. another showed total regression of one of his metastases. 19. Begent RHJ, Green Al, Bagshawe KD, et al. Liposomally entrapped second antibody Overall, 4 of 11 patients were still in partial remission 1—5mo improves tumor imaging with radiolabeled (first) antitumor antibody. Lancet I982;ii: after therapy. 739—742. 20. Rowlinson G, Snook D, Busza A, Epenetos AA. Antibody guided localization of A Phase I/Il clinical pretargeted @°YRAIT trial is currently intraperitoneal tumors following intraperitoneal or intravenous antibody administra being performed by NeoRx Corporation (Seattle, WA) in tion.CancerRes1987;47:6528—6531. human colon, lung and ovarian cancer. The pretargeted pancar 21. Haseman ML Goodwin DA, Meares CF. et al. Metabolizable ‘‘‘In-chelateconjugated anti-idiotype monoclonal antibody for radioimmunodetection of lymphoma in mice. cinoma MAt, NR-LU-lO-streptavidin conjugate is being used Eur J Nuc! Med I986;12:455—460. with 111In-labeled (for dosimetry) and 9°Y-labeled(for therapy) 22. Shi XB, Gumerlock PH, Wellman AA, et al. Rapid PCR construction of a gene biotin-DOTA. The goal is to demonstrate the pretargeting containing Lym-l antibody variable regions. PCR Methods App! 1993:3:46—53. 23. Goodwin DA, Meares CF. McCall Mi, Mcligue M, Chaovapong W. Pre-targeted principle and lack of toxicity in humans and then to see if the immunoscintigraphy of murine tumors with indium-I I I-labeled bifunctional haptens. the excellent preclinical therapy results in tumor mice (43) can J Nuci Med 1988;29:226—234. 24. Hnatowich Di, Virzi F, Rusckowski M. Investigations ofavidin and biotin for imaging be reproduced on scaling up to patients with cancer. applications. JNuc!Med 1987:28:1249—1302. 25. Le Doussal J-M, Martin M, Gautherot E, Delaage M, Barbet J. In vitro and in vivo CONCLUSION targeting of radiolabeled monovalent and divalent haptens with dual specificity monoclonal antibody conjugates: enhanced divalent hapten affinity for cell-bound Biotinyl-LC-DOTA, biotin-P-DOTA and biotin-P-CITC bind antibody conjugate. J Nuc! Med I989;30:1358—1366. the diagnostic and therapeutic radiometals ‘I11n and 88Y with 26. Paganelli GM, Fazio, F. Monoclonal antibody pretargeting techniques for tumor excellent in vivo stability and have the required physiological localization: the avidin/biotin system. Nuc! Med Commun I991;12:2I1—234. 27. Goodwin DA, Meares CF. Pretargeting. General principles. Cancer l997;80:2675— properties for pretargeted diagnosis and therapy. Although 2680. kidney uptake of streptavidin may be a potential problem in 28. Yuan F, Baxter LT, Jam RK. Pharmacokinetic analysis of two-step approaches using RAIT using the biotin/streptavidin system, renal uptake was bifunctional and enzyme-conjugated antibodies. Cancer Res 1991 51 :3 119—3130. 29. Wilchek M, Bayer EA. The avidin/biotin complex in immunology. Immuno! Today much lower when streptavidin was conjugated to MAb. l984;5:39—43. 30. Wilchek M, Bayer EA. The avidin/biotin complex in bioanalytical applications. Ana! Biochem 1988;17I:1—32. ACKNOWLEDGMENTS 31. Green NM. Avidin. Adv Protein Chem 1975;29:85—I33. We thank Carol I. Diamante, Warak Chaovapong, Mm Li, 32. Green NM. The use of [‘4Cjbiotinfor kinetic studies and for assay. Biochemistry Michael J. McCall and Douglas P. Greiner for expert technical 1963;89:585-591. 33. Paganelli G, Riva P. Deleide G, et al. In vivo labeling of biotinylated antibodies by assistance and helpful discussions. These studies were supported in radioactive avidin: a strategy to increase tumor localization. ml .1Cancer I988;2:121— part by a grant from the Veterans Administration, Public Health 125. Service Grants CA 28343 and CA 48282 and CA1686 1. 34. Paganelli G, Chinol M, C G, et al. Therapy trials in cancer patients using an improved 3-step pretargeting approach. In: Bergman H, Kaa SH, ed. Radioactive isotopes in c!inica! medicine and research. Basel: BirkhãuserVerlag; 1996:513—517. REFERENCES 35. Renn 0, Goodwin DA, Snider M, Moran JK, Jacques V. Meares CF. New approaches I. Yuanfang L, Chuanchu W. Radiolabeling of monoclonal antibodies with metal to delivering metal-labeled antibodies to tumors: synthesis and characterization of new chelates. Pure App! Chem I991;3:427—463. biotinyl chelate conjugates for pre-targeted diagnosis and therapy. J Contro!!ed 2. MearesCF. Moi MK, Diril H, et al. Macrocyclicchelatesof radiometalsfor diagnosis Release I996;39:239—249. and therapy. Br J Cancer I990;62:2I—26. 36. Bos ES, Kuijpers WHA, Meesters-Winters M, et al. In vitro evaluation of DNA-DNA 3. Bloomer WD, Lipsztein R, Dalton iF. Antibody mediated radiotherapy. Cancer hybridization as a two-step approach in radioimmunotherapy of cancer. Cancer Res 1985;55:2229—2233. 1994;54:3479—3486. 4. DeNardo Si, DeNardo GL, Peng J-S, Coicher D. Monoclonal antibody radiopharma 37. Senter PD. Activation of prodrugs by antibody-enzyme conjugates: a new approach to ceuticals for cancer radioimmunotherapy. In: Burchiel S, Rhodes B, eds. Radioimmu cancer therapy. FASEB J 1990;4:188—193. noimaging and radioimmunozherapy. New York: Elsevier; 1983:409—417. 38. Goodwin DA, Meares CF. Watanabe N, et al. Pharmacokinetics of pretargeted 5. DeNardo GL, DeNardo SJ, O'Grady LF, Levy NB, Adams GP, Mills SL. Fractionated monoclonal antibody 2D12.5 and Y-88-JANUS-2-(p-nitrobenzyl)-1,4,7,l0-tetraazacy radioimmunotherapy of B-cell malignancies with ‘31I-Lym-A. Cancer Res 1990;50: clodecanetetraacetic acid (DOTA) in BALB/c mice with KHJJ mouse adenocarci 10145—10165. noma: a model for Y-90 radioimmunotherapy. Cancer Res 1994:54:5937—5946. 6. Halpem SE, Hagen PL, Garver PR, et al. Stability, characterization and kinetics of 39. Paganelli G, Grana C, Chinol M, et al. Antibody guided three-step radiotherapy with indium-I I I-labeled monoclonal antitumor antibodies in normal animals and nude 90-Y-biotin in glioma patients. In: Rowlinson-Busza DO, ed. The 14h internationa! mouse-human tumor models. Cancer Res I983;43:5347—5355. conftrence on advances in the app!ication of monoc!ona! antibodies in coincidenta! 7. Hnatowich Di, Layne RL, Lanteigne D, Davis MA. Radioactive labelling of antibody: onco!ogy. Santorini, Greece: Thira; 1997:47. a simple and efficient method. Science 1982;220:6I3—6I5. 40. Schechter B, Silberman R, Arnon R. Wilchek M. Tissue distribution of avidin and 8. Deshpande SV, DeNardo Si, Kukis DL, et al. Yurium-90-labeled monoclonal antibody streptavidin injected to mice. Effect ofavidin carbohydrate, streptavidin truncation and for therapy: labeling by a new macrocyclic bifunctional chelating agent. J Nuc! Med exogenous biotm. Eur J Biochem 1990;189:327—331. I990;31:473—479. 41. Rosebrough SF. Plasma stability and pharmacokinetics ofradiolabeled deferoxamine 9. Kozak RW, Raubitschek A, Mirzadeh 5, et al. Nature of the bifunctional chelating biotin derivatives. J Pharmaco! Exp Ther 1993;265:408—415. agent used for radioimmunotherapy with yttrium-90 monoclonal antibodies: critical 42. Paganelli G, Magnani P. Zito F, et al. Three-step monoclonal antibody tumor targeting factors in determining in vivo survival and organ toxicity. Cancer Res 1989;49:2639— in carcinoembryonic antigen-positive patients. Cancer Res I991;51 :5960—5966. 2644. 43. Axworthy DB, Beaumier PL, Bottino Bi, et al. Preclinical optimization of pretargeted 10. Wessels BW, Rogus RD. Radionuclide selection and model absorbed dose calculations radioimmunotherapy components: high-efficiency, curative @°Ydelivery to mouse for radiolabeled tumor associated antibodies. Med Phvs 1984;I 1:638—645. tumor xenografts [Abstract]. Tumor Targeting 1996:2:156.

1818 THEJOURNALOFNUCLEARMEDICINE•Vol. 39 . No. 10 . October 1998