Actinium-225 Conjugates of MAb CC49 andHumanized DCH2CC49
Stephen J. Kennel,1 Martin W. Brechbiel,2 Diane E. Milenic,2 Jeffrey Schlom,2 and SaedMirzadeh1
1Life Sciences Division, Oak Ridge National Laboratory (ORNL), Oak Ridge, Tennessee, 2Center forCancer Research, National Cancer Institute, National Institutes of Health, Bethesda, Maryland
Radioisotopes with moderate half-lives are essential for conventional radioimmunotherapy using tumor-selective MAbs which require days for localization. Actinium-225, with a half-life of 10 days and a yieldof 4 alpha particles in its decay chain, may be an ideal choice for tumor-targeted radioimmunotherapy.Release of daughter radioisotopes from the primary chelator after the first decay has been a complica-tion with the use of 225Ac. It has been reported that the domain-deleted product of MAb CC49, Hu-DCH2CC49, is able to extravasate and penetrate more deeply into tumors than the parent IgG molecule. Wereasoned that once the 225Ac-chelate-MAb had penetrated into the tumor, the daughter radioisotopeswould remain trapped even if they had been released from the primary chelator. Actinium-225 HEHAMAb CC49 conjugates were tested for distribution, micro-distribution and therapy in immunocompro-mised mice which had LS174T tumors growing at subcutaneous or intramuscular sites. Both 125I and225Ac CC49 and Hu-DCH2 CC49 were efficient in delivery of the radioisotopes to tumor sites. Tissue mi-cro-autoradiography for the two antibody forms did not demonstrate any differences in micro-distribu-tion of either 125I or 225Ac in the tumor. Furthermore, there was no detectable difference for the two car-riers in the tumor retention of daughter radioisotopes from 225Ac. Therapy experiments with 225Ac werecomplicated by radiotoxicity of the conjugates. The lethal dose was about 0.5 mCi in two strains of miceregardless of the carrier. At injected doses of 0.5 and 0.25 mCi, CC49 was slightly active in tumor sta-sis, whereas no consistent significant effect of 225Ac- Hu-DCH2 CC49 on growth of tumors was observed.The potential of 225Ac in radioimmunotherapy is limited by the radiotoxicity of its daughter radioisotopes.Its potential will only be realized if stable conjugates, capable of daughter radioisotope retention, can bedevised.
Key Words: MAb CC49, 225Ac, HEHA, micro-distribution, therapy, radiotoxicity
219
INTRODUCTION
Alpha particle-emitting radioisotopes have thepotential advantages in cancer therapy of veryhigh linear energy of transformation (LET) radi-ation and a very short path length.1 These prop-erties can be exploited to treat small tumors inareas adjacent to essential normal tissue such as
is the case in metastases of cancer. The propertyof high LET can result in cell death by a singletraversal of an alpha particle through the cell nu-cleus.2 With agents possessing this level of tox-icity, specific delivery of the radioisotope to tar-get tissue is important.
One limitation of most a-emitters for radioim-munotherapy is their relatively short half-lives.Thus 211At (t1/2 5 7.2 hrs) and 213Bi (t1/2 5 46min) can be useful only in systems wherein tar-geting to the tumor site is extremely fast.3–5 Incontrast, conventional radioimmunotherapy uti-lizing radioisotopes coupled directly to mono-
Address reprint requests to Dr. Steve Kennel, Life SciencesDivision, Bldg 4500S, Rm F150, Oak Ridge National Lab-oratory, Oak Ridge, TN 37831-6101 Tel: (865) 574-0825Fax: (865) 576-7651.
clonal antibodies (MAbs) requires days for spe-cific localization of the targeting agent in the tu-mor.6,7 The half-life of 225Ac (10 days) is moreappropriate for this type of application. 225Ac hasthe further advantage that it decays to radioactivedaughters, which in turn emit more alpha parti-cles increasing the alpha dose at the site of de-cay.8 One recent advance is the synthesis of a bi-functionalized form of the 225Ac chelator,1 ,4 ,7 ,10,13,16-hexaazocyc looctadecane-N,N9,N0,N-,N9-, N0--hexaacetic acid, HEHA.9–11
HEHA coupled to MAbs has been shown to be arelatively stable carrier for 225Ac.12,13 When225Ac-HEHA MAb 201B was tested in a vascu-lar targeting model system, it was found to betoxic to normal, non-target tissue. This was duemostly to loss of daughter radioisotopes after thefirst a decay of 225Ac to 221Fr13 and the redistri-bution of unbound daughter radioisotopes whichoccurs rapidly in circulating blood.
A major problem with MAbs as carriers for ra-dioisotopes is the fact that they do not penetrateuniformly throughout tumor tissues.6,7 For vari-ous reasons, antibodies bind to tumors at or nearblood vessels creating a “perivascular” pattern.7,14
This microdistribution can also enable redistri-bution of daughter radioisotopes because of theproximity of blood vessels. Recent work with dif-ferent molecular forms of MAbs has shown thatantibody fragments of lower molecular weightcan penetrate tumors faster than whole IgG mol-ecules14,15; however, efficient delivery of ra-dioisotopes to tumors is limited by the rapid re-nal clearance of these fragments. An attractivealternative is use of the domain-deleted form ofthe carcinoma specific16 antibody CC49.17–19
Data show that Hu-DCH2-CC49 is a stable vari-ant of the parent IgG which can penetrate tumorsbetter7 and has a clearance time intermediate be-tween that of whole IgG and the smaller IgG frag-ments and scFvs.7,17–19 We hypothesized that ifthe 225Ac chelate were buried deeply in tumor tis-sue relatively far from the nearest blood vessel,daughter radioisotopes released from the chelate-targeting agent complex during the first decaywould be trapped in the intercellular space andredistribution to non-target sites would be re-duced. To test our hypothesis, we used the HEHAconjugates of MAb CC49 and Hu-DCH2-CC49to deliver 225Ac to LS174T tumors growing innu/nu mice.7,16,17 Our results indicate that the do-main-deleted Hu-DCH2-CC49 was no more ef-fective in promoting retention of daughter ra-
dioisotopes of 225Ac at the tumor site than wasMAb CC49. Furthermore, the whole IgG form of225Ac-CC49 was more effective in this form of targeted radioimmunotherapy.
MATERIALS AND METHODS
Animals and Cell Growth
ICR-SCID and NIH nu/nu mice were purchasedfrom Taconic Farms, Inc. (Germantown, NY) at4–6 weeks of age and were housed in a specificpathogen-free, American Association for Labo-ratory Animal Care, approved facility. Animalswere used between 8 and 12 weeks of age andtreatments were in accordance with protocol#0256 approved by the Institutional Animal Careand Use Committee. LS-174T tumor cells wereobtained from NIH and cultured in Minimal Es-sential Medium containing 10% fetal bovineserum (Biowhittaker, Walkersville, MD), non-es-sential amino acids, glutamine and Gentamycin.16
Cells were harvested by mild trypsinization atmid log growth phase, washed in completegrowth medium and re-suspended after centrifu-gation in 0.01 M Na phosphate buffer, pH 7.6,containing 0.15 M NaCl (PBS) at 1 3 107 cellsper mL. Injections of 100 mL were performed in-tramuscularly (im) or subcutaneously, (sc) within30 min of harvest.
Radionuclide Preparations225Ac was prepared from 225Ra by cation ion-ex-change chromatography as described previ-ously.13 Just prior to use, 225Ac was further pu-rified on a cation ion-exchange resin (BioRadAG50X4, 200-400 mesh, NO3 form, 250 mL bedvolume) which was loaded in 0.1 M HNO3. Thecolumn was washed with 500 mL of 1.2 M HNO3to elute Bi, and 225Ac was stripped with 750 mLof 8 M HNO3. The 225Ac/ nitric acid mixture wasevaporated to dryness and converted to the chlo-ride form. Activity was then dissolved in 3 3 250mL of 10 M HCl and passed through a 150 mLbed volume anion ion-exchange resin (BioRadMPI, 200-400 mesh, Cl-form) to eliminate Fe13.Concentrated HNO3 (0.5 mL) was added, the so-lution was collected in a glass vial and evapo-rated to dryness under a heat lamp. The purified225Ac was leached from the glass with 0.03 MHNO3 and used immediately for incorporationinto the HEHA conjugates.
220
Preparation of 225Ac HEHA and 125I MAbRadioimmunoconjugates
The preparation and properties of murine MAbCC49 and Hu-DCH2-CC49 have been describedelsewhere.16,17 MAb 520-16 is a murine MAb toa mycoplasma protein and was used as a controlIgG in these experiments. Purified IgG was dia-lyzed in conjugation buffer (0.05 M Na2CO3/NaHCO3 buffer, pH 8.6, 0.15 M NaCl, and 5 mMEDTA) for 6 h at 4°C and was reacted with bifunctional HEHA.9,10 A solution of SCN-Bz-HEHA in water was added to the antibody solu-tion such that the initial molar ratio of ligand toantibody was 50:1, 40:1 was used for MAb CC49.After 18 h at room temperature, the reaction mix-ture was dialyzed versus two changes of Chelextreated citrate buffer and then versus 2-(N-mor-pholino) ethane sulfonic acid buffer, pH 6.2 (20mM MES, 150 mM NaCl). The ligand to proteinratio for the product was determined spectropho-tometrically.20 The product was analyzed on aHPLC (ISCO, model 2350, and ChemResearch150 software, Lincoln, NE) employing a G3000size exclusion column (Toso Haas, Montgomery-ville, PA). The eluent was pumped isocraticallyat 1 mL/min at room temperature. Chromatogra-phy was monitored at 280 nm.
The conjugated antibodies were radiolabeledwith 125I in order to document the reactivity ofthe HEHA conjugate. For cell binding and biodis-tribution studies, radioiodination was done withthe succinimidyl iodobenzoate method.21 N-succin-imidyl 3-(trimethylstannyl)benzoic acid MeATE,was synthesized and purified by the method ofKoziorowski et. al.22 and was dissolved in meth-anol containing 5% glacial acetic acid. MeATE(0.2–1.0 mg in 2 mL) was added to carrier free125I (1–20 mL in dilute NaOH) and 5 mg of Chlo-ramine T was added in 5 mL of water. The reac-tion was stopped after 5 min at 40°C by additionof 1 mL of 1 mg/mL sodium metabisulfite and themixture was transferred to a syringe barrel at-tached to a Sep-Pak C18 cartridge (Waters). TwomL of water was added and the solution waspushed through the cartridge. The resin waswashed with 5 mL of water and the 125I-succin-imidyl iodobenzoate (125I-SIB) was eluted withpure methanol. The methanol solution was evap-orated to dryness under a stream of air and theresidue was dissolved in 200 mL acetonitrile. Thesolvent was evaporated again and the 125I-SIB,judged to be .75% pure by TLC, was dissolved
in 5 mL of 0.5 M sodium borate buffer (pH 58.6) and added directly to purified, concentratedantibody in 0.01 M Na phosphate buffer (pH 7.6)in 0.15 M NaCl (PBS). The reaction wasquenched after 10 min at 37°C by addition of 0.1M glycine. IgGs radioiodinated with SIB werepurified by gel filtration on Ultragel AcA34 resinin PBS containing 0.1% gelatin.
For chelation of 225Ac to HEHA conjugatedMAb, freshly prepared 225Ac in 0.03 M HNO3 wasneutralized with 3 M NaOAc to pH 5.0 and incu-bated with HEHA MAbs for 10 min at room tem-perature. The mixture was tested for incorporationby centrifugation of an aliquot in a Microcon 30centrifuge filter (Millipore). The remaining reac-tion mixture was purified by gel filtration on Ul-tragel AcA34 resin in PBS containing 5.0 mg/mLBSA as carrier. After purification, testing of prepa-rations on Microcon 30 filtration showed that theyhad .95% of the 225Ac retained in the filter cup.13
Some radiolabeled MAb preparations were ana-lyzed on sodium dodecyl sulfate polyacrylamidegels followed by phosphorimager detection of ra-dioisotopes. These studies confirmed that antibod-ies had both 125I and 225Ac bound to both heavyand light chains (data not shown).
Radiolabeled MAb were tested for binding toviable LS174T cells grown in 25 mm diametertissue culture wells. Radiolabeled MAbs dilutedin cell growth medium were added to wells in afinal volume of 1 mL. Dilution series were doneto ensure that maximum binding at subsaturationlevels could be determined.23 The cells were in-cubated in 5% CO2 at 37°C for 2 hr. The wellswere washed with chilled PBS and the cells wereremoved from the wells and transferred to glass12 3 75 mm tubes after scraping in saline con-taining 0.1 M NaOH. In some experiments,LS174T cells were harvested by trypsinizationand centrifugation. Concentrated cells in 50 mLof medium (3–9 3 107/mL) were mixed with anequal volume of antibody in 200 mL capped tubesand incubated at room temperature with end overend tube rotation for 2 h. Cells were washed freeof unbound antibody and the radioactivity ana-lyzed on a Packard Cobra Quantum gamma spec-trometer.
Animal Studies and Decay Daughter Analyses
Animals bearing tumors of 0.1 to 0.5 g after 8 to9 days of growth were injected via the lateral tail
221
vein with the purified 225Ac or 125I MAbs. Forbiodistribution studies, animals were sacrificed(3–5 per group) and organs collected andweighed as previously described.24 The 225Accontent of 17 tissues per animal was determinedat least 4 h after collection, to allow for 213Bi/225Acequilibrium, using a NaI(Tl) g-scintillationcounter (Packard 2000).
Analyses of daughter activities,25 were per-formed using a g-spectrometer (Wizard, 1480Autogamma counter, EG&G WALLAC, Turku,Finland) with a three inch crystal. The counterwas calibrated and windows set to collect gammaevents from 221Fr (g, 218 keV) for actinium quan-titation and 213Bi (g, 440 keV). Samples werecounted within 20–50 min post-harvest (see be-low) and again after 213Bi/225Ac equilibrium hadbeen reached (.4 h). To quantitate 225Ac, the in-tensity of the 218 keV g-ray of its 4.8-min 221Frdaughter was measured at the secular equilibriumwith the parent. Bi-213 was quantitated by mea-surement of intensities of g-ray emission at 440keV. Corrections for the individual g-ray inten-sities were performed and 213Bi activity at thetime of sacrifice (A2
0) was calculated by extrap-olation using the following equation.23
A20 5 [A2 2 A2(eq).(e2l2 t 2 e-l1 t)]?el2 t
wherein A2 and A2(eq) represents the activity ofthe remaining 213Bi at the time t and at equilib-rium while l1 and l2 are the decay constants of225Ac and 213Bi, respectively. At the time of sac-
rifice, the ratio of 213Bi to 225Ac is equal to theA2
0/A2(eq). This approach eliminates the need forthe absolute detector efficiencies and differencesin sample geometry. An external standard wasalso employed to arrive at the 213Bi to 225Ac ra-tio. The two data treatments provided a check onthe validity of the methods.
Tumor sizes were graded by three independentinvestigators. Due to the irregular shape of tumors,tumor volume was estimated for each animal asgrades 1–4. Grades were approximately equivalentto tumor weights of 0.1, 0.3, 0.6 and 1.0 g, re-spectively. All animals in moribund condition intherapy experiments were necropsied after sacri-fice. Tissues were fixed in 10% neutral bufferedformalin for 24 h and washed into PBS. The sam-ples were then trimmed and processed for paraffinembedding. Sections (5.0 m) were cut and stainedwith hematoxylin and eosin. Some tissue sectionswere collected and dipped in NTB-2 Emulsion forautoradiography as previously described.2 Tissuesection images were acquired with a Sony 3CCDvideo camera mounted on a Nikon microscope andinterfaced with Adobe Photoshop software on aMacintosh Power PC.
RESULTS
Antibody Conjugation and Radiolabeling
MAbs CC49, Hu-DCH2 CC49 and control 520-16 were conjugated with the isothiocyanate func-tion of bifunctionalized HEHA at reaction ratios
222
of 40:1, 50:1 and 50:1, respectively.12,13 We foundthat it was necessary to use slightly loweramounts of HEHA with MAb CC49 to preservereactivity and prevent aggregation. Several dif-ferent buffer concentrations and reaction condi-tions were tested for incorporation of 225Ac intoHEHA conjugated antibodies. It was found thatthe reaction was dependent on IgG concentrationand was inhibited by higher salt concentrations(data not shown). The conditions reported in theMethods section are optimal for the relatively lowamounts of antibody used in these experiments.In a standard reaction of 200 mg of protein with500 mCi of freshly purified 225Ac, final recoveryof 225Ac in purified, radiolabeled antibody wasonly 15–30%. Antibodies purified carrier free onHPLC gel filtration or in BSA carrier protein gelfiltration on Ultragel Aca34 were tested for pu-rity and reactivity. Autoradiography of SDS-PAGE gels of the purified, radiolabeled IgGs in-dicated that .95% of the 225Ac was bound to
either light or heavy chains of the antibodies (datanot shown). Studies with Microcon-30 filtrationindicated that the 225Ac remained bound to theantibody for at least 5 days after conjugation evenwhen the samples were stored at room tempera-ture.
Radioiodinated or 225Ac radiolabeled antibod-ies were tested for activity in binding to viableLS174T cells growing as monolayers. At limit-ing antibody concentrations the 125I CC49, Hu-DCH2 CC49 and control 520-16 bound 9%, 14%and 4%, respectively. The corresponding bindingvalues for the 225Ac radiolabeled antibodies were11%, 9% and 1%, respectively. Although theseabsolute values for binding are low, they are con-sistent with the binding affinity constants of theantibodies and the antigens concentration as dis-played on monolayer grown cells.23 To substan-tiate these claims, the three MAbs which had beencoupled with bifunctional HEHA were radio-iodinated using the 125I SIB reagent. The
223
Figure 1. Biodistribution of225Ac Mabs. ICR-SCID mice bear-ing LS174T tumors at both im andsc sites were injected iv with 225AcHEHE (0.8 mCi, 4.2 mg CC49, 2.8mg Hu-DCH2 CC49, and 1.2 mg520-16).
radioiodinated MAbs were tested for binding toLS174T cells which had been trypsinized andconcentrated to 6 3 107 cells per mL. Maximumbinding levels of the antibodies tested at 40 and10 ng per 100 mL of cells were: 29%, 58%, and2%, for 125I CC49, Hu-DCH2 CC49 and controlMAb respectively. These data indicate that theHEHA conjugated antibodies retained the major-ity of their binding activity.
BIODISTRIBUTION ANDMICRODISTRIBUTION
Biodistribution of the 125I SIB radiolabeled, HEHAconjugated MAbs was evaluated at 24 h in ICR-SCID mice bearing both subcutaneous and intra-muscular tumors. The sc tumors in the nape ofthe neck ranged from 0.02 g to 0.15 g and the imtumors were from 0.10 to 0.35 g. Data in Table1 show that accumulation of both 125I CC49 andHu-DCH2 CC49 MAbs was relatively high(21–27 %ID/g) in both sc and im tumors. In con-trast, accumulation of the control MAb was ap-proximately 1–2%ID/g in the tumors. The lackof accumulation in the throat fraction which con-tains the thyroid is an indication that the ra-dioiodine stayed bound to the MAbs and was not
released as free I- by the action of dehalogenases.The biodistribution of these reagents reportedhere agrees well with data published by others.19
Biodistribution of the 225Ac radiolabeled anti-bodies was evaluated in ICR-SCID mice bearingboth subcutaneous and intramuscular tumors. Thesc tumors in the nape of the neck ranged from0.02 g to 0.15 g and the im tumors were from0.10 to 0.35 g. Data for all three antibodies areshown in Figure 1. The values for %ID/g accu-mulation of MAbs CC49, Hu-DCH2 CC49 and520-16 at the tumor sites at 24 hrs were: 24.5,18.2 and 10.1 for the sc tumors and 8.1, 9.2, and4.9 for the im tumors, respectively. The valuesare relatively low for the tumors growing im duethe their larger size and the fact that it is difficultto dissect tumor completely free of muscle Thesevalues indicate that the 225Ac radiolabeled anti-bodies performed as well as the radioiodinatedantibodies in tumor accumulation. Actinium-225accumulation in spleen and liver increased overtime from 1–8 days.
One critical issue in targeting 225Ac is the re-tention of the daughter radioisotopes at the siteof antibody delivery. To test the relative reten-tion of the daughters for the different antibodypreparations, the ratio of the last alpha emittingdaughter, 213Bi to 225Ac, was determined at sev-
224
eral times after antibody injection. A ratio of 1.0is obtained in a closed system (test tube) whereno daughter radioisotopes can escape. The datain Table 2 show that over the course of the ex-periment, there is little difference between CC49and Hu-DCH2 CC49 in daughter retention.
Both sc and im tumors show low ratios of213Bi/225Ac indicating a deficiency of up to 50%of 213Bi. The control antibody shows ratios nearer1.0 since the 225Ac found in the tumor after in-jection of this antibody is not specifically bound,but is either circulating or trapped in the extra-cellular space. The kidney is the major site of up-take of free 213Bi, thus ratios greater than 1.0 rep-resent the binding of 213Bi that has been releasedfrom 225Ac at other sites. The liver is the majorsite of capture of free 225Ac and so the ratio inthe liver is complicated by the fact that any free225Ac in the circulation (i.e., that lost from theHEHA chelate elsewhere in the body) is contin-uously being captured. Thus the ratios in the liverwill be skewed toward lower values dependingon the extent of leakage of 225Ac and its subse-quent recapture in the liver.
The micro-distribution of both 225Ac radiola-beled antibodies and 125I radioiodinated antibod-ies was determined by autoradiography of tumorhistologic sections at various times after antibody
injection. Biodistribution data were also collectedfor the individual animals which were injectedwith 125I MAb and then sacrificed for autoradio-graphy. The corresponding biodistribution valuesfor 125I radioiodinated CC49 and Hu-DCH2CC49, from these individual animals were: 12and 10 for sc tumors and 17 and 12 for im tu-mors, respectively. Autoradiographic images forboth radioisotopes were qualitatively similar. Acomposite of the autoradiograms from 225Ac isshown in Figure 2. Two images from differentsections of the tumors demonstrate that for bothCC49 and Hu-DCH2 CC49, the radioisotope isdeposited close to the blood vessels in the tumor.No residual isotope was detected by autoradiog-raphy in tumors from mice injected with radiola-beled control MAb 520-16. If the Hu-DCH2CC49 antibody penetrated the tumor faster thanthe whole IgG molecule, a more even distribu-tion of silver grains would be expected. Carefulanalyses of the images at higher magnificationshow that there is little or no difference in the mi-cro-distributions of the two antibody prepara-tions.
The experiment was repeated and autoradi-ograms from frozen sections as well as paraffinsections were collected at both 1 and 3 days af-ter radiolabeled antibody injection. The data (not
225
Figure 2. Autoradiography of tumor sec-tions from ICR-SCID mice bearing sc (leftpanels) or im (right panels) LS174T tumors10 days after implant. MAbs, radiolabeledwith 225Ac (0.75 mCi on 1 mg MAb), were in-jected iv and the animals sacrificed after 24hr. Tumor sections were incubated underemulsion for 30 days. Black areas are silvergrains indicating the presence of 225Ac
shown) confirm the conclusion that no differencein micro-distribution between the two antibodyforms could be detected.
Therapy in ICR-SCID Mice
Initial therapy experiments were performed onICR-SCID mice bearing both sc and im tumorsof LS174T cells which had been injected 9 daysearlier. Groups of ten animals were treated with0.8 mCi of 225Ac conjugates of CC49, Hu-DCH2-CC49 or control 520-16 per animal. Although thishad been determined to be the maximum toler-ated dose, by day 6 all animals developed signsof radiotoxicity including weight loss and failureto groom. Thus the experiment had to be termi-nated 8 days after injection of 225Ac. Grading oftumor size indicated that for both the im and scinjected cells, animals treated with 225Ac CC49IgG had smaller tumors than those treated with225Ac Hu-DCH2-CC49 or control MAb 520-16;however, due to the large variation in tumor sizes,statistically significant differences could not bedemonstrated. Even though 225Ac CC49-treatedtumors were smaller, they were composed ofhealthy tumor cells and showed signs of activegrowth. Tumors implanted im were larger at treat-ment (ave 5 0.3 g) than those growing sc (ave 50.1 g). In either case, the tumors were probablytoo large for effective treatment with a-emitters,due to the uneven micro-distribution of radioiso-tope and the short path-length of the 5-8 MeV aparticles.
Therapy in nu/nu Mice
Since SCID mice are known to have a deficiencyin DNA repair,26 LD50 determinations were per-formed on nu/nu mice bred on a NIH Swissmouse background. Animals treated with 0.75mCi of 225Ac-CC49 survived, whereas the grouptreated with 1.0 mCi all died by day 8. A secondtherapy experiment was done in which animalshad either im tumors or sc tumors, but not both.The tumor sizes at treatment were about 0.1 geach. Data in Figure 3 show the size distributionof the im tumors in treated mice. Animals (group1) treated with 0.5 mCi of 225Ac-CC49 showedsignificant tumor stasis relative to growth of tu-mors in the other groups. Statistical analysis ofthe data showed that group 1 tumors were sig-nificantly smaller than those in the other groups:e.g., group 1 versus group 4, days 14–19, p ,0.0004. Large tumors in all the other treatmentgroups: group 2, Hu-DCH2 CC49; group 3, con-
226
Figure 3. Therapy of LS174T tumors growing for 12 daysin NIH Swiss nu/nu mice. Animals were treated by iv injec-tion of 225Ac HEME MAbs containing 0.5 mCi on 1.8 mgCC49 (group 1, 10 animals), 0.8 mg Hu-DCH2-CC49 (group2, 9 animals), 0.6 mg control 520-16 (group 3, 10 animals),or cold Mab CC 49 (group 4, 10 animals). Tumor sizes weregraded as described in Methods and animals were sacrificedwhen moribund or when tumors reached 1 g in size.
trol 520-16; or group 4, unlabeled CC49 neces-sitated sacrifice at day 12 after treatment. Al-though tumors in the radiolabeled CC49 treatedanimals grew slower, at sacrifice on day 27 post-therapy, histologic examination indicated that thetumors contained healthy, actively growing cells.
The experiment with animals bearing sc tumorswas again complicated by radiotoxicity. Animalsin groups 2, 3 and 4 all developed wasting syn-dromes and blotchy skin lesions which warrantedtermination. Curiously, animals in group 1, treatedwith 225Ac-CC49, did not develop signs of signif-icant radiotoxicity. Data for tumor sizes are shownin Figure 3. In this experiment animals treated withany of the 225Ac conjugates showed slower tumorgrowth relative to the cold CC49 control group. Atday 14, differences were not statistically signifi-cant; however, for day 16 comparisons of groups1, 2, or 3 versus group 4 tumor sizes showed sig-
nificant differences (p , 0.02) and for days 17–19(p , 0.0004) for all three group comparisons. An-imals in the control group 4 had to be sacrificeddue to tumor size by day 19.
In light of the continued problems with radio-toxicity, a third therapy experiment was con-ducted with Hu-DCH2-CC49 on groups of 10 an-imals each. The animals bearing either sc and imtumors were treated 6 days after implant with0.25 and 0.5 mCi of 225Ac per animal. Data forthe growth of tumors are presented in Figure 4.For sc tumors, group 1 treated with 0.5 mCi andgroup 2 treated with 0.25 mCi of 225Ac Hu-DCH2CC49 showed very slightly slower tumor growththan group 4, unlabeled Hu-DCH2 CC49; how-ever, group 3 animals, treated with 0.5 mCi ofcontrol MAb 16 showed this effect. Due to thelarge variation in size of the sc tumors, data werenot judged to be statistically significant. For im
227
Figure 4. Therapy of LS174T tumors grow-ing for 6 days in NIH Swiss nu/nu mice (10 an-imals per group). Tumor size was graded from1–4 as described in Methods. Animals weretreated by iv injection of 225Ac HEHA Mabs ata specific activity of 0.5 mCi/mg. Treatment with0.25 mCi Hu-DCH2 CC49 (groups 2&6), 0.5mCi control 520-16 (groups 3&7), or cold Hu-DCH2 CC49(groups 4&8).
tumors, tumor growth was more uniform and tu-mor growth of treated groups 5 (0.25 mCi) and 6(0.5 mCi) was retarded significantly when com-pared to control group 4 (p , 0.001 for days11–16). As for the experiment with sc tumors,animals bearing im tumors, treated with 0.5 mCicontrol MAb 16 (group 7) also had a significantretardation in tumor growth (p , 0.001) whencompared with group 8. Animals in control groups4 and 8 had to be sacrificed due to tumor size,while animals treated with 0.5 mCi of 225Ac con-jugated to either antibody or controls were sacri-ficed with signs of radiotoxicity.
Finally, a low dose experiment was performedwith 225Ac-CC49 IgG. As for the previous ex-
periment, animals with smaller tumors weretreated with 0.25 mCi injected doses (Figure 5).Animals treated with 225Ac MAb CC49 IgG hadsignificantly smaller sc tumors at days 14–16(group 1 versus group 3, p , 0.05) as well as imtumors at days 12–16 (group 4 versus 6, p ,0.001) than animals treated with 225Ac-HEHAand mixed with CC49 (groups 3 and 6).
DISCUSSION
Alpha emitting radionuclides have many proper-ties which make them potentially useful for ra-dioimmunotherapy.1,2,8 In particular, 225Ac hasthe advantages of relatively long half-life and ayield of 4 a particles per targeted nuclide.8,13,27
To put these potential advantages to use, at leastthree criteria must be met. First, it is necessaryto have a targeting agent that binds to sites at ornear all of the target cells that must be destroyed.Second, 225Ac must be bound firmly to the agentso that it is not released from the target site. Third,the daughter radioisotopes that result from 225Acdecay must be retained at the target site.
The first point has been addressed in that al-pha emitter-based therapy should be used to treatsmall tumors—either micrometasteses,24 or min-imal residual disease subsequent to some primarydebulking therapy.3,4 The second question is oneof stability of the chelator that binds 225Ac to thetargeting agent. Recent progress has been madein development of specific chelators and one, the12 coordinating ligand, HEHA, has recently beensynthesized in a bifunctional format suitable forcoupling to proteins such as MAbs.9,10,13 Com-plexes of HEHA with 225Ac seem relatively sta-ble in vivo with biological half lives of 2 days orgreater12,13 although it is still not clear that thislevel of stability is adequate for radioim-munotherapy, primarily because free 225Ac is ex-tremely toxic in animals11,13 and that only verysmall injected doses can be tolerated. The pri-mary cause of toxicity is not known11 and acutetoxicity is likely due to destruction of some component of the gastrointestinal system whichwould not be predicted by dosimetry calculationsbased on 225Ac biodistribution data. It is, how-ever, likely that the daughter radioisotopes of225Ac are the mediators of the high toxicity. Mea-surement of the biodistribution of each of theseradionuclides is impossible due to their half-lives.It is clear that retention of the daughter radioiso-topes at the target site is important to limit ra-
228
Figure 5. Therapy of LS174T tumors growing sc (groups1-3) or im (groups 4-6) for 7 days in NIH Swiss nu/nu mice(5 animals per groups). Animals were treated by iv injec-tion of 0.23 mCi 225Ac on 1 mg of CC49 (groups 1 and 4),0.23 mCi 225Ac on 1 mg control (groups 2 and 5), or 1 mgof CC49 mixed with 0.25 mCi 225Ac completed with freeHEHA (groups 3 and 6). Tumors were graded and animalssacrificed as described for Figure 3.
diotoxicity to normal tissue and to allow for max-imum absorbed dose at the tumor site.
A simplified 225Ac-decay chain shows the de-cay daughters and their half-lives.
225Ac (10d) R 221Fr (4.8m) R 217At(32ms) R 213Bi (46m) R 209Pb (3.3 h) R209Bi (stable).
The first decay results in generation of a 221Fratom that is not efficiently chelated with HEHA.Thus, the 221Fr with a half-life of 4.8 min is freeto redistribute (as are subsequent daughters)within the body. The biodistribution properties of221Fr have not been reported. One potentialmethod of holding the daughters at the target sitemay be to use a cage molecule such as a fullereneto hold the radioisotopes independent of theirchemistry. This has not yet been accomplishedand it is not known if the fullerene molecule willbe strong enough to withstand nuclear recoil fromthe high energy decays.28 The logistics and pro-tocols for preparation of such constructs are notavailable as yet.
We chose evaluation of another method to re-tain daughter radioisotopes at the target site. Thehypothesis was that if the 225Ac decay would oc-cur deep in tumor tissue, the daughters, at leastthe ones with the shortest half-lives (221Fr and217At), could be trapped in tissue extracellularspace and would then redistribute more slowly.To test this hypothesis, the domain-deleted, Hu-DCH2 CC49, which was reported to penetrateinto tumors, was employed.7 However, data forredistribution of 213Bi relative to the 225Ac dis-tribution (Table 2) showed no significant differ-ences when either whole IgG CC49 or the do-main-deleted versions were used. These dataindicate that there was little improvement indaughter retention with Hu-DCH2 CC49. We alsowere not able to demonstrate that radiolabled Hu-DCH2 CC49 penetrated more deeply into tumors.Other experimental systems will need to be de-vised to test the hypothesis thoroughly. It shouldbe noted that both the CC49 antibody and the Hu-DCH2 CC49 bind to a target molecule that re-mains at the cell surface. Thus the MAbs accumulate and stay at the cell surface. This po-sitioning of the a-emitter delivers a lower doseof a particles to the nucleus than would antibod-ies which are taken into the cell. It is also possi-ble that antibodies which are internalized wouldsequester the daughter radioisotopes retardingtheir release back into the circulation. We re-
cently have become aware of a study which dem-onstrates this effect. 225Ac bound to internalizingMAbs through the chelator DOTA were used ef-fectively to treat solid tumor xenografts growingin nu/nu mice.29 The investigators showed thatonce the 225Ac was internalized into the cell alongwith the antibody that the daughter radioisotopesremained bound in that cell. Furthermore, the in-ternalization brings the a-emitters closer to thenucleus increasing the effective absorbed dose tothe target.29 This increase in efficiency allowedanimals to be treated with very small injecteddoses thus bypassing the systemic toxicity prob-lems that we encountered our experiments. Sincethe stability of the 225Ac-DOTA complex was notassessed in vivo, it can not be concluded thatDOTA is a better chelator for 225Ac than isHEHA. To our knowledge, HEHA conjugates us-ing the internalizing antibodies have not beentested.
Finally, we observed only marginal therapeu-tic effects with 225Ac MAb CC49 and no signif-icant effect with 225Ac-Hu-DCH2 CC49. Thetherapy with the intact IgG was definitely limitedby radiotoxicity to normal organs. Approximatedosimetry estimates cosidering only the a parti-cle from 225Ac indicate that injection of 0.5 mCiof either of the conjugates would deliver onlyabout 1 Gy to the tumor which is not likely to bea high enough dose to kill all the tumor cells. Itmay be possible to use even lower injected dosesof reagents, but fractionate the dose to limit ra-diotoxic side effects as has been done clinically.The levels of accumulation of CC49 and Hu-DCH2 CC49 at the tumor site were similar, thusit is not clear that why the CC49 IgG was moreeffective therapeutically than was Hu-DCH2CC49. The most significant difference in biodis-tribution was the rapid clearance of the Hu-DCH2CC49 from the circulation. If there were a sig-nificant therapuetic effect of the decay of circu-lating alpha emitters as has been observed inother systems13,21,30,31 it may explain the differ-ences in therapy observed in this study. The ac-tual mechanism of cell killing by alpha irradia-tion is still not completely understood.
ACKNOWLEDGMENTS
The authors thank Trish Lankford, Linda Footeand Arnold Beets for technical assistance. JimWesely of Ridge Microtome prepared the sec-tions and performed the autoradiography. Drs
229
Russ Knapp and Brynn Jones were helpful in re-view of the manuscript. The Oak Ridge NationalLaboratory is managed by UT-Battelle, LLC, forthe U.S. Department of Energy under contractDE-AC05-00OR22725.
The submitted manuscript has been authoredby a contractor of the U. S. Government undercontract DE-AC005-00OR22725. Accordingly,the U. S. Government retains a nonexculsive, roy-alty-free license to publish or reproduce the pub-lished form of this contribution, or allow othersto do so, for U. S. Government purposes.
REFERENCES
1. McDevitt MR, Sgouros G, Finn RD, Humm JL, JurcicJG, Larson SM and Scheinberg, DA. Radioim-munotherapy with alpha-emitting nuclides. Eur J NuclMed 1998;25:1341.
2. Kennel SJ, Stabin M, Roeske, JC, Foote LJ, LankfordPK, Terzaghi-Howe M, Patterson H, Barkenbus J, PoppDM and Mirzadeh S. Radiotoxicity of Bismuth-213bound to membranes of monolayer and spheroid cul-tures of tumor cells. Rad Res 1999;151:244.
3. Scheinberg DA, et al. Phase 1 trial of 213Bi monoclo-nal antibody M195 in acute mylogenous leukemia. JClin Oncol 1999;18:May 16.
4. McLendon RE, Archer GE, Larsen RH, Akabani G,Bigner DD and Zalutsky MR. Radiotoxicity of system-atically administered 211At-labeled human/mouse chi-meric monoclonal antibody: a long term survival studywith histologic analysis. Int J Radiat Oncol Biol Phys1999;45:491.
5. Zalutsky MR, Akabani G, Cokgor I, Friedman HS,Coleman RE, Friedman AH, McLendon RE, Reist CJ,Pegram CN, Zhau X-G and Bugner DD. Phase I trialof astatine-211 labeled human/mouse chimeric anti-tenascin antibody administerd into malignant brain tumor resection cavities. Neuro-Oncology 1999;1:S65.
6. Behr TM, Goldenberg DM and Becker WS. Radioim-munotherapy of solid tumors: a review “of mice andmen.” Hybridoma 1999;16:101.
7. Buchsbaum DJ, Rogers BE, Khazaeli MB, Mayo MS,Milenic DE, Kashmiri SVS, Anderson CJ, Chappell LL,Brechbiel MW, and Curiel DT. Targeting strategies forcancer radiotherapy Clin Cancer Res 1999;5:3048s.
8. Geerlings MW, Caspersen FM, Apostolidis C and VanDer Hout R. The feasibility of 225Ac as a source of a-particles in radioimmunotherapy. Nucl Med Commun1993;14:121.
9. Brechbiel MW and Gansow OA. Synthesis of C-func-tionalised transcyclohexyldiethyleneaminopenta-aceticacids for labeling of monoclonal antibodies with thebismuth-212 a particle emitter. J Chem Soc Perkin TranI 1992;1:1173.
10. Chappell LL, Deal KA, Dadachova E, and BrechbielMW. Synthesis, conjugation and radiolabeling of anovel bifunctional chelating agent for 225Ac for appli-cations in radioimmunotherapy. Bioconjugate Chem2000;11:530.
11. Davis IA, Glowienka KA, Boll RA, Deal KA, Brech-biel MW, Stabin M, Bochsler PN, Mirzadeh S and Ken-nel SJ. Comparison of 225Actinium chelates: Tissue dis-tribution and radiotoxicity. Nucl Med Biol 1999;26: 581.
12. Deal KA, Davis IA, Mirzadeh S, Kennel SJ and Brech-biel MW. The improved stability of Actinium-225macrocyclic complexes. J Med Chem 1999;42:2988.
13. Kennel SJ, Chapell LL, Dadachova E, Brechbiel MW,Lankford PK, Davis IA, Stabin M, and Mirzadeh S.Evaluation of 225Ac for vascular targeted radioim-munotherapy of lung tumors. Cancer Biotherapy Ra-diopharm 2000:15:235.
14. Yokota T, Milenic DE, Whitlow M and Schlom J. Rapidtumor penetration of a single-chain Fv and comparisonwith other immunoglobulin forms. Cancer Res 1992;52:3402.
15. Adams GP, Shaller CC, Chapell LL, Wu C, Horak EM,Simmons HH, Litwin S, Marks JD, Weiner LM, andBrechbiel MW. Delivery of the alpha-emitting ra-dioisotope bismuth-213 to solid tumors via single-chainFv and diabody moleules. Nucl Med Biol 2000;27:339.
16. Milenic, DE, Yokota T, Filpula DR, Finkleman MAJ,Dodd SW, Wood JF, Whitlow M, Snoy P, and SchlomJ. Construction, binding properties, metabolism, and tu-mor targeting of a single-chain Fv derived from the pan-carcinoma monoclonal antibody CC49. Cancer Re-search 1991;51:6363.
17. Slavin-Chiorini DC, Horan Hand P, Kashmiri SVS,Calvo B, Zaremba S, and Schlom J. Biologic propertiesof a CH2 domain-deleted recombinant immunoglobulinInt J Cancer 1993;53:97–103.
18. Slavin-Chiorini, DC, Kashmiri SV, Schlom J, Calvo B,Shu LM, Schott ME, Milenic DE, Snay P, CarrasquilloJ, Anderson K, et al. Biological properties of chimericdomain-deleted anticarcinoma immunoglobulins. Can-cer Research 1995;55 (23 suppl):5957s.
19. Slavin-Chiorini, DC, Kashmiri SVS, Lee H-S, MilenicDE, Poole DJ, Beron E, Schlom J, and Horan Hand P.A CDR-grafted (humanized) domain-deleted antitumorantibody. Cancer Biotherapy Radiopharm 1997;12:305.
20. Dadachova E, Chappell L, and Brechbiel MW. Spec-trophotometric method for determination of bifunc-tional macrocyclic ligand-protein conjugates. Nucl MedBiol 1999;26:977.
21. Zalutsky MR, Narula AS. A method for the radiohalo-genation of proteins resulting in decreased thyroid uptake of radioiodine. Appl Radiat Isot 1987:38:1051–1055.
22. Koziorowski J, Henssen C, Weinreich R. A new con-venient route to radioiodinated N-succinimidyl 3- and4-iodobenzozte, two reagents for radioiodination of pro-teins. Appl Radiat Isot 1998:49:955–959.
230
23. Kennel SJ, Foote LJ, Lankford PK, Johnson M, MitchellT and Braslawsky GR. Direct binding of radioiodinatedmonoclonal antibody to tumor cells: Significance of an-tibody purity and affinity for drug targeting or tumorimaging. Hybridoma 1983;2:297.
24. Kennel SJ, Boll R, Stabin M, Schuller H and MirzadehS. Radioimmunotherapy of micrometastases in lungwith vascular targeted 213Bi. Brit J Cancer 1999;80:175.
25. Mirzadeh S, Kumar K and Gansow OA. The chemicalfate of 212Bi-DOTA formed by b-decay of 212Pb(DOTA)2-complex. Radiochemica Acta 1993;60:1.
26. Blunt T, Gell D, Fox M, Taccioli GE, Lehmann AR,Jackson SP and Jeggo PA. Identification of a nonsensemutation in the carboxyl-terminal region of DNA-de-pendent protien kinase catalytic subunit in the scid mice.Proc Natl Acad Sci USA 1996;93:10285.
27. Beyer GJ, Offord RE, Kunzi G, Jones RML, Ravin U,Aleksandrova Y, Werlen RC, Macke H, Lindros M,Jahn S and Tengblad D. Biokinetics of monoclonal an-
tibodies labelled with radio-lanthanides and 225-Ac inxenografted nude mice. J Label Cmpds Radiopharm1995;37:529.
28. Pauling L. Nature of chemical bond., 3rd Ed., CornellUni. Press, Ithaca, NY 1960.
29. McDevitt MR, Ma D, Lai LT, Simon J, Borchardt P,Frank K, Wu K, Pellegrini V, Curcio MJ, Mieder M,Bander NH, and Scheinberg D. Tumor therapy with tar-geted nano-generators. Science 2001;294:1537.
30. Blumenthal RD, Sharkey RM, Haywood L, NataleAM, Wong GY, Siegel JA, Kennel SJ and Golden-berg DM. Targeted therapy of athymic mice bearingGW-39 human colonic cancer micrometastases with131I-labeled monoclonal antibodies. Cancer Res 1992;52:6036.
31. Cooper MW, Robbins RC, Goldman CK, Mirzadeh S,Brechbiel MW, Stone CD, Gansow OA, Clark RE andWaldmann TA Use of Yttrium-90-labeled antibody inprimate xenograft transplantation. Transplantation1990;50:760.
