galactosylated antibodies and antibody-enzyme conjugates in antibody-directed enzyme prodrug therapy

1114

Galactosylated Antibodies and Antibody-Enzyme Conjugates in Antibody-Directed Enzyme Prodrug Therapy Surinder K . Sharma, Ph.D.,* Kenneth D. Bagshawe, F.R.S.,* Philip J. Burke, Ph.D.,* Joan A. Boden, L.I.Biol.,* Gordon T. Rogers, Ph.D.,* Caroline 1. Springer, Ph.D.,* Roger G. Melton, Ph.D.,t and Roger F. Shenuood, Ph.D.7

Antibody directed enzyme prodrug therapy (ADEPT) has been studied as a two- and three-phase system in which an antibody to a tumor-associated antigen has been used to deliver an enzyme to tumor sites where it can convert a relatively nontoxic prodrug to a cytotoxic agent. In such a system, it is necessary to allow the enzyme activity to clear from the blood before prodrug injection to avoid toxicity caused by prodrug activation in plasma. To accelerate plasma clearance of enzyme activity, two approaches have been studied. The studies have been performed with a monoclonal anticarcinoembryonic-antigen antibody fragment A5B7-F(abl2 conjugated to a bac- terial enzyme, carboxypeptidase G2 (CPGZ), in LS174T xenografted mice. In the first approach, a monoclonal antibody (SB43), directed at CPG2, was used, which inactivates CPG2 in vitro and in vivo. SB43 was galactosylated so that it had sufficient time to form a complex with plasma CPG2, resulting in the inactivation and clearance of the complex from plasma via the carbohydrate-specific receptors in the liver. Injection of SB43gal 19 hours after administration of the radiolabeled conjugate re-

Presented at the Fourth Conference on Radioimmunodetection and Radioimmunotherapy of Cancer, Princeton, New Jersey, Sep- tember 17-19, 1992.

From *CRC Laboratories, Department of Medical Oncology, Charing Cross Hospital, London, United Kingdom; and tPublic Health Laboratory Service, Centre for Applied Microbiology and Re- search, Division of Biotechnology, Porton Down, Salisbury, United Kingdom.

Supported by the Cancer Research Campaign, United Kingdom. The authors thank Dr. Pan Antoniw for CPG2 measurement in

tumors by prodrug conversion in vitro and Mr. Robert Boden for skilled technical assistance.

The authors thank the Cancer Research Campaign (United Kingdom) for grant support.

Address for reprints: Surinder K. Sharma, Ph.D., CRC Laborato- ries, Department of Medical Oncology, Charing Cross Hospital, Ful- ham Palace Road, London W6 8RF, United Kingdom.

Accepted for publication September 24, 1993.

duced the percentage of injected dose per gram in blood without affecting levels in the tumor.

The second approach involved galactosylation of the conjugate so that it cleared rapidly from blood via the asialoglycoprotein receptors in the liver. Localization of the radiolabeled conjugate was achieved by blocking this receptor for about 8 hours with a single injection (8 mg/ mouse) of an inhibitor that binds competitively to the receptor. This allowed tumor localization of the conjugate followed by a rapid clearance of the galactosylated conjugate from blood as the inhibitor was consumed. A tumor- to-blood ratio of 45:l was obtained at 24 hours, which increased to 100:1 at 72 hours after the conjugate injection. These accelerated clearance mechanisms have been applied in antitumor studies in ADEPT. Cancer 1994; 73:1114-20.

Key words: antibody directed enzyme prodrug therapy, enzymes, inactivation, clearance, antibodies, conjugates, galactosylation, prodrugs.

Monoclonal antibodies (MoAb) to human tumor-associated antigens have been conjugated to drugs, toxins, and radioisotopes to achieve a relatively selective kill- ing of tumor cells without much systemic toxicity. The limited effectiveness of this approach, however, has mainly been due to heterogeneity in tumor cell antigen expression, the fact that most of the cytotoxic agents require internalization by the cell to exert their toxic effect and the pharmacokinetic characteristics of radiolabeled antibodies.

The ADEPT, or antibody-directed enzyme prodrug therapy approach, has been designed to overcome some of the limitations of the antibody-drug/toxin conjugates. In the ADEPT system,'-3 antibodies directed at human tumor-associated antigens are used to target enzymes to tumors. The enzyme converts a rela-

Galactosylated Antibodies and Antibody-Enzyme Conjugates/Sharma et al. 1115

tively nontoxic prodrug into a cytotoxic active drug at the tumor site. The drug formed is designed to diffuse into nearby cells, some of which may be antigen nega- tive. Several other workers have studied similar ap- pro ache^.^-'

The targeting of enzymes to tumors is dependent on the pharmacokinetics and biodistribution of the antibody to which it is conjugated, and therefore is in- fluenced by such factors as nature of the antigen to be targeted, intact antibody or fragments, conjugation methods, and the size of the tumors. It is also desirable to reduce plasma enzyme activity after tumor localization has occurred to avoid prodrug activation in plasma leading to toxic effects. High levels of enzyme activity in plasma and nonspecific retention of conjugate in normal tissues result in prodrug activation in these tissues. Therefore, some form of enzyme inactivation or accelerated clearance mechanism is desirable to mini- mize toxicity in the normal tissues. Also, because plasma volume is usually greater than the total tumor tissue volume, the enzyme activity must be very low before prodrug injection. Reducing plasma enzyme levels reduces the level of nonspecific enzyme retention in normal tissues.

In this report, we describe two methods in which accelerated clearance of plasma CPG2 may be achieved. In the first study, a monoclonal anti-CPG2 antibody (SB43) that inactivates CPG2 was used. To avoid inactivation of CPG2 at tumor sites, SB43 was galactosylated' to facilitate the clearance of complexes formed with CPG2 via carbohydrate receptors in the liver.g An alternative approach has been studied in which the conjugate was galactosylated such that it cleared rapidly from blood. To allow localization of the conjugate, an inhibitor" that binds competitively to the asialoglycoprotein receptor was injected. This main- tained high blood levels of the conjugate for a period to allow tumor localization, followed by rapid clearance of the conjugate from blood. The applications of both accelerated clearance mechanisms in ADEPT have been studied.

Materials and Methods

Antibodies and Antibody-Enzyme Conjugates

SB43, an MoAb, was raised against CPG2 and shown to inactivate CPG2 in vitro and in vivo."

Monoclonal anti-CEA antibody, A5B7 fragments, A5B7-F(ab)2 were conjugated to CPG2 through a stable thio-ether linkage," and the conjugate was shown to retain functional properties of both compo- nents.

SB43 and A5B7-F(ab)2-CPG2 were galactosylated according to a modification of a method by Mattes.' Typically, 10 111 activated sugar were used per 200 p g SB43 and 4 pl activated sugar per 200 p g conjugate.

Asialo-Bovine submaxillary gland mucin (a-BSM) was obtained from Sigma chemicals (U.K.; A0789).

The prodrug, 4[(2-chloroethyl)]2-mesyloxy)ethyl]- aminolbenzol-L-glutamic acid, (code-named CMDA) used in experimental therapy studies was synthesized by C. J. S~ringer . '~

Radiolabeling of Antibodies and Conjugates

The SB43 antibody and the conjugate, both in native and the galactosylated form, were radiolabeled by the Chloramine T method.I4 Typically, 1 mCi Na lZ5I (from Amersham Int., UK) was mixed with the protein. The reaction was started by addition of 100 pl chloramine T (2 mg/ml) and the mixture stirred for 45 seconds at room temperature. The reaction was terminated by addition of 200 pl saturated solution (4 mg/ml) of tyro- sine. The mixture was diluted in phosphate buffer to a total volume of 2.5 ml and loaded onto a Sephadex G25 column (PDlO, Pharmacia, U.K.) to separate the free iodide. The radiolabeled protein was eluted with 3.2 ml saline and the appropriate amounts injected into ani- mals.

Blood Clearance Studies wi th SB43

Normal A2G mice (4 mice/time point) were given intravenous injections (20 pg/mouse) of radiolabeled SB43 or SB43gal. Blood samples were collected (under anesthesia) from the retroorbital sinus, at intervals of time, in pre-weighed tubes.

Biodistribution Studies

Nude mice bearing LSl74T xenografts (300-500 mg) were used in tissue distribution studies. Mice were injected with radiolabeled A5B7-F(ab)2-CPG2 (25 units CPG2/mouse) at time 0 and groups of mice (4 mice/ group) were given either saline (0.1 ml), SB43gal (250 p g , 0.1 ml anti-CPG2 antibody), or SBlOgal(250 p g , 0.1 ml irrelevant antibody), 19 hours after the conjugate injection. Mice were killed at 20 hours after conjugate injection and the tissues collected in preweighed tubes.

For biodistribution studies with galactosylated conjugates, groups of mice were injected with either radiolabeled galactosylated conjugate (20 pg/mouse) alone or given a-BSM injection (8 mg/mouse) intraperitoneally, 30 minutes before the galactosylated conjugate injection. Groups of mice (4 mice/group) were killed at 6,

1116 CANCER Supplement February 2, 2994, Volume 73, No. 3

11 0 1 2 3 4 5 6

llME(hrs)

Figure 1. Blood clearance of iodine-125-SB43 and iodine-125- SB43gal. Each A2G mouse was injected intravenously with 20 pg antibody, and the percentage of injected dose per gram blood plus standard deviation was calculated from four mice per time point.

24, and 72 hours after the conjugate injection and tissues collected in preweighed tubes.

All tissues were digested in 7 M KOH before count- ing in a gamma counter.

CPGZ in Blood and Tumor by In Vitro Activation of Prodrug

Mice were injected with A5CP-F(ab')2-CPG2 (25 units/mouse) at time 0 and one group of mice was given SB43gal injection (250 pg/mouse) at 19 hours after the conjugate. Blood and tumors were collected from both groups at 20 hours after conjugate injection and e~tracted. '~ CPG2 was measured by the capacity of these tissues to turn prodrug into drug in vitro.

Therapy Studies in LSI 74T-Xenograf ted Mice

Nude mice bearing small palpable tumors were injected with either native conjugate (50 units CPG2/mouse) followed by SB43gal (250 pg/mouse) at 19 hours or galactosylated conjugate (50 units CPGZ/mouse), 30 minutes after a-BSM injection (8 mg/mouse). The CMDA prodrug was injected to both sets of mice at 24 and 48 hours after conjugate injection (20 mg/mouse). The tumors were measured in three dimensions, twice weekly. The tumor volume was calculated relative to volume on day 1 (relative tumor volume).

Results

Blood Clearance of SB43 and SB43gal

Radiolabeled galactosylated SB43 cleared from blood within 10 minutes after injection (Fig. 1) compared with

native SB43, where %ID/g was more than 40% at that time. Some increase in radioactivity was detected in the blood of SB43gal mice after 15 minutes, which is likely to be associated with free iodide being released from the liver.

Biodistribution Studies wi th 1251-Conjugate and SB43gal

The effect of galactosylated anti-CPG2 antibody (SB43gal) on the tissue distribution of '251-A5B7- F(ab')2-CPG2 was compared with that of a galactosylated irrelevant antibody (anti-hCG antibody, SBlOgal) in nude mice bearing LS174T xenografts at 1 hour after injection of the second antibodies, i.e., 20 hours after conjugate injection. The %ID/g blood in SB43gal injected mice was approximately threefold lower (as measured by radioactivity) than in the control mice (Fig. 2). No decrease in %ID/g blood was observed in the SBlOgal-injected mice. The liver value in SB43gal-injected mice was twice that of the control value, whereas in the SBlOgal-injected mice it was only slightly higher than in the control mice. The "/.ID/g tumor was almost unaffected.

CPGZ Activity in Plasma and Tumors

CPG2 activity in both plasma and tumors was measured by capacity of these tissues to convert prodrug into active drug in vitro. At 1 hour after SB43gal injec-

2

% IW

1

0

Figure 2. Effect of SB43gal and SBlOgal on clearance and tissue distribution of iodine-125-[A5B7-F(ab)Z-CPG2] in LSl74T- xenografted mice. Each mouse was injected with conjugate (25 U CPG2, intravenously) at time 0, and groups of mice (four mice per group) were given either SB43gal (250 pg per mouse, intravenously) or SBlOgal (250 pg per mouse, intravenously) at 19 hours after receiving the conjugate. The percentage of injected dose per gram tissue was calculated at 20 hours after conjugate injection.

Galactosylated Antibodies and Antibody-Enzyme Conjugates/Sharma et al. 1117

PLASMA TUMOUR

Figure 3. CPG2 in plasma and tumor. Mice bearing LS174T xenografts were injected with 25 U ASB7-F(ab’)2-CPG2. One group of mice was given SB43gaI (250 pg/mouse, intravenously) at 19 hours. CPG2 was measured at 20 hours by capacity of these tissues to convert prodrug into drug in vitro.

tion (20 hours after conjugate), the %ID/g blood value for CPG2 was almost 10-fold lower in the SB43gal-injected mice than the control mice with conjugate alone, whereas the tumor CPG2 levels were almost the same for both groups (Fig. 3) . This gave tumor to plasma ratio of 277 for the conjugate and SB43gal-injected mice compared with a ratio of 36 in the mice with conjugate alone.

Biodistribution Studies W i t h the Radiolabeled Galactosylated Conjugate { 1251-[A5B7-F(ab)2- CPGZlgal) in LS174T Xenografts

‘251-[A5B7-F(ab’)2-CPG2]gal cleared from blood and other tissues rapidly. At 6 hours after injection, %ID/g blood was almost 40-fold lower than the native conjugate (Fig. 4, top) and the %ID/g values in other tissues was also lower. There was very little tumor localization observed at all time points studied.

In groups of mice injected with a-BSM, 30 minutes before injection of ‘251-[A5B7-F(ab’)2-CPG2]gal, %ID/ g blood was almost the same as for the native conjugate at 6 hours after injection. The %ID/g tumor was also similar to that in mice injected with the native conjugate. The %ID/g values in other tissues were either higher or similar to those for the native conjugate (Fig. 4, top). At 24 and 72 hours after the galactosylated conjugate injection, however, the conjugate had cleared from blood and other tissues except for the tumor, where it was retained (Fig. 4, middle and bottom). This gave rise to higher tumor-to-blood ratios (Fig. 5) in the

mice given a-BSM followed by galactosylated conjugate.

Therapy Studies in LSl74T Xenografted Mice

Therapy studies were carried out in LS174T xenografted mice using both methods of accelerated clearance of CPG2 activity from plasma, after localization at tumor site was achieved. Prodrug was injected without

:i *Iffi ,1 72hrs

Figure 4. The effect of a-BSM on biodistribution of 125-iodine-[A5B7-F(ab’)2-CFG2]gal in LSl74T-xenografted mice at (top) 6, (middle) 24, and (bottom) 72 hours after conjugate injection. The mice were injected with either iodine-125-[A5B7-F(ab’)2-CPG2] conjugate or iodine-125-[A5B7-F(ab)2-Cl‘G2]ga1(20 pg per mouse, intravenously) alone or with a-BSM (8 mg per mouse, intraperitoneally) 30 minutes before the conjugate injection.


toxicity, 24 hours after the conjugate injection when the plasma CPG2 levels were less than 0.1 U/ml in both cases. Similar growth delay of the tumors was observed (Fig. 6 ) in groups either given native conjugate followed by SB43gal at 19 hours and prodrug at 24 and 48 hours after conjugate or a-BSM followed by galactosylated conjugate 30 minutes later and the prodrug at 24 and 48 hours after the conjugate. Very little effect on tumor growth was observed in mice given either the prodrug alone or the galactosylated conjugate followed by the prodrug at 24 and 48 hours.

Discussion

Our previous studies with radiolabeled A5B7 and A5B7-F(ab')2-CPG2 have shown",'6 that a high con- centration of antibody in blood for up to 8-10 hours provided satisfactory uptake in the tumor. Enzyme activity persisted in blood for several days, however. To optimize use of localized conjugate, it is desirable to accelerate the clearance of enzyme activity from plasma while maintaining high levels of enzyme at the tumor sites. A monoclonal anti-CPG2 antibody, SB43, inactivates CPG2 in vitro and in vivo." Such an antibody also has the potential to inactivate CPG2 at tumor sites because normal immunoglobulins, when injected intravenously, have been shown to accumulate in tumors compared with normal tissues.17 This was avoided by galactosylation of SB43 such that it had sufficient time to complex with CPG2 and this complex together with unbound SB43gal cleared from circulation within minutes after injection.

1

C4 125-I-A5CP 1 w

RATIO 125-CA5CPgal

80 a-BSM+l25-I-A5CPgaJ

60

40

M

20 hrs 72 hrs

Figure 5. Tumor-to-blood ratios from Figure 4.

15 1

Tumaur Volume

5 i

0 1 I

0 lo DAYS 20 30

Figure 6 . Antibody directed enzyme prodrug therapy therapy in LS174T-xenografted mice. Mice were divided into groups of six and given either A5B7-F(ab')Z-CPGZ (50 U CPG2 per mouse, intravenously) at time 0 and SB43gal at 19 hours, or a-BSM (8 mg per mouse, intraperitoneally) followed by [A5B7-F(ab')2-CPG2]gal (50 U per mouse, intravenously) at time 0. The prodrug (400 mg/kg X 2 per mouse over 24 hours, intraperitoneally) was given to all groups beginning 24 hours after the administration of the conjugate.

The clearance of radiolabeled SB43 and SB43gal was studied in A2G mice. lZ5I-SB43gal cleared from blood within 10 minutes after injection. There was a slight increase in the radiolabel in blood about 30 minutes after injection, however. This is probably due to the release of iodide from liver and may be related to the decrease in radioactivity value in the liver from 99% ID/g at 10 minutes to 41% ID/g at 1 hour after injection.

The effect of SB43gal on the clearance of lZ5I- A5B7-F(ab')2-CPG2 was compared with that of a galactosylated irrelevant antibody (SBlOgal) in LS174T xenografted mice. One hour after injection of either SB43gal or SBlOgal, %ID/g blood was approximately threefold lower than the control value, SBlOgal had no effect on the clearance of the conjugate, which indicated that the effect of SB43gal on the clearance of the conjugate is specific for the SB43gal antibody. That the conjugate was taken up by the liver was indicated by the observation that liver values in the SB43gal-injected mice were twice that in the control mice. The tumor values in all groups were almost the same. This was confirmed by measurements on active enzyme concen- trations in blood and excised tumors by their capacity to turn over the prodrug in vitro. The lower %ID/g tumor values with radiolabeled conjugates as compared with the levels of CPG2 obtained by measurements of the localized enzyme in excised tissues by their capacity to convert prodrug into drug in vitro may be due to deha- logenation of the conjugate in vivo. This may lead to a

Galactosylated Antibodies and Antibody-Enzyme Conjugates/Sharma e t al . 1119

decreased effective half-life for the conjugate leading to a lower apparent level in the tumor because free iodide is not targeted to the tumor.

Therapy studies using the three-phase approach have given a better tumor growth delay in the LS174T model compared with the two-phase system.I6

We also have investigated another approach to accelerate the clearance of conjugate from blood and other normal tissues involving galactosylation of the whole conjugate such that it cleared from blood through carbohydrate receptors in the Radiola- beled galactosylated conjugate cleared from blood rapidly, and no localization in the tumor was achieved. It has been s h ~ w n , ' ~ , ' ~ however, that galactose receptors in the liver may be blocked with proteins, such as asialo-fetuin (AF) or asialo-bovine submaxillary gland mucin (a-BSM), which have exposed galactose or galac- tosamine residues. These proteins act as competitive in- hibitors of the receptor in a time- and dose-dependent manner. Our studies with AF proved to be less satisfactory (unpublished observations) than with a-BSM," because AF blocked the galactose-specific receptors in the liver for a short period only. Asialo-bovine submaxillary gland mucin, however, with a single intraperitoneal dose of 8 mg/mouse, 30 minutes before the radiolabeled galactosylated conjugate, inhibited its clearance from blood for up to 10 hours, allowing localization of this conjugate in the tumor.

As the receptor blocking protein (a-BSM) was de- graded, the galactosylated conjugate cleared within hours from blood and other normal tissues, but the conjugate was retained in the tumor. This allowed prodrug injection to be given within 24 hours after the conjugate injection without observed toxicity. The pos- sibility of applying such an approach in experimental ADEPT therapy studies has been demonstrated in the LSl74T xenograft model.

A comparison of clearance mechanism using SB43gal to inactivate and clear CPG2 from plasma 19 hours after conjugate injection, was made with the clearance mechanism using a-BSM and galactosylated conjugate. In both cases, two doses of the prodrug were given (24 and 48 hours after the conjugate).

The two groups showed similar effectiveness in terms of growth delay of the tumor.

Further studies on the level of galactosylation of conjugate and its effect on clearance of CPG2 activity from plasma and tumor localization are needed as well as optimizing the dose and timing of the liver receptor blocking agent, a-BSM.

The potential advantage of this approach in lower nonspecific retention of galactosylated conjugate in normal tissues and therefore possible reduction in nor-

mal tissue toxicity is under investigation. This approach in its current form, however, is unlikely to be applied in the clinical studies with ADEPT because the a-BSM may be highly immunogenic, and its side effects and toxicity are not known. A better alternative to the a- BSM may be an antibody" raised against the asialoglycoprotein receptor membrane. An added advantage of an antibody would be that it may be humanized to reduce immunogenicity .

The antibody SB43 has the advantage that it inactivates CPG2 within minutes. Galactose conjugation has proved to be controlled and reproducible for large-scale preparations required for clinical studies. SB43gal has been applied successfully in the clinical studies with ADEPT" where given as infusion; it reduced CPG2 levels in plasma to less than 0.1 units/ml without toxicity. Enzyme inactivation appears to confer the advantage of being more rapid in its effect than accelerated clearance. Also, with the current conjugate, which is seen to be immunogenic in humans," SB43gal inactivation/ clearance of plasma activity, combined with immuno- suppressive agents, allows a useful time window for several ADEPT cycle treatments.

References

1. Bagshawe KD. Antibody directed enzymes revive anti-cancer prodrugs concept. Br J Cancer 1987; 56:531-2.

2. Bagshawe KD, Springer CJ, Searle F, Antoniw P, Sharma SK, Melton RG, et al. A cytotoxic agent can be generated selectively at cancer sites. Br Cancer 1988; 58:700-3.

3. Springer CJ, Bagshawe KD, Sharma SK, Searle F, Boden JA, Antoniw P, et al. Ablation of human choriocarcinoma xenografts in nude mice by antibody-directed enzyme prodrug therapy (ADEPT) with three novel compounds. Eur J Caticer 1991;

4. Senter PD, Saulnier MG, Schreiber GJ, Hirschberg DL, Brown JP, Hellstrom I, et al. Anti-tumor effects of antibody-alkaline phosphatase conjugates in combination with etoposide phosphate. Proc Natl Acad Sci U S A 1988; 85:4842-6.

5. Senter PD, Schreiber GJ, Hirschberg DL, Ashe SA, Hellstrom KE, Hellstrom I. Enhancement of the in-vitro and in-vivo antitumor activities of phosphorylated mitomycin c and etoposide de- rivatives by monoclonal antibody-alkaline phosphatase conjugates. Cancer Res 1989; 49:5789-92.

6. Roffler SR, Wang SM, Chern JW, Yeh MY, Tung E. Anti-neoplas- tic glucuronide prodrug treatment of human tumor cells targeted with a monoclonal antibody enzyme conjugate. Biochem Pharmacol 1991; 42:2062-5.

7. Bosslet K, Czech 1, Lorenz P, Sedlack HH, Schuermann M, See- mann G. Molecular and functional characterization of a fusion protein suited for tumor specific prodrug activation. Br Cancer

Mattes MJ. Biodistribution of antibodies after intraperitoneal and intravenous injection and effect of carbohydrate modification. Natl Cancer Inst 1987; 79:855-63.

27:1361-6.

1992; 65~234-8. 8.


9. Thornburg RW, Day JF, Baynes JW, Thorpe SR. Carbohydrate mediated clearance of immune complexes from circulation: a role for galactose residues in the hepatic uptake of IgG-antigen complexes. J Bid Chem 1980; 255:6820-5.

10. Ong GL, Ettenson D, Sharkey RM, Marks A, Baumal R, Golden- berg DM, et al. Galactose-conjugated antibodies in cancer therapy: Properties and principles of action. Cancer Re5 1991; 5 1 : 16 19-26.

11. Sharma SK, Bagshawe KD, Burke PJ, Boden RW, Rogers GT. Inactivation and clearance of an anti-CEA carboxypeptidase G2 conjugate in blood after localization in a xenograft model. Br J Cancer 1990; 61:659-62.

12. Melton RG, Boyle JMB, Rogers GT, Burke PJ, Bagshawe KD, Sherwood RF. Optimization of small-scale coupling of A5B7 monoclonal antibody to carboxypeptidase G2. J Immuriol Meth-

13. Springer CJ, Antoniw P, Bagshawe KD, Searle F, Bisset GMF, Jarman M. Novel prodrugs which are activated to cytotoxic alkyl- ating agents by carboxypeptidase G2. ] M u d Chem 1990; 33:677- 81.

ods 1993; 158:49-56.

14. Greenwood FC, Hunter WM, Glover IS. The preparation of l3II-

labelled human growth hormone of high specific radioactivity. J Immunol Methods 1963; 89:114-23.

15. Antoniw P, Springer CJ, Bagshawe KD, Searle F, Melton RG, Rogers GT, et al. Disposition of the prodrug 4-(bis[2-chloroethyl]amino)benzoyl-L-glutamic acid and its active parent drug in mice. Br Cancer 1990; 62:909-4. Sharma SK, Bagshawe KD, Springer CJ, Burke PJ, Rogers GT, Boden JA, et al. Antibody directed enzyme prodrug therapy (ADEPT): a three phase system. Dis Markers 1991; 9:225-31.

17. Bale WF, Countreras MA, Grady ED. Factors influencing localization of labelled antibodies in tumors. Cancer Re5 1980; 40:2965-72.

18. Ashwell G, Morell AG. The role of surface carbohydrates in the hepatic recognition and transport of circulating glycoproteins. A d v Enzymol 1974; 41:99-128. Sarkar M, Liao J, Kabat EA, Tanabe T, Morell G. Binding site of rabbit hepatic lectin. J Biol Chem 1979; 254:3170-4. Bagshawe KD, Sharma SK, Springer CJ, Antoniw P, Boden JA, Rogers GT, et al. Antibody directed enzyme prodrug therapy (ADEPT): clinical report. Dis Markers 1991; 9:233-8.

16.

19.

20.

galactosylated antibodies and antibody-enzyme conjugates in antibody-directed enzyme prodrug therapy

Documents