Cell surface display of a lysosomal enzyme for extracellular ... - Nature

Gene Therapy (2001) 8, 1005–1010  2001 Nature Publishing Group All rights reserved 0969-7128/01 $15.00 www.nature.com/gt

RESEARCH ARTICLE

Cell surface display of a lysosomal enzyme for extracellular gene-directed enzyme prodrug therapy D Heine, R Mu¨ller and S Bru¨sselbach Institute of Molecular Biology and Tumor Research (IMT), Philipps-University Marburg, Emil-Mannkopff-Strasse 2, D-35033 Marburg, Germany

Prodrug conversion is a promising approach to cytotoxic gene therapy if an efficient transfer of the generated drug to adjacent cells can be achieved. To maximize the efficacy of this strategy we sought to develop a system that is based on a human enzyme, acts extracellularly yet in close vicinity of the transduced cell and can be used with multiple prodrugs. Results obtained with a secreted version of human ␤-glucuronidase suggested that this enzyme could be a suitable candidate, although a more stringent retention of the

enzyme at the site of the producer cell, such as its attachment to the cell surface, would be desirable. Here, we show that the fusion of the transmembrane domain of the human PDGF receptor to a C-terminally truncated form of human ␤-glucuronidase results in its surface accumulation at high steady-state levels. Using a doxorubicin prodrug, we demonstrate that this GDEPT system produces a strong bystander effect and has potent antitumor activity in vivo. Gene Therapy (2001) 8, 1005–1010.

Keywords: ␤-glucuronidase; prodrug; HMR 1826; doxorubicin; GDEPT

Introduction The inefficacy of cancer chemotherapy is largely due to dose-limiting, toxic side-effects of conventional drugs and the generation of multidrug-resistant tumor cells due to insufficient drug concentrations at the tumor site.1 One possibility to circumvent this problem is gene-directed enzyme prodrug therapy (GDEPT), also referred to as molecular chemotherapy.2,3 Central to this approach is the tumor-directed delivery of a gene encoding an enzyme that cleaves a systemically applied inactive prodrug to a toxic drug. Since prodrugs usually possess little toxicity, relatively large amounts can be administered, which in turn will lead to high drug concentrations at the tumor. Most GDEPT systems available at present, such as the herpes simplex virus thymidine kinase/ganciclovir system, function intracellularly and thereby limit the efficacy of the system and choice of the prodrug to membrane-permeable molecules.4–7 In an attempt to improve this situation, prokaryotic carboxypeptidase G2 displayed on the cell surface has been developed for GDEPT.8,9 However, the use of a non-human enzyme may provoke undesired immune responses, particularly if multiple applications are required. We therefore sought to design a GDEPT system that is based on a human enzyme, functions extracellularly and can potentially be used with multiple prodrugs. ␤-Glucuronidase seems to be a particularly suitable candidate as a prodrug-converting enzyme, because the endogenous enzyme is located in lysosomes10 and therefore not available for conversion of glucuronide pro-

Correspondence: R Mu¨ller Received 30 January 2001; accepted 2 April 2001

drugs, which are generally hydrophilic and therefore cellimpermeable. In addition, although the pH-optimum of human ␤-glucuronidase is 苲5, it is also strongly active at pH values typical of the tumor microenvironment (around 6.5).11 Like other lysosomal enzymes, ␤-glucuronidase contains a signal peptide for protein synthesis at the rough endoplasmic reticulum and N-glycosylation sites for mannose-6-phosphate-dependent sorting in the trans-Golgi network to the lysosomal compartment.12 We have previously shown that tumor cells transduced to secrete ␤-glucuronidase can efficiently convert a membrane-impermeable doxorubicin prodrug to the toxic drug doxorubicin, and that this GDEPT system can mediate tumor cell killing both in vitro and in vivo.13 A potential problem with this approach, however, lies in the possibility that the secreted enzyme may leak from the tumor with the result of toxic side-effects and decreased enzyme concentrations at the site of the tumor. In the present study, we report the successful conversion of lysosomal human ␤-glucuronidase to an enzymatically active, cell surface-displayed form, and show its functionality as part of a GDEPT approach in animal models of human carcinomas. We propose that analogous approaches will be applicable to other lysosomal enzymes, which may allow for the establishment of a new class of GDEPT systems based on hydrolytic human enzymes.

Results and discussion Design of a functional enzyme with a membrane anchor To express the human ␤-glucuronidase on the cell surface for localized extracellular prodrug activation we engineered different forms of the enzyme and tested these

Extracellular prodrug conversion D Heine et al

1006

Figure 1 Steady-state level of various ␤-glucuronidase fusion proteins after transient transfection of COS-7 cells determined by immunoblot analysis. The constructs are described in detail in Table 1. Numbers at the bottom indicate the enzymatic ␤-glucuronidase activity (MUG conversion) in the cell lysates relative to s-␤Gluc after subtraction of the background activity of non-transfected COS-7 cells (see Materials and methods for details). The double bands in the RSV lanes presumably represent the full-length fusion protein and a cleaved product lacking the transmembrane domain.

for enzymatic activity with the fluorescent substrate 4methylumbelliferyl-␤-glucuronide (MUG) after transient transfection into COS-7 cells (Figure 1 and Table 1). Based on these assays, a fusion protein of the PDGF receptor transmembrane domain14 with a ␤-glucuronidase molecule that was C-terminally truncated at Thr-633 (␤633PDGFR) turned out to be the most promising construct. This C-terminal truncation was essential because of Cterminal processing of the de novo synthesized polypeptide. Deletion of three additional amino acids (␤630PDGFR) reduced overall activity by ⬎90% presumably due to elimination of the fourth N-glycosylation site. We also fused the transmembrane domain of the asialoglycoprotein receptor H1 (ASGPR)15 and the respiratory syncytial virus (RSV) G-protein16 to N-terminally truncated ␤glucuronidase lacking the endogenous signal peptide. The ASGPR and RSV G-protein are type II transmembrane proteins with very short intracellular N-terminal domains and very long extracellular C-termini. Its N-terminal portion naturally acts as a combined signal and anchor sequence and directs the orientation of membrane integration. However, proteins with this signal/anchor sequence fused to human ␤-glucuronidase without (RSV␤G) or with a four-amino acid spacer (RSV-G4-␤G) showed very low, if any, enzymatic activity (Table 1). We therefore performed all further studies with construct ␤633-PDGFR, subsequently referred to as TM-␤Gluc. Table 1

Cell surface expression of TM-␤Gluc The subcellular localization of the recombinant TM␤Gluc was analyzed by two different approaches. First, the detergent saponin was used to extract soluble proteins from transiently transfected COS-7 cells after metabolic labeling with 35S-methionine.15 Saponin-soluble and insoluble proteins were analyzed by immunoprecipitation with a ␤-glucuronidase-specific monoclonal antibody. For comparison, a secreted ␤-glucuronidase lacking a transmembrane domain was included (s-␤Gluc).13 As expected, the majority of the s-␤Gluc protein was in the saponin fraction whereas the TM-␤Gluc was found in the fraction of saponin-insoluble proteins (Figure 2).

Figure 2 Subcellular localization of TM-␤Gluc. Saponin extractability of TM-␤Gluc compared with s-␤Gluc from 35S-labeled COS-7 cells. Saponin-soluble (s) and saponin-insoluble (ins) ␤Gluc protein was detected in cell extracts by immunoprecipitation as described in Materials and methods. The molecular mass of TM-␤Gluc is slightly higher than that of the native enzyme which explains the observed shift in lane 5. The saponin extract of TM-␤Gluc-transfected cells shows a faint band which presumably represents endogenous ␤-glucuronidase (lane 4; compare with lane 1). The origin of the minor amount of s-␤Gluc seen in the saponininsoluble fraction (lane 3) is unclear, but may be due to protein attached to the endoplasmic reticulum. Con, lysate from untransfected control cells showing the position of endogenous ␤-glucuronidase.

Fusion proteins of human ␤-glucuronidase and transmembrane domains

Construct

␤Gluca

Signal peptidea

s-␤Gluc RSV-␤G RSV-G4-␤G ASGPR-␤G ␤633-PDGFR (TM-␤Gluc) ␤630-PDGFR

23–651 23–651 23–651 23–651 23–633 23–630

IgG, 1–19

TM domaina

none RSV-G, 1–65 RSV-G, 1–65 ASGPR, 1–69 IgG, 1–19 PDGFR, 528–563 IgG, 1–19 PDGFR, 528–563

Orientation

Relative activityb

IgG-␤Gluc RSV-␤Gluc RSV-GGGG-␤Gluc ASGPR-␤Gluc IgG-␤Gluc-PDGFR IgG-␤Gluc-PDGFR

++ −c −c − +++ +

Numbers correspond to amino acid positions of the parental wild-type proteins. Position 1 of ␤-glucuronidase corresponds to the Nterminal amino acid of the signal peptide of the unprocessed proform. b Conversion of the ␤-glucuronidase substrate MUG by cell lysates from transiently transfected COS-7 cells. Enzymatic activities (MUG conversion) are indicated relative to ␤-glucuronidase protein levels assessed from the data in Figure 1. Each + corresponds to a difference in activity of 苲3-fold. −, no significant activity detectable. c Very low level of expression. PDGFR, platelet-derived growth factor receptor; ASGPR, asialoglycoprotein receptor; RSV-G, respiratory syncytial virus G-protein; G4, GlyGly-Gly-Gly spacer; TM, transmembrane. a

Gene Therapy


Second, cell surface expression of ␤-glucuronidase was analyzed by indirect immunofluorescence of transiently transfected cells and laser scanning microscopy. Unfixed cells were incubated with a ␤-glucuronidase-specific monoclonal antibody and labeled with a Cy3-conjugated secondary antibody. Figure 3 shows a series of confocal microscopy images representing different levels of the same TM-␤Gluc-transfected cell scanned from top to bottom. The data clearly demonstrate expression of TM␤Gluc on the cell surface. In contrast, no fluorescence was detected with cells transfected with control plasmid or s-␤Gluc (data not shown).

Activation of a doxorubicin prodrug by TM-␤Gluc We next investigated the ability of TM-␤Gluc to cleave the glucuronide prodrug of doxorubicin, ␤-glucuronyldoxorubicin [N-(4-␤-glucuronyl-3-nitrobenzyloxy- carbonyl)-doxorubicin; HMR1826] originally developed for antibody-directed enzyme prodrug therapy17 and also successfully used for experimental monotherapy.18 ␤Glucuronyl-doxorubicin is hydrophilic and cannot pass cell membranes, but after cleavage the generated hydrophobic doxorubicin can enter cells and intercalates into DNA. In the same experiment we sought to assess a potential bystander effect. For this purpose, tumor cells were transiently transfected at a deliberately low transduction efficiency of 苲2%. Two days after transfection the cells were incubated with prodrug for 16 h. Transfected cells were visualized by indirect immunofluorescence for ␤-glucuronidase (Figure 4, green fluorescence), while prodrug conversion and spread of the generated drug were assessed by observing the red autofluorescence of intercalated doxorubicin (emission maximum at 苲600 nm). As shown in Figure 4e, a dramatic bystander effect was seen with the TM-␤Gluc/␤-glucuronyl-doxorubicin system: all cells surrounding the two successfully transduced cells (marked by arrows) were strongly positive for doxorubicin autofluorescence. As expected, untransfected cells showed only weak fluorescence (Figure 4g), whereas all cells in the s-␤Gluc-transfected culture were equally positive (Figure 4b) due to the fact that the secreted enzyme can freely diffuse through the culture medium. In contrast, a gradient of decreasing doxorub-

Figure 3 Confocal microscopy of TM-␤Gluc-transduced cells. COS-7 cells were transiently transfected with TM-␤Gluc using DEAE-dextran and analyzed by indirect immunofluorescence 48 h later. The images show different levels of the same cell scanned from top (A) to bottom (F).

icin autofluorescence was seen around TM-␤Gluc-transduced cells confirming that the enzyme is not released from the cells. This observation also indicates that the observed bystander effect is mediated by the generated doxorubicin.

1007

In vivo functionality of the TM-␤Gluc-based GDEPT system Finally, we investigated the efficacy of the TM-␤Gluc/␤glucuronyl-doxorubicin system in vivo using two different human tumor models, JEG-3 choriocarcinoma and A549 lung adenocarcinoma, stably transduced with TM␤Gluc. These cells, in a mixture with 苲50% non-expressing cells, were injected s.c. into nu/nu mice. Expression of TM-␤Gluc in a similar fraction of cells in the established tumors was confirmed by histochemistry (data not shown). For comparison, tumors were also established from non-expressing control cells. At a tumor size of 苲40 mm3 prodrug was injected into the tail vein and tumor growth was monitored. As shown in Figures 5 and 6, the in vivo growth behavior of TM-␤Gluc-transduced JEG-3 and A549 cells in the absence of prodrug treatment resembled that of the non-transduced parental cells, and prodrug treatment had no significant effect on the nontransduced tumor cells. In contrast, tumors established from TM-␤Gluc expressing JEG-3 cells regressed after a single prodrug application (Figure 5) and some of these reappeared only after ⬎20 days. Two of six mice were still tumor-free after ⬎6 months. Likewise, a single prodrug treatment strongly inhibited the growth of TM␤Gluc expressing A549 tumors for more than 50 days (Figure 6). The stronger therapeutic effect seen with JEG3 tumors is presumably due to the fact that JEG-3 cells are more sensitive to doxorubicin than A549 cells (IC50 in vitro 苲30 nm for JEG-3 versus 苲100 nm for A549 cells). These results clearly demonstrate the functionality of the TM-␤Gluc/␤-glucuronyl-doxorubicin system established in the present study. In summary, we have constructed a cell surfacedisplayed form of the normally lysosomal human ␤-glucuronidase, which is able to convert an inactive glucuronidated prodrug to the cytotoxic doxorubicin with high efficiency. We have demonstrated that the generated drug is efficiently taken up by non-transduced cells resulting in a pronounced bystander effect. Using human xenograft models, we have also shown that TM-␤Gluc in combination with the prodrug efficiently induces tumor cell killing in vivo, even after a single prodrug application. This GDEPT system is characterized by multiple features that collectively might represent a substantial advantage when compared with the systems currently in use. (1) ␤-Glucuronidase is a human enzyme that is nonimmunogenic. (2) The TM-␤Gluc system functions extracellularly, thereby allowing for an extended bystander effect. (3) The TM-␤Gluc is attached to the transduced cell which should minimize leakage from the site of the tumor. With the currently available tools, ie stably transfected tumor cell lines, this hypothesis is, however, difficult to test. Due to the different localization of the enzyme the system can hardly be standardized, since this would require the formidable task of identifying cell lines giving rise to similar levels of secreted and membranedisplayed enzyme, respectively. A comparative analysis can be conclusively addressed once systemically applicable, targeted vectors delivering different forms of Gene Therapy


1008

a

b

c

d

e

f

g

h

Figure 4 Cytotoxic bystander effect of ␤-Gluc-mediated prodrug conversion in cell culture. Cells expressing either s-␤Gluc (a–c) or TM-␤Gluc (d–f) or non-transduced control JEG-3 cells (g and h) were incubated with ␤-glucuronyl-doxorubicin for 16 h. (a, d) detection of ␤-glucuronidase by indirect immunofluorescence (green); (b, e, g) red autofluorescence of nuclear doxorubicin (cleaved prodrug); (c, f, h) staining with Hoechst 33258 to visualize nuclei (blue).

Figure 5 Effect of the TM-␤Gluc/␤-glucuronyl-doxorubicin system on tumor growth on JEG-3 choriocarcinoma in vivo. TM-␤Gluc expressing cells (TM-␤Gluc) and non-expressing control cells (con) were injected s.c. into nude mice (106 cells per mouse). When tumors had reached a size of 苲40 mm3 the mice were divided into subgroups receiving 100 mg/kg (JEG3) of prodrug (+PD) or no treatment. Tumor growth was recorded every 2–3 days. The calculated tumor volume was plotted against the time after prodrug treatment. Data points for each individual animal are shown. Different symbols of the same color represent different animals of the same group. Curves were obtained by applying an exponential curve fit algorithm to the combined data points obtained for each group of mice.

Gene Therapy

Figure 6 Effect of the TM-␤Gluc/␤-glucuronyl-doxorubicin system on tumor growth on A549 lung adenocarcinoma in vivo. Experimental details were as in Figure 5, except that 250 mg/kg of prodrug were applied.

␤-glucuronidase to a tumor are available, since this will obviate the need for any standardization other than the application of similar amounts of vector particles. We have initiated studies towards this goal. (4) The ␤-glucuronyl-doxorubicin prodrug shows little toxicity when delivered systemically,19,20 so that high drug concentrations are achievable locally at the tumor. (5) ␤-Glucu-


ronyl-doxorubicin is hydrophilic which prevents its entry into, and thus its activation of non-transduced cells. (6) The generated drug doxorubicin is lipophilic and therefore rapidly taken up by cells in the vicinity of the transduced cell, thus limiting its escape into the circulation, which is an important consideration in view of its strong cardiotoxicity. (7) Extracellular ␤-glucuronidase is found at high levels in necrotic tumor areas18,21 which will potentiate the efficacy of TM-␤Gluc-based GDEPT approaches. (8) Other ␤-glucuronide products have been described,19,22–29 suggesting that the same enzyme can potentially be used with a plethora of other chemotherapeutics. This also opens up the possibility of combining different prodrugs similar to conventional chemotherapy. (9) Finally, analogous approaches should also be applicable to other hydrolytic lysosomal enzymes thus providing an opportunity for expanding the very small group of human enzymes30 currently available for GDEPT approaches.

Materials and methods Plasmid constructs Human ␤-glucuronidase c-DNA cloned into the pcDNA3 expression vector was used to derive all other constructs by standard cloning procedures.13 All heterologous sequences were inserted using synthetic oligonucleotides and plasmids were verified by DNA sequencing. The precise amino acid sequences and points of insertion into ␤glucuronidase are shown in Table 1. All plasmids were amplified in E. coli and purified according to a standard protocol (Qiagen, Hilden, Germany). For generating stable clones TM-␤Gluc was cloned into the LNCX vector (Clontech, Palo Alto, CA, USA). In all constructs the inserted gene is driven by the CMV promoter. Cell culture The human choriocarcinoma cell line JEG-3 (obtained from A Wellstein, Georgetown University, Washington DC), the human lung adenocarcinoma cell line A549 (kindly provided by K Havemann, Marburg, Germany), and COS-7 cells (ATCC) were cultured and passaged as described previously.13 Stable JEG-3 and A549 clones were isolated after infection with an amphotropic retrovirus generated by transfection of Phoenix A cells according to the manufacturers protocol (Clontech). Determination of enzymatic activity and protein expression COS-7 cells grown on 3-cm dishes were transiently transfected with the different ␤-glucuronidase constructs (Table 1) using the DEAE-dextran method.25 Two days later cells were lysed with 250 ␮l 0.5% sodium-deoxycholate in PBS. Aliquots were used for determination of enzymatic activity and immunoblotting. A solid phase enzyme-linked assay was used for quantification of enzymatic activity. Microtiter plates were coated with a monoclonal antibody against native human ␤-glucuronidase (hybridoma supernatant 2118/157, obtained from HH Sedlacek, Aventis Pharma, Marburg, Germany), blocked with 2% skim milk, and incubated with 50 ␮l cell extracts. Afterwards, samples were processed as described.17 For immunoblot analysis, aliquots of cell extracts were separated by SDS-polyacrylamide gel electrophoresis (10%).

The proteins were transferred on to nitrocellulose by electrophoresis using a semi-dry blotting chamber. The membrane was blocked with 5% skim milk for 2 h and incubated with the primary antibody (hybridoma supernatant 2156/94, obtained from HMR, Marburg, Germany, diluted 1/10) for 2 h at room temperature. Unbound antibody was washed five times with PBS. The membrane was then incubated with the secondary antibody (peroxidase conjugate; Santa Cruz, Heidelberg, Germany) for 2 h at RT, washed, and the enzyme expression was detected upon addition of ECL (Amersham Pharmacia Biotech, Freiburg, Germany).

1009

Saponin extraction and immunoprecipitation COS-7 cells were transfected by the DEAE-dextran technique on 6-cm dishes.31 Forty-eight hours later, the cells were labeled for 3 h with 0.2 mCi/ml 35S-methionin (Amersham Pharmacia Biotech) in DMEM + 10% dialyzed FBS. The cell monolayer was washed with PBS and treated with a buffer containing 0.1% saponin for 10 min.15 The supernatant containing the saponin-soluble proteins was collected, and the remaining insoluble material was scraped from the dish in RIPA buffer (10 mm Tris pH 7.5, 150 mm NaCl, 0.25% SDS, 1% sodium deoxycholate, 1% Nonidet P40, 1 mm DTT, protease inhibitors: 0.5 mm PMSF, 50 ␮g/ml Aprotinin, 50 ␮g/ml Leupeptin). Both fractions were analyzed by immunoprecipitation using hybridoma supernatant 2118/157 (obtained from HMR) as described.13 Immunostaining and confocal microscopy COS-7 cells were transiently transfected with DEAEdextran. Forty-eight hours later, indirect immunostaining was performed using the monoclonal antibody against human ␤-glucuronidase (hybridoma supernatant 2118/157) and a DTAF-labeled secondary anti-mouse antibody (Dianova, Hamburg, Germany). Cells were postfixed with 3.7% formaldehyde and observed under a confocal microscope. Detection of prodrug conversion Two days after transfection cells were incubated for 16 h with 15 ␮m of the ␤-glucuronyl-doxorubicin prodrug HMR1826 (obtained from HH Sedlacek). Afterwards, the cells were fixed with 3.7% formaldehyde. Converted prodrug was detected as red doxorubicin autofluorescence. Transfected cells were detected by indirect immunofluorescence as described above. Cell nuclei were visualized by staining with Hoechst 33258. In vivo studies Nude mice (strain CD-1 nu/nu; three to six animals per group) were injected s.c. with 106 control or TM-␤Gluc cells (in 200 ␮l PBS) using a 21 G needle. After 1–2 weeks tumors were visible. At a size of 苲40 mm3, prodrug (100– 250 mg/kg) was injected i.v., and tumor growth was measured at regular intervals.

Acknowledgements We are grateful to Prof HH Sedlacek for prodrug HMR1826 and monoclonal antibodies to ␤-glucuronidase, to Profs A Wellstein and K Havemann for JEG-3 and A549 cells, respectively, and to Imme Kru¨ger, Tina Stroh, Gene Therapy


1010

Julia Dick and Claudia Cybon for excellent technical assistance. This work is supported by a grant from the DFG to SB (BR 1857/2).

16

References

17

1 Ford JM, Hait WN. Pharmacology of drugs that alter multidrug resistance in cancer. Pharmacol Rev 1990; 42: 155–199. 2 Springer CJ, Niculescu-Duvaz I. Approaches to gene-directed enzyme prodrug therapy (GDEPT). Adv Exp Med Biol 2000; 465; 403–409. 3 Aghi M, Hochberg F, Breakefield XO. Prodrug activation enzymes in cancer gene therapy. J Gene Med 2000; 2: 148–164. 4 Culver KW et al. In vivo gene transfer with retroviral vectorproducer cells for treatment of experimental brain tumors. Science 1992; 256: 1550–1552. 5 Huber BE et al. In vivo antitumor activity of 5-fluorocytosine on human colorectal carcinoma cells genetically modified to express cytosine deaminase. Cancer Res 1993; 53: 4619–4626. 6 Mullen CA, Coale MM, Lowe R, Blaese RM. Tumors expressing the cytosine deaminase suicide gene can be eliminated in vivo with 5-fluorocytosine and induce protective immunity to wild type tumor. Cancer Res 1994; 54: 1503–1506. 7 Bridgewater JA et al. The bystander effect of the nitroreductase/CB1954 enzyme/prodrug sysem is due to a cellpermeable metabolite. Hum Gene Ther 1997; 8: 709–717. 8 Marais R et al. A cell surface tethered enzyme improves efficiency in gene-directed enzyme prodrug therapy. Nat Biotech 1997; 15: 1373–1377. 9 Stribbling SM et al. Regressions of established breast carcinoma xenografts by carboxypeptidase G2 suicide gene therapy and the prodrug CMDA are due to a bystander effect. Hum Gene Ther 2000; 11: 285–292. 10 Paigen K. Mammalian beta-glucuronidase: genetics, molecular biology, and cell biology. Prog Nucleic Acid Res Mol Biol 1989; 37: 155–205. 11 Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000; 407: 249–257. 12 Rosenfeld MG et al. Biosynthesis of lysosomal hydrolases: their synthesis in bound polysomes and the role of co- and post-translational processing in determining their subcellular distribution. J Cell Biol 1982; 93: 135–143. 13 Weyel D, Sedlacek HH, Mu¨ller R, Bru¨sselbach S. Secreted human ␤-glucuronidase: a novel prodrug converting enzyme for gene therapy. Gene Therapy 2000; 7: 224–231. 14 Kontermann RE, Muller R. Intracellular and cell surface displayed single-chain diabodies. J Immunol Meth 1999; 226: 179– 188. 15 Denzer AJ, Nabholz CE, Spiess M. Transmembrane orientation

18

Gene Therapy

19

20

21 22

23

24

25

26

27

28

29

30

31

of signal-anchor proteins is affected by the folding state but not the size of the N-terminal domain. EMBO J 1995; 14: 6311–6317. Lichtenstein DL, Roberts SR, Wertz GW, Ball LA. Definition and functional analysis of the signal/anchor domain of the human respiratory syncytial virus glycoprotein G. J Gen Virol 1996; 77: 109–118. Bosslet K, Czech J, Hoffmann D. Tumor-selective prodrug activation by fusion protein-mediated catalysis. Cancer Res 1994; 54: 2151–2159. Bosslet K, Czech J, Hoffmann D. A novel one-step tumor-selective prodrug activation system. Tumor Targetting 1995; 1: 45–50. Florent JC et al. Prodrugs of anthracyclines for use in antibodydirected enzyme prodrug therapy. J Med Chem 1998; 41: 3572– 3581. Platel D, Bonoron-Adele S, Dix RK, Robert J. Preclinical evaluation of the cardiac toxicity of HMMR-1826, a novel prodrug of doxorubicin. Br J Cancer 1999; 81: 24–27. Bosslet K et al. Elucidation of the mechanism enabling tumor selective prodrug monotherapy. Cancer Res 1998; 58: 1195–1201. Bakinna E, Wu Z, Rosenblum M, Farquhar D. Intensely cytotoxic anthracycline prodrugs: glucuronides. J Med Chem 1997; 40: 4013–4018. Papot S et al. Synthesis and cytotoxic activity of a glucuronylated prodrug of nornitrogen mustard. Bioorg Med Chem Lett 2000; 10: 1835–1837. de Bont DB et al. Synthesis and biological activity of beta-glucuronyl carbamate-based prodrugs of paclitaxel as potential candidates for ADEPT. Bioorg Med Chem 1997; 5: 405–414. Haisma HJ et al. A monoclonal antibody-beta-glucuronidase conjugate as activator of the prodrug epirubicin-glucuronide for specific treatment of cancer. Br J Cancer 1992; 66: 474–478. Haisma HJ et al. Construction and characterization of a fusion protein of single-chain anti-CD20 antibody and human beta-glucuronidase for antibody-directed enzyme prodrug therapy. Blood 1998; 92: 184–190. Houba PH et al. Characterization of novel anthracycline prodrugs activated by human beta-glucuronidase for use in antibody-directed enzyme prodrug therapy. Biochem Pharmacol 1996; 52: 455–463. Leu YL, Roffler SR, Chern JW. Design and synthesis of watersoluble glucuroide derivatives of camptothecin for cancer prodrug monotherapy and antibody-directed enzyme prodrug therapy (ADEPT). J Med Chem 1999; 42: 3623–3628. Wang SM et al. Specific activation of glucuronide prodrugs by antibody-targeted enzyme conjugates for cancer therapy. Cancer Res 1992; 52: 4484–4491. Waxman DJ, Chen L, Hecht JE, Jounaidi Y. Cytochrome P450based cancer gene therapy: recent advances and future prospects. Drug Metab Rev 1999; 31: 503–522. Lucibello FC et al. Periodic cdc25C transcription is mediated by a novel cell cycle-regulated represssor element (CDE). EMBO J 1995; 14: 132–142.