PEDS Advance Access published online on November 22, 2007
Protein Engineering Design and Selection, doi:10.1093/protein/gzm043
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Article |
A33scFv–Green fluorescent protein, a recombinant single-chain fusion protein for tumor targeting
1Medizinische Klinik III—Hematology, Oncology und Transfusion Medicine 2 Institute for Clinical Chemistry and Pathobiochemistry, Charité—Universitätsmedizin Berlin, Campus Benjamin Franklin, Hindenburgdamm 30, D-12200 Berlin, Germany
3 To whom correspondence should be addressed. E-mail: markus.deckert{at}charite.de
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Chemical conjugates of monoclonal antibodies with fluorophores or enzymes have long been used for diagnostic purposes and experimental therapeutic approaches. Recombinant technology allows for the design and expression of tailored genuine fusion proteins, providing defined molecules as to size, molar ratios of the functional components and stability. The production of functional protein, however, is often limited or impossible due to refolding and solubility problems. Here, we report on the production of a soluble recombinant fusion construct, A33scFv–green fluorescent protein (A33scFv::GFP) in Pichia pastoris. A33scFv is a single-chain antibody recognizing the A33 antigen, which is expressed by
95% of colorectal carcinomas and has become a focus of pre-clinical and clinical investigation. The fusion partner GFP was selected both as an experimental tool for functional studies of the A33 antigen and as a potential diagnostic for colon cancer detection and therapy planning. Pichia pastoris yeast strains were transformed with A33scFv::GFP cDNA under the methanol-inducible AOX1 promotor. The construct was properly expressed and secreted into culture supernatants as a soluble protein, which was bifunctional without additional renaturation or solubilization steps. The crude protein solution was purified by affinity chromatography. Surface plasmon resonance, flow cytometry and fluorescence microscopy on sections of normal and cancerous colon tissue revealed specific binding and the applicability of this fusion protein for diagnostic purposes. In addition, the biodistribution of A33scFv::GFP was analyzed in mice bearing A33-positive tumor xenografts, confirming specific tumor targeting.
Keywords: ADEPT/A33 antibody/colon carcinoma/recombinant fusion proteins/tumor targeting
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
In recent years, specific targeting by monoclonal antibodies has become an important modality for the treatment of malignant diseases (Lin et al., 2005
). The antibodies in clinical use today are mostly humanized complete IgG antibodies derived from conventional hybridoma technique. Experimental approaches to antibody-directed enzyme-prodrug therapy have used bifunctional enzyme–antibody constructs produced by chemical conjugation (Kriangkum et al., 2001). Both approaches share the problems of large molecular size, impeding diffusion into tumor tissue, and high immunogenicity, especially if murine antibodies are applied in humans (Ritter et al., 2001). In addition, chemically derived conjugates pose problems in homogeneity and stability of the substance (Cao and Suresh, 1998).
Recombinant single-chain variable fragments (scFv) comprise only the antigen-binding variable chains of one heavy and light chain each, fused with a linker peptide (Wels et al., 1992). Compared with complete IgG, they have only half the avidity because of their monovalent antigen-binding region, and lacking the Fc fragment, they have no intrinsic immunological function (Adams and Schier, 1999). Yet, their small size of 30 kD promises favorable diffusion characteristics in tumor tissue (Jain, 1999).
This makes them particularly interesting for the development of bifunctional antigen-binding fusion proteins instead of conventional conjugates, because they can overcome some of the problems associated with chemical conjugation (Cao and Suresh, 1998). (i) Stability: many linker systems used in chemical conjugation can be instable under changing ambient conditions, whereas single-chain proteins, proteolysis and purification problems discussed below notwithstanding, tend to be stable under physiological conditions and in the absence of specific proteases. (ii) Predictability: most conjugation processes yield conjugate mixtures with a stochiometric variation in the number of partner molecules per antibody, whereas recombinant fusion proteins are of a defined structure. (iii) Size: although complete IgG has been successfully used for tumor targeting, IgG-based conjugates soon become too bulky for efficient diffusion into tumor tissue.
Recombinant fusion proteins based on scFv have been reported before with a couple of fusion partners (Kriangkum et al., 2001), including toxins (Onda et al., 2004), radionuclides (Goshorn et al., 2001) and enzymes (Deckert et al., 2003). However, their production has been technically difficult or expensive. Expression in bacteria tends to lead to insoluble inclusion bodies, which require resolubilization and additional processing to retrieve a functional protein, which may or may not succeed (Lilie et al., 1998). Mammalian expression systems are better suited to this kind of complex proteins, but are of notoriously low yield (Verma et al., 1998).
Yeast expression systems combine some of the advantages of mammalian and bacterial cultures in providing a eukaryotic cell system that can be handled in a similar way as bacteria, allowing for economic production of proteins. Pichia pastoris in particular has been described as a high-yield expression system with the capacity of secreting functional protein into culture medium (Cregg et al., 2000).
The fusion protein introduced here consists of an scFv binding the A33 antigen (A33scFv) (Rader et al., 2000) and green fluorescent protein (GFP) (Daly and McGrath, 2003). The A33 antigen is a transmembrane protein of 43 kD expressed by healthy colon epithelium and by various gastrointestinal tumors and their metastases, including 95% of colorectal carcinomas (Catimel et al., 1996). It is a novel member of the immunoglobulin superfamily (Heath et al., 1997). In a mouse model, native and polyethylene glycol-conjugated huA33 complete IgG showed highly specific binding to the A33 antigen (Deckert et al., 2000). In clinical trials, radiolabeled A33 localized specifically to colon cancer cells, where it was retained for several weeks while clearing within days from normal colon (Welt et al., 1990).
The A33scFv::GFP fusion protein was designed for two purposes: as an experimental tool and as a potential diagnostic. As the first, it would allow the detection of A33 antigen and help in untangling its as yet unclear intracellular lifecycle by fluorescence and laser confocal microscopy. Secondly, with growing interest in the A33 antigen as a therapeutic target, it may be used to detect A33-positivity of a given tumor sample and hence aid in the histological diagnosis of gastrointestinal cancers and in therapy planning. In addition, as A33scFv::GFP is very similar in calculated size and charge to other proposed A33scFv-based fusion proteins, it can be used as a model molecule for the biodistribution and microlocalization of this class of constructs without the need for additional labeling.
Here, we report on the cloning and expression of A33scFv::GFP in P.pastoris, its purification and functional description by flow cytometry, surface plasmon resonance, fluorescence microscopy and in vivo biodistribution studies.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Vector design and host system
A33scFv (Rader et al., 2000) and GFP cDNA (BD-Clontech, Mountain View, CA, USA) were PCR-amplified. Primers were designed based on published sequences to remove start and stop codons and to add flanking restriction sites for vector cloning (5'A33scFvGFP: 5'GCGATGGCCATGGTGAGCCCTAGGGAGCTCCAGATGAC; 3'A33scFvGFP: 5'TGTTCGGATCCTGCGGCCGCTTATTTGTAGAGCTCATCCATGCC). The HP9::GFP control construct, recognizing a human stromal antigen not present on human colon cancer cells or in murine xenograft tumors, was designed and cloned analogously. The fusion constructs in 5'-scFv::GFP-3' orientation were cloned into the methanol-inducible pPIC9K expression vector (Invitrogen, Groningen, The Netherlands). The resulting pPIC9K–scFv::GFP plasmid vectors were ampilified in Escherichia coli and transfected by electroporation into P.pastoris strain GS115 (Invitrogen, Groningen, The Netherlands). Pichia pastoris cultures were grown for 120 h on RDH plates [1 M sorbitol, 2% dextrose, 1.34% yeast nitrogen base (YNB), 4 x 10–5% biotin, 0.005% amino acids] and then selected for multi-copy insert constructs by incubation on RDH plates containing increasing concentrations of the antibiotic G418 up to 4 mg/ml. Clones growing on the highest G418 concentration were used for expression analysis.
Eighty transfected P.pastoris clones growing in the presence of 4 mg/ml G418 were analyzed by PCR for genomic integration of the A33scFv::GFP sequence (5'A33scFvGFP and 3'A33scFvGFP, as described earlier). PCR-positive clones were propagated in 5 ml BMGY medium [1% yeast extract, 2% peptone, 100 mM potassium phosphate (pH 6.0), 1.34% YNB, 4 x 10–5% biotin, 1% glycerol) for 24 h, and stocks of 1 ml were prepared of each clone. Culture aliquots were transferred into 5 ml BMMY medium [1% yeast extract, 2% peptone, 100 mM potassium phosphate (pH 6.0), 1.34%YNB, 4 x 10–5% biotin, 0.5% methanol], and expression of the gene under control of the AOX1 promotor was induced by maintaining the methanol concentration at 0.5% for 72 h. Culture supernatants were then screened for A33scFv::GFP protein by the immunoblot described below, and positive clones were cultured under the same induction conditions, but in a volume of 50 ml, for further analysis. Again, an analogous procedure was applied to the control fusion protein.
Aliquots of 6 µl of protein solution or culture supernatant were analyzed by electrophoresis on 10% SDS–polyacrylamide gels under denaturating conditions. For detection and initial identification of A33scFv::GFP, the proteins were blotted onto a nitrocellulose membrane by electrophoresis and immunostained with a murine anti-GFP monoclonal antibody (Dianova, München, Germany) followed by a peroxidase-conjugated polyclonal anti-mouse IgG (Transduction Laboratories, Lexington, KY, USA) and subsequent fluorogenic substrate development. Blots were documented photographically.
Protein L Plus (Pierce, Rockford, IL, USA) was employed for affinity purification of the crude supernatant following the manufacturer's instructions. Briefly, supernatant was centrifuged at 2500 g for 20 min at 4°C and dialyzed against 0.1 M potassium phosphate, pH 7.2, at 4°C for 16 h using a 25–30 A dialysis membrane (Roth, Karlsruhe, Germany). Protein L gel slurry, buffers and sample were warmed to room temperature before use. A 20 ml column (Biorad, Hercules, CA, USA) was packed with 1 ml Protein L gel. Two hundred milliliters of dialyzed culture supernatant were equilibrated with the same volume of binding buffer (0.1 M potassium phosphate, pH 7.2), and the protein L column was equilibrated with 10 ml of binding buffer before the diluted culture supernatant was loaded onto the gel. The column was washed with 20 ml of the binding buffer, and bound protein was then eluted with 10 ml of ImmunoPure Elution buffer (Pierce, Rockford, IL, USA pH 2–3). Eluted protein was immediately pooled and adjusted to pH 7.0 by 1 M potassium phosphate buffer. This eluate contained two bands of 60 and 27 kD, respectively, which were separated by size exclusion on a Bio-Gel P-30 column in the same phosphate buffer. The protein corresponding to the larger band which represented the A33scFv::GFP fusion construct was eluted in the earliest fractions.
To assess the purity of eluted protein from the Protein L column, SDS–PAGE was performed as described above. The gel was silver stained, developed and documented by digital photography. Overall protein concentration was measured by the BCA-Assay (Pierce, Rockford, IL, USA) following manufacturer's instructions. A33scFv::GFP was quantified digitally as a percentage of total protein using the AlphaEaseFC 4.0 software (Genetic Technologies, Miami, FL, USA).
Surface plasmon resonance assays were performed on chips coated with recombinant gpA33 using a BIAcore instrument (BIAcore AB, Neuchâtel, Switzerland) as described previously (Catimel et al., 1997). After 550 s, sample flow of diluted A33scFv::GFP was replaced by buffer solution. On and off rates and affinity were calculated from the curves using the BIAcore evaluation software.
Cultures of the A33 positive colon carcinoma cell line LIM 1215 and the A33 negative cell line HT 29 were grown under defined conditions until they reached 50% confluence and then suspended in medium and adjusted to 106 cells/ml. Aliquots of 106 cells were centrifuged for 30 s at 500 g. The cell pellet was resuspended with the supernatant or recombinant protein solution to be investigated or with appropriate controls and incubated for 20 min. For inhibition studies, this step was preceded by 20 min of incubation with a complete IgG antibody recognizing the A33 antigen (Ludwig Institute for Cancer Research, New York) or with the HP9::GFP fusion protein as a negative control, and subsequent washing. Flow cytometry was performed on a FACScan instrument (Becton Dickinson). Data were analyzed using the CellQuest sotware (Becton Dickinson), employing a uniform analysis gate set on the core population of the cultured cells so that events of extreme forward or side scatter were excluded.
Frozen sections of 10 µm were obtained with a cryostat and affixed to glass slides with acetone, followed by air drying. These slides were stored frozen at –80°C. Prior to staining, slides were allowed to warm to room temperature. They were then incubated with A33scFv::GFP for 20 min and washed with PBS. Fluorescence-stained sections were examined using an Axioplan fluorescence microscope (Zeiss, Oberkochen, Germany).
A33scFv::GFP was radiolabeled using iodogen beads (Pierce, Rockford, IL, USA), a solid-phase modification of the chloramine T method, as described before (Deckert et al., 2004). Briefly, two beads were submersed in 100 µl of 125I solution (74 MBq, Amersham, Freiburg, Germany). After 5 min, 1.0 mg of protein was added in a volume of 1 ml. After another 7.5 min, the reaction was terminated by removing the iodogen beads from the vial, and the product was purified on a sterile Sephadex G25 column (Amersham) preconditioned with 5% human serum albumin in sterile saline. The protein fractions were pooled and sterile filtered. Specific activity was 18.9 MBq/mg. Immunoreactivity was determined by absorption of 0.1 µg/ml A33scFv::GFP with aliquots of 2 x 107 antigen-positive cells. To determine background activity, cell pellets were pre-treated with a >100-fold excess of unlabeled A33scFv::GFP, and immunoreactivity was calculated by subtracting this background radioactivity from cell-bound 125I activity after washing twice in PBS, and dividing the result by the total activity added. Immunoreactivity measured by this method was 34%.
A33 positive tumor xenografts in mice
CD-1 nude mice were inoculated subcutaneously with the A33 positive pancreatic cancer cell line ASPC-1 or the A33-negative cell line Panc as a control. These cell lines rather than LIM 1215 were selected to reduce the phenomenon of tumor outgrowth, because in our hands they grew slower and more homogenously than LIM 1215. When the largest diameter of the tumor had reached 0.4–0.6 cm, xenograft-bearing animals and naive controls were injected IV with protein-equalized doses of 5–20 µg (0.24–0.97 MBq) of radiolabeled A33scFv::GFP or with 20 µg of radiolabeled HP9::GFP in 100 µl of sterile buffer (0.15 M NaCl, 0.1 M sodium phosphate, pH 7.2). Animals were sacrificed at 3, 6 and 24 h after conjugate injection, and samples of blood, thyroid, lung, liver, spleen and tumor were obtained and weighed. The radioactive dose in counts per minute (CPM) of these samples was measured by an automated gamma counter (1277 Gammamaster, LKB Wallac, Erlangen, Germany) and compared with an aliquot of the injected preparation as a standard. Relative doses were calculated as
|
|
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Soluble A33scFv::GFP fusion protein is secreted into P.pastoris media supernatants
A33scFv::GFP was designed so that the variable regions of the light and heavy chains were connected by an 18 amino acid Gly-Ser linker, whereas the linker between the heavy chain variable region and the ‘payload’ protein, i.e. GFP, consisted of four coherent glycins flanked by a serine on either side. This design was intended to provide high flexibility between the two immunoglobulin chains for correct quarternary folding, but a rather rigid connection between scFv and GFP so as to keep these protein elements from interfering with each other. Appropriate flanking restriction sites allowed for the modular exchange of the individual components. The pPIC9K backbone contained the 5' and 3' regions of the yeast alcohol oxidase gene 1 (AOX1), which could integrate into the chromosomal yeast DNA by omega insertion and thus replace the original AOX1 gene. In addition to the strong AOX1 promotor, the 5' flanking region of this vector contained the alpha mating factor to facilitate extracellular secretion of the expressed protein, and two additional genes, HIS4, which allows for growth on histidine deficient media, and kan, which confers kanamycin resistance. The vector map is shown in Fig. 1 (complete sequence available upon request).
|
-mating factor; AOX-5' and AOX-3', flanking regions of the AOX-I gene sequence; Amp, Kan, resistance genes against ampicillin and G418, respectively; ColE1, E.coli replication factor; HIS4orf, open reading frame of the histidine deaminase gene.After electroporation of P.pastoris strain GS115 with this plasmid vector, transfected MutS (methanol slow-utilizing) clones were further selected by virtue of the kanamycin resistance conferred with genomic integration of the AOX1 replacement sequence. A total of 80 clones were harvested that expressed the kanamycin resistance gene sufficiently to survive at the highest G418 concentration applied, which indicated the presence of multi-copy integrants. These clones were further analyzed by PCR with flanking sequencing primers for A33scFv::GFP cDNA. Of these, eight were rendered positive. They were used for subsequent expression experiments in which the promotor of the AOX1 gene, controlling the gene of interest, was induced by maintaining an optimized methanol concentration, whereas the weaker AOX2 gene was sufficient to sustain the methylotrophic yeast metabolism under these conditions. To avoid complex procedures for yeast cell lysis and protein retrieval, we only screened for A33scFv::GFP secretion into the culture supernatant. Utilizing a polyclonal murine anti-GFP antiserum (BD-Clontech), we detected bands of the predicted molecular size in supernatants of two clones (Fig. 2).
|
A33scFv::GFP is purified by affinity and size exclusion chromatography
After dialysis against phosphate buffer, affinity purification of A33scFv::GFP was performed against sepharose-immobilized protein L and subsequent size exclusion. Although the initial gels of crude supernatants showed considerable degradation at the linker site, yielding a second band of the approximate size of A33scFv, analysis of the purified material by SDS–PAGE and silver staining showed a purity of >95% according to a histogram evaluation of the gel (Fig. 3). The final yield of purified protein ranged from 0.25 to 1 mg per liter of culture.
|
The protein construct binds specifically to A33-positive cells and to the recombinant antigen in vitro
Antigen binding and fluorescence function of the fusion protein were investigated simultaneously by flow cytometry. As shown in Fig. 4a and b, A33-positive LIM1215 cells incubated with crude supernatant displayed approximately 1 log higher fluorescence levels than native cells or the same cells incubated with the HP9::GFP control contruct. Owing to the majority of other proteins, the concentration of A33scFv::GFP in the crude supernatant could only be estimated from quantitative gel analysis to an approximate concentration of 50 µg/ml. Purified A33scFv::GFP in total amounts of 5 and 20 µg per 106 cells displayed quantitatively lower total fluorescence levels than equal volumes of crude supernatant, indicating a loss of specific binding activity during the purification process.
|
Pre-treating cells either with a non-fluorescent A33-construct, A33scFv::Cdy, or with an HP9-fusion protein prior to A33scFv-GFP incubation demonstrated the binding specificity of the fusion protein, as the A33-based, but not the HP9-based construct reduced fluorescence levels on A33-positive cells (Fig. 4c). Finally, comparing cell lines with different constitutive levels of gpA33 surface expression, fluorescence staining by A33scFv::GFP was roughly quantitative (Fig. 4d), i.e. after a33scFv::GFP incubation, the fluorescence of gpA33-negative cells was at the background level of the experiment, whereas the fluorescence of intermediate-expressing HT29 cells was about half that of the gpA33 high-expressing LIM1215 cell line.
A direct assessment of the fusion protein's affinity and specificity was achieved by surface plasmon resonance on immobilized recombinant A33 antigen (Fig. 5). From the on- and off-rates, the dissociation constant of A33scFv::GFP was calculated as Kd = 20.4 nM and that of the unfused A33scFv as Kd = 22.2 nM. The curves for HP9::GFP were almost identical to the buffer controls (not shown) and were subtracted from those for the A33-binding proteins before evaluation.
|
A33scFv::GFP allows morphologic staining of normal and cancerous colon tissue
To apply the fusion protein diagnostically as intended, snap-frozen microsections of healthy and cancerous colon tissue were stained with A33scFv::GFP. As shown in Fig. 6, the structures of crypts were clearly visualized by A33-staining of healthy colon tissue, whereas the colon cancer slides displayed their complete destruction and an intense, irregular fluorescence pattern throughout the specimen. Staining identically treated slides from the same block with the HP9::GFP control construct revealed no specific fluorescence pattern.
|
A33scFv::GFP localizes specifically to A33-positive xenografts in mice
Nude mice carrying A33-positive xenografts from the pancreatic cancer cell line ASPC 1 or an A33-negative control were injected with radiolabeled A33scFv::GFP. At defined time points, groups of three animals were sacrificed and blood, tissue and tumours were analyzed for their relative radioactive doses. As shown in Fig. 7, the highest organ doses were seen in kidney and lung after 3 and 6 h, respectively, and this pattern was reflected by the radioactivity distribution in tumor-free control mice. After 24 h, however, with decreasing radioactivity in the bloodstream, this pattern inverted in favor of the tumor, which became then the tissue with the highest measured activity. In concordance with this finding, the tumor-to-blood ratio increased from <1 at 3 h to 15 at 24 h. Mice treated with the HP9::GFP control construct displayed no discernable or specific tumor uptake.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Although most antibodies and bifunctional constructs in clinical use today are produced by hybridoma technology and chemical conjugation, recombinantly engineered proteins are the next generation of tumor-targeting drugs (Cao and Suresh, 1998
). Their advantages as lined out in the introduction are intriguing, and their design process appears straightforward. Yet, to this date the expression of heterogeneous proteins in a properly folded, functional form appears like an art rather than science. No criteria exist to predict the result for any given expression system, e.g. from the origin species or sequences of the component proteins. Thus protein design, choice of expression system and expression conditions remain largely empirical (Verma et al., 1998). Pichia pastoris in theory combines the uncomplicated handling, scalability, high yield and low cost of bacterial expression systems with a eukaryotic post-translational processing close to that of mammalian cells. A large variety of recombinant proteins has thus been expressed in P.pastoris, including recently reported fusion proteins from heterogeneous sources (Macauley-Patrick et al., 2005).
Here, we report on the functional production of a fusion protein of the A33scFv single-chain antibody fragment with GFP in the P.pastoris system. Only recently has the optimized large-scale expression of the A33 scFv in P.pastoris been published (Damascenco et al., 2004). Our expression strategy was largely parallel to the one described there, but was limited to the less efficient induction conditions in shake-flasks. Under these conditions, P.pastoris allowed for stable and reproducible expression and convenient retrieval of the fusion protein for subsequent purification. With up to 1 mg/l, the total yield of purified protein was considerably lower than that reported for P.pastoris by other groups, and further optimization will be necessary for large-scale fermenter production. However, the final yield here was about 10-fold higher than we previously found for a similar A33-based fusion protein in E.coli (Deckert et al., 2003), and 250-fold higher than the crude protein obtained from expression experiments of the same construct in E.coli (unpublished data). In summary, an expression project practically impossible in various Eschericha coli strains and expression vectors due to low yield and lack of protein function was comparably easily completed in this system. Furthermore, no post-processing was needed (Misawa and Kumagai, 1999), and satisfactory purification to >95% purity was achieved in a simple two-step chromatography procedure, which also stopped the—most probably proteolytic—protein degradation at the linker site observed in the crude material.
The secreted protein displayed bifunctional activity without any further treatment, i.e. even the uncleared culture supernatant could directly be used for the detection of A33-positive cells in flow cytometry.
Binding specificity was demonstrated by blocking experiments in flow cytometry and by surface plasmon resonance on immobilized recombinant A33 antigen. The latter method also allowed for an estimation of actual affinity relative to the complete IgG antibody.
Interestingly, the purified protein showed lower binding activity in flow cytometry than the crude supernatant, despite higher fusion protein concentration and absence of degradation. Along the same line, the affinity calculated from surface plasmon resonance was considerably lower than that reported in the original publication by Rader et al. (2000) even for the unfused A33scFv, whereas relative comparison against the divalent complete IgG antibody was consistent with theoretical prediction (Kortt et al., 2001). This reduced activity of the purified material may be explained by protein denaturation during purification or by the loss of stabilizing salts or proteins in the culture medium.
As one potential diagnostic application of the A33scFv::GFP fusion protein, histological fluorescence staining was investigated in specimens of normal and cancerous colon tissue, confirming specificity and demonstrating the feasibility of this application for routine use.
Biodistribution studies in tumor-xenografted mice were performed both to investigate potential diagnostic in vivo applications and to gain preliminary data to guide in the development of future therapeutic fusion constructs based on A33scFv. The latter approach has its limitations, as GFP as a fusion partner may behave quite differently in vivo than a prodrug-activating enzyme for a number of reasons. Comparing the predicted physical characterization of A33scFv::GFP with that of a potential therapeutic fusion protein, A33scFv::CD, however, we find that despite the difference in size (54 versus 44 kD, respectively), isoelectric point (6.04 versus 6.19) and amino acid composition are similar enough for feasibility testing.
The overall percentage of injected activity detected in tumor tissue was considerably lower than that of bivalent A33 IgG. As in surface plasmon resonance, this may reflect the lower avidity of the monovalent construct compared with divalent IgG. The fact that tumor became the tissue with the highest radioactivity during the elimination phase as well as the absence of any significant accumulation in the A33-negative control tumors or by the HP9scFv-based control protein, however, demonstrate that the accumulation in tumor tissue was due to specific binding rather than mere diffusion and passive attenuation.
A finding not well explained was the non-specific accumulation of radioactivity in other parenchymous organs, especially the kidneys and lungs. Although qualitatively, this finding corresponded to high kidney, lung and spleen uptake in previous investigations with divalent A33 antibody or conjugates thereof, the quantity of this effect was especially high with this monovalent construct. This is at least in part explained by dissociation of the conjugated iodine from the protein as the result of physiologic iodase activity. As a result of this effect, fusion protein presence in blood and blood-rich organs will be overestimated and that of tumor-localized fusion protein be underestimated by this sort of biodistribution assay. Whether in addition this is due to the scFv-construct's size and hence its diffusion characteristics or to the valency of the antigen-binding sites will be the subject of future studies.
Targeting of the A33 antigen has demonstrated high selectivity for colon cancer in clinical phase I/II studies (Welt et al., 1990), yet the clinical effect of the antibody alone was limited (Welt et al., 2003), probably both for insufficient induction of antibody-dependent cytotoxicity and for lack of a downstream signaling pathway triggered or inhibited by antibody binding to gpA33. Still, the targeting specificity of this antigen may be clinically exploited for diagnostic and therapeutic purposes if the antigen-binding function of the antibody can be charged with adequate effector molecules to exert a cytotoxic effect or, as in this case, diagnostic function. Thus, A33scFv::GFP and its production method described here will serve as a model for similarly designed A33-based fusion proteins with components such as prodrug-activating enzymes for antibody-targeted chemotherapy (Deckert et al., 2003).
|
Funding |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Deutsche Forschungsgemeinschaft (De 602/1-1); US Army Breast Cancer Research Program (DAMD17-99-1-9370) to P.M.D.
|
Footnotes |
|---|
Edited by Andrew Bradbury
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
The authors thank Drs Christoph Rader and Carlos F.Barbas III of the Scripps Institute, La Jolla, California, for providing the A33scFv plasmid and for their excellent advice in realizing this project. Sven Enders gave invaluable support with the plasmon surface resonance experiments.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionFundingAcknowledgementsReferences |
|---|
Adams G.P., Schier R. J. Immunol. Meth. (1999) 231:249–260.[CrossRef][Web of Science][Medline]
Cao Y., Suresh M.R. Bioconjug Chem. (1998) 9:635–644.[CrossRef][Web of Science][Medline]
Catimel B., et al. J. Biol. Chem. (1996) 271:25664–25670.
Catimel B., Nerrie M., Lee F.T., Scott A.M., Ritter G., Welt S., Old L.J., Burgess A.W., Nice E.C. J. Chromatogr. A (1997) 776:15–30.[CrossRef][Web of Science][Medline]
Cregg J.M., Cereghino J.L., Shi J.Y., Higgins D.R. Mol. Biotechnol. (2000) 16:23–52.[CrossRef][Web of Science][Medline]
Daly C.J., McGrath J.C. Pharmacol. Ther. (2003) 100:101–118.[CrossRef][Web of Science][Medline]
Damasceno L.M., Pla I., Chang H.J., Cohen L., Ritter G., Old L.J., Batt C.A. Protein Expr. Purif. (2004) 37:18–26.[CrossRef][Web of Science][Medline]
Deckert P.M., Jungbluth A., Montalto N., Clark M.A., Finn R.D., Williams J.C., Richards E.C., Panageas K.S., Old L.J., Welt S. Int. J. Cancer (2000) 87:382–390.[CrossRef][Web of Science][Medline]
Deckert P.M., Renner C., Cohen L.S., Jungbluth A., Ritter G., Bertino J.R., Old L.J., Welt S. Br. J. Cancer (2003) 88:937–939.[CrossRef][Web of Science][Medline]
Deckert P.M., Bornmann W.G., Ritter G., Williams C. Jr, Franke J., Keilholz U, Thiel E., Old L.J., Bertino J.R., Welt S. Int. J. Oncol. (2004) 24:1289–1295.[Web of Science][Medline]
Goshorn S., Sanderson J., Axworthy D., Lin Y.K., Hylarides M., Schultz J. Cancer Biother. Radiopharm. (2001) 16:109–123.[CrossRef][Web of Science][Medline]
Heath J.K., et al. Proc. Natl Acad. Sci. USA (1997) 94:469–474.
Jain R.K. Ann. Rev. Biomed. Eng. (1999) 1:241–263.[CrossRef][Web of Science][Medline]
Kriangkum J., Xu B.W., Nagata L.P., Fulton R.E., Suresh M.R. Biomol. Eng. (2001) 18:31–40.[CrossRef][Web of Science][Medline]
Lilie H., Schwarz E., Rudolph R. Curr. Opin. Biotechnol. (1998) 9:497–501.[CrossRef][Web of Science][Medline]
Lin M.Z., Teitell M.A., Schiller G.J. Clin. Cancer Res. (2005) 11:129–138.
Kortt A.A., Dolezal O., Power B.E., Hudson P.J. Biomol. Eng. (2001) 18:95–108.[CrossRef][Web of Science][Medline]
Macauley-Patrick S., Fazenda M.L., McNeil B., Harvey L.M. Yeast (2005) 22:249–270.[CrossRef][Web of Science][Medline]
Misawa S., Kumagai I. Biopolymers (1999) 51:297–307.[CrossRef][Web of Science][Medline]
Onda M., Wang Q.C., Guo H.F., Cheung N.K.V., Pastan I. Cancer Res. (2004) 64:1419–1424.
Rader C., Ritter G., Nathan S., Elia M., Gout I., Jungbluth A.A., Cohen L.S., Welt S., Old L.J., Barbas C.F. III. J. Biol. Chem. (2000) 275:13668–13676.
Ritter G., Cohen L.S., Williams C. Jr, Richards E.C., Old L.J., Welt S. Cancer Res. (2001) 61:6851–6859.
Verma R., Boleti E., George A.J.T. J. Immunol. Meth. (1998) 216:165–181.[CrossRef][Web of Science][Medline]
Welt S., Divgi C.R., Real F.X., Yeh S.D., Garin-Chesa P., Finstad C.L., Sakamoto J., Cohen A., Sigurdson E.R., Kemeny N. J. Clin. Oncol. (1990) 8:1894–1906.[Abstract]
Wels W., Harwerth I.M., Zwickl M., Hardman N., Groner B., Hynes N.E. Biotechnology (NY) (1992) 10:1128–1132.[CrossRef]
Welt S., Ritter G., Williams C. Jr, Cohen L.S., John M., Jungbluth A., Richards E.A., Old L.J., Kemeny N.E. Clin. Cancer Res. (2003) 9:1338–1346.
Received February 28, 2007; revised July 3, 2007; accepted July 6, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||