PEDS Advance Access first published online on August 10, 2007
This version published online on August 21, 2007
Protein Engineering Design and Selection, doi:10.1093/protein/gzm040
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A fold-back single-chain diabody format enhances the bioactivity of an anti-monkey CD3 recombinant diphtheria toxin-based immunotoxin
1Section on Biophysical Chemistry, Laboratory of Molecular Biology, National Institute of Mental Health, Bldg 10, Room 3D46, 10 Center Drive, Bethesda, MD 20892-1216, USA 2Section of Transplantation Immunobiology, Department of Surgery, University of Alabama Medical Center, Birmingham, AL 35294, USA
4 To whom correspondence should be addressed. E-mail: davidn{at}mail.nih.gov
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
T-cell depleting anti-CD3 immunotoxins have utility in non-human primate models of transplantation tolerance and autoimmune disease therapy. We recently reported that an affinity matured single-chain (scFv) anti-monkey CD3 antibody, C207, had increased binding to T-cells and increased bioactivity in a diphtheria toxin (DT)-based biscFv immunotoxin compared with the parental antibody, FN18. However, FN18 scFvs and their mutant derivatives such as C207 did not exhibit robust bivalent character in the biscFv format. We now report that C207 in a diabody format exhibits a 7-fold increase in binding to T-cells over scFv (C207) indicating considerable divalent character for the diabody. This construct was formed by reducing the VL/VH linker to five residues and was secreted from Pichia pastoris as the non-covalent dimer. An immunotoxin based on this diabody format was secreted as a non-covalent dimer but was devoid of bioactivity and failed to bind T-cells, suggesting steric hindrance from the two large closely positioned truncated DT moieties. We constructed a single-chain diabody immunotoxin by fusing to the truncated DT C-terminus L1-VL-L1-VH-L2-VL-L1-VH where L1 is a five-residue linker and L2 is the longer (G4S)3 linker permitting interactions between the distal and proximal VL/VH domains. This ‘fold-back’ immunotoxin was secreted predominantly as the monomer and exhibited a 5- to 7-fold increase in bioactivity over DT390biscFv(C207) and depleted monkey T-cells in vivo.
Keywords: CD3/diabody/FN18/immunotoxin/monkey
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
Anti-CD3 immunotoxins induce profound but transient T-cell depletion in vivo by inhibiting eukaryotic protein synthesis, and they have utility in non-human primate models of transplantation tolerance and autoimmune disease therapy (Neville et al., 1992
; Fechner et al., 1997; Hu et al., 1997; Thomas et al., 1997; Contreras et al., 2003). The conjugated immunotoxins were formed by chemically cross-linking the antibody or antibody fragment to a diphtheria toxin (DT) binding site mutant, CRM9. In this way, the binding site of the immunotoxin was dictated by the antibody moiety (Neville et al., 1996). The disadvantages of these chemically conjugated immunotoxins are linkage heterogeneity, Fc receptor interactions and a mutated toxin-binding domain in which binding to DT receptors is reduced but not eliminated. These disadvantages were overcome by developing recombinant immunotoxins based on DT truncated at residue 390 which provided optimal translocation of the toxin A chain into the cytosol while eliminating the toxin-binding domain (Williams et al., 1990; Thompson et al., 2001). These DT-based immunotoxins were restricted to having the antibody moiety placed C-terminal to the truncated toxin because antibody domains fused to the N-terminal of the toxin interfered with translocation of biologically active A chain (Madshus et al., 1992; Hexham et al., 2001). Although the affinity of anti-CD3 scFv was reduced to 3% of its initial value when tethered to the C-terminus of DT390, this loss was partially compensated for by forming a biscFv construct by adding a second scFv through a (G4S)3 linker (Thompson et al., 2001). The beneficial effect of the second scFv was ascribed to the divalent character of the biscFv binding, which in the case of the anti-human CD3 antibody UCHT1 was increased 5-fold for the biscFv binding over the scFv binding toward human T cells.
We recently reported the construction of an anti-monkey CD3 immunotoxin utilizing a biscFv format of an affinity-matured mutant of antibody FN18 known as C207 (Wang et al., 2007). Anti-monkey CD3 immunotoxins are important tools in organ transplantation research where the primary goal is to achieve tolerance to allografted organ transplants. Rodents are easily tolerized and are not a good model for human transplantation tolerance protocols compared with monkeys (Fechner et al., 1997; Thomas et al., 1997; Contreras et al., 2003). The FN18 scFv appeared more sensitive than anti-human UCHT1 to positional loss of affinity when placed distal to the DT390 toxin domains compared with the UCHT1 scFv. The divalent character of the C207 biscFv was judged to be suboptimal as determined by the relatively low increase in binding to monkey T-cells of the biscFv relative to the scFv which was only 1.7 ± 0.4 fold, compared with 5-fold for the UCHT1 biscFv/scFv binding ratio. The weaker binding of C207 biscFv could occur through suboptimal associations between the two sets of VL/VH domains within the biscFv. If this were the case, we reasoned that a more rigid structure that forced optimal VL/VH pairing should produce a higher affinity immunotoxin. The current work explores the use of diabody formats to increase the binding of anti-monkey immunotoxins based on the C207 mutant. Diabodies are recombinant non-covalent dimeric molecules with cross-over pairing of the VL and VH domains achieved by linking the VL and VH domains with a peptide linker that is too short to allow the interchain assembly of a functional Fv fragment (Holliger et al., 1993). If the linked VL and VH domains are derived from the same Fv, the diabody will be bivalent. If the linked VL and VH domains are derived from two different Fv structures, the diabody will be bispecific (Le Gall et al., 2004). These structures spontaneously form non-covalent dimers and are stabilized by the presence of four simultaneous variable domain interactions. Single-chain monomeric diabodies have also been generated by linking two diabodies through a longer linker that allows the diabody subunits to fold back head to tail (Kontermann and Muller, 1999).
The initial exploration of anti-monkey CD3 immunotoxin efficacy was done using the continuous cultured monkey T-line HSC-F (Akari et al., 1996). Specific cytotoxicity was assayed by the inhibition of protein synthesis. Equilibrium binding studies were also performed in order to determine the correlation between cytotoxicity and immunotoxin T-cell binding. Previous studies on the equilibrium binding of anti-CD3 antibodies to T-cells and the kinetics of dissociation had revealed a multicomponent binding process consisting of a combination of monovalent and bivalent binding and two equilibrium constants for both monovalent and bivalent binding differing by 10-fold. These were associated with two different dissociation rates also differing by 10-fold (Marano et al., 1989). In order to simplify binding comparisons between antibody fragments and immunotoxins in different formats, we elected to perform equilibrium binding competition studies competing the same labeled tracer, FITC-FN18, and monitor relative binding by the ratios of concentrations at equal tracer displacement. We also assayed cytotoxicity in resting monkey T-cells between different formats. Because the ultimate goal was to have a relevant pre-clinical animal model for the anti-human immunotoxins, we tested efficacy in monkeys by assaying T-cell depletion in the blood and lymphoid compartments. These positive in vivo tests showed that serum clearance was not a limiting factor for in vivo T-cell depletion for these new recombinant immunotoxins.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
Plasmids, bacterial and yeast strains, antibodies, cell lines
Plasmid pPICZ
was purchased from Invitrogen. The hybridoma secreting FN18 was kindly provided by Dr Margreet Jonker, Biomedical Primate Research Center Rijswijk, and was produced and purified by the National Cell Culture Center, Minneapolis, MN, USA. Chemically conjugated anti-monkey CD3 immunotoxin, FN18–CRM9, was prepared as previously described (Neville et al., 1989). The Herpes Saimiri virus transformed cynomolgous cell T-cell line HSC-F (Akari et al., 1996) was supplied by the Centralized Facility for AIDS Reagents supported by EU Programme EVA and the UK Medical Research Council. This cell line exhibits the moderate binding FN18 phenotype (Liu et al., 2007).
Protein synthesis inhibition assay
This was performed by measuring the incorporation of 3H-labeled leucine into HSC-F cells after a 20 h exposure to immunotoxins at different concentrations. Means and standard deviations of six replicates were calculated and divided by the means of non-immunotoxin-treated controls x100 to give per cent control protein synthesis as previously described (Thompson et al., 2001).
T-cell binding assay of recombinant scFv antibodies and immunotoxins
FN18–FITC conjugate was purchased from Biosource and added to 250 000 HSC-F cells at 5 x 10–9 M in the presence of varying concentrations of scFv and immunotoxin binding competitors at 4°C for 30 min. The displacement of FN18–FITC was measured by FACS on a Beckman-Coulter Cytomics FC 500 instrument after washing the cells two times at 400g at 4°C and resuspending in 0.5 ml of 1 x PBS containing 2.5 µg of propidium iodide to exclude dead cells from further analysis. Mean channel fluorescent values, MCI, were corrected (MCF') by subtracting an appropriate FITC-isotope control from both the FN18–FITC tracer and FN18–FITC tracer plus competitor MCF values. The per cent inhibition was determined as [(MCF'FN18–FITC – MCF'FN18–FITC+COMPETITOR)/MCF'FN18–FITC] x 100. Relative binding affinity between any two competitors was determined by dividing their respective concentrations at equal per cent inhibition obtained from plots of per cent inhibition versus concentration as previously described (Thompson et al., 2001).
Immunotoxin-induced loss of CD3+CD4+ resting monkey T-cells
T-cells from normal monkeys were isolated from blood by two centrifugations through Lymphocyte Separation Medium from BioWhittaker following the manufacturer's directions. About 400 000 cells were washed and incubated in RPMI 1640 medium containing 25 mM HEPES (Gibco) with 10% fetal calf serum +2 g/l NaHCO3, adjusted to pH 7.4 + 0.1 mM non-essential amino acids +1.0 mM sodium pyruvate +50 µg/l of gentamicin sulfate with and without varying concentrations of immunotoxins for 72 h at 5% CO2, 37°C. Cells were washed and then stained for 30 min at 4°C with anti-CD3 (SP34-2-PE BD Pharmingen), anti-CD4-FITC (BD Pharmingen), anti-CD20-PE Cy7 (BD Biosciences) and the vital dye 7-ADD (BD Pharmingen) and analyzed on a Beckman-Coulter Cytomics FC 500 instrument counting 104 events. Cells were analyzed for uptake of the vital dye, the per cent of cells within the lymphocyte forward scatter/side scatter gate and the per cent of cells within the lymphocyte gate within the quadrant displaying the CD3+ and CD4+ epitopes by two-color FACS. The per cent of cells within the lymphocyte gate at each immunotoxin concentration was divided by per cent of cells within the lymphocyte gate for the 72 h no-treatment control (G1). The per cent of cells within the CD3+CD4+ quadrant at each immunotoxin concentration was divided by per cent of cells within the CD3+CD4+ quadrant for the 72 h no-treatment control (Q1). The per cent loss of CD3+CD4+ cells was calculated as G1 x Q1 x 100.
|
Expression and purification of biscFv (C207) and DT390-biscFv (C207) |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
C207 is an affinity matured anti-rhesus CD3 mutant antibody selected from a library of FN18 sFvs randomly mutagenized with nucleotide analogs, displayed on yeast and selected by sorting flow cytometry using dye-labeled monkey CD3
ectodomain heterodimer (Wang et al., 2007). Construction of biscFv (C207) and its production in Pichia pastoris has been described as has the construction of the DT-based immunotoxin utilizing biscFv(C207) (Wang et al., 2007) shown in Table I schematic E. In brief, the VL and VH domains of C207 were linked by a (G4S)3 linker to form the scFv and two scFvs were linked by a third (G4S)3 linker to form the biscFv (Table I, D).
|
Expression and purification of diabody (C207) and fold-back diabody (C207) in P. pastoris
For the construction of diabody form of a C207 (Table I, B), the (G4S)3 linker existing between VL and VH scFv(C207) (Table I, A) was replaced with a G4S linker. The scFv(FN18–C207) gene was amplified by PCR using C207 yeast display plasmid as a template DNA and primers GP-095 and WP-052. The primers GP-94, GP-95 and GP-96 are listed in Table II. (All of the remaining GP and WP primers are listed in Table 1 of Wang et al., 2007.) Sequencing of this vector, pGK1903, was performed using two universal primers, T4 and SP6, located on the TA cloning vector, pDrive (Qiagen). Then, this plasmid was used as a template for the PCR in order to construct Db(C207) (Table I, B) and scfbDB(C207) (Table I, C). As described in the previous paper (Wang et al., 2007), the original -factor signal sequence in the vector system was replaced with a killer toxin (kt) signal peptide in order to have a complete cleavage between the signal peptide sequence and the inserted protein sequences of interest. To obtain diabody (C207), the Db(C207) gene was PCR-amplified from the plasmid pGK1903 using the primers GP-102 and WP-052. These vectors were sequenced using two primers located in the pPICZ, 5'-AOX1 primer and 3'-AOX1 primer (Invitrogen). Expression and purification of diabody (C207) and fold-back diabody (C207) was performed using a protein L affinity resin as described previously (Wang et al., 2007).
|
Expression and purification of the diabody and fold-back diabody immunotoxins in P. pastoris
To make A-dmDT390-scfbDb(C207), fold-back diabody bivalent immunotoxin (Table I, G), the Db(C207) gene was PCR amplified from plasmid pGK1903 using two primer sets; the first amplification used primers GP-94 and WP-053 and the second used primers GP-90 and WP-052.
The first PCR product was digested with NcoI and BamHI and the second PCR product was digested with BamHI and EcoRI. Then the two scFv parts in the vector pGK2006 [A-dmDT390-biscFv(C207) in pwPICZ] were sequentially replaced with C207-Db counterparts using NcoI and BamHI for the first scFv and BamHI and EcoRI for the second scFv (Table I, G). The resulting plasmid (pGK2019) was transformed into P.pastoris JW107 and the transformants were selected on YPD plates containing zeocin (100 µg/ml). Because A-dmDT390-scfbDb(C207) expression in strain JW107 was too low to be purified from the induction medium, the expression cassette was moved from pGK2019 to the vector pPIC9K (Invitrogen) by SacI and EcoRI digestion, and then introduced into a new P.pastoris stain (Liu and Neville, unpublished data) that had a much higher expression level. This new strain, which will be described in a forthcoming publication, alleviates some of the negative effects of the unfolded protein response that is induced by the expression of certain heterologous proteins in P.pastoris (Liu et al., 2005).
To make the A-dmDT390-Db(C207) diabody bivalent immunotoxin (Table I, F), the Db(C207) gene from pGK1903 was cut out with NcoI and EcoRI digestion and the 753 bp fragment was purified by gel electrophoresis followed by elution from the gel and then ligated into pGK2019 which was treated with the same restriction enzymes. The resulting plasmid (pGK2017) was transformed into P.pastoris JW107 and transformants were selected as described earlier. The vectors for constructing the diabody and fold-back diabody immunotoxins were sequenced using 5'-AOX1 primer, 3'-AOX1 primer (Invitrogen) and two internal primers Reverse 1 and Forward 1 (Table II). Expression and purification of DT390-Db(C207) was performed as described previously for A-dmDT390-biscFv(C207) (Wang et al., 2007).
DT390-scfbDb(C207)-induced T-cell depletion in monkeys
A-dmDT390-scfbDb(C207) (Table I, G) was administered as an intravenous bolus in two dosing schedules: (A) doses given on day 0 and two doses at a total dose of 0.15 mg/kg (monkey CX3X) and (B) eight divided doses given over 4 days, 6 h apart for a total dose of 0.2 mg/kg (monkey 32972). Lymphocyte phenotype levels were monitored in blood and also in lymph nodes following maceration by flow cytometry on a Beckman-Coulter Cytomics FC 500 instrument counting 104 events. Staining reagents were: CD3+ (CD3 sp34 FITC, Pharmingen, Catalog #556611, clone SP34), CD4+ (CD4 PerCP-Cy5.5, Pharmingen, Catalog #552838, clone L200), CD8b+ (CD8beta ECD, Coulter, Catalog #6607123, clone 2st8.5h7) and CD20+ (CD20 PeCy7, BD/Pharmingen, Catalog #335793). Appropriate IgG isotype controls (isotype for sp34: IgG3 FITC, Southern Biotech, Catalog #0105-02, clone B10 and isotype for FN18: IgG1 FITC, Coulter, Catalog #IM0639U, clone 679.1Mc7) were subtracted from mean fluorescent values. Blood T- and B-cell values were calculated by multiplying the total lymphocyte count per millimeter3 by the per cent of T-cells or B-cells enumerated by flow cytometry. When the total lymphocyte count was not performed on a cytometry day, the count obtained on the previous day was used.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
Purity of fusion protein constructs
Purified diabody immunotoxin A-dmDT390-Db(C207) (Table I, F) ran as a single band by SDS–PAGE coincident with A-dmDT390-scFv(C207) (Table I, C) (Fig. 1) and gave a major peak on Superdex-200 10/300 GL (Amersham Biosciences) corresponding to the dimeric protein and a small trailing shoulder corresponding to the monomer representing 8.3% of the applied protein in 276 nm absorbance area units. On HPLC size exclusion chromatography on a calibrated Superdex-200 10/300 GL (Amersham Biosciences), 85.7% of the applied purified fold-back immunotoxin A-dmDT390scfbDb(C207) (Table I, G) eluted coincident with DT390-biscFv (96 500 Da) while 9.7% of the protein eluted corresponding to a molecular weight of 210 000. An additional band, 4.6% eluted at the column front. Since these higher molecular weight species were not seen on the SDS gel, it is likely that they correspond to head to tail non-covalent dimers and higher oligomers. The purified diabody Db(C207) (Table I, B) and the fold-back diabody scfbDb(C207) (Table I, C), both ran as single bands on SDS gels (data not shown). Their molecular sizes on HPLC are described later.
|
The molecular sizes of diabody and fold-back diabody fusion proteins
On the basis of the literature (Holliger et al., 1993), we expected that the size of the diabody (Table I, B) and the diabody immunotoxin (Table I, F) would correspond to a dimeric protein mediated through molecular associations between the VH and VL domains on neighboring molecules since intermolecular self-associations were prevented by the short G4S linker separating these domains. This is shown to be the case in Fig. 2 where the Superdex 200 size exclusion column elution times of these constructs are plotted versus log molecular weight. The column was calibrated using BSA monomers and oligomers (solid line) and with scFv (M20) and A-dmDT390bisFv(M20) (dashed line). The point for Db(C207) (Table I, B), solid triangle, falls on the dashed line when given the calculated molecular weight of the dimer, 53.7 kDa. The point for the diabody immunotoxin (Table I, F), inverted solid triangle, falls on a linear extension of the dashed line when given the calculated molecular weight of the dimer, 141.1 kDa. In contrast, the fold-back diabody (Table I, C) and the corresponding fold-back diabody immunotoxin with the longer inter-domain (G4S)3 linkers (Table I, G) fall on or near the dashed line when given their respective monomer molecular weight values.
|
Relative binding of scFv, diabody and single-chain fold-back diabody constructs
The relative binding of these constructs to the monkey T-cell line HSC-F is estimated from the inhibition of the anti-monkey anti-T-cell antibody FN18 labeled with FITC by the constructs in a competition assay. The ratio of the construct concentrations at equal levels of inhibition reflects the relative avidity of the constructs and is shown in Fig. 3. The rank order of competitor binding from highest to lowest is scfbDb(C207) (Table I, C), Db(C207) (Table I, B), biscFv(C207) (Table I, D) and scFv(C207) (Table I, A). FN18 competition for FITC-labeled FN18 is shown for a comparative standard. In replicate assays, the ratio of concentrations at equal displacement for Db(C207)/scFv(C207) is 7 ± 2 and the ratio for biscFv(C207)/scFv(C207) is 2.5 ± 0.1 (errors are SD).
|
|
Relative binding of immunotoxin constructs made with scFv, diabody and a single-chain fold-back diabody |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
Although the diabody of C207 (Table I, B) has high T-cell binding, the immunotoxin made in the diabody format (Table I, F) fails to show demonstrable binding judged by lack of inhibition FN18–FITC binding (solid squares, Fig. 4). In contrast, the immunotoxin made in the fold-back diabody format (Table I, G) exhibits robust binding, solid circles, 10-fold greater than that made in the biscFv format (Table I, E), empty circles.
|
Anti-T-cell activity of immunotoxin made in the fold-back diabody format
The anti-T-cell activity of immunotoxin made in the fold-back diabody format was judged by two in vitro assays and by its effects in vivo. In Fig. 5, the per cent control inhibition of protein synthesis for three immunotoxins on the transformed HSC-F monkey T-line is shown. The single-chain fold-back diabody recombinant immunotoxin, A-dmDT390scfbDb(C207) (Table I, G), solid diamonds, is the most potent immunotoxin, exceeding the immunotoxin constructed with the biscFv antibody moiety, A-dmDT390-biscFv(C207) (Table I, E) by 5-fold (empty squares). Similar results are shown in Fig. 6 after a 72 h exposure of immunotoxins to cultured resting monkey T-cells isolated from blood. In Fig. 7, the loss of lymph node T-cells over time enumerated by flow cytometry following bolus intravenous administration of the fold-back diabody immunotoxin is shown for two monkeys. Monkey 32972 received the immunotoxin twice daily, 6 h apart, over 4 days, for a total dose of 0.2 mg/kg. Monkey CX3X received two intravenous bolus doses of immunotoxin, one on day 0 and the other on day 2 for a total dose of 0.15 mg/kg. For a comparison, a third monkey AH44 received another biscFv immunotoxin, A-dmDT390biscFv(M20), in the same dose and schedule as monkey CX3X. This immunotoxin was described previously (Wang et al., 2007) and is 13.5-fold less active against resting T-cells compared with A-dmDT390-biscFv(C207). Lymph nodes were not noticeably deleted of T-cells 4 days after the initial immunotoxin dose. The profound loss of T-cells in the blood for monkey 32972 is shown in Fig. 8. B-cells are only moderately affected. T-cells were present in the blood 1 month later but the per cent of SP34+ T-cells in the lymphocyte gate was reduced from the pre-treatment value of 66% to 18.7% whereas the per cent of CD20+ B-cells rose from 19.2 to 35.3 as did the percentage of CD2+ SP34– cells (probable NK cells).
|
|
|
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
This work shows that a fold-back diabody format for an anti-monkey CD3
immunotoxin (Table I, G) offers improvements in binding and bioactivity toward resting T-cells and transformed T-cell lines ranging from 4- to 10-fold compared with the biscFv format for the affinity matured mutant C207 (Table I, E). This is a substantial improvement for a therapeutic agent and is consistent with the fact that the fold-back diabody format achieves profound in vivo depletion of T-cells in the blood and lymph node compartments when the fold-back immunotoxin is administered twice daily over 4 days (Figs 7 and 8). This is the first report, to our knowledge, of using a fold-back divalent diabody format in an immunotoxin. Krauss et al. (2005) successfully used a diabody format in an RNase-based immunotoxin targeted at B-cells using the CD22 epitope. However, our results showed that the dimeric diabody format was inactive using the larger DT390-based toxin moiety, probably due to steric hindrance. We therefore tried the fold-back diabody format, originally described by Kontermann and Muller (1999). In this format, two identical diabodies are linked through a longer linker that allows the diabody subunits to fold back head to tail.
There are a number of papers that compare the increase in functional affinity (avidity) on going from a scFv format to a divalent diabody format. The increases are varied ranging from 2.8-fold (Nielsen et al., 2000) to 100-fold (Huang et al., 2006). In general, the increased functional affinity in the diabody format is the result of a decrease in the equilibrium dissociation constant (KD) and is due to bivalent binding as demonstrated by methods that independently can measure monovalent and bivalent KD values. Under most conditions of diabody binding to a protein antigen, there exist two measurable KDs, one for monovalent and the other for bivalent binding, and the overall avidity is determined by the fraction of bivalent binding (Huang et al., 2006). Because of steric constraints in diabody-protein bivalent binding, the fraction of multivalent binding never approaches the high values seen in some low molecular weight systems such as tris-vancomycin binding a trivalent D-Ala-D-Ala ligand (Rao et al., 1998). In this case, the trivalent KD (4 x 10–17 M) was approximately equal to the third power of the monovalent KD (1.6 x 10–6 M).
The novel structure of divalent diabodies, enhanced rigidity and a close to 180° angle between the two antigen sites, originally deduced from model building (Holliger et al., 1993) has been documented by X-ray crystallographic studies (Perisic et al., 1994). It is likely that the fold-back divalent diabody shares these features and that they contribute to enhanced improvements over the biscFv format. However, these structural studies are not yet available. The simultaneous and concerted interaction of all four domains undoubtedly stabilizes the structure and provides extra-thermal stability compared with the biscFv format where only two concerted domain interactions occur for each part of the biscFv. The rigid 180° angle between the two antigen sites may permit divalent interactions between monovalently bound diabody-immunotoxin on neighboring cells in closely packed tissue such as lymph nodes, thus expanding the fraction of divalent binding sites beyond that limited on a single cell surface. This would decrease the overall kinetic off-rate and could enhance signaling facilitating endocytosis of immunotoxin. It is likely that the degree of improvement in vivo bioactivity offered by fold-back diabody immunotoxins will vary on a case-by-case basis according to which of the variables outlined above are limiting.
|
Footnotes |
|---|
3 Present address: Department of Animal Science and Technology, Chung-Ang University, Gyeonggi-Do 456-756, Korea
The originally published version of this paper was incorrect. On page 4, line 462 "bi" should appear before "scFv(C207)" so that it reads "A-dmDT390-biscFv(C207)" and on line 462 "Table I, C" should be "Table I, E"
|
Acknowledgements |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
This research was supported by the Intramural Research Program of NIMH, NIH. We thank Jim Nagle and the NINDS/NIH DNA Sequencing Facility for their generous support. K.J.G. and J.M.T. were supported in part by NIDDK 5 U19.DK57958-07.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsExpression and purification of...ResultsRelative binding of immunotoxin...DiscussionAcknowledgementsReferences |
|---|
Akari H., Mori K., Terao K., Otani I., Fukasawa M., Mukai R., Yoshikawa Y. Virology (1996) 218:382–388.[CrossRef][Web of Science][Medline]
Contreras J.L., Jenkins S., Eckhoff D.E., Hubbard W.J., Lobashevsky A., Bilbao G., Thomas F.T., Neville D.M. Jr., Thomas J.M. Am. J. Transplant. (2003) 3:128–138.[CrossRef][Web of Science][Medline]
Fechner J.H. Jr, Vargo D.J., Geissler E.K., Graeb C., Wang J., Hanaway M.J., Watkins D.J., Piekarczyk M., Neville D.M. Jr, Knechtle S.J. Transplantation (1997) 63:1339–1345.[CrossRef][Web of Science][Medline]
Hexham J.M., Dudas D., Hugo R., Thompson J., King V., Dowling C., Neville D.M. Jr, Digan M.E., Lake P. Mol. Immunol. (2001) 38:397–408.[CrossRef][Web of Science][Medline]
Holliger P., Prospero T., Winter G. Proc. Natl Acad. Sci. USA (1993) 90:6444–6448.
Hu H., Stavrou S., Baker B.C., Tornatore C., Scharff J., Okunieff P., Neville D.M. Jr. Cell. Immunol. (1997) 177:26–34.[CrossRef][Web of Science][Medline]
Huang B.C., Davern S., Kennel S.J. J. Immunol. Meth. (2006) 313:149–160.[CrossRef][Web of Science][Medline]
Kontermann R.E., Muller R. J. Immunol. Meth. (1999) 226:179–188.[CrossRef][Web of Science][Medline]
Krauss J., Arndt M.A.E., Vu B.K., Newton D.L., Seeber S., Rybak S.M. Biochem. Biophys. Res. Commun. (2005) 331:595–602.[CrossRef][Web of Science][Medline]
Le Gall F., Reusch U., Little M., Kipriyanov S.M. Protein Eng. Des. Sel. (2004) 17:357–366.
Liu Y.Y., Woo J.H., Neville D.M. Jr. Appl. Environ. Microbiol. (2005) 71:5332–5340.
Liu Y.Y., Wang Z., Thomas J., Goodwin K.J., Stavrou S., Neville D.M. Jr. Immunol. Cell Biol. (2007) 85:357–362.[CrossRef]
Madshus I.H., Olsnes S., Stenmark H. Infect. Immun. (1992) 60:3296–3302.
Marano N., Holowka D., Baird B. J. Immunol. (1989) 143:931–938.[Abstract]
Neville D.M. Jr, Srinivasachar K., Stone R., Scharff J. J. Biol. Chem. (1989) 264:14653–14661.
Neville D.M. Jr., Scharff J., Srinivasachar K. Proc. Natl Acad. Sci. USA (1992) 89:2585–2589.
Neville D.M. Jr, Scharff J., Rigaut K., Hu H., Shiloach J., Slingerland W., Jonker M. J. Immunother. Emphasis Tumor Immunol. (1996) 19:85–92.[Medline]
Nielsen U.B., Adams G.P., Weiner L.M., Marks J.D. Cancer Res. (2000) 60:6434–6440.
Perisic O., Webb P.A., Holliger P., Winter G., Williams R.L. Structure (1994) 2:1217–1226.[Medline]
Rao J., Lahiri J., Isaacs L., Weis R.M., Whitesides G.M. Science (1998) 280:708–711.
Thomas J.M., et al. Transplantation (1997) 64:124–135.[Web of Science][Medline]
Thompson J., et al. Protein Eng. (2001) 14:1035–1041.
Wang Z., Kim G.B., Woo J.H., Liu Y.Y., Mathias A., Stavrou S., Neville D.M. Jr. Bioconjug. Chem. (2007) 18:947–955.[CrossRef][Web of Science][Medline]
Williams D.P., Snider C.E., Strom T.B., Murphy J.R. J. Biol. Chem. (1990) 265:11885–11889.
Received April 17, 2007; revised June 12, 2007; accepted June 28, 2007.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||