A method for preparing chelate-cytokine conjugates with retention of protein structure, biological activity, and pharmacokinetic properties.

The chelating agent diethylenetriaminepentaacetic acid (DTPA) has been conjugated site-specifically to the N-terminus of recombinant human interleukin-2 (rhIL-2) by reaction with DTPA dianhydride at an initial pH of 6.0, thus demonstrating broader application of the conjugation method previously described for the structurally related cytokine rhG-CSF (Ralph et al., 1995). Purity of the DTPA-rhIL-2 conjugate, isolated by cation-exchange FPLC, and chelation of 111In were revealed by cation-exchange HPLC. Purity of the conjugate as well as chelation of radiometal were also demonstrated by SDS-PAGE and TLC, respectively. The stoichiometric molar ratio of DTPA to protein for the conjugate was approximately 1:1 as determined by TLC and mass spectrometry. Localization of the DTPA moiety was resolved by a peptide mapping procedure. The protein retained > 95% secondary structure (alpha helicity) following the conjugation. Addition of metal induced an approximate 22% loss of secondary structure for the conjugate. The in vitro biological activity of the protein was unaffected by the conjugated DTPA, even with chelated metal. Pharmacokinetic analysis of DTPA-conjugated cytokines, following chelated 111In, showed clearance and pharmacokinetic parameter values comparable to those of the corresponding unmodified cytokine. DTPA-conjugated cytokines may prove useful in cytokine research, and furthermore may represent a novel class of molecules for imaging, diagnosing, and/or treatment of malignancies where the cytokine receptor is overexpressed.

Stability of recombinant consensus interferon to air-jet and ultrasonic nebulization.

The stability of recombinant consensus alpha-interferon (rConIFN) to air-jet and ultrasonic nebulization was evaluated. Volumes of 10 mL of 0.5 mg/mL rConIFN in phosphate-buffered saline (PBS) at pH 6.3 were nebulized with a Collison three-jet nebulizer at 40 psig (10 L/min) for up to 25 min. The effects of pH (3.0, 6.3, and 9.0), additive (0.1% w/v Tween 80, 0.1% w/v Tween 20, and 1% w/v PEG 8000), and ionic strength (0, 0.25, and 1.0) were examined. The effects of ultrasonic nebulization were studied using three devices (DeVilbiss "Aerosonic"; Mountain Medical "Microstat", and Medix "Easimist"). Stability of rConIFN was assessed by size exclusion chromatography and native and sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis (PAGE). Structural effects were examined by circular dichroism spectroscopy and bioactivity was assessed by an in vitro mitogenic inhibition bioassay. rConIFN is destabilized by air-jet nebulization. Insoluble noncovalent aggregates are produced rapidly, and only approximately 25% of the initial monomeric protein remains after 25 min of nebulization. This correlates with a decrease in in vitro bioactivity. Aggregation during nebulization is influenced by pH (9.0 < 6.3 < 3.0) but even at the highest pH, > 25% aggregation is observed. Ionic strength does not appear to influence aggregation. rConIFN is also seen to adhere to glass after nebulization. Samples from a rinse of the emptied reservoir with 0.1% w/v SDS, after thorough rinsing with water (three times), show a strong rConIFN band on SDS-PAGE gels. The use of PEG 8000 and Tween mitigate aggregate formation and adhesion (< 20%). The cumulative output collected as a wet or dry aerosol is not aggregated to the same extent as the residual protein remaining in the nebulizer. Ultrasonic nebulization also results in aggregation, but the extent of denaturation is dependent upon the nebulizer used and is related to the heating of nebulizer solutions. Cooling of the nebulizer solution during operation (< 30 degrees C) minimizes aggregation (< 5%), and bioactivity is retained.

Site-specific conjugation of diethylenetriaminepentaacetic acid to recombinant human granulocyte-colony-stimulating factor: preservation of protein structure and function.

The chelating agent diethylenetriaminepentaacetic acid (DTPA) was conjugated site-specifically to the N-terminus of recombinant human granulocyte-colony-stimulating factor (rhG-CSF) by reaction of the protein with DTPA dianhydride at an initial pH of 6.0. The reaction was efficient in that 84% of the starting rhG-CSF was N-terminally modified and could be purified to homogeneity by cation-exchange chromatography. Chelation of 111In by the DTPA-rhG-CSF conjugate was demonstrated by cation-exchange HPLC and thin-layer chromatography. Metal contamination of conjugate preparations, as well as metal-loading onto the conjugate, could be monitored by either cation-exchange HPLC or isoelectric focusing. The 1:1 stoichiometric molar ratio of DTPA to protein for the DTPA-rhG-CSF conjugate was determined by thin-layer chromatography and mass spectrometry, and the localization of the conjugated DTPA moiety was resolved using a peptide mapping procedure. The secondary structure (i.e., alpha-helicity) of the protein was unmodified following conjugation as revealed by circular dichroism. Furthermore, the conjugate induced a similar induction of peripheral WBC counts as unmodified rhG-CSF when injected subcutaneously into hamsters, demonstrating preservation of protein bioactivity. These results reveal a simple and efficient method for conjugating DTPA to protein, via reaction with the dianhydride, to yield a homogeneous and well-defined product. The procedure may prove to be a useful method of labeling growth factors and related proteins while preserving structural and functional integrity.

Pharmacokinetics of pretargeted monoclonal antibody 2D12.5 and 88Y-Janus-2-(p-nitrobenzyl)-1,4,7,10-tetraazacyclododecanetetraacetic acid (DOTA) in BALB/c mice with KHJJ mouse adenocarcinoma: a model for 90Y radioimmunotherapy.

Three-step pretargeting for radioimmunotherapy in BALB/c mice with KHJJ tumors was done with monoclonal antibody (mAb) 2D12.5, which is specific for yttrium-1,4,7,10-tetraazacyclododecanetetraacetic acid (DOTA) but nonspecific for the tumor. Tumor uptake was by passive diffusion of mAb through leaky neovasculature in the tumor. The three steps were: (a) anti-hapten mAb 2D12.5 (0 h); (b) polyvalent haptenprotein conjugate chase (20 h); and (c) 88Y-labeled monovalent DOTA or bivalent Janus-DOTA haptens (21 h) and organ and tumor bioassay (24 h). Rapid tumor (T) uptake and high tumor:blood ratio (T:BL) was seen 3 h after injection after step c. For monovalent 88Y-DOTA, T = 1.7%/g* and T:BL = 16:1; for bivalent 88Y-Janus-DOTA, T = 4.41%/g* and T:BL = 21:1 at 3 h (*, P < 0.001). Blood and bone plus marrow were << 1%/g, and liver was < 1%/g. The 24-h whole body retention was approximately 5% of injected dose with 1% in tumor (20% of total), 1.8% in other organs, and 2.2% in carcass; the 24-h whole body retention of covalent nonspecific antibody conjugates was > 80% of injected dose. The biological half-life in the tumor of 0.9 microCi 88Y-Janus-DOTA was approximately 24 h, measured daily for 5 days. Activity in microCi/g of tumor and blood for 90Y equimolar to the amount of 88Y injected (0.9 microCi 88Y = 0.744 pmol = 36.47 microCi 90Y) was used for calculating the area under the curve of tumor and blood in microCi-h/g of 90Y. The 90Y radiation absorbed dose (RAD) from multiplying microCi-h/g x the 90Y absorbed dose constant, 1.99 RAD-g/microCi-h, gave T = 89 RAD and BL = 3.7 RAD. The therapeutic ratio from RAD T:RAD BL = 24:1. These results indicate that pretargeting 90Y hapten-specific mAb for radioimmunotherapy has considerable promise.

Immunogenicity in rabbits and mice of an antibody-chelate conjugate: comparison of (S) and (R) macrocyclic enantiomers and an acyclic chelating agent.

The macrocyclic bifunctional chelating agent 2-(p-bromoacetamidobenzyl)-1,4,7,10-tetraazacyclododecanetetraa cetic acid (BAD), forms inert metal complexes ideal for radioimmunotherapy. Kosmas et al. (Cancer Res., 52: 904-911, 1992) found 2-imminothiolane linker-(S)-BAD monoclonal antibody HMFG1 highly immunogenic in patients. We studied the immunogenicity of (S) and (R) enantiomers of 2-imminothiolane linker-BAD rabbit IgG, monoclonal antibody Lym-1, and Lym-1 2-imminothiolane linker-(S)-bromoacetamidobenzyl-EDTA in 15 rabbits. Five groups of three each were given 0.1, 1.0, or 10 mg of 111In conjugate i.v., blood samples were taken daily for 14 days and biweekly for 70 days, and the plasma T1/2 was calculated. A drop in plasma 111In at 6-8 days coincided with the appearance of antibody on enzyme-linked immunosorbent assay. Specific anti-(S)-BAD, anti-(R)-BAD, anti-(S)-bromoacetamidobenzyl-EDTA, and anti-mouse IgG were measured. Rabbit IgG conjugates did not elicit an immune response. Mouse IgG conjugates were immunogenic on the first exposure, with both anti-1,4,7,10-tetraazacylododecane N,N',N'',N'''-tetraacetic acid and anti-mouse responses. Anti-1,4,7,10-tetraazacylododecane N,N',N'',N'''-tetraacetic acid was specific for the (S) or (R) enantiomer, but cross-reaction appeared with reboosting. A second injection of the opposite enantiomer gave a response to that enantiomer. Lym-1 bromoacetamidobenzyl-EDTA produced anti-bromoacetamidobenzyl-EDTA and anti-mouse response.

Viability and biodistribution of 68Ga MPO-labelled human platelets.

The viability and biodistribution of 68Ga-mercaptopyridine-N-oxide (MPO)-labelled autologous platelets was studied in 10 patients. The average platelet labelling yield was 36 +/- 12% and injected activity was 2.0 +/- 0.9 mCi 68Ga. The % activity in platelets per ml whole blood was 64 +/- 20% at 15 min-1.0 h postinjection and 76 +/- 14% at 2-4h. The average recovery of platelets (% injected platelets circulating in peripheral blood) was 31 +/- 21% at 15 min-1 h and 39 +/- 20% at 2-4 h. The positron emission tomographic (PET) images showed high circulating vascular background. Two patients had technically inadequate scans, and six were false negative due to high blood background. One patient with a massive pulmonary embolus occurring 24 h prior to scanning had marked uptake of 68Ga platelets in a large clot in the superior branch of the right main pulmonary artery. A second patient, with 68Ga platelets circulating during angioplasty of a left posterior tibial artery stenosis, had intense uptake in the lesion shown on the PET scan obtained 4 h following the procedure. These results indicate good viability of 68Ga-MPO-labelled autologous human platelets, but poor visualization of clots by PET imaging, due to the high blood background at early times.