Rapid Generation of Murine Bispecific Antibodies Using FAST-IgTM for Preclinical Screening of HER2/CD3 T-Cell Engagers.

Bispecific antibodies (BsAbs) can bind to two different antigens, enabling therapeutic concepts that cannot be achieved with monoclonal antibodies. Immuno-competent mice are essential for validating drug discovery concepts, necessitating the development of surrogate mouse BsAbs. In this study, we explored the potential of FAST-IgTM, a previously reported BsAb technology, for mouse BsAb production. We investigated charge-based orthogonal Fab mutations to facilitate the correct assembly of heavy and light chains of mouse antibodies and employed knobs-into-holes mutations to facilitate the heterodimerization of heavy chains. We combined five anti-CD3 and two anti-HER2 antibodies in mouse IgG1 and IgG2a subclasses. These 20 BsAbs were analyzed using mass spectrometry or ion exchange chromatography to calculate the percentages of BsAbs with correct chain pairing (BsAb yields). Using FAST-Ig, 19 out of the 20 BsAbs demonstrated BsAb yields of 90% or higher after simple protein A purification from transiently expressed antibodies in Expi293F cells. Importantly, the mouse BsAbs maintained their fundamental physicochemical properties and affinity against each antigen. A Jurkat NFAT-luciferase reporter cell assay demonstrated the combined effects of epitope, affinity, and subclasses. Our findings highlight the potential of FAST-Ig technology for efficiently generating mouse BsAbs for preclinical studies.

Efficient production of bispecific antibody by FAST-IgTM and its application to NXT007 for the treatment of hemophilia A.

Efficient production of bispecific antibodies (BsAbs) in single mammalian cells is essential for basic research and industrial manufacturing. However, preventing unwanted pairing of heavy chains (HCs) and light chains (LCs) is a challenging task. To address this, we created an engineering technology for preferential cognate HC/LC and HC/HC paring called FAST-Ig (Four-chain Assembly by electrostatic Steering Technology - Immunoglobulin), and applied it to NXT007, a BsAb for the treatment of hemophilia A. We introduced charged amino-acid substitutions at the HC/LC interface to facilitate the proper assembly for manufacturing a standard IgG-type BsAb. We generated CH1/CL interface-engineered antibody variants that achieved > 95% correct HC/LC pairing efficiency with favorable pharmacological properties and developability. Among these, we selected a design (C3) that allowed us to separate the mis-paired species with an unintended pharmacological profile using ion-exchange chromatography. Crystal structure analysis demonstrated that the C3 design did not affect the overall structure of both Fabs. To determine the final design for HCs-heterodimerization, we compared the stability of charge-based and knobs into hole-based Fc formats in acidic conditions and selected the more stable charge-based format. FAST-Ig was also applicable to stable CHO cell lines for industrial production and demonstrated robust chain pairing with different subclasses of parent BsAbs. Thus, it can be applied to a wide variety of BsAbs both preclinically and clinically.

pH Dependence of the Number of Discrete Conformers of Carbonic Anhydrase 2, as Evaluated from Collision Cross-Section Using Ion Mobility Coupled with Electrospray Ionization.

Ion mobility experiments coupled with electrospray ionization (ESI) were conducted to evaluate the folding states of bovine carbonic anhydrase 2 (CA2) under three different pH conditions. Collision cross-section (CCS) of the CA2 ions generated by ESI revealed the presence of six discrete conformers in the gas phase under the conditions employed in this study. The CCS of the most extended conformer was three times larger than that of the most compact one. The charge state distribution of the CA2 ions was indicative of three conformers being present. Although there was consistency in conformer assignment conducted by CCS and charge state distribution, the CCS measurement was shown to be more effective because the information obtained provided more detailed knowledge of the conformation of the protein.

The pH Dependence of Product Ion Spectra Obtained from Precursor Ions with the Same Charge Number in ESI of Carbonic Anhydrase 2.

The effect of solvent conditions, including pH, on product ion spectra obtained from precursor ions produced by electrospray ionization (ESI) was examined. Bovine carbonic anhydrase 2 was used as a model protein and the product ions generated by collision induced dissociation of the whole protein were measured under several different solvent conditions (pH 5.0, 3.7, and 0.1% HCOOH (pH 2.6)/MeCN (1/1)). The product ion spectra from precursor ions with the same charge number, the observed m/z values and the relative intensities of the product ions were similar. It therefore appears that the solvent conditions used have no effect on the product ion that is generated. On the other hand, different profiles of the product ion were obtained from precursor ions having different charge numbers. This indicates that the charge number of the precursor ion appears to be a major determinant of the product ion species and its relative intensity in product ion spectra of proteins.

Identification and multidimensional optimization of an asymmetric bispecific IgG antibody mimicking the function of factor VIII cofactor activity.

In hemophilia A, routine prophylaxis with exogenous factor VIII (FVIII) requires frequent intravenous injections and can lead to the development of anti-FVIII alloantibodies (FVIII inhibitors). To overcome these drawbacks, we screened asymmetric bispecific IgG antibodies to factor IXa (FIXa) and factor X (FX), mimicking the FVIII cofactor function. Since the therapeutic potential of the lead bispecific antibody was marginal, FVIII-mimetic activity was improved by modifying its binding properties to FIXa and FX, and the pharmacokinetics was improved by engineering the charge properties of the variable region. Difficulties in manufacturing the bispecific antibody were overcome by identifying a common light chain for the anti-FIXa and anti-FX heavy chains through framework/complementarity determining region shuffling, and by pI engineering of the two heavy chains to facilitate ion exchange chromatographic purification of the bispecific antibody from the mixture of byproducts. Engineering to overcome low solubility and deamidation was also performed. The multidimensionally optimized bispecific antibody hBS910 exhibited potent FVIII-mimetic activity in human FVIII-deficient plasma, and had a half-life of 3 weeks and high subcutaneous bioavailability in cynomolgus monkeys. Importantly, the activity of hBS910 was not affected by FVIII inhibitors, while anti-hBS910 antibodies did not inhibit FVIII activity, allowing the use of hBS910 without considering the development or presence of FVIII inhibitors. Furthermore, hBS910 could be purified on a large manufacturing scale and formulated into a subcutaneously injectable liquid formulation for clinical use. These features of hBS910 enable routine prophylaxis by subcutaneous delivery at a long dosing interval without considering the development or presence of FVIII inhibitors. We expect that hBS910 (investigational drug name: ACE910) will provide significant benefit for severe hemophilia A patients.

Antibody recycling by engineered pH-dependent antigen binding improves the duration of antigen neutralization.

For many antibodies, each antigen-binding site binds to only one antigen molecule during the antibody's lifetime in plasma. To increase the number of cycles of antigen binding and lysosomal degradation, we engineered tocilizumab (Actemra), an antibody against the IL-6 receptor (IL-6R), to rapidly dissociate from IL-6R within the acidic environment of the endosome (pH 6.0) while maintaining its binding affinity to IL-6R in plasma (pH 7.4). Studies using normal mice and mice expressing human IL-6R suggested that this pH-dependent IL-6R dissociation within the acidic environment of the endosome resulted in lysosomal degradation of the previously bound IL-6R while releasing the free antibody back to the plasma to bind another IL-6R molecule. In cynomolgus monkeys, an antibody with pH-dependent antigen binding, but not an affinity-matured variant, significantly improved the pharmacokinetics and duration of C-reactive protein inhibition. Engineering pH dependency into the interactions of therapeutic antibodies with their targets may enable them to be delivered less frequently or at lower doses.

Ion mobility and collision-induced dissociation analysis of carbonic anhydrase 2.

Folding analysis of the zinc protein, carbonic anhydrase 2 (CA2), was performed using electrospray ionization ion mobility spectrometry coupled with collision-induced dissociation (ESI IMS/CID). Multiply protonated ions with a bimodal charge state distribution were observed indicating the presence of at least two folding states for gas-phase CA2 ions as was described in a previous study (Nabuchi, Y.; Murao, N.; Asoh, Y.; Takayama, M. Anal. Chem. 2007, 79, 8342-8349). In the IMS driftgram, several ions with different mobility were observed for each multiply charged ion, and this suggests that CA2 ions consist of several components with different folding states. IMS/CID spectra were acquired against precursor ions separated by mobility. The CID spectra gave several characteristic product ions including those from the N- and C-terminal region of CA2. A shift to larger charge number for the most abundant of the several product ions was observed for ions having a larger drift time. This charge number shift indicates that the folding state of the ion is more unfolded. Furthermore, differences in the production of an ion corresponding to the N-terminal side fragment gave information about the unfolding process of CA2.

Prediction of drug-drug interactions based on time-dependent inhibition from high throughput screening of cytochrome P450 3A4 inhibition.

A method of assessing the risk of drug-drug interaction (DDI) caused by mechanism-based inhibition (MBI) was developed for early-stage drug development using cytochrome P450 (CYP) 3A4 inhibition screening data. CYP3A4 inhibition was evaluated using a fluorescent substrate with or without preincubation containing an inhibitor. The results showed that five well-known mechanism-based inhibitors, but not the competitive inhibitor ketoconazole, had lower IC(50) after preincubation, suggesting the utility of the IC(50) shift by preincubation to discern mechanism-based inhibitors. A method to approximately predict the change in the area under the concentration-time curve (AUC) of a co-administered drug by MBI was found using IC(50) shift data and the unbound mean plasma concentration of the inhibitor. From our predictions of change in the AUC for 38 drugs using this method, all mechanism-based inhibitors causing change in the AUC of more than 200% were predicted to be high risk. In conclusion, our method provides a simple assessment of the risk of DDI from mechanism-based inhibitors, especially in the early stages of drug development.

Estimation of serum protein binding of compounds metabolized in serum using matrix inhibition.

It is difficult to evaluate the serum protein binding of compounds that are metabolized in rat serum, even when using ultrafiltration. Protein binding was estimated using matrix inhibition, a method that uses the change in metabolic velocity achieved by changing the free fraction of a compound in the incubation mixture by diluting the serum with phosphate buffered saline. The T(1/2) of phenyl nicotinate, benzyl nicotinate, octyl nicotinate, hexyl nicotinate, butyl nicotinate and [(3)H] compound A were 0.165, 0.780, 2.62, 3.94, 5.22 and 135 min, respectively, with protein binding values of 82.1%, 91.6%, 98.8%, 98.5%, 85.5% and 96.9%. The protein binding value of compound A estimated by ultrafiltration was 93.4%, indicating that the two methods give similar values. The matrix inhibition method is thus applicable for the evaluation of compounds metabolized in serum, and provides a simple, useful method to determine protein binding.

New method for the simultaneous estimation of intrinsic hepatic clearance and protein binding by matrix inhibition.

The purpose of this study was to develop a method for estimating the hepatic clearance (CL(h)) without using a protein binding test. This method allows the simultaneous evaluation of the intrinsic hepatic clearance (CL(int)) with a correction for microsomal binding, and the free fraction in the serum (fu). It uses the decrease in metabolic velocity achieved by decreasing the free fraction of a compound in the incubation mixture (fu(inc)) by the addition of serum, and by changing the microsomal protein concentration. This method is denoted as the 'matrix inhibition method', because it uses the inhibition of the metabolic velocity by the incubation matrix. The metabolic rates of eight compounds (diazepam, imipramine, warfarin, and compounds A-E) were evaluated under several incubation conditions using rat serum and microsomes. The correlation of CL(int) evaluated using the method and using equilibrium dialysis after the CL(int) was corrected for microsomal binding was r = 0.968. The correlation of fu . CL(int) was r = 0.996. Although the method required a high enough fu and fu(microsomes) difference among the reaction conditions for each compound, it could evaluate CL(int) and fu simultaneously and easily by adding additional reaction conditions to the metabolic stability tests performed in ADME screening.

Probing the unfolding and refolding processes of carbonic anhydrase 2 using electrospray ionization mass spectrometry combined with pH jump.

A novel method for proving the time course of the unfolding and refolding processes of metalloprotein bovine carbonic anhydrase 2 (CA2) is demonstrated using electrospray ionization mass spectrometry (ESI MS) combined with pH jumps between 3.6 and 4.4. The shift in mass accompanied by the release or coordination of a zinc ion and the change in the charge state distribution were measured to evaluate the folding process. The time course of the ESI mass spectra revealed the existence of four types of ions in the experimental system, i.e., lower charged apo-CA2 and holo-CA2 ions and higher charged apo-CA2 and holo-CA2 ions. The deconvolution spectrum of the ion peak ensemble for each type of ion was processed and time course plots of the relative intensities of the four ions were prepared in order to analyze the folding processes. These analyses revealed the coexistence of two folding states of the lower and higher charged apo-CA2 under the condition of pH 3.6. The lower and higher charged apoproteins spontaneously refolded to the lower charged holoprotein by a pH jump from 3.6 to 4.4 without the addition of an extra zinc ion. The higher charged holoprotein observed during both the unfolding and refolding processes was considered to be an intermediate of the change in folding. The present study indicates that ESI MS combined with pH jump would be a powerful method to probe the unfolding and refolding of proteins. This method simultaneously measures mass spectra and analyzes the folding processes as a function of time using deconvolution spectra constructed by selecting a suitable m/z range for the analysis from the peaks of charge state distributions.

Newly discovered orally active pure antiestrogens.

In order to develop orally active pure antiestrogens, we incorporated the carboxy-containing side chains into the 7alpha-position of the steroid scaffold and found that 17-keto derivative CH4893237 (12b) functioned as a pure antiestrogen with its oral activity much superior to clinically used pure antiestrogen, ICI182,780. Results from the pharmacokinetic evaluation indicated that the potent antiestrogen activity at oral dosing in mice attributed to both improved absorption from the intestinal wall and metabolic stability in liver.

Discovery of thiochroman derivatives bearing a carboxy-containing side chain as orally active pure antiestrogens.

In order to search for alternatives to the sulfoxide moiety in the long side chain of pure antiestrogens, several molecules that may interact with water in a fashion similar to ICI164,384 were designed and it was found that compounds with the carboxy, the sulfamide, or the sulfonamide instead of the sulfoxide moiety also functioned as pure antiestrogens. Interestingly, the compound possessing the carboxy moiety showed superior antiestrogen activity compared to ICI182,780 when dosed orally. Results of the pharmacokinetic evaluation indicated that the potent antiestrogen activity at oral dosing attributed to both the improved absorption from the intestinal wall and the metabolic stability of the compound in liver.

The determination of 2beta-(3-hydroxypropoxy)-1alpha,25-dihydroxy vitamin D3 (ED-71) in human serum by high-performance liquid chromatography-electrospray tandem mass spectrometry.

A liquid chromatography-electrospray ionization tandem mass spectrometric (LC/ESI-MS/MS) method for the determination of 2beta-(3-hydroxypropoxy)-1alpha,25-dihydroxy vitamin D3 (ED-71) in human serum has been developed. ED-71 in human serum was extracted using two solid-phase extraction steps on Bond Elut C18 and NH2 cartridge. The separation of ED-71 and preED-71 isomer was attained by LC using 2 mmol/L ammonium acetate-methanol (15:85, v/v) as a mobile phase on a Symmetry C18 column (5 microm, 150 mm x 2.1mm i.d.). ESI-MS/MS analysis was operated using selected reaction monitoring (SRM) in positive ion mode. The method achieved a lower limit of quantitation of 25 pg/mL. The calibration curve (25-3200 pg/mL) gave acceptable linearity (r>0.9964). Intra-assay precision ranged from 2.3 to 9.7%. Inter-assay precision ranged from 1.0 to 3.4%. The accuracy was within 90.8-107.0%. This highly sensitive and reproducible method is able to determine only biologically active ED-71 by separating it from preED-71, which is considered to be applicable for the determination of serum samples from pharmacokinetic studies in human.

Prediction of oral drug absorption in humans by theoretical passive absorption model.

The purpose of the present study was to examine the oral drug absorption predictability of the theoretical passive absorption model (TPAM). As chemical descriptors of drugs, the octanol/buffer distribution coefficient at pH 6.0 (D(ow)), intrinsic octanol-water partition coefficient (P(ow)), pK(a), and molecular weight (MW) were calculated from the chemical structure. Total passive intestinal membrane permeation consists of transcellular, paracellular and unstirred water layer (UWL) permeation. Transcellular permeation was modeled based on the pH-partition hypothesis with correction for cationic species permeation, and the independent variables were D(ow), P(ow), and pK(a). Paracellular permeation was modeled as a size-restricted diffusion within a negative electrostatic field-of-force, and the independent variables were MW and pK(a). UWL permeation was modeled as diffusion across a water layer, and the independent variable was MW. Cationic species permeation in the transcellular permeation model and the effect of a negative electric field-of-force in the paracellular permeation model were the extensions to the previous TPAM. The coefficients of the paracellular and UWL permeation models were taken from the literature. A data set of 258 compounds with observed values of Fa% (the fraction of a dose absorbed in humans) taken from the literature was employed to optimize four fitting coefficients in the transcellular permeation model. The TPAM predicted Fa%, with root mean square errors of 15-21% and a correlation coefficient (CC) of 0.78-0.88. In addition, the TPAM predicted the effective human intestinal membrane permeability with a CC of 0.67-0.77, as well as the contribution of paracellular permeation. The TPAM was found to predict oral absorption from the chemical structure of drugs with adequate predictability for usage in drug discovery.

Folding analysis of hormonal polypeptide calcitonins and the oxidized calcitonins using electrospray ionization mass spectrometry combined with H/D exchange.

Conformational change of calcitonins has been examined by measuring the rate of hydrogen/deuterium (H/D) exchange in amino acids. Time dependent m/z shift caused by H/D exchange was monitored by electrospray ionization quadrupole ion trap mass spectrometry (ESI-QIT MS). The rate constants of H/D exchange were calculated from apparent first-order kinetics. The time course of H/D exchange exhibited two phases of faster and slower exchange. The smaller rate constant (k2) estimated from the slower H/D exchange was correlated with an alpha-helix content that reflected the folding state. The order of k2 values obtained for human calcitonin (hCT), porcine calcitonin (pCT), salmon calcitonin (sCT), and elcatonin (ECT) was hCT > pCT approximately sCT > ECT. Although the amino acid sequence of sCT is similar to that of ECT, their k2 values were considerably different. The results suggest that ECT is relatively rigid on the N-terminal side cyclic structure in the folded state. Further, the effect of methionine oxidation on k2 has been examined. In the oxidized pCT that possesses similar biological activity with the intact pCT, the k2 values obtained were nearly equal. The k2 of hCT increased via methionine oxidation, and the biological activity was weakened by oxidation. This suggested that methionine oxidation of hCT produced unfolding in the secondary structure and that oxidative unfolding of hCT led to the loss of biological activity. The results indicate that the H/D exchange rate constant may be used as an informative parameter to elucidate the relationship between the folded state and biological activity of polypeptides like calcitonins with secondary structure.

Correction of permeability with pore radius of tight junctions in Caco-2 monolayers improves the prediction of the dose fraction of hydrophilic drugs absorbed by humans.

PURPOSE: To improve predictions of fraction dose absorbed (Fa) for hydrophilic drugs, a correction of paracellular permeability using the pore radius of tight junctions (TJs) in Caco-2 monolayers was performed. METHODS: The apparent permeability coefficient (P9app)) of drugs was measured using the Caco-2 assay and the parallel artificial membrane permeation assay (PAMPA), and values were corrected with the pore radius of TJs. RESULTS: An equation for calculating the pore radius of TJs from the P(app) of lucifer yellow was obtained. The optimal pore radius of TJs in Caco-2 monolayers for predicting human Fa was calculated to be 7 A. The correlation between the actual and predicted Fa was improved by using the P(app) corrected with the pore radius of TJs. Permeability in the PAMPA, which was corrected using the pore radius and membrane potential, was well correlated with that in the Caco-2 assay. Most of the hydrophilic drugs tested in this study were absorbed mainly through the paracellular pathway. CONCLUSIONS: The results suggest the necessity of optimizing paracellular permeation for the prediction of Fa, and also the importance of the paracellular pathway to the absorption of hydrophilic drugs. This method might contribute to the setting of appropriate dosages and the development of hydrophilic drugs.

Permeation characteristics of a hydrophilic basic compound across a bio-mimetic artificial membrane.

In the present study, the permeation characteristics of a hydrophilic basic compound (HBC) in a bio-mimetic parallel artificial membrane permeability assay (bio-mimetic PAMPA) were investigated in detail. The bio-mimetic PAMPA membrane was constructed on a hydrophobic filter by impregnating a lipid solution consisting of phosphatidylcholine (0.8%, w/w), phosphatidylethanolamine (0.8%, w/w), phosphatidylserine (0.2%, w/w), phosphatidylinositol (0.2%, w/w), cholesterol (1.0%, w/w), and 1,7-octadiene (97.0%, w/w). The pH-permeability curve (pH 3-10), the effect of lipid composition, concentration dependency (0.02-2.00 mM), and inhibition by other cationic compounds, were investigated for several HBCs. Ketoprofen and methylchlorpromazine were also employed as an acidic and a quaternary ammonium compound, respectively. At pH 3-6, the permeability of timolol, a HBC, was higher than expected from the pH-partition hypothesis, especially in the PI-containing membrane, whereas the pH-permeability curve of ketoprofen followed the pH-partition hypothesis. Permeation of HBC was saturable and inhibited by basic and quaternary ammonium compounds. Similar results were also found for methylchlorpromazine. The permeation characteristics of HBC observed in the present study are not usually expected in a passive permeation process across an artificial membrane. The participation of facilitated permeation of cationic species was suggested, in addition to a simple passive diffusion of un-dissociated species. Ion pair transport was suggested as a possible permeation mechanism of cationic species. However, further investigation is necessary to clarify the reason for the permeation characteristics of HBC.

Prediction of human intestinal permeability using artificial membrane permeability.

The purpose of the present study was to examine a correlation between the human intestinal permeability (P(eff)) and the bio-mimetic artificial membrane permeability corrected by the paracellular pathway model based on the Renkin function (P(PAMPA-PP-RF)) and to construct a prediction scheme. The effect of the unstirred water layer was incorporated to the prediction scheme. Eighteen P(eff) values of passively absorbed drugs were employed for the analysis. The correlation coefficient (CC) between the predicted and observed logP(eff) was 0.91. P(eff) of furosemide, hydrochlorothiazide and creatinine were underestimated by P(PAMPA-PP-RF). When these compounds were excluded, CC was 0.97. Without the correction for the paracellular pathway, P(eff) of small, cationic and hydrophilic compounds were underestimated. Therefore, P(PAMPA-PP-RF) was found to be an adequate in vitro surrogate for P(eff).

A study of the substrate specificity of Na+-dependent and Na+-independent neutral amino acid transport systems in dog intestinal brush-border membrane vesicles using L-alanine analogues.

Neutral amino acids are mainly transported across the intestinal brush-border membrane by two Na+-dependent systems (system B0 and system B0+) and one Na+-independent system (system b0,+). To investigate potential differences in substrate specificity between these systems, we screened ten different alanine analogues for their ability to inhibit the transport of L-alanine in dog intestinal brush-border membrane vesicles. The results suggested that a phenyl group directly attached to the alpha-carbon has different effects on the Na+-dependent and Na+-independent transport systems, with an increased affinity for the former and a decreased affinity for the latter. Based on these inhibition studies, we investigated [14C]L-phenylglycine transport kinetics in comparison with L-alanine. Similar to L-alanine, L-phenylglycine transport followed at least three routes, however, the Km of the Na+-dependent transport system was lower and the K'm of the Na+-independent system was higher than the corresponding values for L-alanine. These results corroborated the conclusions drawn from the inhibition studies. Based on these data, we conclude that different sizes of immediate parts to the alpha-carbon in functional groups of amino acid analogues have differential effects on the interaction of these amino acid analogues with the Na+-dependent and Na+-independent transport systems for neutral amino acids.