Search

The maytansinoid antibody-drug conjugates (ADCs) in clinical development for cancer therapy each contain a derivative of the microtubule-targeting agent, maytansine, covalently attached to the antibody via an engineered linker. A sample of any of these conjugates contains molecules with different numbers of maytansinoid molecules, or "drug" loads, the relative abundance of which can be determined by mass spectrometry. We examined the accuracy of the Poisson distribution and the binomial distribution in predicting the relative abundance of species with different drug loads for three antibody-maytansinoid conjugates with different antibodies and linker-maytansinoid pairings. We used variance, calculated from the experimental mass distribution data, as the parameter to determine the optimal value n of the binomial distribution number of trials. The accuracy of the Poisson distribution in predicting distribution of the species abundance in these conjugates varied among the conjugates. In contrast, the accuracy of the binomial distribution was similar for all three conjugates and comparable to the best accuracy of the Poisson distribution, as supported by a paired t-test.

Recombinant monoclonal antibody drug products play an increasingly important role in the treatment of various diseases. Antibodies are large, multi-chain proteins and antibody preparations often contain several molecular variants, which renders them heterogeneous. The heterogeneity is further increased in immunoconjugates prepared by covalently linking several drug molecules per antibody molecule. As part of the product characterization, the molecular weights of the antibodies or their drug conjugates need to be measured. Electrospray ionization mass spectrometry (ESI-MS) is well suited for the analysis of recombinant antibodies and immunoconjugates. Sample preparation is an important element of ESI-MS analysis, in particular samples need to be freed of interfering charged species, such as salts and buffer components. In this paper, Amicon centrifugal filters, reversed-phase high-performance liquid chromatography (HPLC), and size-exclusion HPLC were evaluated for sample desalting. Size-exclusion HPLC, using aqueous acetonitrile as the mobile phase, directly coupled to ESI-MS provided the best performance and was optimized for the study of immunoconjugates. The results showed that antibodies carrying covalently linked maytansinoid molecules generated charge envelope profiles that differ from those of the non-conjugated antibody. For the determination of the distribution of the various conjugate species in an immunoconjugate sample prepared by randomly linking in the average 3.6 drug molecules per antibody molecule, the experimental conditions needed to be carefully selected to allow acquisition of the whole spectrum containing the charge envelopes of all species.

Immunoconjugates are being explored as novel cancer therapies with the promise of target-specific drug delivery. The immunoconjugate, huN901-DM1, composed of the humanized monoclonal IgG1 antibody, huN901, and the maytansinoid drug, DM1, is being tested in clinical trials to treat small cell lung carcinoma (SCLC). huN901-DM1 contains an average of three to four DM1 drug molecules per huN901 antibody molecule. The drug molecules are linked to huN901 through random modification of huN901 at epsilon-amino groups of lysine residues, thus yielding a heterogeneous population of conjugate species. We studied the drug distribution profile of huN901-DM1 by electrospray time-of-flight mass spectrometry(ESI-TOFMS), which showed that one to six DM1 drug molecules were attached to an antibody molecule. Both light and heavy chains contained linked drugs. The conjugation sites in both chains were determined by peptide mapping using trypsin and Asp-N protease digestion. Trypsin digestion identified modified lysine residues, since these residues were no longer susceptible to enzymatic cleavage after conjugation with the drug. With respect to Asp-N digestion, modified peptides were identified by observing a mass increase corresponding to the modification. The two digestion methods provided consistent results, leading to the identification of 20 modified lysine residues in both light and heavy chains. Each lysine residue was only partially modified. No conjugation sites were found in complementarity determining regions (CDRs). Using structural models of human IgG1, it was found that modified lysine residues were on the surface in areas of structural flexibility and had large solvent accessibility.

The aim of this study was to perform structural characterization of a recombinant monoclonal antibody (MAb), huN901, by electrospray time-of-flight mass spectrometry (ESI-TOFMS) using both “top-down” and “bottom-up” approaches. In the top-down approach, the molecular masses of the deglycosylated huN901 and the light and heavy chains of the antibody were measured by direct infusion MS and liquid chromatography–mass spectrometry (LC–MS). In the bottom-up approach, trypsin and Asp-N protease were used to digest the separated, reduced and alkylated light and heavy chains followed by LC–MS analysis of the digests. The primary structure and post-translational modifications of huN901 were characterized by both top-down and bottom-up MS approaches. Modifications of N-terminal pyroglutamate formation, cleavage of C-terminal lysine, glycosylation, and deamidation were identified in the antibody heavy chain by both protein mass measurement and peptide mapping. No modifications were found in the complementarity determining regions (CDRs) of both chains. Both trypsin and Asp-N protease digestion had an average sequence recovery of 97%, and generated complimentary mapping results with complete sequence recovery. ESI-TOFMS is a superior tool to characterize MAb and other complex protein pharmaceuticals.

A reversed-phase high-performance liquid chromatography (RP-HPLC) method was developed for detection and quantification of free maytansinoid drug in disulfide-linked conjugates between monoclonal antibodies and the maytansinoid drug DM1 (MAb-DM1). Mobile phases and gradient conditions were optimized for separation of several DM1-related free drug species from MAb-DM1 conjugates. The selectivity, linearity, and reproducibility of the method are reported. Reduction of the disulfide-linked DM1 followed by RP-HPLC allowed estimation of purity of MAb-linked DM1 as well as recovery of L-DM1. The method was also used to estimate drug per MAb ratios, which were consistent with those determined by UV spectroscopy.

The binding of 45Ca2+ to bovine plasma protein C (PC) and to activated bovine plasma protein C (APC) has been examined by equilibrium ultrafiltration at pH 7.4 and 25 degrees C. Under these conditions, PC possesses 16.0 plus or minus 2.0 equivalent Ca2+ binding sites, of average KD (8.7 plus or minus 1.5) x 10(-4) M, and APC contains 9.0 plus or minus 1.0 equivalent Ca2+ binding sites, with an average KD of (4.3 plus or minus 1.1) x 10(-4) M. Both Mn2+ and Sr2+ were capable of ready displacement of Ca2+ from a Ca2+-PC complex, while Mg2+ was less effective in this regard. The alpha-thrombin-catalyzed activation of PC was inhibited by the presence of Ca2+. A kinetic analysis of this effect demonstrated that it was, in large part, due to an increase in the Km of the reaction. Addition of other divalent cations, e.g. Mn2+, Sr2+, and Mg2+, in place of Ca2+ also resulted in inhibition of the alpha-thrombin-catalyzed activation of PC in a manner which paralleled their ability to displace Ca2+ from a Ca2+-PC complex. On the other hand, the activation of PC by the coagulant protein from Russell's Viper venom was augmented by the presence of Ca2+. Other divalent metal ions, such as Sr2+ and Mn2+, in the absence of Ca2+, also weakly stimulated this reaction. Mg2+ was without notable effect.

Binding isotherms of Ca2+ to the bovine Factor IX activation intermediates and products, i.e. Factor IXalpha, Factor IXa alpha, and Factor IXa beta have been examined. At pH 7.4, Factor IX alpha possesses at least two strong Ca2+ sites, with an average KD of 0.1 mM, and an additional 11 weaker sites, with an average KD of 3.7 mM. Bovine Factor IXa alpha also contains at least two Ca2+ binding sites, with an average KD of 0.1 mM, and an additional 11 weaker sites, with an average KD of 1.3 mM. Factor IXa beta, the ultimate activation product of Factor IX, in the intrinsic system, likewise contains at least two strong Ca2+ sites, of average KD 0.1 mM, as well as seven additional weaker sites, possessing an average KD of 1.0 mM. The Ca2+-binding properties of the above proteins are similar to those of their precursor molecule, Factor IX, which we have earlier shown to possess at least two strong Ca2+ sites, with an average KD of 0.1 mM, and 11 weaker sites, of average KD 1.3 mM (Amphlett, G.W., Byrne, R., and Castellino, F.J. (1978) J. Biol. Chem. 253, 6774-6779). Circular dichroism analysis of all of the above proteins was consistent with the molecules possessing a low alpha-helical content, and a high quantity of beta structure and random coil conformations.

The binding isotherms of Ca2+ and Mn2+ to bovine factor IX have been determined at pH 6.5 and pH 7.4, at 25 degrees C. At pH 7.4, at least 2 strong Ca2+ sites, with an average KDISS of 0.1 +/- 0.02 mM, are found. An additional 10 to 11 weaker Ca2+ binding sites, with an average KDISS of 1.3 +/- 0.2 mM are noted, at high levels of Ca2+. At pH 6.5, again at least 2 strong Ca2+ sites on factor IX are evident, with an average KDISS of 0.11 +/- 0.02 mM; but slightly fewer (7 to 8) weaker sites, with an average KDISS of 1.9 +/- 0.3 mM, are obtained. Qualitatively, the binding of Mn2+ to bovine factor IX appears similar to that of Ca2+. At pH 6.5, approximately 2 strong Mn2+ binding sites, with an average KDISS of 13 +/- 1.5 micrometer, and an additional 5 to 6 weak sites, with an average KDISS of 160 +/- 15 micrometer, are present. Manganese does not completely displace Ca2+; and Ca2+ does not completely displace Mn2+ from their respective binding sites. On the other hand, Tb3+ and Sm3+ readily displace Ca2+, at pH 6.5, from its sites on factor IX. The rate and extent of activation of bovine factor IX, by bovine factor XIa, is dependent on the Ca2+ concentration, up to concentrations of Ca2+ which saturate its effect on the system. Substitution of Sr2+ for Ca2+ leads to approximately 50% of the maximum rate of factor IXa formation, and final yield of factor IXa, in this activation system. Manganese does not substitute for Ca2+ in this activation, but does inhibit the stimulatory effect of Ca2+. Both Tb3+ and Sm3+ are effective inhibitors of Ca2+ in factor IX activation by factor XIa.

Summary The rate of hydrolysis of suboptimal concentrations of N-benzoyl- L ¯ ¯ -phenylanalyl- L ¯ ¯ -valyl- L ¯ ¯ -arginine- p- nitroanilide hydrochloride (S-2160) by activated bovine plasma protein C is greatly enhanced by the presence of monovalent cations. Of the cations tested, at their maximal effective concentrations, the order of the ability to enhance the S-2160 amidase activity of activated protein C is Cs+ > Rb+ > K+ > NH4+ > Na+ > Li+. Progressive substitution of methyl groups for hydrogen in NH4+ leads to a concomitant decrease in the ability of these compounds to enhance the amidase activity of activated protein C. Divalent cations, at low concentrations, weakly enhance the amidase activity of activated protein C and show a decreased tendency in this regard at high concentrations. The above observations also apply to hydrolysis of the substrate α -N-tosyl- L ¯ ¯ -arginine methyl ester.

A complete copy of the mRNA sequences encoding human coagulation factor VIII:C has been cloned and expressed. The DNA sequence predicts a single chain precursor of 2,351 amino acids with a relative molecular mass (Mr) 267,039. The protein has an obvious domain structure, contains sequence repeats and is structurally related to factor V and ceruloplasmin.

The binding of calcium to human plasma fibronectin has been measured by equilibrium dialysis at 25 degrees in 0.1 M NaCl 50mM Tris HCL, pH 7.4. Curve fitting of the binding data indicates that fibronectin has two strong calcium binding sites per chain (Mr 220,000), KD = 1.3 mM and approximately 12 weak sites, KD = 2.3 mM. No significant displacement of bound calcium by magnesium was observed at magnesium concentrations up to 1 mM. Calcium binding to a pair of tryptic fragments of fibronectin (Mr congruent to 160,000 and 180,000) that bind to gelatin has also been investigated. These fragments have a single class of calcium binding sites, with 2.2 sites per chain, KD = 1.1 mM. Negligible calcium binding to tryptic fragments derived from other regions of the fibronectin molecule was observed.

Factor VIII is a high molecular weight plasma glycoprotein that functions in the blood clotting cascade as the cofactor for factor DCa proteolytic activation of factor X. Factor VIII does not function proteolytically in this reaction hut itself can be proteolytically activated by other coagulation enzymes such as factor Xa and thrombin. In the plasma, factor VIII exists as a 200 kDa amino-terminal fragment in a metal ion stabilized complex with a 76 kDa carboxy-terminal fragment. The isolation of the cENA for human factor VIII provided the deduced primary amino acid sequence of factor VIIT and revealed three distinct structural domains: 1) a triplicated A domain of 330 amino acids which has homology to ceruloplasmin, a plasma copper binding protein, 2) a duplicated C domain of 150 amino acids, and 3) a unique B domain of 980 amino acids. These domains are arranged as shown below. We have previously reported the B domain is dispensible far cofactor activity in vitro (Toole et al. 1986 Proc. Natl. Acad 5939). The in vivo efficacy of factor VIII molecules harboring the B domain deletion was tested by purification of the wildtype and modified forms and infusion into factor VIII deficient, hemophilic, dogs. The wildtype and the deleted forms of recombinant derived factor VIII exhibited very similar survival curves (Tl/2 = 13 hrs) and the cuticle bleeding times suggested that both preparations appeared functionally equivalent. Sepharose 4B chromatography indicated that both factor VIII molecules were capable of binding canine plasma vWF.Further studies have addressed what cleavages are necessary for activation of factor VIII. The position of the thrombin, factor Xa, and activated protein C (AFC) cleavage sites within factor VIII are presented below, site-directed ENA medicated mutagenesis has been performed to modify the arginine at the amino side of each cleavagesite to an soleucine. In all cases this modification resulted in molecules that were resistant to cleavage by thrombin at the modified site. Modification of the thrombin cleavage sites at 336 and 740 and modification of the factor Xa cleavage site at 1721 resulted in no loss of cofactor activity. Modification of the thrombin cleavage site at either 372 or 1689 destroyed oofactor activity. Modification of the thrombin cleavage site at 336 resulted in a factor VIII having an increased activity, possibly due to resistance to inactivation. These results suggest the requirement of cleavage at residues 372 and 1689 for cofactor activity.

The binding isotherms of Ca 2+ and Sr 2+ to human blood coagulation Factor IX have been obtained at 25 °C and pH 7.4. In the case of both cations, a Scatchard plot of the data reveals that a single class of binding sites exist. For Ca 2+ , a total of 16.0 ± 1.0 sites, of K D 7.3 ± 0.2 × 10 −4 m , are present on human Factor IX. Similar analysis of the Sr 2+ data indicates that Factor IX contains 11.0 ± 1.0 binding sites, with a K D of 1.9 ± 0.1 × 10 −3 m . Both Sr 2+ and Mn 2+ effectively displace Ca 2+ from human Factor IX; whereas Mg 2+ is considerably less potent in this regard. Conversely, Ca 2+ is capable of nearly complete displacement of Sr 2+ from its binding sites on human Factor IX. The activation of human Factor IX, by human Factor XIa, shows a complex dependence on the Ca 2+ concentration. Sr 2+ can substitute for Ca 2+ in this activation process. Mn 2+ cannot, in itself, substitute for Ca 2+ in activation of Factor IX, but does significantly enhance the activation of Factor IX by Factor XIa at suboptimal levels of Ca 2+ . The rate of activation of human Factor IX by the coagulant protein of Russell's viper venom also shows a dependence on the presence of divalent cations. Here, however, a rigid specificity is not noted, since Ca 2+ , Sr 2+ , and Mn 2+ all allow activation to proceed equally well.