Comprehensive Analyses of the Intracellular and in Vivo Disposition of Fab- Small Interfering RNA Conjugate to Identify Key Issues to Improve Its in Vivo Activity.

Comprehensive analyses of intracellular disposition and in vivo pharmacokinetics were performed for small interfering RNA (siRNA) conjugated with the Fab fragment of panitumumab, a fully humanized monoclonal antibody against epidermal growth factor receptor (EGFR). The Fab-siRNA conjugate was internalized into EGFR-expressing cancer cells in an antigen-dependent manner. Intracellular disposition was quantitatively evaluated using fluorescent-labeled panitumumab and confocal microscopy. The majority of internalized panitumumab was suggested to be transferred into lysosomes. In vivo pharmacokinetics were evaluated in EGFR-expressing tumor-bearing mice. Intact Fab-siRNA was measured by immunoprecipitation using anti-Fab antibody followed by quantitative polymerase chain reaction. The Fab portion was measured by a ligand binding assay. Intact Fab-siRNA concentrations rapidly decreased in the plasma and tumor, although the Fab portion concentration remained high, suggesting extensive degradation in the linker-siRNA portion. After incubation of Fab-siRNA in mouse plasma, samples were digested with proteinase K, and extracted siRNA tagged with Fab-derived peptide was subjected to an ion-pair reversed-phase liquid chromatography with mass spectrometry analysis. Results suggested that hydrolysis from the 3' end of the antisense strand of siRNA is the major metabolizing pathway. Based on these findings, endosomal escape and stability in lysosomes, blood, and tumor are key factors to improve to achieve efficient target gene knockdown in tumors, and stabilizing the 3' end of the antisense strand was suggested to be most efficient. Our approaches clearly identified the key issues of Fab-siRNA from a pharmacokinetics aspect, which will be useful for improving the in vivo activity of siRNA conjugated with not only Fab but also other immunoproteins. SIGNIFICANCE STATEMENT: The intracellular and in vivo disposition of Fab-small interfering RNA (siRNA) conjugate was comprehensively investigated using various approaches, including newly developed analytical methods. This study clearly shows that improvements in siRNA stability in lysosomes, blood, and tumor are needed for target gene knockdown in tumors. The major metabolic pathway of Fab-siRNA is 3' exonuclease degradation, suggesting that optimization of the conjugation site to Fab might help improve stability.

An expanded genetic code facilitates antibody chemical conjugation involving the lambda light chain.

Most of the currently approved therapeutic antibodies are of the immunoglobulin gamma (IgG) κ isotype, leaving a vast opportunity for the use of IgGλ in medical treatments. The incorporation of designer amino acids into antibodies enables efficient and precise manufacturing of antibody chemical conjugates. Useful conjugation sites have been explored in the constant domain of the human κ-light chain (LCκ), which is no more than 38% identical to its LCλ counterpart in amino acid sequence. In the present study, we used an expanded genetic code for site-specifically incorporating Nε-(o-azidobenzyloxycarbonyl)-l-lysine (o-Az-Z-Lys) into the antigen-binding fragment (Fab) of an IgGλ, cixutumumab. Ten sites in the LCλ constant domain were found to support efficient chemical conjugation exploiting the bio-orthogonal azido chemistry. Most of the identified positions are located in regions that differ between the two light chain isotypes, thus being specific to the λ isotype. Finally, o-Az-Z-Lys was incorporated into the Fab fragments of cixutumumab and trastuzumab to chemically combine them; the resulting bispecific Fab-dimers showed a strong antagonistic activity against a cancer cell line. The present results expand the utility of the chemical conjugation method to the whole spectrum of humanized antibodies, including the λ isotype.

Development and evaluation of a novel antibody-photon absorber conjugate reveals the possibility of photoimmunotherapy-induced vascular occlusion during treatment in vivo.

Photodynamic therapy (PDT) utilize a photosensitizing agent and light for cancer therapy. It exerts anti-cancer effect mainly by inducing vascular occlusion at the irradiated site. By controlling the irradiation area, PDT can be used in a tumor-specific manner. However, the non-specific cellular damage in the surrounding normal tissue is still a serious concern. Photoimmunotherapy (PIT) is a new type of targeted cancer therapy that uses an antibody-photon absorber conjugate (APC). The superiority of PIT to PDT is the improved target specificity, thereby reducing the damage to normal tissues. Here, we developed a novel APC targeting epithelial cell adhesion molecule (EpCAM) as well as a negative control APC that does not bind to the EpCAM antigen. Our in vitro analysis of APC cytotoxicity demonstrated that the EpCAM APC, but not the negative control, was cytotoxic to EpCAM expressing COLO 205 cells after photoirradiation, suggesting that the cytotoxicity is antigen-dependent. However, in our in vivo analysis using a mouse xenograft tumor model, decreased volume of the tumors was observed in all the mice treated with irradiation, regardless of whether they were treated with the EpCAM APC or the negative control. Detailed investigation of the mechanism of these in vivo reveal that both APCs induce vascular occlusion at the irradiation site. Furthermore, the level of vascular occlusion was correlated with the blood concentration of APC, not the tumor concentration. These results imply that, similar to PDT, PIT can also induce non-targeted vascular occlusion and further optimization is required before widespread clinical use.

Extensive Survey of Antibody Invariant Positions for Efficient Chemical Conjugation Using Expanded Genetic Codes.

The site-specific chemical conjugation of proteins, following synthesis with an expanded genetic code, promises to advance antibody-based technologies, including antibody drug conjugation and the creation of bispecific Fab dimers. The incorporation of non-natural amino acids into antibodies not only guarantees site specificity but also allows the use of bio-orthogonal chemistry. However, the efficiency of amino acid incorporation fluctuates significantly among different sites, thereby hampering the identification of useful conjugation sites. In this study, we applied the codon reassignment technology to achieve the robust and efficient synthesis of chemically functionalized antibodies containing Nε-(o-azidobenzyloxycarbonyl)-l-lysine (o-Az-Z-Lys) at defined positions. This lysine derivative has a bio-orthogonally reactive group at the end of a long side chain, enabling identification of multiple new positions in Fab-constant domains, allowing chemical conjugation with high efficiency. An X-ray crystallographic study of a Fab variant with o-Az-Z-Lys revealed high-level exposure of the azido group to solvent, with six of the identified positions subsequently used to engineer "Variabodies", a novel antibody format allowing various connections between two Fab molecules. Our findings indicated that some of the created Variabodies exhibited agonistic activity in cultured cells as opposed to the antagonistic nature of antibodies. These results showed that our approach greatly enhanced the availability of antibodies for chemical conjugation and might aid in the development of new therapeutic antibodies.

Novel anticarcinoembryonic antigen antibody-drug conjugate has antitumor activity in the existence of soluble antigen.

Carcinoembryonic antigen (CEA) is a classic tumor-specific antigen that is overexpressed in several cancers, including gastric cancer. Although some anti-CEA antibodies have been tested, to the best of our knowledge, there are currently no clinically approved anti-CEA antibody therapies. Because of this, we have created the novel anti-CEA antibody, 15-1-32, which exhibits stronger binding to membrane-bound CEA on cancer cells than existing anti-CEA antibodies. 15-1-32 also shows poor affinity for soluble CEA; thus, the binding activity of 15-1-32 to membrane-bound CEA is not influenced by soluble CEA. In addition, we constructed a 15-1-32-monomethyl auristatin E conjugate (15-1-32-vcMMAE) to improve the therapeutic efficacy of 15-1-32. 15-1-32-vcMMAE showed enhanced antitumor activity against gastric cancer cell lines. Unlike with existing anti-CEA antibody therapies, antitumor activity of 15-1-32-vcMMAE was retained in the presence of high concentrations of soluble CEA.

One-Step Conjugation Method for Site-Specific Antibody-Drug Conjugates through Reactive Cysteine-Engineered Antibodies.

Engineered cysteine residues are particularly convenient for site-specific conjugation of antibody-drug conjugates (ADC), because no cell engineering and additives are required. Usually, unpaired cysteine residues form mixed disulfides during fermentation in Chinese hamster ovarian (CHO) cells; therefore, additional reduction and oxidization steps are required prior to conjugation. In this study, we prepared light chain (Lc)-Q124C variants in IgG and examined the conjugation efficiency. Intriguingly, Lc-Q124C exhibited high thiol reactivity and directly generated site-specific ADC without any pretreatment (named active thiol antibody: Actibody). Most of the cysteine-maleimide conjugates including Lc-Q124C showed retro-Michael reaction with cysteine 34 in albumin and were decomposed over time. In order to acquire resistance to a maleimide exchange reaction, the facile procedure for succinimide hydrolysis on anion exchange resin was employed. Hydrolyzed Lc-Q124C conjugate prepared with anion exchange procedure retained high stability in plasma. Recently, various stable linkage schemes for cysteine conjugation have been reported. The combination with direct conjugation by the use of Actibody and stable linker technology could enable the generation of stable site-specific ADC through a simple method. Actibody technology with Lc-Q124C at a less exposed position opens a new path for cysteine-based conjugation, and contributes to reducing entry barriers to the preparation and evaluation of ADC.

Identification of highly reactive cysteine residues at less exposed positions in the Fab constant region for site-specific conjugation.

Engineered cysteine residues are currently used for the site-specific conjugation of antibody-drug conjugates (ADC). In general, positions on the protein surface have been selected for substituting a cysteine as a conjugation site; however, less exposed positions (with less than 20% of accessible surface area [ASA]) have not yet been evaluated. In this study, we engineered original cysteine positional variants of a Fab fragment, with less than 20% of ASA, and evaluated their thiol reactivities through conjugation with various kinds of payloads. As a result, we have identified three original cysteine positional variants (heavy chain: Hc-A140C, light chain: Lc-Q124C and Lc-L201C), which exhibited similar monomer content, thermal stability, and antigen binding affinity in comparison to the wild-type Fab. In addition, the presence of cysteine in these positions made it possible for the Fab variants to react with variable-sized molecules with high efficiency. The favorable physical properties of the cysteine positional variants selected in our study suggest that less exposed positions, with less than 20% of ASA, provide an alternative for creating conjugation sites.

Site-specific DNA cleavage by artificial zinc finger-type nuclease with cerium-binding peptide.

The addition of a new function to native proteins is one of the most attractive protein-based designs. In this study, we have converted a C(2)H(2)-type zinc finger as a DNA-binding motif into a novel zinc finger-type nuclease by connecting two distinct zinc finger proteins (Sp1 and GLI) with a functional linker possessing DNA cleavage activity. As a DNA cleavage domain, we chose an analogue of the metal-binding loop (12 amino acid residues), peptide P1, which has been reported to exhibit a strong binding affinity for a lanthanide ion and DNA cleavage ability in the presence of Ce(IV). Our newly designed nucleases, Sp1(P1)GLI and Sp1(P1G)GLI, can strongly bind to a lanthanide ion and show a unique DNA cleavage pattern, in which certain positions between the two DNA-binding sites are specifically cleaved. The present result provides useful information for expanding the design strategy for artificial nucleases.

Swapping of the beta-hairpin region between Sp1 and GLI zinc fingers: significant role of the beta-hairpin region in DNA binding properties of C2H2-type zinc finger peptides.

The recent design strategy of zinc finger peptides has mainly focused on the alpha-helix region, which plays a direct role in DNA recognition. On the other hand, the study of non-DNA-contacting regions is extremely scarce. By swapping the beta-hairpin regions between the Sp1 and GLI zinc fingers, in this study, we investigated how the beta-hairpin region of the C(2)H(2)-type zinc finger peptides contributes to the DNA binding properties. Surprisingly, the Sp1 mutant with the GLI-type beta-hairpin had a higher DNA binding affinity than that of the wild-type Sp1. The result of the DNase I footprinting analyses also showed the change in the DNA binding pattern. In contrast, the GLI zinc finger completely lost DNA binding ability as a result of exchanging the beta-hairpin region. These results strongly indicate that the beta-hairpin region appears to function as a scaffold and has an important effect on the DNA binding properties of the C(2)H(2)-type zinc finger peptides.

NMR structure of transcription factor Sp1 DNA binding domain.

To understand the DNA recognition mechanism of zinc finger motifs of transcription factor Sp1, we have determined the solution structure of DNA-binding domain of the Sp1 by solution NMR techniques. The DNA-binding domain of Sp1 consists of three Cys(2)His(2)-type zinc finger motifs. They have typical betabetaalpha zinc finger folds and relatively random orientations. From DNA-binding analysis performed by NMR and comparison between structures determined here and previously reported structures of other zinc fingers, it was assumed that DNA recognition modes of fingers 2 and 3 would be similar to those of fingers of Zif268, in which each finger recognizes four base pairs strictly by using residues at positions -1, 2, 3, and 6 of the recognition helix. On the contrary, finger 1 can use only two residues for DNA recognition, Lys550 and His553 at positions -1 and 3 of the helix, and has more relaxed sequence and site specificity than other Cys(2)His(2) zinc fingers. It is proposed that this relaxed property of finger 1 allows transcription factor Sp1 to bind various DNA sequences with high affinity.

Exchange of histidine spacing between Sp1 and GLI zinc fingers: distinct effect of histidine spacing-linker region on DNA binding.

In the DNA recognition mode of C(2)H(2)-type zinc fingers, the finger-finger connection region, consisting of the histidine spacing (HX(3-5)H) and linker, would be important for determining the orientation of the zinc finger domains. To clarify the influence of spacing between two ligand histidines in the DNA binding, we exchanged the histidine spacing between Sp1 and GLI zinc fingers, which have an HX(3)H-TGEKK linker (typical) and an HX(4)H-SNEKP linker (atypical), respectively. A significant decrease in the DNA binding affinity and specificity is found in Sp1-type peptides, whereas GLI-type peptides show a mild reduction. To evaluate the effect of the linker characteristics, we further designed Sp1-type mutants with an SNEKP linker. As a result, the significant effect of the histidine spacing in Sp1-type peptides was reduced. These results demonstrate that (1) the histidine spacing significantly affects the DNA binding of zinc finger proteins and (2) the histidine spacing and the following linker regions are one effective target for regulating the DNA recognition mode of zinc finger proteins.

Total synthesis of artificial zinc-finger proteins: Problems and perspectives.

The total synthesis of a peptide segment corresponding to the DNA-binding segment of Sp1 (positions 532-623) using a native chemical ligation approach is described. The folding of the synthetic segment in the presence of Zn(II) gave a zinc-coordinated protein. The dissociation constant (K(D)) for the DNA binding of the resulting protein, determined by a gel mobility shift assay, was 130 nM, almost nine times higher than that of the genetically prepared protein. However, methylation interference assay showed an identical sequence specificity of both proteins in DNA recognition. The chemical ligation method to connect the respective zinc-finger units was also accomplished. Successive ligation between a cysteine-containing peptide segment and a chloroacetylated peptide segment gave an artificial three-finger protein, which corresponds to the above DNA-binding segment of Sp1. However, this protein failed to bind DNA, even at 1.25 mM. Assessment of their folding structure based on the absorption spectra of their Co(II) complexes showed that the linker design to connect the respective finger units is critical for the proper folding of the proteins as well as the occurrence of the DNA-binding function.

Influence of TFIIIA-type linker at the N- or C-terminal of nine-zinc finger protein on DNA-binding site.

The central three-zinc finger connection of the native nine-zinc finger protein transcription factor IIIA (TFIIIA) is composed of unique linker sequences, -NIKICV-, -TQQLP-, -AG-, and -QDL-. New artificial nine-zinc finger proteins, Sp1ZF9TC and Sp1ZF9TN, which use the TFIIIA-type linker for their C- and N-terminal three-zinc finger connections have been created. To investigate the influence of TFIIIA-type linker sequences by their different locations in the proteins, gel mobility shift assays (GMSA), DNase I footprinting assays, methylation interference analyses, and hydroxyl radical footprinting assays were performed. The GMSA revealed similar DNA-binding affinities of these two proteins. The footprinting analyses indicated that the two zinc finger proteins recognize the same part of GCII or GCIII DNA. Moreover, the specific base contacts were observed in the same sites of the substrate DNA. In the present proteins, Sp1ZF9TC and Sp1ZF9TN, the four zinc fingers (fingers 1-4 or 5-9) situated in the site opposite to the TFIIIA-type linker position participate in their DNA bindings. The position of the TFIIIA-type linker is important in DNA recognition by multi-zinc finger proteins.

Interconversion between serine and aspartic acid in the alpha helix of the N-terminal zinc finger of Sp1: implication for general recognition code and for design of novel zinc finger peptide recognizing complementary strand.

In the typical base recognition mode of the C(2)H(2)-type zinc finger, the amino acid residues at alpha-helical positions -1, 3, and 6 make a contact with the base in one strand (the primary strand), and the residue at position 2 interacts with the base in a complementary strand (the secondary strand). The N-terminal zinc finger of the three-zinc-finger domain of Sp1 has inherently a unique five-base-pair binding mode in which the guanine bases are recognized in both strands. To clarify the effect of the amino acid at position 2 on DNA binding affinity and base specificity, we have created a library of the mutants by the interconversion between serine and aspartic acid in the N-terminal zinc finger of Sp1 and recombinant variants of finger order. Gel mobility shift and methylation interference assays showed that the combination of arginine and serine at positions -1 and 2, respectively, provides a newly strong guanine contact in the secondary strand and a higher binding affinity than that of wild-type Sp1. Of special interest are the facts that the mutant with lysine and aspartic acid at positions -1 and 2 in the alpha helix predominantly recognizes the bases in the secondary strand and that its DNA binding affinity is higher than that of the wild-type. The aspartic acid or serine at position 2 independently contributes to the DNA binding affinity and base specificity. The present results provide useful information for the design of a novel zinc finger protein with priority for the bases in the secondary strand.