Antibody Co-Administration Can Improve Systemic and Local Distribution of Antibody-Drug Conjugates to Increase In Vivo Efficacy

Several antibody-drug conjugates (ADC) showing strong clinical responses in solid tumors target high expression antigens (HER2, TROP2, Nectin-4, and folate receptor alpha/FRα). Highly expressed tumor antigens often have significant low-level expression in normal tissues, resulting in the potential for target-mediated drug disposition (TMDD) and increased clearance. However, ADCs often do not cross-react with normal tissue in animal models used to test efficacy (typically mice), and the impact of ADC binding to normal tissue antigens on tumor response remains unclear. An antibody that cross-reacts with human and murine FRα was generated and tested in an animal model where the antibody/ADC bind both human tumor FRα and mouse FRα in normal tissue. Previous work has demonstrated that a "carrier" dose of unconjugated antibody can improve the tumor penetration of ADCs with high expression target-antigens. A carrier dose was employed to study the impact on cross-reactive ADC clearance, distribution, and efficacy. Co-administration of unconjugated anti-FRα antibody with the ADC-improved efficacy, even in low expression models where co-administration normally lowers efficacy. By reducing target-antigen-mediated clearance in normal tissue, the co-administered antibody increased systemic exposure, improved tumor tissue penetration, reduced target-antigen-mediated uptake in normal tissue, and increased ADC efficacy. However, payload potency and tumor antigen saturation are also critical to efficacy, as shown with reduced efficacy using too high of a carrier dose. The judicious use of higher antibody doses, either through lower DAR or carrier doses, can improve the therapeutic window by increasing efficacy while lowering target-mediated toxicity in normal tissue.

Quantifying ADC bystander payload penetration with cellular resolution using pharmacodynamic mapping

With the recent approval of 3 new antibody drug conjugates (ADCs) for solid tumors, this class of drugs is gaining momentum for the targeted treatment of cancer. Despite significant investment, there are still fundamental issues that are incompletely understood. Three of the recently approved ADCs contain payloads exhibiting bystander effects, where the payload can diffuse out of a targeted cell into adjacent cells. These effects are often studied using a mosaic of antigen positive and negative cells. However, the distance these payloads can diffuse in tumor tissue while maintaining a lethal concentration is unclear. Computational studies suggest bystander effects partially compensate for ADC heterogeneity in tumors in addition to targeting antigen negative cells. However, this type of study is challenging to conduct experimentally due to the low concentrations of extremely potent payloads. In this work, we use a series of 3-dimensional cell culture and primary human tumor xenograft studies to directly track fluorescently labeled ADCs and indirectly follow the payload via an established pharmacodynamic marker (γH2A. X). Using TAK-164, an anti-GCC ADC undergoing clinical evaluation, we show that the lipophilic DNA-alkylating payload, DGN549, penetrates beyond the cell targeted layer in GCC-positive tumor spheroids and primary human tumor xenograft models. The penetration distance is similar to model predictions, where the lipophilicity results in moderate tissue penetration, thereby balancing improved tissue penetration with sufficient cellular uptake to avoid significant washout. These results aid in mechanistic understanding of the interplay between antigen heterogeneity, bystander effects, and heterogeneous delivery of ADCs in the tumor microenvironment to design clinically effective therapeutics.

Preclinical Antitumor Activity and Biodistribution of a Novel Anti-GCC Antibody-Drug Conjugate in Patient-derived Xenografts

Guanylyl cyclase C (GCC) is a unique therapeutic target with expression restricted to the apical side of epithelial cell tight junctions thought to be only accessible by intravenously administered agents on malignant tissues where GCC expression is aberrant. In this study, we sought to evaluate the therapeutic potential of a second-generation investigational antibody-dug conjugate (ADC), TAK-164, comprised of a human anti-GCC mAb conjugated via a peptide linker to the highly cytotoxic DNA alkylator, DGN549. The in vitro binding, payload release, and in vitro activity of TAK-164 was characterized motivating in vivo evaluation. The efficacy of TAK-164 and the relationship to exposure, pharmacodynamic marker activation, and biodistribution was evaluated in xenograft models and primary human tumor xenograft (PHTX) models. We demonstrate TAK-164 selectively binds to, is internalized by, and has potent cytotoxic effects against GCC-expressing cells in vitro A single intravenous administration of TAK-164 (0.76 mg/kg) resulted in significant growth rate inhibition in PHTX models of metastatic colorectal cancer. Furthermore, imaging studies characterized TAK-164 uptake and activity and showed positive relationships between GCC expression and tumor uptake which correlated with antitumor activity. Collectively, our data suggest that TAK-164 is highly active in multiple GCC-positive tumors including those refractory to TAK-264, a GCC-targeted auristatin ADC. A strong relationship between uptake of ⁸⁹Zr-labeled TAK-164, levels of GCC expression and, most notably, response to TAK-164 therapy in GCC-expressing xenografts and PHTX models. These data supported the clinical development of TAK-164 as part of a first-in-human clinical trial (NCT03449030).

Increased Tumor Penetration of Single-Domain Antibody-Drug Conjugates Improves In Vivo Efficacy in Prostate Cancer Models

Targeted delivery of chemotherapeutics aims to increase efficacy and lower toxicity by concentrating drugs at the site-of-action, a method embodied by the seven current FDA-approved antibody-drug conjugates (ADC). However, a variety of pharmacokinetic challenges result in relatively narrow therapeutic windows for these agents, hampering the development of new drugs. Here, we use a series of prostate-specific membrane antigen-binding single-domain (Humabody) ADC constructs to demonstrate that tissue penetration of protein-drug conjugates plays a major role in therapeutic efficacy. Counterintuitively, a construct with lower in vitro potency resulted in higher in vivo efficacy than other protein-drug conjugates. Biodistribution data, tumor histology images, spheroid experiments, in vivo single-cell measurements, and computational results demonstrate that a smaller size and slower internalization rate enabled higher tissue penetration and more cell killing. The results also illustrate the benefits of linking an albumin-binding domain to the single-domain ADCs. A construct lacking an albumin-binding domain was rapidly cleared, leading to lower tumor uptake (%ID/g) and decreased in vivo efficacy. In conclusion, these results provide evidence that reaching the maximum number of cells with a lethal payload dose correlates more strongly with in vivo efficacy than total tumor uptake or in vitro potency alone for these protein-drug conjugates. Computational modeling and protein engineering can be used to custom design an optimal framework for controlling internalization, clearance, and tissue penetration to maximize cell killing. SIGNIFICANCE: A mechanistic study of protein-drug conjugates demonstrates that a lower potency compound is more effective in vivo than other agents with equal tumor uptake due to improved tissue penetration and cellular distribution.

Site-Specific Conjugation of the Indolinobenzodiazepine DGN549 to Antibodies Affords Antibody-Drug Conjugates with an Improved Therapeutic Index as Compared with Lysine Conjugation

Antibody-drug conjugates have elicited great interest recently as targeted chemotherapies for cancer. Recent preclinical and clinical data have continued to raise questions about optimizing the design of these complex therapeutics. Biochemical methods for site-specific antibody conjugation have been a design feature of recent clinical ADCs, and preclinical reports suggest that site-specifically conjugated ADCs generically offer improved therapeutic indices (i.e., the fold difference between efficacious and maximum tolerated doses). Here we present the results of a systematic preclinical comparison of ADCs embodying the DNA-alkylating linker-payload DGN549 generated with both heterogeneous lysine-directed and site-specific cysteine-directed conjugation chemistries. Importantly, the catabolites generated by each ADC are the same regardless of the conjugation format. In two different model systems evaluated, the site-specific ADC showed a therapeutic index benefit. However, the therapeutic index benefit is different in each case: both show evidence of improved tolerability, though with different magnitudes, and in one case significant efficacy improvement is also observed. These results support our contention that conjugation chemistry of ADCs is best evaluated in the context of a particular antibody, target, and linker-payload, and ideally across multiple disease models.

Antibody-Drug Conjugates with Indolinobenzodiazepine Dimer Payloads: DNA-Binding Mechanism of Indolinobenzodiazepine Dimer Catabolites in Target Cancer Cells

DNA-targeting indolinobenzodiazepine dimer (IGN) payloads are used in several clinical-stage antibody-drug conjugates. IGN drugs alkylate DNA through the single imine moiety present in the dimer in contrast to the pyrrolobenzodiazepine dimer drugs, such as talirine and tesirine, which contain two imine moieties per dimer and cross-link DNA. This study explored the mechanism of binding of IGN to DNA in cells and to synthetic duplex and hairpin oligonucleotides. New, highly sensitive IGN-DNA binding enzyme-linked immunosorbent assay methods were developed using biotinylated IGN analogues (monoimine, diimine, and diamine IGNs) and digoxigenin-labeled duplex oligonucleotides, which allowed the measurement of drug-DNA adducts in viable cells at concentrations below IC₅₀. Furthermore, the release of free drug from the IGN-DNA adduct upon treatment with nuclease ex vivo was tested under physiological conditions. The monoimine IGN drug formed a highly stable adduct with DNA in cells, with stability similar to that of the diimine drug analogue. Both monoimine and diimine IGN-DNA adducts released free drugs upon DNA cleavage by nuclease at 37 °C, although more free drug was released from the monoimine compared to the diimine adduct, which presumably was partly cross-linked. The strong binding of the monoimine IGN drug to duplex DNA results from both the noncovalent IGN-DNA interaction and the covalent bond formation between the 2-amino group of a guanine residue and the imine moiety in IGN.

Synthesis and Evaluation of Camptothecin Antibody-Drug Conjugates

Antibody-drug conjugates (ADCs) that incorporate the exatecan derivative DXd in their payload are showing promising clinical results in solid tumor indications. The payload has an F-ring that also contains a second chiral center, both of which complicate its synthesis and derivatization. Here we report on new camptothecin-ADCs that do not have an F-ring in their payloads yet behave similarly to DXd-bearing conjugates in vitro and in vivo. This simplification allows easier derivatization of camptothecin A and B rings for structure-activity relationship studies and payload optimization. ADCs having different degrees of bystander killing and the ability to release hydroxyl or thiol-bearing metabolites following peptide linker cleavage were investigated.

Synthesis of Highly Potent N-10 Amino-Linked DNA-Alkylating Indolinobenzodiazepine Antibody-Drug Conjugates (ADCs)

Indolinobenzodiazepine DNA alkylators (IGNs) are the cytotoxic payloads in antibody-drug conjugates (ADCs) currently undergoing Phase I clinical evaluation (IMGN779, IMGN632, and TAK164). These ADCs possess linkers that have been incorporated into a central substituted phenyl spacer. Here, we present an alternative strategy for the IGNs, linking through a carbamate at the readily available N-10 amine present in the monoimine containing dimer. As a result, we have designed a series of N-10 linked IGN ADCs with a wide range of in vitro potency and tolerability, which may allow us to better match an IGN with a particular target based on the potential dosing needs.

A Case Study Comparing Heterogeneous Lysine- and Site-Specific Cysteine-Conjugated Maytansinoid Antibody-Drug Conjugates (ADCs) Illustrates the Benefits of Lysine Conjugation

Antibody-drug conjugates are an emerging class of cancer therapeutics constructed from monoclonal antibodies conjugated with small molecule effectors. First-generation molecules of this class often employed heterogeneous conjugation chemistry, but many site-specifically conjugated ADCs have been described recently. Here, we undertake a systematic comparison of ADCs made with the same antibody and the same macrocyclic maytansinoid effector but conjugated either heterogeneously at lysine residues or site-specifically at cysteine residues. Characterization of these ADCs in vitro reveals generally similar properties, including a similar catabolite profile, a key element in making a meaningful comparison of conjugation chemistries. In a mouse model of cervical cancer, the lysine-conjugated ADC affords greater efficacy on a molar payload basis. Rather than making general conclusions about ADCs conjugated by a particular chemistry, we interpret these results as highlighting the complexity of ADCs and the interplay between payload class, linker chemistry, target antigen, and other variables that determine efficacy in a given setting.

Abstract 231: Optimizing lysosomal activation of antibody-drug conjugates (ADCs) by incorporation of novel cleavable dipeptide linkers

Peptides constitute a major linker class in antibody-drug conjugates (ADCs), designed to connect antibodies to cytotoxic drug molecules. An optimal linker for an ADC should be stable in circulation and be cleaved efficiently in lysosomes upon binding and internalization of the ADC in target cells, releasing drug molecules that exit lysosomes and inhibit key cellular functions. Previous studies have focused on a limited selection of peptide linkers and specific biochemical tools, such as a cathepsin B cleavage assay, leading to the selection of a valine-citrulline dipeptide linker with a p-aminobenzyloxycarbonyl (PABC) spacer, which is now often used in candidate ADCs.

In this study, we screened a panel of dipeptide linkers for efficient lysosomal proteolysis. Dipeptide linkers bearing different amino acids (typically both L, with D used as a control) were synthesized with a fluorogenic leaving group, 7-amino-4-methylcoumarin, which provided a convenient model system for screening of peptide cleavage using lysosomal extracts from cancer cells as well as several individual cathepsins. The linkers were also tested for stability in mouse, rat, cynomolgus, and human plasma.

Based on these screens, we identified several novel, previously unreported peptide linker designs and incorporated them into ADCs bearing a DNA-alkylating indolinobenzodiazepine (IGN) payload. These ADCs with different peptide linkers were assayed for in vitro cytotoxicity in multiple cancer cell lines, in vivo efficacy in human tumor xenograft models in mice, and ex vivo plasma stability. In addition, we measured the impact of peptide linkers on the kinetics of proteolytic processing of ADCs in cancer cell lines. We observed that several dipeptide linker designs were superior in rates of lysosomal processing compared to a reference standard L-Ala-L-Ala dipeptide linker.

Citation Format: Paulin Salomon, Luke Harris, Emily E. Reid, Erin K. Maloney, Alan J. Wilhelm, Michael L. Miller, Ravi V. Chari, Thomas A. Keating, Rajeeva Singh. Optimizing lysosomal activation of antibody-drug conjugates (ADCs) by incorporation of novel cleavable dipeptide linkers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 231.

Abstract 230: Antibody-drug conjugates (ADCs) with indolinobenzodiazepine dimer (IGN) payloads: DNA-binding mechanism of IGN catabolites in target cancer cells

A DNA-interacting indolinobenzodiazepine dimer (IGN) payload was designed with a single reactive imine group towards the goal of eliminating DNA cross-linking and avoiding related toxicities, while conferring a strong binding of the IGN scaffold to duplex DNA. Several IGN ADCs, wherein the payload is linked via a peptide or hindered disulfide, are currently being evaluated in the clinic. In contrast to our lead IGNs, DNA-interacting pyrrolobenzodiazepine (PBD)-based ADC payloads, such as talirine and tesirine, contain two reactive imine groups that can cross-link DNA.

Here, we investigated the mechanism of binding of IGN catabolites with DNA in target cancer cells, and with model duplex DNA or hairpin oligonucleotides. Hairpin and duplex oligonucleotides, designed for high melting temperatures (around 50-60 °C), were custom synthesized with labels. Model IGN catabolites bearing a single imine (mono-imine) or two imine groups (di-imine) were synthesized with a biotin label. Sensitive assays were developed to measure IGN-DNA binding in cells at sub-cytotoxic concentrations (lower than IC50) to allow studies of DNA adduct stability and repair.

The mono-imine IGN molecules bind readily to oligonucleotides, generating stable adducts as determined by gel filtration and reversed phase HPLC analysis. To investigate the binding of unconjugated IGNs with cellular DNA, cancer cells were incubated with mono- and di-imine IGNs for a short-term, followed by wash and further incubation in fresh media. Both mono-and di-imine IGN molecules remained bound to genomic DNA even at 2 days, suggesting a potent interaction with cellular DNA. The time course of binding of IGN to DNA in cells was slower than that observed with model oligonucleotides, as expected because the tightly coiled cellular DNA presumably binds IGN only after the unwinding of DNA during cell cycle or transcription.

Upon DNA cleavage by an added nuclease, free IGN was released from IGN adducts of model oligonucleotides and from genomic DNA of cells that had been treated with unconjugated IGN or IGN ADC. This dissociation of IGN from IGN-DNA adducts only upon cleavage with nuclease suggests that a strong non-covalent interaction between IGN and duplex DNA stabilizes the adduct. The amount of free IGN released from cellular DNA adduct upon nuclease treatment was about 2-fold greater for mono-imine IGN than di-imine IGN, presumably because di-imine IGN was partly cross-linked to cellular DNA. Mono-imine IGN-DNA adducts could potentially be repaired by cellular endonucleases via a DNA cleavage mechanism. In conclusion, the mono-imine IGN payload molecules form highly stable adducts with DNA, which dissociate upon DNA cleavage at physiological temperature.

Citation Format: Rajeeva Singh, Luke Harris, Paulin Salomon, Emily E. Reid, Michael L. Miller, Ravi V. Chari, Thomas A. Keating. Antibody-drug conjugates (ADCs) with indolinobenzodiazepine dimer (IGN) payloads: DNA-binding mechanism of IGN catabolites in target cancer cells [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 230.

Abstract 229: Utilizing a mouse cross-reactive model system to better understand antibody-drug conjugate pharmacokinetics, biodistribution and efficacy

Antibody-drug conjugates (ADCs) are designed to deliver a potent cytotoxic payload directly to tumors, thus limiting exposure in normal tissues. However, target antigen expression on normal tissues can lead to lower systemic ADC exposures, resulting in sub-efficacious concentrations at the tumor site as well as heterogeneous distribution within tumors. Traditional preclinical efficacy studies performed in rodent models using ADCs with non-cross-reactive antibodies have been of limited translational relevance, and a better understanding of the factors that impact ADC dose and activity remains to be elucidated. To examine relationships between variables that could influence ADC efficacy, we generated a cross-reactive model system that utilized a chimeric anti-murine folate receptor α (FRα) antibody (designated rmFR1-12) that binds both mouse and human FRα, and can be conjugated to either maytansinoid (DM) or indolinobenzodiazepine (IGN) payloads. This model system was predicted to have substantial target-mediated drug disposition (TMDD) due to normal tissue expression of FRα. An rmFR1-12-s-SPDB-DM4 ADC was made using tritium-labeled DM4 and administered to tumor-bearing mice in order to assess ADC pharmacokinetics (PK), biodistribution, and efficacy. This approach allowed tracking of the ADC at multiple levels – whole animal, organ, tumor, and cell. Studies were undertaken that assessed the impact of xenograft antigen expression, ADC dose, and ADC drug-to-antibody ratio (DAR) on the PK, biodistribution, and efficacy of the rmFR1-12-s-SPDB-DM4 conjugate. The results showed that TMDD significantly impacted the PK, biodistribution, and activity of the conjugate relative to a non-cross-reactive ADC, with lower ADC doses being more severely impacted than higher doses. Antigen expression positively correlated with local ADC exposure and efficacy. Decreasing the DAR (by co-dosing naked antibody with the ADC) increased systemic exposure. A positive correlation between systemic exposure and dose of naked antibody was observed. Factors that impacted local exposure included: the type of tissue (normal vs tumor), presence or absence of FRα expression on normal tissue, and dose of naked antibody. Of note, in the naked antibody plus ADC study, efficacy did not correlate with local exposure - suggesting that ADC distribution within tumors is of similar importance as the amount of ADC delivered. Overall, these findings underscore the importance of accounting for site, and extent of, normal tissue target expression with respect to ADC PK/PD, and the data generated from these studies are currently being used to build a multiscale physiologically based PK model of a cross-reactive ADC.

Citation Format: Leanne Lanieri, Rassol Laleau, Bahar Matin, Jenny Lee, Steven Boule, Paulin Salomon, Luke Harris, Michael Miller, Nicholas C. Yoder, Yulius Setiady, Neeraj Kohli, Thomas A. Keating, Jan Pinkas, Richard Gregory. Utilizing a mouse cross-reactive model system to better understand antibody-drug conjugate pharmacokinetics, biodistribution and efficacy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 229.

Abstract 224: Antibody-drug conjugates (ADCs) of a new class of N-10 amino linked DNA alkylating indolino-benzodiazepines (IGNs)

We have previously reported on our antibody-drug conjugates (ADCs) that incorporate a highly potent novel DNA alkylating indolino-benzodiazepine (termed IGN) dimer. ADCs of these DNA alkylating IGNs were found preclinically to demonstrate better tolerability and an improved overall therapeutic index (TI) compared with those of the corresponding DNA crosslinking IGNs (Miller, et al., Mol. Cancer Ther. 2016, 2018). IGN-containing ADCs that have reached the clinical stage bear DNA alkylating IGN compounds linked to the antibody via a cleavable linker, incorporated at the amino group of a centrally located anilino spacer moiety. In our ongoing effort to further explore the structure-activity relationship (SAR) of DNA alkylating effector molecules for ADCs, we disclose here a new class of IGNs possessing a self-immolative peptide linker attached at the N-10 amine of the imine-reduced IGN monomer subunit. We explored the impact of modifying the central spacer group connecting the IGN monomers, and site of linkage on in vitro potency and the interaction with DNA. A set of IGN molecules that met our potency criteria were identified and linkable forms of these compounds were prepared and conjugated to a folate receptor-α (FRα)-binding antibody. These ADCs displayed potent, antigen-specific in vitro activity across a panel of FRα-expressing cell lines. A lead ADC selected for in vivo studies demonstrated high stability in plasma and potent efficacy in various xenograft models, at doses well below the maximum tolerated dose. Thus, IGNs modified with an N-10 self-immolative peptide linker are promising DNA alkylating effector molecules for use in ADCs.

Citation Format: Michael L. Miller, Emily E. Reid, Katie E. Archer, Luke Harris, Erin K. Maloney, Laura M. Bartle, Olga Ab, Alan J. Wilhelm, Jose F. Ponte, Rajeeva Singh, Thomas A. Keating, Ravi V. Chari. Antibody-drug conjugates (ADCs) of a new class of N-10 amino linked DNA alkylating indolino-benzodiazepines (IGNs) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr 224.

Abstract 75: Comparison of site-specific and lysine-linked indolino-benzodiazepine antibody-drug conjugates (ADCs)

ADCs are a promising modality for cancer therapy enabled by chemical conjugation of potent cytotoxic compounds to monoclonal antibodies. While many ADCs in clinical evaluation employ heterogeneous conjugation chemistries where the payload is linked through lysine or endogenous cysteine residues, there has recently been considerable interest in site-specific conjugation. ADCs prepared using site-specific methods are believed to have a wider therapeutic index compared to heterogeneous ADCs. We have previously shown that site-specific ADCs incorporating the maytansinoid-based tubulin inhibitor DM1 were less efficacious in vivo when compared to analogous lysine-linked conjugates (Yoder et al AACR 2015 Abstract #545). More recently, we presented results from the evaluation of 2.5-3.0 drug-to-antibody ratio (DAR) heterogeneous lysine-linked and 2 DAR site-specific ADCs using antibodies with engineered cysteines at position 442 in the heavy chains (known as CYSMABTM). These ADCs used the peptide-linked indolino-benzodiazepine DNA alkylator DGN549 (also known as IGN-P1) as the effector. Unlike the DM1 case, site-specific DGN549 ADCs were at least as active in vivo when compared to lysine-linked ADCs (Yoder et al AACR 2016 Abstract #2960). We have made further pharmacological comparisons between CYSMAB and lysine-linked DGN549 ADCs at matched payload doses using two different antibodies targeting distinct cell surface receptors. In the case of mAb1, CYSMAB and lysine-linked ADCs were comparably active. For mAb2, the CYSMAB ADC was more active than the lysine-linked ADC in some models and similarly active in others. The mAb1 CYSMAB ADC exhibited a significantly higher maximum tolerated dose (MTD) compared to the lysine-linked ADC. In contrast, the MTDs of the mAb2 ADCs were similar. However, the mAb2 CYSMAB conjugate was better tolerated in terms of median lethal dose. The mechanism for the improved tolerability of the mAb1 CYSMAB conjugate is not apparent. In an effort to understand whether it is a consequence of conjugation chemistry or DAR, we compared the tolerability of the mAb1 CYSMAB conjugate to that of a ~4 DAR site-specific analog and found that both factors contribute. To determine if our observations can be rationalized in terms of in vivo disposition, we compared the pharmacokinetics of mAb1 CYSMAB and lysine-linked ADCs at matched antibody doses. Intriguingly, the CYSMAB ADC showed slightly greater exposure. These results, along with our previous work on site-specific DM1 ADCs, indicate that in some cases site-specific conjugation can lead to improved efficacy and tolerability. However, generalizations cannot be made across different combinations of antibody, linker, and payload. We conclude that the advantages and disadvantages of site-specific conjugation should be carefully considered for every candidate ADC.

Citation Format: Chen Bai, Nicholas C. Yoder, Alan Wilhelm, Sharlene Adams, Kathleen Whiteman, Jenny Lee, Katie O'Callaghan, Erin Maloney, Manami Shizuka, Yelena Kovtun, Thomas A. Keating. Comparison of site-specific and lysine-linked indolino-benzodiazepine antibody-drug conjugates (ADCs) [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 75. doi:10.1158/1538-7445.AM2017-75

Abstract 2960: Potent in vivo activity of site-specific indolino-benzodiazepine antibody-drug conjugates (ADCs) generated via engineered cysteine conjugation

ADCs are widely studied for cancer therapy, with numerous agents in preclinical and clinical development embodying a wide array of targets, linker chemistries, and cytotoxic effector classes. A fourth element of ADC design that has received much attention recently is the site of conjugation of the cytotoxic molecule to the antibody. Historically, lysine- or interchain cysteine-directed conjugation has been used, but site-specific chemistries have become increasingly popular. Our previous evaluation of site-specific and lysine-linked ADCs utilizing a tubulin-acting maytansinoid effector molecule found the lysine-linked version was more active in vivo (Yoder et al., AACR 2015 #645). Here we present a comparison of engineered cysteine site-specific and lysine-linked ADCs utilizing the previously described indolino-benzodiazepine (henceforth referred to as IGN) effector IGN-P1 (Miller et al., AACR 2015 #652) which is designed to undergo proteolytic cleavage upon cell uptake to release a potently cytotoxic DNA alkylator.

We show that HC-S442C mutants of human IgG1 can be conjugated via maleimide chemistry to IGN-P1 to give stable, potent, and homogeneous ADCs with drug to antibody ratio (DAR) of 2. The in vitro potency of engineered-cysteine IGN-P1 ADCs is largely dependent on the DAR of the ADC, although some difference is observed between HC-S442C and other cysteine mutants used for conjugation.

Pharmacokinetic study of C442 maleimide conjugates suggests that the chemical linkage between effector and antibody is stable upon administration in mice. Further, and in contrast to our previous observations utilizing maytansinoid ADCs, the site-specific and Lys-linked IGN-P1 ADCs showed comparable efficacy in vivo on a molar drug basis. This effect was observed across two different antibodies targeting two different cell surface antigens. These results suggest that, in certain cases, site-specific conjugation chemistry can offer comparable activity to heterogeneous conjugation at well-tolerated doses.

Citation Format: Nicholas C. Yoder, Chen Bai, Alan Wilhelm, Erin K. Maloney, Olga Ab, Emily E. Reid, Manami Shizuka, Daniel Tavares, Rassol Laleau, Xiuxia Sun, Megan E. Bogalhas, Lintao Wang, Jan Pinkas, Michael L. Miller, Ravi Chari, Thomas A. Keating. Potent in vivo activity of site-specific indolino-benzodiazepine antibody-drug conjugates (ADCs) generated via engineered cysteine conjugation. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2960.

Understanding How the Stability of the Thiol-Maleimide Linkage Impacts the Pharmacokinetics of Lysine-Linked Antibody-Maytansinoid Conjugates

Antibody-drug conjugates (ADCs) have become a widely investigated modality for cancer therapy, in part due to the clinical findings with ado-trastuzumab emtansine (Kadcyla). Ado-trastuzumab emtansine utilizes the Ab-SMCC-DM1 format, in which the thiol-functionalized maytansinoid cytotoxic agent, DM1, is linked to the antibody (Ab) via the maleimide moiety of the heterobifunctional SMCC linker. The pharmacokinetic (PK) data for ado-trastuzumab emtansine point to a faster clearance for the ADC than for total antibody. Cytotoxic agent release in plasma has been reported with nonmaytansinoid, cysteine-linked ADCs via thiol-maleimide exchange, for example, brentuximab vedotin. For Ab-SMCC-DM1 ADCs, however, the main catabolite reported is lysine-SMCC-DM1, the expected product of intracellular antibody proteolysis. To understand these observations better, we conducted a series of studies to examine the stability of the thiol-maleimide linkage, utilizing the EGFR-targeting conjugate, J2898A-SMCC-DM1, and comparing it with a control ADC made with a noncleavable linker that lacked a thiol-maleimide adduct (J2898A-(CH2)3-DM). We employed radiolabeled ADCs to directly measure both the antibody and the ADC components in plasma. The PK properties of the conjugated antibody moiety of the two conjugates, J2898A-SMCC-DM1 and J2898A-(CH2)3-DM (each with an average of 3.0 to 3.4 maytansinoid molecules per antibody), appear to be similar to that of the unconjugated antibody. Clearance values of the intact conjugates were slightly faster than those of the Ab components. Furthermore, J2898A-SMCC-DM1 clears slightly faster than J2898A-(CH2)3-DM, suggesting that there is a fraction of maytansinoid loss from the SMCC-DM1 ADC, possibly through a thiol-maleimide dependent mechanism. Experiments on ex vivo stability confirm that some loss of maytansinoid from Ab-SMCC-DM1 conjugates can occur via thiol elimination, but at a slower rate than the corresponding rate of loss reported for thiol-maleimide links formed at thiols derived by reduction of endogenous cysteine residues in antibodies, consistent with expected differences in thiol-maleimide stability related to thiol pKa. These findings inform the design strategy for future ADCs.

Role of the Outer Membrane and Porins in Susceptibility of β-Lactamase-Producing Enterobacteriaceae to Ceftazidime-Avibactam

This study examined the activity of the novel antimicrobial combination ceftazidime-avibactam against Enterobacteriaceae exhibiting different outer membrane permeability profiles, specifically with or without porins and with or without expression of the main efflux pump (AcrAB-TolC). The addition of the outer membrane permeabilizer polymyxin B nonapeptide increased the antibacterial activities of avibactam alone, ceftazidime alone, and ceftazidime-avibactam against the characterized clinical isolates of Escherichia coli, Enterobacter aerogenes, and Klebsiella pneumoniae. This enhancement of activities was mainly due to increased passive penetration of compounds since inhibition of efflux by the addition of phenylalanine-arginine β-naphthylamide affected the MICs minimally. OmpF (OmpK35) or OmpC (OmpK36) pores were not the major route by which avibactam crossed the outer membranes of E. coli and K. pneumoniae. In contrast, Omp35 and Omp36 allowed diffusion of avibactam across the outer membrane of E. aerogenes, although other diffusion channels for avibactam were also present in that species. It was clear that outer membrane permeability and outer membrane pore-forming proteins play a key role in the activity of ceftazidime-avibactam. Nevertheless, the MICs of ceftazidime-avibactam (with 4 mg/liter avibactam) against the ceftazidime-resistant clinical isolates of the three species of Enterobacteriaceae studied were ≤ 8 mg/liter, regardless of outer membrane permeability changes resulting from an absence of defined porin proteins or upregulation of efflux.

Abstract 4531: Effects of drug load on therapeutic index for antibody-maytansinoid conjugates

Antibody-drug conjugates (ADCs) are being actively pursued as a new modality to treat cancer following the regulatory approval of Adcetris@ and Kadcyla@. ADCs consist of a cytotoxic agent, or drug, conjugated to a targeting antibody (Ab) through a linker. The two approved ADCs (and most ADCs now in the clinic) are heterogeneous conjugates with an average molar drug to Ab ratio (DAR) of 3-4 (potentially ranging from 0-8 for individual molecules). To understand the effects of different DAR ranges on the preclinical properties of ADCs using a maytansinoid cytotoxic agent, we prepared a series of conjugates with a cleavable linker (M9346A-sulfo-SPDB-DM4 targeting folate receptor alpha) or an uncleavable linker (J2898A-SMCC-DM1 targeting epidermal growth factor receptor) with varying DAR, and evaluated their biochemical characteristics, in vivo stability, efficacy, and tolerability.

Both M9346A-sulfo-SPDB-DM4 and J2898A-SMCC-DM1 conjugates with low (average ∼2, range 0-4) to very high DAR (average 10, range 7-14) were prepared in good yield, high monomer content and low free drug levels. At constant Ab concentration, the in vitro potency consistently increased with increasing DAR. We then characterized the in vivo disposition of these ADCs. First, pharmacokinetic analysis showed that conjugates with an average DAR below ∼6 had comparable clearance rates, but at an average DAR around 9-10 rapid clearance was observed. Biodistribution studies in mice showed that these 9-10 DAR conjugates rapidly distribute to the liver, with the maximum%ID/g for this organ at 24-28% compared with 7-10% for lower DAR conjugates (all at 2-6 h post-injection). We further studied the efficacy of these ADCs in mouse xenograft models at both constant antibody and constant maytansinoid doses. At constant Ab dose, conjugates with average DAR ∼6 are more efficacious than conjugates with DAR ∼3. However, DAR 9-10 conjugates are either less or similarly active compared to DAR 6 conjugates, consistent with the observed rapid clearance. At constant maytansinoid dose, conjugates with 2 - 6 DAR show similar efficacy, while those with 9-10 DAR were again less active. Tolerability studies monitoring body weight loss show that conjugates with different DAR, even as high as 9-10, have comparable tolerability based on total administered maytansinoid dose.

In summary, our preclinical findings on tolerability and efficacy suggest that maytansinoid conjugates with DAR in the range between 2 - 6 have better therapeutic index than conjugates with very high DAR (∼9-10). These very high DAR ADCs suffer from decreased efficacy likely due to faster clearance. These results support the use of DAR 3-4 for maytansinoid ADCs, but suggest the exploration of lower or higher DAR depending on the biology of the target antigen.

Citation Format: Xiuxia Sun, Jose F. Ponte, Nicholas C. Yoder, Jennifer Coccia, Leanne Lanieri, Rassol Laleau, Qifeng Qiu, Rui Wu, Erica Hong, Megan Bogalhas, Lintao Wang, Erin K. Maloney, Olga Ab, Hans K. Erickson, Thomas A. Keating, Ravi Chari, John M. Lambert. Effects of drug load on therapeutic index for antibody-maytansinoid conjugates. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4531. doi:10.1158/1538-7445.AM2015-4531

Abstract 645: Stability and efficacy comparison of site-specific and lysine-linked maytansinoid antibody-drug conjugates

As with any therapeutic molecule, antibody-drug conjugates (ADCs) exhibit structure-activity relationships, and medicinal chemistry efforts in this field strive to optimize structure to give the maximum therapeutic index. Recent interest in ADCs as cancer therapy has led to a number of different combinations of linker, payload, and conjugation chemistry.In particular, site-specific methods of payload conjugation have been suggested to generally improve therapeutic properties as compared with more established approaches directed toward lysines or endogenous cysteines.We have investigated the preparation, stability, and activity of anti-folate receptor alpha (FRα) ADCs carrying the microtubule inhibitor, DM1, and conjugated to engineered cysteine mutants utilizing different sites, and compared these ADCs with lysine-directed heterogeneous conjugates. In both embodiments, the DM1 is linked with a protease-cleavable linker. We show that highly homogeneous DM1 ADCs can be produced using engineered cysteine chemistry, enabling assessment of the effects of site-specific conjugation in cells and in animal models. We find that in vitro potency of both lysine-linked and engineered cysteine-linked ADCs against FRα-positive KB cells scales with the total DM1 delivered to cells. Buffer stability experiments in the presence of excess thiol suggest that most engineered cysteine conjugates are comparable in stability to the lysine-linked ADC. A notable exception shows about twice as much fractional DM1 loss upon 3 days of incubation as the other conjugates. Comparison of in vivo activity of two site-specific DM1 ADCs in a KB xenograft model shows measurable activity differences between different conjugation sites. However, a lysine-linked conjugate using almost identical linker chemistry shows approximately 2-fold superior activity to either site-specific construct on a molar DM1 basis. We conclude that, while site-specific conjugation of ADCs may provide a benefit in certain contexts, in other contexts, it may lead to decreased activity, such as in the anti-FRα/KB model examined here. We also observe that different conjugation sites may offer significant differences in activity. It is therefore advisable to evaluate each unique combination of payload, linker, drug:antibody ratio, conjugation site(s), and antibody to the maximum extent possible.

Citation Format: Nicholas C. Yoder, Chen Bai, Daniel Tavares, Wayne C. Widdison, Olga Ab, Kathleen R. Whiteman, Alan Wilhelm, Erin K. Maloney, Hans K. Erickson, Thomas A. Keating. Stability and efficacy comparison of site-specific and lysine-linked maytansinoid antibody-drug conjugates. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 645. doi:10.1158/1538-7445.AM2015-645

Abstract 4504: Plasma pharmacokinetics and tumor accumulation in mice of IMGN779, an antibody-drug conjugate for acute myeloid leukemia

IMGN779, an antibody-drug conjugate (ADC) consisting of the anti-CD33 antibody, Z4681A, linked to the potent DNA-alkylating agent, DGN462, via a charged disulfide linker, sulfo-SPDB, is in development for the treatment of acute myeloid leukemia (AML). IMGN779 is highly active in vitro against AML cell lines and primary patient AML cells and causes complete regression of AML xenograft tumors at non-toxic doses in vivo.

To understand the fate of IMGN779 upon uptake, we performed metabolism studies on cultures of ADC-treated AML cells in vitro, and distribution studies in vivo, following the plasma clearance and tumor accumulation of the ADC in mice. The antigen-mediated binding, uptake, and degradation of IMGN779 by AML cells in culture were measured using a radiolabeled conjugate. CD33-targeted degradation of IMGN779 was observed in treated cultures, with metabolites detected within the cells, some as DNA-adducts, and in the media following efflux from cells. Approximately 40% of initially-bound IMGN779 was measured as protein-free degraded species after 22 h incubation at 37 °C, with an additional ∼40% measured as DNA-associated species.

To evaluate plasma pharmacokinetics (PK), mice were injected with [3H]propionate radiolabeled Z4681A antibody (Ab), or with IMGN779 that was radiolabeled either on the Ab portion of the ADC ([3H]propionate) or on the DGN462 moiety. Total Ab and intact ADC concentrations were determined in plasma samples collected from 2 min to 28 days post-injection. Overlapping clearance profiles for Z4681A and Ab-labeled IMGN779 were observed with half-lives (t½) of ∼18 d, which demonstrated that conjugation with DGN462 did not alter Ab PK. A faster clearance of [3H]DGN462-labeled IMGN779 was observed, with a plasma t½ of 4.5 d, indicating that DGN462 is released from IMGN779 in circulation. Plasma samples were CD33-affinity captured and analyzed by LC/MS to determine the mass-distribution profile of IMGN779. A decrease in relative abundances of antibody species associated with a high number of DGN462 molecules from 2 min- 3 d supported the plasma clearance results.

To determine the tumor localization of IMGN779 and its metabolites, the radioactivity in excised HL60/QC xenografts dosed with 5 mg/kg [3H]IMGN779 or non-targeting [3H]ADC was measured at 6, 24 and 48 h after treatment. The total amount of ADC (intact and metabolized) present was similar, but slightly higher for IMGN779 versus non-targeting ADC samples. However, the amount of both extractable metabolites and DNA-bound species was higher in tumor samples from mice treated with IMGN779 versus non-targeting ADC, indicating CD33-targeted metabolism in the tumor.

From these studies, we conclude IMGN779 demonstrates CD33 target-mediated generation of DGN462 metabolites, and also exhibits DNA-modification consistent with the mechanism of action of its effector molecule, DGN462.

Citation Format: Katharine C. Lai, Prerak Shah, Surina Sikka, XiuXia Sun, Rassol LaLeau, Kathleen R. Whiteman, Holly Johnson-Modafferi, Alan Wilhelm, Charlene Audette, Lintao Wang, Megan E. Bogalhas, Thomas A. Keating, Ravi Chari. Plasma pharmacokinetics and tumor accumulation in mice of IMGN779, an antibody-drug conjugate for acute myeloid leukemia. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 4504. doi:10.1158/1538-7445.AM2015-4504

Abstract 5442: Microscale methods for preparation and screening of antibody-drug conjugates

Antibody-drug conjugates (ADCs) are hybrid biotherapeutics that combine the targeting specificity of monoclonal antibodies with chemically conjugated, highly potent small molecule chemotherapeutics. Using established research scale approaches, the amount of antibody material needed to prepare candidate ADCs far exceeds the quantities required for initial in vitro screening. The need to scale up production across many antibodies slows down early lead selection efforts and wastes material. We have therefore developed methods for conjugating multiple antibodies with ADC payloads in parallel at the 50-150 μg scale in 96-well plates. Pilot reactions show that antibodies can be titrated to different final drug:antibody ratios (DARs) with different payloads, and that differences in pH can alter the reaction kinetics with useful effects. We show that 96-well centrifugal ultrafiltration enables highly parallel ADC purification while maintaining the rigorous removal of residual cytotoxic impurities observed with more established methods such as gel filtration chromatography. In addition, we describe characterization of microscale ADCs using a single chromatographic assay requiring ∼5 μg of material. The resulting platform reduces the required input quantity of antibody required for in vitro ADC screening by at least 5-10 fold. It also enables much higher conjugation throughput with concomitant decrease in time needed to generate and characterize ADCs.

To assess the utility of the platform for ranking candidate antibodies, we compare in vitro cytotoxicity results for a panel of ADCs produced by both microscale and research scale methods. We also present three example screens in which antibody libraries of 10-85 members against different targets were conjugated using microscale methods and the resulting ADCs ranked by in vitro potency. For each antibody library, the screen size, conjugation conditions, and target DAR range were adjusted to suit the target biology, antibody type,and payload class. For example, we present evidence suggesting that, for targets in which functional antibody activity is not observed, normalization of DAR to the 2-6 range is adequate for screening. Across the screens, the success rates for producing ADC in quantity and quality suitable for screening were in the 75-90% range, using 200-600 μg of input antibody. Cytotoxic potencies ranging over 2-3 orders of magnitude were observed in the resulting ADC libraries, suggesting that microscale conjugation can rapidly focus ADC discoverycampaigns on high potency molecules.

For early stage antibody and ADC screening efforts, we find that microscale conjugation methods yield ADCs that can substitute for traditionally prepared conjugates. We expect these methods will be applicable across many different ADC targets and payloads, and possibly applicable more generally to conjugated macromolecule therapeutic or diagnostic reagents.

Citation Format: Nicholas C. Yoder, Kalli C. Catcott, Molly A. McShea, Carl Uli Bialucha, Parmita Saxena, Chen Bai, Kathy L. Miller, Thomas G. Gesner, Mikias Woldegiorgis, Stuart W. Hicks, Megan E. Lewis, Michael S. Fleming, Hans K. Erickson, Seth E. Ettenberg, Thomas A. Keating. Microscale methods for preparation and screening of antibody-drug conjugates. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr 5442. doi:10.1158/1538-7445.AM2015-5442

Antibacterial inhibitors of Gram-positive thymidylate kinase: structure-activity relationships and chiral preference of a new hydrophobic binding region

Thymidylate kinase (TMK), an essential enzyme in bacterial DNA biosynthesis, is an attractive therapeutic target for the development of novel antibacterial agents, and we continue to explore TMK inhibitors with improved potency, protein binding, and pharmacokinetic potential. A structure-guided design approach was employed to exploit a previously unexplored region in Staphylococcus aureus TMK via novel interactions. These efforts produced compound 39, with 3 nM IC50 against S. aureus TMK and 2 μg/mL MIC against methicillin-resistant S. aureus (MRSA). This compound exhibits a striking inverted chiral preference for binding relative to earlier compounds and also has improved physical properties and pharmacokinetics over previously published compounds. An example of this new series was efficacious in a murine S. aureus infection model, suggesting that compounds like 39 are options for further work toward a new Gram-positive antibiotic by maintaining a balance of microbiological potency, low clearance, and low protein binding that can result in lower efficacious doses.

In vivo validation of thymidylate kinase (TMK) with a rationally designed, selective antibacterial compound

There is an urgent need for new antibacterials that pinpoint novel targets and thereby avoid existing resistance mechanisms. We have created novel synthetic antibacterials through structure-based drug design that specifically target bacterial thymidylate kinase (TMK), a nucleotide kinase essential in the DNA synthesis pathway. A high-resolution structure shows compound TK-666 binding partly in the thymidine monophosphate substrate site, but also forming new induced-fit interactions that give picomolar affinity. TK-666 has potent, broad-spectrum Gram-positive microbiological activity (including activity against methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus), bactericidal action with rapid killing kinetics, excellent target selectivity over the human ortholog, and low resistance rates. We demonstrate in vivo efficacy against S. aureus in a murine infected-thigh model. This work presents the first validation of TMK as a compelling antibacterial target and provides a rationale for pursuing novel clinical candidates for treating Gram-positive infections through TMK.

In vitro validation of acetyltransferase activity of GlmU as an antibacterial target in Haemophilus influenzae

GlmU is a bifunctional enzyme that is essential for bacterial growth, converting D-glucosamine 1-phosphate into UDP-GlcNAc via acetylation and subsequent uridyl transfer. A biochemical screen of AstraZeneca's compound library using GlmU of Escherichia coli identified novel sulfonamide inhibitors of the acetyltransferase reaction. Steady-state kinetics, ligand-observe NMR, isothermal titration calorimetry, and x-ray crystallography showed that the inhibitors were competitive with acetyl-CoA substrate. Iterative chemistry efforts improved biochemical potency against gram-negative isozymes 300-fold and afforded antimicrobial activity against a strain of Haemophilus influenzae lacking its major efflux pump. Inhibition of precursor incorporation into bacterial macromolecules was consistent with the antimicrobial activity being caused by disruption of peptidoglycan and fatty acid biosyntheses. Isolation and characterization of two different resistant mutant strains identified the GlmU acetyltransferase domain as the molecular target. These data, along with x-ray co-crystal structures, confirmed the binding mode of the inhibitors and explained their relative lack of potency against gram-positive GlmU isozymes. This is the first example of antimicrobial compounds mediating their growth inhibitory effects specifically via GlmU.

Correction for Interference by Test Samples in High-Throughput Assays

In high-throughput biochemical assays performed in multiwell plates, the effect of test samples on the activity of the biochemical system is usually measured by optical means such as absorbance, fluorescence, luminescence, or scintillation counting. The test sample often causes detection interference when it remains in the well during the measurement. Interference may be due to light absorption, fluorescence quenching, sample fluorescence, chemical interaction of the sample with a detection reagent, or depression of the meniscus. A simple method is described that corrects for such interference well by well. The interference is measured in a separate artifact assay plate. An appropriate arithmetic correction is then applied to the measurement in the corresponding well of the activity assay plate. The correction procedure can be used for single-point screening or potency measurements on serial dilutions of test samples.

Exploitation of structural and regulatory diversity in glutamate racemases

Glutamate racemase is an enzyme essential to the bacterial cell wall biosynthesis pathway, and has therefore been considered as a target for antibacterial drug discovery. We characterized the glutamate racemases of several pathogenic bacteria using structural and biochemical approaches. Here we describe three distinct mechanisms of regulation for the family of glutamate racemases: allosteric activation by metabolic precursors, kinetic regulation through substrate inhibition, and D-glutamate recycling using a d-amino acid transaminase. In a search for selective inhibitors, we identified a series of uncompetitive inhibitors specifically targeting Helicobacter pylori glutamate racemase that bind to a cryptic allosteric site, and used these inhibitors to probe the mechanistic and dynamic features of the enzyme. These structural, kinetic and mutational studies provide insight into the physiological regulation of these essential enzymes and provide a basis for designing narrow-spectrum antimicrobial agents.

The structure of VibH represents nonribosomal peptide synthetase condensation, cyclization and epimerization domains

Nonribosomal peptide synthetases (NRPSs) are large, multidomain enzymes that biosynthesize medically important natural products. We report the crystal structure of the free-standing NRPS condensation (C) domain VibH, which catalyzes amide bond formation in the synthesis of vibriobactin, a Vibrio cholerae siderophore. Despite low sequence identity, NRPS condensation enzymes are structurally related to chloramphenicol acetyltransferase (CAT) and dihydrolipoamide acyltransferases. However, although the latter enzymes are homotrimers, VibH is a monomeric pseudodimer. The VibH structure is representative of both NRPS condensation and epimerization domains, as well as the condensation-variant cyclization domains, which are all expected to be monomers. Surprisingly, despite favorable positioning in the active site, a universally conserved histidine important in CAT and in other C domains is not critical for general base catalysis in VibH.

Yersiniabactin synthetase: a four-protein assembly line producing the nonribosomal peptide/polyketide hybrid siderophore of Yersinia pestis

Yersiniabactin synthetase comprises four proteins, YbtE, HMWP1, HMWP2, and YbtU, encompassing seventeen functional domains, twelve catalytic and five carrier, to select, activate, and incorporate salicylate, three cysteines, and one malonyl moiety into the iron chelator yersiniabactin (Ybt). In the present study, yersiniabactin has been reconstituted in vitro from the 4 protein assembly line by the use of eight biosynthetic precursors. The rate of one turnover, comprising 22 chemical operations performed by the assembly line to release the completed Ybt molecule, was determined at 1.4 min(-1). During the course of Ybt production, the elongating acyl-S-enzyme chain was shown to transfer across a nonribosomal peptide synthetase/polyketide synthase (NRPS/PKS) interprotein interface and then a PKS/NRPS intraprotein interface. This study on the Ybt synthetase assembly line represents the first complete in vitro reconstitution of a nonribosomal peptide/polyketide hybrid system.

Tailoring enzymes that modify nonribosomal peptides during and after chain elongation on NRPS assembly lines

Nonribosomal peptide synthetases are large enzyme complexes that synthesize a variety of peptide natural products through a thiotemplated mechanism. Assembly of the peptides proceeds through amino acid loading, amide-bond formation and chain translocation, and finally thioester lysis to release the product. The final products are often heavily modified, however, through methylation, epimerization, hydroxylation, heterocyclization, oxidative cross-linking and attachment of sugars. These activities are the province of specialized enzymes (either embedded in the multidomain nonribosomal peptide synthetase structure or standalone).

Heterocycle formation in vibriobactin biosynthesis: alternative substrate utilization and identification of a condensed intermediate

The iron-chelating peptide vibriobactin of the pathogenic Vibrio cholerae is assembled by a four-subunit nonribosomal peptide synthetase complex, VibE, VibB, VibH, and VibF, using 2,3-dihydroxybenzoate and L-threonine as precursors to two 2,3-dihydroxyphenyl- (DHP-) methyloxazolinyl groups in amide linkage on a norspermidine scaffold. We have tested the ability of the six-domain VibF subunit (Cy-Cy-A-C-PCP-C) to utilize various L-threonine analogues and found the beta-functionalized amino acids serine and cysteine can function as alternate substrates in aminoacyl-AMP formation (adenylation or A domain), aminoacyl-S-enzyme formation (A domain), acylation by 2,3-dihydrobenzoyl- (DHB-) S-VibB (heterocyclization or Cy domain), heterocyclization to DHP-oxazolinyl- and DHP-thiazolinyl-S-enzyme forms of VibF (Cy domain) as well as transfer to DHB-norspermidine at both N(5) and N(9) positions (condensation or C domain) to make the bis(oxazolinyl) and bis(thiazolinyl) analogues of vibriobactin. When L-threonyl-S-pantetheine or L-threonyl-S-(N-acetyl)cysteamine was used as a small-molecule thioester analogue of the threonyl-S-VibF acyl enzyme intermediate, the Cy domain(s) of a CyCyA fragment of VibF generated DHB-threonyl-thioester products of the condensation step but not the methyloxazolinyl thioesters of the heterocyclization step. This clean separation of condensation from cyclization validates a two-stage mechanism for threonyl, seryl, and cysteinyl heterocyclization domains in siderophore and antibiotic synthetases. Full heterocyclization activity could be restored by providing CyCyA with the substrate L-threonyl-S-peptidyl carrier protein (PCP)-C2, suggesting an important role for the protein scaffold component of the heterocyclization acceptor substrate. We also examined heterocyclization donor substrate specificity at the level of acyl group and protein scaffold and observed intolerance for substitution at either position.

Reconstitution and characterization of the Vibrio cholerae vibriobactin synthetase from VibB, VibE, VibF, and VibH

Vibriobactin [N(1)-(2,3-dihydroxybenzoyl)-N(5),N(9)-bis[2-(2, 3-dihydroxyphenyl)-5-methyloxazolinyl-4-carboxamido]norspermidine] , is an iron chelator from the cholera-causing bacterium Vibrio cholerae. The six-domain, 270 kDa nonribosomal peptide synthetase (NRPS) VibF, a component of vibriobactin synthetase, has been heterologously expressed in Escherichia coli and purified. VibF has an unusual NRPS domain organization: cyclization-cyclization-adenylation-condensation-peptidyl carrier protein-condensation (Cy(1)-Cy(2)-A-C(1)-PCP-C(2)). VibF activates and covalently loads its PCP with L-threonine, and together with vibriobactin synthetase proteins VibE (adenylation) and VibB (aryl carrier protein) condenses and heterocyclizes 2, 3-dihydroxybenzoyl-VibB with L-Thr to 2-dihydroxyphenyl-5-methyloxazolinyl-4-carboxy-VibF in the first demonstration of oxazoline formation by an NRPS cyclization domain. This enzyme-bound aryl oxazoline can be transferred by VibF to various amine acceptors but most efficiently to N(1)-(2, 3-dihydroxybenzoyl)norspermidine (k(cat) = 122 min(-1), K(m) = 1.7 microM), the product of 2,3-dihydroxybenzoyl-VibB, norspermidine, and VibH. This diacylated product undergoes a second aryl oxazoline acylation on its remaining secondary amine, also catalyzed by VibF, to yield vibriobactin. Vibriobactin biosynthesis in vitro has thus been accomplished from four proteins, VibE, VibB, VibF, and VibH, with the substrates 2,3-dihydroxybenzoic acid, L-Thr, norspermidine, and ATP. Vibriobactin synthetase is an unusual NRPS in that all intermediates are not covalently tethered as PCP thioesters and in that it represents an NRPS pathway with two branch points.

Vibriobactin biosynthesis in Vibrio cholerae: VibH is an amide synthase homologous to nonribosomal peptide synthetase condensation domains

The Vibrio cholerae siderophore vibriobactin is biosynthesized from three molecules of 2,3-dihydroxybenzoate (DHB), two molecules of L-threonine, and one of norspermidine. Of the four genes positively implicated in vibriobactin biosynthesis, we have here expressed, purified, and assayed the products of three: vibE, vibB, and vibH. All three are homologous to nonribosomal peptide synthetase (NRPS) domains: VibE is a 2,3-dihydroxybenzoate-adenosyl monophosphate ligase, VibB is a bifunctional isochorismate lyase-aryl carrier protein (ArCP), and VibH is a novel amide synthase that represents a free-standing condensation (C) domain. VibE and VibB are homologous to EntE and EntB from Escherichia coli enterobactin synthetase; VibE activates DHB as the acyl adenylate and then transfers it to the free thiol of the phosphopantetheine arm of VibB's ArCP domain. VibH then condenses this DHB thioester (the donor) with the small molecule norspermidine (the acceptor), forming N(1)-(2, 3-dihydroxybenzoyl)norspermidine (DHB-NSPD) with a k(cat) of 600 min(-1) and a K(m) for acyl-VibB of 0.88 microM and for norspermidine of 1.5 mM. Exclusive monoacylation of a primary amine of norspermidine was observed. VibH also tolerates DHB-acylated EntB and 1,7-diaminoheptane, octylamine, and hexylamine as substrates, albeit at lowered catalytic efficiencies. DHB-NSPD possesses one of three acylations required for mature vibriobactin, and its formation confirms VibH's role in vibriobactin biosynthesis. VibH is a unique NRPS condensation domain that acts upon an upstream carrier-protein-bound donor and a downstream amine, turning over a soluble amide product, in contrast to an archetypal NRPS-embedded C domain that condenses two carrier protein thioesters.

Expression, purification, and characterization of HMWP2, a 229 kDa, six domain protein subunit of Yersiniabactin synthetase

The six domain, 229 kDa HMWP2 subunit of the Yersinia pestis yersiniabactin (Ybt) synthetase has been expressed in soluble, full-length form in E. coli as a C-terminal His8 construct at low growth temperatures and with attenuated induction. All six domains of this nonribosomal peptide synthetase subunit, three phosphopantetheinylatable carrier protein domains (ArCP, PCP1, PCP2), one adenylation (A) domain, and two cyclization domains (Cy1, Cy2), have been assayed and are functional. Mutants that convert the phosphopantetheinylatable serine residue to alanine in each of the carrier protein domains accumulate acyl-S-enzyme intermediates upstream of the blocked apo carrier protein site. The ArCP mutant cannot be salicylated by the adenylation protein YbtE; the PCP1 mutant releases salicyl-cysteine from thiolysis of the Sal-S-ArCP intermediate; and the PCP2 mutant releases hydroxyphenyl-thiazolinyl-cysteine from the HPT-S-PCP1 acyl enzyme intermediate, all of which demonstrates processivity and directionality of chain growth. Restoration of the ArCP mutant's function was accomplished with the native ArCP fragment added in trans. The wild-type HMWP2 subunit accumulates hydroxyphenyl-4, 2-bithiazolinyl-S-enzyme at its most downstream PCP2 carrier site, presumably for transfer to the next subunit, HMWP1. The A domain was found to activate and transfer to PCP1 and PCP2 not only the natural L-Cys but also S-2-aminobutyrate, L-beta-chloroalanine, and L-Ser, enabling testing of the substrate specificity of the Cy domain. Probes of Cy domain function include mutagenesis of the Cy1 domain's conserved signature motif DX(4)DX(2)S to show that both D residues but not the S are crucial for both amide bond formation and heterocyclization. Also the Cy1 domain would accept an alternate upstream electrophilic donor substrate (2,3-dihydroxybenzoyl-S-ArCP) but would not process any of the three alternate downstream nucleophilic acceptors in place of Cys-S-PCP1, even for the amide bond-forming step in chain elongation.

Selectivity of the yersiniabactin synthetase adenylation domain in the two-step process of amino acid activation and transfer to a holo-carrier protein domain

The adenylation (A) domain of the Yersinia pestis nonribosomal peptide synthetase that biosynthesizes the siderophore yersiniabactin (Ybt) activates three molecules of L-cysteine and covalently aminoacylates the phosphopantetheinyl (P-pant) thiols on three peptidyl carrier protein (PCP) domains embedded in the two synthetase subunits, two in cis (PCP1, PCP2) in subunit HMWP2 and one in trans (PCP3) in subunit HMWP1. This two-step process of activation and loading by the A domain is analogous to the operation of the aminoacyl-tRNA synthetases in ribosomal peptide synthesis. Adenylation domain specificity for the first step of reversible aminoacyl adenylate formation was assessed with the amino acid-dependent [(32)P]-PP(i)-ATP exchange assay to show that S-2-aminobutyrate and beta-chloro-L-alanine were alternate substrates. The second step of A domain catalysis, capture of the bound aminoacyl adenylate by the P-pant-SH of the PCP domains, was assayed both by catalytic release of PP(i) and by covalent aminoacylation of radiolabeled substrates on either the PCP1 fragment of HMWP2 or the PCP3-thioesterase double domain fragment of HMWP1. There was little selectivity for capture of each of the three adenylates by PCP3 in the second step, arguing against any hydrolytic proofreading of incorrect substrates by the A domain. The holo-PCP3 domain accelerated PP(i) release and catalytic turnover by 100-200-fold over the leak rate (<1 min(-1)) of aminoacyl adenylates into solution while PCP1 in trans had only about a 5-fold effect. Free pantetheine could capture cysteinyl adenylate with a 25-50-fold increase in k(cat) while CoA was 10-fold less effective. The K(m) of free pantetheine (30-50 mM) was 3 orders of magnitude larger than that of PCP3-TE (10-25 microM), indicating a net 10(4) greater catalytic efficiency for transfer to the P-pant arm of PCP3 by the Ybt synthetase A domain, relative to P-pant alone.

Assembly of the Pseudomonas aeruginosa nonribosomal peptide siderophore pyochelin: In vitro reconstitution of aryl-4, 2-bisthiazoline synthetase activity from PchD, PchE, and PchF

Three Pseudomonas aeruginosa proteins involved in biogenesis of the nonribosomal peptide siderophore pyochelin, PchD, PchE, and PchF, have been expressed in and purified from Escherichia coli and are found to produce the tricyclic acid hydroxyphenyl-thiazolyl-thiazolinyl-carboxylic acid (HPTT-COOH), an advanced intermediate containing the aryl-4,2-bis-heterocyclic skeleton of the bithiazoline class of siderophores. The three proteins contain three adenylation domains, one specific for salicylate activation and two specific for cysteine activation, and three carrier protein domains (two in PchE and one in PchF) that undergo posttranslational priming with phosphopantetheine to enable covalent tethering of salicyl and cysteinyl moieties as acyl-S-enzyme intermediates. Two cyclization domains (Cy1 in PchE and Cy2 in PchF) create the two amide linkages in the elongating chains and the cyclodehydrations of acylcysteine moieties into thiazolinyl rings. The ninth domain, the most downstream domain in PchF, is the chain-terminating, acyl-S-enzyme thioester hydrolase that releases the HPTT-S-enzyme intermediate to the observed tandem bis-heterocyclic acid product. A PchF-thioesterase domain active site double mutant fails to turn over, but a monocyclic hydroxyphenyl-thiazolinyl-cysteine (HPT-Cys) product continues to be released from PchE, allowing assignment of the cascade of acyl-S-enzyme intermediates involved in initiation, elongation, and termination steps.

Initiation, elongation, and termination strategies in polyketide and polypeptide antibiotic biosynthesis

Progress in sequence analysis of biosynthetic gene clusters encoding polyketides and nonribosomal peptides and in the reconstitution of in vitro activities continues to reveal new insights into the growth of these natural products' acyl chains, which have been revealed as a series of elongating, covalent, acyl enzyme intermediates on their multimodular scaffolds. Studies that focus on the three stages of natural product biosynthesis - initiation, elongation, and termination - have yielded crucial information on monomer substrate specificity, domain and module portability, and product release mechanisms, all of which are important not only for an understanding of this exquisite enzymatic machinery, but also for the rational construction of new, functional synthetases and synthases that are a goal of combinatorial biosynthesis.

Identification of a Mycobacterium tuberculosis gene cluster encoding the biosynthetic enzymes for assembly of the virulence-conferring siderophore mycobactin

BACKGROUND

Many pathogenic bacteria secrete iron-chelating siderophores as virulence factors in the iron-limiting environments of their vertebrate hosts to compete for ferric iron. Mycobacterium tuberculosis mycobactins are mixed polyketide/nonribosomal peptides that contain a hydroxyaryloxazoline cap and two N-hydroxyamides that together create a high-affinity site for ferric ion. The mycobactin structure is analogous to that of the yersiniabactin and vibriobactin siderophores from the bacteria that cause plague and cholera, respectively.

RESULTS

A ten-gene cluster spanning 24 kilobases of the M. tuberculosis genome, designated mbtA-J, contains the core components necessary for mycobactin biogenesis. The gene products MbtB, MbtE and MbtF are proposed to be peptide synthetases, MbtC and MbtD polyketide synthases, MbtI an isochorismate synthase that provides a salicylate activated by MbtA, and MbtG a required hydroxylase. An aryl carrier protein (ArCP) domain is encoded in mbtB, and is probably the site of siderophore chain initiation. Overproduction and purification of the mbtB ArCP domain and MbtA in Escherichia coli allowed validation of the mycobactin initiation hypothesis, as sequential action of PptT (a phosphopantetheinyl transferase) and MbtA (a salicyl-AMP ligase) resulted in the mbtB ArCP domain being activated as salicyl-S-ArCP.

CONCLUSIONS

Mycobactins are produced in M. tuberculosis using a polyketide synthase/nonribosomal peptide synthetase strategy. The mycobactin gene cluster has organizational homologies to the yersiniabactin and enterobactin synthetase genes. Enzymatic targets for inhibitor design and therapeutic intervention are suggested by the similar ferric-ion ligation strategies used in the siderophores from Mycobacteria, Yersinia and E. coli pathogens.

A Remarkable Two-Step Synthesis of Diverse 1,4-Benzodiazepine-2,5-diones Using the Ugi Four-Component Condensation

A two-step, general synthesis of 1,4-benzodiazepine-2,5-diones (BZDs) is presented. This synthesis employs an Ugi four-component condensation using a convertible isocyanide (1-isocyanocyclohexene), followed by an acid-activated cyclization reaction. This synthesis represents a dramatically improved route to BZDs over those currently in the literature. In addition, since amino acids are not used as inputs, the potential for molecular diversity is much greater than that with existing syntheses. It was also found that BZDs substituted with methylenes at the C-3 and N-4 positions display conformational isomerism in the NMR spectra at room temperature. Variable-temperature NMR experiments support this observation and offer the interesting conclusion that the BZD core structure, in certain examples, might not be as rigid as previously supposed.