Skip to main content

"Generation and Evaluation of ""Next-generation"" Antibody-Toxin-Conjugates for Cancer Therapy"

Final Report Summary - MABTOX (Generation and Evaluation of "Next-generation" Antibody-Toxin-Conjugates for Cancer Therapy)

Executive Summary:
The SME and RTD partners in the Research for SMEs Project MABTOX had the objective to develop and to validate two different and novel enzymatic conjugation technologies for the generation of site-specifically conjugated "next-generation" Antibody-Drug-Conjugates (ADCs) for safer and more effective therapy of cancer. At the beginning of the project, the three leading SMEs had only the concept of using enzymatic approaches, involving on one hand sortase enzymes (background technology of NBE-Therapeutics) and, on the other hand, split-inteins (background technology of ERA Biotech) to generate site-specifically conjugated ADCs. Traditionally, ADCs are produced by chemical coupling of small molecular weight toxins to antibodies using specific linker chemistry that either couples the toxins to free lysine or cysteine residues of an antibody. This chemical conjugation is a random process and leads to heterogeneous ADCs with variable number of toxins being conjugated to the antibody and therefore results in drugs that are composed of mixtures of ADC molecules. Depending on the conjugation sites, the properties and efficacies of these chemically conjugated ADCs differ, which represents a challenge for developing safe and effective ADCs. By enzymatic conjugation the specificity of enzymes are exploited to precisely couple the toxin to defined sites in the ADC. This requires that both the antibody and the toxin are modified with “handles” for enzyme recognition. In case of sortase-enzyme mediated antibody conjugation (NBE’s SMAC-Technology), the toxins need to be modified with a short stretch of glycine residues. In the case of the split-intein technology the toxin moiety needed to be coupled to a C-intein peptide of ca. 40 amino acid length. These modifications of the small molecular weight toxin payloads required the expertise and know how of the third SME (TUBE Pharma) in the project, which is specialized in chemical modification and the synthesis of toxin compounds.
The activities of the 3 SMEs were supported by 7 RTD performers providing contract research services in the areas of protein optimization and design, peptide manufacturing, recombinant enzyme and antibody manufacturing, protein analytics, chemical synthesis and functional testing of oncology drugs in tissue culture cancer models. During the project all components required for the site-specific conjugation of toxic payloads using sortase enzymes and split inteins have successfully been manufactured and quality controlled. This included monoclonal antibodies with C-terminal sortase recognition peptides or N-intein domains, glycine modified monomethylauristatin F (MMAF) toxin with various functional chemical groups, recombinant sortase enzyme and C-intein peptides with a unique cysteine handle for toxin conjugation for toxins with maleimide linker functionality.
The method for ADC manufacturing by sortase enzyme mediated conjugation could successfully be developed and validated. This way, ADCs specific for the tumor targets CD30 and gpNMB targets have been generated. These next-generation ADCs were shown to exhibit comparable potencies for tumor cell killing in vitro, demonstrating the feasibility to manufacture potent, enzymatically and site-specifically conjugated ADCs. ADCs via the split-intein conjugation approach could not be manufactured based on technical challenges in preparing toxin-conjugated C-inteins. A patent application aimed at protecting the functionally validated site-specifically conjugated anti-CD30 and anti-gpNMB MMAF ADCs has been drafted and will be submitted in 2016.

Project Context and Objectives:
The MABTOX project was entirely focused on developing and validating two novel enzymatic conjugation platforms for the development of site-specifically conjugated, “next-generation” antibody drug conjugates (ADCs). One conjugation platform had been conceived by NBE-Therapeutics and involved the use of so-called sortase enzymes. Sortases are a unique class of trans-peptidase enzymes that can be found in gram-positive bacteria and that can catalyze the formation of peptide bonds between proteins when certain amino acid recognition motifs are present at each of the proteins. Therefore, sortase enzymes are also called protein ligases, because they can mediate the covalent coupling of two different proteins. The second conjugation platform represented background technology ERA-Biotech (replaced by ZIP solutions in February of 2015). Their technology platform was based on the activity of so-called split-inteins that can mediate the coupling of two protein domains. In this technology two proteins can be conjugated, if one protein is attached to an N-terminal part of an split-intein (the so-called N-intein) and if the other protein is coupled to a C-terminal part of a split-intein (the so-called C-intein).
While both the sortase enzyme and split-intein technologies have been used for protein engineering in the past, they have never been used in the context to attach non-protein/non-peptide small molecular weight toxin molecules to antibodies.
At the beginning of the project it was absolutely unclear, whether or not such a strategy would be feasible, efficient enough and would result in the generation of ADCs with similar potency for tumor cell killing as generated by classical chemical conjugation. In addition, it was unclear whether or not the use of enzymes in the context of conjugations of small molecular weight compounds, which are unrelated to proteins or peptides, might not result in unexpected side-reactions that would yield side products resulting from conjugations to unwanted amino acid residues of the manufactured ADC. Lastly, at the beginning of the project it was also uncertain, whether enzymatically conjugated ADCs, either generated by the sortase-enzyme technology (NBE background) or by the split-intein technology (ERA/ZIP background), would display equivalent potency for tumor cell killing than chemically conjugated ADCs.
In order to avoid additional tumour target or toxin-related risks, the consortium had decided to validate the novel approaches with antibodies, targets and toxins that are already validated in the context of traditional ADCs. Originally, the grant application contained the proposal to generate enzymatically conjugated ADCs against the liquid tumour (leukemia/lymphoma) targets CD30 and CD19, for which chemically conjugated ADCs had already been developed into the clinic. For the CD30 target, the benchmark ADC would be the FDA-approved MMAE-toxin conjugated ADC Adcetris® (from Takeda) for the treatment of Hodgkin lymphoma, which is also known as brentuximab-vedotin. For the CD19 target the benchmark ADC would have been a CD19-specific MMAF-conjugated ADC by Seattle Genetics (SGN-CD19A), which is currently in early stage clinical testing. However, before the start of the project, the joint steering committee (JSC) for the project have decided to implement 2 modifications. First, it was decided to replace the CD19 target with a validated solid tumor target, gpNMB, against which Pfizer had developed an MMAE-conjugated ADC, which was already in phase II clinical testing in melanoma.
Second, instead of using the MMAE toxin for validating enzymatically conjugated ADCs, the close structural variant MMAF was used. The reason for this second modification was based the fact that standard chemically conjugated ADCs often contain linker structures in connection with toxin payloads, ensuring efficient release of the toxin from the ADC upon tumour cell binding and internalisation. Based on data that was published shortly before the start of the project work, it became clear that MMAE toxins require cleavable linker structures for optimal function, which are normally not present in enzymatically conjugated ADCs generated by sortase-enzyme or split-intein conjugation. Conversely, published data suggested that the structural derivative of MMAE, MMAF, can potently kill tumour cells in the absence of cleavable linker structures.

Main Objectives:
Therefore, the final goal of the MABTOX project was to develop enzymatically conjugated anti-CD30 and anti-gpNMB ADCs with MMAF as a payload and to determine, if these next-generation ADCs would be functional for tumour cell killing. In order to achieve this goal the following main objectives needed to be achieved. 1.) Synthesis routes for MMAF toxin derivatives with chemical modifications allowing sortase enzyme conjugation needed to developed and the modified MMAF toxins needed to be manufactured. 2.) C-intein variants needed to be designed and manufactured allowing specific coupling of MMAF to produced C-inteins, such that by C-intein/N-intein interaction, the MMAF can be conjugated to an antibody. 3.) Recombinant antibodies needed to be manufactured containing modifications at the C-termini of the heavy and light chains of the antibody allowing sortase enzyme mediated conjugation or split-intein mediated conjugation to the C-terminal ends of the antibody heavy and light chains. 4.) Catalytically active recombinant sortase enzyme needed to be manufactured for sortase conjugation, because sortase enzymes were not available from regular commercial sources. 5.) Sortase-conjugated and split-intein conjugated MMAF ADCs against the CD30 and gpNMB tumor targets needed to be manufactured. 6.) Eventually the enzymatically conjugated ADCs needed to be tested for their cell killing activity against Hodgkin Lymphoma and melanoma tumour cells in tissue culture in vitro, in comparison to traditional, chemically conjugated ADCs.

Main Achievements:
In order to achieve the above-mentioned main objectives of the project, 8 different work-packages have been defined and for the majority of these work-packages the goals and deliverables have been achieved. Most notably, all components for NBE-Therapeutics' sortase-mediated antibody conjugation technology (SMAC-technology) could successfully be manufactured, and the generation of SMAC-technology conjugated MMAF ADCs specific for the CD30 and gpNMB tumor targets could be accomplished. Lastly, it could be demonstrated that SMAC-technology conjugated “next-generation” MMAF-ADCs displayed essentially identical anti-tumor activity against Hodgkin lymphoma and melanoma cells, in comparison to chemically conjugated benchmark ADCs, including the FDA-approved and marketed ADC Adcetris®. Because site-specifically, SMAC-technology conjugated anti-CD30 and anti-gpNMB ADCs with MMAF toxins represent novel and innovative composition of matter, a patent application describing the manufacturing and functional characterization of these next-generation ADCs has been drafted and will be submitted soon. The submission is planned to occur during the first half of 2016. Unfortunately, several technical difficulties were encountered in the realisation of the split-intein mediated toxin conjugation to antibodies that was based on the background technology of ERA Biotech / Zip solutions. First of all, only one of two different N-intein domains proved to be expressable and intact, if fused to the C-terminal ends of an antibody expressed in CHO cells. One of the N-intein domains selected and analysed (NrdJ.1) showed quantitative hydrolysis, as determined by mass-spectrometry analysis. Due to this hydrolysis the NrdJ.1 N-intein domain did not remain intact during purification of the recombinant N-intein-antibody fusion protein, and rendered these constructs unusable for the split-intein conjugation. However, the second N-intein domain selected and evaluated (GP41.1) as determined by mass-spectrometry analysis appeared to remain intact, if expressed as a genetic fusion to the C-termini of antibody heavy chains in CHO cells. This antibody-N-intein fusion protein would have been potentially usable for split-intein mediated conjugation using GP41.1 C-intein. As a second technical difficulty in connection with the split-intein conjugation approach, the chemical synthesis of the ca. 40 amino acid C-intein domains by peptide synthesis proved to be inefficient and did not result in production of the GP41.1 C-intein in sufficient quantity and quality, causing a first major delay of work-package 2 (generation of a functional C-intein peptide domain available for toxin coupling). As a solution to overcome the bottleneck of chemically synthesising C-intein peptides by RTD partner JPT peptide technologies, the JSC decided that the SME-partner ERA Biotech, in charge of this work-package (later replaced by its successor in the project, ZIP solutions) would attempt to produce the peptide as a recombinant peptide in E.coli. However, also that strategy proved to be inefficient and only insufficient quantities of the GP41.1 C-intein peptide could be produced by this backup strategy. In addition, the low quantities of the GP41.1 C-intein peptide did not display any catalytic activity, causing a further delay. In a last attempt, the GP41.1 C-intein peptide was expressed as a fusion protein, intended to increase expression yield and proper folding of the C-intein. However, also this approach failed to result in catalytically active C-intein domain of the GP41.1 split-intein. Lastly, chemical coupling of maleimide-linker modified MMAF to the GP41.1 C-intein protein was also associated with poor yield. Eventually, ERA Biotech (and later its successor, ZIP solutions), responsible for the split-intein mediated conjugation approach, was unable to reach the objectives to deliver functionally active and MMAF-coupled GP41.1 split-intein peptide to NBE-Therapeutics. Therefore, split-intein conjugated CD30 and gpNMB ADCs cpuld not be manufactured. Despite significant efforts by ERA Biotech and later ZIP solutions, who were in charge of the delivery of MMAF-conjugated C-intein peptides, this work-package (WP2) could not successfully be completed within the framework of the project. Therefore, split-intein conjugated MMAF ADCs could not be evaluated functionally.
In summary, it can be concluded that the first enzymatic conjugation technology, based on NBE-Therapeutics’ sortase-enzyme mediate antibody conjugation technology (SMAC-technology) could successfully be established and it could be shown that the resulting “next-generation” ADCs are fully functional, thereby validating this approach. The second enzymatic conjugation technology, based on ERA Biotech’s/ZIP-solution’s split-intein approach, despite significant efforts and execution of experimental backup strategies, however, could not be established and validated.

Project Results:
Overview of Objectives/Achievement of Milestones for the project:

During the 2-year MABTOX project, which was focused on establishing and validating novel enzymatic antibody conjugation approaches for generating site-specifically conjugated ADCs with either sortase enzymes or split-inteins, 8 technical work-packages have been defined, in order to reach the final objective of the project (please compare to the last version of the description of work (DoW), and its attached Annex I, in its last version of 23. Oct. 2015). The main objectives and achievements, including joint-steering committee-agreed adjustments for the 8 different work-packages are provided below:

1.) WP1 (main responsible: SME-partner TUBE):

Objective: WP1 originally had the objective to provide glycine-modified MMAE toxin that could be used for the site-specific conjugation of the MMAE toxin to C-terminally sortase-tagged antibodies obtained from WP4. This work-package mainly involved the RTD-partner TBD providing the synthesis services for the different toxin compounds.

Achievements & adjustment: Based on published data related to cleavable linkers required for MMAE containing ADCs and reports that the MMAE variant MMAF, which is a closely related derivative of MMAE, may be fully functional also in the absence of cleavable linkers, by unanimous JSC decision, and before the start of activities, it was decided to produce MMAF toxin payloads in this work-package, instead of the MMAE toxin. As a consequence, 9 different glycine-modified variants of MMAF and an O-methyl ester derivative, called MMAF-OMe have been synthesized and delivered in this work-package (details see below, and deliverable reports D1.1 – D1.4 of the project). Eventually, as part of WPs 7 & 8, it could be demonstrated that the synthesised toxins were fully functional as payloads and resulted in highly potent ADCs with anti-tumor cell activity comparable to benchmark ADCs manufactured by chemical conjugation. Therefore, it can be concluded that WP1 has successfully been completed. Not all of the MMAF and MMAF-OMe variants could be evaluated, but this will be continued after the end of the project, and any additional data from these experiments will be included within the priority year in the patent application claiming the composition of matter of such novel anti-CD30 and anti-gpNMB-MMAF and MMAF-OMe ADCs (see WP 8). Details for this WP are provided in the 4 deliverable reports D1.1 - D1.4 provided for this WP.

2.) WP2 (main responsible SME-partner: ERA, replaced by SME-partner ZIP on February 13, 2015):

Objective: The goal of WP2 was to provide an optimized C-intein polypeptide to be produced by chemical peptide synthesis based on optimized design by RTD-partner INNO and eventually manufactured by RTD-partner JPT.

Achievements & adjustments: Based on published functional and structural data, two C-int domains have been selected by ERA and optimized C-int variants have been designed by INNO. However, chemical peptide synthesis of these C-int peptides by JPT proved to be more challenging as expected. The main reason for this was that the length of the C-int peptides (ca. 40 amino acids) is close to the limit for chemical synthesis at larger scale. In addition, the particular amino acid sequence of the C-int peptide appeared to cause unanticipated problems for the efficient peptide synthesis and required additional optimizations, which caused major delays. Therefore, as a backup, the JSC decided to produce wild-type C-int domains for the two selected split-inteins, GP41-1 and NrdJ.1 in a validated E.coli expression system to provide initial material for evaluation of split-intein mediated conjugation. However, also that strategy proved to be inefficient and only insufficient quantities of the GP41.1 C-intein peptide could be produced by this backup strategy. In addition, the low quantities of the GP41.1 C-intein peptide did not display any catalytic activity, causing a further delay. In an last attempt, the GP41.1 C-intein peptide was expressed as a fusion protein, intended to increase expression yield and proper folding of the C-intein. However, also this approach failed to result in catalytically active C-intein domain of the GP41.1 split-intein. In summary, despite significant efforts for optimization of the chemical synthesis of C-intein peptides and implementation of a backup-strategy by expressing the C-int peptides in E.coli functional C-int domains could not be obtained. Details for this WP are provided in the 3 deliverable reports D2.1 - D2.3 provided for this WP.

3.) WP3 (main responsible: SME-partner TUBE):

Objective: WP3 had the goal to provide MMAE conjugated C-intein peptides that could be used and evaluated for conjugation to N-int modified antibodies from WP5.

Achievements & adjustments: Due to the considerations detailed in WP1, and based on unanimous JSC decision, before the start of the project it was decided to rather generate MMAF conjugated C-intein peptides, rather than MMAE conjugated C-intein peptides. The detailed structures and synthesis route of these linker modified MMAF compounds have been disclosed in the final MABTOX Review Meeting with the EU scientific expert and they are also available in an updated deliverable report D3.3. However, due to the failure to provide functional C-intein peptides from WP2, the final MMAF-conjugated C-intein peptides could not be manufactured. Details for this WP are provided in the 5 deliverable reports D3.1 - D3.5 provided for this WP.

4.) WP4 (main responsible: SME-partner NBE):

Objective: WP4 originally had the goal to provide two sortase motif modified antibodies specific for CD30 and CD19 tumor antigens. The production of the modified recombinant proteins was mainly the task of RTD-partner EVI and the analytical quality control involved RTD-partner PROT.

Achievements & adjustments: Based on best scientific rationale and unanimous JSC decision, the CD19 mAb was replaced by an antibody against the solid tumor (breast & melanoma) target gpNMB, which is also a validated target, against which an ADC with conventional chemical conjugation to MMAE is being evaluated in phase II clinical trials by Pfizer for the indication of melanoma. Ca. 85mg and 70 mg of anti-CD30 antibody clone Ac10 and anti-gpNMB antibody clone CR011, respectively, containing a C-terminal sortase recognition motif could successfully be produced and quality controlled. These antibody clones correspond exactly to the antibody clones that form the basis for the chemically conjugated MMAE benchmark ADCs Adcetris®/brentuximab-vedotin (of Takeda) and glembatumumab-vedotin (of Pfizer). Therefore, it can be concluded that the objective for WP4 has been achieved. Details for this WP are provided in the 5 deliverable reports D4.1 - D4.5. provided for this WP.

5.) WP5 (main responsible: SME-partner NBE):

Objective: WP5 had the goal to provide two N-intein modified CHO-cell expressed recombinant antibodies specific for CD30 and CD19 and to provide the analysis regarding the structural integrity of the C-terminal fusions of the N-intein domains.

Achievements & adjustment: As described above, originally it was proposed to work with antibodies specific for the CD19 and CD30 tumour targets. However, based in unanimous decision by the JSC before the start of the project, it was decided to replace the CD19 antibody with an antibody specific for the solid tumour (breast & melanoma) target gpNMB, which is also a validated ADC target. Based on published patent applications disclosing the specifications of marketed ADC Adcetris® from Takeda (also called brentuximab-vedotin, and based on antibody clone Ac10), and of clinical stage ADC glembatumumab-vedotin (antibody clone CR011) from Pfizer, mammalian expression vectors for C-terminally N-intein-tagged antibodies have successfully been cloned. The expression vectors were of human IgG1/kappaL isotype. For each antibody clone, two different N-intein domains, namely that of split intein Gp41.1 and NrdJ.1 have been fused to the anti-CD30 and the anti-gpNMB mAbs. This resulted in the production of four different N-intein-antibody fusion antibodies (Ac10-GP41.1 Ac10-NrdJ.1 CR011-GP41.1 and CR011-NrdJ.1). The N-intein modified recombinant antibodies have successfully been expressed in CHO cells and purified at ca. 40 mg scale each. Detailed protein analytics including mass-spectrometry analysis revealed that the NrdJ.1 N-intein domain is a target for proteolytic cleavage, rendering this split intein fusin antibody useless for further split-intein mediated conjugations. Indeed, a substantial part of the N-intein domain of 19 kD size had proteolytically been cleaved off in CHO cells. In contrast, there was no evidence of degradation for the Gp41.1 N-intein domain, suggesting that this split-intein antibody fusion construct could be used for coupling with the corresponding GP41-1-C-intein-toxin conjugates that were the deliverable of WP 3. Therefore, the objective and the final milestone of WP5 has successfully been achieved. Details for this WP are provided in the 4 deliverable reports, D5.1 – D5.4. provided for this WP.

6.) WP6 (main responsible: SME-partner ERA, replaced by SME-partner ZIP on February 13, 2015):

Objective: WP6 had the goal to establish a manufacturing process for recombinant expression of sortase A enzyme from Staphylococcus aureus in E.coli and to eventually provide 1 gram of catalytically active sortase A enzyme for SMAC-technology mediated conjugations. This activity mainly involved the services of the RTD-partner BIO in the project, a service provider specialised in the production of recombinant E.coli proteins.

Achievements: Because sortase A enzyme was not available from commercial sources at the start of the project, a manufacturing process for the enzyme needed to be established. Meanwhile, there is indeed one single commercial source that offers E.coli produced recombinant sortase A enzyme form Staphylococcus aureus (BPS Bioscience, U.S.) however, at the cost of US$ 425 for 50 micro-gram (Catalogue price for item #71086), this still would not be economically viable for the project. In fact, 1 gram quantity that was the deliverable for this work-package would translate into cost of US$ 8.5 Mio. Therefore, even considering the most recent commercial availability for sortase A enzyme by one commercial supplier, it still would be required to establish an internal manufacturing process for the sortase enzyme. In a first step, E.coli expression vectors for recombinant sortase A from Staphylococcus aureus have been cloned and quality controlled. Then, E.coli production clones have been established and a shake-flask production process for production the sortase A enzyme has been established, resulting in an acceptable yield of 80mg/liter shake culture. From this process development run, a test batch of 13 mg of sortase A enzyme has been produced as affinity purified, recombinant sortase A for quality control purposes and for the characterization of the catalytic activity of the enzyme. For the assessment of the catalytic activity a fluorescent substrate (Glycine-5 modified FITC) had been generated and an assay has been established to quantitate the catalytic activity of sortase A enzyme by measuring the conjugation of the Glycine-5 modified FITC substrate to sortase-motif tagged antibodies. The purified recombinant sortase A protein produced from the process development run displayed sufficiently high activity resulting in up to 80% conjugation of the fluorescent dye substrate to sortase-motif tagged antibodies. Further protein analytical characterisation confirmed the identity and high quality of the produced test batch of sortase A. Based on these results an E.coli production clone has been up-scaled for the production of the desired 1gram batch of sortase A enzyme in 25 liter shake flask cultures. From this process, 1’500 mg (1.5 g) of sortase A could be produced as > 95% pure recombinant protein following Ni-NTA affinity purification. This corresponds to a final yield after expression, processing and purification of 60mg/liter shake culture, and only suggests a loss of 25% of the material as a consequence of processing and affinity purification. The final production batch was also quality controlled for purity, structural integrity and catalytic activity and this batch was of the same high quality as the original process development test batch. Therefore, it can be concluded that the goal of this project has been completely achieved. If the catalogue-price of US$ 425 for 50 micro-gram of sortase A by the one commercial source (BPS Bioscience in the U.S.) would be taken as a basis to calculate the value of the 1.5 g batch of sortase A, this would correspond to a value of US$ 12.75 Mio. This illustrates that it was absolutely required to establish the manufacturing process as part of this project, because larger amounts of high-quality sortase A enzyme are of critical importance for the manufacturing of site-specifically SMAC-technology conjugated ADCs of work-package 7. Details for this WP are provided in the 4 deliverable reports, D6.1 – D6.4. provided for this WP.

7.) WP7 (main responsible: SME-partner NBE):

Objective: Using the developed components of WPs 1, 2, 3, 4, 5 and 6, MABTOX WP7 originally had the goal to provide site-specifically conjugated MMAE-ADCs against the CD19 and CD30 tumor antigens using the sortase-mediated antibody conjugation technology (SMAC-technology), proprietary to NBE-Therapeutics, and the split-intein approach, proprietary to ERA-Biotech (later in the project replaced by ZIP-solutions).

Achievements: As outlined in the previous WP descriptions 1, 2, 3, 4 and 5 further above, before the start of the project and based on best scientific rationale and by unanimous decision of the MABTOX Joint Steering Committee (JSC), the CD19 tumour target had been replaced by the solid tumour target gpNMB. In addition, as also outline before, it was decided to replace the toxin payload MMAE with its close chemical derivative MMAF, based on expected higher potency of ADCs in the context of stable peptide linkers (none of these changes have changed the scope of the project). Due to the failure to provide sufficient quantity and quality of C-intein peptides and therefore also the failure to provide MMAF-conjugated C-intein peptides from WP 3, the goal to produce site-specifically conjugated split-intein conjugated MMAF-ADCs could not be achieved.
However, in contrast, sortase-motif modified anti-CD30 and anti-gpNMB antibodies had successfully been produced with required quality and quantity from WP4. Also, glycine-modified MMAF toxins have successfully been manufactured from WP1, and a large batch of high-quality sortase A enzyme from Staphylococcus aureus has been manufactured from WP6. Therefore, all components for NBE-Therapeutics' sortase-mediated antibody conjugation technology (SMAC-technology) were available. First the available component were used to successfully establish an optimized conjugation process for the generation sortase conjugated anti-CD30-MMAF and anti-gpNMB-MMAF reaching a high efficiency of >80 % conjugation. The final goal of this WP was to thoroughly characterize the enzymatically conjugated ADCs for the proper conjugation of the MMAF toxin payload to the C-termini of the IgH and the IgL chains of the antibodies, and to exclude the presence of any “mis-conjugations”. These analysis included HPLC analysis by reverse-phase HPLC, hydrophobicity interaction chromatography (HIC), but also a detailed mass-spectrometric analysis, in order to confirm the detailed structural features of the SMAC-technology manufactured ADCs. From these analyses it could be concluded that the toxin payload had indeed been conjugated to the expected conjugation site determined by the position of the sortase tag, LPXTG, positioned at the C-termini of the IgH and IgL chains of the conjugated antibodies. In addition, peptide mapping by mass spectrometric analysis proved that identiy of the ADCs with the theoretical amino acid sequence. Also, mass-spectrometric analysis of the entire antibody and of isolated antibody heavy and light chains demonstrated that the antibody was fully intact and displayed a molecular weight as expected, including the mass of the toxic payload as expected. Therefore, it can be concluded that at least for sortase enzyme mediated conjugation of payloads to antibodies the goals of this WP have been fully achieved. Due to the technical difficulties in connection with C-intein peptide manufacturing and conjugation of WP 2, the split-intein conjugation approach could not be performed. Details for this WP are provided in the 3 deliverable reports, D7.1 – D7.3. provided for this WP.

8.) WP8 (main responsible: SME-partner NBE):

Objective: WP8 had the objective to provide the final proof-of-concept that the manufacturing of site-specifically conjugated ADCs by either sortase enzymes or split-inteins would result in functionally active ADCs with high potency for tumour cell killing. In case of positive validation of these “next-generation” ADC formats, final goals of WP8 were to file intellectual property rights for the characterised novel ADCs and after submission of IP to also submit the results to peer-reviewed journals for the publication of a paper. The functional testing involved the services of the RTD partner ONCO in the project.

Achievements: In order to be able to compare our site-specific, enzymatically conjugated anti-CD30-MMAF and anti-gpNMB-MMAF ADCs against traditional chemically conjugated MMAE ADCs, first chemically conjugated benchmark ADCs needed to be procured. Because the anti-CD30-MMAE ADC Adcetris® is a marketed product since 2011, we could procure commercial material from an Oncologist at the Swiss Tumor Center at Hirslanden Clinic in Zürich. However, the anti-gpNMB-MMAE ADC glembatumumab-vedotin from Pfizer is not yet market approved and therefore not available at the market. Therefore, we had to perform a chemical conjugation with exactly the same linker-toxin structure and process as disclosed in the patent application for Pfizer's glembatumumab-vedotin. The process for chemical conjugation has successfully been set-up at NBE, and two batches of the anti-gpNMB-MMAE ADC glembatumumab-vedotin have been manufactured as benchmark ADCs. The only two commercially approved ADCs Adcetris® (from Takeda) and Kadcyla® (from Roche) are chemically conjugated and have a drug-to-antibody ratios (DAR) of ca. 3.5. This appears to be a sweet-spot for the number of drugs coupled on average to an antibody in chemical antibody conjugates. Because it is not exactly known, what the exact average drug-to-antibody ratio is of chemically conjugated glembatumumab-vedotin from Pfizer, we produced two batches, one with an average DAR of 3.2 and one with an average DAR of 4.6. Both of these batches were used as benchmarks for our in vitro tumour cell killing experiments in this WP. As a next step, using the benchmark ADCs, as well as regular chemotherapy drugs, in vitro tumour cell models needed to be set up for relevant tumour cells. The RTD-partner Oncotest has extensive experience and access to human tumor cell lines representing various human tumour cell types. Eventually, 6 Hodgkin lymphoma lines, 8 melanoma and 2 breast cancer cell lines were used to establish tumour cell killing assays, which requires adaptation of the cell culture conditions for each individual cell line, in order to be able to obtain meaningful dose-response curves upon drug exposure. Once the tumour cell models had been set-up using the standard chemotherapy drug Paclitaxel, the established conditions were also used for evaluation of the enzymatically manufactured ADC and the chemically conjugated benchmark ADCs. The 6 Hodgkin lymphoma line included CD30-negative lymphomas as well as CD30-negative controls, in order to assure that a potential anti-lymphoma effect would be specifically mediated via the specific binding of the ADCs via CD30 and not mediated via non-specific binding or uptake of the ADCs. Results from the Hodgkin lymphoma tumour killing experiments showed clearly that the SMAC-technology conjugated anti-CD30 ADCs exhibited the same, if not a hint higher potency for tumor cell killing, than commercially available Adcetris®. In addition, this tumour killing was only detectable in 3 out of 6 CD30-expressing Hodgkin lymphoma cell lines, but not in 3 out of 6 CD30-negative Hodgkin lymphoma cell lines. Therefore, this first result with the sortase-enzyme conjugated anti-CD30 ADCs clearly demonstrated an exquisite specific potency of site-specific, sortase enzyme conjugated CD30-MMAF ADCs, which were in no way inferior to the marketed product Adcetris®.
Next, we also evaluated the potency of the SMAC-technology conjugated anti-gpNMB-MMAF ADCs for tumor killing on a set of 2 breast cancer and 8 melanoma cells lines. Only 4 of the melanoma cell lines were clearly positive for the gpNMB tumour marker. Also in this experiment, none of the gpNMB negative cells was tractable to anti-gpNMB ADC killing - neither with the chemically conjugated ADCs, nor with the sortase enzyme conjugated ADCs. This indicates that there is no non-specific killing of the melanoma cells by any of the tested ADCs. Two of the four melanoma cell lines responded strongly to treatment with the chemically conjugated anti-gpNMB-MMAE ADC glembatumumab-vedotin of Pfizer, but only for one of them a comparably strong anti-tumor activity was detectable upon treatment with the sortase-conjugated anti-gpNMB-MMAE ADC. While this proves that the feasibility of site-specific sortase-enzyme conjugated ADCs also for the gpNMB target, it is not entirely clear, why a differential effect was detectable in two different responsive tumour cell lines. This eventually needs to be followed up by further experiments and analyses. At the same time another clearly gpNMB positive melanoma cell line did not respond to any of the ADC treatments, not even the chemically conjugated ADCs, despite the fact that it was generally responsive to paclitaxel treatment. This highlights that the tumour target gpNMB might be associated with a more complex target biology that is not yet fully understood. One of the four gpNMB positive cell lines only responded weakly to anti-gpNMB ADC treatment, again here similarly to the chemically conjugated benchmark ADC, as well as to the site-specifically, sortase enzyme conjugated ADCs. In summary, it can be concluded that the format of SMAC-technology conjugated ADCs could be validated in two different tumor cell models - one for a liquid tumour and one for a solid tumour target. Therefore, the final objective of the MABTOX project has been achieved by demonstrating clear anti-tumour activity of our novel, “next-generation” ADC format produced by NBE-Therapeutics’ SMAC-technology. Because site-specifically conjugated ADCs are believed to be associated with improved pharmacological properties (increase stability, more homogeneous PK/PD paramneters, etc.) and because the format represents a novel composition of matter for both the produced anti-CD30-MMAF and anti-gpNMB-MMAF ADCs, we have drafted a first patent application protecting the novel composition of matter. It is the goal to file the patent application as a European priority application before the end of the second quarter 2016. Following this, the data disclosed in the patent application can also be submitted for publication in a peer-reviewed scientific journal, which is also planned for 2016.
Therefore, in summary, it can be concluded that at least for the sortase-enzyme conjugation approach, the MABTOX project has achieved all of the objectives that have been defined in the project proposal that has been submitted back in 2012, when the enzymatic ADC-manufacturing strategies have only been at concept stage. Details for this WP are provided in the 6 deliverable reports, D8.1 – D8.6 provided for this WP.

Potential Impact:
Cancer is still one of the most frequent causes of premature death that can affect people of all age-groups. Cancer has a significant socio-economic impact on our society, due to the high costs associated with cancer prevention, diagnosis and eventually treatment. The economic cost associated with time and effort spent by the health systems, but also by patients and their families affected by cancer and the economic cost associated with lost productivity due to cancer-related disability and premature death is a significant burden for any society. Therefore, we need to relentlessly work towards improving the quality and effectiveness of cancer prevention and treatment strategies. Despite significant improvements in the areas of cancer prevention, diagnosis and therapy during recent years, standard therapy of cancer is still mostly based on classical chemotherapy in conjunction with surgery and radiotherapy. As we all know, these therapies are often associated with significant side-effects and strong suffering for the patients that need these therapies. Unfortunately, in most cases standard therapies are only life-prolonging and not really curative. Clearly, targeted therapies that specifically address the tumour via tumour specific features and the use of biotechnology to create cancer drugs with high potency and selectivity that may reduce side-effects and that eventually improve the chance to cure cancer, are of high relevance for society. In recent years, antibody-based therapies, developed by Biotech companies around the world, and which are based on the specific function of the immune system's own defence system (e.g. antibodies) have resulted in breakthrough development in the cancer field. However, these therapies are still not in the mainstream of cancer therapy and only available for selected cancer indications. In addition, it has been shown recently, that the anti-tumour potency of antibodies can be dramatically increased, if antibodies are “armed” with potent drugs (toxins). This way, the toxic drugs are specifically taregeted to tumour cells. In clinical trials this class of drugs is currently saving the lives of patients that would normally die, if only standard of care was available to them. However, as clinical trials with ADCs advance, and a more extensive data basis is building up, the field is recognising that traditional ADCs are not optimal, because the chemical conjugation process for the manufacturing of the first generation ADCs leads to heterogeneous drug substances with variable and sub-optimal properties. The most important liability of chemically conjugated ADCs is a certain degree of pre-mature release of the toxic drug from the antibody before the ADC acts on the tumour. This is due to the fact that the stability of the drug coupled to the antibody very much depends on where in the antibody structure the drug-coupling has occurred. In the marketed product Kadcyla® from Roche-Genentech, it is known that toxin drugs can be attached to more than 90 different positions in the antibody, of which some may keep the drug very tightly attached, whereas at other sites a quick release of the drug occurs. In clinical practice it is observed that indeed shortly after injection of a chemically ADC, a surge of detached free toxin is detectable in serum of ADC treated patients, which apparently leads to the classical side-effects of chemotherapy. This notion is supported by analyses of the side-effects of ADC therapies in clinical trials using different ADCs, different targets and addressing different indications. The side-effects appear to be always identical and related to the toxin moiety that is attached to the ADCs. Therefore, it is of greatest importance to establish technologies allowing the manufacturing of site-specifically conjugated ADCs, that can be optimised for optimal attachment of the toxin-drugs to the antibody, and to achieve maximum safety with highest specific potency (i.e. to achieve an optimal therapeutic window). This trend has been recognised in the industry, but the paradigms in ADC development are only slowly shifting. The demonstration in the MABTOX project that it is possible to generate site-specifically conjugated ADCs with enzymatic technologies that are equally effective (safety parameters still to be investigated) than classical ADCs, is already a huge step forward.

On the market level, it is also noteworthy that the field of ADC development is very much dominated by U.S. based Biotech and Pharma companies, e.g. Genentech, Pfizer, Seattle Genetics and Immunogen. More than 80% of ADC drugs currently in clinical trials have been developed and pioneered by U.S. based corporations. The exposure of European top innovators and companies in this field is clearly underrepresented. In this sense the foundations for establishing the SMAC-technology at NBE-Therapeutics that was supported by the MABTOX project, has created one internationally recognised European player that continues to pursue the vision to develop the early concepts further into clinical practice.

On a micro-economic level, the Research for SMEs grant MABTOX had an immensely positive impact on the development of the coordinating company. At the beginning of the project, the organisation had 4 employees and it has meanwhile grown to 12 employees. Currently, NBE-Therapeutics is at the brink of raising investor’s funds for a further expansion and entering into the phase of a clinical stage company. In addition, the interactions between the Euroepan partners of the project, on one hand between the SMEs, and also between the SMEs and the RTD partners have tightened and several collaborations are continued or have been started in addition to the MABTOX project.

Therefore, in conclusion, it can be stated that the MABTOX project has and will have a sustained value and a positive impact 1.) on the micro-economic level for each of the partners in the project, 2.) on the competitiveness of the European SMEs versus the U.S. corporations dominating this field, and 3.) for the society and the patients, if eventually the improved ADC drug format finds its way into the clinic and to cancer patients in need.

List of Websites: