Automating ADC Manufacturing with Electricity

Abstract
Combining the specificity and stability of antibodies with the potency of small molecules has always been at the core of the antibody drug conjugate (ADC) field. The technological challenge to this therapeutic approach is in bringing two disparate molecules in union with one another. Once this was achieved, the next technological hurdle was, in essence, automation [1].

Early versions of an ADC were the product of technical skill which also required trial and error. Bioconjugation experts investigated many combinations of antibodies, small molecule drugs, conditions, and reaction times to produce stable and efficacious final drugs. This work serves as a technical knowledge base from which to convert new antibodies into ADCs. Even with compendium textbooks like Bioconjugate Techniques, trial and error is still required especially to manufacture an ADC at clinical scale.

Technological innovations by ADC giants like Genentech and Seagen sought to automate the union of antibodies and small molecule drugs with protein manipulation and drug linker chemistry [1].  While not typically considered automation, these advances sought to limit the need for human intervention in the ADC manufacturing process for more consistent and scalable curative drugs. Applying electrochemistry to bioconjugation, we hope to automate the creation of the building blocks of drug discovery. This would allowing scientists to apply their creativity to what they can create and not if they can.

The World’s First Bioconjugation Instrument
In this article we present early development of the world’s first bioconjugation instrument as furtherance of the central goal of automating ADC manufacturing. Bioconjugation instrumentation does exist, [2] however, this instrument only automates the mixing of the essential components of ADC manufacturing (antibody, drug/linker and reducing agent). The recipe of how all of these components work in concert still needs to be predetermined through expert intervention and trial and error.

Our invention is a true bioconjugation instrument as it is an active participant in the bioconjugation process using electrochemistry to replace reductants like TCEP (Figure 1A). Furthermore, our development of this instrument has determined that bioconjugation conditions can be universally applied to protein types and concentrations. For example, all human IgG1 antibodies at 2 mg/mL will require the same electrochemical conditions to produce a consistently labeled final ADC.

Electricity as a Tool for Consistent Bioconjugation
Using electricity to reduce disulfide bonds in proteins has been previously reported as a method to prevent foaming and to keep whey protein emulsions stable in food chemistry applications [3]. A commercial instrument is also available to use electricity to reduce disulfide bonds prior to mass spectrometry analysis. [4][5] Our goal was to utilize simple engineering to generate an electrochemical bioconjugation cell (eChem) for the expressed purpose of keeping the protein intact while reducing the disulfide bonds with electricity.

Figure 1.0 [A] General diagram of our initial proof of concept electrochemistry device. [B] Photograph of the working device with a 3D printed scaffold to hold the device in place. [C] Overview of potential development for new ADCs: 96 well screening, small scale protection for in vivo mouse studies post in vitro screening, and large scale lead drug manufacturing for clinical trials.
Our first innovation was to separate the reducing and oxidizing chambers of the electrochemical cell (Figure 1A). A disposable polypropylene cup with a dialysis membrane served to isolate the reduction of the protein from the oxidative charge balancing that occurs in the other chamber. We have tested multiple metal anodes in the reduction chamber, and primarily use a carbon rod cathode in the oxidation chamber. The separation of the oxidation and reduction chambers also allows for higher throughput and no cross contamination as the reduction chamber is disposed of between reductions.

Our initial proof of concept instrument (Figure 1B) represents a medium scale model that could be used to generate enough material for an in vivo mouse study in as little as 48 hours. We are also working on a 96 well version which would help automate the building and screening thousands of ADCs (Figure 1C). It is our expectation that, as we develop this technology, the small scale reagents would then be scaled up using the same buffers and electrochemistry conditions to produce the same ADC at each stage of drug development (Figure 1C).

Figure 2.0 Step by step process for reducing disulfide bonds in our electrochemical chamber.

Proof of Concept
For this experiment we used a bivalent mouse (Fab)2 as our model protein. With disulfide bonds linking the two Fab portions, we knew we could easily identify disulfide bond reduction by the change in protein size observed at specified time intervals on a standard protein gel (Figure 3). Our initial focus was on testing the voltage applied to our electrochemical cell as it was the key to the function of the instrument.

With voltage A, the higher voltage, a nonfunctional result was produced. The pH of the buffer was noticed to decrease significantly, indicating that acidic conditions had formed due to hydrolysis. No reduction was observed, and the protein was degraded after 2 minutes.

With voltage B, the lower voltage, the desired disulfide bond reduction was achieved. The 100 kDa (Fab)2 was cleanly split into 50 kDa Fab and 25 kDa heavy/light chain subunits. Maximum subunit concentrations were achieved after approximately 10 minutes, however the test was run out to 30 minutes, by which point notable subunit degradation occurred.

For a range of constant voltages, various electrical and buffer conditions were investigated, with similar results. Above a certain voltage, hydrolysis becomes dominant and degradation occurs. Whereas, at lower voltages, stable reduced proteins are produced. Standard protein buffers like PBS and HEPES were tested with similar results.

Figure 3. [A] Schematic of the expected reduction steps. [B] Two separate voltages supplied by a DC powersource, A and B, were applied constantly to 5mL of 2 mg/mL mouse (Fab)2 protein in the reduction chamber. The reduction chamber was disposed of and the larger oxidation chamber buffer replaced between experiments. 10 µL samples were taken from the reduction chamber every 2 min and ran on a non-reduced precast 4-12% Bis-Tris protein gel which was then stained with a Coomassie blue protein stain.
Putting the Pieces Together
Building on what we learned with our POC (Fab)2 protein we began testing on an intact antibody which, after disulfide reduction, was mixed with a maleimide linked toxin or fluorophore. All the same buffers and electrochemical conditions that were optimized on the mouse (Fab)2 were applied to a humanized IgG1 herceptin-like antibody.

The goals of the experiments were: analyze the location and the extent of the antibody labeling and  ensure a product with selective cell toxicity in functional assays. Automation of ADC manufacturing requires, at minimum, achieving these goals to meet industry and regulatory standards. The methods included:

  • Using mass spectrometry to verify maleimide linker attachment and location post electrochemical disulfide bond reduction.
  • Testing of a bioconjugated Herceptin-like antibody in functional assays to ensure retention of biological activity.

For the analytical mass spec analysis, the result was quite surprising in that we are only seeing fluorophore additions to the Fab region of the antibody and never, in repeat testing, have we seen additions to the FC portion (Figure 4). This Fab specific labeling has repeated in follow-up experiments. This method of specific disulfide reduction may help to automate the creation of dual labeled antibodies for drug combination screening and manufacturing.[7]

Figure 4.0 Prior to injection into the MALDI-TOF, the antibody was digested with papain which will cleave the antibody at the hinge region reducing the antibody into its Fab and FC fragments.

Retained Specificity and Activity
With the Herceptin-like antibody conjugates analyzed, the next step was to test them in various functional assays. With the IgG1 antibody reduced in the eChem instrument, we made two different antibody conjugates. One with a maleimide tethered fluorophore (AlexaFluor 488) for Cell surface staining and imagining.

The other, a maleimide linked toxin (MMAF) for in vitro cell killing experiments. From these experiments we were able to confirm that the eChem reduced antibody had retained its function and specificity

Figure 5.0 [A] An overview of the workflow in generating two different reagents from the initial electrochemical disulfide reduction. Ten-fold molar excesses of the respective maleimide linked toxin or fluorophore were incubated overnight at room temperature with reduced antibody (either eChem or TCEP). The antibody conjugates were then purified with a 50 kDA Zeba desalting column and ran on reduced and non-reduced protein gel to confirm the antibody had not degraded by either TCEP or eChem (data not shown). [B] Inverted microscope image of cells stained with fluorescently labeled antibody. HER2 positive (HCC-1954) or HER2 negative cells (HCC-1187) were stained with Herceptin-like antibody conjugated with Alexa Fluor 488 and the nuclear stain DAPI. [C] 72 hour Cytotoxicity assay with antibodies labeled with toxin. Either SKBR3, Her2 positive, or MDA-MB-231, Her2 negative, were separately seeded into a 96 well tissue culture plate. The controls used were unlabeled Herceptin-like antibody, human IgG1 isotype control labeled with MMAF after TCEP reduction, and Herceptin-like antibody labeled with MMAF after TCEP reduction. The Herceptin-like antibody labeled with MMAF after eChem reduction, along with the controls, were then titrated and added to the 96 well plate of cells for three days. Cells were lysed using Celltiter-Glo with higher luminescence values indicating live cells and no cell killing. Of note, our TCEP reduced control did demonstrate some off-target toxicity in the MDA-MB-231 cells at the highest concentration.
A New Chemistry to be Explored
In this report we have demonstrated some of our current understanding of how this electrochemical device is able to harness electricity for bioconjugation.

This technology removes guesswork and years of expertise required by other industry standard methods. We do imagine that one day making your own tagged antibodies will be an easy push-button procedure at any scale allowing new antibody therapeutic companies to enter the ADC field. This kind of automation has already had a tremendous impact on biotechnology fields like Cell Therapy allowing scientists to build and screen thousands of small scale drugs to find the transformative unicorn. [8]

Development of this technology will continue along two tracks.

The first will be to graduate our POC instrument to a benchtop instrument to automate the bioconjugation process at small, medium and large scales.

The second track is unlocking the new linker chemistries made possible by starting the bioconjugation process with an electric spark.

References
[1] Tsuchikama K, An Z. Antibody-drug conjugates: recent advances in conjugation and linker chemistries. Protein Cell. 2018 Jan;9(1):33-46. Epub 2016 Oct 14. PMID: 27743348; PMCID: PMC5777969, DOI: 10.1007/s13238-016-0323-0.
[2] Yuichi Nakahara, Brian A. Mendelsohn, and Yutaka Matsuda. Antibody–Drug Conjugate Synthesis Using Continuous Flow Microreactor Technology. Organic Process Research & Development 2022 26 (9), 2766-2770, DOI: 10.1021/acs.oprd.2c00217.
[3] Philippe Cayot, Hélène Rosier, Loı̈c Roullier, Thomas Haertlé, Gérard Tainturier, Electrochemical modifications of proteins: disulfide bonds reduction. Food Chemistry,Volume 77, Issue 3, 2002, Pages 309-315, DOI: 10.1016/S0308-8146(01)00352.
[3] Simon Mysling, Rune Salbo, Michael Ploug, and Thomas J. D. Jørgensen, Electrochemical Reduction of Disulfide-Containing Proteins for Hydrogen/Deuterium Exchange Monitored by Mass Spectrometry, Analytical Chemistry 2014 86 (1), 340-345, DOI: 10.1021/ac403269a
[4] Martijn M. Vanduijn, Hendrik-Jan Brouwer, Pablo Sanz de la Torre, Jean-Pierre Chervet, and Theo M. Luider, Online Electrochemical Reduction of Both Inter- and Intramolecular Disulfide Bridges in Immunoglobulins, Analytical Chemistry 2022 94 (7), 3120-3125. DOI: 10.1021/acs.analchem.1c04261.
[5] Agnieszka Kraj, Hendrik-Jan Brouwer, Nico Reinhoud, and Jean-Pierre Chervet, A novel electrochemical method for efficient reduction of disulfide bonds in peptides and proteins prior to MS detection, Anal Bioanal Chem (2013) 405:9311–9320, DOI: 10.1007/s00216-013-7374-3.
[6] Nervig CS. and Owen SC Advances in the Development of Dual-Drug Antibody Drug Conjugates –J. ADC. January 5, 2023. DOI: 10.14229/jadc.2023.01.05.001.
[7] Moutsatsou P, Ochs J, Schmitt RH, Hewitt CJ, Hanga MP. Automation in cell and gene therapy manufacturing: from past to future. Biotechnol Lett. 2019 Nov;41(11):1245-1253, Epub 2019 Sep 20. PMID: 31541330; PMCID: PMC6811377, DOI: 10.1007/s10529-019-02732-z.

Authors:

Corresponding Authors: Derrick Houser. E-mail: Derrick Houser  and Scott Beaver, Ph.D. E-mail: Scott Beaver, Ph.D

Key terms: ADC, bioconjugation, automation, bioconjugation techniques
Published In: ADC Review| Journal of Antibody-drug Conjugates

DOI: https://doi.org/10.14229/jadc.2023.04.10.001.

How to cite:
Houser D. 1 and Beaver S. 1 2 Automating ADC Manufacturing with Electricity – J. ADC. April 10, 2023. DOI: 10.14229/jadc.2023.04.10.001.

1 Express biolabs
2 ChemTalk

Last Editorial Review: March 31, 2023

Creative Commons License

Article History:

  • Original Manuscript Received January 24, 2023
  • Review results received March 23, 2022
  • Manuscript accepted for publication April  10, 2023

Featured image: Antibody. Photo courtesy: © 2016 – 2023 – Fotolia/Adobe. Used with permission.