AbDEN

Why Antibodies?

Monoclonal antibodies are at the forefront of treating many diseases due to their high specificity and potency as compared to other drugs. For viral diseases, they can interrupt the pathogenic cycle at many steps. For Flaviviruses like Dengue, mAbs work by mechanisms like blocking viral attachment to cell surface receptors, interrupting viral membrane fusion, or calling for effector functions.

For a disease like Dengue, using monoclonal antibodies is a challenge due to the phenomenon of Antibody-dependent Enhancement, which may be because of sub-neutralizing concentrations of neutralising antibodies, and the presence of non-neutralizing antibodies that signal immune function. This serves to deliver the virus to one’s cells, and allows the virus to infect and proliferate in our cells. However, this does not mean that the favourable qualities of monoclonal antibodies cannot be utilised at all. In addition to their other properties, mAbs are also incredibly modular. Most aspects of their structure and function can be manipulated through protein engineering - which is what we decided to do.

Engineering the Antibody

We posited that our final drug should:

  • Neutralise the virus
  • Not cause antibody-dependent enhancement

Our first proposal was to synthesise an avian homolog of IgG antibodies: Immunoglobulin Y - targeted towards an appropriate epitope of the Dengue virus that ensures neutralisation. There is evidence that IgY antibodies targeted towards Dengue work as therapeutic agents and do not cause Antibody-dependent Enhancement. This would also be the first instance of production of an avian antibody in bacteria.

To be sure of what we were trying to do, we met immunologists Dr Satyajit Rath and Dr Vineeta Bal from our institute in April to discuss the proposal. They raised some concerns:

  • IgY is from an avian system, and had the potential to be very immunogenic in humans, inducing an immune response against itself. Therefore, it would be fair to assume that the patient would have a strong humoral response against it upon second administration.
  • It may be quite expensive to synthetically manufacture, compared to extracting it from chicken egg yolk, which, despite its problems, is quite simple and low-tech as a method.

We also met Dr Alexander Taylor, who has worked extensively on IgY antibodies. At the time, we were thinking of an intravenous treatment, and were advised that immunogenicity would be a major issue that would arise for IV-delivered IgY. A usual immune response breaks down the antigen to be processed by Antigen-presenting cells (APC) using MHC Class II pathway. IgY, a foreign body, would be treated as an antigen when introduced intravenously and hence dosing with IgY would generate an IgG response.

Fig: A visual representation of the immune cascade response initiated by the immune system of our body upon the recognition of IgY antibodies as a foreign antigen leading to the production of anti-IgY IgG antibodies resulting in the destruction of the IgY antibodies.



We then shifted to aglycosylated versions of IgG to achieve the same functionality as IgY: neutralisation, but no effector function, thus preventing Antibody-dependent Enhancement. Glycans in the CH2 domains of IgGs are critical for the immune effector function of IgG antibodies. Removing the glycan via mutation of glycosylation sites, or production in bacterial systems, would generate aglycosylated antibodies. Aglycosylated IgGs have been proven to be effective against Dengue in mouse models. Currently, some aglycosylated IgGs are even in clinical development for other diseases.

However, as we learnt more about antibody production over the summer, we realised that the production of our antibody - a large, complex, multimeric protein - would still be fairly costly. Dengue primarily affects poor tropical countries, so creating an expensive therapeutic would contribute close to nothing toward solving the problem.

We finally settled on a single chain antibody fragment called an scFv (single chain fragment variable). These are the isolated two variable regions (VH, VL) of the antibody connected with a short, flexible linker. They possess the same antigen binding properties and no effector function. They will act as purely neutralising agents, potentially with higher potency due their ability to access inaccessible epitopes. They are also simple, monomeric proteins, which would make them cheap and to produce in bacterial systems. Their small size would also reduce the immunogenic response they would trigger in the human body. The problem with scFvs, we realised, was their half-life. They are below the renal filtration limit, and can thus get filtered out of the body at very high rates, giving them a half-life of a few hours in mice. To combat this, we started investigating avenues to increase the half life of antibody fragments.

The half-life of both the long lived serum proteins - albumin and IgG - are governed by the neonatal Fc receptor (FcRn). Binding with the receptor occurs under conditions of low pH within proteolytic vesicles, and lets the protein be transported back to the surface of the cell, where the protein dissociates at physiological pH.

Fig: A visual representation of the FcRn mediated recycling of IgG and serum albumin that contributes toward their high serum half-life.

We found a small peptide that we could attach to our scFv that would bind to the FcRn receptor in the same pH-dependent manner, leading to the recycling of the scFv in the body. We decided to call this assembly a NeoFv.

This molecule ticks all the boxes for a good therapeutic.

Choosing the Epitope

For our antibody therapeutic, we had to find an epitope linked to the dengue virus for it to target. We found that most neutralising antibodies target the E protein, or the envelope of the virus.

Fig: A representation of the viral envelope of DENV. The outer envelope layer consists of homodimers of E proteins arranged as trimeric rafts. There are three domains in the E protein which is preceded by the membrane layer consisting of the M protein. These domains could be potential targets for antibodies.

There are multiple domains of the E protein. Based on our literature review, we found that we could target the fusion loop epitope (FLE), which is highly conserved across all dengue virus serotypes and strains, and is vitally important for the virus to be able to escape and infect cells. We also discovered Ab513, an antibody that was engineered to target the third domain of the E protein - EDIII - to obtain a cross-neutralising antibody. We reached out to experts in Dengue virology to confirm our findings. We met Dr Vidya Mangala Prasad from the Indian Institute of Science, Bangalore, who works on flaviviral structural virology.

She informed us that the FLE tends to be quite sequestered on the surface of the virus, which might make it challenging for the antibody to access it. She suggested that we target a quaternary epitope on the Dengue virus: at the interface of the dimer of the E protein. Antibodies that target the E-dimer epitope (EDE) are rare, highly potent, and cross-neutralising.

We finalised the choice of epitope and antibody - EDE and the EDE-targeting antibody, C10.

We took the variable regions of the C10 antibody and designed an scFv construct from it, adding the FcRn-binding peptide. We used the gene synthesis service from IDT to synthesise the gene for expression in E.coli.

We also synthesised genes for Ab513-derived fragments as a backup, in case we would only be able to access the E-protein to assay our protein.

All these genes are available in the biobrick registry.

Chassis

For the purposes of this project, we decided to use bacterial production for our antibodies. Bacteria provide several advantages as compared to eukaryotic cells: a shorter growth time, high genetic manipulability and cost-effectiveness.

Bacterial cytoplasmic expression is not generally suited for large, multimeric disulfide-bonded proteins like antibodies because of their reducing cytoplasm. Periplasmic expression is usually preferred when making such proteins, when bacteria are used at all.

We came across SHuffle E. coli. SHuffle is a strain of E.coli that has been engineered to be able to form stable disulfide-bonded proteins in the cytoplasm, by mutating various reductive pathways (trx-, gor-, ahPc*) and over-expressing disulfide bond isomerase cDsbC, creating an oxidising cytoplasm where complex proteins can be folded correctly. The inventors of SHuffle have successfully produced full-length antibodies in the cytoplasm of the cells.



Fig: A description of the various pathways introduced within the cytoplasm of SHuffle to ensure the proper production of antibodies there at higher yields.



scFvs are smaller and simpler than full-length antibodies, but they still have internal disulfide bonds, so SHuffle still offers the advantage of being able to produce a higher fraction of soluble protein, as compared to other strains of E.coli.

We reached out to Dr Mehmet Berkmen of New England Biolabs, the inventor of the strain. He very generously provided us with answers to our technical questions regarding the handling and functionality of the strain. He also went on to become our secondary PI and provided us with guidance and support throughout the iGEM cycle.

Construct

Guided by existing literature on full-length antibodies in SHuffle E.coli, we created pET21b constructs, with the heavy and light chain of full-length antibodies in a single operon under a strong T7/Lac promoter.
We also used this same pET21b framework on the final construct for our scFv and our NeoFv. All our proteins had 6XHis-tags, so they could be purified by an Ni-NTA affinity column, due to the ease of the method rather than any therapeutic concerns. We attached the His-tags at the C-termini of our proteins. For full-length antibodies, they were attached at the C-terminus of the heavy chain, as has been done in literature before.

Fig: A representation of the construct of the plasmid used to produce scFvs and NeoFvs within SHuffle.

In-silico Screening of NeoFv Variants

The half-life modulating peptide on our scFv can be arranged in multiple combinatorial possibilities. The peptide itself can be mutated to have a cyclising disulfide bond and to bind to the neonatal receptor in a pH-dependent manner. It can be put at the N-terminus, C-terminus, or within the linker of the scFv. We performed Molecular Docking simulations on all the NeoFv (scFv-FcRn binding peptide) variants generated through AlphaFold 2.0 against the FcRn receptor and identified the best variants through the docking score.

To validate our results, we performed Molecular Dynamics simulations of the complexed structures to calculate the energy of separation. A detailed description of this is available on our modelling page.

Hence, we identified the best NeoFv variant to synthesise, clone and assay, which was - VH-linY12H-VL

This represents the scFv with the binding peptide within its linker. The peptide lacks the cyclising mutation, it is linear in conformation. It also has a tyrosine-to-histidine mutation at the 12th position in the peptide, which allows the formation of a salt-bridge at pH 6, leading to pH-dependent binding of the peptide to the receptor.

Optimising Parameters for Production

In order to obtain the maximum yield of our protein and make the process as efficient as possible, we designed experiments for multifactorial analysis of the values of the variables that govern the yield of production of our antibodies. A detailed description of this is available on our Modelling page.

References

  1. Campbell LP, Luther C, Moo-Llanes D, Ramsey JM, Danis-Lozano R, Peterson AT. Climate change influences on global distributions of dengue and chikungunya virus vectors. Philos Trans R Soc Lond B Biol Sci. 2015 Apr 5;370(1665):20140135. DOI
    PMID: 25688023; PMCID: PMC4342968.
  2. Symptom control for the treatment of dengue
  3. Neglected tropical diseases
  4. Dengue vaccine
  5. Dengvaxia controversy
  6. Antibody dependant enhancement
  7. Kulkarni R. Antibody-Dependent Enhancement of Viral Infections. Dynamics of Immune Activation in Viral Diseases. 2019 Nov 5:9–41. DOI
    PMCID: PMC7119964.
  8. Phadke R, Mohan A, Çavdaroğlu S, Dapke K, Costa ACDS, Riaz MMA, Hashim HT, Essar MY, Ahmad S. Dengue amidst COVID-19 in India: The mystery of plummeting cases. J Med Virol. 2021 Jul;93(7):4120-4121. DOI
    Epub 2021 Apr 6. PMID: 33788285; PMCID: PMC8250513.
  9. Yan G, Lee CK, Lam LTM, Yan B, Chua YX, Lim AYN, Phang KF, Kew GS, Teng H, Ngai CH, Lin L, Foo RM, Pada S, Ng LC, Tambyah PA. Covert COVID-19 and false-positive dengue serology in Singapore. Lancet Infect Dis. 2020 May;20(5):536. DOI
    Epub 2020 Mar 4. PMID: 32145189 PMCID: PMC7128937.
  10. Bates A, Power CA. David vs. Goliath: The Structure, Function, and Clinical Prospects of Antibody Fragments. Antibodies (Basel). 2019 Apr 9;8(2):28. DOI
    PMID: 31544834; PMCID: PMC6640713.
  11. Fink AL, Williams KL, Harris E, Alvine TD, Henderson T, Schiltz J, Nilles ML, Bradley DS. Dengue virus specific IgY provides protection following lethal dengue virus challenge and is neutralizing in the absence of inducing antibody dependent enhancement. PLoS Negl Trop Dis. 2017 Jul 7;11(7) :e0005721. DOI
    PMID: 28686617; PMCID: PMC5517069.
  12. Kelly, V. W., & Sirk, S. J. (2022). SHORT FCRN-binding peptides enable salvage and transcytosis of scFv antibody fragments. ACS Chemical Biology, 17(2), 404–413. DOI
  13. Balsitis SJ, Williams KL, Lachica R, Flores D, Kyle JL, Mehlhop E, Johnson S, Diamond MS, Beatty PR, Harris E. Lethal antibody enhancement of dengue disease in mice is prevented by Fc modification. LoS Pathog. 2010 Feb 12;6(2):e1000790. DOI
    PMID: 20168989; PMCID: PMC2820409.
  14. Ju MS, Jung ST. Aglycosylated full-length IgG antibodies: steps toward next-generation immunotherapeutics. Curr Opin Biotechnol. 2014 Dec;30:128-39. DOI
    Epub 2014 Jul 16. PMID: 25035939.
  15. Lobstein, J., Emrich, C.A., Jeans, C. et al. SHuffle, a novel Escherichia coli protein expression strain capable of correctly folding disulfide bonded proteins in its cytoplasm. Microb Cell Fact 11, 753 (2012). DOI
  16. Robinson MP, Ke N, Lobstein J, Peterson C, Szkodny A, Mansell TJ, Tuckey C, Riggs PD, Colussi PA, Noren CJ, Taron CH, DeLisa MP, Berkmen M. Efficient expression of full-length antibodies in the cytoplasm of engineered bacteria. Nat Commun. 2015 Aug 27;6:8072. DOI
    PMID: 26311203; PMCID: PMC4560801.
  17. Immune response