Welcome to the Abvance II wiki home page.
This page was created on 30 April 2018.
This 'landing page' provides a top level view of the project and includes links to subsections and other relevant external content.
The AbVance II wiki is a 'live' document which will evolve in line with the project and is the 'go to' place for all the latest information.
Use the "edit" icon at the top of this page to edit and the "create" icon on the left to create a new page.
Contact richard.norman@pistoiaalliance.org if you have any questions about this wiki, including if you do not see the "edit" icon.
Proposal
The successful development of antibody molecules into approved drugs relies on optimizing a series of biophysical and pharmacological properties before the start of clinical trials. Failure during clinical trials or during the associated later stages of development is expensive. As such, early prediction of ‘developability’ risk is of key importance.
Our confidence in the ability to predict developability problems early by surrogate biophysical assays is limited by the low number of molecules with associated biophysical data and an agreed standard way of measuring this data across the industry. In addition, key pharmacological properties such as human and cyno PK (clearance), viscosity and solubility, and an understanding of whether these endpoint data could be predicted using biophysical surrogate data is lacking. Similar assays are used across the industry to predict developability risk (see Jain et al., 2017: http://www.pnas.org/content/114/5/944), but institutions use different methodologies. Nevertheless, commonality can be expected among differently implemented assays of the same type. For example we can expect some level of consistency in rank ordering of molecules by Hydrophobic Interaction Chromatography (HIC) even if different resins are used.
There is an opportunity to collectively identify the most predictive biophysical surrogate assays to understand variability among these, benchmark and establish best practices. In the first instance, this would act as an incentive for methods standardization and could subsequently encourage the building and growth of a common database to impact future antibody drug development.
Objectives
Based on the proposal we are working on our approach to achieving our objectives:
>Agree a standard/common way across industry of measuring data for 'well-known' biophysical assays.
>Share biophysical data generated by the above assays for sets of molecules contributed by industrial partners.
>Share associated endpoint data (e.g. human PK) for the above sets of molecules.
>Understand whether there is line of sight (LoS) from early selection (high-throughput) assay data to endpoint data.
>Understand whether we can predict endpoint data from sequence/structure.
The team
Project Team Meeting: Agendas & Minutes
Our Project Team Meetings are held every Tuesday, 17:00-18:00 (Central European Time)
If you would like to participate, contact the Project Manager: richard.norman@pistoiaalliance.org
Name | Organisation | Location | Role | Relevant experience & interests |
|---|---|---|---|---|
Qing Chai | Eli Lilly | San Diego, US | ||
Craig Dickinson | Eli Lilly | San Diego, US | ||
Francisco J. Fernández | Abvance Biotech | Madrid, ES | Chief Executive Officer | CEO. Structural biologist. Scientific leadership to the antibody drug discovery, business strategy and partnering programmes. |
Laura Hook | GSK | Stevenage, UK | ||
Bryan Jones | Eli Lilly | San Diego, US | Research Fellow | Group Leader for Protein Bio-Sciences group, responsible for expression, purification, biochemical and biophysical characterization of potential therapeutic proteins and antibodies. |
Helen Li | Eli Lilly | San Diego, US | ||
Chris Lloyd | MedImmune | Cambridge, UK | Protein engineer. Research Developability lead for MedImmune. Works on mAbs, bispecific's, ADCs and other novel formats across various therapy areas | |
Carmen Nitsche | Pistoia Alliance | Business Development Lead | ||
Richard Norman | Pistoia Alliance | Gjøvik, NO | Consultant Project Manager | Project Manager. Background in structural biology, and small molecule drug discovery and early development at AstraZeneca. Interested in drug development of biologics, data sharing and the impact which Machine Learning and Artificial Intelligence will have on the Life Sciences. |
Abhinandan Raghavan | Novartis | Basel, CH | ||
Kieran Todd | GSK | Stevenage, UK | Senior Investigator | Bioinformatician providing support to all aspect of the antibody development process, ranging from target identification through to patent support but most heavily focused on lead molecule identification and/or design. |
Problem statements (and scale)
“Our confidence in the ability to predict developability problems early by surrogate biophysical assays is limited by the low number of molecules with associated biophysical data and an agreed standard way of measuring this data across the industry.” | 100% industry members polled AGREE | 0% industry members polled DISAGREE |
|---|
“Key pharmacological properties such as human and cyno PK (clearance), viscosity and solubility, and an understanding of whether these endpoint data could be predicted using biophysical surrogate data is lacking.” | 43% industry members polled AGREE | 57% industry members polled DISAGREE |
|---|
Value proposition
>Sharing data from mAbs which are poorly behaved would provide added value at least beyond the Jain et al., 2017 publication as they considered only reasonably well behaved mAbs.
>Developing a minimum set of ‘value’ assays, which can be used going forward, having previously understood the data output overlaps and correlations between assays currently used in each organisation.
>Organisations have a variety of early high-throughput assays which provide a qualitative sense of whether a molecule should be progressed but what is lacking is rigour in endpoint prediction from these assays. Generating enough data from enough molecules with varying properties/behaviours will enable higher confidence in endpoint data prediction early thus enabling earlier decision-making on developability of a molecule.
Scope
| In scope | Out of scope |
|---|---|
>Molecules: standard and non-standard mAbs (i.e. bi-/multi-specific's, ADCs, IgG/Fc fusions). >Data: Chemical & structural stability, Aggregation, Toxicity, PK and Immunogenicity. >Assays: early selection (high-throughput), detailed profiling (bespoke, low-throughput) and endpoint (pre-candidate nomination). | >Data: Manufacturability (expression yields). |
Strategy
>Creating a database of 'value' assays and associated data (focus on standard mAbs)
- Agree important criteria (associated with developability risk) e.g. from Target Product Profile (TPP).
- Identify assays which provide information on criteria.
- Prioritise which assays provide most 'value' (bang for buck) information with respect to predictive power (critical, decision-making, risk data).
- Select a set of molecules including some with interesting/poor properties
- Share SOPs, guideline metrics and data for 'value' assays on selected set of molecules.
- Test, validate and show correlation with endpoint data
- Generate standard assay data for selected (public or internal from 'dead' projects) molecules (index and normalise using standard molecules)
- Test if correlation hold for non-standard mAbs
>Predicting endpoint data (e.g. in vivo nonspecific clearance rates) from sequence/structure
- Establish current internal practices for use of sequence/structure information and structural descriptors to predict developability risk.
- Highlight 'gaps' which could be addressed through, more data, better data or better understanding of the data (i.e. correlations)
- Combine sequence/structure data with standard 'value' assay data
- Seek correlations between sequence/structure and assay data > assay data and endpoint data
Early risk assessment (challenges)
>Likely biggest challenge is IP and what is feasible to share.
Mitigation: identify low to high sensitivity data and pressure test with Legal departments (internal legal sponsor).
>First hurdle may be willingness within organisations to share data and assay method details.
Mitigation: target appropriate management level (internal executive sponsor) early and pressure test using 'basecase/upscale' scenarios and survey questions.
>Related to the above; willingness to share assay data correlation with endpoint data.
Mitigation: select 1-3 assays for which data is being shared and establish their correlation with PK, viscosity, solubility etc. internally, i.e. build model which can be shared thus circumventing the requirement for sharing primary data.
>Willingness to share sequence/structure information on molecules.
Mitigation: use information from molecules in the public domain and/or 'dead' internal projects. Define and use structure descriptors.
>Reaching agreement on how and where data is stored.
>Reaching agreement on model for ownership of data.
Questions
>What level of agreement is there with the problem statements?
>What data/analysis is there to support/refute the problem statements?
>What level of detail are organisations prepared to share with regards the assays, data, established correlations and sequence/structure?
>How much data do we need to make a difference and how are we going to test this?
>Who will ‘own’ and manage data?
>What internal related efforts are ongoing?
Blog stream
Blog stream
Create a blog post to share news and announcements with your team and company.
Recently updated
