IT and Database Workshop 01 NOVEMBER 2017

Minutes from GAiT Website and Database Workshop, Berlin, 1 NOV 2017

Delegates

Andreas Kurtz (AK)

Dave Turner (DT) [On phone]

Stephen Sullivan (SJS)

Stefanie Seltman

Robert Muller

Johannes Dewender

Nancy Mah

Amanda Mack [On phone]

Keren Abberton [On phone]

Rick Monsma [On phone]

Tenneille Ludwig [On phone]

Summary

• SJS reiterated the GAiT mission, the role of the website and database play in achieving that mission, and suggested a structure for new GAiT website. He highlighted that the website would have an interface to access a tailored database for clinical grade iPSCs. After a brief update of recent GAiT working group and Board activities, he introduced AK as Database working group lead/Database Workshop Chair and yielded the floor.
• Workshop Chair AK reviewed the rationale for a clinical grade iPSC database tailored to GAiT specifications, then reviewed how the hPSCreg database structure might be changed, augmented, and/or supplemented to meet the need of the GAiT community for a clinical grade iPSC database. He requested that the GAiT Quality Standards and Regulatory working groups decide together with the Registry on Heading, Tab, Field and Data Format edits required and described design options he wished considered by potential GAiT database users.
• SJS highlighted that QC Release Criteria had been agreed at the Quality Standards workshop in Sheffield 5 OCT 2017, and the database would list Release Criteria as mandatory fields. Additional criteria ‘For information only’ would also be listed in the relevant fields but these would not be essential to complete a database record. However, if Release Criteria were amended in future, in response to, for example, technological advancement or a change in scientific interpretation, the database structure would need to reflect such evolution.
• AK and DT listed Clinical Grade iPSC fields that would need to be included (these included immunogenetic factors separated into 4 categories: blood group, human leukocyte antigen (HLA) Classes I and II, and killer cell immunoglobulin-like receptor (KIR) receptor ligand.
• How data would be deposited and accessed within the database was discussed. Further attention to identify and instigate appropriate measures to prevent data misuse were warranted and how such measures would be enacted required further consideration, especially with respect to the European GDPR.
• A list of all mandatory fields for the database would be needed before data validation of records can take place. This will include GAiT QC Release Criteria, but all other mandatory fields will also need to be identified (e.g. mandatory criteria for regulators).
• Use of components within MIRACAM II matrix will allow easier data conversion and copying with partner databases / interoperability (e.g. those of CIRA and Rutgers).
• Database user-management will be managed through the website and associated security ramifications should therefore be assessed preemptively. Both the interface and database should be updateable and users of the database and website should both be independently trackable. How this is managed in other databases should be assessed.

AK estimated GAiT would need to employee one database technician/programmer for one year to modify the hPSCreg platform to GAiT’s purpose, test and validate it. Estimated Cost: 75.000 – 80.000 EUR.

Actions:

• SJS to continue to develop the description of the website and liaise with AK and the Board before putting the website development out to tender.
• SJS to solicit expert advice on security requirements for the website and database. While the issue of policing entrance to the GAiT community (and as a direct consequence the ability to register lines in the database) was discussed at the workshop, it was subsequently clarified that the database would have no personal or imputable information, and this would simplify security issues. Blood, Tissue and Cell Services will not divulge the identities of their donors nor medical information – they provide a donation number as the linkage and the recipient organisation can be assured that the donation complies with the donor selection and screening criteria for that institution (which are in the public domain).

In addition, there are strict national and international data protection laws around the sharing of personal information with heavy penalties for transgression. GAiT cannot act as custodian for this kind of information.]

• SJS will circulate the finalised Minutes and associated manuscript from the GAiT Quality Standards workshop, Sheffield 5 OCT 2017 to better inform on mandatory and optional fields for the GAiT database.

With regards to the database, generally speaking the GAiT Donor Selection and Regulatory working groups needs to provide a detailed list of Subject Fields and information on the data format expected to be submitted under each field.

• SJS to ask GAiT Quality Standards working group to review hPSCreg fields under Derivation and Culture headings and suggest modifications.
• SJS to ask Quality Standards and Regulatory working groups:

o To review fields under ‘Donor Selection’ and ‘Regulatory’ headings and determine if these are sufficient, require more refinement, and to identify what is missing. AK can provide a list of existing fields in a readable format (e.g. Excel)

o Are LMS/MCB/WCB data to be stored independently in supplementary GAiT database or are links to other Banks sufficient?

o What field changes are required under the ‘Characterisation’ heading for clinical grade iPSC lines (e.g. those for Release criteria, Batch associated test data, or Cell line history file data).

o What assays are to be listed within the database? For example, what are the accepted criteria for demonstrating clinical grade iPSC pluripotency? What are the appropriate fields to use in the database? There are a large range of tests and GAiT must remain neutral.

o Which fields are to be included under Genotype subheadings: Karyotype, STR / SNP authentication, Genome data, and HLA data?

• SJS to solicit expert advice on security ramifications for accessing the database through the website as part of the development plan for the website. [Update: While AK later added his team has a data consultant in hPSCreg to interact with, MLT highlighted that security ramifications would be extensive and profound if users put personal medical information associated with traceability and imputability into the database. It is best to not do this, and is unnecessary to achieve GAiT’s mission in any case.]

Detailed Outline

1. Introduction, meeting objectives, and general update on GAiT activity
2. Rationale & Design for GAiT website
3. Rationale for GAiT clinical grade iPSC database and design options
4. Augmentation of Database Structure to include Immunogenetic Factors pertenent to Donor-Recipient Matching
5. Database-related issues arising from recent Quality Standards workshop
6. Data conversion/transfer and data user management
7. Close and Supplementary Questions

1. Introduction, meeting objectives, and general update on GAiT activity

SJS thanked AK and his team for hosting this GAiT Website and Database workshop. The workshop was arranged to precede the EBiSC consortium “Scalability of iPSC technology for future drug discovery & therapy”.

The objectives for this GAiT workshop were summarised as (i) discussing of development options for building the GAiT website and the GAiT clinical grade iPSC database with particular emphasis on what GAiT associates actually require and (ii) compiling a list of action items to progress their further development and instigation.

The mission and structure of GAiT were reiterated. A review of recent GAiT activity was undertaken. The recommendations from the Quality Standards Workshop 23 OCT 2017 would be circulated shortly as part of wider consultation, a detailed manuscript was being prepared and would be circulated in due course with a view to submitting it for publication early in the new year All those contributing to the effort either through meeting attendance for assisting with developing the manuscript would be listed as authors.

At the most recent GAiT Board meeting, Prof Johan Hyllner retired. He was replaced by Dr Jacqueline Barry, Chief Clinical Officer of the Cell and Gene Therapy Catapult (CGT). Other associates had been appointed to fill Dr Barry’s previous role as Regulatory working group lead.

GAiT is seeking to build a consensus across many professions and functions and this it can only do by listening to all stakeholders and developing holistic strategies. The Board considers all stakeholders equal even if their proportion of contribution or ability to assist GAiT may be different. The success of GAiT hinges on building a framework that is useful to all stakeholders involved in bringing new iPSC-based cellular therapeutics to the clinic.

A GAiT Industry Liaison is scheduled to coincide with Phacilitate/WSCS conference22-25 JAN 2018. This will be chaired by Director Michael May. More details would be included in the December GAiTpost newsletter.

A Donor Selection meeting would be held at CGT, London on or around 12 DEC 2017. This will focus on developing registry relationships. While GAiT has engaged previously on donor characterisation, there is also much to consider when seeking to identify appropriate donors from registries.

The organisational structure of GAiT was reviewed. The role of Member organisations (NYSCF, CGT, KHiB, CCRM) and Chair Prof Marc Turner (Medical Director of the SNBTS) was described as well as the present configuration of the current working groups: Quality Standards, Database, Donor Selection, Regulatory. It was noted that Working GAiT Group structures would evolve to suit efficient tackling of different projects.

1. Rationale & Design for GAiT website

A putative design developed by SJS was reviewed. Most of the website, except for the homepage, contact page, and social media tie-ins, would be private (i.e. password protected for GAiT associates only). The primary functions of the website are to assist communication and coordination of effort between GAiT associates worldwide. The website would also contain a secure interface to the GAiT clinical grade iPSC database.

Portals would host current/past newsletters, and describe future planned events. The website would have 5 increasing ranges of user access (allowing greater access to all information on the website and database interface – (i) public (access to home and contact pages), (ii) associate (password protected access to portals, searchable discussion boards, and the database, (iii) Board member (password protected access to associate pages as well as the Board portal which would contain Financial Reports and Board meeting minutes), Chairperson/ILO (allowing Board member access with additional editing capacity), and Administrator (complete access to the sites pages and code editing functions for the website administrator and database interface administrator).

Thus, any GAiT associate will be able to access the website and database with a password, reducing the necessity for regular update emails. Searchable discussion boards allow GAiT associates to assist and inform each other on topics pertinent to GAiT’s mission. Additionally, Board members would be able to access Board minutes and financial reports. Board meeting minutes and other materials.

No objections were raised to this design. SJS introduced the GAiT Database Working Group Lead Prof Andreas Kurtz chair for this meeting and yielded the floor to him for the remainder of the Workshop.

1. Rationale for GAiT clinical grade iPSC database and design options

AK stated that the rationale for the GAiT clinical grade iPSC database to allow the GAiT community to collect, organize, access, validate and compare data from clinical grade iPSCs admitted into the database. Data would span the foci of GAiT working groups (e.g. The Donor Selection committee would focus on data relevant to donor selection, ethics and regulation, the Quality Standards working group would focus on data relevant to safety and QC, characterisation and banking data, and the Regulatory working group would focus on data relevant to management and expedition of clinical trials, as well as those related to applications and uses.

The release criteria agreed by the Quality Standards workshop 5 OCT 2017 would be mandatory fields in the database, the ‘For information only’ criteria would be optional but all fields mentioned in the meeting would be included, in addition a supplementary field would also be added for criteria not mentioned explicitly at the Quality Standards workshop but which may be used by associates to add other information voluntarily.

Options for designing the Database Interface would depend on whether or not only GAiT associates could register lines. SJS stated that GAiT currently had a policy that anyone interested in producing clinical grade iPSC lines could join, so this could not be an issue as anyone wishing to register lines in the database could become a GAiT associate.

AK stated the database will be accessed through a sub-interface on the GAiT website, similar to what has been done with EBiSC. The internal part of the database is where data will be stored and can be accessed by GAiT associates that have access to the associate related private sections of the website via a username/password security log in. Tools for validation will be part of the internal component of the data base and these allow the deposited data to be validated for completeness. External users will be able to search for cell lines and data in the website but only GAiT associates with a password can input data. The record would only go ‘live’ to external users when the data in the record has been validated and deemed complete, with consent by the depositors.

The database structure will reflect that of the Human Pluripotent Stem Cell Registry (hPSCreg) but will have additional fields for pertinent Clinical Data. In the Registry, fields are listed under the following headings: General Information, Donor Information, Ethics/Usage, Derivation, Culture Conditions, Characterisation, Genotyping and Genetic Modification. The Human Pluripotent Stem Cell Registry allows searching for particular research grade iPSCs and for users to register projects to particular lines. AK noted that database users could also detail their own experiences with particular lines (feedback), thereby fostering greater awareness of the lines robustness, consistency, and applicability within the GAiT community.

AK described how cell lines would be registered in the GAiT database. [Update: MLT highlighted that the information individual records would have in this database would be significantly different from that of one set up for research purposes. He repeated personal medical information associated with traceability and imputability would not be contained in the database].

There might come a time where the Board would seek to ensure the legitimacy of those seeking to join the GAiT community, but presently the community is so small that the GAiT Board & Chair know all associates personally. The clinicaltrials.gov registry currently has an issue with stem cell tourism fraudsters legitimising their activity by registering ‘trials’. SJS will bring this issue to JB’s and the Regulatory working group’s attention.

AK then reviewed the fields in the hPSCreg database under the Cell Derivation and Culture headings: Under the Cell Derivation heading, they have Source cell, Vectors/vector clearance (i.e. Vector Specifications, Integration loci, Selection conditions used, Integration anaylsis fields. Under the Cell Culture heading, they have Media used, Materials used in cultivation, Feeder or coating/substrate used, Expansion rates, Associated viability rates of the cell, Culture labratory specifications and accrediation (GMP/non-GMP). Cryopreservation, Shipping, or Licensing and Management System (LIMS) data do not currently have Fields in the HPSCreg database.

AK stated that while LMS data could be stored in its own database, in his opinion, the GAiT database should just provide a link to the data in other banks, users can then request data from them directly should it be required. The same would apply to Batch data, Master Cell Bank data or Working Cell Bank data.

Under Characterisation, Quality Control and Regulation Heading, on the Characterisation Tab, the hPSCreg database has the following fields: Expression of Markers (Markers associated with Pluripotent Stem Cells), Differentiation potency, Viability, Microbiology, Sterility and Virology. The GAiT database will need to be developed to be more eleborate in derms of the breath and depth of data that should be included. It presently does not have fields for Release criteria, Batch associated test data or Cell line history file data. Cell line history files could be automatically co-delineated based on available data or a suitable link could just be provided.

The data are mapped as much as possible using standard terms, taxonomies or ontologies with the aim of increased searchability and comparability. Further codofication will be developed for new data fields defined by GAiT. Data output needs input from GAiT as hPSCreg is in the process of revising the current data output / filter format (e.g. similar to HiPSci).

1. Augmentation of Database Structure to include Immunogenetic Factors pertinentpertinent to Donor-Recipient Matching

Under the Genotyping Tab, AK highlighted that immunogenetic factors would need to be added and suggested the following additional categories added: HLA – type I, HLA -type II, Non HLA, then there would be fields where ninformationinformation on 1st and 2nd allele can be inputtedinputted. AK suggested the following Fields: Blood Group A, B, and O. Under HLA class I AK suggested fields for HLA-A,B and C. Under HLA-classclass II, AK suggested Fields DRB1, DQA1, DQB1, DPA1, DPB1, DM, DO and under Non HLA, he suggested Fields KIR and MICA.

DT introduced himself and his part within the GAiT Donor Characterisation effort. He had given consideration to the HLA data required for lines to be included the the GAiT haplobank system and these would assist identification of which homozygous donors would be useful in different populations worldwide. His list for additional fields to be initially included in the GAiT database the following immunogenetic factors (i) Blood group ABO (ii) HLA groups: full types, ideally 4th field (high resolution) across all loci, A,B,C,DRB1,3,4,5,DQA,DQB,DPA,DPB and (iii) KIR Ligands, C1/C2, Bw4/6. DT suggested when it came to the HLA data, he thought it pertinent to collect as much information as feasible. He mentioned recent publications and discussions concerning homozygous donors and the risk of activation of NK cells in recipients, so he feels it worthwhile to describe the KIR ligands that are present in the donor. While one could extrapolate from the HLA data, DT feels to have them separately listed in the database would be useful. Both AK and DT acknowledged there was significant overlap in what they had independently suggested.

DT then proposed nomenclature for the HLA types to be included in the database. Ideally, DT suggested GAiT records the highest resolution data (4th field resolution type) for HLA types, these can be derived by Next Generation Sequence platforms that are now commonplace. DT also suggested additional fields to highlight (i) that the HLA type had been confirmed in the iPSC line in addition to the Donor Material i.e. confirmation that the HLA type remains consistent after culture adaptation and/or nuclear reprogramming procedures where spontaneous mutations can arise. (ii) the method used to identify the HLA type, (iii) the laboratory where typing was carried out along with its accreditation status. Abbreviated code names for what haplotypes individual iPSC lines represent would make searching for particular HLA type easier. DT gave an example of such code names, at AK’s request. AK stated that software could do this automatically when the more detailed HLA form data are added.

1. Database-related issues arising from recent Quality Standards workshop

Issues arising from the GAiT Quality Standards Workshop 5 OCT 2017 that were relevant to the development of the database were then discussed. AK stated that a list of mandatory fields was available for research grade iPSCs in the hPSCreg database, but a more extensive one would be needed for clinical grade iPSCs. What data should be integrated and visible in the GAiT database to evaluate and select a cell line for clinical use? GAIT needs to differentiate between mandatory field inputs, and optional inputs that may be of scientific interest, may be of interest for regulators, or potentially useful in another fashion but not critical at this stage.

AK asked if there should be a patient matching function within the database. SJS stated that there might be insurance-related and legal consequences if such an instrument was developed and used for patient-tissue matching. DT stated it was difficult to develop algorithms for matching HLA because there are so many permutations and he does not think it needs to be something that needs to integrated into the GAiT database presently. [Update: MLT agreed with this, saying such decisions need expert Histocompatibility & Immunogenetics opinion on a case by case basis.] There are other methods for identifying an appropriate cell line for a specific patient or population of interest.

The Quality Standards workshop also highlighted that which assays are to be listed in the database needs clarification.

GAiT should not reinvent the wheel when it came to data standards. A review of what data standards already exist that can be utilised and developed by GAiT? How should assays be defined within the database? AK reiterated the previously discussed question as to what data should be in a central database and what data should remain only at the banks (LIMS, Bank stocks, GMP/Materials certificates)? Finally, AK asked what data should be shared with authorities via the central registry and what data are needed for internal validation?

Data validation to be stored within the database requires minimal acceptance criteria. Would data validation be automated or manually done? Should there be different levels of acceptance (perfect, all is there; acceptable but perhaps certain non-essential data for better risk assessment are not provided; rejected)? An assessment of Regulatory restrictions (use, transfer/exchange) should be undertaken. AK assumes GAiT will issue Data Validation Certificates (cell line file) for Regulators.

AK showed how the hPSCreg database allows intuitive searching for cells lines and data. The lines are represented and organised by predefined data categories. He showed how other important search items / filters (e.g. previous use in clinial trials) could be added. Furthermore, he stated regulatoy restrictions should be considered.

Regarding storing information on Registration and Feedback on Applications, AK stated the EU and other regulators wanted cell records to include a list of clinical trials that a line was currently being used in. Consequently, it would be useful to have a field where a GAiT line can be registered for use in different clinical trails. Finally, he also recommended that a field where user feedback on specific lines be added, however such information could also be shared on the website’s discussion boards.

1. Data conversion/transfer and data user management

MIACARM (Minimal Information About a Cellular Assay for Regenerative Medicine] is a matrix that can be adopted to GAiT’s purpose. MIACARM eases the transfer of data between different platforms.

There is a checklist for stem cell production and QC items in the Source Cell module in MIACARM-II and this was tailored by CIRA and Rutgers to their own requirements. AK wants to use parts of MIACARM-II in the GAiT database but not directly implement the whole matrix. It is a matrix that is open-source, and would make data conversion and copying easier for users.

Database user management will be managed through the website. Security implications will need to be considered carefully. It is important that the interface and database be updateable and that users of the website and database can be independently tracked.

Finally, AK estimated GAiT would need to employ one database technician / programmer for one year to modify the hPSCreg platform to GAiT’s purpose and closed the workshop. Details / workplan can be provided when requested.

1. Close and Supplementary Questions

SJS thanked everyone for their input and reminded delegates that questions could always be submitted after the meeting. The following questions were received from Amanda Mack after the meeting and answered below by AK:

1. At what point is it desirable to register a line? Information may continue to be gathered over time, so is it anticipated lines will be registered and amenable to additional information as it becomes available?

Cells are amendable to new data or changed data any time. The changes will be reviewed and if necessary, validated (if release criteria are affected). There is a mechanism in place to solve conflicting data input. In addition, all changes and amendments to data are stamped traceable.

1. Will the database be designed to accommodate the production of multiple banks for the same line (e.g. MCB and WCBs) to reflect potential differences in characterization criteria?

Yes, the database can accommodate the same line under different synonyms and reflect the connection in the database and to the user. This is currently only implemented for sublines (with the derived-from logic). To implement it for different banks will need some modification in the database.

1. It will take time to gain clarity from regulatory authorities regarding absolute acceptance criteria. So how will this be determined with regard to accepting registered lines within the database or labelling them as acceptable or not? Given the range of regulatory perspective globally, what “standard” would be established? Would it be based on the most rigorous perspective?

This can only be answered by interacting with regulators and persons/researchers with experience in having lines accepted for clinical use. This does not necessarily mean to apply the most stringent criteria. The goal is to have a synthesis of the different regulators and then define the criteria, which would fulfil the requirements in all participating states (the highest).

As a first step, it is possible to define the necessary and sufficient criteria within the GAiT group and ‘grade’ the lines according to their compliance to these GAiT standards. The standards should be high.

We thank the GAiT working groups (Database, Quality Standards & Controls, Histocompatability & Immunogenetics, and Regulatory Compliance) for their efforts developing the structure for the clinical grade iPSC-database.