New lab photo

Graduation week for Dr Dashwood, Dr Gentry and Dr Ali!

Huge congratulations to Dr Amy Dashwood, Dr Ntombizodwa Gentry and Dr Magda Ali! All three graduating this week with their PhDs from University of Cambridge! Our first Cambridge PhD students, who I find out from reading the acknowledgements were known as "Adrian's Angels" or "The Three Musketeers". Fantastic scientists all, I'm really proud to have been part of their career journey. I look forward to following their successes into the future, already started with a postdoc at the University of Manchester, a postdoc at the MRC Laboratory of Molecular Biology (LMB), and a commercialisation position at Cambridge Enterprise. Well done!

Understanding tissue migration

We have an exciting new bioRxiv story that just went live! This one takes a computational immunology approach to understanding tissue-resident lymphocytes. The story highlights the extra value mathematical modelling can bring to biology.

It starts with a large multi-tissue multi-timepoint parabiosis experiment we ran to understand tissue Tregs. Václav Gergelits, lead author on the study, saw greater potential in this dataset to understand the kinetics of lymphocyte migration broadly.

We extracted turnover data for CD4, CD8, Treg, B cells and NK cells from 17 tissue sources, and generated a sophisticated model of migration, activation and death for each lineage and tissue. The Markov chain modelling found high-confidence solutions that matched the empirical data beautifully.

This tells us a probabilistic model and three distinct states (resting/activated/resident) are sufficient to recapitulate the complex migratory and tissue-residency behaviour of these cells. The cell states change probabilities, but the behaviour is still *probabilistic*.

This means lymphocytes do not have a residency "clock". We can measure the average dwell times for resident cells, but if this average residency time is 3 weeks, it does not mean cells have a 3-week timer. It means the cells have a daily probability of leaving that gives a 3 week average. The dice roll comes up earlier for some cells than for others, within those cells being intrinsically different. Like radioactive decay of atoms, it is just probability - there is nothing intrinsically different about the uranium atoms that decay after a week vs those that decay after a million years, they just had different rolls of the dice.

This approach can explain much of the variation in cell fate without needing to invoke cellular heterogeneity! Two identical cells can have highly divergent outcomes simply because of probability, without different underlying biology. In fact, we can create thousands of identical "digital cells", model them with these simple rules, and we get the empirically-observed range of dwell-times. There is no need to invoke TCR clonality or the like - it is simply an emergent property of cells with probabilistic kinetics!

A great example of applied mathematics informing biology!

Take a read of the pre-print here.

Understanding vaccination in transplant patients

We have a new systems vaccinology paper out at npj Vaccines!

The study tackles the problematic question of why transplant patients responded so poorly to the COVID vaccine. While most people had great antibodies from a single dose, only half of transplant patients have responded even after three!

We took blood from 20 healthy, 31 lung transplant and 59 kidney transplant patients prior to vaccination, and profiled 444 immunological parameters, to get a comprehensive systems immunology profile. We then followed who did and didn't respond to the vaccine, to find the immunological associates.

First up, there are clinical effects: Vaccine response was especially poor soon after transplantation, and in patients on immunosuppressive cocktails, especially those including MMF. Even taking this into account, there were immunological drivers associated with poor response.

As you might predict, the patients that responded best were those with an immune profile that had returned closer to normal post-transplantation. In fact, you could predict vaccine response with 93% accuracy just based on 10 immune parameters.

 

Oddly though, some patients were able to hobble together a poor but detectable response after two shots. These patients didn't have a more normal immune profile, and had quite unusual relationships between immune populations, suggesting that they had put together a poor-but-functional "kludge".

This study was a joint initiative from our lab, Arnaud Marchant's lab at at ULB and Stephanie Humblet-Baron's lab at KU Leuven. 

Huge thanks to all team members, especially Nicolas Gemander, Julika Neumann, Rafael Veiga and Isabelle Etienne for their leadership roles.

Biggest thanks of all to the patients who volunteered for the study!

Read the full paper here

Lab punting

Congratulations Dr Magda Ali!

Congratulations Dr Magda Ali! Our latest successful PhD viva was from the amazing Magda, finishing off a great PhD.

Well done Magda!

Moments before thesis submission...

Lab retreat 2024

Congratulations Dr Ntombizodwa Makuyana!

Congratulations Dr Ntombizodwa Makuyana! Our 24th PhD completion from the lab, and our 1st Cambridge PhD defence! Way to go Tombi! 

Designer IL2 mutations for clinical use

We have a new pre-print out, on novel designer IL2 mutations!

IL2 is a powerful immunomodulator, but the dual roles make it complex to use, and many designer mutations reduce bioactivity or result in poor production. This makes it much harder to move these muteins to the clinic.

Rob van der Kant, Joost Schymkowitz and Frederic Rousseau from the VIB Switch lab took up the challenge of designing new IL2 mutants that not only improve the specificity for Tregs or CD8 T cells, but also maintain bioactivity and actually improve production capacity. They sent the designs over to us to test!

Great work from Amy Dashwood screened these mutations designs in vitro, with in vivo testing by Ntombizodwa Makuyana resulting in a set of mouse and human IL2 muteins with the desired biological properties. In particular we solve some of the common issues with IL2 muteins by considering the bound and unbound structures. For example, to make IL2 specific for Tregs, the approach is to allow binding to IL2RA, the high affinity receptor sub-unit used by Tregs, while block binding to the IL2RB used by CD8 T cells. The problem is that Tregs also need the full IL2RA-IL2RB-IL2RG trimer to assemble for optimal signal. So the typical Treg mutein is more specific, but also has poorer bioactivity. We solved this by creating a block between IL2 and IL2RB that moves out of the way after IL2 binds IL2RA, allowing the full trimer to form. These muteins are not only superior to the original IL2 in terms of cellular specificity, but by removing the aggregation-prone regions. By identifiying the aggregation gateway residues and changing them to be aggregation-resistant, we can further improve these muteins by making them aggregation resistance. The net effect is that the IL2 muteins we made are more specific for either Tregs or CD8 T cells, and will also be cheaper and easier to produce - the perfect combo for biologic drugs!

Our take-home message: if you are engineering proteins for therapeutic use, remember to take into account production, aggregation and bioactivity. These factors count when it comes to making a drug!

Take a look at the full paper on BioRxiv.