What’s new in immunology?
Insights from COMPASS 2024

Where to from here? Future thinking

Presented by Dr Tim Raine | Prof. Kevin Winthrop | Prof. Iain McInnes

What does the future hold for the treatment of inflammatory diseases? Each of the three international keynote speakers gave the COMPASS audience their views on what developments they are excited about in clinical practice across immunology.

“What can we do to work more closely together?” was the key question from Dr Tim Raine. He paid tribute to his late colleague Dr Deepak Jadon, with whom he had a fruitful clinical collaboration, and put forward that pragmatic, collaborative interactions between gastroenterologists and rheumatologists, can be hugely beneficial for both patients, and for health systems. He described how the well-informed cross-disciplinary discussions their department instituted not only assisted with clinical decisions, such as the management of difficult pain or challenging patients; but also had the potential to reduce clinic visits and costs for shared patients with immune-mediated inflammatory diseases. “It may be a formal MDT, or it may be just swapping numbers,” he said, but the key message was that communication is the way forward from his point of view.

Prof. Kevin Winthrop said that he is intrigued by the ever-expanding range of new immunological targets that are being examined. What keeps things interesting, he suggested, is that regardless of how sure people are about what will be found, very often things don’t turn out as expected, or interactions are discovered that no-one had any knowledge of. His vision of the future is that therapies are going to become increasingly more targeted. Immunological therapies have become safer over the last 20 years as they have become more selective, he stated – both in terms of the molecular target, but also in terms of which patient will benefit most from which mode of action.

Dr Raine explained that there are already biomarkers in development that may be able to predict not just efficacy, but also the likelihood of adverse events; giving him confidence that the risk-benefit ratio will become even more favourable as therapies become more highly tailored to individuals. “I see the world in a much better place in terms of patient care than it was 20 years ago. And I think that trend is going to continue, whether artificial intelligence is involved or not!”

Prof. Iain McInnes said that, as a self-described ‘mechanistic reductionist’, he treats each patient as a whole system. However, he finds this is often a ‘bland and superficial’ view, which can be unsatisfactory when making clinical decisions. For this reason, he’s enthusiastic about the exciting methodologies on the horizon, which may help us understand why and how diseases are operating, how these processes can be interrupted, and ultimately how they can be reversed.

He ended with the story of a patient with rheumatoid arthritis (RA) who was part of a clinical trial Prof. McInnes was engaged in during his PhD. When the trial was over, the patient came to his office and enquired why he was never asked ‘why’ he took part. Prof. McInnes admitted that patients were always asked for permission, but their motivation could not be questioned for ethical reasons. Nevertheless, he said, when a large Glaswegian wants to tell you something, it’s advisable to listen. So, the man took a photograph out of his jacket pocket and placed it on the desk. It was a picture of a baby’s hand; his first grandchild. He then placed his deformed, RA-affected hand beside it and said, “Promise me that THIS will never happen to THAT.”

“Where to from here? I say it’s a call to arms. The technologies we are gifted with will take us to a cure.”

– Prof. Iain McInnes

Co-management of shared patients and the role of RINVOQ

Presented by Dr Tina Racunica | Dr Emma Flanagan

Other significant history? Mr A had been diagnosed with Reiter’s in 1973 and experienced iritis of the left eye in 1986. This resulted in residual scarring and restricted pupillary constriction which caused photophobia. He also reported previous episodes of psoriasis of the face and scalp. In 2006, he was diagnosed with diverticulosis and had a left hemicolectomy for colo-vesical fistula in 2006, but in the Rheumatology consultation said that his bowels had not been an issue since surgery. A colonoscopy in 2019 had revealed internal haemorrhoids, but was otherwise normal.

Treatment decision: IL17A (May 2019)
The patient had a reasonable response to secukinumab 150 mg monthly, and that response improved when the dose was increased to 300 mg monthly in March 2020. His BASDAI score dropped from 8.9 in March 2019 to 2.0 in October 2020, and his CRP decreased from 11 to 5. However, in February 2022 the patient had an iritis flare followed by a diagnosis of coronary artery disease that required a drug-eluting stent. Then in March 2022 he was hospitalised with severe rectal bleeding.

Co-management begins
In March 2022, Mr A had been experiencing abdominal pain and diarrhoea and had undertaken a bowel prep prior to a planned outpatient colonoscopy when he had noticed bleeding that became severe, leading to his presentation at the emergency department (ED).

Initial co-management decision: infliximab + MTX
Discussions between the managing specialists determined a treatment course of intravenous infliximab (5 mg/kg at weeks 0, 2 and 6, every 6 weeks) combined with methotrexate (MTX). Mr A initially experienced clinical remission with normal faecal calprotectin and inflammatory markers. However, in May 2022 he developed chest pain and raised inflammatory markers (CRP, 55; ESR, 48) with examination revealing bilateral pulmonary infiltrates. This was believed to be Pneumocystis jiroveci pneumonia (PJP) due to treatment response. Infliximab was withheld until he recovered then recommenced 6-weekly with MTX. But by September 2022 he was experiencing worsening AS features with severe pain in the thoracic and lumbar spine.

Revised management: RINVOQ® (upadacitinib)
Mr A was immunised for herpes zoster and commenced RINVOQ® (upadacitinib) in September 2022: 45 mg induction for 4 weeks, followed by 30 mg maintenance for 4 months, then switching to an ongoing 15 mg maintenance dose. Both his AS and CD symptoms responded well, and he remains in clinical remission: his inflammatory markers have normalised and his BASDAI is 4.7.

What did the experts think of Mr A’s case?
Did Mr A really have AS? Having confirmed with Dr Racunica that Mr A was HLA-B27-positive, Prof. Iain McInnes suggested that some of Mr A’s previously noted symptoms, such as psoriasis and the 1973 diagnosis of Reiter’s syndrome, could classify him as having an “MHC-Class-I-opathy”. That is, exhibiting at different points in his life, overlapping clinical manifestations that are linked to the major histocompatibility complex I (MHC-I) antigen-presentation pathway.1 Prof. McInnes noted that an accurate diagnosis would be very difficult in this gentleman and speculated that the noted failure to respond to previous treatments may, in part, relate to diagnostic uncertainty given Mr A’s wide-ranging, and changing, symptomatology.

Did IL17 inhibition have a role to play in the development of Mr A’s bowel symptoms? Prof. McInnes said that large-scale population data is emerging that suggests that, “We are not seeing a clear correlation between IL17A blockade and emergence of inflammatory bowel disease over and above what we would expect to see in a ‘B27 psoriatic patient population – which this guy is somehow in.” So he suggested he wouldn’t necessarily point the figure at secukinumab for the ensuing gastrointestinal symptoms of Mr A.

Dr Raine however, felt himself duty-bound to point out that there is significant evidence of inflammatory bowel disease (IBD) worsening in patients treated with IL17 blockade.2 So both specialists did agree that IL17A inhibitors should be avoided in people with existing gastrointestinal (GI) symptoms. Although Dr Raine did implore that one should not necessarily be sending everyone with ‘funny GI symptoms’ for faecal calprotectin testing. He explained that, when dealing with a population that have a high pre-test probability of GI inflammation – as is the case with patients with inflammatory syndromes such as Mr A’s –the performance characteristics of the faecal calprotectin test are different. Such tests are designed to perform in a general population, and will definitely be confounded by current use of non-steroidal anti-inflammatory drugs (NSAIDs) for example. So he suggested that even in cases where faecal calprotectin would still be considered useful “…if you can’t get the patient off NSAIDs for at least 2 weeks, don’t send the test, because it will be elevated and won’t tell you anything.”

“This is the kind of case that demonstrates why we need to move on from our historical clinical classifications to molecular tissue-based classifications of what the patient actually has,” said Prof. McInnes, and he expressed the hope that 10-15 years from now that will be the case.

How should potential PJP be approached? Prof. Winthrop suggested that it is unusual to see PJP in patients on infliximab alone and indicated that it is more common in patients who are also being treated with steroids. He explained his usual procedure would be to conduct a bronchoscopy to get diagnostic confirmation, and he wouldn’t usually suggest prophylaxis in patients who are not on steroids.

He also indicated that once PJP therapy was decided on, he would leave them on that treatment for a minimum of three months before re-introducing disease-modifying therapy. His preference would also be to try a different key therapy than the one that was being used when the patient developed the PJP.

Immunology update: beyond cytokines to clinical practice

Presented by Prof. Iain McInnes

“There’s been a decade of progress in inflammatory diseases with many new MOAs…” said Prof. Iain McInnes, “…and yet remission is not the norm, repair is challenging, immune homeostasis eludes us, and it’s all rather expensive. So how do we break through the therapeutic ceiling in IMIDs?”

Prof. McInnes believes we need to better understand immune-mediated inflammatory diseases (IMIDs) before they can be effectively treated. In order to do this, his suggestion is that we starting looking at clinical classifications by molecular state,3 as opposed to broad IMID class or most recent drug failure. Techniques such as single-cell sequencing and transcriptomic analysis are beginning to reveal what individual cell types are doing, and where they are doing it, in different disease states;^4,5 and this is where Prof. McInnes thinks research should be focused.

For example, single-cell analysis of synovial tissue macrophages from RA patients with active disease (versus those who were in TNFi-induced remission) revealed a particular MerTK subpopulation that is able to regulate remission by producing anti-inflammatory proteins, such as resolvin.⁵ Patients who had high levels of these macrophages stayed in remission after treatment withdrawal, and have subsequently remained in remission for several years.

Beyond this, spatial transcriptomic analysis is now allowing researchers to look at individual cells without taking them out of the tissue.⁶ Distinct subsets of myeloid dendritic cells have recently been characterised that reveal differences in cell signalling at the synovium between the states of active RA and remission. An AXL-positive dendritic cell subset has been identified that fails to replenish in RA remission, and researchers are exploring whether the AXL protein may provide a key to moving from ‘RA in remission’ into ‘health’.⁶

However, Prof. McInnes pointed out that RA patients don’t just struggle with their joints. They also have heart disease, pain, periodontal disease, osteoporosis, and other comorbidities. In particular, he noted that autoimmune diseases and obesity are both increasing in prevalence, and he suggested, are very likely linked.⁷ Indeed, current thinking suggests that treating chronic diseases without tackling excess adiposity may promote multimorbidity.⁸ “Multimorbidity, regardless of contributing condition, confers poorer outcome, poorer response to therapy, and poorer quality of life,” he stated. So, taking a holistic approach to treating chronic diseases is the only way to break through that therapeutic glass ceiling.

The final example, he gave was that of the emerging evidence for crosstalk between the peripheral immune response and the brain. With studies showing that immune hyperactivity can modulate neurological processes and may be implicated in the pathophysiology of depression.

Ultimately, while we’re getting closer to figuring out how RA works, we mustn’t forget the bigger picture, Prof. McInnes insisted, and we need to start treating patients as a whole, rather than concentrating on just their joints.

Immunisation update: what’s new and what’s worth it?

Presented by Prof. Kevin Winthrop

Vaccination fatigue after the COVID-19 pandemic is very real, said Prof. Kevin L. Winthrop – for both patients and for doctors. Despite this, vaccines remain important for immune-compromised populations; and he suggested that, in general, the specialist who is responsible for prescribing an immune-modulating drug should oversee the vaccinations relevant to their patient.

RSV vaccinations. Two respiratory syncytial virus (RSV) vaccines, both indicated in adults ≥60 years, have been released in Australia in 2024.^10,11 Data indicates both RSV vaccines have similar safety and efficacy profiles as well as reassuringly low rates of severe reactogenicity (<4%).^12,13 However, both vaccines have induced serious neurologic events such as Guillain-Barre syndrome and inflammatory demyelinating disease at low rates (~1 in 10,000). Although, more than half such events were in patients who were simultaneously vaccinated for influenza and RSV; and Prof. Winthrop shared that, so far, he is avoiding using both vaccines at once in his own practice. In addition, having a booster shot one year later did not make a substantial difference to the 2-year efficacy of the vaccination.¹⁴

Should DMARDs be put on hold for vaccinations? The 2022 ACR guidelines suggest that it’s best to hold disease-modifying anti-rheumatic drugs (DMARDs) at the time of influenza vaccination for one week.¹⁵ For the COVID-19 vaccine, recent studies with MTX, abatacept and tumour necrosis factor inhibitors (TNFis) all indicate that patients have improved responses if at least one dose of the DMARD is withheld.^16,17 So Prof. Winthrop suggested that, in general, holding DMARDs for 1-2 weeks after a vaccination makes sense.

However, for rituximab (RTX), given its half-life of >6 months, ‘holding’ a dose is not practical. So what should be done? Humoral immunity (measured by seropositivity) after mRNA-based COVID-19 vaccination is severely attenuated in the first 3-6 months after an RTX dose, particularly in combination with MTX. But cell-mediated immune responses may still occur in a proportion of patients treated with RTX;¹⁹ so, Prof. Winthrop suggested that, while vaccinations shouldn’t be withheld, timing is crucial.

Updates on other vaccinations. Pneumococcal vaccination is strongly recommended for patients aged <65 years who are on immunosuppressive medication. Prof. Winthrop recommended transitioning to a single-dose strategy with the newer PCV20 vaccine, including for patients who may have been part-way through a two-dose course with older vaccines.^15,20

In terms of herpes zoster, observational studies in IMID patients,^21-23 show that the recombinant zoster vaccine (RZV) is well-tolerated and disease incidence is effectively reduced; but the efficacy rates do not reach the same impressive levels seen in healthy patients. DMARD use, in particular abatacept, but also the use of Janus kinase inhibitors (JAKis) and MTX in combination may markedly reduce immune responses.²³ There are some encouraging results for RINVOQ®, with one study (n=95) showing that patients treated with UPA 15 mg QD plus MTX achieved satisfactory humoral (71.4%) and cell-mediated (64.7%) immune responses to RZV at Week 60.²⁴ But these are still far lower than expected in a heathy population, and further data on efficacy and safety of RZV in IMID populations is a currently a high unmet need.

Is there such a thing as rational therapeutic selection?

Presented by Prof. Iain McInnes | Dr Tim Raine

The essential question, said Prof. McInnes, is “Do we know which of the many options we have available, is the right one for the human being sitting in front of us?” At this point in time, he suggested, the answer is ‘No’. But how close are we to this holy grail of rational therapeutic selection?

Of course a huge number of rational factors are already taken into account by both Gastroenterologists and Rheumatologists when making treatment decisions. The ‘dark art’ of choosing the right treatment, for the right patient, at the right time, is something both Prof. McInnes and Prof. Raine agreed specialists are already very good at. It is the aspect of taking an individual patient’s biology into account that leaves something to be desired.

The current use of personalised biomarkers for diagnosis and ongoing assessment in inflammatory arthritis incorporates serology (e.g. rheumatoid factor, anti-citrullinated protein autoantibodies) acute phase reactants (e.g. C-reactive protein, erythrocyte sedimentation factor), imaging, and physical assessment.^25,26 Prof. McInnes highlighted that while imaging can’t be used to guide clinical decisions, it can be used to assess if there is inflammation in the first place.

As for biomarkers in IBD, Dr Tim Raine said there is no easy-to-measure biomarker of drug response available in clinics right now. So treatment decisions rely on known factors such as efficacy, sustained response rates, safety profiles, convenience for the patient, and reimbursement challenges etc. Using genetic information to direct clinical practice, however, remains incredibly difficult. Dr Raine explained that both penetrance (how likely a mutation will have an effect) and effect size must be considered: high penetrance with a large odds ratio is clinically useful, whereas low penetrance with a small odds ratio isn’t, he stated.

However, some useful biomarkers are emerging. Dr Raine gave an example from the PANTS 3-year prospective study (n=1,240), which revealed that the risk of immunogenicity to TNFi monotherapy was doubled if the genetic marker HLA-DQA1*05 was present (hazard ratio 1.9; 95% CI 1.6-2.25).²⁷ Similarly, the NUDT15 mutation has been associated with an elevated risk of thiopurine-induced myelosuppression and it is estimated that 95 patients must be genotyped to prevent 1 case (95% CI, 62-143 patients). These types of studies, he suggested, are starting to impact gastroenterological clinical management, but it’s a slow process.

Prof. McInnes discussed the idea of developing new biomarkers based on current unmet needs and explained that there are multiple challenges in developing ‘biology-informed’ choices in IMID. RMDs are frequently multifactorial and have heterogeneous pathophysiology, disease course and therapeutic responses, yet are classified in therapy-defined populations. Biomarker discovery in diseased tissues may require invasive access, but how do we decide which tissue to sample? Tissue(s) that inform systemic immune dysregulation include the genome/epigenome, blood, skin, entheses, urine, and synovial membranes.5,29 However, a specific tissue response does not necessarily represent what is happening in other tissues. Age is a further factor that can influence IMID development through channels such as epigenetic changes, differences in immune maturation and the accumulation of exposures such as smoking and dysbiosis that impact immune dysregulation over time.30 Attempting to manage these disordered disease kinetics with rational therapeutics is difficult, because among all these potential influences there is no way to tell where or why an individual’s disease started. Finally, the large data sets generated with ‘omic studies require the use of complex technologies, and potential biomarkers require rigorous validation that can take several years to complete.

Dr Raine went on to review some attempts to make omics-informed biomarkers work in IBD. He showed that genetics, so far, are poor at predicting IBD phenotype or prognosis.^31,32 A transcriptomic predictor of IBD outcome (based on a CD8 T-cell transcriptional signature) showed early promise.³³ But when it came to a clinical trial, provided no predictive potential, with patients doing equally well with an aggressive treatment strategy regardless of their biomarker status.³⁴ There may still be hope, Dr Raine says, with drug-specific, stable metagenomic predictors of response to common biologics; and the current OMICROHN study is working to determine if this can be validated.³⁵

A better approach, Dr Raine suggested, might be to design therapies from the ground up, by proposing rational drug targets based on molecular pathology. For example, a 2021 study identified a subpopulation of ulcerative colitis and CD patients who have a specific mRNA gene expression profile related to IL1.³⁶ These patients were non-responders to current biologic therapies (TNFi, vedolizumab); and signature-positive patients were characterised by neutrophil infiltrates, activated fibroblasts, epithelial cell loss and crypt formation. Thus providing a biological basis for investigating IL1 signalling blockade in this patient population.³⁶

Prof. McInnes said the landscape is similar in rheumatology, and emphasised that pathogenesis-led drug discovery should be the norm in 2025. He believes that in the future, we should look at the state of the patient at a molecular level, then consider what the chosen perturbation (drug) is, and from there determine what biosignature would be associated with that perturbation. “Coincident development of target selection, drug agent and biosignature will mean we’re no longer guessing,” he suggested. He ended by saying that the only reason it’s complicated right now is because we don’t understand it – yet.

The impact of AI on future practice

Presented by Anders Sörman-Nilsson

“The rate of change has never been this fast and will never be this slow again,” said Anders Sörman-Nilsson in his presentation introduction. His overall prediction being that artificial intelligence (AI) will augment every clinician’s practice in the foreseeable future.

There is no doubt that AI-based technology is transforming the healthcare industry.³⁷ The next generation of our healthcare workforce is predicted to include such advances as virtual medical assistants that can triage patients and identify appropriate referrals; and AI-assisted pharmacists that can analyse patient and medication-level data to supplement medical decision making.³⁸

AI-driven virtual therapists are already making large inroads into the field of mental health, and a recent study has shown that young people often preferred AI-generated responses over responses from human therapists. But only on topics such as relationships, self-expression and physical health; notably, this preference did not hold true for more sensitive topics such as suicide.^39,40

AI-driven drug development is also starting to bear fruit. Applying modern, machine-learning techniques to problems such as protein folding is helping researchers more effectively determine the structure of biological targets and design small molecules that have the potential to interact with them without causing undue toxicity. Indeed, several AI-designed drugs are already in, or are entering, clinical trials, including a PDE-10 inhibitor that is designed to treat ulcerative colitis.⁴¹

Nevertheless, the signs so far are that AI can’t ‘go it alone’. At least not yet. In the field of radiology, in which machine-learning-enhanced image analysis has become increasingly well used in recent years,⁴² studies suggest that humans and AI working collaboratively can outperform either AI or humans individually.⁴³ The same synergy is true for the field of drug discovery, with AI considered by experts to complement, but certainly not replace, human scientists.⁴¹

So in a world where AI has the ability to plan your appointments, take clinical notes and write reports for you; the only question is, will you embrace AI and its benefits, or continue to work harder without it.