As a group of progressive, highly heterogenous diseases, Immune-mediated Inflammatory diseases (IMIDs), are immensely challenging to treat. This is because when diagnosed through phenotypic assessment such as clinical parameters, physiological properties, and imaging data, patients who appear to share common features (e.g., symptoms), respond very differently to the same treatment or not at all. Solely treating patients according to phenotype has proven unsuccessful to date even with the most effective therapeutics, biologics (1). In combination with the variability in disease onset, course, and the failure to respond to treatment, it is clear there is much more to uncover as patients suffering from the same disease may not share identical underlying mechanisms (pathophysiology).
Determining the most suitable treatment for IMID patients requires an in-depth understanding of their disease mechanism. For inflammatory diseases, this is required taking into consideration the individuality of the immune system, and similarities between specific patient groups. The understanding and grouping of patients according to disease mechanism based on genetic predisposition, treatment response, and altered molecular pathways is a concept defined as “endotyping” (2). Widely recognised IMIDs such as Rheumatoid Arthritis (RA), Psoriasis, and Systemic Lupus Erythematosus (SLE) are well researched, however, more is yet to be understood about diseases recently classified as IMIDs including Asthma, Type 1 Diabetes Mellitus, and the neurodegenerative disorder characterized by dysregulated neuroinflammatory pathways, Alzheimer’s disease (3,4,5). This article will discuss how endotyping is transforming the treatment of such progressive inflammatory diseases towards a precision medicine approach.
Endotypes in Rheumatoid Arthritis
For a long time, a trial-and-error approach has been used to treat patients suffering from RA, an autoimmune and chronic inflammatory disease characterized by inflammation of the synovial joints (6). Today, clinical assessment, disease activity and radiographic scores are used to direct which treatment a patient will receive. Yet, it is still unclear why there is a reduced or lack of response over time to the most successful types of treatment available; biologics targeting TNF (etanercept, adalimumab, infliximab, and certolizumab), targeting B cell depletion (rituximab), or the IL-6 receptor (tocilizumab), or targeting T cell stimulation (abatacept) (7). This indicates the urgent need to further understand the underlying mechanisms of this disease. Given the significantly high cost of biologics and the progressive nature of RA, its distinctive mechanisms must be defined to provide tailored treatment to patients before joint damage results in complete disability.
In the effort to understand underlying pathological mechanisms of RA, five distinct putative endotypes have been identified through clustering disease profiles of RA patients (8). To identify these disease profiles, serum biomarkers related to cartilage degradation, bone resorption, CRP metabolites, interstitial matrix degradation, and macrophage activity were measured and analyzed. The clusters identified were then investigated for their differences in clinically relevant variables such as disease duration, disease activity, and disability. Further analysis of these endotypes is required using advanced digital tools such as machine learning to determine the ability to predict treatment response and future complications. This would be valuable in guiding improved and efficient treatment of RA patients, reducing further disease progression, while enabling the development of novel therapies.
Endotypes in Asthma
Over the years, systems biology and learning across studies has also enabled the understanding of molecular mechanisms of asthma revealing two endotypes: Type 2 and non-Type 2 (9). The type 2 (T2 high) endotype encompasses the secretion of Th2 cytokines (IL-4, IL-5, and IL-13) and eosinophilic inflammation. These individuals also demonstrate a marked loss of tight junction proteins reducing epithelial barrier integrity and inducing epithelial barrier remodeling due to the secretion of Alarmins; (Thymic stromal Lymphopoietin (TSLP), IL-25 and IL-33) which activate lineage native cells (ILC2) to propagate the release of IL-5, the eosinophil-specific regulatory inflammatory cytokine and IL-13, which has been shown to have synergistic effects with TNF-α (10).
Contrarily, the non-type 2 (T2 low) endotype (observed in 50% of asthma patients), has been associated with the activation of Th1 and/or Th17 cells which are associated with neutrophilic, steroid-resistant asthma. Unlike the first endotype (T2 high) which is already well understood with several treatments available such as inhaled/oral corticosteroids (11) and biologics targeting IgE, IL-13, IL-4Rα, IL-5 and IL-5Rα (12), there is a need to further characterize the T2 low endotype and identify biomarkers to guide novel therapy development. The role of endoplasmic reticulum (ER) stress, the NLRP3 inflammasome (cytoplasmic oligomeric signaling complexes formed within inflammatory cells) (13), and phosphoinositide-3-kinase-δ (PI3-K) in the steroid resistance of this severe asthma endotype are yet to be determined. These factors are regarded as relevant as they are involved in the activation of proinflammatory pathways through the production of proinflammatory cytokines (ER stress and the NLP3 inflammasome) (14) or the activation of macrophages to further initiate proinflammatory cytokine production (PI3-K in the Akt signaling pathway) (15) thus, could contribute to the disease mechanism of IMIDs.
Due to the discovery of such differing underlying mechanisms, asthma is no longer considered a single disease. The identification of various disease subgroups enabled to stratify patients accordingly and move away from a “one size fits all” treatment approach. As a result, there have been improvements in the management of this inflammatory disease with biological therapies targeting IgE (omalizumab), IL-4Rα (dupilumab) and IL-5Rα (benralizumab) IL-5 (mepolizumab and reslizumab), and investigation of a new class of biologics to treat the corticosteroid-resistant asthma endotype (16). This endotype-driven approach is being applied to manage a broad range of allergic diseases (2). Currently, there are ongoing efforts to precisely characterize and understand the different asthma endotypes which will in turn enable the identification of novel personalized therapies.
Endotypes in Alzheimer’s Disease
The complex neurodegenerative disease Alzheimer’s remains a challenge in the clinic with all patients receiving the same treatment despite the indication that the underlying mechanism of their disease is varied. This disease, characterized by chronic neuroinflammation, requires the differentiation of endotypes to uncover new promising drug targets. Clinical data has demonstrated three subgroups: class 1 (rapid decliners), class 2 (slow decliners), and class 3 (severely impaired- slow decliners). The investigation of underlying biological mechanisms between the different subgroups could confirm whether these are indeed representative of different disease endotypes. This would enable the stratification of patients and identification of those likely to be more responsive to existing treatment, and therapies in development.
Using an integrative multi-omics approach, mechanistic alterations occurring at the transcription factor and chromatin dynamics level in early-onset familial Alzheimer’s disease revealed 6 modulated gene programs involved in pluripotency, dedifferentiation, cell cycle reentry, inflammation, lineage miRNA, and neuronal specification (17). Further investigation of the pathways of the molecules involved could enable the delineation of underlying mechanisms and identification of biomarkers to determine the most suitable treatment for the prevention of further complications such as progressive memory loss and cognitive decline.
Endotypes in Diabetes
Although the metabolic disorder diabetes mellitus has already been classified into two subtypes (type 1 & type 2), there may be further endotyping required for type 1 diabetes as there is still no successful treatment for this autoimmune disease (18). Type 1 diabetes is considerably heterogeneous in onset, insulin secretory capacity, complication rates, and efficacy of treatment leading to caveats in mechanistic and immunological studies. In failed clinical trials where teplizumab, the CD3 antibody was investigated for its ability to preserve β-cell function and hence insulin secretion, some patients demonstrated a good response (particularly younger patients (<12) who had a higher level of c-peptide, the byproduct of insulin, and those from North America and Europe). This also introduces the concept of regiotypes; the regional difference between endotypes due to environmental influence.
In another teplizumab clinical trial, patients grouped by Human Leukocyte Antigen (HLA) immunity and zinc transporter 8 autoantibodies were also responsive to the drug indicating the importance of stratifying patients according to endotype to identify the true benefit of a treatment (18). These concepts should also be considered when differentiating patients by disease mechanism and using biomarkers to guide both clinical trial design and treatment.
Prediction of Treatment Response Using Biomarkers
Successful attempts to provide biomarker-guided treatment are being taken by pharmaceutical companies such as Egis Pharmaceuticals PLC, a member of the Servier group. The Hungarian pharmaceutical company is introducing a precision medicine approach to the treatment of RA by developing a blood-based in-vitro diagnostic test (PREDYSTIC® Infliximab RA Kit) (19) based on genomic biomarkers. This test will differentiate between responders and non-responders of infliximab, the anti-TNF-α inhibitor enabling efficient, targeted treatment of Rheumatoid Arthritis.
IBD, the progressive gastrointestinal inflammatory disorder which comprises Crohn’s disease and Ulcerative colitis, can lead to extensive irreversible damage to the wall of the bowel leading to severe future complications and loss of bowel function. There is thus a critical need to optimize treatment early on to improve the prospects of suffering patients (20). One way to do this is to use validated biological signatures (biomarkers), to infer future complications or response to treatment. Prognostic biomarkers based on gene expression can be used to predict future complications whereas predictive biomarkers can be used to identify early on which patients would respond to a specific treatment (20). A research organization making strides in this area is the Crohn’s and Colitis Foundation who, in collaboration with Life Arc, is advancing the development of a clinical test (21) based on prognostic and predictive biomarkers to improve long-term outcomes and patient’s quality of life.
The Importance of Stratifying Patients During Clinical Trials
Using biomarkers to guide clinical trials can provide immense benefit as the design of a clinical trial can affect whether it ultimately succeeds or fails. As highlighted in the Astra Zeneca framework for improving R&D productivity, the right patient and target must be selected to improve the success of a clinical trial (22). Where patients are not stratified by endotype, the investigated treatment is likely to demonstrate minimal efficacy compared to an endotype-driven trial (7). A more targeted clinical trial prevents the masking of a treatments’ benefit due to a non-specific patient population. The incorporation of newer, innovative clinical trial designs such as master protocols and leveraging data across different clinical development phases has been recommended by the FDA (23).
Master protocol clinical trials are an innovative new type of clinical trial design that comprises 3 different types: basket, umbrella, and platform trials (24). Basket trials involve treating subjects with different diseases but related endotypes with a combination or a single treatment after stratification using a biomarker linked to the drug target and/or disease endotype. Where multiple treatments or treatment combinations are investigated for a single disease in subjects stratified into several groups, this is known as an umbrella trial. In this case, subjects may be stratified according to molecular signatures identifiable using biomarkers.
Platform trials, (an extension of the adaptive trial), however, involve the investigation of multiple treatments or combinations for a single disease (or its subtypes) in a sequential manner that is constantly evaluated and adjusted accordingly (25). As previously applied in breast cancer programs such as I-SPY2, this trial design could be used to investigate multiple treatments in biomarker defined groups of patients with distinct endotypes (26). Such clinical trials provide a flexible, yet highly targeted approach allowing to establish therapeutic efficacy of treatment candidates in specific groups of patients.
What is Required to Identify Endotypes?
Identifying endotypes for the development of biomarkers and their validation during clinical trials is challenging as IMIDs share common overlapping inflammatory pathways. Defining such a connected system composed of different types of molecular components can be immensely complicated. Although some endotypes have been identified and predictions to treatment can be made using advanced statistical methods, there may be unexpected outcomes in biological scenarios. Efforts need to be made to understand this as these endotypes may be dynamic, changing with time (not all interactions may be present in all patients at the same time), and may differ due to other factors including sex, ethnicity, age group, etc.
A collaborative effort between industry, healthcare institutions, researchers, and funders, is needed to leverage advanced technology and methods (including modeling) in an integrative approach to investigate diseases and treatments. Combining molecular information (genetic, epigenetic, RNA, proteomic, metabolomic) with clinical and demographic data as well as other (e.g., environmental) factors influencing disease progression and health outcomes, would enable the identification of drug endotypes; the pathways activated by drugs, and disease endotypes; pathways activated in disease. The key is to examine the complementarity between the two endotypes to identify biomarkers that will indicate the drug most suitable for a particular patient.
It is time to progress from phenotype-driven diagnosis and treatment towards an endotype-driven approach, particularly for IMIDs which include widely researched diseases such as RA and under-researched inflammatory diseases such as Asthma. This will enable a better understanding of the mechanisms underlying each disease and the identification of specific patient subsets using validated biomarkersto guide treatment delivery. The resulting outcomes of such precision medicine approaches could include cost savings in biopharmaceutical development, more targeted and thereby effective treatments, and most importantly, improved outcomes for patients suffering from IMIDs.