Q: What are the recent developments in the pursuit of a more personalized, pathophysiologically targeted approach to RA treatment?

A lot of time and money have been spent on using biomarkers or genetics to guide targeted therapy, but, so far, it has been a dark hole. However, we know what does not work. Genetic genotyping is not going to be the salvation in terms of predicting treatment response to a specific drug in RA. While there are some treatment-response associations with genetic markers, I do not think that genetics will be able to identify a specific targeted therapy for a given patient with RA. Some companies are trying to use gene expression profiling to identify patients who will respond to treatment. As a first step, however, it may be more important to identify those who will not  respond to a therapy. For example, since anti–tumor necrosis factor (anti-TNF) agents are usually the default biologics that are used first, it would be especially helpful to know whether a patient is unlikely to respond to these drugs. The next step would be to identify those who will respond to an agent with a different mechanism of action (MOA). You want to make sure that you are not simply picking out refractory patients who will not respond to anything. You need empiric, if not causal, evidence that they are going to do better on something with a different MOA. Some of the early work that I have seen suggests that you will likely need 90% or greater accuracy for a test to be acceptable to rheumatologists. We have seen many potential single predictors of response, but we need validation in a separate cohort. We must show not only that a patient will not respond to a particular agent, but also that they will respond better to a different type of agent.

We have been looking at serum biomarkers ever since we began using anti-TNF agents, but nothing has really panned out. We still have not identified the population of patients that is not going to have the type of response we are looking for. Using synovial biopsies to identify the pathobiology that is occurring locally is a promising approach. This is analogous to using tumor biopsies to identify characteristics of malignancies that drive precision treatment in oncology. Perhaps we can learn more from the synovial tissue than we can from the blood, and perhaps this will tell us more about the response to treatment, since the synovial cells are so central in that respect. Such analyses would go beyond simply examining the histology that identifies the inflammatory aspect of the disease. The next step would be to further analyze the macrophages, T cells, fibroblasts, and subtypes of cells present in the synovium to see distinct cellular profiles that can give us clues to guide our treatment approaches. This may also allow us to predict who will respond to a particular treatment paradigm. Although this technology is in the early stages, it holds a lot of promise. Centers in the United States are performing ultrasound-guided synovial biopsies more and more. This is a relatively noninvasive procedure that ensures that you get true synovial tissue. We then have laboratory techniques of single-cell sequencing and metabolomics that can help us to learn more from that tissue. I think that this is the hope for truly targeted therapy at some point in the future.

This is an interesting area, but how close we are to achieving the goal of pathophysiologically targeted therapy in RA is a matter of debate. I am increasingly interested in using molecular markers to assess disease control so that I can identify the patients who have had an excellent response to therapy. I would investigate further by using quality-of-life measures to determine where social behavioral interventions are needed. The role of pivotal cytokines such as interleukin 6 (IL-6) in domains such as fatigue, pain, mood, and sleep is provocative, and whether this can be used to predict outcomes is still a work in progress. There are some preliminary data suggesting that targeting IL-6 could be superior to anti-TNF therapy in some of these domains, but we need larger collaborative studies to confirm this. There are also preliminary data indicating that baseline IL-6 levels may predict response to anti–IL-6 agents, but that also needs validation. Although there are advances in the use of biomarkers, we need high throughput, noninvasive tools to make this work in clinical practice.