Patients with rheumatoid arthritis (RA) experience joint pain and inflammation that may hinder their social, occupational, and recreational activities. If left untreated, this chronic autoimmune inflammatory disease may lead to long-term disability and premature death. 

Therapeutic strategies for RA have considerably improved in the past decades. “However, these new treatments are only good for so many patients. Usually, when we do clinical trials, about 60 or 70 percent of the patients will respond to treatment. The other ones will not,” said Christian Lood, an immunologist at the University of Washington. 

Diana Abdueva, the founder and chief executive officer of the precision medicine company Aqtual, emphasized the difficulties faced by patients who often undergo multiple therapies before finding an effective one. This is because for RA and many other chronic diseases, clinicians do not have accurate diagnostic tests for predicting therapeutic response for each patient, she said. 

According to Lood, who is not associated with Aqtual, for any rheumatic or inflammatory disease, early treatments can get patients into remission faster and improve their prognosis, preventing downstream effects such as organ damage. “So early and precise treatment is helpful,” he said.

Abdueva believes that precision medicine requires a sensitive platform that captures specific signals from patient samples using noninvasive procedures. Her team at Aqtual is developing such a platform wherein they will identify biomarkers from a single blood draw and leverage the information in DNA circulating in the bloodstream. The innovative approach holds promise for tailoring RA treatments based on individual patient profiles. 

Special DNA

Abdueva has predominantly focused her career on designing diagnostic tools using DNA technology. “For the past 12 years, I’ve been working on very special DNA,” she said. This special DNA is circulating cell-free DNA (cfDNA), which is extracellular DNA that originates primarily from the death of cells. “We have a normal cell turnover, so our organs, whether they are healthy or diseased, shed all of the DNA,” she explained. But when there is chronic inflammation, higher rates of cell death occur at a particular site, shedding DNA remnants into the blood.

The levels of cfDNA fragments in the blood, their quality, and their methylation and mutation frequencies reveal specific details about a patient’s condition, which researchers can potentially use for diagnostic and prognostic tools in cancer and other pathologies, including RA (1-3). Abdueva, for instance, initially applied cfDNA technology to detect deletions in maternal plasma associated with chromosomal syndromes and to improve tumor characterization in patients with cancer (4,5). Recognizing its potential as a noninvasive biomarker for those other diseases, Abdueva founded Aqtual four years ago with the plan to revolutionize the chronic disease space by analyzing cfDNA from a simple routine blood draw.

RA is a complex heterogeneous disease. Although patients may experience similar pain, there might be very different mechanisms at play in each case. These mechanisms influence therapy responses. By looking at each patient’s cfDNA, Aqtual’s test aims to identify the underlying RA mechanism and use that information to match the patient with the best drug.

“The beauty of the technologies that we’ve developed is that we actually amplify the signal that comes from outside of the blood by finding only DNA that is regulatory active,” Abdueva explained. “That's the advantage that we have with our technology that no one else has today,” she said. 

Aqtual’s team uses machine learning to train its cfDNA platform. First, Abdueva and her colleagues collected samples from hundreds of patients with RA. Then they collected samples from patients who did not have RA but had another similar autoimmune condition. The machine learning approach attempted to find the features that differentiate patients with RA from patients with diseases adjacent to RA. 

In November 2023, Aqtual announced that their platform could identify key gene expression signatures at the synovial joints, the tissue between bones, through a routine blood draw (6). They looked at samples from 89 patients with RA and 102 patients without it using a group that included healthy individuals and patients with osteoarthritis, psoriatic arthritis, and other inflammatory conditions. Based on the genetic signatures in the blood, their machine learning algorithm identified RA signatures with 96.1 percent specificity and 90.8 percent sensitivity.

“[This] is a fantastic milestone for us as a proof-of-concept,” said Abdueva. “The machine learning application, in this particular context of well defined features, translates in a remarkable ability to deliver both sensitivity and specificity for conditions that are very close clinically.” 

“There is a lot of exciting academic work that shows that if only you can look into the joint, and specifically into synovial, which is inflamed in rheumatoid arthritis, you can pick up all of the heterogeneity of rheumatoid arthritis,” said Abdueva. Aqtual’s platform promises to do so, and consequently provide adequate therapy selection for patients with RA. “They are the ones who are in need of this precision medicine solutions the most.” 

Lood agreed that the platform achieved excellent sensitivity and specificity. Yet, he was skeptical about the novelty. “They [demonstrated] that this blood draw can help stratify patients whether they have rheumatoid arthritis, whether they are healthy, or whether they have other inflammatory diseases,” he said. “That’s a good thing to do, though maybe not as helpful because usually the clinician will be able to tell the diagnosis even without that blood draw; we have currently very good diagnostic markers for rheumatoid arthritis.”

While Lood does not think that the strength of Aqtual’s platform will be in diagnosis, he is excited about its potential for identifying responses to treatment in different patients. “That is something we currently don’t know much about in the clinical practice, and that could be a very good thing,” he said. 

Learning from nonresponders

While the identification of RA signatures using Aqtual’s platform is an important milestone for Abdueva and her colleagues, their next focus is to conduct a clinical trial to develop a signature classifier that may reveal the best drug choice for each patient. 

“On average, a patient switches therapy every eight months,” Abdueva explained, so every eight months, clinicians must pick a new drug. “This process in rheumatoid arthritis is called trial and error . . . because there are no biomarkers that can lead you to one or another.” 

It’s an excellent starting point to screen and see if they can find any signal for any of the drugs... I think it’s the right approach and I think we’ll get some very interesting data from that. 
- Christian Lood, University of Washington

Abdueva hopes that cfDNA in the blood of patients with RA may provide genetic, transcriptomic, and epigenetic data that serve as biomarkers for predicting the best therapeutic choices for each case. Their first clinical approach to test this is to follow patients with RA who showed an inadequate response or intolerance to a previous therapy and are currently starting a new treatment. The team will analyze the circulating cfDNA of the patients at the time of the therapy switch and 12 weeks later. 

“It’s an excellent starting point to screen and see if they can find any signal for any of the drugs,” said Lood. “This should guide them towards what the second trial would be to validate whatever they can find, so I think it’s the right approach, and I think we’ll get some very interesting data from that.”

Abdueva emphasized the need for a classified training study with a larger and more diverse cohort of patients to then move to clinical validation. If the team successfully identifies signatures that inform therapeutic choices, Abdueva hopes that this test may be available for clinicians and patients in early 2025. For the first test versions, the doctors will obtain a blood sample from their patients and then send it to Aqtual for processing. “All of the pain points associated with the patient journey start with a clinician making a decision that the therapies that the patient is [taking are] not working,” she said. “We have a perfect opportunity to come with our actual test at the moment when the therapy change is made.” 

Lood agreed on the importance of finding biomarkers such as those sought by the Aqtual team. “It would be fantastic because we are struggling so much with patients and finding the best treatment for them,” he said. “This would not only apply for rheumatoid arthritis, but if they find a good approach, then that could be applicable to any kind of disease.” 

References

1. Jung, K. et al. Cell-free DNA in the blood as a solid tumor biomarker—A critical appraisal of the literature. Clin Chim Acta  411, 1611-24 (2010).
2. Silva de Miranda, F. et al. Properties and Application of Cell-Free DNA as a Clinical Biomarker. Int J Mol Sci  22, 9110 (2021).
3. Hashimoto, T. et al. Cell-Free DNA in Rheumatoid Arthritis. Int J Mol Sci  22, 8941 (2021).
4. Chitty, L. et al. 201: Detection of DiGeorge and Cri du Chat syndrome deletions from maternal plasma by deep sequencing cell free DNA (cfDNA). Am J Obstet Gynecol  210, S110 (2014).
5. Abdueva, D. et al. Abstract 3350: Cell-free DNA fragmentation patterns analyzed in over 15000 cancer patients reveal changes associated with tumor somatic mutations and result in improved sensitivity and specificity of somatic variant detection. Cancer Res  77, 3350 (2017).
6. Taylor, P. et al. A Novel Blood-Based Assay Differentiates Seropositive and Seronegative Rheumatoid Arthritis from Healthy Individuals and Those with Other Inflammatory Diseases or Osteoarthritis [abstract]. Arthritis Rheumatol  75 (2023).