Late Dupilumab Responders Have Unique Molecular Profile

AMSTERDAM — A proteomics study designed to determine why some patients with atopic dermatitis (AD) respond quickly to dupilumab, others respond more slowly, and the remainder do not respond at all demonstrated that molecular responses in these three groups are very different.

A discovery that could lead to personalizing therapies, the data identified “distinct systemic biomarker profiles,” according to Ester Del Duca, MD, an instructor in the Laboratory of Inflammatory Skin Diseases at the Icahn School of Medicine at Mount Sinai, New York City.

The study was conducted with 67 patients with AD and 16 healthy controls. Serum was collected at two timepoints: An average of 20 weeks after starting dupilumab, then at a mean interval of about 9 months later. At these timepoints, called follow-up 1 and 2, a panel of more than 600 proteins, including unique markers for immunologic, cardiovascular, and neurologic activity, were evaluated.

The criterion for differentiating the three response groups was an Investigator Global Assessment (IGA) score of 0 or 1, signifying clear or almost clear skin (or at least a 2-point IGA reduction from baseline). Early responders were those who met the criterion at both follow-ups, late responders were those who met this criterion only at the second follow-up, and nonresponders never met the criterion.

“There were no significant differences at baseline in clinical severity, past medical history, or patient characteristics,” said Del Duca, who presented these data in a late breaking news session at the European Academy of Dermatology and Venereology (EADV) 2024 Congress.

Of the 67 patients with AD, 39 were early responders, 11 were late responders, and 17 were nonresponders.

The differences in proteomics were marked. For early responders, there was an early normalization of the proteome, reported Del Duca, illustrating the differences in the proteome of the three groups with a color-coded chart of protein upregulation and downregulation relative to healthy controls. The normalization of the proteome persisted in early responders when assessed at the second follow-up.

In the late responders, the proteome dysregulation was substantial relative to healthy controls at the first follow-up, but there was considerable improvement by the second follow-up. Although the change at the second follow-up was still not as robust as that seen in the early responders at either follow-up, Del Duca described the proteomic profile as a 45% improvement from the first follow-up.

In contrast, nonresponders showed worsening in their blood proteome from follow-up 1 to 2. Nonresponders at first follow-up showed upregulation relative to healthy controls for many proteins associated with the Th1 response, such as interferon gamma, CXCL9, and CXCL10, and Th2 response, such as interleukin-4 and Th17/22, and these did not normalize or worsen by the second follow-up.

“Uniquely to nonresponders, key Th1 biomarkers remained significantly upregulated relative to controls at both follow-up 1 and 2,” with a P value

To achieve normalization of the proteome as defined by healthy controls, both upregulation and downregulation of protein activity were required, although more upregulations than downregulations were observed.

When evaluating the proteome changes most implicated in immunoregulation, the investigators were able to show a correlation between worsening in the proteome and greater severity of AD as defined by IGA, Eczema Area and Severity Index, and body surface area involvement.

“Spearman analysis revealed strong and positive correlations between improvements in biomarkers at follow-up 1 and 2 with improvements in clinical markers,” Del Duca said. As examples, she noted favorable changes in biomarkers specifically associated with T cells, dendritic cells, and natural killer cells as clinical outcomes improved.

Conversely, the worsening in T-cell activation among nonresponders, particularly Th1 biomarkers, also tracked with increasing AD symptoms over time.

The implications of the research are broad, and most importantly, it shows that therapeutic targets are likely to differ between patients with AD, according to Del Duca. Although proteomic studies have not yet been conducted with other treatments, these might provide further insight about how patients with AD differ in response across other drugs.

This is important work, according to Brigitte Dréno, MD, PhD, head of the Department of Dermatology, Nantes University Hospital, Nantes, France. As moderator of the late breaking news session, she suggested that there are many potential messages from these data, not least that treatment of AD likely involves targeting cytokines beyond those affected by dupilumab in at least some patients.

When Dréno asked Del Duca about what could be surmised about changes from baseline before treatment to the first follow-up, Del Duca said that the study was retrospective, so baseline data were not available.

This is an important missing piece of this investigation, according to Dréno.

“As you move this work forward,” she said that it would be “very important” to determine “if there are predictive markers for evaluating which patients will respond.”

This is a small study with many additional variables to consider in order to develop a clinically useful tool, Del Duca noted. However, this work not only has the potential to guide treatment selection but the biomarkers upregulated in nonresponders are already “suggesting potential targets for refining therapeutic strategies,” she said.

The study received funding from Bristol Myers Squibb. Del Duca reported no financial relationships with industry. Dréno reported financial relationships with La Roche-Posay, Pierre Fabré, and Galderma.

Ted Bosworth is a medical journalist based in New York City.