0563: High-Dimensional Immunophenotyping with Mass Cytometry Reveals Unique Immune Cell Aberrations in Patients with Undiagnosed Inflammatory and Autoimmune Diseases

Alisa Mueller¹, Takanori Sasaki², Joshua Keegan¹, Jennifer Nguyen¹, Alec Griffith¹, Elizabeth Feig¹, Alice Horisberger³, Ye Cao¹, Gregory Keras¹, Lauren Briere⁴, Laurel Cobban¹, Daimon Simmons¹, Juan Pallais¹, Jeffrey Sparks², V. Michael Holers⁵, Undiagnosed Diseases Network¹, David Sweetser⁴, Joel Krier¹, Joseph Loscalzo¹, James Lederer¹ and Deepak Rao¹, ¹Brigham and Women's Hospital, Boston, MA, ²Brigham and Women's Hospital and Harvard Medical School, Boston, MA, ³Harvard Medical School and Lausanne University Hospital, Boston, MA, ⁴Massachusetts General Hospital, Boston, MA, ⁵University of Colorado, Denver, CO

Background/Purpose: Few tools are available to evaluate the immune dysregulation in patients with severe autoimmune or inflammatory conditions that do not conform to well-defined rheumatologic diseases. The ability to identify features of the aberrant immune response in patients with undiagnosed inflammatory conditions may reveal underlying etiology and guide therapy.

Methods: We used mass cytometry to study the immune dysregulation in patients with severe immunologic conditions evaluated in the Undiagnosed Diseases Network (UDN), a multi-center program to evaluate patients with rare conditions that defy diagnosis. We analyzed PBMCs from 16 UDN patients with inflammatory conditions plus 87 non-inflammatory controls, 24 lupus patients, and 30 rheumatoid arthritis (RA) patients as comparators using two 39-marker panels.

Results: Our clustering analyses of mass cytometry panels containing T or B cell-related markers revealed major cell populations and subsets (Fig. 1A). We identified patients who appeared as unique individual outliers in either of 2 types of measurements: frequency of a cluster either (1) as a proportion of total PBMCs or (2) as a proportion of T cells, B cells, or combined monocytes and natural killer cells. We defined an extreme upper outlier as a case in which the single highest patient value is >2-fold higher than the next highest patient. This simple approach identified outlier features in 5/16 UDN patients (31.3%) and in 0/121 (0.0%) of comparator patients, indicating significant enrichment of outlier features in UDN patients (p< 0.001 Fisher's exact test) (Fig. 1B). UDN #1 is a patient with an 8-year history of debilitating global erythroderma, anhidrosis, and alopecia. We discovered that this patient had a markedly expanded population of CD25^hiCD127^- Tregs (Fig. 2A) comprising 54% of all T cells (average=4.7±4.5%) (Fig. 2B-D). Subsequent TCR clonality studies of blood and skin samples revealed the Treg population to be polyclonal, and T cell-directed therapies were trialed. Longitudinal single-cell RNA sequencing analyses revealed a reduction in Treg frequency with abatacept without clinical improvement, but with subsequent clinical improvement with JAK inhibitors. UDN #2 is a patient with a 4-year history of worsening cranial nerve palsies who failed multiple immunosuppressive therapies. We identified a population of CD5⁺Bcl-2⁺ B cells (Fig. 2E) comprising 22% of PBMCs (Fig. 2F-H). Strikingly, 87% of all cells in this unique cluster originated from UDN #2 (Fig. 2I). Clinical evaluation subsequently identified chronic lymphocytic leukemia and this patient was directed to oncology. Other outlier patients include: CNS disease with markedly increased Vδ2 T cells (UDN #3); progressive polyneuropathy with CD138⁺CD56⁺ expanded monocytes; and lymphadenopathy and erythema nodosum with an expanded monocyte subset.

Conclusion: High-dimensional cytometry can identify marked abnormalities in patients with inflammatory diseases that do not meet clinical criteria for rheumatic or other described disorders. Integrating cytometric analyses with clinical findings and genetic variants offers a promising approach to defining pathogenic immune mechanisms in undiagnosed patients.
Figure 1: Outlier cellular aberrations are identified in UDN patients. (A) UMAP of samples analyzed by mass cytometry panels consisting of T and B cell-related biomarkers reveal major cell populations (T marker panel shown). (B) For all UDN and comparator samples, the frequency of each cluster as a proportion of total PBMCs or as a proportion of the indicated cell populations was analyzed. In cases where an outlier cluster frequency was detected in a sample, all the samples from that cluster are shown in red and the outlier patient designation is labeled. Here, T cell clusters and B cell clusters from the T cell marker panel are displayed.

Figure 2: Outlier analyses reveal abnormalities in defined cell types as well as previously uncharacterized populations. (A) T cell marker panel analysis of UDN patient #1 reveals a marked expansion of a cluster 12, consisting of activated CD25hiCD127- Tregs with high ICOS and Ki67 expression. (B-D) This Treg population is markedly expanded comprising 54.4% of all T cells as shown in the bar chart (B), biaxial gating (C), and dot plot comparison with established rheumatologic conditions of RA and SLE as well as non-inflammatory controls (D). (E) Analysis of the B cell marker panel for UDN patient #2 reveals the expansion of a unique CD5+Bcl-2+ B cell population, cluster 23. (F-H) This cluster comprises 22.2% of all PBMCs as shown in the bar chart (F), biaxial gating (G), and dot plot depicting comparator groups (H). (I) Cluster 23 is not a typical B cell subset, and the vast majority of cells in this cluster (87.22%) are derived from UDN patient #2, with many samples containing minimal or no cells contributing to this cluster.
Disclosures: A. Mueller, None; T. Sasaki, None; J. Keegan, Alloplex Biotherapeutics; J. Nguyen, None; A. Griffith, None; E. Feig, None; A. Horisberger, None; Y. Cao, None; G. Keras, Takeda Pharmaceuticals; L. Briere, None; L. Cobban, None; D. Simmons, None; J. Pallais, None; J. Sparks, Bristol Myers Squibb, AbbVie/Abbott, Amgen, Boehringer Ingelheim, Gilead, Inova Diagnostics, Janssen, Optum, Pfizer; V. Holers, Janssen; U. Network, None; D. Sweetser, Merck Sharp and Dohme Corp.; J. Krier, None; J. Loscalzo, None; J. Lederer, VeloceBio, LLC, Alloplex Biotherapeutics, Inc; D. Rao, Janssen, Merck, Bristol-Myers Squibb, Scipher Medicine, HiFiBio, Inc., AstraZeneca, Pfizer.