Combining Flow and Mass Cytometry in the Search for Biomarkers in Chronic Graft-versus-Host Disease

Abstract

Chronic graft-versus-host disease (cGVHD) is a debilitating complication arising in around half of all patients treated with an allogeneic hematopoietic stem cell transplantation. Even though treatment of severe cGVHD has improved during recent years, it remains one of the main causes of morbidity and mortality in affected patients. Biomarkers in blood that could aid in the diagnosis and classification of cGVHD severity are needed for the development of novel treatment strategies that can alleviate symptoms and reduce the need for painful and sometimes complicated tissue biopsies. Methods that comprehensively profile complex biological systems such as the immune system can reveal unanticipated markers when used with the appropriate methods of data analysis. Here, we used mass cytometry, flow cytometry, enzyme-linked immunosorbent assay, and multiplex assays to systematically profile immune cell populations in 68 patients with varying grades of cGVHD. We identified multiple subpopulations across T, B, and NK-cell lineages that distinguished patients with cGVHD from those without cGVHD and which were associated in varying ways with severity of cGVHD. Specifically, initial flow cytometry demonstrated that patients with more severe cGVHD had lower mucosal-associated T cell frequencies, with a concomitant higher level of CD38 expression on T cells. Mass cytometry could identify unique subpopulations specific for cGVHD severity albeit with some seemingly conflicting results. For instance, patients with severe cGVHD had an increased frequency of activated B cells compared to patients with moderate cGVHD while activated B cells were found at a reduced frequency in patients with mild cGVHD compared to patients without cGVHD. Moreover, results indicate it may be possible to validate mass cytometry results with clinically viable, smaller flow cytometry panels. Finally, no differences in levels of blood soluble markers could be identified, with the exception for the semi-soluble combined marker B-cell activating factor/B cell ratio, which was increased in patients with mild cGVHD compared to patients without cGVHD. These findings suggest that interdependencies between such perturbed subpopulations of cells play a role in cGVHD pathogenesis and can serve as future diagnostic and therapeutic targets.

Keywords: flow cytometry; graft-versus-host disease; hematopoietic stem cell transplantation; immunophenotyping; mass cytometry.

Figures

Figure 1

Serum protein phenotype. (A) B-cell activating factor (BAFF) levels in peripheral blood and (B) BAFF/B-cell ratios for the four chronic graft-versus-host disease (cGVHD) patient groups. Statistical analysis was done with the Mann–Whitney U test. n = 11, no cGVHD; n = 6, mild cGVHD; n = 5, moderate cGVHD; and n = 8, severe cGVHD.

Figure 2

Conventional flow cytometry results. (A) The percentage of mucosal-associated T (MAIT)-cells in peripheral blood, defined as CD161+ TCRVα7.2+ T-cells, in CD4−, CD4− CD8+ and CD4− CD8− gates. Representative flow cytometry figures of a patient with severe chronic graft-versus-host disease (cGVHD) are shown below the graphs for each gating strategy. Statistical analysis was done with the Mann–Whitney U test (MW). n = 9, no cGVHD; n = 5, mild cGVHD; n = 8, moderate cGVHD; n = 9, severe cGVHD. (B) CD38 expression in CD3+, CD3+ CD8+, and CD3+ CD4+ T-cells. Representative flow cytometry figures of a patient without cGVHD and with mild cGVHD are shown to the right of the graphs gated on T-cells. Statistical analysis was done with the MW. n = 11, no cGVHD and n = 7, mild cGVHD.

Figure 3

Mass cytometry analysis in patients without chronic graft-versus-host disease (cGVHD) versus patients with mild cGVHD. Results after automated cell clustering software Citrus and ACCENSE (n = 11, no cGVHD and n = 9, mild cGVHD). The statistical analysis was performed by the Citrus software. The boxplots indicate the spread of the abundancy of the separate clusters, and the histograms depict the expression of specific cellular markers (blue depicts background expression, and red indicates expression of the cluster). (A) Multidimensional depiction of some main cellular subsets. (B) Two B-cell subsets. Cluster 399963 expressed CD19, HLA-DR, CD39, CXCR5, CCR4, and CXCR3. Cluster 399970 expressed CD19, HLA-DR, CD39, CXCR5, and Ki-67. (C) Two NKT-cell subsets. Cluster 399962 expressed CD3, Granzyme B (GzB), CD57, CD44, Ki-67, and CD8. Cluster 399954 expressed CD3, GzB, CD57, CCR4, PD-1, and to a lesser degree CD8. (D) An NK-cell subset, cluster 399979, expressed CD57, GzB, CD39, CD11c, and CD161. (E) A CD4+ T-cell subset, cluster 399983, expressed CD3, TCRαβ, CD4, CD5, CCR4, CD127, CD27, and CD28.

Figure 4

Mass cytometry analysis in patients with moderate chronic graft-versus-host disease (cGVHD) versus patients with severe cGVHD. Results after automated cell clustering software Citrus and ACCENSE (n = 10, moderate cGVHD and n = 10, severe cGVHD). The statistical analysis was performed by the Citrus software. The boxplots indicate the spread of the abundancy of the separate clusters, and the histograms depict the expression of specific cellular markers (blue depicts background expression, and red indicates expression of the cluster). (A) Multidimensional depiction of some main cellular subsets. (B) A B-cell subset, cluster 399948, expressed CD19, HLA-DR, CD39, CXCR5, CCR4, CXCR3, and to a slight degree CD38. (C) An NKT-cell subset, cluster 399981, expressed CD3, CD57, Granzyme B (GzB), and dimly expressed CD44, PD-1, CD8, and CCR4.

Figure 5

Confirmatory flow cytometry results. Analyzed flow cytometry results for a selection of markers for 37 patients [n = 15, no chronic graft-versus-host disease (cGVHD); n = 7, mild cGVHD; n = 10, moderate cGVHD; and n = 5, severe cGVHD]. GzB, Granzyme B. Statistical analysis was done with the Mann–Whitney U test. (A) A potential NKT-cell subset, based on cluster 399954 (Figure 3C). Representative flow cytometry figures of a patient without cGVHD and with mild cGVHD are shown to the right of the graphs gated on T-cells. (B) CD38 expression on CD39+ CXCR5+ HLA-DR+ B-cells. Representative flow cytometry figures of a patient without cGVHD and with mild cGVHD are shown below the graphs. (C) CD39 expression on CXCR5+ HLA-DR+ B-cells. Representative flow cytometry figures of a patient with moderate and severe cGVHD are shown below the graphs. (D) CD8 expression on CD57+ GzB+ T-cells. Representative flow cytometry figures of a patient with moderate and severe cGVHD are shown below the graphs.