Vacuoles in neutrophil precursors in VEXAS syndrome: diagnostic performances and threshold

Abstract

VEXAS (vacuoles, E1 enzyme, X-linked, autoinflammatory, somatic) syndrome is a recently described adult-onset inflammatory disease,1 responsible for inflammatory and haematological disorders.2 The presence of bone marrow vacuoles appeared to be greatly significant in the diagnosis of VEXAS syndrome.1 However, myeloid and erythroid vacuolization has been described in other disorders: copper deficiency,3 acute myeloid leukaemia,4 Menkès disease5 and transcobalamin II deficiency.6 Nevertheless, these disorders are not associated with chronic inflammatory signs. Vacuolated precursors have also been described in myelodysplastic syndromes (MDS),7 which may be associated with inflammatory features even in the absence of VEXAS syndrome.8 We aimed to determine whether the intensity of vacuolization in myeloid precursors was associated with VEXAS syndrome, in comparison with patients with similar features and with patients with MDS, and to determine the best threshold for differentiating these conditions. This study was approved by the ethics committee of Angers University Hospital (2021-027). We fully revised the medical files of patients attending the Angers University Hospital between January 2007 and December 2020 encoded with the codes corresponding to the main differential diagnoses of VEXAS syndrome.1 To be included, patients’ features needed to strongly suggest VEXAS syndrome by fulfilling the following criteria: (i) unexplained chronic fever or biological inflammatory signs, (ii) ≥1 inflammatory involvement described in VEXAS syndrome,1 and (iii) ≥1 chronic cytopenia. We excluded women as the VEXAS is an X-linked genetic disorder, and patients without bone marrow aspiration. The patients were genetically screened for VEXAS syndrome and separated into VEXAS or UBA1-WT (wild type) groups according to the presence or absence of a UBA1 mutation, respectively. Patients with no available blood or marrow samples for genetic assay were excluded. The MDS group included patients paired by age (±5 years) with the patients of the VEXAS group, with a 1:3 ratio. Only patients with MDS-EB (excess blasts) or MDS-RS (ring sideroblasts) of the 2016 revision to the WHO classification were included, as these forms of MDS have been previously associated with vacuolated precursor cells.7 The patients in the MDS group were randomly selected among men who never received MDS treatment (notably no azacitidine and no erythropoietin) before the bone marrow assessment. The MDS prognosis was evaluated according to the MDS International Prognostic Scoring System (IPSS). The previously described somatic mutations in UBA1 (NM_003334.3) were screened by Sanger sequencing to confirm the diagnosis.1, 9 The polymerase chain reaction (PCR) primers used were forward 5′-AAAGCCGGGTTCTAACTGCT and reverse 5′-CAGGACACTGGATGTCTGGA. In cases of multiple bone marrow aspirations, the first specimen was considered for analysis. Each bone marrow aspiration was blindly evaluated by two independent, specialized haematologists. Each haematologist counted the number of vacuoles (rounded optically empty gap of at least 1 µm) in each of 400 neutrophil myeloid precursor cells (myeloblasts, promyelocytes and myelocytes). Each analysed cell was classified as follows: ≤1 vacuole or >1 vacuole. Quantitative data are presented in medians and quartiles and qualitative data are presented as absolute values and percentages. The type I error was set at 5%. The percentages of vacuolated cells were compared using the Mann–Whitney test, the Wilcoxon matched-pairs signed-rank test and the Kruskal–Wallis test as appropriate. The reliability between the two raters for the proportion of cells containing >1 vacuole was evaluated by means of intraclass correlation (ICC) after log transformation. We assessed the diagnostic performances of the proportions of neutrophil precursors having >1 vacuole for the diagnosis of VEXAS syndrome with the area under the curve (AUC) of receiver operating characteristic (ROC) curves. We identified 11 patients who met the criteria for screening VEXAS syndrome, and 9/11 underwent both bone marrow analysis and had available blood or marrow samples for genetic assay. 6/9 patients presented an UBA1 mutation, constituting the VEXAS group. The three others constituted the UBA1-WT group. The MDS group included 18 patients with MDS: 5 MDS-EB-1, 10 MDS-EB-2, 1 MDS-RS-SLD (single lineage dysplasia) and 2 MDS-RS-MLD (multilineage dysplasia). In the MDS group, median IPSS was 2 [1–2] with one low-risk, five intermediate-1 risk, eight intermediate-2 risk and four high-risk MDS. Characteristics of patients are detailed in Table I and Tables SI and SII. Figure S1 presents examples of vacuoles observed in the patients. Evaluation of the proportion of cells containing >1 vacuole showed a very good reliability between the raters (ICC = 0·893). The proportion of neutrophil precursors with >1 vacuole differed among groups: 18·0% [13·4–20.3] in the VEXAS group, 4·5% [4·5–9.0] in the UBA1-WT group and 1·6% [0·9–5.4] in the MDS group (P = 0·0005, Fig 1). The AUC of the ROC curves assessing the diagnostic performances of the proportion of neutrophil precursors with >1 vacuole for the diagnosis of VEXAS syndrome was 1·00 compared to both the UBA1-WT and the MDS group as controls. A threshold of ≥10% of neutrophil precursors with >1 vacuole was associated with the diagnosis of VEXAS syndrome with a sensitivity of 100% and a specificity of 100%. The presence of vacuoles in myeloid precursors was observed in our two control conditions that represent the main differential diagnoses for VEXAS syndrome. However, an important proportion of neutrophil precursors was highly associated with VEXAS syndrome. The proportion of vacuolated precursors is therefore an accessible, cheap and efficient way to select patients who need UBA1 sequencing. We may not exclude that some mutations leading to VEXAS syndrome have not yet been identified. We think that special attention for in-depth genetic investigations should be given to patients with a phenotype of VEXAS syndrome and with ≥10% of neutrophil precursors with >1 vacuole but without a previously described UBA1 mutation. This could lead to uncovering unknown UBA1 mutations. The main limitation of our study is the relatively limited number of subjects and we cannot exclude that the sensitivity and specificity would be slightly lower than 100% in a larger cohort. However, we described six new patients with a VEXAS syndrome whereas the current number of reported patients in the literature is very low. In conclusion, our study confirmed that the sole presence of vacuoles in myeloid precursors is not specific to VEXAS syndrome. However, the presence of ≥10% of neutrophil precursors with >1 vacuole was associated with VEXAS syndrome with excellent sensitivity and specificity. None. Fig S1. Vacuoles in neutrophil precursor cells in bone marrow analysis of a patient in the VEXAS group (A) and a patient in the MDS group (B). Table SI. Detailed clinical and biological features during the course of the disease, treatments received and evolution of patients in the VEXAS and UBA1-wild type groups. Table SII. Number of cells with vacuoles, biological features and treatments of patients in the three groups. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

Topics & Concepts