Unsupervised cluster analysis and subset characterization of abnormal erythropoiesis using the bioinformatic Flow-Self Organizing Maps algorithm.

BACKGROUND: The Flow-Self Organizing Maps (FlowSOM) artificial intelligence (AI) program, available within the Bioconductor open-source R-project, allows for an unsupervised visualization and interpretation of multiparameter flow cytometry (MFC) data. METHODS: Applied to a reference merged file from 11 normal bone marrows (BM) analyzed with an MFC panel targeting erythropoiesis, FlowSOM allowed to identify six subpopulations of erythropoietic precursors (EPs). In order to find out how this program would help in the characterization of abnormalities in erythropoiesis, MFC data from list-mode files of 16 patients (5 with non-clonal anemia and 11 with myelodysplastic syndrome [MDS] at diagnosis) were analyzed. RESULTS: Unsupervised FlowSOM analysis identified 18 additional subsets of EPs not present in the merged normal BM samples. Most of them involved subtle unexpected and previously unreported modifications in CD36 and/or CD71 antigen expression and in side scatter characteristics. Three patterns were observed in MDS patient samples: i) EPs with decreased proliferation and abnormal proliferating precursors, ii) EPs with a normal proliferating fraction and maturation defects in late precursors, and iii) EPs with a reduced erythropoietic fraction but mostly normal patterns suggesting that erythropoiesis was less affected. Additionally, analysis of sequential samples from an MDS patient under treatment showed a decrease of abnormal subsets after azacytidine treatment and near normalization after allogeneic hematopoietic stem-cell transplantation. CONCLUSION: Unsupervised clustering analysis of MFC data discloses subtle alterations in erythropoiesis not detectable by cytology nor FCM supervised analysis. This novel AI analytical approach sheds some new light on the pathophysiology of these conditions.

Analysis of erythroid maturation in the nonlysed bone marrow with help of radar plots facilitates detection of flow cytometric aberrations in myelodysplastic syndromes.

BACKGROUND: Accumulating data support the role of flow cytometry (FCM) in diagnostic work-up of myelodysplastic syndromes (MDS). Changes in erythropoiesis are less documented than in granulopoiesis. However, most studies were performed on bone marrow samples (BMSs) after red blood cell lysis. We have established a FCM protocol for erythropoiesis, following a no-lysis approach and live gate acquisition of nucleated cells using DNA dye DRAQ5. METHODS: The ERY tube consisted of CD36, CD71, CD105, CD117, CD13, and CD45. Comparison with cytomorphological differential counts was carried out in a learning cohort of 80 BMS. To detect aberrations, we analyzed 208 BMS from 135 patients and five normal donors, divided into three cohorts: MDS (n = 68), nonclonal cytopenia (n = 43), and normal controls (n = 29). Radar plot (RP) was created for an overview of normal and aberrant patterns. RESULTS: The proportion of erythropoiesis in the ERY tube showed better agreement with the cytomorphology, compared to FCM panels on lysed BMS. We confirmed that aberrations in coefficient of variation (CV) of CD36 fluorescence intensity (p < .001), mean fluorescence intensity of CD36 (p = .012), and CV of CD105 (p < .001) can distinguish between MDS and nonclonal cytopenia. RP facilitated evaluation of erythropoietic maturation patterns and aberrant patterns were identified in 85% of MDS patients. CONCLUSION: This study provides evidence that a no-lysis approach and RP analysis allow a more reliable evaluation of erythropoiesis and erythroid dysplasia, supporting the integration of FCM erythroid panels in the standard work-up of MDS.

Improved minimal residual disease detection by targeted quantitative polymerase chain reaction in Nucleophosmin 1 type a mutated acute myeloid leukemia.

Multicolor flow cytometry (MFC) and real-time quantitative PCR (RQ-PCR) are important independent techniques to determine minimal residual disease (MRD) in acute myeloid leukemia (AML). MFC is the standard method, but may be unreliable. Therefore, MFC-based determination of MRD with an RQ-PCR-based approach targeting the nucleophosmin 1 (NPM1) type A mutation was set out to compare. Since most current NPM1 RQ-PCR MRD protocols suffer from clear definitions of quantifiability, we sought to define quantifiability in a reproducible and standardized manner. The limit of quantifiability of our RQ-PCR protocol for the NPM1 type A mutation varied between 0.002% and 0.04% residual leukemic cells depending on the features of the standard curve for each PCR experiment. The limit of detection was close to 0.001% leukemic cells. The limit of detection by MFC ranged from 0.01% to 1% depending on the phenotype of the leukemic cells as compared with non-leukemic bone marrow cells. Forty-five MRD samples from 15 patients using both NPM1 mutation specific RQ-PCR and MFC were analyzed. In 32 of the 45 samples (71%), an MRD-signal could be detected with RQ-PCR. A quantifiable NPM1 mutation signal was found in 15 samples (33%) (range 0.003%-2.6% leukemic cells). By contrast, only two follow-up samples (4%) showed residual leukemic cells (0.04% and 0.3%, respectively) by MFC. Thus, RQ-PCR of the NPM1 type A mutation was more sensitive and reliable than MFC for determination of MRD, which might have clinical implications. © 2016 Wiley Periodicals, Inc.

Intermittent maple syrup urine disease: two case reports.

The presenting symptoms and clinical course of 2 cases of intermittent maple syrup urine disease (MSUD) are described. Intermittent MSUD is a potentially life-threatening metabolic disorder caused by a deficiency of branched-chain α-keto acid dehydrogenase, the enzyme complex that decarboxylates the 3 branched-chain amino acids. In contrast to classic MSUD, children with the intermittent form show normal development with normal intelligence and, when asymptomatic, normal levels of branched-chain amino acids. Symptoms usually appear between 5 months and 2 years of age, when a trivial infection such as otitis media or viral gastroenteritis triggers catabolism of muscle protein. Intermittent MSUD should be suspected in cases of common infections with a clinically atypical course, especially in children displaying ataxia or marked drowsiness.

HDL stimulates apoM secretion.

Apolipoprotein M (apoM) in human plasma is mainly associated with HDL. A retained signal peptide anchors apoM to the lipoproteins. To investigate the role of the signal peptide in the transfer of apoM from the synthesizing cell to the lipoproteins, wildtype apoM cDNA and the Q(22)A mutant, introducing a signal peptidase cleavage site, were used to stably transfect HEK293 cells, which intrinsically do not express apolipoproteins. When cultured under serum-free conditions, wildtype apoM was, in contrast to Q(22)A, poorly secreted. Addition of serum or purified HDL stimulated secretion of wildtype apoM, which was recovered in the medium incorporated in HDL. The liver cell line HepG2, which synthesizes HDL, was cultured under serum-free conditions and found to secrete apoM as part of an HDL-like particle. In conclusion, due to its retained signal peptide, apoM is poorly secreted unless HDL is either coexpressed or added to the culture medium.

Levels of apolipoprotein M are not associated with the risk of coronary heart disease in two independent case-control studies.

Apolipoprotein M (apoM), a 25 kDa plasma protein belonging to the lipocalin protein family, is predominantly associated with HDL. Studies in mice have suggested apoM to be important for the formation of pre-beta-HDL and to increase cholesterol efflux from macrophage foam cells. Overexpression of human apoM in LDL receptor-deficient mice reduced the atherogenic effect of a cholesterol-rich diet. The aim of the present study was to investigate whether the apoM levels in man predict the risk for coronary heart disease (CHD). ApoM was measured in samples from two separate case-control studies. FINRISK '92 consisted of 255 individuals, of whom 80 developed CHD during follow-up and 175 were controls. The Copenhagen City Heart Study included 1,865 individuals, of whom 921 developed CHD during follow-up and 944 were controls. Correlation studies of apoM concentration with several analytes showed a marked positive correlation with HDL and total cholesterol as well as with apoA-I and apoB. There was no significant difference in mean apoM level between CHD and control subjects in either study. In conditional logistic regression analyses, apoM was not a predictor of CHD events, [odds ratio (95% CI) 0.97 (0.74-1.27) and 0.92 (0.84-1.02), respectively]. In conclusion, no association between apoM and CHD could be found in this study.

The signal peptide anchors apolipoprotein M in plasma lipoproteins and prevents rapid clearance of apolipoprotein M from plasma.

Lipoproteins consist of lipids solubilized by apolipoproteins. The lipid-binding structural motifs of apolipoproteins include amphipathic alpha-helixes and beta-sheets. Plasma apolipoprotein (apo) M lacks an external amphipathic motif but, nevertheless, is exclusively associated with lipoproteins (mainly high density lipoprotein). Uniquely, however, apoM is secreted to plasma without cleavage of its hydrophobic NH(2)-terminal signal peptide. To test whether the signal peptide serves as a lipoprotein anchor for apoM in plasma, we generated mice expressing a mutated apoM(Q22A) cDNA in the liver (apoM(Q22A)-Tg mice (transgenic mice)) and compared them with mice expressing wild-type human apoM (apoM-Tg mice). The substitution of the amino acid glutamine 22 with alanine in apoM(Q22A) results in secretion of human apoM without a signal peptide. The human apoM mRNA level in liver and the amount of human apoM protein secretion from hepatocytes were similar in apoM-Tg and apoM(Q22A)-Tg mice. Nevertheless, human apoM was not detectable in plasma of apoM(Q22A)-Tg mice, whereas it was easily measured in the apoM-Tg mice. To examine the plasma metabolism, recombinant apoM lacking the signal peptide was produced in Escherichia coli and injected into wild-type mice. The apoM without signal peptide did not associate with lipoproteins and was rapidly cleared in the kidney. Accordingly, ligation of the kidney arteries in apoM(Q22A)-Tg mice resulted in rapid accumulation of human apoM in plasma. The data suggest that hydrophobic signal peptide sequences, if preserved upon secretion, can anchor plasma proteins in lipoproteins. In the case of apoM, this mechanism prevents rapid loss by filtration in the kidney.

Apolipoprotein M associates to lipoproteins through its retained signal peptide.

Apolipoprotein M (apoM) is predominantly associated with HDL. In this study, it was investigated whether apoM's uncleaved signal peptide is necessary for the protein's ability to associate with lipoproteins. ApoM with a cleavable signal peptide, Q22A, was expressed, together with wild-type apoM, in HEK293 cells. On size-exclusion chromatography, the elution profile of wild-type apoM was similar to that of human HDL-associated plasma apoM. In contrast, the size of the Q22A mutant corresponded to free, unassociated apoM. This strongly indicates that the signal peptide is indeed necessary for apoM's ability to associate with lipid.

An ELISA for apolipoprotein M reveals a strong correlation to total cholesterol in human plasma.

Apolipoprotein M (apoM) is a 188 amino acid, 25 kDa protein belonging to the lipocalin protein superfamily. Although predominantly associated with high density lipoprotein, apoM is found in all major lipoprotein classes. To facilitate clinical studies of apoM, we have developed a sandwich ELISA for the measurement of apoM in human plasma. This method has been used to investigate normal apoM variation and to establish reference values for healthy individuals through the measurement of 598 samples from the Nordic Reference Interval Project Bio-bank and Database (NOBIDA) biobank. For women 18-49 years old, the reference interval for apoM was 0.58-1.18 micromol/l, whereas for women 50+ years and for men, the reference range was 0.61-1.30 micromol/l. Correlation studies of apoM with 26 common clinical chemical analytes from the NOBIDA database revealed a marked positive correlation with plasma total cholesterol (r = 0.52) and LDL and HDL cholesterol (r = 0.43 and 0.36, respectively). There was no statistically significant correlation with HDL/total cholesterol ratio or body mass index. In conclusion, a sandwich ELISA for the measurement of apoM in human plasma shows that apoM concentration is strongly correlated to total cholesterol in healthy individuals.

Hydrophobic ligand binding properties of the human lipocalin apolipoprotein M.

Apolipoprotein M (apoM) is a plasma protein associated mainly with HDL. ApoM is suggested to be important for the formation of prebeta-HDL, but its mechanism of action is unknown. Homology modeling has suggested apoM to be a lipocalin. Lipocalins share a structurally conserved beta-barrel, which in many lipocalins bind hydrophobic ligands. The aim of this study was to test the ability of apoM to bind different hydrophobic substances. ApoM was produced both in Escherichia coli and in HEK 293 cells. Characterization of both variants with electrophoretic and immunological methods suggested apoM from E. coli to be correctly folded. Intrinsic tryptophan fluorescence of both apoM variants revealed that retinol, all-trans-retinoic acid, and 9-cis-retinoic acid bound (dissociation constant = 2-3 microM), whereas other tested substances (e.g., cholesterol, vitamin K, and arachidonic acid) did not. The intrinsic fluorescence of two apoM mutants carrying single tryptophans was quenched by retinol and retinoic acid to the same extent as wild-type apoM, indicating that the environment of both tryptophans was affected by the binding. In conclusion, the binding of retinol and retinoic acid supports the hypothesis that apoM is a lipocalin. The physiological relevance of this binding has yet to be elucidated.

Isolation and characterization of human apolipoprotein M-containing lipoproteins.

Apolipoprotein M (apoM) is a novel apolipoprotein with unknown function. In this study, we established a method for isolating apoM-containing lipoproteins and studied their composition and the effect of apoM on HDL function. ApoM-containing lipoproteins were isolated from human plasma with immunoaffinity chromatography and compared with lipoproteins lacking apoM. The apoM-containing lipoproteins were predominantly of HDL size; approximately 5% of the total HDL population contained apoM. Mass spectrometry showed that the apoM-containing lipoproteins also contained apoJ, apoA-I, apoA-II, apoC-I, apoC-II, apoC-III, paraoxonase 1, and apoB. ApoM-containing HDL (HDL(apoM+)) contained significantly more free cholesterol than HDL lacking apoM (HDL(apoM-)) (5.9 +/- 0.7% vs. 3.2 +/- 0.5%; P < 0.005) and was heterogeneous in size with both small and large particles. HDL(apoM+) inhibited Cu(2+)-induced oxidation of LDL and stimulated cholesterol efflux from THP-1 foam cells more efficiently than HDL(apoM-). In conclusion, our results suggest that apoM is associated with a small heterogeneous subpopulation of HDL particles. Nevertheless, apoM designates a subpopulation of HDL that protects LDL against oxidation and stimulates cholesterol efflux more efficiently than HDL lacking apoM.

Characterization of apoM in normal and genetically modified mice.

A novel human apolipoprotein [apolipoprotein M (apoM)] was recently described and demonstrated to be a lipocalin. We have now examined apoM in wild-type mice and mice with genetically altered lipoprotein metabolism. Liver and kidney showed high mRNA expression, whereas spleen, heart, brain, and testis demonstrated low expression. ApoM gene expression was initiated on embryonic day 10. Western blot analysis of plasma suggested that mouse apoM, like its human counterpart, is secreted with a retained signal peptide, but unlike human apoM it is not glycosylated. Gel filtration of plasma showed apoM to be associated with HDL-sized particles in wild-type and apoA-I-deficient mice and with HDL- and LDL-sized particles in LDL receptor-deficient mice, whereas apoM was mainly found in VLDL-sized particles in high-fat, high-cholesterol-fed apoE-deficient mice. The plasma concentration of apoM was similar in wild-type, LDL receptor-deficient, and apoE-deficient mice but was reduced to 33% in apoA-I-deficient compared with wild-type mice (P = 0.007). These data suggest that apoM mainly associates with HDL in normal mice but also with the pathologically increased lipoprotein fraction in genetically modified mice. The substantially decreased apoM levels in apoA-I-deficient mice suggest a connection between apoM and apoA-I metabolism.