Erythroid precursors from patients with low-risk myelodysplasia demonstrate ultrastructural features of enhanced autophagy of mitochondria.

Recent studies in erythroid cells have shown that autophagy is an important process for the physiological clearance of mitochondria during terminal differentiation. However, autophagy also plays an important role in removing damaged and dysfunctional mitochondria. Defective mitochondria and impaired erythroid maturation are important characteristics of low-risk myelodysplasia. In this study we therefore questioned whether the autophagic clearance of mitochondria might be altered in erythroblasts from patients with refractory anemia (RA, n=3) and RA with ringed sideroblasts (RARS, n=6). Ultrastructurally, abnormal and iron-laden mitochondria were abundant, especially in RARS patients. A large proportion (52+/-16%) of immature and mature myelodysplastic syndrome (MDS) erythroblasts contained cytoplasmic vacuoles, partly double membraned and positive for lysosomal marker LAMP-2 and mitochondrial markers, findings compatible with autophagic removal of dysfunctional mitochondria. In healthy controls only mature erythroblasts comprised these vacuoles (12+/-3%). These findings were confirmed morphometrically showing an increased vacuolar surface in MDS erythroblasts compared to controls (P<0.0001). In summary, these data indicate that MDS erythroblasts show features of enhanced autophagy at an earlier stage of erythroid differentiation than in normal controls. The enhanced autophagy might be a cell protective mechanism to remove defective iron-laden mitochondria.

Megakaryocytic dysfunction in myelodysplastic syndromes and idiopathic thrombocytopenic purpura is in part due to different forms of cell death.

Platelet production requires compartmentalized caspase activation within megakaryocytes. This eventually results in platelet release in conjunction with apoptosis of the remaining megakaryocyte. Recent studies have indicated that in low-risk myelodysplastic syndromes (MDS) and idiopathic thrombocytopenic purpura (ITP), premature cell death of megakaryocytes may contribute to thrombocytopenia. Different cell death patterns have been identified in megakaryocytes in these disorders. Growing evidence suggests that, besides apoptosis, necrosis and autophagic cell death, may also be programmed. Therefore, programmed cell death (PCD) can be classified in apoptosis, a caspase-dependent process, apoptosis-like, autophagic and necrosis-like PCD, which are predominantly caspase-independent processes. In MDS, megakaryocytes show features of necrosis-like PCD, whereas ITP megakaryocytes demonstrate predominantly characteristics of apoptosis-like PCD (para-apoptosis). Triggers for these death pathways are largely unknown. In MDS, the interaction of Fas/Fas-ligand might be of importance, whereas in ITP antiplatelet autoantibodies recognizing common antigens on megakaryocytes and platelets might be involved. These findings illustrate that cellular death pathways in megakaryocytes are recruited in both physiological and pathological settings, and that different forms of cell death can occur in the same cell depending on the stimulus and the cellular context. Elucidation of the underlying mechanisms might lead to novel therapeutic interventions.

Mitochondrial disruption and limited apoptosis of erythroblasts are associated with high risk myelodysplasia. An ultrastructural analysis.

AIMS: To investigate the ultrastructural characteristics of erythroblasts in myelodysplasia (MDS) which might be of additional importance in understanding its pathogenesis. METHODS AND RESULTS: 22 patients were classified according to FAB (French-American-British classification), IPSS (international prognostic score system), cytogenetic risk factors and transfusion dependency. Using electron microscopy, in 77% of the cases, nuclear abnormalities consisting of disrupted membranes and cystic/dilated perinuclear spaces were noted. In a limited number of patients (n=7), a low percentage of apoptosis in the erythroid lineage (mean 3.1+/-1.6%; median 3%: range 1-6) (normal controls: <0.5%) could be noted, primarily in mature erythroblasts and significantly associated with spongiform nuclear features. In all patients extensive cytoplasmic vacuolization and myelin figures in erythroblasts were demonstrated. In 55% of the cases, enlarged and abnormal mitochondria were observed, significantly associated with iron-accumulation. A significant inverse relation existed between the absence of apoptosis and more advanced, or high risk disease and cytogenetic risk factors. Mitochondrial abnormalities were significantly correlated with high risk disease as well with an increase in transfusion dependency. CONCLUSIONS: These data indicate that in MDS apoptosis may play a role in early stages of disease. The overall prominent defects in mitochondria might be an additional defect that is involved in ineffective erythropoiesis.

The in vitro effects of cytokines on expansion and migration of megakaryocyte progenitors.

Increasing the number of megakaryocyte progenitors in stem cell transplants by ex vivo expansion culture may be an approach to accelerate platelet recovery in patients undergoing high-dose chemotherapy. We evaluated the effect of three different cytokine combinations on expansion, with special emphasis on the type of colony formation and migration of megakaryocytic cells. The number of clonogenic megakaryocyte progenitors (colony-forming units-megakaryocyte; CFU-Mk) with high- (> 20 cells/colony) and low-proliferative capacity (5-20 cells/colony) and the number of megakaryocytic (CD61+) cells were significantly increased by including interleukin 3 (IL-3) or IL-3 + IL-6 + IL-11 + Flt3-ligand to cultures containing megakaryocyte growth and development factor (MGDF) plus stem cell factor (SCF). No difference in the maturation of megakaryocytes from all three cytokine combinations to platelets were observed, as demonstrated by electron microscopy. In chemotaxis experiments, the migration towards stromal cell-derived factor 1 (SDF-1) was shown to be reduced for CD61+ cells and megakaryocyte progenitors cultured in other cytokines besides MGDF + SCF. The reduced migration was related to a lower expression of CXCR4, the receptor for SDF-1, on megakaryocytes from the proliferating cultures. These in vitro results demonstrate that expansion in IL-3 and other cytokines besides MGDF + SCF significantly impair the capacity of megakaryocytic cells to migrate.

Excessive apoptosis of bone marrow erythroblasts in a patient with autoimmune haemolytic anaemia with reticulocytopenia.

We report a patient with autoimmune haemolytic anaemia (AIHA) with reticulocytopenia, who showed excessive apoptosis of erythroblasts. Ultrastructural analysis of bone marrow cells showed that 50% of erythroblasts had characteristic features of apoptosis, which was confirmed by staining with Annexin-V. In addition, in contrast to normal erythroblasts, Ig staining of the entire erythroblast population could be shown. These data show that apoptosis may contribute to the mechanism of reticulocytopenia in AIHA.

A simple procedure for in situ immunolabeling, embedding and sectioning of layers of cultured endothelial and smooth muscle cells for both light and electron microscopy.

We present a simple procedure for in situ immunolabeling, embedding and sectioning of layers of cultured endothelial and smooth muscle cells for both light and electron microscopy. Endothelial and smooth muscle cells were seeded in tissue culture chambers/slides precoated with 30% (w/v) gelatin drops fixed with 0.5% glutaraldehyde. Live endothelial cell layers were labeled with an antibody against the surface membrane protein, anti-CD13. After labeling, the cell layers were fixed and separated from the chambers/slides by lifting all of the samples with a spatula. Sections (1-2 mm) were cut, embedded and processed further for light or electron microscopy. Because of the delicate cell layers and the importance of preserving maximum integrity, labeling was performed under standard culture conditions and treated in situ during the entire procedure. Moreover, the small chamber size of the tissue culture dishes generated the additional advantages of requiring only a limited number of cells, small volumes of media, and little antibody.

Association of smooth muscle cell tissue factor with caveolae.

There is still no satisfactory explanation for the low catalytic activity of tissue factor (TF)/factor VII(a) complexes towards coagulation factor X, as found on the apical surface side of cell layers. It has been hypothesized that TF exists in a latent form. Layers of cultured human smooth muscle cells, constitutively expressing TF, were immunogold-labeled for TF in situ and processed for electron microscopy. We showed that, besides internalization and accumulation in lysosomal-like structures, TF remained associated with noncoated, flask-shaped microinvaginations of the plasma membrane. These invaginations were identified as caveolae. In regions in which intercellular contacts were interrupted, more TF-positive caveolae were observed. Enzymatically detached smooth muscle cells exhibited a similar enlargement of caveolar structures. Concomitantly, an increase of catalytic activity of apically formed TF/VIIa complexes towards factor X was found on the suspended cells. We speculate that caveolae-associated TF may function as a latent pool of procoagulant activity, which can rapidly be activated at sites in which vessel wall integrity is lost.

Basal tissue factor expression in endothelial cell cultures is caused by contaminating smooth muscle cells. Reduction by using chymotrypsin instead of collagenase.

A discrepancy exists between basal tissue factor (TF) expression found in endothelial cell cultures and the failure to detect TF in unpertubated endothelial cells in vivo. We demonstrated that basal TF expression in endothelial cell cultures originated from contaminating cells. These cells were ultrastructurally and flowcytometrically identified as smooth muscle cells. The cell cultures had been obtained from collagenase-treated human umbilical cord vessels. Histologic studies revealed that after collagenase treatment the basement membrane was digested and underlying structures were disrupted at some areas of the vein. We selected chymotrypsin as an alternative for the isolation of endothelial cells. Using chymotrypsin, the endothelial lining was selectively lost leaving the basement membrane undisturbed. Furthermore, use of chymotrypsin instead of collagenase minimized the level of basal TF activity. Taken together, we demonstrated that basal TF expression in endothelial cell cultures is caused by contaminating smooth muscle cells. This contamination can strongly be reduced using chymotrypsin instead of collagenase for isolation of endothelial cells.

Tumor necrosis factor alpha-induced endothelial tissue factor is located on the cell surface rather than in the subendothelial matrix.

Because there is no consensus regarding the precise distribution of induced endothelial tissue factor (TF), we studied TF activity in and on tumor necrosis factor alpha-stimulated cultured human umbilical vein endothelial cells (ECs) and their underlying matrix. TF was mainly expressed on the cell surface. Only small traces were found on the apical surface suggesting that TF is predominantly located on the basolateral side of the cell membrane. The presence of TF on the cell surface was confirmed by flow cytometry. Subendothelial TF activity appeared to be dependent upon the procedure used to remove the stimulated EC monolayer. Whereas ammonium hydroxide or hypotonic lysis resulted in relatively high levels of matrix-associated TF, virtually no TF was found on the matrix after mild enzymatic detachment of stimulated ECs. Cell removal with EDTA resulted in intermediate levels of matrix-associated TF. Neither the enzymatic treatment nor EDTA degraded or removed this TF activity. Similar patterns were observed for matrix-associated TF antigen and EC surface markers. Electron microscopic analysis showed cell fragments on the matrix after monolayer lysis. The findings strongly suggest that induced endothelial TF associated with the subendothelial matrix actually represents TF on EC remnants.

Natural killer cell surface markers on cells containing parallel tubular structures.

Parallel tubular structures (PTS) and/or the associated electron-dense granules, thought to contain molecules responsible for target cell lysis, can be detected in the cytoplasm of lymphocytes with large granular lymphocyte (LGL) morphology. In the present study we compared PTS presence in freshly isolated peripheral blood lymphocytes and peripheral blood lymphocytes incubated overnight in the presence of human pooled serum, sera from patients with Hodgkin's disease and interferon-alpha. Under all conditions we found PTS in the majority of CD16+ cells (64.3-74.8%) but less than 41.8% in CD57+ cells. In the case of double-labeled lymphocytes, 41.0-61.7% CD16+/8+ but only 24.2-27.5% CD57+/8+ cells were PTS+. Thus, in all cases where lymphocytes expressed CD16 antigen there was a high percentage of PTS positivity. Although the PTS+ cells exhibited phenotypic heterogeneity there was, except for a proportion of CD57+ lymphocytes which exhibited less of the characteristic LGL features, generally LGL morphological homogeneity. CD16 lymphocytes are potentially more cytotoxic than CD57 and CD8 cells. Taking this into consideration, the presence of the PTS in the majority of CD16 cells suggests an important role for PTS in target cell lysis.

Immunoelectron microscopic characterization of lymphocytes containing parallel tubular structures induced by natural but not recombinant interferon-alpha.

Activation of natural killer (NK) cell activity is one of the immune functions which can be altered by the interferons (IFNs). We previously incubated healthy donor peripheral blood lymphocytes (PBMC) in the presence of natural (n) IFN-alpha and found induction of granular lymphocytes, a proportion of which expressed CD 8 and CD 57. In the present study we further delineated the membrane characteristics of these induced granular lymphocytes and observed the greatest proportion to be CD 16+, with lesser proportions positive for CD 8, CD 57, or coexpressing CD 16 and CD 8. Thus, nIFN-alpha-induced granular lymphocytes have both the morphological and membrane characteristics of functional NK cells. However, in contrast to nIFN-alpha, incubations with recombinant (r) IFN-alpha and n- and rIFN-gamma were found to only enhance NK activity; the concomitant increase in granular lymphocytes as observed after nIFN-alpha incubation was absent. Therefore, even though the different IFNs applied had comparable effects on the function of cytotoxic effector cells, taking into account the observed morphological discrepancy, the unknown mechanisms or pathways by which this is achieved are apparently not identical.

Gold probe choice in simultaneous detection of human lymphocyte surface antigens at the ultrastructural level.

A double immunogold-labeling method in immunoelectron microscopy was used for simultaneous detection of two antigens by monoclonal antibodies [OKT 8 (CD 8), anti-Leu-7, anti-Leu-11b (CD 16)] on lymphocytes in suspension. The combination of gold probe size (5 nm and 15 nm) and monoclonal antibody was found to be decisive for detecting double-labeled cells with the OKT 8+, Leu-11b+ phenotype. The combinations of OKT 8 labeled with the 5-nm gold probe (OKT 8(5] and anti-Leu-11b with the 15-nm gold probe (Leu-11b15) gave double-labeled cells; the reverse situation, using OKT 8 with a 15-nm gold probe (OKT 8(15] and anti-Leu-11b with a 5-nm gold probe (Leu-11b5), did not. Double-labeled OKT 8+, Leu-7+ cells were detected irrespective of which gold probe combination was applied. Our findings indicate that although the double immunogold-labeling method is well suited for study of lymphocyte subsets, it is important to determine suitable combinations of gold probe sizes and monoclonal antibodies for the lymphocyte subset under study, taking into account surface antigen density, so that double labeling ensues.

Interferon-alpha induction of lymphocytes containing parallel tubular structures.

The induction by IFN-alpha in peripheral blood lymphocytes of parallel tubular structures (PTS) and/or electron-dense granules occurring in a minority of peripheral blood lymphocytes was examined. IFN reportedly augments natural killer (NK) cell activity of large granular lymphocytes (LGL); these cells contain PTS and/or electron-dense granules. Normal peripheral blood mononuclear cells were incubated with IFN-alpha and surface antigen expression was measured by means of indirect immunofluorescence and, at the ultrastructural level, using gold labelled monoclonal antibodies. Surface antigen reactivity with the monoclonal antibodies OKT 3, 4, 8 and Anti-Leu-7 (HNK-1) showed no difference between the IFN-alpha incubation and non-IFN-alpha groups. However, electron microscope investigation revealed significant absolute increases in the percentage of OKT 8+ and Anti-Leu-7+ cells which were PTS-positive after IFN-alpha treatment compared with the control groups. The cytotoxicity assay using the K562 cell line showed enhanced lytic activity. Our results suggest that cells coexpressing the OKT 8 and Leu-7 antigens may be responsible for a minor proportion of the increase in PTS but that IFN-alpha mainly induces PTS and/or associated structures in cells which express the OKT 8+ antigen. These PTS+/OKT 8+ cells may contribute to enhanced cell cytotoxicity.

T cells in patients with chronic T gamma lymphocytosis: morphology, cytochemistry, ultrastructure and immunological characteristics.

We have studied the morphology and cytochemistry in relation to the immunological phenotyping and functional properties of T cells from eight patients with chronic T gamma lymphocytosis. At the light microscopic level the morphology of the patients' lymphocytes was similar to that described for large granular lymphocytes. Ultrastructurally, a division into two groups could be made on differences in the amount of cytoplasm and the location and the more irregular form of the nuclei. The lymphocytes of one group (five patients) had in common the phenotype Fc gamma +, T3 +, T4 -, T8 +, Ia -, M1 - and demonstrated (with the exception of one patient) the same functions: presence of K-cell activity, absence of NK, helper and suppressor cell activities. In the other group (three patients), the lymphocytes of one patient showed the same phenotype and functions as those indicated above. The other two patients both lacked the T8 antigen on their lymphocytes but were different with regard to other surface markers. In addition, their cells were functionally identical: both demonstrated NK- and K-cell activity. Thus in this group of eight patients with chronic T gamma lymphocytosis, the immunological and functional subdivision paralleled in part a morphological division at the ultrastructural level.

Effect of sera from patients with Hodgkin's disease on normal donor lymphocytes containing parallel tubular structures.

Normal donor peripheral blood mononuclear cells incubated for 24 h with sera from patients with Hodgkin's disease were investigated by electron microscopy for the presence of parallel tubular structures (PTS) and/or amorphous electrondense granules (large granular lymphocytes = LGL). In comparison with normal human serum, 14 out of 29 sera of the patients induced a marked increase in the percentage of LGL. From a limited number of experiments it was likely that this increase is paralleled by an increase in Fcgamma receptor-bearing cells after the incubation. This serum effect did not show a correlation with the number of Fcgamma receptor-positive lymphocytes in the peripheral blood of the patients. A difference in the induction effect could be demonstrated between the sera from patients with a favourable and those with an unfavourable clinical course, but this distinction was not absolute. The presence or absence of splenic involvement by Hodgkin's disease does not apparently influence the effect of the sera. From experiments using sera positive for immune complexes or anti-Epstein-Barr virus antibodies, it seems unlikely that these factors are responsible for the observed increase in LGL.

Susceptibility of the expression of parallel tubular structures in lymphocytes to the exposure to ammonium chloride buffer.

In isolation procedures for lymphocytes, isotonic ammonium chloride buffers are frequently used to lyse erythrocytes. In addition to evidence for an effect on lymphocyte function we now report a morphological alteration after treatment of lymphocytes with ammonium chloride buffer. Expression of parallel tubular structures was lost and only amorphous, electron-dense granules could be observed.

Lymphocytes with parallel tubular structures: morphologically a distinctive subpopulation.

Peripheral blood lymphocytes were fractionated in T, B and Null cell enriched subsets by means of sheep red blood cell rosette (ESRBC) sedimentation and nylon wool adherence. The ultrastructural features of these subpopulation were investigated. The T cell fraction in which the sheep erythrocytes were removed from the ESRBC rosette-forming cells (ESRBC-RFC) by lysis with ammonium chloride, consisted mainly of two morphologically distinctive subsets. The majority of the cells (80%) displayed a smooth surface membrane and had a high nuclear to cytoplasmic ratio with few cytoplasmic organelles. The other cell type (18%) had a relatively rough surface membrane, a low nuclear to cytoplasmic ratio, often an indented nucleus and numerous cytoplasmic organelles such as characteristic amorphous granules and sometimes parallel tubular structures (p.t.s.). If the T cells were obtained after mechanical vibration of the ESRBC-RFC, the majority of these cells appeared morphologically identical to this latter cell type. Cells with p.t.s. and amorphous granules were also demonstrated within the Null and B cell enriched fractions (50% and 25% respectively), though in the B cell enriched fraction this cell type is probably due to a contamination of Null cells. Previous observations had already demonstrated that these cells in the three fractions represent the Fc gamma receptor-bearing lymphocytes. The similarities suggest that the Fc gamma receptor-bearing and p.t.s. containing lymphocytes form a morphologically distinct subpopulation.

Parallel tubular structures in T,B and null lymphocyte subpopulations.

T, B and Null lymphocyte-enriched subpopulations were isolated by means of sheep red blood cell (E SRBC) sedimentation and nylon wool adherence. In these subsets the proportion of Fc receptor-bearing cells was determined by the antibody-coated human and ox erythrocyte rosette assays (EA Hu and Ox). In the T-cell fraction the proportion of Fc-bearing lymphocytes was dependent on the procedure by which the sheep erythrocytes were removed from the E SRBC rosette-forming cells. Mechanical vibration resulted in considerably higher percentages of EA rosette-forming cells (EA-RFC) than osmotic shock or lysis with ammonium chloride. In all three procedures the number of EA Hu-RFC was slightly higher than the number of EA Ox-RFC. In the B-cell fraction the proportion of EA Ox-RFC was higher than that of EA Hu-RFC, mean values 43 and 33%, respectively. In the Null cell fraction the reversed was seen: mean values 54% EA Hu-RFC and 45% EA Ox-RFC. In all three lymphocyte fractions the majority of the EA-RFC contained parallel tubular structures and/or associated amorphous granules. Non-Fc receptor-bearing lymphocytes only occasionally showed parallel tubular structures or the amorphous granules.