Activation of mononuclear phagocytes and its relationship to asplenia and phosphatidylserine exposing red blood cells in hemoglobin E/β-thalassemia patients

Authors

  • Wansa Banyatsuppasin,

    1. Department of Pathology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
    Search for more papers by this author
  • Punnee Butthep,

    1. Department of Pathology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
    Search for more papers by this author
  • Vichai Atichartakarn,

    Corresponding author
    1. Division of Hematology, Department of Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
    • Division of Hematology, Department of Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Rama VI Road, Bangkok 10400, Thailand
    Search for more papers by this author
  • Ammarin Thakkinstian,

    1. Section for Clinical Epidemiology and Biostatistics, Research Center, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
    Search for more papers by this author
  • Napaporn Archararit,

    1. Laboratory Section, Research Center, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
    Search for more papers by this author
  • Kovit Pattanapanyasat,

    1. Office for Research and Development, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand
    Search for more papers by this author
  • Suporn Chuncharunee

    1. Division of Hematology, Department of Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand
    Search for more papers by this author

  • Conflict of interest: Nothing to report

Activation of Mononuclear Phagocytes and Its Relationship to Asplenia and Phosphatidylserine Exposing Red Blood Cells in Hemoglobin E/β-Thalassemia Patients

Aged or abnormal red blood cells with exposed phosphatidylserine (PS-RBCs) are cleared from the circulation by splenic macrophages. In asplenic patients, other mononuclear phagocytic cells in tissues and in circulation may function in this capacity. To better understand these changes and the relationship among splenic status, PS-RBCs, blood monocytes, and serum tumor necrosis factor (TNF-α), a product of mononuclear phagocyte activation, patients with hemoglobin E/β-thalassemia (E/β-Thal) were studied. Whole blood of 20 nonsplenectomized, 20 splenectomized E/β-Thal patients, and 20 healthy subjects was assayed for PS-RBCs; for monocytes, activated monocytes, and monocyte response to lipopolysaccharide stimulation; and serum was assayed for TNF-α. Asplenic E/β-Thal patients had significantly increased (P < 0.05) amounts of PS-RBCs, monocytes, activated monocytes, and levels of serum TNF-α. The amount of PS-RBCs correlated with levels of serum TNF-α, but the amount of activated monocytes did not correlate with either the amount of PS-RBCs or levels of serum TNF-α. Monocyte response to lipopolysaccharide stimulation in asplenic patients was not as efficient as in the other patients or in normals (77 vs. 404, and 304 folds increment, respectively). The results suggest that splenectomy in E/β-Thal patients led to an increased amount of PS-RBCs and activation in the mononuclear phagocytic system.

Thalassemia (Thal), an autosomal recessive hereditary hemolytic anemia, is the most common known human genetic disorder. It is caused by mutations of the globin gene clusters resulting in varying degrees of decreased globin chain synthesis. It is classified into α and β according to the globin chain involved [1]. It can be co-inherited with an abnormal hemoglobin (Hb), such as Hb E, resulting in Hb E/β-Thal (E/β-Thal), which is the most common form of the symptomatic thalassemic syndromes globally [2, 3]. The produced defective red blood cells (RBC) are cleared prematurely by various mechanisms during their passages through the spleen [4]. One important mechanism is through splenic macrophages' recognition of phosphatidylserine (PS)-exposing cells [5]. From overwork, the spleen may become very large, and splenectomy is usually performed for hypersplenism and/or symptomatic splenomegaly. These defective and pathologic RBCs must then be cleared by other mononuclear phagocytic or reticuloendothelial (RE) cells, such as Kupffer cells [6], whose function is not as effective as shown by an increased amount of such cells in circulation in splenectomized E/β-Thal patients [5, 7].

Monocyte activation has previously been reported in both Thal [8–10] and sickle cell disease (SCD) [11, 12]. Increased antibody-dependent cellular cytotoxicity [8] and up-regulation of Fcγ receptor I (FcγRI or CD64) expression [9] of monocytes in response to their clearance of thalassemic RBCs were reported in both α- and β-Thal. Macrophage colony-stimulating factor, interferon-γ, and tumor necrosis factor (TNF-α) also play roles as inducers and effectors of this activation [8, 9]. Increased serum TNF-α, interleukin-6, and interferon-γ levels were reported in E/β-Thal patients, particularly after splenectomy [13, 14]. Increased amounts of TNF-α and interleukin-1β per cell, indicative of monocyte activation, and increased serum C-reactive protein levels were also reported in SCD [11], who invariably have functional asplenia.

To better understand the inter-relationship among splenic status, PS-exposing RBCs (PS-RBCs), monocyte activation, and serum TNF-α, a product of mononuclear phagocyte activation, we studied these parameters in E/β-Thal patients with and without intact spleen, and compared them with those of normal controls (NC).

In total, 40 ambulatory and well E/β-Thal patients who attended the adult hematology clinic at Ramathibodi Hospital were studied. Their age ranged from 19 to 57 years, and 19 were men. They were free from medication and blood transfusion for at least four preceding weeks. Half of them had undergone splenectomy, with a median interval since splenectomy of 13.5 years (range, 5–35). In nonsplenectomized (NS) patients, median vertical measurement from splenic tip on palpation to left costal margin was 3.5 cm (range, 0.5–12.5). Blood transfusion and iron chelation therapy were usually modest, and at the discretion of the attending physicians. Controls were 20 consenting healthy subjects, who had not taken any drugs for at least four preceding weeks. Study protocol was approved by the institutional ethics committee for studies in humans (11-46-31). Written informed consents were obtained from all patients.

Patients' characteristics, mean (standard deviation) or median (range) of hematological data, absolute amounts of PS or annexin V (AV)+ RBCs, activated monocytes (CD14+ expressing intracellular TNF-α and CD14+/CD11b+), and levels of serum TNF-α of the patients and normal controls (NC) are shown in Table I. Total amount of packed RBCs transfusion in the two patients' groups almost reached statistically significant difference (P = 0.055, independence t-test). Values of platelet count in the NS group had a wide variation because of the varied spleen sizes. Splenectomized (S) patients had significantly lower amounts of RBCs, and significantly higher amounts of reticulocytes and NRBCs than the others (P < 0.05), indicating more severe hemolysis. They also had significantly higher amounts of white blood cells (WBCs), monocytes, platelets, PS-RBCs, activated monocytes, and levels of serum TNF-α than the others (P < 0.05). Similar to the amount of monocytes, values of the latter three in the NS and NC groups were not significantly different, suggesting an influential role of the spleen on these parameters.

Table I. Characteristics and Results of Studied Blood Profiles Expressed as Means (SD) or Median (Range) in Each Group of Subjects
CharacteristicHemoglobin E/β-thalassemiaNormal controls (n = 20)
Splenectomized (n = 20)Nonsplenectomized (n = 20)
  • Kruskal–Wallis with corrected Mann–Whitney test:

  • a

    significant difference compared with nonsplenectomized group (P < 0.05).

  • b

    significant difference compared with normal controls (P < 0.05).

  • One-way analysis of variance with Bonferroni test:

  • c

    significant difference compared with normal controls (P < 0.05).

  • d

    significant difference compared with nonsplenectomized group (P < 0.05).

Age (year)26.0 (20.0–48.0)a33.0 (19.0–57.0)b24.0 (18.0–44.0)
Gender (male:female)9:1110:109:11
Total packed red blood cells transfusion (unit)83.5 ± 11122.4 ± 41.40
Hematologic parameters   
 RBC (1012/L)2.8 ± 0.4c,d3.8 ± 1.0c4.7 ± 0.6
 Hb (g/L)60 (45–80)a,b71 (55–117)b131 (106–176)
 Hct (proportion of 1.0)0.209 (0.164–0.288)a,b0.229 (0.193–0.368)b0.396 (0.324–0.547)
  MCV (fL)75.4 ± 7.0c,d65.3 ± 7.8c86.5 ± 5.9
  MCH (fmol/cell)1.4 ± 0.1c1.3 ± 0.2c1.8 ± 0.2
  MCHC (mmol/L)18.0 ± 1.6c19.2 ± 1.2c20.3 ± 1.9
 Reticulocyte (109/L)239.1 (154.0–509.6)a,b75.9 (48.4–193.2)b41.2 (8.5–76.7)
 NRBC/100 WBC598.0 (61.0–1,418.0)a,b4.0 (0.0–66.0)b0
 WBC (109/L)10.2 ± 2.2c,d7.6 ± 2.2c5.7 ± 1.2
  Monocyte (106/L)976.0 (223.0–2,650.0)a,b295.0 (100.0–909.0)372.0 (180.0–990.0)
 Platelet (109/L)735.0 (450.0–1,056.0)a,b227.5 (31.0–423.0)280.0 (163.0–381.0)
Absolute amount of annexin V + RBCs (109/L)96.3 ± 44.9c,d46.9 ± 21.640.2 ± 19.3
Absolute amount of CD14+-expressing intracellular TNF-α (106/L)4.3 (1.0–67.9)a,b0.5 (0.1–10.5)1.2 (0.1–2.9)
MFI of CD14+/CD11b+65.7 ± 28.2c,d44.1 ± 12.051.4 ± 10.8
Serum TNF-α (ng/L)11.3 ± 3.6c,d8.4 ± 3.37.1 ± 1.6

Correlation among absolute amounts of activated monocytes, PS-RBCs, and levels of serum TNF-α were assessed by multiple linear regression analysis after controlling for types of subjects. In consideration of skewness, data were first transformed to a log-scale. The only statistically significant correlation (coefficient = 0.914, standard error = 0.188, P < 0.001) was between levels of serum TNF-α and absolute amounts of PS-RBCs in the S group.

Values of median (range) of percentage of CD14+-expressing intracellular TNF-α before and after lipopolysaccharide (LPS) stimulation in each group of subjects are shown in Table II. Before stimulation, the S group had a significantly larger number of such cells than the others (P < 0.05). A significant increase in amounts of such cells after LPS stimulation was shown in all groups (P < 0.001). However, the increment in the S was less than those in the NS and NC groups (77 vs. 404, and 304 folds increment, respectively).

Table II. Percentage of CD14+-Expressing Intracellular TNF-α Before and After Lipopolysaccharide (LPS) Stimulation in Each Group of Subjects
Type of subjectsNPercentage of CD14+-expressing intracellular TNF-α, median (range)P-valuea
Before stimulationAfter LPS stimulation
  • a

    Wilcoxon signed rank test.

  • b

    Significant difference from normal controls and nonsplenectomized patients (P < 0.05) (Mann–Whitney U test).

  • c

    Significant difference from normal controls and nonsplenectomized patients (P < 0.01) (Mann–Whitney U test).

Splenectomized hemoglobin E/β-thalassemia200.45 (0.15–3.42)b34.65 (0.31–67.16)c<0.001
Nonsplenectomized hemoglobin E/β-thalassemia200.21 (0.02–3.50)84.93 (42.94–94.83)<0.001
Normal controls200.26 (0.02–1.23)78.93 (55.25–93.45)<0.001

The increased amount of circulating PS-RBCs in the S group (Table I) is in agreement with previous reports [5, 7] and likely due to the less-efficient function of other RE cells than splenic macrophages in clearing these pathologic cells. However, until we can do pre- and postsplenectomy studies on the same patient, there will always be an argument that the finding could also be due to a more severe disease in the splenectomized one. Findings of a nonstatistically significant difference between the amounts of PS-RBCs in the NS and NC groups (Table I) reinforce a recognition of the better performance by the spleen [4]. Taken together, these findings in the S and NS groups in the context of overlapping disease severity between them suggest that the final amount of circulating PS-RBCs is determined more by status of the spleen than the inherent RBC defects alone.

Monocytosis in the S group could be due to reactive production and/or inability to home to the spleen. However, it was not found in another study despite elevated serum macrophage colony-stimulating factor levels, which were unrelated to splenic status [8]. This discrepancy requires further studies.

Monocyte activation in the S group (Table I) was similar to previous reports in SCD patients [11, 12], who invariably have autosplenectomy, suggesting its close relationship with splenic absence in these two hereditary hemolytic anemia. The smaller amounts, but not statistically different, of activated monocytes in the NS compared with the NC groups (Table I) is quite interesting. It could be misleading because of a small number of patients causing skewness of data, or because of ancillary monocytes having limited roles in the presence of more efficiently functioning (in clearing defective RBCs) and abundant splenic macrophages in an enlarged hyperplastic spleen. Compared with the S group, the higher increment of CD14+-expressing intracellular TNF-α after an ex vivo LPS stimulation in the NS one (Table II) with reserved capacity would lend some support to the latter possibility. It also suggests that the impaired response of monocytes in splenectomized patients could be due to “overwork,” at least to help clear the now abundant circulating pathologic RBCs in the absence of splenic macrophages. Because chronic iron overload can also contribute to this impairment [15, 16], we searched for data on serum ferritin levels near the time of studies and found no statistically significant difference between the S and NS groups [median (range) = 2,599 (424–7,180) ng/mL and 1,150 (266–4,450) ng/mL, respectively. P = 0.136, Mann-Whitney U test].

Levels of serum TNF-α were significantly increased in the S compared with the other two groups (Table I). Their correlation with the amounts of activated monocytes and PS-RBCs were assessed by multiple linear regression analysis after controlling for types of subjects. The only correlation among the three variables in the three groups was between serum TNF-α levels and amounts of PS-RBCs in the S group (coefficient = 0.914, standard error = 0.188, P < 0.001). The results suggest that, after splenectomy, an abnormally large amount of PS-RBCs activated many types of RE cells, all of which can produce TNF-α [6]. Finding hepatosplenomegaly in Thal disease and chronic hemolytic anemia suggests the Kupffer cell being the prominent one.

Our findings of monocyte activation in splenectomized E/β-Thal patients help explain certain clinical features pertaining to this group. Monocyte activation together with elevated serum TNF-α levels could activate vascular endothelial cells, leading to an expression of adhesion molecules and tissue factor on their surfaces [11]. Together with increased amounts of thrombogenic PS-RBCs [5, 7] and activated platelets [17], this could further facilitate the formation of thrombotic pulmonary arteriopathy, the basis of pulmonary arterial hypertension, which is more prevalent after splenectomy [18]. The blunted inflammatory cytokine response to LPS, in keeping with previous reports of reduced monocyte phagocytic activity because of a “compensatory” increased erythrophagocytic activity in splenectomized β-Thal patients [19, 20], may affect host defense against infection, which is known to be more prevalent and severe among these patients [21–23].

In conclusion, our studies suggest role of spleen in controlling the amount of PS-RBCs and mononuclear phagocytic activity in E/β-Thal patients. Validation in a larger cohort and preferably on the same patients before and after splenectomy would strengthen our conclusion. In the meantime, splenectomy must be judiciously applied in E/β-Thal patients to avoid the undesirable consequences.

Methods

After an overnight fast, venous blood was drawn from the antecubital vein into a plastic syringe. Seven, 3, and 3 mL each of blood were then transferred into plain (Becton Dickinson Bioscience (BDB), Franklin Lakes, NJ), potassium ethylenediaminetetraacetic acid (EDTA)- (BDB), and sodium heparin- (Greiner Bio-One, Monroe, NC) containing vacutainer tubes, respectively. Blood anticoagulated with EDTA was used for determination of complete blood count, reticulocyte count, Hb typing, and monocyte surface expression of CD11b. Blood anticoagulated with sodium heparin was used for determination of monocytes expressing intracellular cytokine and PS-RBCs. Serum was used for the TNF-α assay.

Hematologic and serum TNF-α assay

Complete blood count and reticulocyte count were done by automated hematological analyzer (Technicon H*3; Bayer Diagnostics, Tarrytown, NY). Hb typing was done by high-performance liquid chromatography (Variant II; Bio-Rad Laboratories, Hercules, CA). Serum TNF-α assay was done by enzyme-linked immunosorbent assay kits (R&D Systems, Minneapolis, MN).

Flow cytometric measurement of monocyte activation. Monocyte surface expression of CD11b

Blood samples were kept at 4°C before processing. One hundred microliters of EDTA blood was lysed with 2 mL of fluorescence-activated cell sorting (FACS) lysing solution (BDB) for 10 min at 4°C. After washing, the remaining WBCs were incubated with 2 μL of allophycocyanin-conjugated anti-CD14 and 5 μL of phycoerythrin (PE)-conjugated anti-CD11b (BDB) for 20 min at 4°C in the dark. The stained cells were washed and resuspended in 350 μL of 1% paraformaldehyde. Monocyte activation was measured as mean fluorescence intensity of CD11b+ by BDB FACSCalibur flow cytometer (BDB). A total of 5,000 cells were collected.

Monocytes expressing intracellular cytokine

Five hundred microliters of sodium heparinized whole blood was incubated with 10 μL of Brefeldin A (BFA) (Sigma-Aldrich, St. Louis, MO) under 5% CO2 for 4 hr at 37°C with and without 1 μg/mL of LPS (Sigma-Aldrich). BFA prevents cytokine from leaking out of cytoplasm [24]. Two hundred microliters each of unstimulated and LPS-stimulated blood samples were lysed with 2 mL of FACS lysing solution for 10 min at room temperature. After centrifuging and decanting the supernatant, WBCs were washed and stained with 2 μL of allophycocyanin-conjugated anti-CD14 for 15 min. Samples were then washed and incubated with 500 μL of FACS permeabilizing solution (BDB) for 10 min. After washing, cells were incubated with 5 μL of fluorescein isothiocyanate (FITC)-conjugated anti-human TNF-α (BDB) for 30 min. The last three steps were done at room temperature in the dark. FITC-conjugated anti-mouse IgG1 (BDB) was used as a negative marker. The amount of stained cells was determined by BDB FACSCalibur flow cytometer. A total of 5,000 cells were collected.

Flow cytometric measurement of PS-RBCs

Two microliters of sodium heparinized whole blood was incubated with 2 μL of FITC-conjugated AV (BDB), 2 μL of PE-conjugated glycophorin A (DAKO, Hamburg, Germany), and 94 μL of 1× AV binding buffer (BDB) at room temperature for 15 min in the dark. FITC-conjugated anti-mouse IgG1 was used to set a negative control. After a similar incubation process, 300 μL of 1× AV binding buffer was added to both samples. The amount of stained cells was determined by BDB FACSCalibur flow cytometer. A total of 100,000 cells were collected.

Acquisition and data analysis were performed by CellQuest Software (BDB). Stained cells were excited with 488-nm light from a 15-mW argon ion laser. Logarithmic green and orange-red fluorescence of FITC for PS expression and PE for RBCs were measured through 530/30 and 585/42 band pass filters, respectively. The RBCs were gated on the basis of their logarithmic amplification of the forward scatter and 90° light scatter signals.

Statistical analysis

Data were described by means (standard deviation) or median (range) where appropriate. Multiple linear regression analysis was applied to determine the correlation between absolute amounts of CD14+/CD11b+- or CD14+-expressing intracellular TNF-α and those of AV+ RBCs, and levels of serum TNF-α, respectively in each group. Data were transformed to be log-scale where appropriate. Assumptions of linear regression, normality, and constant variances of residuals were checked. All analyses were performed using STATA 10.0 (Stata Corp., College Station, TX). P value less than 0.05 was considered to be statistically significant.

Author Contributions

V.A., P.B. (W.B.'s preceptor for her M.S. thesis), K.P., W.B., and S.C. designed the research; W.B. and N.A. performed the research; W.B. and A.T. performed statistical analyses; W.B., P.B., A.T., and V.A. wrote the article.

Wansa Banyatsuppasin*, Punnee Butthep*, Vichai Atichartakarn†, Ammarin Thakkinstian‡, Napaporn Archararit§, Kovit Pattanapanyasat¶, Suporn Chuncharunee†, * Department of Pathology, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, † Division of Hematology, Department of Medicine, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, ‡ Section for Clinical Epidemiology and Biostatistics, Research Center, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, § Laboratory Section, Research Center, Faculty of Medicine, Ramathibodi Hospital, Mahidol University, Bangkok, Thailand, ¶ Office for Research and Development, Faculty of Medicine, Siriraj Hospital, Mahidol University, Bangkok, Thailand.

Ancillary