Introduction

Antibodies to red blood cells (RBCs) may show various effects in vivo on circulating RBCs compared with their reactivity with RBCs in serological testing. In addition, their clinical significance often depends on the clinical condition of individual patients. Indeed, some patients may develop a severe haemolytic transfusion reaction (HTR) due to an incompatible RBC transfusion, whereas other patients may develop only mild or no reactions under identical serological findings [1, 2]. Therefore, the clinical significance of detectable antibodies cannot always be determined using standard serological testing.

During the last decades, numerous methods have been developed to measure or predict RBC survival. Currently, chromium-51 or biotin-labelled tests for survival measurement [3, 4] are most reliable; however, they cannot be used routinely. Alternative techniques, including the biological cross-match, antibody-dependent cellular cytotoxicity (ADCC) assay [5], chemiluminescence test and the monocyte monolayer assay (MMA) [6], are often predictive but do not invariably exclude HTRs. The MMA is the most widely used assay to evaluate the clinical significance of alloantibodies to RBCs. This test is based on the use of autologous or allogeneic monocytes and RBCs opsonized with recipient serum. It is yet not unanimously clear if the outcome of patients transfused with serologically incompatible RBCs demonstrates a sound correlation for all antibodies with results of the MMA. Since patients with alloantibodies potentially causing clinically relevant haemolysis are usually not transfused with incompatible blood, data correlating with results of the MMA with haemolysis parameters are difficult to obtain. A recent study [7] used a MMA to cross-match 61 RBC alloantibodies with RBC units. Thirty-one out of 61 patients with no or variable significant antibodies were transfused successfully with RBC units and negative MMA cross-match. Unlike macrophages, circulating monocytes are not known to exhibit erythrophagocytosis in alloantibody-mediated haemolytic anaemia [6]. Hence, it remained questionable whether the use of macrophages might more accurately reflect the in vivo outcome than the use of monocytes [6, 8, 9].

The human leukemic cell line THP-1 has several advantages over human peripheral monocytes and is commonly used for investigating the function and regulation of monocytes and macrophages. THP-1 cells express FcRI and FcRII receptors [10]. THP-1 monocytes can convert to macrophage-mimicking cells in the presence of phorbol 12-myristate 13-acetate (PMA) [11, 12]. With increasing cell adherence, FcRI and FcRII receptor expression is reduced [10]. Interestingly, phagocytosis of IgG-coated sheep RBCs increases much stronger than that of uncoated RBCs (64 % vs. 35 %) [12]. In a previous study that applied the fluorometric quantitative erythrophagocytosis assay using human THP-1 cells and PKH26-labelled RBCs, high sensitivity and good reproducibility were demonstrated [13]. We modified this assay by using PMA-treated THP-1 cells and PKH26-labelled RBCs in a shaking assay. Thus, we used adherent macrophage-like cells instead of monocytes, and gentle shaking may favour antibody-mediated interaction between macrophage-like cells and RBCs. Accordingly, a trypsinization step was needed. All other steps and analysis of the data were done as described by Healey et al. [13]. The test was initially used to assess the clinical relevance of several known alloantibodies. To demonstrate a correlation between the assay and antibody-dependent haemolysis, RBCs from three groups of patients were analysed as follows: patients with clinically significant AIHA, patients with AIHA in remission and pregnant women with detectable autoantibodies to RBCs.

Materials and methods Patients

The research was approved by the local ethics committee (No EA2/058/12). Sixteen patients with AIHA of warm type were included. Routine haemolysis parameters (haemoglobin, LDH, haptoglobin, reticulocytes) were determined. In addition, five pregnant women with detectable autoantibodies without haemolysis were investigated. EDTA blood samples for the control group were obtained from healthy blood donors.

Serological testing

Serological testing, including antibody screening and the monospecific direct antiglobulin test (DAT), was performed using standard gel technology as described elsewhere [14, 15]. Acid eluates were performed according to the manufacturer's recommendation (BAG Health Care GmbH, Lich, Germany).

THP-1 cells culture

THP-1 human monocytic leukaemia cells (TIB-202, American Type Culture Collection (ATCC), Manassas, VA, USA) were cultured at a density of 1–5 × 105 cells/ml in RPMI-1640 medium (GIBCO, Grand Island, NY, USA) containing 10% fetal bovine serum (FCS; GIBCO, Invitrogen, Carlsbad, California, USA), and 1% penicillin 100 U/ml and 1% streptomycin 100 µg/ml (GIBCO Life Technologies, Carlsbad, California, USA). Cells were maintained at 37°C in a humidified atmosphere with 5% CO2. Cell suspensions were pelleted at 900 g for 2 min, resuspended and transferred at a density of 1 × 105 cells/ml in refreshed medium twice weekly.

Differentiation of THP-1 cells to macrophage-mimicking cells

THP-1 cells were resuspended at a density of 1 × 106 cells/ml in fresh growth medium and PMA (Calbiochem, Darmstadt, Germany, hazardous substance, please acknowledge the safety data sheet) to a final concentration of 10−8 M [16]. Five hundred microlitres of the suspension was transferred into each well of 6-well culture plates (Thermo Fisher Scientific, Waltham, MA, USA) and incubated at 37°C in a humidified atmosphere with 5% CO2. Differentiation medium with 10-8 m PMA was refreshed after 24 h. Following 48 h, the medium was replaced with fresh medium without PMA and incubated for a further 3 h. The adherent, macrophage-like THP-1 cells were almost confluent before use in the phagocytosis assay.

Verification of THP-1 cell differentiation by flow cytometry

After a 10-min trypsinization period (GibcoTM TrypLE Express), PMA-treated THP-1 cells were resuspended in cell culture medium. PMA-treated and non-treated THP-1 cells (0·5 × 106) were diluted in 200 μl PBS and stained with 10 µl of the fluorescent labelled anti-human CD11b / MAC-1 (FITC) or anti-human HLA-DR (APC) antibodies (both from BD Biosciences Pharmingen TM). Cells were subsequently analysed by flow cytometry [16] with MACSQuant® Flow Cytometer (Miltenyi Biotech, Bergisch Gladbac, Germany).

RBC preparation and labelling by PKH26 and the phagocytosis assay

RBCs from patients with AIHA, pregnant women with detectable autoantibodies and from healthy donors were washed three times with saline (0·9% NaCl) and pelleted at 900 g for 2 min before PKH26 staining. In contrast, patients’ plasma containing alloantibodies (anti-D, anti-K, anti-Lu(b), anti-Yt(a), anti-Ch (200µl)) were incubated with 50 µl pelleted donor RBCs (expressing the correspondent antigen) for 2 h at 37°C and then washed three times with saline (0·9% NaCl). The alloantibody-coated RBCs were further tested in an identical way as patients’ and healthy donors’ RBC, respectively.

All RBCs were membrane-stained using PKH26 according to the manufacturer's recommendations (Sigma-Aldrich, St. Louis, MO, USA). Briefly, 10 µl of RBCs from pellet was diluted in 125 µl diluent C (supplied with the PKH26 kit), added to 0·5 µl PKH26 in 125 µl diluent C, followed by immediate mixing and incubation at room temperature for 4 min. The reaction was terminated by the addition of 100 μl of human AB plasma. After 1 min, 2 ml of cell culture medium was added. Cells were pelleted at 1700 g for 10 min, resuspended in 100 µl of cell culture medium and added to PMA-differentiated THP-1 cells in a 6-well plate. Subsequently, plates were incubated at 37°C for 45 min under soft continuous agitation.

Flow cytometry

After incubation at 37°C for 45 min, non-phagocytic RBCs were removed by washing three times with 2 ml of ice-cold isotonic saline (0·9% NaCl). Residual RBCs were haemolysed with ice-cold hypotonic saline (0·2% NaCl) for 2 min, followed by the addition of 2 ml of ice-cold hypertonic saline (1·6% NaCl) for 2 min to restore isotonicity [13, 17]. Subsequently, the supernatant was removed and the adherent macrophage cells were detached by trypsin. Before and after trypsinization, THP cells were checked microscopically for remaining RBCs or detritus THP cells were diluted with 500 µl PBS (1 x 106 cells/ml) and kept on ice until analysis by flow cytometry (MACSQuant® Flow Cytometer, Miltenyi Biotech). At least 10 000 events were collected for each sample. Data were analysed using the FlowJo® software (FlowJo LLC, Ashland, OR, USA). The mean fluorescence and percentage of cells that were strongly positively stained for PKH26-red-labelled RBCs compared with control group were calculated. Controls using RBCs from two or three healthy blood donors with no selection of age, sex or blood group were included in each experiment.

Analysis and statistics

Clinical relevance of allo- or autoantibodies was determined by comparing patients’ results (MFI or % phagocytosis) with arithmetic means ± standard deviation (SD) of the same day control group. As a cut-off, antibodies generating results (MFI or % phagocytosis) higher than arithmetic means + 3 SD of controls were considered clinically relevant.

For statistical analysis of patient groups and control groups, data were calculated as arithmetic means ± standard error of the mean (SEM). Statistical analysis was performed using Microsoft Excel and SSPS. Significance between two groups was determined using Mann–Whitney U-test.

Results Serological and clinical data

In total, RBCs from 16 patients with long-term AIHA were studied. All patients had a positive eluate (Table 1). Fourteen of 16 patients showed IgG-positive DAT and 11 patients C3d-positive DAT. One patient predominantly had an IgM-wAIHA (Pat. W7). Another patient (Pat. W12) had both warm and cold autoantibodies (mixed-type IgM-AIHA). Based on haemoglobin, reticulocyte, lactate dehydrogenase and haptoglobin levels, as well as the clinical picture, 12 patients showed clear signs of haemolysis. All five pregnant women had an IgG-positive DAT and detectable autoantibodies in the eluate (Table 1). There were no signs of haemolysis in the pregnant cohort [18].

Table 1. Most relevant data of patients with AIHA Patient No. Age DAT Eluate Hb LDH Hp RETR Medication (years) Sex IgG C3d IgM IgA (g/dl) (U/l) (mg/dl) 10E3/μl W1 8 M 4+ – (+) 1+ P 10∙9 281 nd nd Rituximab / Prednisolone W2 64 F 4+ 2+ – – P 13∙4 583 <5∙8 nd Prednisolone 1 mg/d W3 49 F 4+ 4+ – – P 10 443 <5∙8 140 Cyclosporin A 300 mg /d, Prednisolone 10 mg/d, Darbepoetin alfa W4 82 F 3+ 3+ – – P 11∙4 348 122 50 Prednisolone 5 mg/d, Darbepoetin alfa W5 48 F 4+ (+) – 2+ P 8∙4 1326 <5∙8 454 Prednisolone 30 mg/d, Cyclophosphamide 150 mg/d W6 60 F 4+ 4+ – – P 13∙0 483 <5∙8 nd Prednisolone 5 mg/d W7 36 M (+) 4+ 4+ (+) P 12∙4 281 <5∙8 169 Prednisolone 75 mg/d W8 43 M 4+ 3+ – – P 10∙4 460 <5∙8 112 Prednisolone 10 mg/d, Azathioprine 200 mg/d W9 68 F 4+ 3+ 2+ – P 8∙6 1214 nd 419 Mycophenolic acid; 2000 mg/d, Prednisolone 10 mg W10 50 M 4+ – – – P 12∙4 397 <5∙8 nd Prednisolone 7∙5 mg/d, Cyclophosphamide 150 mg/d W11 27 M 3+ (+) – – P 12∙9 252 <5∙8 nd – W12 51 F – 4+ – – P 12∙5 314 nd nd – W13 55 F 4+ 1+ – (+) P 11∙5 227 116 nd Prednisolone 10 mg/d, Cyclophosphamide 150 mg/d W14 73 F 1+ – – – P 11∙3 234 149 101 Cyclophosphamide 25 mg/d W15 27 F 4+ – – – P 14∙8 235 95∙6 79 Prednisolone 10 mg/d, Azathioprine 200 mg/d W16 56 M – – – – P 13∙9 242 110 81 Prednisolone 2∙5 mg/d G1 36 F 4+ – (+) – P G2 36 F 4+ 2+ – – P G3 29 F 3+ – – – P G4 32 F 1+ – – – P G5 40 F 1+ – – – P DAT, direct antiglobulin test; G, pregnancy with autoantibodies of warm type; Hb, haemoglobin; Hp, haptoglobin; LDH, lactate dehydrogenase; nd, no data; P, positive; RETR, reticulocytes; W, patient with AIHA of warm type. PMA-treated THP-1 cells transform into macrophage-like cells

Phorbol 12-myristate 13-acetate-differentiated THP cells were adherent, demonstrating increasing CD11b expression, as well as a downregulation of HLA-DR (Fig. 1a). The strongest changes in CD11b/HLA-DR expression were observed from 48 to 72 h. Therefore, THP cells were used after 48 h of differentiation in further experiments. Scatter plots of THP-1 cells and differentiated THP-1 cells (macrophage-like cells; blank control) were similar and showed only background fluorescence in channel B2 B585 (Fig. 1b) and a very low number of cells in region P3 (<0·1 %). After the phagocytosis assay, trypsinated THP-1 cells were checked for remaining RBCs by the use of Hayem’s solution and the improved Neubauer haemocytometer showing an almost complete haemolysis of non-phagocyted RBCs (data not shown). Figure 1c shows an overlay of the histograms of PMA-treated THP-1 cells and PKH26-stained RBCs from healthy donors. After incubation with PKH26-stained RBCs, PMA-treated THP-1 cells showed increased fluorescence from ingested PKH26-stained RBCs (control group; Fig. 1d). This reflects already published results showing that non-opsonized RBCs were also digested in larger amounts [13, 17].

Characterization of PMA-treated THP-1 cells. (a) Effect of PMA differentiation on CD11b (left) and HLA-DR (right) expression in THP-1 cells at different times (0 and 48 h). (b) Scatter plot of PMA-treated THP-1 cells (blank control). Region P3 has been selected as contains ≤ 0·1% of events. (c) Histogram overlay of PMA-treated THP-1 cells (black) and PKH26-stained RBCs (blue). (d) Histogram of PMA-treated THP-1 cells after phagocytosis of PKH26 stained RBCs from healthy blood donors (control group).

RBCs coated with alloantibodies

RBCs loaded with significant (anti-D, anti-K) or potentially significant alloantibodies for haemolysis (anti-Lu(b), anti-Yt(a)) were observed to further enhance phagocytosis by PMA-treated THP-1 cells (Fig. 2a–j, m). However, two of three anti-Yt(a) did not enhance phagocytosis significantly by more than mean ± 3 SD of the control group, which corresponds to the clinical experience with this antibody. In contrast, pre-incubation of RBCs with the non-haemolytic alloantibody anti-Ch did not affect phagocytosis (Fig. 2k,l,m).

Erythrophagocytosis of RBCs coated with alloantibodies. (a-l) Scatter plot (a) and histogram (b) of PMA-treated THP-1 cells after phagocytosis of RBCs from a healthy blood donor (control group). Scatter plots (c, e, g, i, k) and overlay of histograms with a control group (d, f, h, j, l) of PMA-treated THP-1 cells after phagocytosis of alloantibody-coated RBCs [anti-D, anti-K, anti-Lu(b), anti-Yt(a) and anti-Ch (blue).]. (m) Results of THP-1 phagocytosis of RBCs pre-incubated with alloantibodies in comparison with control group RBCs.

Enhanced phagocytosis of RBCs from patients with active wAIHA

A typical result obtained from patients with significant wAIHA is depicted in Fig. 3a–c using the example of patient no. W9 (Table 2). Erythrophagocytosis was significantly increased (> average mean of control + 3 SD) in patients with active AIHA compared with healthy donors (control group). This significant increase in the mean fluorescence of all THP-1 cells and percentage of strongly positive stained cells in region P3 indicates a strong avidity of macrophages to patient RBCs. In comparison, erythrophagocytosis of RBCs from a pregnant woman (G4, Table 2) was not increased (Fig. 3d–f). Patients with clinically relevant AIHA of warm type showed a highly significant enhanced erythrophagocytosis as shown by mean fluorescence (Fig. 3g). In contrast, a similar erythrophagocytosis was observed in patients with AIHA of warm type in remission or patients with autoantibodies due to pregnancy and healthy blood donors, respectively (Fig. 3g). Further analysis did not demonstrate a correlation with the amount of bound IgG or C3d (Tables 1 and 2).

Phagocytosis of RBCs from a patient with AIHA and a pregnant woman with warm autoantibodies. Scatter plots and overlay histogram of PMA-treated THP-1 cells after phagocytosis of RBCs from a healthy blood donor (a, c: red line) and of RBCs from a patient with clinically relevant AIHA (b, c: blue line). Scatter plots and overlay histogram of PMA-treated THP-1 cells after phagocytosis of RBCs from a healthy blood donor (d, f: red line) and of RBCs from a pregnant woman with warm autoantibodies (e, f: blue line). (g) Comparison of THP-1 phagocytosis of RBCs from healthy blood donors (CG, first column), patients with significant wAIHA (second column), patients with wAIHA in complete remission (third column) and pregnant patients with autoantibodies (fourth column). Each point indicates one THP-1 sample after incubation with the respective RBCs. The grey lines represent the mean of each group. The mean fluorescence for the CG was normalized at 1. ***P < 0·001 significant difference from control group.

Table 2. In vitro erythrophagocytosis by PMA-treated THP-1 cells Patient no

Haemolysisa

Yes/No

MFI % Phagocytosis Patient CG (mean ± SD) Patient CG (mean ± SD) W1 Yes 3271 2119 ± 195 37 19∙5 ± 3∙5 W2 Yes 3943 2119 ± 195 42 19∙5 ± 3∙5 W3 Yes 2745 1738 ± 51 28 14∙3 ± 0∙6 W4 Yes 3540 2148 ± 188 33 16∙3 ± 1∙5 W5 Yes 5908 3431 ± 169 60 33∙7 ± 5∙9 W6 Yes 4301 2723 ± 44 34 20∙5 ± 1∙2 W7 Yes 6774 4140 ± 585 63 42∙0 ± 4∙2 W8 Yes 5096 3431 ± 169 56 33∙7 ± 5∙9 W9 Yes 12032 4247 ± 596 82 45∙7 ± 7∙5 W10 Yes 3513 2700 ± 186 41 27∙3 ± 3∙2 W11 Yes 1912 1076 ± 192 41 20∙3 ± 6∙7 W12 Yes 4189 1076 ± 192 70 20∙3 ± 6∙7 W13 No 2507 1076 ± 192