Morphological and flow-cytometric analysis of haemin-induced human neutrophil activation: implications for transfusion-related acute lung injury

Abstract

Background

Transfusion-related acute lung injury (TRALI) is associated with vascular endothelial cell injury following neutrophil activation. Recently, it has been suggested that haem-related molecules induce activation of neutrophils and that erythrocyte-derived substances contained in blood preparations are involved in TRALI. We observed the morphological effects and reactive oxygen species (ROS) production of haem-related molecules and investigated the effects of signal transduction inhibitors on haem-induced neutrophil activation.

Materials and methods

The polymorphonuclear cell fraction was isolated and stimulated using a control stimulant, PMA or fMLP, or by haem-related molecules, haemin, ferric citrate, or protoporphyrin IX. After stimulation, neutrophil was analysed using electron microscopy, a flowcytometer (FCM) and confocal laser scanning microscope to determine the fluorescent intensity of aminophenyl fluorescein (to detect ROS).

Results

In FCM analysis, haemin and protoporphyrin IX, both of which have a porphyrin ring, induced ROS production in neutrophils. Ferric citrate, which has no porphyrin ring, did not induce neutrophil activation. Haemin alone induced ROS production at relatively high concentrations, whereas low-level fMLP acted as an agonist in the presence of low concentrations of haemin. Haem-related molecules induced ROS production in neutrophil granules through signal transduction in a way similar to PMA. However, in electron microscopy studies, haemin stimulated neutrophils showed minute process at their surface and did not show the vacuolation observable following stimulation with PMA or fMLP.

Discussion

We suggest that low concentrations of haem-related molecules with porphyrin rings in the presence of other stimulating agent are important for ROS production and possibly the onset of TRALI. The ROS production induced by these molecules is dependent on a signal transduction pathway in a way similar to PMA.

Introduction

Transfusion-related acute lung injury (TRALI) is the leading cause of transfusion-related death^1,2. Most episodes of TRALI develop within 6 hours after transfusion and are manifested by dyspnoea, fever, hypoxia and bilateral non-cardiogenic pulmonary oedema^3,4. According to a US report, the frequency of TRALI is 1/55,000 to 1/260,000 transfusions, with a mortality rate of 6% to 20%^4,5. The pathogenesis of TRALI is summarised by a two-event model or threshold theory and is related to the patient’s condition^5–9. Important factors are neutrophil and pulmonary endothelial cell activation induced by transfusion, resulting in increased vascular permeability via the production of reactive oxygen species (ROS) and other molecules⁴. The production of ROS, including superoxides, in neutrophils is increased through the activation of NADPH oxidase^10–12.

Implicated factors include anti-leucocyte antibodies such as anti-human leucocyte antigen (HLA) class I and class II, anti-neutrophil antibodies^{1,2,9,13–15}, and soluble biological molecules such as soluble CD40 ligand¹⁶, lipids^6–8 and cytokines17,18. However, there are many reports regarding TRALI induction by red blood cell (RBC) components containing little plasma^19,20 and acute respiratory distress syndrome is partially related to oxidative stress induced by abnormal iron metabolism^21,22. Increased haemolysis occurs in RBC components stored for long periods^23–25. A rat lung perfusion model indicated that older RBC products increase the risk of TRALI^8,26. Furthermore, respiratory failure after cardiac surgery has been reported to be more frequent in patients transfused with older blood²⁷. We and Graca-Souza et al. have shown that haem-related molecules may activate human neutrophils^28–30. It is possible that haem-related molecules derived from haemolysis in transfusion bags and in patients’ blood vessels may be a trigger of TRALI. However, there are few reports regarding the mechanism of neutrophil activation by haem-related molecules.

Here, we report the characteristics of haemin-induced neutrophil activation determined using ultrastructural analysis and flow cytometric detection of ROS. The effects of signal transduction inhibitors affecting neutrophil function on haemin-induced neutrophil activation were investigated and the pathogenic significance of haem-related molecules in TRALI are discussed.

Materials and methods

Cell preparation

Heparinised peripheral blood was collected from healthy volunteers after obtaining their written informed consent. Neutrophils were separated at room temperature using a density gradient method in a Mono-Poly separation medium (Dainippon-Sumitomo Pharma Biomedical, Osaka, Japan) according to the manufacturer’s instructions; purity was greater than 90%. Neutrophils were suspended in phosphate-buffered saline supplemented with 0.1% bovine serum albumin (Sigma-Aldrich, St. Louis, United States of America), and kept at 4 °C until use.

Pre-treatment for the detection of reactive oxygen species

For pre-treatment of collected cells, 1 μL of anti-CD16b conjugated to phycoerythrin (Becton Dickinson and Company, [BD], Franklin Lakes, USA) or control antibody was added to 100 μL of 2×10⁶/mL of cells. After incubation for 10 minutes, 2 μL of aminophenyl fluorescein (APF; Sekisui Medical, Tokyo, Japan) were added and incubated for 30 minutes at room temperature³¹. Stimulation with agonists was performed for 10 minutes at room temperature without agitation.

Stimulators

Neutrophils were stimulated using 0.32 nM of phorbol myristate acetate (PMA; Sigma-Aldrich), or 25 or 50 nM of N-formyl methionyl leucyl phenylalanine (fMLP; Sigma-Aldrich) as positive controls. Dimethyl sulphoxide (DMSO) was used to make solutions of 7.6 or 76 μM of Fe³⁺ (ferri)-protoporphyrin IX (haemin; molecular weight 651.95; Sigma-Aldrich), 760 nM of protoporphyrin IX (PPIX; molecular weight 626.03; Sigma-Aldrich), and 40 μM of ferric citrate (molecular weight 244.94; Sigma-Aldrich). DMSO alone had no effect on neutrophil ROS production.

Haemoglobin (molecular weight 64,500) derived from erythrocytes has four haem groups, each of which contains a porphyrin ring and an iron atom. Therefore, 7.6–76 μM haemin corresponds to 1.9–19 μM (120–1,200 mg/L) haemoglobin and to 0.10–1.02% haemolysis.

Flow cytometric analysis

Flow cytometric analysis was conducted using a FACSCalibur (BD) to determine the mean APF fluorescent intensity (geometric mean) in CD16b-positive cells.

Confocal laser scanning microscopy

Neutrophils were applied to a glass-bottom dish coated with poly-L-lysine (Sigma-Aldrich). The adhered cells were then observed using confocal laser scanning microscopes IX81 (Olympus, Tokyo, Japan) and CSU-X1 (Yokogawa Electric, Tokyo, Japan) and electron microscopy (Hamamatsu Photonics, Shizuoka, Japan).

Electron microscopy

Neutrophils were fixed with 1% glutalaldehyde for 16 hours at 4 °C.

For transmission electron microscopy the fixed neutrophils were spun down onto a silane-coated glass slide using a Cytospin (ThermoFisher Scientific, Waltham, United States of America) for 10 minutes at 120 g. After fixation with OsO₄ for 30 minutes at 4 °C, samples were dehydrated using a graded ethanol series, and invert-embedded in Quetol 812 (Nisshin EM, Tokyo, Japan). Ultrathin sections of 80 to 100 nm were prepared using an ultramicrotome (Leica Microsystems, Wetzlar, Germany), and the sections were observed using an H-7500 (Hitachi High-Technologies, Tokyo, Japan).

For scanning electron microscopy the samples were attached to poly-L-lysine-coated glass slides. After fixation and dehydration using the same method as for transmission electron microscopy, ethanol was replaced with t-butylalcohol (Wako Pure Chemicals, Osaka, Japan). These samples were dried using a freeze-drier (Hitachi High-Technologies) and observed with a JSM-7500F (Nippon-Denshi, Osaka, Japan) without decoration.

Inhibitors

NADPH oxidase inhibitor (apocynin, Sigma-Aldrich), mitogen-activated protein kinase (MAPK) p38 pathway inhibitor (SB203580; Calbiochem, San Diego, CA, USA), extracellular signal-related kinase (ERK) inhibitor (PD98059; Calbiochem), and phosphatidylinositol 3-kinase (PI3K) inhibitor (wortmannin; Sigma-Aldrich) were dissolved in DMSO and used at final concentrations of 300 μM, 20 μM, 20 μM, and 100 nM, respectively. These inhibitors were applied for 10 minutes at room temperature, followed by APF treatment and stimulation with neutrophil agonists.

Statistical analysis

The statistical analysis was performed with the JMP-7 package (SAS Institute Inc, Cary, NC, United States of America). The t-test was employed for comparisons of two groups. A P-value <0.05 was considered statistically significant.

Results

Stimulator-induced production of reactive oxygen species by neutrophils

Isolated neutrophils were treated with the APF probe to show ROS formation and any increased production following stimulation. As demonstrated by flow cytometry scatter plots of APF fluorescence intensity (Figure 1A), neutrophils stimulated by PMA, fMLP, or haemin showed higher fluorescence than control cells (unstimulated neutrophils pre-treated or not with APF). The effects of the various stimulators on mean APF fluorescent intensity in neutrophils are shown in Figure 1B. The mean APF fluorescent intensity in neutrophils stimulated with 0.32 nM PMA was significantly different from that in control neutrophils; the effect of PMA was concentration-dependent (data not shown). fMLP (25 nM or 50 nM) also induced significant differences in APF fluorescent intensity from that in control, unstimulated neutrophils, although this effect was not concentration-dependent. Haemin induced ROS formation at a relatively high concentration of 76 μM. The effect of haemin was concentration-dependent, and treatment with 7.6 μM of haemin did not induce significantly different ROS production from that in control cells. PPIX induced ROS formation in neutrophils in a concentration-dependent manner, whereas ferric citrate, which does not have a porphyrin ring, did not stimulate ROS formation. DMSO alone did not affect ROS production.

Figure 1.

ROS formation, determined by an APF probe, in isolated neutrophils treated with stimulators.

(A) FCM scatter plots of APF fluorescence in isolated neutrophils. Forward scatter is represented on the horizontal axis and fluorescence intensity on the vertical axis. Grey spots are CD16b-positive neutrophils. Control (unstimulated) neutrophils not pretreated with APF showed very low APF fluorescence (far left), while control (unstimulated) neutrophils pretreated with APF showed some APF fluorescence (middle left). Neutrophils stimulated by PMA, fMLP, or haemin showed higher APF fluorescence than control cells (middle, middle right, and far right, respectively).

(B) The mean APF fluorescent intensity values of CD16b-positive neutrophils after stimulation with PMA, fMLP, or haemin are indicated and compared with the values of control neutrophils. n=4 (mean±SE). *P <0.05.

Morphology

Confocal laser scanning microscopy of neutrophils was performed after the cells had been stimulated with PMA, fMLP, haemin or PPIX (Figure 2). These stimulators produced granular APF-induced fluorescence in the cytoplasm with the highest intensity being produced following PMA treatment.

Figure 2.

Confocal laser scanning microscopy of ROS formation in isolated neutrophils pre-treated with an APF probe.

Isolated neutrophils were stimulated using PMA (middle-left), fMLP (middle-right), or haemin (right) after APF treatment and observed using confocal laser scanning microscopy. Top panels are bright fields (BF). Bottom panels are APF intensity (APF). Bar=2 μm.

We also used transmission electron microscopy (Figure 3, top) and scanning electron microscopy (Figure 3, bottom) to observe neutrophils after stimulation. Both PMA and fMLP induced cytoplasmic granular polarisation and vacuolation, as reported previously (Figure 3)^32–37. Although haemin, PPIX, and ferric citrate induced granular polarisation, vacuolation was less apparent and vacuoles were smaller than those induced by PMA or fMLP (Figure 3). Haemin and PPIX characteristically induced the formation of minute processes from the cell membrane (Figure 3); this feature was not observed for any of the other stimulators tested.

Figure 3.

Electron microscopy of neutrophils after stimulation. Isolated neutrophils were unstimulated (left) or stimulated using 0.32 nM of PMA (second from left), 50 nM of fMLP (third from left), 76 μM of haemin (third from right), 760 nM of PPIX (second from right), or 40 μM of ferric citrate (right) and observed using transmission electron microscopy (TEM, top panels) and scanning electron microscopy (SEM, bottom panels). Bar=2 μm.

The effects of simultaneous addition of fMLP and haemin

We evaluated the effects of combination treatment with haem-related molecules and fMLP and examined the biological significance of haemin-induced neutrophil activation. Figure 4 shows the effects of the combination of haemin and low-concentration fMLP treatment on ROS production. Although the mean APF intensity induced by 7.6 μM haemin alone was only 115.26±7.22 (Figure 4, black small circle), following combination treatment with 25 nM fMLP it was 198.92±21.22% (Figure 4, black square). Table I shows the values of the mean APF intensity induced by each of the compounds alone and in combination. The combination effect was observed with lower concentrations of hemin and was not dependent on fMLP concentrations between 5 and 25 nM.

Figure 4.

Effects of the combination of haemin and fMLP. Isolated neutrophils were stimulated using combination treatment with haemin and fMLP after APF treatment and analysed using flow cytometry. n=4 (mean±SE).

Table I.

Effects of combined treatment with haemin and fMLP.

The mean APF fluorescent intensity values of CD16b-positive neutrophils stimulated with various combinations of haemin and fMLP are shown and compared with the values produced by haemin alone. n=4 (mean±SE).

fMLP (nM)	Haemin (μM)
	0	3.8	7.6	15.2	38.4
0	132.83±9.07	123.68±6.76	115.26±7.22	115.15±6.02	129.26±15.43
5	166.04±20.71	198.13±23.04**	191.42±16.04**	229.68±44.52**	246.49±63.67*
10	164.12±14.56	162.92±10.10**	171.23±9.95**	170.64±12.67**	191.95±24.89*
25	183.13±15.12*	208.46±18.23**	198.92±21.22**	218.99±47.74*	224.08±42.34*

^*:P <0.05;

^**:P <0.01.

Effects of inhibitors of neutrophil function on haemin-induced production of reactive oxygen species

To investigate signal transduction pathways, the effects of NADPH oxidase inhibitors, MAPK inhibitors or a PI3K inhibitor were examined (Figure 5). Haemin-induced ROS was inhibited by NADPH, p38MAPK, ERK, and PI3K inhibitors, in a similar way to the inhibition of PMA-induced ROS production.

Figure 5.

Effects of signal transduction inhibitors on ROS production. Isolated neutrophils were incubated with signal transduction inhibitors before stimulation and analysed using flow cytometry. Controls cells were incubated with DMSO before stimulation. The final concentrations of signal transduction inhibitors were 300 μM apocynin (NADPH oxidase inihibitor), 20 μM SB203580 (p38MAPK inhibitor), 20 μM PD98059 (ERK inhibitor) and 100 nM wortmannin (PI3K inhibitor). n=4 (mean±SE),*P<0.05 versus each control.

Discussion

One of the most important pathogenic factors resulting in acute lung injuries such as TRALI is neutrophil activation⁴. In cases of TRALI, several clinical and experimental studies have shown that anti-leucocyte antibodies are major inducers of neutrophil activation^{1,2,9,13–15}. However, since there are many reports regarding TRALI induced by RBC components containing few anti-leucocyte antibodies^19,20), haem-related molecules derived from haemolysis in transfusion bags and in patients’ blood vessels may trigger TRALI. Haem-related molecules are present not only in RBC but also in myocytes in the form of myoglobin, and haem-containing enzymes (peroxidases, cytochromes, catalase, etc.)⁵. The reported effectiveness of universal leucoreduction for the prevention of TRALI might be related to lower levels of haem-containing proteins in blood products³⁸. Furthermore, haemin can be produced and is biologically functional after massive tissue damage or major surgical operations⁵. It is, therefore, important to investigate the effects of haem-related molecules such as haemin on the pathogenesis of TRALI and neutrophil activation. Graca-Souza et al. have already reported that haemin exerts a chemotactic effect on human neutrophils, and we have previously shown, using chemiluminescence, that haemin induces ROS production in human neutrophils^28–30.

In this study, we observed neutrophil activation using three haem-related molecules, haemin, PPIX, and ferric citrate (Figure 1). We demonstrated that PPIX, which contains no iron in the porphyrin ring, induced ROS production, but ferric citrate, which contains bare iron, did not induce ROS production. These results suggest the importance of the porphyrin ring in ROS production. The importance of the porphyrin ring in activating neutrophils was also observed by Porto et al.²⁹.

TRALI or TRALI-like events may also occur even if anti-leucocyte or HLA antibodies are not detectable in the blood product³⁹; other factors, such as patients’ preconditions or molecular factors must, therefore, also be investigated⁵.

We examined ROS production following combination treatment with low concentrations of fMLP and 3.8 –38.4 μM haemin to investigate the effect of multiple factors, and the additive effect of these stimulators was proven (Figure 4). A very low concentration of fMLP (5 nM) increased the ROS production stimulated by haemin, and this increase was not concentration-dependent. These results suggest that combination treatment with a low concentration of factors is sufficient to increase the ROS production of haem-related molecules.

A concentration of 3.8–38.4 μM haemin is achievable in several conditions. The haemin concentration in plasma and blood products has not yet been reported. However, free haemoglobin, especially alpha-haemoglobin is structurally unstable and tends to denature^40,41. Haemoglobin can be converted to methaemoglobin by oxidative stress and subsequently into haemin⁴². Furthermore, it has been reported that haemoglobin can be denatured in the presence of low concentrations of haemin⁴³. Thus, haemin is readily released into the plasma. As haemolysis in the transfusion bags of RBC concentrates is reported to be a maximum of 0.8%⁴⁴, which is equivalent to 20 μM of haemoglobin, this corresponds to 80 μM of haemin. Haem-related molecules derived from leucocyte enzymes are also expected to be present and the plasma concentration of haem-related molecules can be increased to pathological levels during massive transfusions^5,38. Alternatively, a concentration of 20 μM haemoglobin can be observed during haemolytic disorders⁴⁵. Moreover, aged blood products are reported to suffer from oxidative injury^25,26, and have been discussed as one of the inducers of TRALI²⁶. In an animal model, plasma from outdate blood easily caused acute lung injury⁸, and Koch et al. reported a higher incidence of respiratory failure after cardiac surgery in the patients transfused with older blood²⁷. Patients requiring red blood cells are shown to be under strong oxidative stress^46–48. Severe inflammation is also known to be related to NF-kappa B or tumour necrosis factor activation and oxidative stress^49,50. Thus, significant amounts of haemin can be expected to be present in patients or blood products under certain conditions.

Confocal laser scanning microscopy revealed that haemin induced the production of ROS in granules (Figure 2), and flow cytometry analysis revealed that ROS production depends on the levels of PI3K, p38, and NADPH oxidase (Figure 5). These results are similar to those of studies using PMA, indicating that haemin activates ROS production through a physiological signal transduction pathway and not through its toxic effects. However, morphological analysis by electron microscopy revealed that activation by haemin is different from induction by PMA or fMLP, suggesting different mechanisms of ROS production. In particular, the formation of minute processes from the cell membrane is rare in cases subjected to PMA or fMLP stimulation (Figure 3). Although haemin is considered to penetrate through the cell membrane because of its hydrophobic character⁴², protoporphyrin ring transporter, HCP1, is reportedly present and functions in duodenal cells⁵¹, haematopoietic stem cells⁵², astrocytes⁵³ and so on. Thus, the role of HCP1 in neutrophils warrants further exploration.

In conclusion, haemin, alone or in combination with fMLP, can induce neutrophil activation, resulting in ROS production and morphological changes. Haemin induces neutrophil activation through a signal transduction system similar to that used for PMA. The porphyrin ring of haemin was important for these reactions. Haem-related molecules, which are derived not only from haemolysis but also from leucocyte enzymes in transfusion bags or in patients’ blood vessels, may induce neutrophil activation and be related to TRALI in cases in which multiple factors are present. The biological significance of haem-related molecules in the pathogenesis of acute lung injuries, including TRALI, requires further investigation.