Article Text

This article has a correction. Please see:

Download PDFPDF

Apolipoprotein H, a new mediator in the inflammatory changes ensuing in jeopardised human myocardium
  1. H W M Niessen1,
  2. W K Lagrand2,
  3. H J A M Rensink3,
  4. Ch J L M Meijer1,
  5. L Aarden3,
  6. C E Hack3
  1. 1Department of Pathology, Free University Hospital, PO Box 7057, De Boelelaaan 1117, 1081 HV Amsterdam, The Netherlands
  2. 2Department of Cardiology, Free University Hospital
  3. 3CLB, Sanguin Blood Supply Foundation Service, PO Box 9190, 1006 AD Amsterdam, The Netherlands
  1. Dr Niessen jwm.niessen{at}


Aim—To investigate the presence of membrane “flip flop” in ischaemic human myocardium, we assessed depositions of apolipoprotein H (apoH; β2-glycoprotein 1) in ischaemic myocardium. Serum protein apoH can bind to negatively charged phospholipids and can also inhibit blood coagulation in vitro. We hypothesised that, because of its affinity for phosphatidyl serine, apoH might bind to “flip flopped” cells and would therefore be useful as a marker for membrane flip flop in vivo.

Methods—Myocardial tissue specimens were obtained from patients who had died within 14 days of acute myocardial infarction.

Results—Immunohistochemical analysis of these specimens revealed that apoH was selectively deposited in infarcted areas of human myocardium of at least one day's duration. Depositions of apoH were not found in non-ischaemic myocardial tissue samples obtained from patients who died from other (extracardial) causes. In vitro experiments with the human leukaemia T cell line Jurkat, subjected to apoptosis by etoposide, showed that apoH was bound to the membrane of apoptotic cells. However, these experiments also indicated that flip flop itself is not sufficient for apoH binding. In addition, Jurkat cells that bound apoH were positive for activated complement complexes, as was also found in the human heart.

Conclusions—These results suggest that apoH is involved in the inflammatory processes that occur in ischaemic myocardium.

  • myocardium
  • apolipoprotein H
  • inflammation
  • complement

Statistics from

Apolipoprotein H (apoH; β2-glycoprotein 1) is a 50 kDa phospholipid binding serum protein that preferentially binds to negatively charged phospholipids such as phosphatidyl serine (PS).1,2 Because of this property, apoH might inhibit blood coagulation and ADP dependent platelet aggregation, at least in vitro. Its precise physiological role remains to be elucidated.3 Chonn et al found that the fast clearance of PS containing liposomes by the liver in mice is mediated via plasma derived apoH, suggesting that this protein might have a role in removing PS containing microvesicles.4

Recently, we have shown co-deposition of complement with the acute phase protein C reactive protein (CRP) in infarcted sites of human myocardium.5 The ligand for CRP in infarcted myocardium, however, remains to be established. One possibility6 is that binding sites for CRP are generated in cells that have undergone so called “flip flop”.7 In normal cells, various phospholipids are asymmetrically distributed between the inner and outer leaflet of the membrane, PS being mainly located in the inner leaflet. In damaged cells (ischaemic, apoptotic, or necrotic cells) phospholipids of the inner and outer leaflet exchange, a phenomenon known as flip flop,8,9 leading to the exposure of PS in the outer leaflet. We hypothesised that because of its affinity for PS, apoH could be used as an in vivo marker for these “flip flopped” cells. In our present study we tested this hypothesis and searched for apoH depositions in the infarcted myocardium, in relation to those of activated complement.

Patients, materials, and methods


Patients, referred to the department of pathology for necropsy, were included in this study when at necropsy they showed signs of a recently developed acute myocardial infarction; that is, decreased lactate dehydrogenase (LD) staining (decolouration) of the affected myocardium. Most of the patients had participated in earlier studies on the involvement of CRP and complement in infarcted myocardium.5 Our study was approved by the ethics committee of the Free University Hospital Amsterdam.


Myocardial tissue specimens were obtained from the infarcted as well as from adjacent sites. These latter sites showed normal LD staining patterns and were studied as internal controls. Before being prepared as cryosections, the tissue specimens were stored at −196°C (liquid N2). The glass slides used for microscopy were pretreated with 0.1% poly-L-lysine (Sigma Chemical Company, St Louis, Missouri, USA) to enhance the adherence of the frozen tissue sections.


We used a monoclonal antibody (C3-9; IgG-1 subclass) against activated complement factor C3 that has been used previously for immunohistochemical studies.5 Cof23, directed against apoH, was a gift of Dr Koike (Department of Medicine II, Hokkaido University School of Medicine, Sapporo, Japan). Polyclonal rabbit antihuman apoH/34, was obtained by immunising rabbits with purified apoH (Department of AutoImmune Diseases, CLB, Sanguin Blood Supply Foundation, The Netherlands).

For FACS analysis, fluorescein isothiocyanate (FITC) conjugated annexin V (Boehringer Ingelheim, Heidelberg, Germany) and phycoerythrin (PE) conjugated streptavidin (Amersham International, Amersham, UK) were used.

Biotinylation of monoclonal and polyclonal antibodies using L-C-biotin-N-hydroxisuccimide ester (Pierce Chemical Co., Rockford, IL) was performed according to the manufacturer's instructions.


The human leukaemia T cell line Jurkat was maintained in Iscove's modified Dulbecco's medium (IMDM) (Gibco, Grand Island, New York, USA) supplemented with 5% (vol/vol) fetal calf serum (FCS), 100 U/ml penicillin (Gibco), 100 μg/ml streptomycin (Gibco), 50 μM 2-mercaptoethanol, and 20 μg/ml human transferrin (Sigma) (culture medium) at 37°C in a humid atmosphere and 5% CO2.

Apoptosis was induced in the Jurkat cells by incubation with 25 μM etoposide (Sigma) overnight. The cells were then separated from the medium by centrifugation at 800 ×g for six minutes, resuspended in IMDM (no FCS) in a 96 well roundbottom plate (100 μl/well, 200 000 cells/well), and washed once with IMDM. Thereafter, cells were incubated with serum (or purified apoH) diluted in IMDM (50 μl/well), for 30 minutes at 37°C. After washing four times with BB (1.19 g Hepes, 4.4 g NaCl, 0.19 g KCl, 0.10 g CaCl2, 0.1 g MgCl2 in 500 ml, pH 7.4) and BA (0.5% (wt/vol) bovine serum albumin (BSA) (Boehringer Mannheim), 0.02% (wt/vol) azide), biotinylated anti-apoH antibodies (5 μg/ml in buffer BB-BA, 50 μl/well) or anti-C3 monoclonal antibody (C3-9) were added to the cells for 30 minutes at 4°C. After three washes with BB-BA buffer, a mixture of PE labelled streptavidin (Becton Dickinson, San Jose, California, USA) and FITC labelled annexin V (Boehringer Ingelheim), diluted 1/200 and 1/2000 in BB-BA buffer, respectively, was added to the cells (final volume 50 μl/well). After 20 minutes of incubation at 4°C in the dark, the cells were washed three times with BB-BA. Finally, cells were resuspended in 100 μl BB-BA in epicstubes and analysed on a FACScan (Becton Dickinson, San Jose, California, USA).


Frozen sections (4 μm thick) were mounted on to glass slides, dried for one hour by exposure to air, and fixed in acetone (“Baker analysed reagent”; Mallinckrodt Baker BV, Deventer, Holland). The slides were incubated at room temperature for 10 minutes with normal rabbit serum (Dakopatts A/S, Glostrup, Denmark) and diluted 1/50 in phosphate buffered saline (PBS) containing 1% (wt/vol) BSA (PBS-BSA) after rinsing in PBS. Incubation of the slides with specific antibody solutions was performed for 60 minutes at room temperature (C3-9 diluted 1/1000 in PBS-BSA; biotinylated anti-apoH polyclonal antibody diluted 1/500; monoclonal antibody against apoH diluted 1/400). In control experiments, similar incubations were performed with irrelevant control monoclonal antibodies: IgG1 and mouse myeloma protein, MOPC (Cappel, Organon Teknika, Turnhout, Belgium).

The slides incubated with antibodies against complement were washed for 30 minutes with PBS and incubated with horseradish peroxidase conjugated rabbit antimouse immunoglobulins (RaM-HRP; Dakopatts A/S), diluted 1/25 in PBS-BSA. The slides incubated with biotinylated anti-apoH monoclonal or polyclonal antibodies were washed for 30 minutes and incubated with streptavidin horseradish peroxidase conjugates (Dakopatts), diluted 1/500 in PBS-BSA for one hour. Thereafter, the slides were washed again in PBS and incubated for four minutes in 0.5 mg/ml 3,3`-diaminobenzidinetetrahydrochloride (DAB; Sigma) in PBS, pH 7.4, containing 0.01% (vol/vol) H2O2. The slides were then washed again, counterstained with haematoxylin for 40 seconds, dehydrated, cleared, and finally mounted.

Microscopic criteria10,11 were used to estimate infarct duration in all myocardial tissue specimens. Jeopardised fibres were characterised by the intensity of eosinophilic staining of involved myofibres, loss of nuclei and cross striation, polymorphonuclear neutrophil and lymphocyte infiltration, and fibrosis. However, because morphological judgement is more reliable in paraffin wax embedded slides, corresponding paraffin wax embedded tissue slides were also made, to confirm the determination of jeopardised versus non-jeopardised tissue. Two independent investigators (HWMN, WKL) each judged and scored all slides for infarct duration and anatomical localisation of specific antibody as visualised by immunohistochemical staining. Anatomical localisations examined were myofibre (membrane, cytoplasm, cross striations) and bloodvessel elements. For the final scoring results, a consensus was achieved by the two investigators.



Myocardial tissue specimens were obtained from 17 patients who had died after acute myocardial infartion as confirmed by necropsy performed within 24 hours after death (table 1). Specimens were obtained from the infarcted as well as from the unaffected myocardial tissue. The infarct age, assessed by microscopical criteria,10,11 varied from less than 12 hours to more than 2 weeks.

Table 1

Duration of infarctions in the patients included


As shown recently, complement was localised in infarcted sites of human myocardium.5 ApoH, as detected by the biotinylated monoclonal antibody, was found in the same parts that stained positive for complement (fig 1). Notably, staining for apoH was most intense at the plasma membrane of cardiomyocytes. Cytoplasmatic localisation of apoH was also found, although staining for apoH in the cytoplasm was clearly less than that of the plasma membrane. Some staining of cross striations was also found (fig 2). In larger infarcts, the border zone appeared to stain more intensively for apoH than the centre of the infarcted region.

Figure 1

(A) Immunohistochemical localisation of the activated complement component C3 (monoclonal antibody C3-9) and (B) apolipoprotein H (biotinylated monoclonal antibody) in infarcted myocardium (J, jeopardised; N, normal). Specimens were from a patient with an infarction of five to nine days' duration. The frozen tissue sections shown were from the same myocardial infarction site (magnification, ×100).

Figure 2

Positive staining of myofibre elements (sarcolemma, (A) cytoplasm, and (B) cross striation (arrows)) in infarcted myocardium for apolipoprotein H (biotinylated monoclonal antibody; magnification, ×630). The specimen is from a patient with an infarction of one to three days' duration.

Similar to complement, apoH was not detected in infarctions of less then 24 hours duration. Furthermore, apoH was not found on endothelium, in contrast to complement.5 Similar staining results were obtained when a biotinylated rabbit polyclonal antibody was used against apoH (fig 3).

Figure 3

Immunohistochemical localisation of apolipoprotein H (biotinylated polyclonal antibody) in infarcted myocardium. Specimens were from a patient with infarction of five to nine days' duration (magnification, ×630).

Staining of the myocardial tissue specimens with irrelevant control antibodies yielded negative results. In addition, internal controls—specimens taken from non-infarcted sites of the myocardium of the same patient—did not stain for C3 or apoH. Furthermore, myocardial tissue specimens from an immature child who died in utero at an amenorrhoea duration of 22 weeks (these specimens were taken to represent a pure, non-ischaemic myocardial control) did not stain for C3 or apoH, and neither did an old infarction (> 1 year old).

Unfortunately, immunohistochemical studies have the limitation that the well known marker of membrane flip flop, annexin V, cannot be used reliably in tissue slides of the heart. For this reason, we performed in vitro studies using Jurkat cells. These cells were incubated with etoposide overnight, which causes them to die as a result of apoptosis.12 Subsequently, the cells were incubated with anti-apoH or anti-C3. The flip flop phenomenon was assessed by staining for annexin V. The cells that were annexin V negative did not bind apoH (table 2), whereas annexin V positive cells did. Recently Manfredi et al have shown that Jurkat cells, positive for annexin V, stained for apoH.13,14 However, when we characterised these cells in more detail, it appeared that only cells that also stained with propidium iodide (and thus were beyond the early phase of apoptosis as characterised by the flip flop phenomenon) bound apoH. This indicates that staining for apoH is not discriminative in detecting cells that only underwent flip flop of the membrane. However, similar to the immunohistochemical staining results in the heart, apoH positive cells also bound complement, suggesting a role of both mediators in the inflammatory mediated process of cell death.

Table 2

Binding of annexin V, activated complement, and apolipoprotein H to apoptotic cells


The precise mechanisms contributing to myocardial cell death in the human myocardium after infarction are still not fully understood. We have hypothesised that in jeopardised myocardium the membrane of cardiomyocytes might become perturbed, thereby becoming a target for acute phase proteins such as serum phospholipase A2 and CRP.6 This would result in fixation of CRP to the cells and the subsequent activation of complement.6 Indeed, colocalisation of complement and CRP in jeopardised human myocardium was recently demonstrated, supporting this hypothetical mechanism.5

Perturbation of the cell membrane involves an exchange of phospholipids of the inner and outer leaflets, resulting in flip flop of the membrane. As a consequence of this process, cells expose PS in the outer leaflet. Flip flop of cell membranes in vivo is difficult to demonstrate. It has been shown that in vitro apoH binds to PS.15 Because of this property, apoH deposition in tissues may reflect flip flop of membranes in vivo. In our present study, we indeed found that apoH is localised to jeopardised human myocardium. Moreover, apoH in particular appeared to bind to the membranes of cardiomyocytes, supporting its supposed function as a marker for membrane flip flop. However, in vitro experiments in human leukaemia T cell Jurkat cells indicate that flip flop itself is not sufficient for apoH binding.

In addition to binding to PS, apoH might also bind to cardiolipin.16–18 Cardiolipin is a constituent of the inner19,20 and outer mitochondrial membrane.21 Thus, cardiolipin might also serve as a binding site for apoH. This could explain the cytosolic localisation of apoH that we saw in a few of the infarctions. The identity of the ligands for apoH on (apoptotic) cells is currently under investigation.

In addition, another apolipoprotein, apoliprotein J (clustrin) has been found to be colocalised in ischaemic human hearts together with complement membrane attack complex (MAC).22 It has been suggested that apolipoprotein J might be involved in the clearance of damaged and necrotic tissue, together with MAC. A similar role might be played by apoH, especially because the clearance of liposomes by the liver in mice is in part mediated by plasma derived apoH. Interestingly, after interaction with serum, liposomes not only become positive for apoH but also for complement.4 In the human myocardium and in in vitro experiments with Jurkat cells, we have shown colocalisation of activated complement and apoH. This raises the possibility that apoH, bound to cells, might trigger or enhance complement activation, directly or indirectly via CRP–ligand complexes.

In conclusion, for the first time we have shown that apoH is localised in infarcted human myocardium to areas that also contain CRP and activated complement.5 We therefore hypothesise that apart from complement, CRP, and apolipoprotein J, apoH is a new player in the inflammatory changes ensuing in infarcted myocardium.


This study was financially supported by the Netherlands Heart Foundation, grant number 93–119. Dr Niessen is a recipient of the Dr Dekker programme of the Netherlands Heart Foundation (D990025).


Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Linked Articles

  • BMJ Publishing Group Ltd and Association of Clinical Pathologists