Low-molecular weight and unfractionated heparins induce a downregulation of inflammation: decreased levels of proinflammatory cytokines and nuclear factor-κB in LPS-stimulated human monocytes


  • Hélène Hochart,

    1. Thrombosis and Haemostasis Research Group, Institute of Molecular Medicine, Trinity Centre for Health Sciences, Trinity College Dublin, St James's Hospital, Dublin
    Search for more papers by this author
  • P. Vincent Jenkins,

    1. Thrombosis and Haemostasis Research Group, Institute of Molecular Medicine, Trinity Centre for Health Sciences, Trinity College Dublin, St James's Hospital, Dublin
    2. National Centre for Hereditary Coagulation Disorders, St James's Hospital, Dublin
    Search for more papers by this author
  • Owen P. Smith,

    1. Department of Paediatric Haematology and Oncology, Our Lady's Hospital for Sick Children, Crumlin, Dublin, Ireland
    Search for more papers by this author
  • Barry White

    1. Thrombosis and Haemostasis Research Group, Institute of Molecular Medicine, Trinity Centre for Health Sciences, Trinity College Dublin, St James's Hospital, Dublin
    2. National Centre for Hereditary Coagulation Disorders, St James's Hospital, Dublin
    Search for more papers by this author

Helene Hochart, Thrombosis and Haemostasis Research Group, Institute of Molecular Medicine, Trinity Centre for Health Sciences, Trinity College Dublin, St James's Hospital, James's Street, Dublin 8, Ireland.
E-mail: hocharh@tcd.ie.


Unfractionated heparin (UFH) and low-molecular weight heparin (LMWH) are well defined anticoagulant agents. Recent data suggest that both LMWH and UFH may also have potent anti-inflammatory properties; however, their mechanism of action responsible for the anti-inflammatory effect is not yet fully elucidated. This study was designed to assess the effect of LMWH and UFH on human monocytes production of inflammatory markers and nuclear translocation of nuclear factor (NF)-κB. Cultured monocytes were pretreated for 15 min with LMWH or UFH (10 μg and 1 μg/million cells) before stimulation with lipopolysaccharide (LPS) at a dose of 1 ng/million cells. Proinflammatory cytokines tumour necrosis factor (TNF)-α, interleukin (IL)-8, IL-6 and IL-1β release were subsequently measured by enzyme-linked immunosorbent assay at 6 h, and nuclear translocation of the proinflammatory NF-κB was assessed at 2 h. Treatment with pharmacological doses of LMWH and UFH significantly attenuated LPS-induced production of TNF-α, IL-8, IL-6 and IL-1β as well as NF-κB translocation. These results indicate equivalent and significant heparin anti-inflammatory properties at low doses on monocyte-mediated immune response. The inhibition of NF-κB activation certainly represents one of the mechanisms by which heparin exerts its anti-inflammatory effect. LMWH and UFH therefore appear as potential therapeutic inhibitors of inflammation.

Unfractionated heparin (UFH) and low-molecular weight heparin (LMWH) are highly sulphated proteoglycans that are widely applied as anticoagulant drugs. They represent pivotal agents for the prevention and treatment of thromboembolic disorders, pulmonary embolism, disseminated intravascular coagulation and unstable angina (Hirsh et al, 2001). Their anticoagulant mechanism of action resides in their ability to increase the activity of an endogenous coagulation factor, antithrombin III, which inhibits two serine proteases involved in the coagulation cascade: factor IIa or thrombin and factor Xa (Bourin & Lindahl, 1993). Recently, a number of clinical and experimental studies have suggested that, aside from their anticoagulant capacity, heparins affect inflammatory responses. The accumulation of leucocytes in inflamed brain (Yanaka & Nose, 1996), skin (Teixeira & Hellewell, 1993) and lung (Seeds & Page, 2001) have been reduced by heparin in animal studies, and in several clinical trials heparin appeared to have the potential to treat inflammatory bowel disease (IBD) (Gaffney et al, 1991, 1995; Evans et al, 1997), arthritis (Gaffney & Gaffney, 1996), rhinitis (Vancheri et al, 2001) and human asthma (Page, 1991; Diamant et al, 1996), with an anti-inflammatory effect dissociable from its anticoagulant activity (Lever & Page, 2002). Experimental studies have shown that UFH and a variety of heparin derivatives inhibit poinflammatory cytokine gene expression by lipopolysaccharide (LPS)-stimulated human mononuclear cells (Hogasen & Abrahamsen, 1995; Attanasio et al, 1998; Gori et al, 2004); however, their precise mechanism of action is still unknown. In addition, heparin and o-desulphated heparin have been reported to inhibit the nuclear factor (NF)-κB activation in a tumour necrosis factor (TNF)-α-stimulated human endothelial cell line and in ischaemic-reperfused rat myocardium (Thourani et al, 2000). LMWH also appeared to inhibit the nuclear translocation of NF-κB in high glucose-stimulated human endothelial cells (Manduteanu et al, 2003) and in T cells (Hecht et al, 2004).

The addition of LPS to cells results in activation of the transcription factor NF-κB, which plays crucial roles in regulating the expression of many proinflammatory cytokine genes involved in the inflammatory responses, such as those encoding for TNF-α, interleukin (IL)-8, IL-6 and IL-1β (Baldwin, 1996; Yao et al, 1997). The increased production of TNF-α and IL-1β by NF-κB translocation causes its activation in return, leading to a positive regulatory loop that amplifies local inflammation (Barnes & Karin, 1997). NF-κB activation, a critical phenomenon in host inflammatory response, is implicated in a wide range of diseases, such as sepsis, IBD, asthma and rheumatoid arthritis (Bohrer et al, 1997), and therefore represents an ideal molecular target. The aim of this study was to evaluate and compare the effect of an unfractionated and a LMWH on the production of proinflammatory cytokines and NF-κB activation in LPS-stimulated human monocytes.

Materials and methods

Unfractionated heparin and low-molecular weight heparin

The UFH Minihep and the LMWH Innohep (Tinzaparin sodium) were kindly provided by Leo Pharma (Dublin, Ireland). UFH was extracted from porcine intestinal mucosa; and LMWH was produced by enzymatic depolymerisation of UFH using the heparinase I from Flavobacterium heparinum, followed by preparative gel filtration on Sephadex. Both heparins were certified as sterile and preservative-free by the manufacturer.


Lipopolysaccharide from Escherichia coli (serotype 0127:B8 purified by phenol extraction) was supplied by Sigma Chemicals (St Louis, MO, USA).

Human monocytes isolation and FACS staining

All blood donors were recruited in this study with informed consent in accordance with local ethical guidelines. Peripheral blood mononuclear cells (PBMCs) were obtained from citrate-anticoagulated venous blood by Lymphoprep density-gradient centrifugation (Nycomed, Oslo, Norway). Briefly, blood diluted in two parts in phosphate-buffered saline (PBS) was slowly layered over Lymphoprep; and platelets removed after centrifugation at 400 g for 40 min. PBMC layers were collected and washed twice with PBS at 400 g for 10 min to eliminate further platelet contaminants and remaining Lymphoprep. Viable cells (%≥95) were counted by means of a haemocytometer using the trypan blue exclusion technique. The non-monocyte fraction of the PBMCs was then depleted using Monoclonal Antibody Cell Sorter (MACS) microbeads (Milteny Biotec, Bergish Gladbach, Germany) conjugated to antibodies directed against lymphocytes, basophils, eosinophils and erythrocytes, allowing a negative selection of monocytes. A VarioMACS magnet (Milteny Biotec) was used, with a separation of up to 2 × 109 cells. Isolated monocytes were identified by flow cytometric analysis (FACsCalibur, Becton Dickinson, Oxford, UK) after staining with a specific monoclonal antibody [R-phycoerythrin (RPE)-conjugated mouse antihuman CD14; DakoCytomation Ireland Ltd, Galway, Ireland]. Data analysis were performed using cellquest software (Becton Dickinson); the percentage of viable pure monocytes was ≥90% in the final suspension.

Monocytes preparation for cytokine and NF-κB measurement experiments

Isolated monocytes were resuspended at a density of 1 × 106 cells/ml for cytokine and 2 × 106 cells/ml for NF-κB measurement, in Roswell Park Memorial Institute (RPMI) medium supplemented with 10% platelet-free autologous plasma, in order to maintain high cell viability. Cells were then either exposed to LPS alone (1 ng/million cells) or in combination with 1 IU (10 μg/ml) or 0·1 IU (1 μg/ml) heparin. Experimental controls consisted of untreated and heparin alone treated cells. UFH or LMWH was added to cells 15 min prior to stimulation with LPS (according to a preliminary time course experiment carried out on the THP-1 cell line, data not shown). Those samples not receiving LPS, UFH or LMWH received an equal volume of PBS. After 2 or 6 h incubation at 37°C and 5% CO2, each sample was aspirated into a microcentrifuge tube and spun at 400 g for 5 min. Cell-free culture supernatants were removed and stored at −80°C for subsequent cytokine measurements and cell pellets were collected for nuclear protein extraction.

Cell viability

Monocyte viability under all experimental conditions was determined by two different methods. Quantification of cell viability was based on the cleavage of the Water-Soluble-Tetrazolium Salt (WST-1) by mitochondrial dehydrogenases in viable cells using the Cell Proliferation Reagent WST-1 (Roche Molecular Biochemicals, Basel, Switzerland) according to the manufacturer's instructions. Quantification of cell death and cell lysis was based on the measurement of lactate dehydrogenase (LDH) activity in the supernatant using the Cytotoxicity Detection Kit (Roche) according to the manufacturer's instructions.

Enzyme-linked immunosorbent assay

Solid-based sandwich enzyme-linked immunosorbent assays (ELISAs) were carried out on cell culture supernatants using R&D Systems (Minneapolis, MN, USA) DuoSet ELISA Development Systems for human TNF-α, IL-8, IL-6 and IL-1β. The protocol was followed for each assay according to the manufacturer's instructions.

Nuclear protein extraction

Nuclear extracts were prepared for NF-κB analysis from whole cell pellets following a Nuclear Extract Kit (Active Motif, Rixensart, Belgium) methodology, in which proteins contained in the cytoplasmic and the nuclear compartments of the cell are obtained separately. Briefly, 2 × 106 cells were washed once with PBS/phosphatase inhibitors and centrifuged at 400 g for 5 min at 4°C. The cell pellets were then resuspended in 160 μl 1X hypotonic buffer and incubated on ice for 15 min. After addition of 8 μl detergent, the lysed cellular suspensions were mixed by vortexing, incubated on ice for 5 min and then centrifuged at 14 000 g for 30 s at 4°C. Cytoplasmic extracts were removed and stored at −80°C. Nuclear extracts were prepared by resuspending the 2 × 106 cell nuclei in 40 μl complete lysis buffer, incubating it for 30 min on ice on a rocking platform (150 rpm), vortex-mixing it for 30 s and then centrifuging it at 14 000 g for 10 min at 4°C. Supernatants (nuclear fraction) were stored at −80°C. The nuclear protein content of each extract was determined by the Bradford assay (Sigma-Aldrich) using bovine serum albumin (BSA) as standard.

Nuclear transcription factor-κB detection

The NF-κBp50 detection was carried out following the sensitive and specific chemiluminescent ELISA-based TransAM NF-κBp50 Transcription Factor Assay (Active Motif) methodology. Briefly, 30 μl of complete binding buffer was added to microwells coated with an immobilised oligonucleotide containing an NF-κB consensus binding site (5′-GGGACTTTCC-3′). An equal amount of nuclear cell extract (1·25 μg) diluted in 20 μl of complete lysis buffer was added to microwells. Positive control (Jurkat nuclear extract) and blank (complete lysis buffer) wells were also prepared. After 1 h incubation with mild agitation on a rocking platform, the microwells were washed three times with 200 μl 1X washing buffer, and 50 μl of diluted (1:1000 in 1X Ab binding buffer) NF-κBp50 antibody added to all microwells. After 1 h incubation, the microwells were washed three times with 200 μl 1X washing buffer, and 50 μl of diluted (1:10 000 in 1X Ab binding buffer) horseradish peroxidase-conjugated antibody added to all microwells. Finally, after 1 h incubation, the microwells were washed four times with 200 μl 1X washing buffer, and 50 μl of chemiluminescent working solution added to all microwells. Chemiluminescence results were obtained in Relative Luminescence Units (RLU), using a Tecan SpectraFluor Plus luminometer (RetiSoft Inc., Toronto, ON, Canada).


Cytokine data (ng/ml) are expressed as mean ± SD of six replicates from three different experiments. NF-κB data (RLU) are expressed as mean ± SD of four replicates from two different experiments. All monocyte experiments were performed on blood samples from different individuals. Statistical analyses were performed on raw data (ng/ml and RLU) using a paired t-test (data desk 4·0 software, Data Description, Ithaca, NY, USA). The P-values lower than 0·05 were considered statistically significant.


Cell viability

Monocyte viability under all experimental conditions was checked in preliminary experiments. Incubation of cells with unfractionated and LMWHs did not show any detrimental effect on cells (data not shown).

Effect of LPS and heparin on cytokine production by monocytes

The effect of UFH and LMWH on LPS-induced proinflammatory cytokines production by monocytes was measured at 6 h after LPS stimulation. Secreted levels (ng/ml) in cell culture supernatants of TNF-α, IL-8, IL-6 and IL-1β from six replicates were determined by ELISA and pooled together. Overall, levels of TNF-α, IL-8, IL-6 and IL-1β were found to be low in untreated and heparin-treated samples (data not shown), strongly increased upon LPS stimulation, and reduced by heparin addition prior to LPS compared with LPS treatment alone (Table I). LPS at 1 ng per 1 × 106 cells appeared to induce a 42-fold increase in TNF-α, a 5-fold in IL-8, a 30-fold in IL-6 and a 5·5-fold in IL-1β compared with unstimulated cells. Heparin addition to cells resulted in significant inhibition (P < 0·05) of all proinflammatory cytokines production compared with LPS alone. TNF-α, IL-8, IL-6 and IL-1β levels were decidedly decreased, by 17·8%, 9·9%, 11·2% and 33·3%, respectively, with the addition of 1 IU/ml LMWH, and by 27·3%, 13·5%, 17% and 43·9%, respectively, with 0·1 IU LMWH, also by 17·4%, 12·7%, 22·7% and 45·6%, respectively, with 1 IU/ml UFH, and by 15·4%, 12·1%, 28·3% and 61·4%, respectively, with 0·1 IU UFH. LMWH and UFH treatments were not statistically different from each other for any of the cytokines investigated. Treatment of monocytes with either LMWH or UFH (1 IU and 0·1 IU) alone had little effect on TNF-α, IL-6 and IL-1β cytokine protein levels, with results equivalent to untreated samples.

Table I.  Effect of LMWH and UFH on proinflammatory cytokine production (ng/ml/106 cells) by LPS (1 ng/ml/106 cells) stimulated human monocytes.
  1. Data are presented as mean ± SD of three monocyte experiments realised in duplicate. The amount of cytokine production after heparin treatment was measured by ELISA (R&D system), and the percentage inhibition of each cytokine after heparin treatment compared with the LPS control was evaluated for statistical significance using a paired t-test (*P < 0·05 vs. LPS; **P < 0·01 vs. LPS; ***P < 0·001 vs. LPS).

  2. LMWH, low-molecular weight heparin; UFH, unfractionated heparin; LPS, lipopolysaccharide; TNF, tumour necrosis factor; IL, interleukin; ELISA, enzyme-linked immunosorbent assay.

LPS 1 ng/ml1·836 ± 1·0875·582 ± 1·511·832 ± 1·2210·057 ± 0·024
LMWH 1 IU + LPS1·509 ± 0·893**5·032 ± 1·571*1·626 ± 1·193*0·038 ± 0·022*
LMWH 0·1 IU + LPS1·335 ± 0·652*4·831 ± 1·549**1·520 ± 1·082*0·032 ± 0·014**
UFH 1 IU + LPS1·517 ± 0·925**4·876 ± 1·422***1·417 ± 0·893*0·031 ± 0·018*
UFH 0·1 IU + LPS1·553 ± 0·957**4·909 ± 1·207**1·314 ± 0·959**0·022 ± 0·009*

Effect of LPS and heparin on nuclear factor-κBp50 translocation in monocytes

We explored a possible mechanism for the cytokine inhibition in LPS-stimulated monocytes by examining the effect of UFH and LMWH on the nuclear translocation of activated NF-κBp50, a central element involved in the gene expression of proinflammatory cytokines. A preliminary time course experiment on THP-1 cells defined the optimal cell culture incubation time at 2 h after LPS stimulation (data not shown). NF-κBp50 RLU levels in the cell nuclei obtained from four replicates were pooled together. Treatment of monocytes with LPS at 1 ng per 1 × 106 cells strongly activated NF-κBp50, inducing a 3·3-fold increase in RLU compared with unstimulated cells (Fig 1). Preincubation with heparin prior to LPS stimulation significantly decreased NF-κBp50 levels (P < 0·05), by 26% on average compared with LPS alone. Treatment with LMWH or UFH alone did not show any downregulation of basal NF-κBp50 compared with unstimulated cells. LMWH and UFH treatments were not statistically different from each other and no difference in dose–response was noted. To reinforce our finding on NF-κBp50 downregulation by heparin in LPS-activated monocytes, TNF-α levels from the same experiments were subsequently measured by ELISA (data not shown). Preincubation with heparin prior to LPS stimulation of monocytes significantly decreased TNF-α levels, by 23% on average.

Figure 1.

Effect of low-molecular weight heparin (LMWH) and unfractionated heparin (UFH) on lipopolysaccharide (LPS)-induced nuclear factor (NF)-κBp50 activation in human monocytes. Monocytes (1 × 106/ml) were preincubated for 15 min with LMWH and UFH at two different concentrations (1 IU = 10 μg and 0·1 IU = 1 μg/ml) before the addition of LPS (1 ng/ml). The nuclear NF-κBp50 activation was determined at 2 h by nuclear protein extraction followed by TransAM NF-κBp50 Chemi Transcription Factor Assay. Data are expressed as mean Relative Luminescence Units (RLU) ± SD of four replicates from two different experiments. The percentage inhibition of NF-κB translocation was analysed by paired t-test (*P < 0·05; **P < 0·01).


Heparin and related glycosaminoglycans are increasingly recognised as modulators of the inflammatory process. Heparin and its derivatives inhibit the activation of inflammatory cells, their cytokine expression and production (Gori et al, 2004) and also their adhesion to vascular endothelial cells (Lever et al, 2000; Smailbegovic et al, 2001). Monocytes play a pivotal role in the host inflammatory response, by producing a wide variety of proinflammatory cytokines. It was therefore of interest to elucidate and understand the mechanism by which heparin exerts its effect on these cells.

We examined the regulation of LPS-induced proinflammatory cytokines release by LMWH and UFH pretreated monocytes. As expected, the production of TNF-α, IL-8, IL-6 and IL-1β was upregulated by a nanomolar dose of LPS as a result of its binding to the monocyte surface and the subsequent intracellular signalling leading to proinflammatory cytokines gene transcription. The cytokines measured in our experiments are largely involved in the upregulation of inflammatory reactions and thus in numerous inflammatory diseases. The level of cytokine produced by LPS-stimulated monocytes was compared to that of cells pretreated with LMWH and UFH at two different, pharmacologically relevant, concentrations (1 and 0·1 IU/ml) prior to LPS addition [therapeutic heparin levels are equivalent to 0·3–0·7 IU/ml for the treatment of venous thrombosis (Hirsh & Raschke, 2004)]. All cytokine amounts measured were significantly reduced when the cells were treated with heparin, up to 18% on average with 1 IU/ml LMWH, 26% with 0·1 IU/ml LMWH, 24·6% with 1 IU/ml UFH and 29·3% with 0·1 IU/ml UFH. It appeared that the lower heparin concentration (0·1 IU compared with 1 IU) induced a stronger decrease in inflammation. Heparin molecules, either LMWH or UFH, are negatively charged and may aggregate or interact with other components in the medium during incubation, preventing their efficient interaction with monocytes. Conversely, smaller and less numerous molecules possess better access to cell receptors. However, no significant statistical difference was apparent between different heparin treatments.

These results demonstrated an inhibition of LPS induction of proinflammatory cytokines in heparin pretreated cells, and are in agreement with recent similar studies performed on mononuclear cells (Attanasio et al, 1998; Gori et al, 2004). Heparin may therefore represent a potential inhibitory agent in chronic inflammatory diseases, by reducing the release of proinflammatory cytokines.

Lipopolysaccharide activation of monocytes leads to the initiation of proximal signalling events, which result in the increased transcription and stability of proinflammatory cytokine messenger RNA. Increased transcription requires the concerted binding of NF-κB, c-jun and Egr-1 to the promoter region of a wide variety of proinflammatory cytokine genes, such as TNF-α gene (Yao et al, 1997). We showed previously that activated protein C, an anticoagulant that protects against sepsis, inhibited the NF-κB translocation in the LPS-stimulated monocytic cell line THP-1 (White et al, 2000). To further investigate the anti-inflammatory mechanism of heparin action, we therefore examined its effect on monocyte LPS-mediated NF-κB activation. An accurate monitoring of NF-κB activation in cells is crucial for signal transduction pathway analysis, and we demonstrated here that LPS-induced NF-κB translocation from the cytoplasm to the nucleus was significantly reduced in heparin pretreated cells, even at low heparin doses, as shown by the 30% reduction induced with 0·1 IU/ml UFH. These results were further supported by the observation of a downregulation in TNF-α levels expressed by monocytes under the exact experimental conditions. Our findings on the downregulation of NF-κB activation in monocytes by heparin are similar to findings on heparin effects on NF-κB activation in stimulated endothelial cells. Thourani et al (2000) showed that o-desulphated heparin downregulated NF-κB-mediated activation of TNF-α-stimulated human umbilical vascular endothelial cells. Similarly a study by Manduteanu et al (2003) demonstrated a significant inhibition of NF-κB expression by a LMWH in high glucose-stimulated human endothelial cells.

Nuclear factor-κB activation is considered to be an amplifying and perpetuating mechanism of the inflammatory process, and is implicated in a wide range of inflammatory diseases. NF-κB regulates the expression of many genes whose products are chemokines, immune receptors and adhesion molecules, such as intercellular adhesion molecule-1, vascular cell adhesion molecule-1 and E-selectin, involved in the recruitment of inflammatory cells from the circulation to the site of inflammation. Its activation is also vital for proinflammatory cytokines regulation (Baeuerle & Baichwal, 1997). This transcription factor therefore represents an excellent pharmacological target as it promotes the transcription of the TNF-α gene, among others; and anti-TNF-α therapy has been found to be beneficial in rheumatoid arthritis and Crohn's disease (Feldmann et al, 1994). The inhibition of NF-κB would interrupt the positive inflammatory feedback loop generated by TNF-α and IL-1β, and globally downregulate the production of a wide range of proinflammatory molecules. It is this particular property of NF-κB that makes it such an attractive molecular target for novel anti-inflammatory therapies and offers an explanation for the heparin effects. The ability of heparin to downregulate LPS-induced NF-κB activation may be of considerable clinical significance according to the importance of the latter in inflammatory disorders.

In conclusion, these results demonstrate that pharmacologically relevant doses of heparin possess the ability to inhibit LPS-induced proinflammatory cytokines and nuclear translocation of NF-κB in human monocytes; and indicate a potential mechanism responsible, in part, for the protective effect of the drug in inflammatory disorders. The heparin inhibition of NF-κB translocation identifies a novel and important immunomodulatory pathway. In clinical practice, heparin use as an anti-inflammatory agent is restricted by its potential to induce bleeding complication, though novel, anti-inflammatory non-haemostatically active heparin derivatives are being developed (Gori et al, 2004). Further investigations are required in this area, in order to optimise the development of such inhibitory agents of the inflammatory process and also to elucidate the molecular mechanism of action of heparin and its derivatives.


The authors would like to thank Leo Pharma for provision of low-molecular weight heparin Innohep (Tinzaparin sodium) and unfractionated heparin (Minihep). Also wish to thank all the generous blood donors. The authors acknowledge Mr Ciaràn Murphy and Mrs Jodie Bruneau for their advice and technical assistance. This work was supported by an unrestricted grant from Novo Nordisk.