bacterial sepsis
Berbée, J.F.P.
Citation
Berbée, J. F. P. (2007, May 24). The role of apolipoprotein CI in lipid metabolism and bacterial sepsis. Retrieved from
https://hdl.handle.net/1887/11973
Version: Corrected Publisher’s Version
License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden
Downloaded from: https://hdl.handle.net/1887/11973
Note: To cite this publication please use the final published version (if applicable).
Chapter 5
Apolipoprotein CI Stimulates the Response to
Lipopolysaccharide and Reduces Mortality in Gram-
Negative Sepsis
Jimmy F.P. Berbée1,2, Caroline C. van der Hoogt1,2, Robert Kleemann1,3, Emile F. Schippers4, Richard L. Kitchens6,Jaap T. van Dissel4, Irma A.J.M. Bakker-
Woudenberg7, Louis M. Havekes1,2,5, Patrick C.N. Rensen1,2
From the department of 1Biomedical Research, TNO-Quality of Life, P.O. Box 2215, 2301 CE Leiden, The Netherlands; Departments of 2General Internal Medicine,
3Vascular Surgery, 4Infectious Diseases, and 5Cardiology, Leiden University Medical Center, P.O. Box 9600, 2300 RC, Leiden, The Netherlands; 6Department of Internal Medicine, Div. of Infectious Diseases, UT Southwestern Medical Center, 5323 Harry Hines Blvd, Dallas, TX 75390-9113, USA; 7Department of Medical Microbiology and Infectious Diseases, Erasmus Medical Center, P.O. Box 1738, 3000 DR Rotterdam, The
Netherlands.
FASEB J 2006, 26 (12): 2162-2164
Abstract
Objective: Gram-negative sepsis is a major death cause in intensive care units.
Accumulating evidence indicates the protective role of plasma lipoproteins such as HDL in sepsis. It has recently been shown that septic HDL is almost depleted from apoCI, suggesting that apoCI may be a protective factor in sepsis.
Methods and Results: sequence analysis revealed that apoCI possesses a highly conserved consensus KVKEKLK binding motif for LPS, an outer-membrane component of Gram-negative bacteria. Through avid binding to LPS involving this motif, apoCI improved the presentation of LPS to macrophages in vitro and in mice, thereby stimulating the inflammatory response to LPS. Moreover, apoCI dose-dependently increased the early inflammatory response to Klebsiella pneumoniae-induced pneumonia, reduced the number of circulating bacteria, and protected mice against fatal sepsis.
Conclusions: Our data support the hypothesis that apoCI is a physiological protector against infection by enhancing the early inflammatory response to LPS, and suggest that timely increase of apoCI levels could be used to efficiently prevent and treat early sepsis.
Introduction
Sepsis affects more than 700,000 people annually and accounts for about 210,000 deaths per year in the US alone. The incidence is still rising at rates between 1.5% and 8% per year1. Despite the active search for novel therapeutic agents2, sepsis remains a serious cause of morbidity and mortality in intensive care units. Many cases of sepsis are caused by Gram-negative bacteria, which evoke immune responses mainly via their outer membrane component LPS through cellular activation via TLR43. TLR4 is important in host defense against many Gram-negative bacteria, as shown by an impaired defense of TLR4- deficient mice during urinary tract infection with Escherichia coli4, intratracheal inoculation of Klebsiella pneumoniae5,6, intranasal administration of Haemophilus influenzae7, and intraperitoneal infection with Klebsiella8, Neisseria9, and Salmonella10 species.
TLR4 signaling results in the production of various proinflammatory mediators including TNFα. These mediators plays an essential role in the early host defense to infection and sepsis by generating an adequate response to bacterial infections, as shown in several animal models11-19. In fact, neutralization of TNFα by antibodies impaired bacterial clearance19, aggravated mortality from bacterial infection in mice11,12,17,19, and was harmful in those septic patients with low risk of death20. Similarly, an innate anti-inflammatory cytokine profile in humans has
been associated with fatal outcome of meningococcal disease21,22 and febrile illness caused by community-acquired infection23. Therefore, it is evident that a timely proinflammatory response to infections is crucial for generating an efficient anti-bacterial attack, thereby preventing the onset of the systemic inflammatory response syndrome (SIRS), septic shock, and eventually death24.
Recently, it has been recognized that sepsis is closely linked to plasma lipoprotein metabolism. Sepsis affects plasma lipoprotein levels by modulating lipolytic enzymes and lipoprotein receptors25. In addition, all lipoprotein classes can bind LPS, and, as such, affect the biological response to LPS26,27. Previously, we have demonstrated that protein-free emulsion particles that mimic lipoproteins are unable to affect the biological fate of LPS, despite the fact that they bind LPS in vitro28. This indicates that the protein moieties of lipoproteins (i.e. apolipoproteins) are responsible for their LPS-modulating effects. Indeed, we were able to show that apoE binds LPS directly, thereby redirecting LPS from macrophages to hepatocytes in vivo28. Consequently, apoE prevents the LPS-induced production of cytokines (including TNFα) and subsequent death in rodents29. However, although reduced cytokine production may be of benefit in severe sepsis by preventing ongoing excessive inflammation, it can be detrimental in early sepsis by impairing the anti-bacterial response.
ApoCI is encoded by the same gene cluster as apoE and has opposite effects on lipoprotein metabolism. Whereas apoE has a classical function in facilitating the hepatic clearance of triglyceride (TG)-rich lipoproteins30, apoCI is able to block the clearance of lipoproteins by inhibiting the lipoprotein lipase-mediated hydrolysis of their core triglycerides in the periphery31 and by interfering with the apoE-dependent recognition of lipoproteins by their hepatic receptors32. Since apoCI has recently been shown to be virtually absent from HDL (the main carrier of apoCI in plasma) in human sepsis33, we hypothesized that apoCI may also play a role in modifying the biological response to LPS. We indeed demonstrate that apoCI contains a consensus LPS-binding motif, enhances the biological response to LPS, and reduces mortality in Gram-negative sepsis in mice.
Materials and Methods
Animals − Male apoCI-deficient (apoc1-/-) mice34, human apoCI-transgenic (APOC1) mice32, and wild-type (WT) littermates (all C57Bl/6 background) were housed at the breeding facility of TNO-Quality of Life in a temperature- and humidity-controlled environment and allowed free access to water and chow. All experiments were approved by the animal ethics committee of TNO. Experiments were conducted at 10-12 weeks of age unless stated otherwise.
Agarose Gel Electrophoresis − ReLPS (Sigma) and wtLPS (List) (both from Salmonella minnesota) were radioiodinated (specific activities 800-1000 cpm/
ng, respectively) and dialyzed against PBS pH 7.4. 125I-ReLPS or 125I-wtLPS (150 ng) was incubated (30 min at 37°C) with human isolated apoCI (purity ≥ 95%, Labconsult, Brussels, Belgium), synthesized apoCI or apoCI mutant (apoCImut; purity ≥ 95%, Protein Chemistry Technology Center, UT Southwestern Medical Center, Dallas, USA), and the mixtures were subjected to electrophoresis in an 0.75% (w/v) agarose gel28. 125I-activity was detected by phosphor imaging and apoCI by immunoblotting.
Binding of ApoCI to Immobilized LPS − EIA plates were coated with 0.3 mg/
mL LPS (Escherichia coli 0111:B4), washed, and blocked with 10 mg/mL BSA.
Duplicate control wells and LPS-coated wells were incubated with apoCI in 1 mg/mL BSA (1 h at 37°C). Bound apoCI was detected with an anti-human apoCI- biotin conjugated antibody (Academy Biomedical Company, Houston, Texas) for 1 h at room temperature followed by streptavidin-HRP for 30 min. Plates were developed with tetramethylbenzidine (TMB) substrate and read at 450 nm on an ELISA plate reader. The optical density resulting from specific apoCI binding to LPS was calculated by subtracting the readings of control wells (non-specific) from those of the LPS-coated wells at each apoCI concentration.
FITC-LPS Dequenching − ReLPS was fluorescently labeled with FITC with a molar labeling efficiency of 1:135. FITC-ReLPS (100 ng) was incubated (140 min at 37°C) with mouse plasma (0.125% v/v) in PBS pH 7.4 (total volume 100 µL), and fluorescence was measured before and after addition of 0.5%
sodium deoxycholate (to determine maximum dequenching) in a CytoFluor II Fluorescence Multi-Well Plate Reader (PerSeptive Biosystems) (λex 485 nm, λem 530 nm).
Kinetic Studies in Mice − 125I-LPS (10 µg/kg) was incubated in PBS pH 7.4 with isolated or synthesized human apoCI, apoCImut, isolated apoCIII (purity ≥ 95%, Labconsult, Brussels, Belgium), or BSA. 125I-wtLPS was incubated with synthesized apoCI or apoCImut. WT mice were anesthetized (dormicum, 5 mg/kg;
domitor, 0.5 mg/kg; fentanyl, 0.05 mg/kg), the abdomens were opened, and the incubation mixtures were injected via the vena cava inferior. After 30 min, blood plasma and liver samples were collected and counted for radioactivity36.
TNFα Induction − LPS (25 µg/kg) in PBS with 0.1% BSA was injected via the tail vein into conscious mice. After 60 min, plasma was taken, and TNFα was determined using the mouse TNFα-specific OptEIATM ELISA (BD Biosciences
Pharmingen) according to the manufacturer’s instructions. For in vitro studies, RAW 264.7 cells were seeded into 24-well plates (1x106 cells/well) and cultured overnight at 37°C in DMEM with 10% FBS. Cells were washed, incubated with LPS (S. minnesota or K. pneumoniae; 1 ng/mL) or the TLR2 agonist zymosan (100 ng/mL) with or without apoCI, apoCImut, or apoCIII in DMEM with 0.01%
human serum albumin (4 h at 37°C), and TNFα was determined in the medium.
Murine Sepsis Study − Mice (age 19-22 weeks; body weight 26-28 g) were anesthetized as described above. A 5 mm vertical incision was made in the center of the anterior neck to allow for visual inspection of subsequent intratracheal intubation followed by instillation of a Klebsiella pneumoniae suspension (ATCC 43816, capsular serotype 2; logarithmic growth phase; 500-750 colony forming units in 20 µL PBS). Following bacterial inoculation, mice received a subcutaneous injection of antisedan (2.5 mg/kg), anexate (0.5 mg/kg), and naloxon (1.2 mg/kg) to recover from anesthesia. In a first study, blood was taken after 24 h to determine plasma TNFα and macrophage migration inhibitory factor (MIF) by Western blotting37 and E-selectin by ELISA. After 48 h, another blood sample was taken to quantify Klebsiella pneumoniae in plasma by plating of serial dilutions on tryptone soy agar. Mortality was scored twice daily after bacterial inoculation, and the presence of only Klebsiella pneumoniae in lung and blood specimens of all succumbed mice was verified by culture on tryptone soy agar plates. The effect of Klebsiella pneumoniae inoculation on mortality was repeated in a subsequent experiment.
Statistical Analysis − Data are expressed as mean ± SEM. Statistical significance was assessed by two-tailed Mann-Whitney nonparametric test for two independent samples (SPSS version 11.5) unless stated otherwise.
Results
ApoCI Avidly Binds to LPS − To investigate whether apoCI could bind bacterial LPS, we first performed sequence alignment analysis between apoCI and established LPS-binding proteins, in which alternating sequences of cationic and hydrophobic amino acids are frequent38. ApoCI contains the most cationic lysine (K)-residues (i.e. 16 mol%) as compared to all other apolipoproteins, as well as a number of cationic/hydrophobic amino acid sequences throughout its structure.
Moreover, apoCI contains a KVKEKLK motif in its C-terminal domain (apoCI48-54) that is highly homologous to the LPS-binding sequence of the Limulus anti-LPS factor (LALF43-49; KWKYKGK)39 and the cationic antibacterial protein (CAP18117-
123; KIKEKLK)40. Interestingly, this KVKEKLK sequence belongs to one of the
Rf 1
Rf 0
ApoCI : LPS (mol/mol)
A B C
ApoCI (ng/mL)
apoc1-/- WT APOC1 0 50 100 150 200
0 .1 .2 .5 1 2 0.0 0.2 0.4 1.2 1.0 0.8 0.6
40
0 10 20 30
ApoCI binding (A450) FITC-LPS fluorescence (% increase)
apoCI apoCImut
*
*
three highly conserved amino acid intervals within apoCI among various species (i.e. human, baboon, dog, rat, and mouse) as shown by matrix/median filter analysis41.
Human apoCI was indeed able to fully disaggregate micelles of radiolabeled ReLPS [which is a truncated LPS containing the lipid A moiety and some 2- keto-3-deoxyoctonic acid (KDO) sugars] already at a 1:1 molar ratio, as evident from an LPS mobility shift assay on agarose gel (Fig. 1A). ApoCI colocalized with LPS (Rf 0.9) as evidenced by immunoblotting (not shown). To evaluate the contribution of the positively charged K residues in the putative LPS-binding motif KVKEKLK to the LPS-binding properties of apoCI, we synthesized the full-length (57 amino acids) mature apoCI peptide and an apoCImut in which the K residues were replaced by neutral alanine (A) residues (AVAEALA). This modification eliminates the positive charges within this domain without affecting the overall structure of apoCI. ApoCImut appeared far less efficient to bind immobilized LPS than apoCI (Fig. 1B), indicating that the KVKEKLK motif is indeed mainly responsible for the binding of LPS, and that the charged K residues are directly involved in LPS binding.
We next evaluated whether the apoCI content of mouse plasma affects LPS monomerization (Fig. 1C). We used FITC-labeled ReLPS, the fluorescence of which is strongly reduced (i.e. 93%) due to fluorescence self-quenching resulting from the micellar structure of LPS. On one hand, Incubation of FITC-ReLPS with serum from APOC1 mice resulted in 40% increased monomerization of ReLPS
Figure 1. ApoCI binds to LPS. 125I-ReLPS (A) (150 ng) was incubated (30 min at 37°C) with increasing amounts of human apoCI (ratio apoCI:LPS ranges from 0 to 2) and subjected to agarose gel electrophoresis. The resulting gel was scanned for 125I-activity by phosphor imaging. (B) LPS was coated onto EIA plates, incubated with apoCI (black circles) or apoCImut (white circles) in BSA (1 h at 37°C), and the binding of apoCI was quantified by spectrophotometry. (C) FITC-LPS (100 ng) was incubated (140 min at 37°C) with plasma (0125% v/v) from apoc1-/-, WT, and APOC1 mice (n=4), and fluorescence was determined. *P<005.
micelles as compared to serum WT mice, indicating that human apoCI competes with other serum components for LPS binding. This monomerization-enhancing effect is not due to the elevated VLDL levels per se as observed in APOC1 transgenic mice, since serum from apoE-deficient mice is 2-fold less effective in LPS monomerization studies, despite highly elevated VLDL levels (unpublished data). On the other hand, serum from apoc1-/- mice showed a 2-fold reduced LPS monomerization as compared to WT mice, whereas total lipoprotein and lipid levels were virtually identical. Mouse plasma apoCI, which also contains the KVKEKLK motif, is thus also able to effectively induce association of LPS with HDL.
ApoCI Prevents the Clearance of LPS by the Liver and Spleen − To establish whether the binding of apoCI to LPS has consequences for the metabolic fate of LPS, radiolabeled ReLPS was incubated with or without human apoCI and injected i.v. into WT mice (Fig. 2). ReLPS alone was rapidly cleared from serum, and predominantly associated with the liver (68%) and spleen (10%). ApoCI, isolated from human plasma, caused a dose-dependent inhibition of the serum clearance of LPS (Fig. 2A) by stimulating the association of LPS with HDL (not shown). This was accompanied by a strong reduction in the uptake of LPS by the liver (Fig. 2B) and spleen (Fig. 2C). At molar equilibrium, apoCI almost completely abolished the clearance of LPS from serum by the liver and spleen.
In fact, apoCI was already effective in significantly attenuating the clearance of a 5-fold molar surplus of LPS.
ApoCIII, which has a similar molecular weight, structure, and lipoprotein distribution pattern as apoCI, but does not contain typical LPS-binding sequences, only marginally affected the clearance of LPS as compared to apoCI. Whereas, at a 16-fold molar surplus, apoCI almost completely inhibited the uptake of LPS by the liver, apoCIII reduced the liver uptake by only 15% (Fig. 2E). A 16-fold molar surplus of BSA also had no effect on the clearance of LPS (Fig. 2D-F).
We next evaluated the contribution of the LPS-binding domain KVKEKLK to the effect of apoCI on the serum clearance of LPS by using synthesized apoCI and apoCImut (Fig. 3). Synthesized wild-type apoCI had a similar effect on ReLPS kinetics (Fig. 3A-C) as isolated apoCI (Fig. 2A-C). In addition to ReLPS, apoCI also prevented the clearance of wtLPS from serum by the liver and spleen (Fig.
3D-F). Compared to apoCI, apoCImut was much less efficient in preventing the serum clearance of both ReLPS (Fig. 3A-C) and wtLPS (Fig. 3D-F), indicating the crucial involvement of the LPS-binding domain KVKEKLK.
ApoCI Stimulates the LPS-induced Proinflammatory Response − We next evaluated the effect of apoCI on the LPS-induced proinflammatory response
(Fig. 4). I.v. injection of free ReLPS into apoc1-/-, WT, and APOC1 mice resulted in an apoCI-dependent increase in plasma TNFα levels (Fig. 4A). Apoc1-/- mice showed a trend towards 2-fold lower TNFα levels as compared to WT mice, and APOC1 mice showed 7-fold higher plasma TNFα levels as compared to WT mice. This effect was not caused by the elevated plasma lipid levels in APOC1 mice, since hyperlipidemic LDL receptor (LDLr) and LDLr-related protein (LRP) double-deficient mice42 did not show an enhanced TNFα response to LPS as compared to WT mice (2.9±1.1 vs 3.2±1.0 x103 pg/mL, respectively), despite a more extreme hyperlipidemia as compared to APOC1 mice31. Likewise, incubation of LPS with apoCI before injection into mice enhanced the LPS- induced production of TNFα 3.8-fold as compared to injection of LPS alone (not shown). Subsequent in vitro studies showed that apoCI greatly stimulates the TNFα response to ReLPS (Fig. 4B) and wtLPS (Fig. 4C) in RAW 264.7 mouse macrophages. ApoCIII stimulated the LPS-induced TNFα response to only a
A Serum B C
20 0 125I-ReLPS (% of dose)
Liver Spleen
*
* *
*
* *
40 60 80 100
20 0 40 60 80
0 2 4 6 8 10 12
0 .2 .5 1
ApoCI : LPS (mol/mol)0
.2 .5 1 0 .2 .5 1
D Serum E F
20 0 125I-ReLPS (% of dose)
Liver Spleen
*
* *
40 60 80 100
20 0 40 60 80
0 2 4 6 8 10 12
ApoCI − + − − − + − − − + − − nd ApoCIII − − + − − − + − − − + − BSA − − − + − − − + − − − +
Figure 2. ApoCI dose-dependently prevents the serum clearance of LPS by the liver and spleen. 125I- ReLPS (10 µg/kg) was injected i.v. into WT mice without and with preincubation with isolated apoCI (molar ratio apoCI: LPS = 0-1) (A-C) or with isolated apoCI, apoCIII, and BSA (molar ratio protein:
LPS = 16) (D-F) (n=2-4). At 30 min after injection, radioactivity was determined in serum (A, D), liver (B, E), and spleen (C, F). *P<0.05. nd, not determined.
small extent (Fig. 4B-C). Apolipoproteins alone did not trigger TNFα production (Fig. 4B). Again, the LPS-binding domain KVKEKLK is crucial for the enhanced response to LPS, since apoCImut was ineffective. The effect of apoCI was also specific for LPS, since apoCI did not stimulate the response to the TLR2 agonist zymosan (Fig. 4D).
ApoCI Increases the Inflammatory Response to Klebsiella pneumoniae and Protects Against Fatal Sepsis − We next investigated whether an apoCI-dependent increased inflammatory response is associated with a higher rescue rate from death due to sepsis. To do so, apoc1-/-, WT, and APOC1 mice were challenged intrapulmonally with Klebsiella pneumoniae, an established procedure leading to Gram-negative bacterial pneumonia, sepsis, and eventually death43. At 24 h after inoculation of 500-750 colony forming units of Klebsiella pneumoniae, the apoCI status of the mice positively correlated with markers
Figure 3. The effect of apoCI on the serum clearance of LPS involves the LPS-binding motif KVKEKLK. 125I-ReLPS (A-C) or 125I-wtLPS (D-F) (10 µg/kg) was injected i.v. into WT mice without and with preincubation with synthesized apoCI or apoCImut (molar ratio apoCI:LPS = 1) (n=2-4). At 30 min after injection, radioactivity was determined in serum (A, D), liver (B, E), and spleen (C, F).
*P<0.05.
A Serum B C
20 125I-ReLPS (% of dose) 0
Liver Spleen
* * * * * *
40 60 80
20 0 40 60 80
0 2 4 6 8
ApoCImut − − + − − + − − + ApoCI − + − − + − − + −
D Serum E F
20 125I-ReLPS (% of dose) 0
Liver Spleen
* * * * * *
40 60 80
20 0 40 60 80
0 2 4 6 8
ApoCImut − − + − − + − − + ApoCI − + − − + − − + −
of the inflammatory response, i.e. the plasma levels of TNFα, MIF (a TNFα- inducible pro-inflammatory cytokine which is more stable than TNFα itself), and E-selectin (a TNFα-inducible inflammation marker) (Fig. 5A-C). At 48 h after inoculation, the number of circulating Klebsiella pneumoniae was strongly and dose-dependently decreased by apoCI. APOC1 mice displayed a 25-fold lower bacterial count as compared with apoc1-/- mice (Fig. 5D). In line with our in vitro findings, these data point to an enhanced activation of macrophages by apoCI during Gram-negative sepsis, as judged from the collective rise in TNFα, MIF, and E-selectin, and show that the apoCI-dependent increased proinflammatory
Figure 4. ApoCI enhances the LPS-triggered TNFα induction. (A) LPS (25 µg/kg) was injected into apoc1-/-, WT, or APOC1 mice (n=6) and TNFα was determined in plasma after 60 min. (B-D) RAW 2647 cells were incubated (4 h at 37°C) with ReLPS (1 ng/mL) (B), wtLPS (1 ng/mL) (C), or zymosan (ZS; 100 ng/mL) (D) in DMEM supplemented with 001% human SA, without or with preincubation (30 min at 37°C) with apoCI, apoCImut, or apoCIII (molar ratio apo: ReLPS = 10:1; apo: wtLPS = 100:1;
apo: zymosan = 3450:1), and TNFα was determined in the medium (n=4). *P<0.05, **P<0.01. nd, not detected.
D
B ReLPS C
Medium TNFα (x 103 pg/mL)
wtLPS Zymosan
* *
*
0 2 1 3 4
ApoCI − + − − + − − − + − − − + − − nd
ApoCImut − − + − − + − − − + − − − + − ApoCIII − − − + − − + − − − + − − − +
* nd nd
LPS + + + + − − − 0 + + + + + + + + 2
1 3 4
0 0.2 0.1 0.3 0.4 Plasma TNFα (x 103 pg/mL)
A
32 24 16 8 0
** **
apoc1-/- WT APOC1 +LPS +LPS +LPS
status is associated with a reduced bacterial outgrowth to plasma. Accordingly, apoCI protected mice against fatal sepsis. After bacterial inoculation in WT mice, a gradual loss of animals occurred, with 39% survival after 2 weeks. Whereas apoc1-/- mice were more susceptible to death following sepsis (only 16% survival), expression of human apoCI in APOC1 mice improved survival (53% survival) (Fig. 5E). All mice that died had ongoing infection, as verified by Klebsiella pneumoniae-positive cultures from lungs and blood obtained postmortem.
Together, these data indicate that elevated apoCI levels are associated with an enhanced inflammatory response to LPS, a more efficient bacterial killing, and that apoCI can prevent mice from septic death.
B
1.0
Plasma TNFα (relative to WT)
WT, sham apoc1-/- WT APOC1
A C
Plasma MIF (relative to WT) Plasma E-selectin (ng/mL)
apoc1-/- WT APOC1 apoc1-/- WT APOC1
Survival of mice (%)
Blood K. pneumoniae (x103 cfu/mlL)
D E
apoc1-/- WT APOC1
**
**
*
*
**
***
APOC1 WT apoc1-/-
Time after inoculation (days) 1.5
0.5 0
2.0 1.5
1.0 0.5
0 0
250 200 150 100 50
500 400 300 200 100 0
100 80
40 60
20
00 2 4 6 8 10 12 14
Figure 5. ApoCI associates with increased inflammatory response and survival of mice with experimental sepsis. (A-D) Apoc1-/- (n=13), WT (n=15), and APOC1 mice (n=11) were inoculated intratracheally with 500-750 cfu of Klebsiella pneumoniae. (A-C) After 24 h, plasma levels of TNFα (pooled plasma, immunoblot), MIF (individual mice, immunoblot), and E-selectin (individual mice, ELISA) were determined. *P<005, **P<001. Inserts show immunoblots of pooled plasma. (D) After 48 h, plasma levels of Klebsiella pneumoniae were determined. **P<001. (E) Apoc1-/- (white circles, n=25), WT (black circles, n=29), and APOC1 (black square, n=19) mice (i.e. including the mice described above and a second set of mice) were inoculated intratracheally with the same dose of Klebsiella pneumoniae or saline (white squares, WT mice, n=6) and their two-week survival was assessed. *P<005, **P<001 (log-rank test, Graph Pad software Inc).
Discussion
Until recently, apolipoproteins have been assigned classical roles in mediating certain aspects of lipoprotein metabolism, either by affecting plasma factors that modify lipoprotein composition, or by mediating the uptake of lipoproteins or their constituents by various tissues. ApoCI is the smallest apolipoprotein, consisting of only 57 amino acids, and is unusually rich in lysine residues. We and others have described that apoCI attenuates TG-rich lipoprotein clearance, either by inhibiting the peripheral lipoprotein lipase-mediated TG hydrolysis31 or by interfering with the hepatic uptake of these lipoproteins32,44, but the physiological relevance of such an action is unclear. The structure of apoCI is remarkably conserved among species (i.e. mouse, rat, dog, baboon, human), with three highly conserved sequences41, and no structural mutations have been reported in humans thus far. These findings suggest that apoCI could have a specific function, which should not necessarily relate to lipid metabolism. In the present study, we report that apoCI does have an important function in facilitating the inflammatory response to Gram-negative bacterial infections.
We discovered by sequence alignment analysis that apoCI contains a number of alternating sequences of cationic and hydrophobic residues that are frequent in established LPS-binding proteins38. In particular, one of the highly conserved intervals within the C-terminal domain of apoCI appears to contain the LPS-binding motif KVKEKLK (apoCI48-54), which is identical between mice and humans. Indeed, human apoCI avidly bound to LPS in vitro, and the resulting complex was resistant to dissociation in the blood despite the abundance of potentially destabilizing plasma transfer factors (e.g. LPS-binding protein, phospholipids transfer protein). As a consequence, upon i.v. injection in mice, apoCI markedly prolonged the residence time of LPS in serum by association with long-circulating HDL, reaching a maximum effect already at an apoCI:LPS
= 1:1 molar ratio. Importantly, the LPS-monomerizing capacity of murine plasma was increased by the presence of moderate levels of human apoCI (i.e. 35 mg/
dL in APOC1 mice as compared to 10 mg/dL in human plasma31). Reciprocally, the monomerization of LPS was decreased in the absence of apoCI, indicating that the endogenous plasma level of apoCI per se can have a profound impact on the LPS distribution in vivo.
Physiological amounts of both exogenous and endogenous apoCI strongly enhanced the TNFα production upon i.v. LPS administration in mice. These TNFα- increasing effects are not caused by the effects of apoCI on plasma lipid levels.
APOC1 mice are hyperlipidemic, but an even more pronounced hyperlipidemia as observed in LDLr/LRP double-deficient mice did not increase the LPS-induced TNFα response. In fact, we showed that apoCI can directly stimulate the LPS-
induced TNFα production by macrophages in vitro. The increased apoCI-induced inflammatory response to LPS may thus be caused by a prolonged residence of LPS in the serum and/or direct activation of macrophages.
ApoCI was able to affect the serum clearance and TNFα-stimulating properties of both full-length wtLPS (composed of the toxic lipid A moiety, as well as the inner core, outer core, and O-antigen) and truncated ReLPS (merely composed of the lipid A moiety and some KDO sugars), indicating that the lipid A/KDO moiety of the LPS molecule contains the crucial elements for interaction with apoCI. Mutation of lysine residues to alanine residues within the KVKEKLK domain of apoCI markedly reduced these LPS-modulating effects of apoCI without affecting the helical structure, indicating that this highly conserved domain within the C-terminal helix of apoCI is indeed responsible for LPS binding. Taken together, these data suggest that the interaction between LPS and apoCI is likely to involve electrostatic interaction between the lysine residues within apoCI and electronegative elements (presumably phosphate groups) within the lipid A moiety of LPS. Since lipid A is the common determinant of LPS molecules from all bacterial species, apoCI is likely to bind a wide array of wild-type and mutant LPS molecules. The fact that apoCI is able to enhance the binding of LPS to HDL further indicates that such an electrostatic interaction between apoCI and LPS can still occur on the lipoprotein surface.
Since sepsis has initially been regarded as an excessive systemic proinflammatory response to infections as largely mediated by TNFα, anti- inflammatory strategies e.g. aimed to neutralize TNFα have been widely employed as potential therapeutic tools for the treatment of sepsis20. However, although such approaches have been successful to inhibit LPS-induced toxicity and mortality, a recent metaregression analysis of 22 clinical sepsis trials with anti-inflammatory agents, including 9 trials with anti-TNFα antibodies, showed that beneficial therapeutic effects of anti-inflammatory therapies could only be demonstrated in patients with a high risk of death, whereas anti-inflammatory agents were harmful in those patients with a low mortality risk20. Therefore, we hypothesized that the observed apoCI-mediated increased inflammatory reaction to LPS may enhance survival from Gram-negative infections as a result of a timely and effective host response.
As a model for human sepsis caused by Gram-negative bacteria, we applied an experimental pneumonia model in which a local Klebsiella pneumoniae infection causes a lethal sepsis43. By using intratracheal inoculation, the bacterial inoculum can be precisely controlled which guarantees a reproducible infectious disease model with low variation between animals43. The protective role of LPS- induced TLR4 signaling in this septic model has been conclusively established,
as TLR4-deficiency shortened survival together with an enhanced bacterial outgrowth5,6. This model is advantageous over cecal ligation and puncture (CLP)-induced polymicrobial septic peritonitis, in which the survival of mice is independent of the ability to respond to LPS, as the absence of TLR4 did not affect survival45. The pneumonia-related sepsis model is also preferred over other models that employ bolus i.v. infusion of large amounts of live bacteria or LPS, leading to conditions of intoxication rather than sepsis46,47.
Consistent with our finding that apoCI directly stimulated the LPS-induced TNFα response by macrophages in vitro, apoCI expression in mice also dose- dependently increased the initial proinflammatory response towards Klebsiella pneumoniae, as judged from increased plasma inflammation markers (TNFα, MIF, E-selectin) reflecting an activation of macrophages. These effects were accompanied by lower bacterial counts in plasma and reduced mortality resulting from sepsis, confirming that an increased early inflammatory response effectuates an efficient anti-bacterial response. ApoCI mainly reduced mortality in the early phase after infection (i.e. within 4 days), during which the anti-bacterial attack is mainly mediated by a non-specific cellular response. By inducing selective granulocytopenia in mice, it has previously been demonstrated that circulating granulocytes play an important role in the defense against K. pneumoniae infection in the lung48. In fact, treatment of K. pneumoniae-infected C57Bl/6 mice with cyclophosphamide, leading to depletion of granulocytes, increased the mortality rate from approx. 20% to nearly 90% within the first 4 days after infection49. Collectively, these data thus strongly suggest a crucial involvement of granulocytes in the apoCI-stimulated antibacterial attack.
MIF has previously been implicated as a critical mediator of septic shock50. On the other hand, MIF has been shown to facilitate the detection of LPS-containing bacteria, enabling cells that are at the forefront of the host antimicrobial defense system, such as macrophages, to respond rapidly to invasive bacteria by rapid production of proinflammatory cytokines51. The present study indeed shows that elevated levels of MIF during the initial response to bacteria are associated with increased survival. These seemingly conflicting data are inherent to the complex etiology of sepsis: whereas a proinflammatory response is often harmful in severe sepsis, an early inflammatory response is crucial to combat bacterial infections24.
Based on a large body of evidence, current models of innate immune defenses postulate a central role for TLR-mediated processes in pathogen detection, initiation of a rapid immune defense, and regulation of subsequent adaptive immune responses5. Our observation that apoCI increases the inflammatory response to LPS and Klebsiella pneumoniae, and reduces mortality
in Klebsiella pneumoniae-induced sepsis fully supports the concept that an adequate immune response is essential to protect mice against mortality from Gram-negative infections4-10, and suggests that apoCI may play a central role in innate immunity.
Currently, the synthetic lipid A analogue E5564 that has been designed to antagonize the effects of LPS by interacting with TLR4, is clinically being investigated for the treatment of severe sepsis and septic shock. E5564 has been demonstrated to block the induction of LPS-induced cytokines as well as lethality in LPS-sensitized mice upon intraperitoneal injection of E. coli, which led to rapid death of the vast majority (85%) of mice within 12 hours54. Likewise, a high dose of E5564 reduced the rapidly induced mortality of rats induced by intravascularly injected E. coli55. These data confirm that anti-inflammatory strategies may be beneficial in preventing lethality due to acute and severe inflammation in acute toxicity models such as intraperitoneally54 and intravascularly55 administered E.
coli, resulting in death of the majority of animals within 24 hours after injection.
However, a high dose of E5564 did not rescue rats from a clinically more relevant infection model in which E. coli was administered extravascularly55, and even showed a slight trend towards higher mortality, confirming that TLR4 antagonism at an early stage of an actual infection model may be harmful. Clearly, these studies confirm the importance of route of infection and timing of treatment, and suggest that the benefit of apoCI may be limited to early stages of infection.
Based on our present data, we propose the following model for the protective effects of apoCI in Gram-negative sepsis. Upon the entry and proliferation of bacteria in the blood, LPS is released into the plasma and binds to apoCI, which involves the interaction between the LPS-binding motif KVKEKLK within apoCI and presumably phosphate groups within lipid A, the shared moiety of all LPS species. Although apoCI may bind LPS in the lipid-free and lipid-bound state, the resulting apoCI-LPS complex is mainly associated with HDL, which is the main carrier of apoCI in mice and humans56. ApoCI effectively presents the LPS to responsive cells such as macrophages (the mechanistic basis of which is under current investigation), leading to a rapid and enhanced production of proinflammatory cytokines, among which TNFα and MIF. These cytokines are essential for effective eradication of the bacterial infection thereby preventing infection-related mortality. Therefore, plasma apoCI protects against fatal sepsis by effectuating an early and adequate anti-bacterial response. We speculate that apoCI may provide a therapeutic handle in the ongoing search for strategies that are aimed to prevent or treat sepsis at an early stage.
Acknowledgements
This study was supported in part by the Netherlands Organization for Scientific Research (NWO RIDE 014-90-001 to LMH, NWO VENI 016.036.061 to RK, NWO VIDI 917-36-351 to PCNR), by the Leiden University Medical Center (Gisela Thier Fellowship to PCNR), by ZorgOnderzoek Nederland, formerly the Dutch Foundation for Preventive Medicine PraeventieFonds (grant 28-2875,23 to JTvD), and by the NIAID (grant AI045896 to RLK). Funders did not have any role in the design, analysis, interpretation and report of the present study.
The authors have declared that no relationships exist may pose a conflict of interest.
We thank Elly P. de Wit (TNO-Quality of Life), Marian T. ten Kate (Erasmus Medical Center), and Jason Gillman (UT Southwestern Medical Center) for excellent technical assistance.
References
1. Martin GS, Mannino DM, EatonS, Moss M. The epidemiology of sepsis in the United States from 1979 through 2000. N Engl J Med 2003;348:1546-1554.
2. Riedemann NC, Guo RF, Ward PA. Novel strategies for the treatment of sepsis. Nat Med 2003;9:517-524.
3. Poltorak A, He X, Smirnova I, Liu MY, van Huffel C, Du X, Birdwell D, Alejos E, Silva M, Galanos C, Freudenberg M, Ricciardi-Castagnoli P, Layton B, Beutler B. Defective LPS signaling in C3H/HeJ and C57BL/10ScCr mice: mutations in Tlr4 gene. Science 1998;282:2085-2088.
4. Hagberg L, Hull R, Hull S, McGhee JR, Michalek SM, Svanborg Eden C. Difference in susceptibility to gram-negative urinary tract infection between C3H/HeJ and C3H/HeN mice. Infect Immun 1984;46:839-844.
5. Branger J, Knapp S, Weijer S, Leemans JC, Pater JM, Speelman P, Florquin S, van der Poll T. Role of Toll-like receptor 4 in gram-positive and gram-negative pneumonia in mice. Infect Immun 2004;72:788-794.
6. Schurr JR, Young E, Byrne P, Steele C, Shellito JE, Kolls JK. Central role of toll-like receptor 4 signaling and host defense in experimental pneumonia caused by Gram-negative bacteria.
Infect Immun 2005;73:532-545.
7. Wang X, Moser C, Louboutin JP, Lysenko ES, Weiner DJ, Weiser JN, Wilson JM. Toll-like receptor 4 mediates innate immune responses to Haemophilus influenzae infection in mouse lung. J Immunol 2002;168:810-815.
8. Wang M, Jeng KC, Ping LI. Exogenous cytokine modulation or neutralization of interleukin- 10 enhance survival in lipopolysaccharide-hyporesponsive C3H/HeJ mice with Klebsiella infection. Immunology 1999;98:90-97.
9. Woods JP, Frelinger JA, Warrack G, Cannon JG. Mouse genetic locus Lps influences susceptibility to Neisseria meningitidis infection. Infect Immun 1988;56;1950-1955.
10. Bernheiden M, Heinrich JM, Minigo G, Schutt C, Stelter F, Freeman M, Golenbock D, Jack RS. LBP, CD14, TLR4 and the murine innate immune response to a peritoneal Salmonella infection. J Endotoxin Res 2001;7:447-450.
11. Echtenacher B, Falk W, Mannel DN, Krammer PH. Requirement of endogenous tumor necrosis factor/cachectin for recovery from experimental peritonitis. J Immunol 1990;145:3762-3766.
12. Bagby GJ, Plessala KJ, Wilson LA, Thompson JJ, Nelson S. Divergent efficacy of antibody to tumor necrosis factor-alpha in intravascular and peritonitis models of sepsis. J Infect Dis 1990;163:83-88.
13. Alexander HR, Sheppard BC, Jensen JC, Langstein HN, Buresh CM, Venzon D, Walker EC, Fraker DL, Stovroff MC, Norton JA. Treatment with recombinant human tumor necrosis factor- alpha protects rats against the lethality, hypotension, and hypothermia of gram-negative sepsis. J Clin Invest 1991;88:34-39.
14. Mastroeni P, Arena A, Costa GB, Liberto MC, Bonina L, Hormaeche CE. Serum TNF alpha in mouse typhoid and enhancement of a Salmonella infection by anti-TNF alpha antibodies.
Microb Pathog 1991;11:33-38.
15. Rothe J, Lesslauer W, Lotscher H, Lang Y, Koebel P, Kontgen F, Althage A, Zinkernagel R, Steinmetz M, Bluethmann H. Mice lacking the tumour necrosis factor receptor 1 are resistant to TNF-mediated toxicity but highly susceptible to infection by Listeria monocytogenes. Nature 1993;364:798-802.
16. van Furth R, van Zwet TL, Buisman AM, van Dissel JT. Anti-tumor necrosis factor antibodies inhibit the influx of granulocytes and monocytes into an inflammatory exudate and enhance the growth of Listeria monocytogenes in various organs. J Infect Dis 1994;170:234-237.
17. van der Poll T, Keogh CV, Buurman WA, Lowry SF. Passive immunization against tumor necrosis factor-alpha impairs host defense during pneumococcal pneumonia in mice. Am J Respir Crit Care Med 1997;155:603-608.
18. Nara C, Tateda K, Matsumoto T, Ohara A, Miyazaki S, Standiford TJ, Yamaguchi K. Legionella- induced acute lung injury in the setting of hyperoxia: protective role of tumour necrosis factor- alpha. J Med Microbiol 2004;53:727-733.
19. Moore TA, Lau HY, Cogen AL, Monteleon CL, Standiford TJ. Anti-tumor necrosis factor-alpha therapy during murine Klebsiella pneumoniae bacteremia: increased mortality in the absence of liver injury. Shock 2003;20:309-315.
20. Eichacker PQ, Parent C, Kalil A, Esposito C, Cui X, Banks SM, Gerstenberger EP, Fitz Y, Danner RL, Natanson C. Risk and the efficacy of antiinflammatory agents: retrospective and confirmatory studies of sepsis. Am J Respir Crit Care Med 2002;166:1197-1205.
21. Westendorp RG, Langermans JA, de Bel CE, Meinders AE, Vandenbroucke JP, van Furth R, van Dissel JT. Release of tumor necrosis factor: an innate host characteristic that may contribute to the outcome of meningococcal disease. J Infect Dis 1995;171:1057-1060.
22. Westendorp RG, Langermans JA, Huizinga TW, Elouali AH, Verweij CL, Boomsma DI, Vandenbroucke JP, Vandenbrouke JP. Genetic influence on cytokine production and fatal meningococcal disease. Lancet 1997;349:170-173.
23. van Dissel JT, van Langevelde P, Westendorp RG, Kwappenberg K, Frolich M. Anti-inflammatory cytokine profile and mortality in febrile patients. Lancet 1998;351:950-953.
24. Netea MG, van der Meer JW, van Deuren M, Kullberg BJ. Proinflammatory cytokines and sepsis syndrome: not enough, or too much of a good thing? Trends Immunol 2003;24:254-258.
25. Harris HW, Gosnell JE, Kumwenda ZL. The lipemia of sepsis: triglyceride-rich lipoproteins as agents of innate immunity. J Endotoxin Res 2000;6:421-430.
26. Flegel WA, Baumstark MW, Weinstock C, Berg A, Northoff H. Prevention of endotoxin-induced monokine release by human low- and high-density lipoproteins and by apolipoprotein A-I.
Infect Immun 1993;61:5140-5146.
27. Harris HW, Grunfeld C, Feingold KR, Rapp JH. Human very low density lipoproteins and chylomicrons can protect against endotoxin-induced death in mice. J Clin Invest 1990;86:696- 28. Rensen PC, Oosten M, Bilt E, van Eck M, Kuiper J, Berkel TJ. Human recombinant apolipoprotein 702.
E redirects lipopolysaccharide from Kupffer cells to liver parenchymal cells in rats In vivo. J Clin Invest 1997;99;2438-2445.
29. van Oosten M, Rensen PC, van Amersfoort ES, van Eck M, van Dam AM, Breve JJ, Vogel T, Panet A, van Berkel TJ, Kuiper J. Apolipoprotein E protects against bacterial lipopolysaccharide- induced lethality. A new therapeutic approach to treat gram-negative sepsis. J Biol Chem 2001;276:8820-8824.
30. Mahley RW. Apolipoprotein E: cholesterol transport protein with expanding role in cell biology.
Science 1988;240:622-630.
31. Berbee JF, van der Hoogt CC, Sundararaman D, Havekes LM, Rensen PC. Severe hypertriglyceridemia in human APOC1 transgenic mice is caused by apoC-I-induced inhibition of LPL. J Lipid Res 2005;46:297-306.
32. Jong MC, Dahlmans VE, van Gorp PJ, van Dijk KW, Breuer ML, Hofker MH, Havekes LM. In the absence of the low density lipoprotein receptor, human apolipoprotein C1 overexpression in transgenic mice inhibits the hepatic uptake of very low density lipoproteins via a receptor- associated protein-sensitive pathway. J Clin Invest 1996;98:2259-2267.
33. Barlage S, Frohlich D, Bottcher A, Jauhiainen M, Muller HP, Noetzel F, Rothe G, Schutt C, Linke RP, Lackner KJ, Ehnholm C, Schmitz G. ApoE-containing high density lipoproteins and phospholipid transfer protein activity increase in patients with a systemic inflammatory response. J Lipid Res 2001;42:281-290.
34 van Ree JH, Hofker MH, van den Broek WJ, van Deursen JM, van der Boom H, Frants RR,
Wieringa B, Havekes LM. Increased response to cholesterol feeding in apolipoprotein C1- deficient mice. Biochem J 1995;305(Pt 3):905-911.
35 Troelstra A, Antal-Szalmas P, de Graaf-Miltenburg LA, Weersink AJ, Verhoef J, van Kessel KP, van Strijp JA. Saturable CD14-dependent binding of fluorescein-labeled lipopolysaccharide to human monocytes. Infect Immun 1997;65:2272-2277.
36 Rensen PC, Herijgers N, Netscher MH, Meskers SC, van Eck M, van Berkel, TJ. Particle size determines the specificity of apolipoprotein E-containing triglyceride-rich emulsions for the LDL receptor versus hepatic remnant receptor in vivo. J Lipid Res 1997;38:1070-1084.
37 Kleemann R, Gervois PP, Verschuren L, Staels B, Princen HM, Kooistra T. Fibrates down-regulate IL-1-stimulated C-reactive protein gene expression in hepatocytes by reducing nuclear p50- NFkappa B-C/EBP-beta complex formation. Blood 2003;101:545-551.
38 de Haas CJ, van der Zee R, Benaissa-Trouw B, van Kessel KP, Verhoef J, van Strijp, JA.
Lipopolysaccharide (LPS)-binding synthetic peptides derived from serum amyloid P component neutralize LPS. Infect Immun 1999;67:2790-2796.
39 Hoess A, Watson S, Siber GR, Liddington R. Crystal structure of an endotoxin-neutralizing protein from the horseshoe crab, Limulus anti-LPS factor, at 15 A resolution. Embo J 1993;12:3351- 3356.
40 Larrick JW, Hirata M, Zheng H, Zhong J, Bolin D, Cavaillon JM, Warren HS, Wright SC. A novel granulocyte-derived peptide with lipopolysaccharide-neutralizing activity.J Immunol 1994;152:231-240.
41 Weinberg RB. Identification of functional domains in the plasma apolipoproteins by analysis of inter-species sequence variability. J Lipid Res 1994;35:2212-2222.
42 Espirito Santo SM, Rensen PC, Goudriaan JR, Bensadoun A, Bovenschen N, Voshol PJ, Havekes LM, van Vlijmen BJ. Triglyceride-rich lipoprotein metabolism in unique VLDL receptor, LDL receptor, and LRP triple-deficient mice. J Lipid Res 2005;46:1097-1102.
43 Bakker-Woudenberg IA. Experimental models of pulmonary infection. J Microbiol Methods 2003;54:295-313.
44 Shachter NS, Ebara T, Ramakrishnan R, Steiner G, Breslow JL, Ginsberg HN, Smith JD.
Combined hyperlipidemia in transgenic mice overexpressing human apolipoprotein Cl. J Clin Invest 1996;98:846-855.
45 Echtenacher B, Freudenberg MA, Jack RS, Mannel DN. Differences in innate defense mechanisms in endotoxemia and polymicrobial septic peritonitis. Infect Immun 2001;69:7271- 7276.
46 Deitch EA. Animal models of sepsis and shock: a review and lessons learned. Shock 1998;9:1- 47 Cross AS, Opal SM, Sadoff JC, Gemski P. Choice of bacteria in animal models of sepsis. Infect 11.
Immun 1993;61:2741-2747.
48 Rehm SR, Gross GN, Pierce AK. Early bacterial clearance from murine lungsSpecies-dependent phagocyte response. J Clin Invest 1980;66:194-199.
49 de Bont N, Netea MG, Demacker PN, Kullberg BJ, van der Meer JW, Stalenhoef AF. Apolipoprotein E-deficient mice have an impaired immune response to Klebsiella pneumoniae. Eur J Clin Invest 2000;30:818-822.
50 Calandra T, Echtenacher B, Roy DL, Pugin J, Metz CN, Hultner L, Heumann D, Mannel D, Bucala R, Glauser, MP. Protection from septic shock by neutralization of macrophage migration inhibitory factor. Nat Med 2000;6:164-170.
51 Calandra T, Roger T. Macrophage migration inhibitory factor: a regulator of innate immunity. Nat Rev Immunol 2003;3:791-800.
52 Akira S, Takeda K. Toll-like receptor signalling. Nat Rev Immunol 2004;4:499-511.
53 Beutler B. Inferences, questions and possibilities in Toll-like receptor signalling. Nature 2004;430:257-263.
54 Mullarkey M, Rose JR, Bristol J, Kawata T, Kimura A, Kobayashi S, Przetak M, Chow J, Gusovsky F, Christ WJ, Rossignol DP. Inhibition of endotoxin response by e5564, a novel Toll- like receptor 4-directed endotoxin antagonist. J Pharmacol Exp Ther 2003;304:1093-1102.
55 Solomon SB, Cui X, Gerstenberger E, Danner RL, Fitz Y, Banks SM, Natanson C, Eichacker PQ.
Effective dosing of lipid A analogue E5564 in rats depends on the timing of treatment and the route of Escherichia coli infection. J Infect Dis 2006;193:634-644.
56 Cohn JS, Tremblay M, Boulet L, Jacques H, Davignon J, Roy M, Bernier L. Plasma concentration and lipoprotein distribution of ApoC-I is dependent on ApoE genotype rather than the Hpa I ApoC-I promoter polymorphism. Atherosclerosis 2003;169:63-70.
