Synergistic Inflammation Is Induced by Blood Degradation Products with Microbial Toll-Like Receptor Agonists and Is Blocked by Hemopexin

Abstract

Detection of microbial components by immune cells via Toll-like receptors (TLRs) with subsequent induction of inflammation
is essential for host defense. However, an overactive immune response can cause tissue damage and sepsis. The endogenous molecule
hemoglobin and its derivative heme are often released into tissue compartments where there is infection in the presence of
degrading blood. We found that hemoglobin synergizes with multiple TLR agonists to induce high levels of tumor necrosis factor
and interleukin-6 from macrophages and that this synergy is independent of TLR4 and MyD88. In contrast, heme synergized with
some but not all TLR agonists studied. Furthermore, the synergy of both hemoglobin and heme with lipopolysaccharide was suppressed
by hemopexin, a plasma heme-binding protein. These studies suggest that hemoglobin and heme may substantially contribute to
microbe-induced inflammation when bacterial or viral infection coexists with blood degradation and that hemopexin may play
a role in controlling inflammation in such settings.

Full text

Synergistic Inflammation Is Induced by Blood Degradation
Products with Microbial Toll-Like Receptor Agonists and Is
Blocked by Hemopexin
Tian Lin1,a, Young Ho Kwak1,a, Fatima Sammy1, Ping He1, Sujatha Thundivalappil1,
Guangjie Sun1, Wei Chao2, and H. Shaw Warren1
1 Infectious Disease Unit, Massachusetts General Hospital, Boston
2 Department of Anesthesia and Critical Care, Massachusetts General Hospital, Boston
Abstract
Detection of microbial components by immune cells via Toll-like receptors (TLRs) with
subsequent induction of inflammation is essential for host defense. However, an overactive
immune response can cause tissue damage and sepsis. The endogenous molecule hemoglobin and
its derivative heme are often released into tissue compartments where there is infection in the
presence of degrading blood. We found that hemoglobin synergizes with multiple TLR agonists to
induce high levels of tumor necrosis factor and interleukin-6 from macrophages and that this
synergy is independent of TLR4 and MyD88. In contrast, heme synergized with some but not all
TLR agonists studied. Furthermore, the synergy of both hemoglobin and heme with lipo-
polysaccharide was suppressed by hemopexin, a plasma heme-binding protein. These studies
suggest that hemoglobin and heme may substantially contribute to microbe-induced inflammation
when bacterial or viral infection coexists with blood degradation and that hemopexin may play a
role in controlling inflammation in such settings.
Much of the pathophysiology that occurs early and during microbial infection is believed to
be due to the induction of inflammation in tissues that has evolved as an essential part of the
defense against early microbial challenge. Integral to this concept is the early identification
of microbes in tissues by specialized cells, such as macrophages, with the production of
secondary mediators, such as cytokines, that amplify the signal and communicate with other
local and distant tissues. It is now appreciated that certain molecules on microorganisms
known as microbial-associated molecular pattern molecules (MAMPs) interact with a
limited number of pattern recognition receptors called Toll-like receptors (TLRs), to initiate
a cascade of events that ultimately result in a signal being transmitted to the nucleus to
produce cytokines. Each TLR has receptor-specific ligands, such as lipopolysaccharide
(LPS), which signals through TLR4; lipoteichoic acid and peptidoglycan, which signal
through TLR2; viral nucleic acid structures, such as double-stranded RNA poly I: C, which
signals through TLR3; the guanosine analogue loxoribine, which signals through TLR7; and
bacterial DNA (CpG), which signals through TLR9 [1,2].
In some clinical situations, a hyperactive immune response can cause tissue damage. Sepsis
syndrome, which is defined by certain parameters of systemic inflammation in the setting of
Reprints or correspondence: H. Shaw Warren, Infectious Disease Unit, Massachusetts General Hospital East, 149 13th St, 5th Fl,
Charlestown, MA 02129 (swarren1@partners.org).
aThese two authors contributed equally to this work.
Potential conflicts of interest: in accordance with institutional policy, H.S.W. has reported the use of hemopexin as a potential anti-
inflammatory agent, and an application for patent protection has been filed. All other authors: no conflicts.
NIH Public Access
Author Manuscript
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
Published in final edited form as:
J Infect Dis
. 2010 August 15; 202(4): 624–632. doi:10.1086/654929.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

infection [3], is characterized at early stages by high levels of proinflammatory cytokines
[4]. Although the precise mechanisms underlying sepsis syndrome are not well understood,
it is widely believed that excessive production of cytokines may be a driving force, a
situation sometimes referred to as a “cytokine storm.” A better understanding of the
mechanisms responsible for the liberation of these cytokines may permit the development of
effective control strategies.
Exogenous microbial TLR ligands synergize with one another to activate signaling pathways
with subsequent induction of proinflammatory cytokines from macrophages and other
immune cells [5–10]. Data from our laboratory suggest that the outcome of stimulation with
different microbial TLR ligands is dependent on differential engagement of MyD88-
dependent and MyD88-independent pathways [11]. Recent studies suggest that endogenous
host molecules also can act as TLR ligands and that molecules released or induced during
tissue damage may contribute to the induction of inflammatory cytokines in sepsis [4]. For
example, such oxidants as hydrogen peroxide [12], heat-shock proteins (HSPs; ie, HSP-60,
HSP-70, and Gp-96), and self-messenger RNA have been proposed to synergize with
exogenous TLR agonists [13].
Visible or microscopic blood is often present in tissues where there is infection and necrosis,
so that hemoglobin and microorganisms coexist in infected microenvironments. This
situation is particularly common when there is invasive bacterial or viral infection with
tissue necrosis; after trauma, burns, or surgery; or in any infection where there is breach of
capillaries. Older studies revealed that blood and hemoglobin enhanced growth of bacteria
by providing heme as a nutrient source [14]. However, hemoglobin is also known to
synergize with LPS to augment macrophage induction of tumor necrosis factor (TNF) [15–
18]. It has been proposed that hemoglobin preparations increase the biological activity of
LPS through physically interacting with LPS [19,20]. Many of the studies involving the
interactions of hemoglobin and LPS have focused on the development of artificially cross-
linked hemoglobin for use as a cell-free blood transfusion substitute, where such an
interaction has major potential implications [15,16,21,22].
Given the ubiquity of blood in infected tissues and the broad array of bacterial and viral
infections in which blood might play a role, we studied the activation of macrophages in the
presence of hemoglobin and multiple different TLR ligand agonists (ie, ligands for TLR2,
TLR3, TLR4, TLR7, and TLR9). We observed extensive synergy with all of these TLR
agonists. We found that this synergy is not TLR4 dependent and not MyD88 dependent and
that the degraded hemoglobin product, hemin, synergizes with some (but not all) TLR
agonists. Finally, we observed that hemopexin (Hx), a plasma protein that binds heme with
an extremely high affinity, blocks the synergy of both hemoglobin and hemin with LPS,
raising the possibility that Hx may be involved in local regulation of this synergy. Our
findings also suggest that exogenously administered Hx might be a candidate for treating
inflammation induced by microorganisms in tissues that contain blood breakdown products
through a novel mechanism of blocking this synergistic activation of inflammation.
MATERIALS AND METHODS
Materials
The following TLR agonists were purchased: smooth LPS from Escherichia coli O55:B5
(List Biologicals), Pam3Cys (EMC Microcollections), poly I:C, loxoribine, and CpG DNA
(Invivogen). All these TLR agonists were dissolved in pyrogen-free H2O and saved as
aliquots at −80°C. E. coli O4 was the kind gift of Alan Cross (University of Maryland), and
it was heat killed and washed before use, as described elsewhere [23]. Heat-killed
Staphylococcus aureus was purchased from Invivogen. Hemin chloride was purchased from
Lin et al. Page 2
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Frontier Scientific. Hemin solutions were made immediately before use in the dark, as
described elsewhere [24]. In brief, the powder was dissolved in pyrogen-free 0.1 N NaOH at
10 mg/mL and was further diluted in serum-free medium as desired; these steps were
followed by filtration through 0.22-μm Millipore membranes. C57BL/6, C3H/HeN, and
C3H/HeJ mice were obtained from Charles River Laboratories. MyD88 knockout
(MyD88−/−) mice were generated by Kawai and colleagues [25] and had been backcrossed
>10 generations into the C57BL/6J strain.
Purification of mouse hemoglobin
Hemoglobin was purified as previously described elsewhere [26], with modifications, under
pyrogen-free conditions. In brief, mouse blood was collected from C57BL/6 mice by means
of cardiac puncture, and it was washed with an equal weight of isotonic saline solution
(0.9% NaCl, wt/vol) three times by centrifugation at 1000 g, to remove serum proteins.
Equal volumes of saline were added to the pellet containing red blood cells, and this solution
was sonicated for 5 × 10 s at 40% amplitude with 1-min laps between pulses in a Branson
450 sonicator from Branson Ultrasonics. The hemoglobin solution was diluted with an equal
volume of saline and subjected to a second centrifugation at 2000 g for 1 h. The resulting
hemoglobin solution, removed from the center layer, was filtered through 0.22-μm Millipore
membranes and saved at −20°C in the dark. The concentration of purified mouse
hemoglobin was measured using Micro-BCA. The purity of the hemoglobin was confirmed
to be >99%, by means of nondenaturing polyacrylamide gel electrophoresis and high-
pressure liquid chromatography.
Purification of Hx from mouse or human serum
Mouse serum Hx (mHx) or human serum Hx (hHx) was purified using heme affinity
chromatography essentially as we have described elsewhere [27]. In brief, after filtration
through 0.22- μm Millipore membranes, the abundant albumin in the serum was precipitated
and removed by use of cold 1.68% rivanol solution (pH 8.0). The sample obtained after
rivanol precipitation was dialyzed against pyrogen-free phosphate-buffered saline. Protease
inhibitors (0.5 mmol/L 4-[2-aminoethyl] ben-zenesulfonyl fluoride, 10 μmol/L E-64, 2 μg/
mL aprotinin, and 1 μmol/L pepstatin A) were added, and they interacted with the dialyzed
sample obtained after rivanol precipitation for 15 min by gentle agitation at 4°C. The
mixture was applied to 6 mL of heminagarose column (Sigma) 3 times, followed by
extensive washing with 1200 mL of phosphate-buffered saline containing 0.5 mol/L NaCl
overnight at 4°C to remove unbound proteins. Hx bound to the column was eluted by 0.2
mol/L citric acid (pH 2.0), followed by immediate neutralization with 10 mol/L NaOH.
Proteins in the buffer were exchanged, concentrated in PBS at 4°C by use of Centriprep
YM- 30 (Millipore), and saved in aliquots at −80°C.
Limulus amebocyte lysate assay
The limulus amebocyte lysate assay was performed as previously described elsewhere [28].
Preparation of macrophages
Bone marrow–derived macrophages (BMDMs) were prepared from mice, as described
elsewhere [29], with minor modifications [11,27]. BMDMs were seeded at 1.28 × 105 cells/
well in 96-well tissue culture plates and were allowed to adhere overnight before use in
assays.
Macrophage assays and cytokine production
BMDMs were washed 3 times in serum-free medium, followed by incubation overnight with
different TLR agonists with or without hemoglobin or hemin chloride at indicated
Lin et al. Page 3
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

concentrations. Purified mHx was added to the culture in some experiments, as noted.
Levels of TNF and interleukin (IL)–6 in the supernatants were quantitated by enzyme-linked
immunosorbent assay (ELISA; R & D Systems), in accordance with the manufacturer’s
instructions.
Human whole-blood assay
Whole-blood stimulations were performed as previously described elsewhere [30].
Heparinized whole blood, collected aseptically from healthy human volunteers, was diluted
1:4 in pyrogen-free RPMI 1640 (Cellgro; Mediatech) and placed in a 96-well plate at 100
μL/well. The diluted blood was incubated with desired concentrations of stimuli at 37°C,
followed by centrifugation at 500 g for 5 min. The supernatants were saved for
quantification of cytokines by ELISA.
Statistics
Except where indicated, representative data from at least three experiments are presented in
the figures. Data are expressed as means, and error bars denote standard errors. The data
were analyzed using Prism 5 software (GraphPad). One-way analysis of variance followed
by Dunnett’s post hoc test was used to assess cytokine levels produced by macrophages
treated with different TLR agonists in the absence or presence of hemin or hemoglobin at
desired concentrations. Student’s t test was used to compare the cytokine levels produced by
macrophages from C3H/HeN with C3H/HeJ or C57BL/6 with MyD88−/− mice in the
presence of TLR agonists and hemin or hemoglobin at desired concentrations. Two-way
analysis of variance followed by Bonferroni’s post hoc test was used to assess cytokine
levels from macrophages treated with LPS with or without hemin or hemoglobin at different
concentrations in the presence of Hx or phosphate-buffered saline. P <.05 was considered to
denote statistical significance.
RESULTS
Strong synergy of hemoglobin with TLR2, TLR3, TLR4, TLR7, and TLR9 agonists and with
bacteria to induce TNF and IL-6 from macrophages
Differing concentrations of hemoglobin were incubated with a predetermined optimized
concentration of each TLR agonist in cell culture with BMDMs, as described in Materials
and Methods. Levels of the proinflammatory cytokines TNF and IL-6 induced in the
supernatant were measured by ELISA. Hemoglobin significantly enhanced the induction of
TNF and IL-6 from BMDMs by all TLR agonist ligands studied (LPS, Pam3Cys, Poly I:C,
loxoribine, and CpG) in a dose-dependent manner (Figure 1). Concentrations of hemoglobin
as low as 30 μg/mL led to synergistic induction of the cytokines with LPS and Pam3Cys
(Figure 1A–D), whereas higher concentrations (~300 μg/mL) of hemoglobin were required
for loxoribine (Figure 1G, H) and CpG DNA (Figure 1I and 1J). Similar results were
obtained with different concentrations of the TLR agonists (data not shown) as well as with
killed E. coli and S. aureus (Figure 2A, 2B, 2E, and 2F). Synergy was observed as early as 3
h after the cells were stimulated (Figure 2C, 2D, 2G, and 2H).
TLR4- and MyD88-independent synergistic induction of proinflammatory cytokines by
hemoglobin and various TLR agonists
Because heme has been described as a TLR4 ligand agonist [24], we next addressed the
question of whether hemoglobin acts through TLR4-signaling pathways to induce the
synergy that we observed. For these experiments, the synergistic induction of cytokines
induced in BMDMs from C3H/HeJ mice, which are deficient in TLR4, was compared with
induction of cytokines from control BMDMs from C3H/HeN mice, by use of the same TLR
Lin et al. Page 4
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

agonists shown in Figure 1. As expected, neither LPS nor LPS with hemoglobin induced
TNF and IL-6 from C3H/HeJ macrophages (data not shown). The synergy for non-LPS
ligands with hemoglobin was present despite the deficiency of TLR4 (Figure 3), suggesting
that heme ligation of TLR4 [24] was not a prominent mechanism in the synergy. To be able
to compare between cell preparations from the different strains, these results are presented
as the percentage change.
We previously noted that the synergy between different microbial TLR agonists is altered
depending on whether the agonists function through the MyD88-dependent signaling
pathway [11]. MyD88 is an essential adaptor molecule in the signaling pathway of many
TLRs, with the exception that TLR3 is exclusively and TLR4 is partially, but not
predominantly, MyD88 independent [11]. Accordingly, we assessed the synergy of
hemoglobin with ligands of TLR3 and TLR4 in BMDMs from MyD88−/− mice. As shown
in Figure 4, hemoglobin induced a similar percentage increase in TNF from both wild-type
(C57BL/6) and MyD88-deficient BMDMs, when stimulated with TLR4 agonist LPS and
TLR3 agonist poly I:C. There was no significant IL-6 produced from MyD88-deficient
BMDMs (not shown). As expected, other TLR agonists (ie, Pam3Cys, loxoribine, and CpG
DNA) did not induce TNF and IL-6 from MyD88-deficient BMDMs, with or without
hemoglobin (data not shown).
Synergistic induction of proinflammatory cytokines from macrophages with hemin and
agonists of TLR3, TLR4, and TLR7—but not with TLR2 and TLR9—and the role of TLR4
and MyD88
Hemoglobin consists of 2 moieties: hemin and globin. It has been hypothesized but not
directly shown that hemin is the active moiety that synergizes with LPS [31]. To directly
study this question, we evaluated the synergy of hemin chloride in the presence and absence
of different TLR agonists. As previously reported, we found that hemin alone weakly but
significantly enhanced production of TNF (data not shown) [24]. Hemin synergized with
LPS (Figure 5A), poly I:C (Figure 5B), and loxoribine (Figure 5C). We also found that the
synergy of hemin with poly I:C and loxoribine was not TLR4 dependent (Figure 5B and 5C)
and that synergy with LPS was partially but not completely MyD88 dependent (Figure 5D
and 5E). Of note, we did not find any hemin synergy with Pam3Cys and CpG DNA,
although a large range of hemin chloride concentrations (0.01 to ~30 μmol/L) was tested
(data not shown).
Suppressive effect of Hx on the synergistic induction of pro-inflammatory cytokines from
macrophages by hemin and hemoglobin with LPS
Hx is the major heme scavenger in plasma and has been described to bind free heme in a 1:1
ratio with remarkably high affinity [32,33]. In addition, Hx has some immunomodulatory
activities in that it modestly down-regulates proinflammatory cytokines from macrophages
[27] and functions as an anti-inflammatory component of serum high-density lipoprotein in
atherosclerosis [34]. It was therefore of interest to assess whether Hx would alter the
induction of pro-inflammatory cytokines that were synergistically induced by hemin with
LPS. We found that Hx abrogated the synergistic induction of TNF (Figure 6A). Similar
results were obtained with IL-6 (data not shown). Unexpectedly, Hx also significantly
down-regulated the synergistic induction of TNF by hemoglobin with LPS (Figure 6B).
Similar results were obtained with IL-6 (data not shown). As previously described elsewhere
[27], in both situations, Hx decreased LPS-induced TNF when studied in the absence of
hemin or hemoglobin (as designated by the zero point in Figures 6A and 6B).
Lin et al. Page 5
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Suppressive effect of the induction of proinflammatory cytokines by LPS and E. coli in
whole human blood
A human whole-blood stimulation assay has been proposed as a more physiological system
in which to assess anti-inflammatory drugs [35]. We found that induction of TNF by LPS
and E. coli was significantly reduced by the addition of mHx and hHx in this system (Figure
6C and 6D).
DISCUSSION
Inflammation induced by microbes through activation of TLRs is believed to play a critical
role in host defense; however, in some circumstances, this same inflammation is also
believed to cause sepsis and tissue injury. We previously studied the synergy between
different exogenous microbial TLR agonists, in an effort to dissect the initiation process of
this inflammation [11]. In the present study, we report that multiple TLR agonists synergize
remarkably with 2 endogenous molecules, hemoglobin and its derivative heme. This finding
likely has broad importance in the clinic, because microscopic blood and blood degradation
products are present in many infected local microenvironments. Furthermore, we found that
this synergy is blocked by Hx, suggesting that Hx may be important in controlling the
inflammation in which even small amounts of degraded blood are present.
Older studies revealed that blood and hemoglobin enhance growth of bacteria by providing a
source of heme [14]. Our findings extend prior available information related to the induction
of cytokines by microbial cell walls in the setting of blood. Much of this information comes
from studies to evaluate the safety of using soluble cross-linked hemoglobin preparations as
a potential substitute for red blood cells, although enhancement of LPS-induced
inflammation by native hemoglobin has also been reported in multiple previous studies
[16,18,31]. In addition to the synergy with LPS, hemoglobin has been reported to increase
IL-6, when present in combination with lipoteichoic acid [36], and several cytokines, when
present in combination with the cell wall of Streptococcus suis [37]. To our knowledge, this
is the first study to systematically evaluate the synergy of native hemoglobin on
inflammation induced by multiple TLR agonists, including ligands for TLR3, TLR7, and
TLR9, which are involved in inducing inflammation after viral infections [38,39], to report
synergy of TLR agonists with heme and to report that the synergy is suppressed with Hx.
The mechanisms responsible for the synergy of hemoglobin with different TLR agonists
remain unclear. It has been proposed that one mechanism for hemoglobin synergy with LPS
is through direct binding of hemoglobin to LPS leading to better presentation of LPS to
TLR4 or is by inducing the conformational changes in the lipid A moiety of LPS [19,20].
Other studies suggest that hemoglobin may produce cell hypersensitivity to LPS by
increasing the intracellular load of heme iron, which catalyze cellular redox changes and
oxidant damage without direct interacting with LPS [40]. In a separate study, it was reported
that hemoglobin synergy with lipotechoic acid was partially dependent on TLR4 [36].
Although we cannot exclude the possibility that each of these different TLR agonist
molecules bound to hemoglobin or heme and that presentation was enhanced to the TLR,
this seems unlikely given their markedly different molecular properties. We also found that
TLR4 is not required for hemoglobin synergy with the synthetic TLR2 agonist Pam3Cys as
well as with TLR 3, TLR7, and TLR9 agonists and that MyD88 signaling is not required for
synergy with LPS and poly I:C, which activate cells via partial and exclusive MyD88-
independent pathways, respectively. The reason for the partial dependence on MyD88 for
the heme synergy with LPS is unclear. Taken together, these findings suggest that the role of
hemoglobin itself in the synergy may be independent of TLR activation. One possible
mechanism may be through quenching of the anti-inflammatory action of nitric oxide [41].
Lin et al. Page 6
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Blood and/or blood degradation products in infected body fluids are common. Sites at which
this occurs include but are not limited to pleural infection, peritoneal infection, central
nervous system infection, surgical wounds, burn wounds, trauma wounds, pancreatitis, and
necrotizing infections of any tissue, including lung tissue. Numerous bacterial, fungal,
parasitic, and viral infections also directly lead to hemorrhagic lesions or are present
together with bloody secretions. In addition, microbial TLR agonists are present with
hemoglobin in circulating blood in blood vessels if there is even minor hemolysis during
bacteremia, parasitemia, fungemia, or viremia. It is difficult to estimate the concentrations of
available free hemoglobin or heme in such infected tissues, drainage fluids, or blood. The
total concentration of hemoglobin in red blood cells in whole blood is ~150 mg/mL. We
found significant synergy with several TLR agonists (LPS, Pam3Cys, and Poly I:C) at 30
μg/mL, which is 1/5000th of this concentration. Estimation of the effective concentration of
heme is even more difficult. The approximate total heme concentration in red blood cells in
whole blood is approximately 9 mmol/L. We found significant synergy with 1 and 3 μmol/L,
which is approximately 1/3000th of this concentration. However, much of the free
hemoglobin and heme that is released from damaged red blood cells is likely to be bound
and blocked by haptoglobin [42,43] or by Hx [32] (Figure 6).
One of our most striking results was the suppressive effect of the Hx on both heme and
hemoglobin-LPS synergy. The suppression of the synergy of hemoglobin with LPS was
unexpected. There may be several potential mechanisms involved. First, Hx might simply
sequester free heme from interaction with macrophages. Indeed, this might also explain the
suppression of synergy with hemoglobin as well, because heme can be directly transferred
from hemoglobin to Hx [44]. Alternatively, heme-Hx complexes might induce
hemoxygenase-1, which has anti-inflammatory effects through its enzymatic products
carbon monoxide and biliverdin [45–47]. Finally, some other unknown mechanisms could
be involved. Elucidation of the mechanisms for the suppression of hemoglobin synergy with
LPS by Hx will need more studies.
A next step in evaluating the relevance of the findings will be to study the synergy of
hemoglobin and heme in some sort of in vivo system. Unfortunately, choosing an
appropriate model system that is reflective of human disease may be challenging [48].
Intrinsically, mice are multiple orders of magnitude more resilient to TLR4 agonist
challenge than humans, and the same relationship likely is true for other TLR agonists [23].
Therefore, assessment of the synergistic inflammation caused by blood products and
microbes might be better performed in a species that is similar to humans in sensitivity, such
as rabbits [23].
In summary, our studies suggest that hemoglobin and/or heme could contribute substantially
to the initiation of the inflammatory response induced by different microbial TLR agonists
in settings where trace amounts of blood coexist with bacterial or viral infection. The latter
is of special relevance given the current epidemic of H1N1 influenza and the knowledge that
severe disease can be accompanied by blood in the alveolar spaces [49]. Further work will
be needed to evaluate whether local concentrations of Hx are depleted in infected tissue
microenvironments and whether infusion of Hx to bind, neutralize, or clear hemoglobin or
hemin could be beneficial in some settings.
Acknowledgments
Financial support: National Institutes of Health (grants AI059010 and GM59694) and the Shriners Hospital for
Crippled Children (grant 8720).
Lin et al. Page 7
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

References
1. Akira S, Takeda K, Kaisho T. Toll-like receptors: critical proteins linking innate and acquired
immunity. Nat Immunol. 2001; 2:675–80. [PubMed: 11477402]
2. Akira S, Uematsu S, Takeuchi O. Pathogen recognition and innate immunity. Cell. 2006; 124:783–
801. [PubMed: 16497588]
3. Bone RC, Balk RA, Cerra FB, et al. Definitions for sepsis and organ failure and guidelines for the
use of innovative therapies in sepsis. The ACCP/SCCM Consensus Conference Committee.
American College of Chest Physicians/Society of Critical Care Medicine. Chest. 1992; 101:1644–
55. [PubMed: 1303622]
4. Rittirsch D, Flierl MA, Ward PA. Harmful molecular mechanisms in sepsis. Nat Rev Immunol.
2008; 8:776–87. [PubMed: 18802444]
5. Lombardi V, Van Overtvelt L, Horiot S, Moingeon P. Human dendritic cells stimulated via TLR7
and/or TLR8 induce the sequential production of Il-10, IFN-γ, and IL-17A by naive CD4+ T cells. J
Immunol. 2009; 182:3372–9. [PubMed: 19265114]
6. Strandskog G, Skjaeveland I, Ellingsen T, Jorgensen JB. Double-stranded RNA- and CpG DNA-
induced immune responses in Atlantic salmon: comparison and synergies. Vaccine. 2008; 26:4704–
15. [PubMed: 18602433]
7. Vanhoutte F, Paget C, Breuilh L, et al. Toll-like receptor (TLR)2 and TLR3 synergy and cross-
inhibition in murine myeloid dendritic cells. Immunol Lett. 2008; 116:86–94. [PubMed: 18166232]
8. He H, Genovese KJ, Nisbet DJ, Kogut MH. Synergy of CpG oligo-deoxynucleotide and double-
stranded RNA (poly I:C) on nitric oxide induction in chicken peripheral blood monocytes. Mol
Immunol. 2007; 44:3234–42. [PubMed: 17339052]
9. Roelofs MF, Joosten LA, Abdollahi-Roodsaz S, et al. The expression of toll-like receptors 3 and 7
in rheumatoid arthritis synovium is increased and costimulation of Toll-like receptors 3, 4, and 7/8
results in synergistic cytokine production by dendritic cells. Arthritis Rheum. 2005; 52:2313–22.
[PubMed: 16052591]
10. Napolitani G, Rinaldi A, Bertoni F, Sallusto F, Lanzavecchia A. Selected Toll-like receptor agonist
combinations synergistically trigger a T helper type 1-polarizing program in dendritic cells. Nat
Immunol. 2005; 6:769–76. [PubMed: 15995707]
11. Bagchi A, Herrup EA, Warren HS, et al. MyD88-dependent and MyD88-independent pathways in
synergy, priming, and tolerance between TLR agonists. J Immunol. 2007; 178:1164–71. [PubMed:
17202381]
12. Paul-Clark MJ, Sorrentino R, Bailey LK, Sriskandan S, Mitchell JA. Gram-positive and gram-
negative bacteria synergize with oxidants to release CXCL8 from innate immune cells. Mol Med.
2008; 14:238–46. [PubMed: 18231574]
13. Tsujimoto H, Ono S, Efron PA, Scumpia PO, Moldawer LL, Mochizuki H. Role of Toll-like
receptors in the development of sepsis. Shock. 2008; 29:315–21. [PubMed: 18277854]
14. Bornside GH, Bouis PJ Jr, Cohn I Jr. Hemoglobin and Escherichia coli, a lethal intraperitoneal
combination. J Bacteriol. 1968; 95:1567–71. [PubMed: 4870275]
15. Roth RI, Levin FC, Levin J. Distribution of bacterial endotoxin in human and rabbit blood and
effects of stroma-free hemoglobin. Infect Immun. 1993; 61:3209–15. [PubMed: 8335351]
16. Su D, Roth RI, Yoshida M, Levin J. Hemoglobin increases mortality from bacterial endotoxin.
Infect Immun. 1997; 65:1258–66. [PubMed: 9119460]
17. Su D, Roth RI, Levin J. Hemoglobin infusion augments the tumor necrosis factor response to
bacterial endotoxin (lipopolysaccharide) in mice. Crit Care Med. 1999; 27:771–8. [PubMed:
10321668]
18. Bodet C, Chandad F, Grenier D. Hemoglobin and LPS act in synergy to amplify the inflammatory
response. J Dent Res. 2007; 86:878–82. [PubMed: 17720859]
19. Kaca W, Roth RI, Levin J. Hemoglobin, a newly recognized lipopolysaccharide (LPS)-binding
protein that enhances LPS biological activity. J Biol Chem. 1994; 269:25078–84. [PubMed:
7929195]
Lin et al. Page 8
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

20. Brandenburg K, Garidel P, Andra J, et al. Cross-linked hemoglobin converts endotoxically inactive
pentaacyl endotoxins into a physiologically active conformation. J Biol Chem. 2003; 278:47660–
9. [PubMed: 13679376]
21. Bornside GH, Bouis PJ Jr, Cohn I Jr. Enhancement of Escherichia coli infection and endotoxic
activity by hemoglobin and ferric ammonium citrate. Surgery. 1970; 68:350–5. [PubMed:
4915899]
22. Su D, Roth RI, Levin J. Hemoglobin infusion augments the tumor necrosis factor response to
bacterial endotoxin (lipopolysaccharide) in mice. Crit Care Med. 1999; 27:771–8. [PubMed:
10321668]
23. Warren HS, Fitting C, Hoff E, et al. Resilience to bacterial infection: difference between species
could be due to proteins in serum. J Infect Dis. 2010; 201:223–32. [PubMed: 20001600]
24. Figueiredo RT, Fernandez PL, Mourao-Sa DS, et al. Characterization of heme as activator of Toll-
like receptor 4. J Biol Chem. 2007; 282:20221–9. [PubMed: 17502383]
25. Kawai T, Adachi O, Ogawa T, Takeda K, Akira S. Unresponsiveness of MyD88-deficient mice to
endotoxin. Immunity. 1999; 11:115–22. [PubMed: 10435584]
26. Andrade CT, Barros LA, Lima MC, Azero EG. Purification and characterization of human
hemoglobin: effect of the hemolysis conditions. Int J Biol Macromol. 2004; 34:233–40. [PubMed:
15374679]
27. Liang X, Lin T, Sun G, Beasley-Topliffe L, Cavaillon JM, Warren HS. Hemopexin down-regulates
LPS-induced proinflammatory cytokines from macrophages. J Leukoc Biol. 2009; 86:229–35.
[PubMed: 19395472]
28. Novitsky TJ, Roslansky PF, Siber GR, Warren HS. Turbidometric method for quantifying serum
inhibition of limulus amoebocyte lysate response. J Clin Micro. 1985; 20:211–6.
29. Schilling D, Thomas K, Nixdorff K, Vogel SN, Fenton MJ. Toll-like receptor 4 and Toll–IL-1
receptor domain-containing adapter protein (TIRAP)/myeloid differentiation protein 88 adapter-
like (Mal) contribute to maximal IL-6 expression in macrophages. J Immunol. 2002; 169:5874–80.
[PubMed: 12421970]
30. Liang MD, Bagchi A, Warren HS, et al. Bacterial peptidoglycan-associated lipoprotein: a naturally
occurring Toll-like receptor 2 agonist that is shed into serum and has synergy with
lipopolysaccharide. J Infect Dis. 2005; 191:939–48. [PubMed: 15717270]
31. Yang H, Wang H, Bernik TR, et al. Globin attenuates the innate immune response to endotoxin.
Shock. 2002; 17:485–90. [PubMed: 12069185]
32. Tolosano E, Altruda F. Hemopexin: structure, function, and regulation. DNA Cell Biol. 2002;
21:297–306. [PubMed: 12042069]
33. Paoli M, Anderson BF, Baker HM, Morgan WT, Smith A, Baker EN. Crystal structure of
hemopexin reveals a novel high-affinity heme site formed between two beta-propeller domains.
Nat Struct Biol. 1999; 6:926–31. [PubMed: 10504726]
34. Watanabe J, Grijalva V, Hama S, et al. Hemoglobin and its scavenger protein haptoglobin
associate with ApoA-1–containing particles and influence the inflammatory properties and
function of high density lipoprotein. J Biol Chem. 2009; 284:18292–301. [PubMed: 19433579]
35. Wilson BMG, Severn A, Rapson NT, Chana J, Hopkins P. A convenient human whole blood
culture system for studying the regulation of tumour necrosis factor release by bacterial
lipopolysaccharide. J Immunol Methods. 1991; 139:233–40. [PubMed: 1904465]
36. Cox KH, Ofek I, Hasty DL. Enhancement of macrophage stimulation by lipoteichoic acid and the
costimulant hemoglobin is dependent on Toll-like receptors 2 and 4. Infect Immun. 2007;
75:2638–41. [PubMed: 17296755]
37. Tanabe S, Gottschalk M, Grenier D. Hemoglobin and Streptococcus suis cell wall act in synergy to
potentiate the inflammatory response of monocyte-derived macrophages. Innate Immun. 2008;
14:357–63. [PubMed: 19039059]
38. Takeda, K.; Akira, S. Toll-like receptors. In: Coligan, JE., et al., editors. Current Protocols in
Immunology. Vol. chapter 14. 2007.
39. Takeda K, Akira S. Toll-like receptors in innate immunity. Int Immunol. 2005; 17:1–14. [PubMed:
15585605]
Lin et al. Page 9
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

40. Crowley JP, Metzger J, Gray A, Pivacek LE, Cassidy G, Valeri CR. Infusion of stroma-free cross-
linked hemoglobin during acute gram-negative bacteremia. Circ Shock. 1993; 41:144–9.
[PubMed: 8269643]
41. Bogdan C. Nitric oxide and the immune response. Nat Immunol. 2001; 2:907–16. [PubMed:
11577346]
42. Schaer DJ, Alayash AI, Buehler PW. Gating the radical hemoglobin to macrophages: the anti-
inflammatory role of CD163, a scavenger receptor. Antioxid Redox Signal. 2007; 9:991–9.
[PubMed: 17508920]
43. Polticelli F, Bocedi A, Minervini G, Ascenzi P. Human haptoglobin structure and function—a
molecular modelling study. FEBS J. 2008; 275:5648–56. [PubMed: 18959750]
44. Hrkal Z, Vodrazka Z, Kalousek I. Transfer of heme from ferrihemoglobin and ferrihemoglobin
isolated chains to hemopexin. Eur J Biochem. 1974; 43:73–8. [PubMed: 4209590]
45. Chung SW, Liu X, Macias AA, Baron RM, Perrella MA. Heme oxygenase-1-derived carbon
monoxide enhances the host defense response to microbial sepsis in mice. J Clin Invest. 2008;
118:239–47. [PubMed: 18060048]
46. Otterbein LE, Soares MP, Yamashita K, Bach FH. Heme oxygenase-1: unleashing the protective
properties of heme. Trends Immunol. 2003; 24:449–55. [PubMed: 12909459]
47. Pamplona A, Ferreira A, Balla J, et al. Heme oxygenase-1 and carbon monoxide suppress the
pathogenesis of experimental cerebral malaria. Nat Med. 2007; 13:703–10. [PubMed: 17496899]
48. Warren HS. Editorial: Mouse models to study sepsis syndrome in humans. J Leukoc Biol. 2009;
86:199–201. [PubMed: 19643738]
49. Perez-Padilla R, Rosa-Zamboni D, Ponce DL, et al. Pneumonia and respiratory failure from swine-
origin influenza A (H1N1) in Mexico. N Engl J Med. 2009; 361:680–9. [PubMed: 19564631]
Lin et al. Page 10
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 1.
Synergy of hemoglobin (Hb) with lipopolysaccharide (LPS), Pam3Cys, Poly I:C, loxoribine,
and bacterial DNA (CpG). Bone marrow–derived macrophages from C57BL/6 mice were
washed 3 times with serum-free medium and then were cultured overnight with either Hb
alone or Hb with various Toll-like receptor agonists: LPS (1 ng/mL) (A and B ), Pam3Cys
(P3C; 10 ng/mL) (C and D ), Poly I:C (150 μg/mL) (E and F ), loxoribine (250 μmol/L) (G
and H ), and CpG (0.1 μmol/L) (I and J ). Concentrations of tumor necrosis factor (TNF) (A,
C, E, G, and I ) and interleukin (IL)–6 (B, D, F, H, and J ) in the supernatants were
determined by enzyme-linked immunosorbent assay. The results denote the mean ± standard
error and are representative of 4 independent experiments. *P <.05 and **P <.01, compared
between cells treated with and without Hb.
Lin et al. Page 11
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 2.
Synergy of hemoglobin (Hb) with killed Escherichia coli and Staphylococcus aureus. Bone
marrow–derived macrophages from C57BL/6 mice were washed 3 times with serum-free
medium and then cultured overnight (A, B, E, and F ) or at different times (C, D, G, and H )
with heat-killed E. coli at indicated concentrations (A and B ) or at 105 cfu/mL (C and D ) or
with S. aureus at indicated concentrations (E and F ) or at 107 cfu/mL (G and H ) in the
absence or presence of Hb at 300 μg/mL. Concentrations of tumor necrosis factor (TNF) (A,
C, E, and G ) and interleukin (IL)–6 (B, D, F, and H ) in the supernatants were determined
by enzyme-linked immunosorbent assay. The results denote the mean ± standard error and
Lin et al. Page 12
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

are representative of 4 independent experiments. *P <.05, **P <.01, ***P <.001, compared
between cells treated with and without Hb.
Lin et al. Page 13
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 3.
Synergy of hemoglobin (Hb) with Toll-like receptor (TLR) agonists is not TLR4 dependent.
Bone marrow–derived macrophages from HeN and HeJ mice were washed 3 times with
serum-free medium and then cultured overnight with TLR agonists alone, including
Pam3Cys (P3C; 10 ng/mL) (A and B ), Poly I:C (P[I:C]; 150 μg/mL) (C and D ), loxoribine
(Lox; 250 μmol/L) (E and F ), and bacterial DNA (CpG; 0.1 μmol/L) (G and H ), or with Hb
at 100 μg/mL (A–D ) or 300 μg/mL (E and F ) or 1000 μg/mL (G and H ) indicated
concentrations. Concentrations of tumor necrosis factor (TNF) (A, C, E, and G ) and
interleukin (IL)–6 (B, D, F, and H ) in the supernatants were determined by enzyme-linked
immunosorbent assay. Each graph denotes the percentage of cytokine production when
cytokine production by TLR agonist alone is assigned to 100%. Results denote the mean ±
the standard error and are representative of 4 independent experiments. *P <.05 and **P <.
01, compared between cells treated with and without Hb.
Lin et al. Page 14
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 4.
Synergy of hemoglobin (Hb) with lipopolysaccharide (LPS) and Poly I:C is not MyD88
dependent. Bone marrow–derived macrophages from C57BL/6 and MyD88 knockout
(MyD88KO) mice were washed 3 times with serum-free medium and then cultured with
Toll-like receptor (TLR) agonists LPS (1 ng/mL) (A) or Poly I:C (P[I:C]; 150 μg/mL) (B )
and Hb at 100 μg/mL. Concentrations of tumor necrosis factor (TNF) in the supernatants
were determined by enzyme-linked immunosorbent assay. The data denote the percentage of
cytokine production, when cytokine production by TLR agonist alone is assigned to 100%.
The results denote the mean ± the standard error and are representative of 3 independent
experiments. *P <.05 and **P <.01, compared between cells treated with and without Hb.
Lin et al. Page 15
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 5.
Synergy of hemin with lipopolysaccharide (LPS), Poly I:C, and loxoribine, and the role of
Toll-like receptor (TLR)–4 and MyD88. Bone marrow–derived macrophages from C57BL/
6, C3H/HeN, C3H/HeJ, or MyD88 knockout (MyD88KO) mice were washed 3 times with
serum-free medium and then cultured with TLR agonists. A, C57BL/6 cells, hemin
concentrations as shown (the open bar denotes no hemin; vertical lines, 1 μmol/L hemin; and
the solid bar, 3 μmol/L hemin) and, where indicated, LPS (1 ng/mL). The tumor necrosis
factor (TNF) concentrations in the culture are shown. B and C, C3H/HeN (HeN) and C3H/
HeJ (HeJ) cells cultured with Poly I:C (P[I:C], 150 μg/mL) (B ) or loxoribine (Lox; 250
μmol/L) (C ) in the absence or presence of hemin (3 μmol/L). D and E, C57BL/6 or MyD88
KO cells cultured with Poly I:C (P[I:C]; 150 μg/mL) (D ) or LPS (1 ng/mL) in the absence
or presence of hemin (3 μmol/L) (E ). B–E, The percentage of TNF production when TNF
by the same type of cells stimulated by TLR agonist alone is assigned to 100%. The results
denote the mean ± the standard error and are representative of four independent experiments.
*P <.05, **P <.01, and ***P <.001, compared between cells treated with and without
hemin.
Lin et al. Page 16
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript

Figure 6.
Effect of mouse or human hemopexin (mHx or hHx) on hemin or hemoglobin (Hb) synergy
with lipopolysaccharide (LPS) or Escherichia coli on bone marrow–derived macrophages
(BMDMs) or human whole blood. BMDMs from C57BL/6 mice were washed 3 times with
serum-free medium and then cultured overnight with 1 ng/mL of LPS in the absence (●) or
presence (▲) of 300 μg/mL mHx and indicated concentrations of hemin (A) or Hb (B ).
Concentrations of tumor necrosis factor (TNF) in the supernatants were determined by
enzyme-linked immunosorbent assay. Fresh human whole blood was diluted 1:4 in serum-
free medium and cultured for 1 h and then were coincubated overnight with 2.5 ng/mL LPS
or 105 cfu/mL E. coli in the absence or presence of mHx (300 μg/mL) (C ), or 105 cfu/mL of
E. coli with or without hHx (400 μg/mL) (D ). The results denote the mean ± the standard
error and are representative of more than 4 independent experiments. *P <.05, **P <.01,
***P <.001, compared between samples treated with and without Hx.
Lin et al. Page 17
J Infect Dis. Author manuscript; available in PMC 2011 August 15.
NIH-PA Author Manuscript NIH-PA Author Manuscript NIH-PA Author Manuscript