Proinflammatory Effects of Bacterial Recombinant Human
C-Reactive Protein Are Caused by Contamination With
Bacterial Products, Not by C-Reactive Protein Itself
Mark B. Pepys, Philip N. Hawkins, Melvyn C. Kahan, Glenys A. Tennent, J. Ruth Gallimore,
David Graham, Caroline A. Sabin, Arturo Zychlinsky, Juana de Diego
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Abstract—Intravenous administration to human volunteers of a commercial preparation of recombinant human C-reactive
protein (CRP) produced in Escherichia coli was recently reported in this journal to induce an acute phase response of
serum amyloid A protein (SAA) and of CRP itself, and to activate the coagulation system. The authors concluded that
CRP is probably a mediator of atherothrombotic disease. Here we confirm that this recombinant CRP preparation was
proinflammatory both for mouse macrophages in vitro and for mice in vivo, but show that pure natural human CRP had
no such activity. Furthermore mice transgenic for human CRP, and expressing it throughout their lives, maintained
normal concentrations of their most sensitive endogenous acute phase reactants, SAA and serum amyloid P component
(SAP). The patterns of in vitro cytokine induction and of in vivo acute phase stimulation by the recombinant CRP
preparation were consistent with contamination by bacterial products, and there was 46.6 EU of apparent endotoxin
activity per mg of CRP in the bacterial product, compared with 0.9 EU per mg of our isolated natural human CRP
preparation. The absence of any proinflammatory activity in natural CRP for macrophages or healthy mice strongly
suggests that the in vivo effects of the recombinant preparation observed in humans were attributable to proinflammatory
bacterial products and not human CRP. (Circ Res. 2005;97:0-0.)
Key Words: C-reactive protein 䡲 inflammation 䡲 atherosclerosis 䡲 atherothrombosis cardiovascular disease
T
here is enormous current interest in the predictive association between the circulating concentration of
C-reactive protein (CRP) and future coronary events. This
followed our original description of such a relationship in
patients with established coronary disease,1 and the subsequent finding by others that single baseline measurements of
CRP are also significantly associated with risk of future
coronary events in individuals in the general population. With
increasing recognition that atherosclerosis and atherothrombosis are inflammatory processes, the capacity to precisely
measure a sensitive systemic marker of inflammation has
proved irresistible. Unfortunately, enthusiasm for what was
widely perceived as an important new risk marker has swept
away decades old knowledge of CRP and submerged critical
analysis of both the epidemiological evidence and in vitro
studies with CRP.
CRP, the classical acute phase protein, is an extremely
sensitive nonspecific systemic marker of toxicity, inflammation, infection, and tissue damage.2,3 There are very few
inflammatory and tissue damaging conditions in which CRP
production does not increase, and there is no specific associ-
ation of CRP with cardiovascular disease. CRP is an invariant
protein that has long been assayed with high sensitivity,4,5 and
there is no novel entity of high sensitivity CRP (hs-CRP) that
is somehow specific for cardiovascular disease. Furthermore
all other nonspecific systemic markers of inflammation that
have been measured are similarly associated with future
atherothrombotic events.6 – 8 With more than 7000 coronary
events now analyzed in prospective studies in general populations, it has become clear that the strength of the association
with CRP is notably less than was apparent from the smaller
initial studies,8,9 and there are compelling arguments against
both the rationale for, and utility of, measuring CRP as a risk
marker in healthy individuals.10,11 Nevertheless, the intense
focus on CRP has provoked pervasive speculation that the
epidemiological association between CRP values and atherothrombotic events implies a causal relationship.12,13
The idea that CRP might contribute to the pathogenesis of
atherosclerosis and atherothrombotic events is not new. We
first suggested it more than 20 years ago when we discovered
the specific binding of CRP to low density lipoproteins
(LDL) and very low– density lipoproteins (VLDL),14,15 and it
Original received July 8, 2005; revision received August 24, 2005; accepted October 19, 2005.
From the Centre for Amyloidosis and Acute Phase Proteins, Department of Medicine (M.B.P., P.N.H., M.C.K., G.A.T., J.R.G.), Hampstead Campus,
and the Department of Primary Care and Population Sciences (C.A.S.), Royal Free and University College Medical School, London, UK; Scottish
National Blood Transfusion Service (D.G.), Pennicuik, Midlothian, Scotland; and the Department of Cellular Microbiology (A.Z., J.d.D.), Max Planck
Institute for Infection Biology, Berlin, Germany.
Correspondence to Prof M.B. Pepys FRS, Centre for Amyloidosis and Acute Phase Proteins, Department of Medicine, Royal Free and University
College Medical School, Rowland Hill Street, London NW3 2PF, United Kingdom. E-mail [email protected]
© 2005 American Heart Association, Inc.
Circulation Research is available at http://circres.ahajournals.org
DOI: 10.1161/01.RES.0000193595.03608.08
1
2
Circulation Research
November 25, 2005
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
is supported by the potentially proinflammatory capacity of
human CRP to activate complement,16,17 and the presence of
CRP in atherosclerotic plaques.18 However, many plasma
proteins enter the plaques,19,20 and presence at the scene of a
crime is not proof of guilt or complicity; CRP in plaques
could equally well have an atheroprotective function.21 Transgenic expression of human CRP in apoE knockout mice does
not affect development or severity of atherosclerosis or the
nature of the lesions,22 but because spontaneous atherothrombotic arterial occlusion does not occur in this model it is not
informative about a possible role of CRP in atherothrombosis.
A variety of proinflammatory and prothrombotic effects of
human CRP preparations, mostly obtained from commercial
suppliers, have been reported on cultured cells in vitro
(reviewed in refs 12,13). However none of the preparations
were properly characterized, nor did these studies include
adequate controls. Bisoendial et al have now reported that
intravenous infusion into human subjects of recombinant
human CRP, produced in bacteria for in vitro use in CRP
immunoassays, elicited a massive acute phase response and
activated the coagulation system.23 Are these really the
effects of human CRP?
Materials and Methods
CRP
Human CRP produced in recombinant Escherichia coli (Oriental
Yeast Co Ltd, Shiga, Japan)24 was supplied in solution at 1 mg/mL
in 140 mmol/L NaCl, 2 mmol/L CaCl2, 20 mmol/L Tris, 0.05%
NaN3, pH 7.5 (BiosPacific). Our natural human CRP, purified from
malignant ascitic fluid,25 and finally in solution in 140 mmol/L NaCl,
10 mmol/L Tris, 2 mmol/L CaCl2, 0.1% NaN3, pH 8.0, was ⬎99%
pure by SDS reduced 8% to 18% PAGE (50 ␮g sample loading,
Brilliant blue R-350 stain, single band sensitivity limit ⬇0.1 ␮g).
Both CRP preparations at 1 mg/mL, confirmed by latex enhanced
immunoturbidimetric assay (Roche)26 and by BCA protein assay
(Pierce), were extensively dialyzed against the same batch of
140 mmol/L NaCl, 10 mmol/L Tris, 2 mmol/L CaCl2, pH 8.0, to
remove azide and any other dialyzable substances, and their structural and functional integrity were confirmed by gel filtration on
Superose-12 (Amersham), calcium dependent binding to immobilized phosphocholine, and electrospray mass spectrometry.
Endotoxin Assay
Endotoxin was assayed, according to the 2005 European Pharmacopoeia monograph 2.6.14, by Limulus kinetic turbidimetric method
(Cambrex), in pyrogen-free tubes with pyrogen-free water as diluent,
at final CRP concentrations of 100, 10, and 1 ␮g/mL. Results in two
separate assays were consistent.
Mice
Mice deficient respectively in Toll-like receptor 4 (TLR4), TLR2,
and myeloid differentiation factor 88 (MyD88)27 were bred under
specific pathogen–free conditions (Bundesinstitut für Risikobewertung, Berlin, Germany). Conventional female wild type C57BL/6
mice aged 7 weeks, C57BL/6 mice transgenic for human CRP,22 and
their wild-type litter mates, were used for in vivo experiments.
Cell Lines, Reagents, and In Vitro Assays
Bone marrow– derived macrophages obtained respectively from
wild-type, TLR2-, TLR4-, and MyD88-deficient mice, were seeded
at 50 000 per well in 96-well plates in DMEM containing 10% FCS,
5% horse serum, 10 mmol/L HEPES, 1 mmol/L pyruvate, and
10 mmol/L L-glutamine at 37°C and 5% CO2.28 After adherence
overnight, stimulation was compared in 200 ␮L fresh medium
containing 10% FCS using highly purified Shigella flexneri lipopoly-
saccharide (LPS),29 bacterial lipopeptide, Pam3CysSerLys4 (BLP;
EMC Microcollections), and the CRP preparations. Supernatants
collected after 24 hours were frozen until assayed for murine TNF-␣
by ELISA (R&D Systems), with a lower detection limit of 10 pg/mL.
The NF-␬B– dependent luciferase reporter assay29 used human
embryonic kidney (HEK) 293 cells that stably express human TLR2
or that had been transfected with empty vector as a control. The
THP-1–I␬B enhanced green fluorescent protein (EGFP) cell line was
generated by stable transfection of THP-1 with the pI␬B–EGFP
expression cassette (Clontech) and selected for neomycin resistance
at a final concentration of 0.5 ␮g/mL. A stable population of cells
was isolated by fluorescence activated cell sorting. THP-1–I␬B–
EGFP cells remained fluorescent when the NF-␬B pathway was
inactive whereas degradation of I␬␤–EGFP in response to NF-␬B
activators was quantifiable by fluorescence activated cell sorting
analysis. Soluble CD14 and LPS binding protein (LBP) were from
Biometec GmbH.
In Vivo Acute Phase Responses
Two days after bleeding for baseline values, 3 different groups of
mice (n⫽7 to 8) received intravenous injections of 100 ␮g of,
respectively, commercial recombinant human CRP exactly as supplied by BiosPacific, this product after extensive dialysis, and natural
human CRP dialyzed extensively against the same buffer; a fourth,
control, group received buffer alone. All mice were bled 24 hours
later for measurement of human CRP by latex enhanced immunoturbidimetric assay (Roche),26 mouse serum amyloid P component
(SAP), and complement component C3 by electroimmunoassay,30,31
and serum amyloid A protein (SAA) by ELISA (BioSource UK,
Nivelles, Belgium). An otherwise identical preliminary experiment
but with 3 mice per group gave closely similar results.
Results
Authenticity of the CRP Preparations
The recombinant and the natural human CRP preparations
were both composed entirely of intact native pentamers by
analytical gel filtration chromatography and were 100%
functionally intact by specific calcium-dependent binding to
phosphocholine. In both preparations the CRP protomer mass
was 23 028 Da by electrospray mass spectrometry, corresponding to the mass of 23 029 Da predicted from the known
amino acid sequence including the N-terminal pyroglutamic
acid.32
Endotoxin Content of the CRP Preparations
Assays of the natural CRP, with acceptable recoveries (106.5
to 146.6%) of spiked endotoxin at all dilutions of CRP tested,
detected 0.9080 endotoxin units (EU)/mg of CRP when tested
at 100 ␮g/mL of CRP; nothing was detected at higher
dilutions of CRP. Recovery of spiked endotoxin was also
acceptable (126.7 to 138.3%) at all dilutions of recombinant
CRP tested, but the apparent endotoxin content was substantial: 46.62 EU/mg of CRP at 100 ␮g/mL of CRP, 30.58
EU/mg of CRP at 10 ␮g/mL of CRP, and 15.13 EU/mg of
CRP at 1 ␮g/mL of CRP. The falling apparent endotoxin
content at increasing dilutions of CRP is bizarre as the
endotoxin value per mg of CRP should remain constant
regardless of dilution. Either interfering material is present,
despite the acceptable recoveries observed, or the activity in
the assay does not behave like authentic bacterial LPS.
Whatever the explanation, there is a very striking difference
from the natural human CRP preparation.
Pepys et al
TABLE 1.
CRP or Not CRP? That Is the Question
3
Effect of Intravenous Injection of Different Human CRP Preparations on Acute Phase Proteins in Mice
Natural CRP
(n⫽8)
Bacterial Recombinant
CRP (n⫽7)
Control Buffer Only
Median (range) Serum Concentrations (mg/L)
Dialyzed Bacterial
Recombinant CRP
(n⫽8)
CRP at 24 hours
Not detected (n⫽7)
3.22 (2.16, 3.98)
2.78 (1.74, 3.22)
2.56 (1.76, 4.40)
SAA at baseline
12.9 (12.5, 16.9)
14.0 (11.0, 16.0)
16.6 (11.8, 20.2)
22.3 (16.8, 32.2)
0.0009*
SAA at 24 hours
12.2 (7.3, 15.3)
10.6 (7.1, 22.5)
460.0 (319.2, 847.1)
410.8 (248.4, 741.2)
0.0001
⫺2.0 (⫺7.9, 2.7) (n⫽6)
⫺2.3 (⫺8.4, 11.2)
443.7 (301.3, 830.5)
391.4 (223.7, 709.0)
0.0001
7.4 (6.6, 11.3)
7.05 (5.6, 8.3)
6.8 (5.0, 11.7)
8.15 (6.0, 11.9)
Difference in SAA
(24 hours to baseline)
SAP at baseline
SAP at 24 hours
P Value
0.39 (excluding control group)
0.13
4.8 (4.2, 12.8)
4.4 (2.3, 11.6)
14.2 (13.1, 19.3)
17.35 (12.5, 23.4)
0.0001
⫺2.5 (⫺7.1, 5.4) (n⫽7)
⫺2.55 (⫺3.7, 3.3)
7.8 (2.5, 10.3)
9.35 (3.0, 13.6)
0.0003
C3 at baseline
513 (457, 651)
410 (292, 504)
499 (355, 621)
471 (443, 626)
0.05⫹
C3 at 24 hours
611 (448, 778)
343 (319, 494)
523 (436, 576)
471 (337,669)
0.002⫹
Difference in C3
(24 hours to baseline)
45 (⫺99, 321)
⫺65 (⫺145, 41)
15 (⫺121, 192)
⫺34 (⫺214, 149)
Difference in SAP
(24 hours to baseline)
0.12
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Preparations containing 100 ␮g of the different human CRP preparations shown, in solution in Tris-buffered saline containing calcium, or this buffer only as a
control, were injected intravenously into each mouse 48 hours after bleeding for baseline values. After 24 hours each animal was bled for assay of human CRP and
murine SAP and SAA. The residual concentrations of human CRP were as expected for the ⬇4 hour plasma half life of human CRP in mice.52
P values by Kruskal-Wallis test.
*Although the baseline SAA values were significantly different between the groups, the absolute difference was trivial and they were all within the normal range
for SAA in these mice. No sample was available for SAA assay from one mouse in the undialyzed recombinant CRP group which therefore comprised 6 animals for
this analyte.
⫹
Baseline and 24 hour C3 values differed between groups but no treatment produced a significant positive or negative acute phase response.
Pure Natural Human CRP Is Not
Proinflammatory In Vivo
Transgenic Production of Human CRP Is Not
Proinflammatory in Mice In Vivo
There was no increase in concentration of mouse SAP, SAA,
or C3 at 24 hours after intravenous injection into mice of pure
natural human CRP (Table 1), at the time when the acute
phase responses of these proteins to a single inflammatory or
toxic stimulus is usually maximal.31,33 The ⬇4.00 mg/kg dose
of CRP was ⬎3-fold higher than given to human subjects by
Bisoendial et al,23 but it did not provoke any acute phase
response and its effect was indistinguishable from injection of
buffer alone (Table 1). In dramatic contrast, injection of this
same dose of recombinant CRP, whether or not dialyzed
beforehand, induced a highly significant 2- to 3-fold rise in
mouse SAP and a 20- to 40-fold rise in mouse SAA (Table 1).
This startling acute phase response is comparable to that
produced in humans by the putatively “highly purified”
recombinant CRP preparation of Bisoendial et al23
Male mice transgenic for human CRP, with median human
CRP concentrations of ⬇2 to 5 mg/L throughout their lives,
maintained normal values of their exquisitely sensitive autologous acute phase proteins, mouse SAP, and SAA, apart from
rare spikes caused by occasional fighting injury22 (Table 2).
TABLE 2
Pure Natural Human CRP Is Not
Proinflammatory In Vitro
LPS-induced TNF-␣ production is dependent on TLR4
wheras TLR2 induces TNF-␣ release in response to bacterial
lipopeptides (BLP), and both pathways involve activation of
NF-␬B through the adaptor MyD88. Therefore TLR4⫺/⫺ and
MyD88⫺/⫺ macrophages do not produce significant amounts
of TNF-␣ in response to LPS, whereas TLR2⫺/⫺ and
MyD88⫺/⫺ macrophages do not respond to BLP (Figure 1A).
Mouse SAP and SAA Values in Wild-Type and Human CRP Transgenic Mice
Age
Protein, mg/L
12 months
14 months
16 months
18 months
CRP
SAP
SAA
CRP
SAP
SAA
CRP
SAP
SAA
CRP
SAP
SAA
Median
0
6.3
6.7
0
4.2
NA
0
7.1
NA
0
8.9
12.5
Range
0–0
5.7–11.9
4.2–16.6
6.0–35.1
10.7–171.0
Wild-type (n⫽7)
0–0
3.0–8.9
0–0
5.2–18.0
0–0
hCRP transgenic (n⫽10)
Median
2.4
7.0
7.6
4.9
5.2
Range
0.7–9.1
5.2–10.3
1.4–16.2
1.9–79.3
3.6–90.2
NA
1.5
7.1
0.7–10.2
4.1–12.7
NA
2.8
8.0
20.4
0.8–14.2
5.3–47.9
10.7–141.7
NA indicates not assayed.
There were no significant differences between wild-type and human CRP transgenic mice with respect to SAP and SAA concentrations at any time point; range
of P values (Mann–Whitney U test) at various ages, 0.52 to 0.77.
4
Circulation Research
November 25, 2005
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Figure 2. Activation of NF-␬B in THP-1 cells. Sh. flexneri LPS at
50 ng/mL (solid bars) activated significantly only in the presence
of both soluble CD14 (sCD14) and lipopolysaccharide binding
protein (LBP). Natural human CRP at 100 ␮g/mL (open bars)
had no activity alone or in the presence of both soluble CD14
(sCD14) and lipopolysaccharide binding protein (LBP). Recombinant bacterial CRP at 100 ␮g/mL (hatched bars) produced substantial activation in the presence of sCD14 and LBP, consistent
with the presence of LPS. Similar results were obtained regardless of whether or not the recombinant CRP material was dialyzed before testing.
human peripheral blood monocytes, whereas the natural CRP
did not.
Discussion
Figure 1. Activation by TNF-␣ production by wild-type and different knockout mouse macrophages. A, Stimulation with Sh.
flexneri LPS at 100 ng/mL (solid bars) or synthetic BLP at 20
ng/mL (open bars), showing the dependence on TLR4 and
TLR2, respectively and the requirement for MyD88 for both,
compared with baseline production in medium alone (hatched
bars). B, Effect of increasing concentrations of the different
human CRP preparations showing complete absence of activity
in the natural material (open bars), compared with marked dose
dependent stimulation by the recombinant product that required
TLR4, consistent with LPS activity. Similar results were obtained
regardless of whether or not the recombinant CRP material was
dialyzed before testing.
The recombinant CRP preparation elicited substantial production of TNF-␣ from both wild-type and TLR2⫺/⫺ cells but
not from TLR4⫺/⫺ and MyD88⫺/⫺ macrophages, regardless of
whether or not it was dialysed before testing (Figure 1B). In
marked contrast, our natural human CRP elicited no detectable production of TNF-␣ from any of the cells tested (Figure
1). NF-␬␤ was activated in THP-1 cells by the recombinant
CRP preparation in the presence of soluble CD14 and LBP,
consistent with the presence of LPS, but the natural human
CRP had no effect (Figure 2). Stimulation by the recombinant
CRP was unaffected by dialysis before testing. HEK293 cells
stably transfected with human TLR2 showed activation of
NF-␬␤ by synthetic BLP at doses as low as 0.1 ng/mL, but
there was no response to either CRP preparation at any
concentration up to 100 ␮g/mL (Figure 3). These results are
all compatible with the presence of LPS but not BLP in the
recombinant preparation. In other experiments, not shown
here, the recombinant CRP also induced IL-1␤ production by
Intravenous injection into mice of highly purified, structurally
intact, and fully functionally active human CRP, isolated
from human subjects, did not induce an acute phase response
of either mouse SAP or SAA. In marked contrast, injection of
the same quantity of CRP from the commercial bacterial
recombinant preparation induced a dramatic increase in the
Figure 3. Activation of NF-␬B in TLR2 transfected HEK 293
cells. Expression of TLR2 (solid bars) enables dose-dependent
NF-␬B activation by bacterial lipopeptides compared with
absence of any effects in control cells transfected with vector
alone (open bars). Neither natural nor recombinant CRP preparations produced any activation at all in either TLR-transfected
or control cells, suggesting absence of lipopeptide. Similar
results were obtained regardless of whether or not the recombinant CRP material was dialyzed before testing.
Pepys et al
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
circulating concentration of both these proteins, compatible
with the well known exquisitely sensitive murine acute phase
response to endotoxin.31,34,35 The recombinant material also
triggered proinflammatory cytokine and signaling responses
in mouse macrophages in vitro, consistent with the presence
of Gram-negative bacterial LPS in the preparation, whereas
our natural human CRP had no effect in these assays. We did
not detect bacterial lipopeptides in the recombinant preparation, but did not exclude the possible presence of bacterial
peptidoglycans. Formal in vitro testing for endotoxin by the
Limulus kinetic turbidimetric assay detected just 0.9 EU/mg
of CRP in the natural human CRP preparation, compared with
46.6 EU/mg in the recombinant product. However the activity
in the commercial material behaved anomalously on dilution,
and it was not possible to assign an estimate of endotoxin
content compatible with European Pharmacopoeia guidelines.
Bisoendial et al report that the Limulus assay detected ⬍1.5
EU/mL in their material but they do not state at what CRP
concentration this result was obtained.23 Because both the
recombinant and the natural human CRP molecules were
structurally and functionally intact, and were indistinguishable from each other, the acute phase response and transient
neutrophilia observed in humans by Bisoendial et al23 are
very unlikely to have been caused by CRP itself. These
effects are consistent with contamination by bacterial or other
extraneous material, despite the apparent lack of pyrogenicity
or increased TNF-␣ production. The present completely
negative result with authentic highly purified human CRP
places the onus on others to prove that the effects observed
with recombinant bacterial or other preparations are really
attributable to CRP itself.
Bisoendial et al23 subjected their material to a single gel
filtration chromatography step to purify it before intravenous
injection, but they showed that thereafter it still remained
proinflammatory both to cells in vitro and to mice as well as
humans in vivo. This is not surprising because LPS, in
particular, is known to bind avidly and nonspecifically to
proteins and surfaces, so that once present as a contaminant it
is exceptionally difficult to remove from biological systems.
Indeed in our experiments, extensive dialysis against physiological buffer, which is likely to be more effective than a
single gel filtration step for removal of low molecular weight
solutes, had no effect on the stimulatory activity of the
recombinant CRP preparation either in vitro or in vivo. The
recombinant CRP provided by BiosPacific is manufactured
by the Oriental Yeast Company in Japan, and, to our
knowledge, this is the only commercial source of bacterial
recombinant human CRP. CRP released by the recombinant
E. coli is isolated from the culture medium by nonspecific
calcium-dependent affinity chromatography on EACASepharose.24 Although this captures the human CRP, it would
be astonishing if all bacterial proinflammatory products were
removed, and we show here that they definitely are not. This
does not matter for the intended in vitro use of the product as
a calibrator for CRP immunoassays, but it is critical for
experiments on living cells in culture and evidently even
more important when administered in vivo. Furthermore,
different batches of the recombinant CRP might be variably
contaminated with different proinflammatory bacterial prod-
CRP or Not CRP? That Is the Question
5
ucts, leading to varying biological effects. The present results
are consistent with recent evidence that most of the proinflammatory activities ascribed to CRP, on the basis of in vitro
studies with commercial CRP preparations, are caused by
either contamination with LPS or by failure to remove the
sodium azide that is universally present as a preservative.25,36 – 41 In particular van den Berg and colleagues have
shown that CRP isolated carefully from human malignant
ascites or produced recombinantly in mammalian cells had no
proinflammatory effects on endothelial cells, and that these
effects of the bacterial recombinant product could all be
replicated by addition of either LPS or azide.41 These results,
together with the present findings, also demonstrate that
earlier observations of cytokine induction by preparations of
natural human CRP in the prerecombinant era, for example
Ballou et al,42 must have been caused by contamination with
proinflammatory substances other than CRP.
Human CRP concentrations encompass a 10 000-fold dynamic range from ⬇0.05 to ⬇500 mg/L, usually reaching a
peak around 24 to 48 hours after a single acute phase
stimulus.3,5,43 This behavior, which has been known for
decades, does not suggest a role for CRP in modulation of
systemic inflammatory, vascular, or coagulation responses.
Patients with many different acute and chronic diseases
display the whole range of CRP concentrations without
associated coagulopathy, vascular dysregulation, or global
upregulation of inflammation. Although CRP and SAA values generally correlate, there can be substantial divergence
between them44 that is inconsistent with the idea that CRP
itself triggers SAA production as proposed by Bisoendial et
al.23 Also, when the stimulus for acute phase CRP production
terminates, for example, after acute myocardial infarction43 or
with successful antimicrobial treatment of sepsis,45– 47 the
plasma CRP concentration falls exponentially with a half
time of ⬇24 hours, which is almost as fast as the rate of CRP
clearance from the circulation.48 There is never a secondary
biphasic spike in CRP concentration as would be expected if
CRP itself were a trigger for its own production as suggested
by Bisoendial et al23
Transgenic expression of human CRP evokes no acute
phase response in mice (Table 2) and intravenous or subcutaneous injection into rats of 40 mg/kg of pure natural human
CRP affects neither acute phase proteins (unpublished observations) nor blood pressure.25 In sharp contrast, despite the
lack of clinical symptoms or major signs other than transient
neutrophilia, the bacterial recombinant CRP preparation induced a massive acute phase response in humans. The SAA
values in the range of 50 to ⬎2000 mg/L23 are comparable to
the SAA responses to severe sepsis, acute myocardial infarction, major surgery, and massive trauma. SAA is the precursor of AA type amyloid fibrils which cause the serious and
often lethal disease of AA amyloidosis that complicates
chronic inflammatory diseases.49 It would be highly undesirable if the therapeutic use of a bacterial recombinant protein
incurred the risk of such a grave complication, and activation
of coagulation is also potentially harmful. The observations
of Bisoendial et al therefore suggest that in addition to
pyrogenicity, safety screening of recombinant proteins for
6
Circulation Research
November 25, 2005
human administration should include testing for acute phase
responses.
The present demonstration that natural human CRP is not
proinflammatory in vivo in healthy animals does not conflict
with the compelling evidence that human CRP can contribute
significantly to the severity of ischemic tissue damage,50,51
and possibly also other forms of tissue injury associated with
major acute phase responses in diseased subjects. The presence in damaged tissues of ligands for CRP, coupled with its
greatly increased concentration, evidently enables CRP to
exert pathogenic effects that are absent in healthy undamaged
individuals. Human CRP is thus a valid therapeutic target, as
we originally suggested,50 but the only responsible basis for
development and application of CRP-inhibitor treatments
must be rigorous evidence specifically implicating CRP itself
in pathogenesis.
16.
17.
18.
19.
20.
21.
Acknowledgments
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
This work was supported by Medical Research Council Program
Grant G97900510 (to M.B.P. and P.N.H.) and National Institutes of
Health Grant 1 R01 HL078578-01 (to M.B.P.). J.d.D. is the recipient
of European Union Marie Curie Fellowship HPMF-CT-2002-01528.
The study was designed by M.B.P., P.N.H., A.Z., and J.d.D.
Experimental work was conducted by M.C.K., G.A.T., J.R.G.,
J.d.D., and J.L. C.A.S. did the statistical analyses. M.B.P. wrote
the manuscript.
References
1. Liuzzo G, Biasucci LM, Gallimore JR, Grillo RL, Rebuzzi AG, Pepys
MB, Maseri A. The prognostic value of C-reactive protein and serum
amyloid A protein in severe unstable angina. N Engl J Med. 1994;331:
417– 424.
2. Pepys MB. C-reactive protein fifty years on. Lancet. 1981;i:653– 656.
3. Pepys MB, Hirschfield GM. C-reactive protein: a critical update. J Clin
Invest. 2003;111:1805–1812.
4. Kindmark C-O, Thorell JI. Quantitative determination of individual
serum proteins by radio-electroimmunoassay and use of 125I-labelled
antibodies (application to C-reactive protein). Scand J Clin Lab Invest.
1972;29(Suppl. 124):49 –53.
5. Shine B, de Beer FC, Pepys MB. Solid phase radioimmunoassays for
human C-reactive protein. Clin Chim Acta. 1981;117:13–23.
6. Danesh J, Collins R, Appleby P, Peto R. Association of fibrinogen,
C-reactive protein, albumin, or leukocyte count with coronary heart
disease. J Am Coll Cardiol. 1998;279:1477–1482.
7. Danesh J, Collins R, Peto R, Lowe GDO. Haematocrit, viscosity, erythrocyte sedimentation rate: meta-analyses of prospective studies of
coronary heart disease. Eur Heart J. 2000;21:515–520.
8. Danesh J, Wheeler JG, Hirschfield GM, Eda S, Eiriksdottir G, Rumley A,
Lowe GDO, Pepys MB, Gudnason V. C-reactive protein and other circulating markers of inflammation in the prediction of coronary heart
disease. N Engl J Med. 2004;350:1387–1397.
9. Danesh J, Whincup P, Walker M, Lennon L, Thomson A, Appleby P,
Gallimore JR, Pepys MB. Low grade inflammation and coronary heart
disease: prospective study and updated meta-analyses. BMJ. 2000;321:
199 –204.
10. Levinson SS, Miller JJ, Elin RJ. Poor predictive value of high-sensitivity
C-reactive protein indicates need for reassessment. Clin Chem. 2004;50:
1733–1735.
11. Woloshin S, Schwartz LM. Distribution of C-reactive protein values in
the United States. N Engl J Med. 2005;352:1611–1613.
12. Labarrere CA, Zaloga GP. C-reactive protein: from innocent bystander to
pivotal mediator of atherosclerosis. Am J Med. 2004;117:499 –507.
13. de Maat MPM, Trion A. C-reactive protein as a risk factor versus risk
marker. Curr Opin Lipidol. 2004;15:651– 657.
14. de Beer FC, Soutar AK, Baltz ML, Trayner I, Feinstein A, Pepys MB.
Low density and very low density lipoproteins are selectively bound by
aggregated C-reactive protein. J Exp Med. 1982;156:230 –242.
15. Rowe IF, Baltz ML, Soutar AK, Trayner IM, de Beer FC, Walker L,
Bowyer D, Herbert J, Feinstein A, Pepys MB. Rabbit and rat C-reactive
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
protein bind apolipoprotein B-containing lipoproteins. J Exp Med. 1984;
159:604 – 616.
Kaplan MH, Volanakis JE. Interaction of C-reactive protein complexes
with the complement system. I. Consumption of human complement
associated with the reaction of C-reactive protein with pneumococcal
C-polysaccharide and with the choline phosphatides, lecithin and sphingomyelin. J Immunol. 1974;112:2135–2147.
Siegel J, Rent R, Gewurz H. Interactions of C-reactive protein with the
complement system. I. Protamine-induced consumption of complement in
acute phase sera. J Exp Med. 1974;140:631– 647.
Zhang YX, Cliff WJ, Schoefl GI, Higgins G. Coronary C-reactive protein
distribution: its relation to development of atherosclerosis. Atherosclerosis. 1999;145:375–379.
Stender S, Zilversmit DB. Transfer of plasma lipoprotein components and
of plasma proteins into aortas of cholesterol-fed rabbits. Molecular size as
a determinant of plasma lipoprotein influx. Arteriosclerosis. 1981;1:
38 – 49.
Zhang Y, Cliff WJ, Schoefl GI, Higgins G. Plasma protein insudation as
an index of early coronary atherogenesis. Am J Pathol. 1993;143:
496 –506.
Bhakdi S, Torzewski M, Paprotka K, Schmitt S, Barsoom H, Suriyaphol
P, Han S-R, Lackner KJ, Husmann M. Possible protective role for
C-reactive protein in atherogenesis: complement activation by modified
lipoproteins halts before detrimental terminal sequence. Circulation.
2004;109:1870 –1876.
Hirschfield GM, Gallimore JR, Kahan MC, Hutchinson WL, Sabin CA,
Benson GM, Dhillon AP, Tennent GA, Pepys MB. Transgenic human
C-reactive protein is not proatherogenic in apolipoprotein E-deficient
mice. Proc Natl Acad Sci U S A. 2005;102:8309 – 8314.
Bisoendial RJ, Kastelein JJP, Levels JHM, Zwaginga JJ, van den Bogaard
B, Reitsma PH, Meijers JCM, Hartman D, Levi M, Stroes ESG. Activation of inflammation and coagulation after infusion of C-reactive
protein in humans. Circ Res. 2005;96:714 –716.
Tanaka T, Horio T, Matuo Y. Secretory production of recombinant
human C-reactive protein in Escherichia coli, capable of binding with
phosphorylcholine, and its characterization. Biochem Biophys Res
Commun. 2002;295:163–166.
Clapp BR, Hirschfield GM, Storry C, Gallimore JR, Stidwill RP, Singer
M, Deanfield JE, MacAllister RJ, Pepys MB, Vallance P, Hingorani AD.
Inflammation and endothelial function: direct vascular effects of human
C-reactive protein on nitric oxide bioavailability. Circulation. 2005;111:
1530 –1536.
Eda S, Kaufmann J, Roos W, Pohl S. Development of a new
microparticle-enhanced turbidometric assay for C-reactive protein with
superior features in analytical sensitivity and dynamic range. J Clin Lab
Anal. 1998;12:137–144.
Weiss DS, Raupach B, Takeda K, Akira S, Zychlinsky A. Toll-like
receptors are temporally involved in host defense. J Immunol. 2004;172:
4463– 4469.
Schaible U, Kaufmann SHE. Studying trafficking of intracellular
pathogens in antigen-presenting cells. In: Sansonetti P, Zychlinsky A,
eds. Molecular Cellular Microbiology. Vol 31. New York: Academic
Press;2002:343–360.
Aliprantis AO, Yang R-B, Mark MR, Suggett S, Devaux B, Radolf JD,
Klimpel GR, Godowski P, Zychlinsky A. Cell activation and apoptosis by
bacterial lipoproteins through toll-like receptor-2. Science. 1999;285:
736 –739.
Pepys MB, Dash AC, Fielder AHL, Mirjah DD. Isolation and study of
murine C3. Immunology. 1977;33:491– 499.
Pepys MB, Baltz M, Gomer K, Davies AJS, Doenhoff M. Serum amyloid
P-component is an acute-phase reactant in the mouse. Nature. 1979;278:
259 –261.
Oliveira EB, Gotschlich EC, Liu T-Y. Primary structure of human
C-reactive protein. J Biol Chem. 1979;254:489 –502.
Pepys MB, Rogers SL. Complement-independence of the acute phase
production of serum amyloid P component (SAP) in mice. Brit J Exp
Pathol. 1980;61:156 –159.
Poole S, Gordon AH, Baltz M, Stenning BE. Effect of bacterial endotoxin
on body temperature, plasma zinc and plasma concentrations of the
acute-phase protein serum amyloid P component in mice. Br J Exp Path.
1984;65:431– 439.
Poole S, Gaines Das RE, Baltz M, Pepys MB. Detection of endotoxin in
mice by measurement of endotoxin-induced changes in plasma concentrations of zinc and of the acute-phase protein serum amyloid P component. J Pharm Pharmacol. 1986;38:807– 810.
Pepys et al
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
36. Nagoshi Y, Kuwasako K, Cao YN, Kitamura K, Eto T. Effects of
C-reactive protein on atherogenic mediators and adrenomedullin in
human coronary artery endothelial and smooth muscle cells. Biochem
Biophys Res Commun. 2004;314:1057–1063.
37. van den Berg CW, Taylor KE, Lang D. C-reactive protein-induced in
vitro vasorelaxation is an artefact caused by the presence of sodium azide
in commercial preparations. Arterioscler Thromb Vasc Biol. 2004;24:
e168 –171.
38. van den Berg CW, Taylor KE. Letter regarding article by Li et al,
“C-reactive protein upregulates complement-inhibitory factors in endothelial cells”. Circulation. 2004;110:e542.
39. Lafuente N, Azcutia V, Matesanz N, Cercas E, Rodrı́guez-Mañas L,
Sánchez-Ferrer CF, Peiró C. Evidence for sodium azide as an artifact
mediating the modulation of inducible nitric oxide synthase by C-reactive
protein. J Cardiovasc Pharmacol. 2005;45:193–196.
40. Swafford AN, Jr., Bratz IN, Knudson JD, Rogers PA, Timmerman JM,
Tune JD, Dick GM. C-reactive protein does not relax vascular smooth
muscle: effects mediated by sodium azide in commercially available
preparations. Am J Physiol Heart Circ Physiol. 2005;288:H1786 –H1795.
41. Taylor KE, Giddings JC, van den Berg CW. C-reactive protein-induced in
vitro endothelial cell activation is an artefact caused by azide and lipopolysaccharide. Arterioscler Thromb Vasc Biol. 2005;25:1225–1230.
42. Ballou SP, Lozanski G. Induction of inflammatory cytokine release from
cultured human monocytes by C-reactive protein. Cytokine. 1992;4:
361–368.
43. de Beer FC, Hind CRK, Fox KM, Allan R, Maseri A, Pepys MB.
Measurement of serum C-reactive protein concentration in myocardial
ischaemia and infarction. Br Heart J. 1982;47:239 –243.
CRP or Not CRP? That Is the Question
7
44. Hogarth MB, Gallimore JR, Savage P, Palmer AJ, Starr JM, Bulpitt CJ,
Pepys MB. Acute phase proteins, C-reactive protein and serum amyloid
A protein, as prognostic markers in the elderly inpatient. Age Ageing.
1997;26:153–158.
45. Hind CRK, Thomson SP, Winearls CG, Pepys MB. Serum C-reactive
protein concentration in the management of infection in patients treated
by continuous ambulatory peritoneal dialysis. J Clin Pathol. 1985;38:
459 – 463.
46. Cox ML, Rudd AG, Gallimore R, Hodkinson HM, Pepys MB. Real-time
measurement of serum C-reactive protein in the management of infection
in the elderly. Age Ageing. 1986;15:257–266.
47. Boralessa H, de Beer FC, Manchie A, Whitwam JG, Pepys MB.
C-reactive protein in patients undergoing cardiac surgery. Anaesthesia.
1986;41:11–15.
48. Vigushin DM, Pepys MB, Hawkins PN. Metabolic and scintigraphic
studies of radioiodinated human C-reactive protein in health and disease.
J Clin Invest. 1993;91:1351–1357.
49. Pepys MB. Pathogenesis, diagnosis and treatment of systemic amyloidosis. Phil Trans R Soc Lond B. 2001;356:203–211.
50. Griselli M, Herbert J, Hutchinson WL, Taylor KM, Sohail M, Krausz T,
Pepys MB. C-reactive protein and complement are important mediators of
tissue damage in acute myocardial infarction. J Exp Med. 1999;190:
1733–1739.
51. Gill R, Kemp JA, Sabin C, Pepys MB. Human C-reactive protein
increases cerebral infarct size after middle cerebral artery occlusion in
adult rats. J Cereb Blood Flow Metab. 2004;24:1214 –1218.
52. Baltz ML, Rowe IF, Pepys MB. In vivo studies of the clearance of
C-reactive protein. Clin Exp Immunol. 1985;59:243–250.
Downloaded from http://circres.ahajournals.org/ by guest on June 18, 2017
Proinflammatory Effects of Bacterial Recombinant Human C-Reactive Protein Are
Caused by Contamination With Bacterial Products, Not by C-Reactive Protein Itself
Mark B. Pepys, Philip N. Hawkins, Melvyn C. Kahan, Glenys A. Tennent, J. Ruth Gallimore,
David Graham, Caroline A. Sabin, Arturo Zychlinsky and Juana de Diego
Circ Res. published online October 27, 2005;
Circulation Research is published by the American Heart Association, 7272 Greenville Avenue, Dallas, TX 75231
Copyright © 2005 American Heart Association, Inc. All rights reserved.
Print ISSN: 0009-7330. Online ISSN: 1524-4571
The online version of this article, along with updated information and services, is located on the
World Wide Web at:
http://circres.ahajournals.org/content/early/2005/10/27/01.RES.0000193595.03608.08.citation
Permissions: Requests for permissions to reproduce figures, tables, or portions of articles originally published
in Circulation Research can be obtained via RightsLink, a service of the Copyright Clearance Center, not the
Editorial Office. Once the online version of the published article for which permission is being requested is
located, click Request Permissions in the middle column of the Web page under Services. Further information
about this process is available in the Permissions and Rights Question and Answer document.
Reprints: Information about reprints can be found online at:
http://www.lww.com/reprints
Subscriptions: Information about subscribing to Circulation Research is online at:
http://circres.ahajournals.org//subscriptions/