C-reactive protein modulates lipid mediators in a pro-inflammatory direction

Abstract

Background

C-reactive protein (CRP) is a risk factor for atherosclerosis. Although inflammation may confound this association, CRP itself has been hypothesized to possess both pro-atherosclerotic and pro-inflammatory properties. In this study, we aimed to elucidate the mechanism by which CRP may modulate bioactive lipid mediators.

Results

We found that the overexpression of human CRP increased plasma IL-6 and TNF-a levels in mice. Moreover, the conditioned medium of CRP-overexpressing HepG2 cells increased the release of these cytokines from RAW264.7 cells to a greater degree than recombinant CRP. Lipidomics analyses then revealed that the overexpression of CRP increased the total levels of plasma lysophosphatidic acid, lysophosphatidylethanolamine, lysophosphatidylglycerol, lysophosphatidylinositol, and sphingosine 1-phosphate in mice. It also increased the levels of pro-inflammatory arachidonic acid derivatives, including PGE2 metabolites, LTA4 metabolites, and oxylipids, and decreased the levels of anti-inflammatory eicosapentaenoic acid- and docosahexaenoic acid-derived mediators. In regard to the mechanisms, analyses of CRP-overexpressing HepG2 cells suggested that CRP may increase the hepatic production of glycero-lysophospholipids, and may also modulate eicosanoids and related mediators outside the liver. Finally, analyses of the fraction separated using anti-CRP IgG suggested that CRP can bind several lipid mediators including sphingosine 1-phosphate, PGE2, and PGF2a.

Conclusions

CRP may modulate lysophospholipids, and eicosanoids, and related mediators in pro-atherosclerotic and pro-inflammatory directions.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12950-025-00459-5.

Background

Inflammation has been established as involved in the pathogenesis of many diseases, including atherosclerosis [1]. C-reactive protein (CRP) is a representative biomarker for inflammation and has been proposed to be a risk factor for atherosclerosis [2, 3]. This association is reasonable, since CRP is a well-known biomarker for inflammation, which is one of the important factors in the pathogenesis of atherosclerosis. However, several studies, have suggested that CRP may be directly involved in the pathogenesis of atherosclerosis [4]. Reportedly, CRP inhibits nitric oxide production [5] and increases the expression of adhesion molecules in endothelial cells [6]. CRP may also promote foam cell formation by binding LDL [7] and promote thrombosis by activating platelets [8].

In circulation, CRP mainly exists as a pentamer and originates from the livers [9]. When pentamer CRP binds to its ligand, it is irreversibly converted into monomer CRP [10]. Moreover, monomer CRP is believed to be responsible for most of the biological properties of CRP described above [4]. However, the ligands of CRP remained to be fully elucidated; one ligand candidate is lysophosphatidylcholine (LPC) [11]. LPC is a lysophospholipid that can, via autotaxin, be converted into lysophosphatidic acid (LPA), which in turn exerts potent biological effects via LPA receptors [12]. Structurally, LPC belongs to the glycero-lysophospholipids family, which also includes LPA, as well as lysophosphatidylethanolamine (LPE), lysophosphatidylglycerol (LPG), lysophosphatidylinositol (LPI), and lysophosphatidylserine (LPS). Sphingosine 1-phosphate (S1P), a sphingolipid-derived lysophospholipid, is also structurally akin to LPC [13].

Numerous studies have found that lysophospholipid mediators are involved in the pathogenesis of inflammation and atherosclerosis [14–17]. Biologically, eicosanoids, which are derived from arachidonic acid (AA), as well as eicosapentaenoic acid (EPA)- or docosahexaenoic acid (DHA)-derived metabolites resemble lysophospholipids and are also involved in inflammation [18, 19]. Moreover, previous studies by our group have demonstrated that these lipid mediators show dynamic modulations in human diseases in which inflammation is deeply involved, such as COVID-19 [20, 21], bladder cystitis [22], cancer [23–25], neuropathic pain [26, 27], Alzheimer’s disease [28], diabetes [29], and atherosclerosis [30–32].

Given the possibility that CRP exerts pro-inflammatory and pro-atherosclerotic effects, this study aimed to elucidate the mechanisms by which it modulates lysophospholipids, eicosanoids, and related mediators. To do so we used both murine and in vitro experiments.

Methods

Construction of an adenovirus vector encoding human CRP

Human CRP cDNA was cloned from a cDNA library for human liver tissue using the following primers: forward, 5’tgaattcaggcccttgtatc3’; and reverse, 5’ tcccagcatagttaacgagc3’. An adenovirus coding CRP (Ad-CRP) and a control blank adenovirus (Ad-Null) were then constructed using the AdEasy system (Stratagene, La Jolla, CA). Viruses were then purified using CsCl gradient centrifugation as described previously [33, 34].

Animal experiments

Ten-week-old C57BL/6 male mice were obtained from CLEA Japan (Tokyo, Japan). Two groups were then administered with Ad-CRP or Ad-Null via the tail vein at a dose of 2.5 × 10⁸ pfu/g of body weight. Mouse experiments were performed on the fifth day after viral administration and all mice fasted for four hours beforehand. All animal experiments were conducted in accordance with the guidelines for Animal Care of The University of Tokyo and were approved by its animal research ethics committee (Approval Nos. P11-074, P17-074, and A23M0174).

Cell experiments

For cellular experiments, HepG2 cells and RAW264.7 cells were first obtained from the American Type Culture Collection (ATCC, Manassas, VA). These cells were then cultured in DMEM (D5796, Sigma-Aldrich) supplemented with 10% fetal bovine serum (FBS; 554–02655; BioSera Inc., Orange, CA) and 1% penicillin/streptomycin (168-23191; WAKO Pure Chemical Industries). Recombinant human CRP was purchased from the manufacture (236608-1MG, Merck & Co. Inc. Rahway, NJ), which was produced in E. coli.

HepG2 cells were then transfected with Ad-CRP or Ad-Null at a multiplicity of infection of 25. After 48 h, the original DMEM was replaced with an FBS-free medium. Next, 24 h later the medium and cellular contents were collected and subjected to downstream analyses. RAW264.7 cells were challenged with an FBS-free medium containing recombinant CRP at concentrations of 0, 1, or 10 mg/dL, a conditioned medium of HepG2 cells infected with Ad-CRP or Ad-Null (i.e., CM-CRP or CM-Null), or a mixture of CM-CRP and CM-Null at a ratio of 1:9, prepared as described below. After 24 h, the medium was then collected to measure cytokine concentration. We prepared CM-CRP and CM-Null as follows: first, HepG2 cells were cultured on a 10 cm petri plate, and were then infected with Ad-CRP or Ad-Null as described above. After 48 h, the medium was replaced with FBS-free medium (12 mL/dish). After another 24 h, the medium was collected and used as CM-CRP or CM-Null for further analyses.

Measurements of protein levels, human CRP, mouse IL-6, and mouse TNF-a

Total protein levels were measured by a colorimetric assay for protein concentration following detergent solubilization (DC protein assay, 500 − 0116, Bio-Rad Laboratories, Inc. Hercules, Calif). Human CRP levels were measured using a DCRP00B ELISA kit (R&D Systems, Inc., MN, USA). The coefficient of variation (CV) of the kit was 2.86% and the analytical detection limits were confirmed from 0.20 mg/dL to 50 mg/dL in our laboratory.

The IL-6 and TNF-a levels in plasma or medium were measured using M6000B and MTB00B ELISA kits (R&D Systems, Inc.), respectively.

Measurement of lysophospholipids, eicosanoids, and related mediators

Next, we measured the levels of lysophospholipids, eicosanoids, and related mediators using two independent liquid chromatography-mass spectrometry (LC-MS/MS) protocols. Both procedures used an LC8060 system consisting of a quantum ultra-triple quadrupole mass spectrometer (Shimadzu, Japan) with experimental procedures following the method described and validated in previous reports from our group [29, 35]. Briefly, we monitored 12 acyl chains (i.e., 14:0, 16:0, 16:1, 18:0, 18:1, 18:2, 18:3, 20:3, 20:4, 20:5, 22:5, and 22:6) for all of LPA, LPC, LPS, LPI, LPG, and LPE. We also measured the levels of AA, EPA, DHA, and 155 eicosanoids and related mediators that are enumerated and described in our previous article [22]. Both the intra-day and inter-day coefficients of variation for almost all the metabolites are below 20%, as we have reported previously [24, 29, 35–37].

Separation of CRP from murine plasma samples

Next, we separated CRP from murine plasma using a procedure similar to that described in our previous article for the separation of apolipoprotein D [38]. We collected 250 µL of plasma from twelve CRP-overexpressing mice. We diluted plasma with 250 µL of PBS and mixed this with anti-CRP IgG (HPA027396, Sigma-Aldrich) or rabbit IgG (553-61281, WAKO Pure Chemical Industries) in 50 µL of Protein G. After incubation overnight, the binding antibody was then washed and collected using an immunoprecipitation kit (11719386001 Immunoprecipitation Kit (Protein G) Roche, Mannheim, Germany). Next, we dissolved the collected lipids directly in methanol containing 0.1% formic acid, then measured the levels of lysophospholipids, eicosanoids, and related mediators via LC-MS/MS as described above.

Western blot analyses

The whole protein fractions of cells were extracted using RIPA lysis buffer containing a protease inhibitor cocktail (04693116001, Roche) was then performed with SDS-PAGE using 2-mercaptoehanol as the reducing agent. We used 2 µg of cellular protein or a volume corresponding to 0.1 µL of plasma, with protein standards included to assess protein size. The anti-CRP antibody used here was No. 235,752(Calbiochem, Gibbstown, NJ).

Statistical analyses

All data were statistically analyzed using SPSS (IBM SPSS Inc., IL, USA). Results are expressed as dot plots or as median ± Interquartile Range (IQR). Statistical comparisons between any two groups were performed using Mann–Whitney U-tests and comparisons among three groups were performed using Kruskal–Wallis tests, followed by Steel–Dwass post-hoc tests. P < 0.05 was used as the threshold of statistical significance for all analyses.

Results

Overexpression of CRP induces inflammatory cytokine release from RAW264.7 cells

First, we investigated whether the overexpression of CRP resulted in the formation of CRP pentamers or monomers. Figure 1A shows that the overexpression of CRP and recombinant CRP used mainly formed monomers, which has been identified as the biologically active form [4]. Next, the plasma human CRP levels of CRP-overexpressing and control mice are shown in Fig. 1B. Finally, as shown in Fig. 1C and D, plasma TNF-a and IL-6 levels were higher in CRP-overexpressing mice.

Fig. 1.

Modulation of inflammatory cytokines by CRP in vivo and in vitro (A–D) Ten-week-old C57/BL6 mice were injected with an adenovirus vector encoding human CRP (CRP) or a control blank adenovirus vector (Null). Five days after the viral injection, modulation analyses were performed (n = 6/group). Shown are (A) The CRP levels of the plasma collected from the mice administered with adenovirus coding CRP or blank adenovirus (Null), as well as recombinant CRP described in the Method section, were determined by Western blotting. We have the western blots twice and this panel is the representative of the totality of all the western blotting results. The bands which were deemed, based on the molecular weights, as CRP were marked with arrows. (B) Plasma CRP levels as determined by ELISA. (C) Plasma TNF-a levels. (D) Plasma IL-6 levels. (E-H) RAW264.7 cells treated with recombinant CRP at concentrations of 0, 1, or 10 mg/dL (E, F) or conditioned medium of HepG2 cells infected with Ad-CRP or Ad-Null (i.e., CM-CRP or CM-Null) or with a mixture of CM-CRP and CM-Null at the ratio of 1:9. These treatments were prepared as described in the Materials and Methods section (G, H) (n = 6). After 24 h, the TNF-a (E, G) and IL-6 (F, H) levels of the medium were determined via ELISA and relative protein values were adjusted to cellular protein levels. ns: not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Next, to investigate whether the induction of inflammatory cytokines from RAW264.7 cells occurs in response to CRP itself or via modulation of other molecules by CRP, we challenged RAW264.7 cells with recombinant CRP or CM-CRP. The concentrations of CRP in the CM-CRP and CM-Null treatments were 3.8 mg/dL and not detected, respectively and the CRP exists as monomers in both cellular and medium components of CRP-overexpressing HepG2 cells (Supplemental Fig. S1). As shown in Fig. 1E and F, recombinant CRP only induced IL-6 at a concentration of 10 mg/dL, while CM-CRP induced both TNF-a and IL-6 to a greater degree than recombinant CRP (Fig. 1G and H), even when the concentration of CRP was lower than recombinant CRP. Taken together, these results suggest that although CRP itself might induce inflammatory cytokines, other molecules modulated by CRP overexpression in HepG2 cells may accelerate the release of inflammatory cytokines from RAW264.7 cells.

Overexpression of CRP modulated the levels of plasma lysophospholipids

Next, we investigated the modulation of the plasma levels of lipid mediators, including lysophospholipids, eicosanoids, and related mediators. With respect to lysophospholipids, we observed elevated levels of plasma total LPA, LPE, LPG, LPI, and S1P, while the plasma levels of total LPC and LPS were not significantly changed (Fig. 2A–G). Furthermore, when we investigated the modulation of specific species, we observed almost unbiased elevation of LPA, LPE, LPG, and LPI species, and observed increased 18:0 LPC, decreased 20:5 and 22:6 LPC, and increased 18:0, 22:5, and 22:6 LPS (Fig. 2H–M).

Fig. 2.

Modulation of plasma lysophospholipids in CRP-overexpressing mice. Modulation of plasma lysophospholipids in the mice described in Fig. 1A–D (n = 6) Shown are (A) Plasma LPC levels, (B) plasma LPA levels, (C) plasma LPE levels, (D) plasma LPG levels, (E) plasma LPI levels, (F) plasma LPS levels, and (G) plasma S1P levels. Also shown are (H) Plasma LPC species, (I) plasma LPA species, (J) plasma LPE species, (K) plasma LPG species, (L) plasma LPI species, and (M) plasma LPS species. ns: not significant, *P < 0.05, **P < 0.01. Red asterisks suggest an increase in CRP-overexpressing mice, while blue asterisks suggest a decrease

Overexpression of CRP-modulated plasma eicosanoids and related mediators

With respect to eicosanoids and related mediators, we observed different patterns among different lipid mediators. The plasma levels of PGE2, 8-iso-PGE2, PGA2, and 8-iso-PGA2, which are involved in the initiation of inflammation and in the regulation of immune responses [39], were higher in CRP-overexpressing mice (Fig. 3A–D). Moreover, the plasma levels of PGF2a, which induces vasoconstriction [40], and TXB2, a thrombogenic TXA2 derivative [41], were also higher in CRP-overexpressing mice, as were the levels of 6-keto-PGF1a, an anti-thrombogenic PGI2 derivative [42] (Fig. 3E–G). We also observed the upregulation of 11-trans-LTE4, LTB4, and N-acetyl-LTE4 in CRP-overexpressing mice (Fig. 3H–J), all of which are pro-inflammatory LTA4 metabolites [39], as well as 11-HETE and 15-HETE (Fig. 3K, L), which are AA-derived oxylipids that play pro-inflammatory roles [43]. In contrast, in CRP-overexpressing mice we also observed lower levels of plasma 8,12-iso-iPF2a-VI-1,5-lactone a derivative of F2-isoprostane [44] whose biological properties remain unknown (Fig. 3M). Moreover, the plasma levels of 8,9-DHET and 2-carboxy-AA, two AA-derived CYP metabolites that induce vasorelaxation and protect against inflammation [45, 46] were higher in CRP-overexpressing mice, as were those of 12-HHT, which facilitates epidermal wound healing [47], (Fig. 3N–P).

Fig. 3.

Modulation of plasma eicosanoids and related mediators in CRP-overexpressing mice. Modulation of plasma eicosanoids and related mediators in the mice described in Fig. 1A–D (n = 6). Only lipid mediators that were modulated significantly are shown (A–V). Also show are AA levels (W), EPA levels (X), and DHA levels (Y). ns: not significant, *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

With respect to EPA and DHA derivatives, which protect against inflammation [19], we found that the plasma levels of 10,17-DiHDoHE, 12-HEPE, 14,15-DiHETE, and 20-HDoHE were lower in CRP-overexpressing mice (Fig. 3Q–T). In contrast, the plasma levels of 8,9-EET-EA, a CYP product of arachidonoyl ethanolamide (AEA) that may possess anti-inflammatory properties [48], were higher (Fig. 3U), while those of Azelaoyl-PAF, which is a PPAR-g agonist [49], were lower (Fig. 3V). Finally, we observed no significant differences in plasma AA, EPA, and DHA levels (Fig. 3W–Y).

Overexpression of CRP modulates the cellular and medium components of lysophospholipids

In the above murine model, we were unable to determine whether the modulation of the lipid mediators resulted from the direct effects of CRP or via consequent conditions caused by CRP. We therefore investigated the modulation of lipid mediators by CRP using in vitro experiments using HepG2 cells.

We found that the overexpression of CRP significantly increased the cellular levels of LPC, LPA, LPE, LPG, LPI, and LPS, but not S1P (Fig. 4A–G). Of these lysophospholipids, only LPC showed significant upregulation in medium levels in response to the overexpression of CRP (Fig. 4H–N). We observed no biased regulation of LPC species in either the cellular components or medium (Supplemental Fig. S2).

Fig. 4.

Modulation of cellular and medium lipid mediators in CRP-overexpressing HepG2 cells. HepG2 cells were infected with an adenovirus vector encoding human CRP (CRP) or a blank adenovirus vector (Null) (n = 4). After 48 h, the culture medium was replaced with a serum-free medium. After a further 24 h, the cellular components (i.e., A–G, O–U) and the medium (i.e., H–N, V, W) were collected for analyses involving lysophospholipids (A–N), eicosanoids, and related mediators (O–W). ns: not significant, *P < 0.05

Overexpression of CRP modulates the cellular and medium components of eicosanoids and related mediators

We found that the overexpression of CRP significantly increased the cellular levels of 18-HETE, a CYP product of AA that acts as a vasoconstrictor [50], 8-isoPGF3a, an isoprostane produced by the free-radical peroxidation of EPA, and whose biological properties remain unknown [51]. CRP overexpression also decreased cellular levels of 15-KEDE, an eicosadienoic acid derivative [51], as well as the levels of 9-HODE and 13-KODE, LOX products of linoleic acid that may be involved in cancer pathology [52, 53], and those of AEA and its product 5,6-EET-EA [54], which may possess vasorelaxant effects [55] (Fig. 4O–U). With respect to the modulation of lipid mediators in the medium, we found that the overexpression of CRP decreased the levels of 19-HETE, a CYP product of AA that acts as a vasodilator [56], as well as 13-KODE (Fig. 4V, W).

CRP may bind some lysophospholipids, eicosanoids, and related mediators

Previous studies have suggested that CRP may bind LPC [11]. Here, we investigated the possibility that CRP binds lipid mediators other than LPC as well. To do so, we first separated CRP proteins from the plasma of CRP-overexpressing mice (Fig. 5A). Compared to fraction precipitated with negative control IgG, the fraction separated with anti-CRP IgG contained significant higher levels of S1P among monitored lysophospholipids (Fig. 5B, Supplemental Fig. S3, S4). With respect to eicosanoids and related mediators, the fraction precipitated with anti-CRP IgG contained significantly higher levels of several AA-derived metabolites, including PGE2, PGF2a, 1a1b-dihomo-PGF2a, 11-dehydro-TXB2, a thrombogenic TXA2 derivative [41], 14,15-LTC4, a 15-lipoxygenase product of AA that is involved in the pathogenesis of asthma [57], PGJ2 and its derivative 15-deoxy-delta-12,14-PGJ2, which may possess anti-inflammatory [58] and PPAR-g agnostic properties [59], 11,12-DHET, an AA-derived CYP metabolite that induces vasorelaxation [45], and 19-HETE (Fig. 5C–K).

Fig. 5.

Possible binding of some lipid mediators to CRP. The plasma of CRP-overexpressing mice was subjected to immunoprecipitation with anti-CRP antibody (anti-CRP) or a negative control IgG (NC), as described in the Materials and Methods section. Precipitated fractions were then extracted for lipidomic analyses. (A) Western blot analysis for CRP of the precipitated fractions. (B) Plasma S1P levels in the precipitated fractions (n = 6). (C–K) Plasma levels of eicosanoids and related mediators in the precipitated fractions (n = 6). Only lipids that were significantly modulated are shown. ns: not significant, *P < 0.05, **P < 0.01

Discussion

In the present study, we investigated how CRP modulates lipid mediators to identify the mechanism underlying the putative direct effect of CRP on the pathogenesis of human diseases, including atherosclerosis [4]. We found that CM-CRP induced inflammatory cytokines in RAW264.7 cells to a greater degree than did recombinant CRP (Fig. 1E–H), suggesting that the overexpression of CRP in HepG2 cells produced molecules which induced the accumulation of pro-inflammatory cytokines. We also found that the overexpression of CRP modulated the levels of lysophospholipids, eicosanoids, and related mediators in both our mice experiments and in our in vitro experiments. In both experiments, the main observed isomer of CRP was a monomer (Fig. 1A, Supplemental Fig. S1). Since monomer CRP may be responsible for the pathological role played by CRP in atherosclerosis (i.e., relative to pentamer CRP, another major CRP isoform [4]), the present model was reasonable for the investigation of the involvement of lipid mediators in the direct effects of CRP. The modulations of lipid mediators by CRP are summarized in Fig. 6.

Fig. 6.

Modulation of lipid mediators by CRP. Scheme depicting the modulation of lysophospholipids, eicosanoids, and related mediators by CRP, as suggested by the results of the present study

In mice, the overexpression of CRP increased the total levels of LPA, LPE, LPG, LPI, and S1P. For pathogenic atherosclerosis, LPA has been proposed to promote atherosclerosis via the activation of platelets, initiation of lymphocyte or monocyte adhesion on the endothelium, migration of smooth muscle cells, and the induction of inflammation [15, 30, 60]. Moreover, LPI and LPG have been found to induce inflammation via GPR55 [23, 61]. In contrast, S1P generally protects against atherosclerosis by exerting anti-apoptosis and vasoprotective effects [14]. The roles of LPE in the pathogenesis of atherosclerosis remain unknown, while we have demonstrated a substantial increase of plasma LPE in acute coronary syndrome [30]. With respect to LPC and LPS, although their total levels were not significantly modulated, we observed that overexpression of CRP was associated with an increase in 18:0 LPC, a decrease in 20:5 LPC and 22:6 LPC, and increases in 18:0 LPS, 20:5 LPS, and 22:6 LPS. At present, the specific roles of LPC and LPS species remain unknown, whereas LPC may induce lipotoxicity [62]. Therefore, except for S1P, our results indicate that the overexpression of CRP caused plasma lysophospholipid levels to be modulated in a pro-atherosclerotic or pro-inflammatory direction.

Next, with respect to eicosanoids and related mediators, we identified modulation of some AA derivatives as well as EPA and DHA derivatives. We observed that the overexpression of CRP resulted in the elevation of some pro-inflammatory AA derivatives, including PGE2 metabolites, LTA4 metabolites, and oxylipids, of vasoconstrictors such as PGF2a, and of prothrombogenic TXA2 metabolites. The overexpression of CRP was also associated with a decrease in anti-inflammatory EPA and DHA derivatives. These modulations suggest that the overexpression of CRP also modulated the levels of eicosanoids and related mediators in a pro-atherosclerotic direction.

To understand the mechanisms underlying the modulations of lipid mediators, we investigated their modulations by the overexpression of CRP in HepG2 cells, since CRP can modulate lipid mediators both directly and indirectly. The overexpression of CRP was associated with significant increases in cellular S1P, LPC, LPA, LPE, LPG, LPI, and LPS and tended to increase other lysophospholipids. The difference in the modulation of LPC between in vivo and in vitro environments may be explained by differences in the species between them (Fig. 2H and Supplemental Fig. S2). Taken together, these results suggest that CRP may enhance the production of lysophospholipids in cells or suppress the degradation of lysophospholipids. Finally, regarding eicosanoids and related mediators, although the overexpression of CRP modulated some of the lipids, these modulations were not similar to those found in the plasma. This result suggests that CRP may modulate lipid mediators outside the liver, where CRP is produced.

We further investigated the role of lipids carried on CRP, since CRP has been reported to bind LPC [11]. Interestingly, the present study could not support the hypothesis that CRP binds LPC, nor other glycero-lysophospholipids; in contrast, we did observe that CRP seemed to bind S1P. Moreover, given that only S1P did not increase in the cellular components of the CRP-overexpressing HepG2 cells, in a similar way to apolipoprotein M [63], CRP might bind and retard the degradation of S1P. This in turn may result in the upregulation of plasma S1P in CRP-overexpressing mice. With respect to eicosanoids and related mediators, CRP may bind PGE2 and PGF2a, both of which are upregulated in the plasma of CRP-overexpressing mice. These pro-inflammatory lipids may in turn contribute directly to the pro-atherosclerotic properties of CRP. We also note that other lipids that were not significantly modulated in CRP-overexpressing mice may also bind to CRP. However, PGJ2 15-deoxy-delta-12,14-PGJ2, 11,12-DHET, and 19-HETE are known to possess anti-inflammatory or vasodilation properties [45, 56, 58], and therefore probably do not explain the direct pro-atherosclerotic effects of CRP. In any case, the modulation of the levels of eicosanoids and related mediators may instead be explained by the indirect effects of CRP, since the overexpression of CRP in HepG2 cells did not replicate the modulation pattern detected for eicosanoids and related mediators in the plasma of the CRP-overexpressing mice.

The major limitations of the present study are that we were not able to fully elucidate the underlying mechanism responsible for the modulation of lipid mediators as described above and that we did not elucidate how the present findings are involved in the pathogenesis of human disease. However, we believe that since CRP is clinically upregulated in many diseases, including Castleman disease [64], the present findings may help researchers to better investigate the pathogenesis such human diseases. Moreover, further research is necessary to determine how CRP is associated with lipid mediators in the various diseases in which CRP levels increase.

Conclusions

CRP may modulate lysophospholipids, eicosanoids, and related mediators in pro-atherosclerotic and pro-inflammatory directions. Although some lipids, such as S1P, showed an opposite trend, other key pieces of data, including the fact that CM-CRP induced inflammatory cytokines, likely explain the direct pro-atherosclerotic properties of CRP.

Abbreviations

AA

Arachidonic acid

Ad-CRP

Adenovirus coding CRP

Ad-Null

Control blank adenovirus

AEA

Arachidonoyl ethanolamide

CM-CRP

Conditioned medium of HepG2 cells infected with Ad-CRP

CM-Null

Conditioned medium of HepG2 cells infected with Ad-Null

CRP

C-reactive protein

DHA

Docosahexaenoic acid

EPA

Eicosapentaenoic acid

FBS

Fetal bovine serum

LC-MS/MS

Liquid chromatography-mass spectrometry

LPA

Lysophosphatidic acids

LPC

Lysophosphatidylcholine

LPE

Lysophosphatidylethanolamine

LPG

Lysophosphatidylglycerol

LPI

Lysophosphatidylinositol

LPS

Lysophosphatidylserine

S1P

Sphingosine 1-phosphate

Author contributions

M. K designed the study; M. K., K. T., H. I. and M. H. performed the experiments; M. K., K. T., M. H., H. I. and Y. Y. wrote the manuscript.

Funding

This work was supported by Public Trust Cardiovascular Research Fund (M. K.), JSPS KAKENHI Grant Number 16H06236 (M.K.), 20H03573 (M. K.), 22K08635 (K. T., M. K.), and 24K02358 (M.K.) and by AMED under Grant Number JP24zf0127006h0003 (M.K.).

Data availability

All relevant data are contained in the manuscript and the datasets generated or analyzed during the current study will be made available upon reasonable request.

Declarations

Ethics approval and consent to participate

All animal experiments were conducted in accordance with the guidelines for Animal Care and were approved by the animal research ethics committee of The University of Tokyo (Approval Nos. P11-074, P17-074, and A23M0174).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.