In vitro, IL-6 is the main inducer of the human C-reactive protein (CRP) gene, and IL-1 and steroids can enhance this effect. However, in mice, IL-6 is necessary but not sufficient for induction of the human CRP transgene, and testosterone is required for its constitutive expression by males. To examine the relative contributions of testosterone and IL-6 in the regulation of CRP gene expression, we produced CRP-transgenic (CRPtg), IL-6-deficient (IL-6−/−) mice. Male CRPtg/IL-6−/− mice expressed CRP constitutively, but CRP levels were not increased after injection of LPS. However, acute-phase CRP levels were attained after injection of IL-6. In contrast, female CRPtg/IL-6−/− mice did not express CRP constitutively or after administration of LPS, IL-6, IL-1, or IL-6 plus IL-1. Like males, testosterone-treated CRPtg/IL-6−/− females expressed CRP constitutively, and their transgene responded to injection of IL-6. The endogenous acute-phase protein serum amyloid P (SAP) was expressed constitutively equally by male and female IL-6−/− mice, responded minimally to LPS, and did not respond to either IL-6 or IL-1 alone. Acute-phase levels of SAP were induced in IL-6−/− mice by injection of IL-6 together with IL-1 or LPS. We conclude that in vivo, both constitutive and IL-6-dependent acute-phase expression of the CRP transgene require testosterone. In contrast, testosterone is not required for expression of the SAP gene, which requires IL-1 plus IL-6 for acute-phase induction.
Tissue injury results in a systemic reaction, termed the acute-phase response (APR)3 (1), that includes changes in the serum levels of several plasma proteins produced by hepatocytes. I.p. injection of LPS elicits the APR and thus serves as a convenient experimental model for the study of induction of acute-phase protein genes (2). The genes encoding TNF-α, IL-1, and IL-6 are induced by LPS (3) and all three cytokines contribute to regulation of the APR (4, 5). Acute-phase protein genes have been divided into two general classes according to their responsiveness to different cytokines and steroids. Class 1 genes are induced mainly by IL-1, IL-1 in combination with IL-6, or both cytokines in combination with glucocorticoids; class 2 genes are induced by IL-6 alone (4, 6). However, considerable variability exists among individual acute-phase proteins, and the exact cytokine and/or hormone requirements for their regulation in vivo are much more complex than suggested by this classification.
The pentraxins comprise a small group of evolutionarily conserved, Ca2+-binding proteins (7, 8) that participate variably in the APR in different species (4). For example, the pentraxin C-reactive protein (CRP) (9) is a major acute-phase protein in humans, while in the mouse it is a trace plasma component and only a minor acute-phase protein (10, 11, 12). On the other hand, serum amyloid P (SAP), a pentraxin structurally similar to CRP (reviewed in 8 , is highly inducible in the mouse but not in humans (13). Numerous studies using freshly isolated human hepatocytes and hepatoma cell lines (14, 15, 16, 17, 18) have established that IL-6 is the major inducer of the CRP gene and that IL-1 (14) and glucocorticoids (15, 18) act only in synergy with IL-6 to increase CRP gene expression. Similar studies using primary mouse hepatocytes (19) have shown that the mouse SAP gene can be induced by the direct action of either IL-1 or IL-6. Thus, both CRP and SAP are usually considered to be class 1 acute-phase gene products.
Ciliberto et al. (20) constructed human CRP transgenic (CRPtg) mice and used them to investigate the regulation of CRP gene expression in vivo. It was established that cis-acting elements within the CRP transgene are responsible for liver-specific and for LPS-induced acute-phase expression of human CRP. It was also established that the trans-acting factors necessary for control of the CRP transgene are conserved from mouse to humans. Further detailed studies of these and additional CRPtg mice (21) revealed that distal elements in both the 5′- and 3′-flanking regions of the gene are required to suppress expression of CRP in the absence of induction. We showed independently that CRPtg mice exhibit a sexually dimorphic pattern for constitutive expression of human CRP, i.e., basal expression of the transgene was observed only in males and was dependent on testosterone (22).
The involvement of IL-6 in the induction of the CRPtransgene in vivo was recently investigated using CRPtg/IL-6-deficient (IL-6−/−) mice (23). Increased expression of serum CRP (40 μg/ml) and mRNA could be induced in CRPtg/IL-6-sufficient mice by injection of LPS, but surprisingly, not by injection of IL-6. In contrast, in IL-6−/− animals the transgene was not induced by LPS, but a slight increase in serum CRP was evident in sera of mice that received a combination of LPS and IL-6. Based on these observations, the authors concluded that IL-6 was necessary, but not sufficient, for induction of the human CRP gene. It appears that only female transgenic animals were used in these experiments. The possibility that IL-6-induced expression of the transgene, like its basal expression (22), exhibits sexual dimorphism or is influenced by testosterone was not determined. In this study, by selective breeding with IL-6−/− mice (24) we generated CRPtg/IL-6−/− hybrids and investigated these possibilities. We show that in both sexes, absence of IL-6 abolishes the ability of the human CRP transgene to respond to LPS. Furthermore, using combinations of passively administered IL-1, IL-6, and testosterone, we show also that in CRPtg/IL-6−/− mice the transgene is not responsive to IL-1 and fully responds to IL-6 only in the presence of testosterone. By comparison, SAP basal or induced expression is not influenced by testosterone. LPS-induced SAP response is reduced but not abolished in IL-6−/− mice and IL-6 is unable to induce SAP unless administered in combination with IL-1.
Materials and Methods
All mice used in this study were fed and watered ad libitum, and barrier maintained under a 12-h light-dark cycle according to protocols established by the Animal Resources Program at the University of Alabama at Birmingham. We have described previously the establishment of a breeding colony of C57BL/6J CRPtg mice (22). CRPtg mice carry a 31-kb ClaI fragment of human genomic DNA comprised of the CRP gene, 17 kb of 5′-flanking sequence, and 11.3 kb of 3′-flanking sequence (20). After injection of LPS into CRPtg mice, peak levels of serum IL-6 (170 ± 35 ng/ml) are attained by 2 h. This is followed by a human CRP response with peak levels reached by 18 h. Expression of the CRP transgene was reported to be independent of copy number and integration site (21). The generation and genetic background of IL-6−/− mice have been detailed elsewhere (24). IL-6−/− mice are homozygous for a disruption of the fourth exon of the murine IL-6 gene, and produce no detectable serum IL-6 after LPS injection (24).
Female CRPtg mice (8–12 wk old) were crossed with IL-6−/− males to produce IL-6+/− F1 hybrids. These were screened for presence of the CRP transgene using human CRP-specific PCR, performed as described previously (25). CRPtg/IL-6+/− hybrids were backcrossed to the IL-6−/− parental strain. CRP transgenic and nontransgenic, IL-6+/− and IL-6−/− F2 progeny were obtained in the expected Mendellian ratios. All appeared normal and healthy. Semiquantitative PCR (data not shown) revealed no gross differences in the number of copies of the CRP transgene in parental, F1, and F2 CRPtg mice.
Administration of LPS, cytokines, and testosterone
LPS (from Escherichia coli serotype 026:B6) was purchased from Sigma (St. Louis, MO), resuspended in sterile pyrogen-free 0.9% NaC1 (Abbott Laboratories, North Chicago, IL), and injected i.p. at a dose of 25 μg per mouse. Recombinant mouse IL-6 (sp. act., 5.6 × 107 U/mg) and recombinant mouse IL-1β (2.2 × 107 U/mg) (Genzyme, Cambridge, MA), were each injected i.p. at a dose of 500 ng per mouse. Mice treated with testosterone (Innovative Research of America, Toledo, OH) received a single, s.c.-implanted pellet (10 mg of testosterone) designed to release physiologic amounts of the hormone over a 21-d period (22). Autoclaved pellets served as placebos.
Measurement of serum IL-6, CRP, and SAP
Sera from blood samples (50 μl) collected before and 2 and 18 h after injection of LPS were used to measure the concentration of IL-6 (2 h sera), CRP (18 h sera), and SAP (18 h sera) by ELISA. Mouse IL-6 ELISA was performed as described (24), using rat mAb MP5-20F3 and biotinylated mAb MP5-32C11 (PharMingen, San Diego, CA) as capture and detection Abs, respectively. Peroxidase-labeled goat anti-biotin (Vector, Burlingame, CA) was used as the reporter, and murine IL-6 was used to generate standard curves. The lower limit of detection was 10 pg of IL-6 per ml serum. F2 mice were categorized as IL-6+/− or IL-6−/− based on the presence or absence of detectable IL-6 in sera collected 2 h after LPS injection (24). ELISA for CRP used sheep anti-human CRP serum (Cappel; Durham, NC) and anti-CRP murine mAb HD2-4 (26) as the capture and detection Ab, respectively, and affinity-purified human CRP (27) as the standard. The assay has been detailed elsewhere (28) and has a lower limit of detection of 20 ng of human CRP per ml serum. ELISA for mouse SAP was performed as described (27) using sheep and rabbit anti-SAP serum as the capture and detection Ab, respectively, and mouse SAP reference standards from Calbiochem-Novabiochem (San Diego, CA). The lower limit of detection was 25 μg of SAP per ml serum.
All values are reported as the mean ± SEM of at least two experiments. Comparisons of mean values were performed using Student’s t tests.
To evaluate the role of IL-6 in the regulation of CRP gene expression, CRP serum levels were measured in age- and sex-matched CRPtg/IL-6−/− mice before and after LPS injection. CRPtg/IL-6+/− littermates were used as controls. IL-6−/− and IL-6+/− CRPtg male mice expressed equal amounts of CRP constitutively, but acute-phase levels were induced by LPS only in the presence of the functional IL-6 gene (Fig. 1,A). In contrast, neither IL-6−/− nor IL-6+/− female transgenics expressed CRP constitutively. LPS-induced levels in the IL-6+/− females were much lower than those in their male counterparts, and IL-6−/− females did not express CRP after treatment with LPS. By comparison, neither basal nor LPS-induced levels of SAP differed substantially between the sexes (Fig. 1,B). LPS-induced levels of SAP were significantly increased over basal levels in both IL-6−/− and IL-6+/− mice, but in the absence of IL-6, LPS induction was attenuated (Fig. 1 B).
We next examined whether IL-6 alone is sufficient for induction of the CRP transgene. As shown in Figure 2,A, male CRPtg/IL-6−/− mice (n = 4) responded to IL-6 administration by increasing their serum CRP levels to the same extent as the control IL-6+/− males (n = 4) used in the same experiment. Importantly, the IL-6-induced CRP levels measured were equal to those induced previously in the same group of four IL-6+/− males by LPS (240 ± 61 μg/ml). In stark contrast to males, both IL-6-sufficient and IL-6-deficient females failed to express CRP in response to IL-6 (Fig. 2,A). In addition, administration of the same amount of IL-6 to these female mice 30 min after injection of LPS resulted in detectable but very low levels of CRP (Fig. 2,A), whereas no CRP could be detected if IL-6 was administered 30 min before or 2 h after LPS (data not shown). Administration of IL-6 alone caused no significant increase in serum SAP in either male or female mice, but when administered 30 min after LPS injection IL-6 induced acute-phase SAP levels (Fig. 2 B). The combined results indicate that in both sexes IL-6 is necessary for LPS induction of the CRP transgene. In addition, IL-6 alone is sufficient for CRP induction in male mice but not in females. By comparison, despite a clear IL-6 requirement for full induction by LPS, IL-6 alone does not cause significant enhancement of SAP expression in either sex.
The failure of IL-6 to induce the CRP transgene in IL-6-deficient female mice raised the possibility that CRP expression in female as opposed to male mice required an additional LPS-inducible cytokine. The experiment summarized in Figure 3,A demonstrates that IL-1, which is known to be induced by LPS and in turn can induce several acute-phase genes, also failed to produce a CRP response in CRPtg/IL-6−/− female mice. Similar results were obtained in male CRPtg/IL-6−/− mice (data not shown). Combining IL-1 with IL-6 also had no effect on expression of CRP by IL-6−/− female mice. However, in CRPtg/IL-6+/− female mice, IL-1 induced detectable albeit low levels of CRP. Administration of IL-1 alone, like IL-6 alone, failed to induce the SAP gene in IL-6−/− female mice, but the two cytokines injected together induced a SAP response similar to that induced by LPS (compare Figs. 3 B and 1B).
The differential induction of the CRP transgene in male and female IL-6−/− mice by LPS and IL-6 or by combinations of these agents made it likely that acute-phase expression of the transgene in male IL-6−/− mice was dependent on the presence of testosterone. To test this hypothesis, female CRPtg/IL-6−/− and control CRPtg/IL-6+/− mice were implanted with testosterone pellets designed to release constant levels of the hormone over a 21-d period. Controls received pellets devoid of active testosterone. As shown in Figure 4,A, testosterone-treated IL-6−/− female mice expressed basal CRP levels equal to those of male transgenic mice (compare Figs. 4,A and 1A). More importantly, administration of IL-6 to these mice induced significantly higher than basal levels of CRP. For unknown reasons, perhaps reflecting low-level inflammation caused by the implantation surgery, control testosterone-treated female IL-6+/− mice expressed relatively high basal CRP levels. Nevertheless, administration of IL-6 to these animals also induced a significant rise in serum CRP. Basal SAP levels were not affected by testosterone, and there was no significant induction of the SAP gene by IL-6 in testosterone-treated female mice (Fig. 4 B).
The results of this study demonstrate that 1) in mice, the basal expression of the CRP transgene requires testosterone, whereas no such requirement is evident for the endogenous gene encoding SAP, a structurally similar acute-phase protein; 2) IL-6 is necessary and sufficient for induction of the CRP transgene in male mice, but in females induction of the transgene requires, in addition to IL-6, a second message, which can be initiated either by LPS or by testosterone; and 3) IL-1 does not appear to have a significant effect on the induction of the CRP transgene. In contrast, full induction of the mouse SAP gene requires IL-6 together with IL-1.
The transgenic mice used in this study carry a 31-kb human genomic DNA fragment, which in addition to the CRP gene contains large 5′ and 3′-flanking regions. These have been shown to contain all elements necessary for liver-specific, LPS-inducible expression of CRP (20). Also, expression of this transgene has been shown to be independent of integration site and copy number (21, 23). We first observed sexual dimorphism in the expression of this transgene in studies investigating the protective effect of CRP against experimental pneumococcal infection (22). Constitutive expression of the transgene was observed only in male mice, but LPS-induced acute-phase expression of CRP was also seen in female transgenic animals, albeit at levels lower than those attained in males. Castration and reconstitution of male mice demonstrated that testosterone is a prerequisite for constitutive expression. The present data extend those previous observations by demonstrating that IL-6 can induce the transgene to express acute-phase levels of CRP in IL-6-deficient male but not in female mice.
Like the CRP transgene, several hepatically expressed endogenous genes, including the gene for the acute-phase protein α2 macroglobulin (α2M) (29), also display sexual dimorphism in rodents. Of these, sex-limited protein (Slp) has been studied extensively. The gene encoding Slp exhibits 95% nucleotide identity with the mouse complement C4 gene, but it expresses no C4 activity (30, 31, 32). Similar to the behavior of the CRPtransgene, expression of the Slp gene depends on the presence of testosterone (33). However, the effect of testosterone is not direct but is exerted through changes in the secretion of growth hormone (GH) (34), which in rodents has a sexually dimorphic pattern (35). In males, testosterone induces high-amplitude, low-frequency GH pulses, while in females the hormone is secreted in a high-frequency low-amplitude pattern. Consequently, mean plasma levels of GH in males are double those in female mice (36). It has been proposed (34) that a negative control on the Slp gene is relieved by a series of events that are initiated by testosterone and include the highly pulsative secretory pattern of GH (35). A similar testosterone-mediated mechanism could explain the sexually dimorphic pattern of constitutive expression of the CRP transgene in mice. Although the DNA element responsible for the testosterone requirement in mice apparently resides within the human transgene, no sexual dimorphism of CRP expression has been reported in humans, which also indicates that testosterone does not have a direct effect on the transgene. Thus, the regulatory events set in motion by testosterone in mice must include secretion of a mediator, perhaps GH, that has a rodent-specific pattern. In fact, in preliminary experiments, treatment of CRPtg female mice with human GH (0.125 IU injected i.p twice daily for 5 days), a procedure known to elicit a male-specific pattern of expression of mouse Slp (34), induced significant basal expression of human CRP (11.1 ± 4.6 μg/ml, n = 3 mice).
Like Slp, constitutive CRP expression in males likely results from attenuation of repression rather than from activation of an enhancer element. Existing published data provide partial indirect support for these arguments. Importantly, it has been shown that the promoter of the human CRP transgene is constitutively competent, but is under strong negative control by two distal elements, one in the 5′- and another in the 3′-flanking region (21). The two elements contribute independently to the low-level constitutive expression. Transgenic mice carrying constructs lacking one of the two negative elements have higher basal expression than animals carrying the entire 31-kb fragment, while maintaining LPS inducibility. Animals carrying transgenes lacking both elements express high constitutive levels of CRP and are only minimally induced by LPS.
IL-6−/− mice have been used previously to investigate the role of IL-6 in the induction of various murine acute-phase genes (37, 38, 39, 40). These studies demonstrated that IL-6 is absolutely necessary for induction of acute-phase genes following tissue damage or infection with Gram-positive bacteria. On the other hand, LPS could induce acute-phase proteins equally well in IL-6-deficient and sufficient animals, indicating that other inflammatory mediators triggered by LPS could substitute for IL-6. The single exception to this pattern was provided by the α2M gene, which could not be induced in IL-6−/− mice either by tissue damage caused by turpentine or by LPS (39). Our results are in agreement with these data. Mouse SAP was induced by LPS in IL-6−/− mice, albeit to a lesser extent than in the IL-6+/− littermates. In contrast, like α2M, the CRP transgene had an absolute IL-6 requirement for induction by LPS.
The role of IL-6 in the induction of the CRP gene in vitro has been demonstrated previously by several investigators (15, 16, 17, 18, 41), and both cis-acting elements and nuclear factors have been defined (42, 43, 44, 45, 46). Two CAAT enhancer binding protein (C/EBP) isoforms, C/EBPβ and C/EBPδ (45), and the signal transducer and activator of transcription-3 (44), all of which are induced by IL-6, have been shown to bind to different response elements in the proximal region of the CRP promoter, resulting in transcriptional activation. Interestingly, several studies on the regulation of CRP expression by hepatoma cells have shown a synergistic effect between IL-6 and IL-1 for maximal induction of the gene. For example, IL-6-induced synthesis of CRP by the hepatoma cell line NPLC/PRF/5 was shown to be enhanced by the addition of IL-1 (47). Similarly, in the hepatoma cell lines Hep3B and HepG2, IL-6 failed to induce CRP synthesis unless combined with IL-1 (18, 48). In contrast, no direct effect of IL-1 on the CRP transgene could be demonstrated in our experiments. As shown in Figure 3, IL-1 either alone or in combination with IL-6 failed to trigger full CRP responses in IL-6−/− female mice. In additional experiments, no effect of IL-1 could be demonstrated in male IL-6−/− mice (data not shown). We cannot exclude the possibility that IL-1 given in a larger quantity and/or at different times with respect to IL-6 may have induced a CRP response. However, we note that the dose of IL-1 and the regime we used were clearly sufficient to induce the endogenous SAP gene (Fig. 3 B). Results similar to ours were previously obtained in studies of CRP induction in human primary hepatocyte cultures (47). Neither IL-1 nor the IL-1 receptor antagonist had an effect on the IL-6-induced secretion of CRP by these cells. The authors suggested that the effects of IL-1 on CRP production by hepatoma cells probably can be attributed to changes related to the malignant transformation (47). Our results are in agreement with this conclusion. On the other hand, the finding in this and a previous study (23) that LPS induces CRP responses in IL-6-sufficient female transgenic mice, whereas IL-6 fails to induce the transgene in IL-6−/− female mice, indicates that, at least in females, an LPS-induced mediator(s) in addition to IL-6 is necessary for induction of the transgene.
A hypothetical model that best accounts for our results and those reported by other investigators is as follows. The human CRP gene is under strong negative control by two distinct elements, one in the distal 5′- and the other in the distal 3′-flanking area (21). Constitutive expression of the gene requires at least partial release of one of the two negative controls. In transgenic mice, this is likely achieved by GH and/or another unknown mediator(s) comparable to GH, displaying a sexually dimorphic pattern of expression. In humans, this mediator(s) apparently is not under androgen control. The other negative control can be relieved by an LPS-induced mediator(s). Acute-phase induction of the gene only occurs when at least one of the two negative controls is relieved and requires only IL-6. In female mice, this condition is satisfied by IL-6 plus another mediator that can be triggered by LPS. In males, IL-6 alone is sufficient, although the putative LPS-induced mediator(s) can enhance the IL-6 effect.
We thank Paula Kiley for secretarial assistance.
This research was supported in part by National Institutes of Health Grant AI 15607.
Abbreviations used in this paper: APR, acute-phase response; CRP, human C-reactive protein; CRPtg, human CRP transgenic mice; SAP, mouse serum amyloid P; α2M, α2 macroglobulin; Slp, sex-limited protein; GH, growth hormone; C/EBP, CAAT enhancer binding protein.