The proinflammatory mediators C3a and C5a are essential for liver regeneration

Abstract

Complement has been implicated in liver repair after toxic injury. Here, we demonstrate that complement components are essential for liver regeneration, and mediate their effect by interacting with key signaling networks that promote hepatocyte proliferation. C3- or C5-deficient mice exhibited high mortality, parenchymal damage, and impaired liver regeneration after partial hepatectomy. Mice with dual C3 and C5 deficiency had a more exacerbated phenotype that was reversed by combined C3a and C5a reconstitution. Interception of C5a receptor signaling resulted in suppression of IL-6/TNFalpha induction and lack of C3 and C5a receptor stimulation attenuated nuclear factor-kappaB/STAT-3 activation after hepatectomy. These data indicate that C3a and C5a, two potent inflammatory mediators of the innate immune response, contribute essentially to the early priming stages of hepatocyte regeneration.

Figure 1.

Histological and immunohistochemical findings in C3^−/−, C5^−/−, C3^−/−C5^−/−, and C5^−/−-treated mice and their controls (Hematoxylin and eosin stainings, magnification 200). (a) The liver section of a control animal 44 h after PHx (representative for all control strains used). A vigorous regenerative response with numerous mitotic figures (▵) and multinucleated hepatocytes (→) can be seen. C3^−/− (b), C5^−/− (c), and C5aRa-treated animals (d) 44 h after PHx display areas of parenchymal damage (♦). Hepatocytes adjacent to necrotic areas display ballooning degeneration (↑). (e–h) Representative areas from sections of C3^−/−C5^−/− livers 44 h after PHx without reconstitution (e), or with reconstitution using C3a (f), C5a (g), or C3a and C5a (h). Severely injured C3^−/−C5^−/− livers (e) are partially protected from this damage after PHx when they are reconstituted with one of the two anaphylatoxins (f and g). When the combination of both is applied, parenchymal damage is completely prevented and the regenerative capacity recovers (h; mitotic figures, ▵).

Figure 2.

Defective hepatocyte S-phase entry after PHx in C3- and C5-deficient mice or after C5aR blockade. Graphs in a, b, d, and f depict numbers of BrdU-labeled hepatocytes (total number of labeled hepatocytes in 10 high-power fields at various times after PHx). (a) BrdU incorporation levels in C3^−/− (n = 9) mice over time are clearly reduced and show a significant difference compared with control (n = 8) at 44 h after PHx (P < 0.05). (b) Reconstitution with either human C3 (n = 3) or C3a (n = 3) is capable of restoring BrdU incorporation to nearly wild-type levels at 44 h after PHx. (c) The low number of BrdU-positive nuclei in C3^−/− mice is paralleled by a delayed gain in liver weight, as measured by the percentage of original liver weight before PHx at harvest time. (d) Numbers of BrdU-labeled hepatocytes in C5^−/− and C5^+/+ mice at various times after PHx reveal a delayed peak of BrdU incorporation. (e) This delay was associated with slower liver weight recovery after PHx. (f) The number of BrdU-positive nuclei is significantly reduced at 44 h after PHx after C5aRa treatment (P < 0.01; BrdU incorporation levels in C5aRa-treated mice [n = 10], white bar; BrdU uptake in control peptide-treated mice [n = 3], lightly shaded bar). The same pattern is observed for animals after anti-C5 antibody treatment (n = 3, black bar) when compared with control IgG treatment (n = 3, darkly shaded bar; P < 0.05).

Figure 3.

S-phase entry by hepatocytes after PHx is abrogated by combined C3 and C5 deficiency and reversed by anaphylatoxin reconstitution. (a) BrdU incorporation into nuclei of hepatocytes from C3^−/−C5^−/− mice is minimal after PHx (n = 16, white bar). Reconstitution with C5a (n = 7) or C3a (n = 4) alone (diagonally striped bar) or with the combination of C3a and C5a (n = 5; shaded bar) reverses the proliferative response partially, as measured by the number of BrdU-positive nuclei. The reconstitution does not reach the proliferative response of C3^+/+C5^+/+ (n = 3, dotted bar) mice. (b) The same pattern can be detected for liver damage at 44 h after PHx. The percentage of parenchymal liver necrosis (quantified with “Scion Image” software; Materials and Methods) is highest in C3^−/−C5^−/− livers without reconstitution (white bar). Single reconstitution with either anaphylatoxin reduces the extent of necrosis slightly (diagonally striped bar). Simultaneous reconstitution with C3a and C5a (black bar) reveals a significant additive protective effect in terms of the extent of necrosis (P < 0.01). (c) The serum levels of the aspartate amino transferase (I.U./ml) mirror and confirm the amount of parenchymal damage measured by histology (b). Combined C3a and C5a reconstitution (black bar) significantly decreases serum transaminase levels when compared with untreated C3^−/−C5^−/− animals (white bar).

Figure 4.

C5aRa blockade decreases TNFa (a) and IL-6 (b) mRNA synthesis in whole liver mRNA extracts during the early phase after PHx. (a) Black bars show the levels of TNFα mRNA synthesis in livers of C5aRa-treated animals in densitometric arbitrary units at 0, 3, and 6 h. Gray bars show the corresponding values for control peptide-treated animals. The quantification was performed on the bases of relative intensities of the corresponding bands in the displayed agarose gels. Normalization was done according to β-actin mRNA intensity in each sample to compensate for loading differences. Control peptide-treated animals showed a constant decrease in TNFα mRNA synthesis over time after PHx. In C5aRa-treated animals, this decrease was more pronounced than in control livers at 3 and 6 h after PHx. (b) Black bars show the relative levels of IL-6 mRNA synthesis in livers of C5aRa-treated animals at 0, 3, and 6 h. Gray bars show the corresponding values for control peptide-treated animals (quantification was performed as described in A). Control peptide-treated animals show a constant approximate fivefold increase in IL-6 mRNA synthesis over time after PHx. C5aRa treatment causes an abrogation of this increase in IL-6 mRNA expression. (mRNA samples from three separate experiments were pooled for each time point).

Figure 5.

C3 deficiency and C5aRa blockade diminishes NF-κB and STAT-3 activation in nuclear liver extracts after PHx. The top depicts NF-κB activation, and the bottom shows STAT-3 activation in both study groups as compared with their controls. Lane 1 represents NF-κB in rabbit reticulocyte lysates (RRL). WT (lanes 3 and 4) and control peptide-treated (lanes 9 and 10) mice show induction of both transcription factors, with the strongest signal detected at 1 or 2 h after PHx. C3 deficiency (lanes 6 and 7) and C5aRa treatment (lanes 12 and 13) diminishes the induction of both factors after PHx, with the most prominent difference at 1 h in C3^−/− mice (NF-κB 54%- and STAT-3 56%-less activation than in control) and 2 h after C5aRa treatment (NF-κB 60%- and STAT-3 37%-less activation than in control). The table summarizes the reduction of activation of both transcription factors in C3^−/− and C5aRa-treated mice when compared with control (% values). The densitometric values from gels of two separate PHx experiments were averaged (Materials and Methods). Lanes 2, 5, 8, and 11 represent the baseline activation of both transcription factors in each group without surgery.

Figure 6.

Summary of the proposed mechanisms by which complement activation products modulate the priming of hepatocytes after PHx. PHx causes among other factors the release of reactive oxygen species (ROS) and LPS. These factors can trigger complement activation either locally, in the portal circulation, or systemically. After complement activation cleavage of C3 or C5 leads to the generation of the potent anaphylatoxins C3a and C5a. Our current results support a regulatory role for these anaphylatoxins in liver regeneration. Their respective G-protein coupled receptors C3aR and C5aR are activated on Kupffer cells (1) and thereby enhance the release of TNFα and IL-6 (2). These cytokines activate NF-κB and STAT-3 in hepatocytes via the corresponding receptors, and, thus, initiate the transcription of immediate early genes. The final transition into G1 phase and the transcription of cell cycle genes is supported by the activation of several transcription factors including AP-1 and CEBP/β (4). Alternatively, the release of IL-6 might induce C5aR activation on hepatocytes after PHx (3). This direct effect of C5a on hepatocytes would further promote their priming.