Synthesis of IL-6 by hepatocytes is a normal response to common hepatic stimuli

Abstract

Exogenous interleukin 6 (IL-6), synthesized at the initiation of the acute phase response, is considered responsible for signaling hepatocytes to produce acute phase proteins. It is widely posited that IL-6 is either delivered to the liver in an endocrine fashion from immune cells at the site of injury, or alternatively, in a paracrine manner by hepatic immune cells within the liver. A recent publication showed there was a muted IL-6 response in lipopolysaccharide (LPS)-injured mice when nuclear NFκB was specifically inactivated in the hepatocytes. This indicates hepatocellular signaling is also involved in regulating the acute phase production of IL-6. Herein, we present extensive in vitro and in vivo evidence that normal hepatocytes are directly induced to synthesize IL-6 mRNAs and protein by challenge with LPS, a bacterial hepatotoxin, and by HGF, an important regulator of hepatic homeostasis. As the IL-6 receptor is found on the hepatocyte, these results reveal that induction of the acute phase response can be regulated in an autocrine as well as endocrine/paracrine fashion. Further, herein we provide data indicating that following partial hepatectomy (PHx), HGF differentially regulates IL-6 production in hepatocytes (induces) versus immune cells (suppresses), signifying disparate regulation of the cell sources involved in IL-6 production is a biologically relevant mechanism that has previously been overlooked. These findings have wide ranging ramifications regarding how we currently interpret a variety of in vivo and in vitro biological models involving elements of IL-6 signaling and the hepatic acute phase response.

Figure 1. IL-6 in serum-free cultured hepatocytes and Kupffer Cells.

(A) Representative RT-PCR depicting IL-6 from a serum-free rat hepatocyte culture (Heps) at 2 h after attachment (also see Figure 4A) and from fresh Kupffer Cells (KCs) at 15 min after attachment. GAPDH was used as a positive control. Cells are from the same animal. (B) Densitometric analyses depicting percent IL-6 mRNA, as compared to GAPDH, with mean ± s.e.m., in hepatocytes and Kupffer cells. n  =  number of independent trials using separate animals. * indicates statistical significance, P = 0.0238, between hepatocytes and Kupffer cells, two-tailed t-test. (C) Immunofluorescent staining showing IL-6 in rat hepatocytes (top, note the comparatively large size and presence of typical bi-nucleated cells) and Kupffer cells (bottom) plated completely serum-free from the same animal. Secondary antibody was conjugated with Cy3 (red). The left panel is a control visualized at the same gating with no primary antibody added. The right panel is a phase contrast image of the cells taken at the same magnification just prior to staining. (D, E) mRNA FISH depicting IL-6 in hepatocytes (D), or Kupffer cells (E) from the same animal. Arrows indicate IL-6 mRNA-negative Kupffer cells. Animal n = 8. Nuclei were stained with Hoescht dye (blue). Probes were conjugated with Alexa 546 (red). Left to right, hybridization with: reagent control (no probe, used for gating), negative control (labeled IL-6 intron), labeled IL-6 cDNA and labeled GAPDH cDNA (positive control). (F) Representative slot blot samples (bottom) and analyses (top, n = 12) of media from freshly plated hepatocytes at 2 h after attachment (T0) or after another 2 h (2 h). * P = 0.0313 significance, using paired two-tailed t-test. Scale bars, 20 µm in images.

Figure 2. IL-6 production by hepatocytes after PHx.

Resting rat livers (A) or remnant livers removed 6 h after PHx (B) were probed for IL-6 mRNA by FISH. Probes, red; Nuclei, blue. Lower panel in B represents a higher magnification from the panel above it. Panels in A and B were taken from hybridizations performed the same day and those at the lower magnifications were gated the same for the picture. Mock hybridizations (no probe) were used as the immunofluorescent gating control. (C) Densitometric analysis of IL-6 on western blots using protein lysates prepared from resting livers (T0) or remnant livers at 6 h post-PHx. Actin was used as a reference control. Animal n = 3. * indicates statistical significance (P = 0.0217), two way t-test. On the right is a representative western blot and RT-PCR from a single animal's whole liver showing the relative changes in protein and mRNA when compared to control (actin for protein, GAPDH for mRNA). Numbers underneath are the numerical change relative to resting liver (T0). (D, E, F) Simultaneous staining for IL-6 (D), albumin (E), or lysozyme C (LysC) (F) proteins (green) and IL-6 mRNA (red) in livers at 6 h after PHx. Co-localization (merge) appears as yellow. Left panels were probed with the IL-6 cDNA and antibody against IL-6, albumin, or lysozyme C. Right panels were probed with the IL-6 intron and secondary antibody only (controls). For D and E, broken line boxes represent the areas magnified in the insets. For F, the initial magnification shown is higher and 2 separate regions containing macrophages have been magnified. Controls were used for gating levels. (G) Immunohistochemical staining for IL-6 in resting rat livers or remnant livers removed 6 h after PHx. Arrows and arrowheads point to hepatocytes and macrophages, respectively. Inset shows magnification of one of the macrophage arrowheads. Scale bars, 20 µm in images.

Figure 3. IL-6 synthesis and NFκB signaling in hepatocytes after LPS injection.

(A) FISH for IL-6 mRNAs in serum-free rat hepatocyte cultures, 15 min after media change with 1 µg LPS/ml or diluent (control). For these photographs, as a baseline level of IL-6 mRNA was known to be present (see 1D), gating was adjusted with the diluent-treated sample serving as the baseline. (B, C) Rat livers were injected with 100 µg/kg LPS or saline (control) and harvested at 4 h post treatment. In B, samples were simultaneously stained for albumin protein (green) and IL-6 mRNA (red). Co-localization (merge) appears as yellow. A presumptive inflammatory cell (expressing IL-6 mRNA but albumin-negative) is indicated by an arrow and featured in magnified inserts. Mock hybridizations (not shown) were used as the immunofluorescent gating control. C shows standard immunohistochemical staining using an antibody against IL-6. (D, E) Representative immunohistochemistry depicting nuclear p65 (D), or p50 (E) with and without LPS treatment. Arrows show nuclei stained with p65 or p50 (brown), arrowheads indicate unstained nuclei (colorless). Simultaneously stained tissue section from liver of ILK-null mice shows positive nuclear localization staining for p50 (left panel of D). Dotted boxes are featured in magnified inserts. Scale bars, 20 µm in all images except for D (40 µm).

Figure 4. HGF-mediated effects on IL-6 production in serum-free hepatocyte cultures.

(A) Representative RT-PCR of IL-6 and GAPDH (control) in serum-free hepatocyte cultures over 25 min, in the presence or absence of 20 or 500 ng HGF/ml. Duplicates shown are separate cultures from a single animal. (B) Summary graph of relative percent IL-6 mRNA to GAPDH in hepatocytes treated with 20 or 500 ng HGF/ml at 1, 15 and 30 min, with mean ± s.e.m. n = 4. * indicates statistical significance, P<0.05, between time point control and designated condition. (C) FISH probing for IL-6 mRNAs in serum-free hepatocyte cultures at 15 min after media change with either 20 ng HGF/ml (bottom panel) or diluent (no treatment, top panel). Mock hybridizations (not shown) were used as the immunofluorescent gating control. (D) Representative western blots for detection of IL-6 in serum-free hepatocyte cultures over a 15 min time period, in the absence or presence of 20 or 500 ng HGF/ml. Animal n = 3. (E) Immunofluorescent staining for IL-6 in serum-free hepatocyte cultures over 15 min, in the presence or absence of 20 or 500 ng HGF/ml. The cells without primary antibody were used as the immunofluorescent gating control. Numbers under figure represent quantifiable increase at 15 min, relative to 1 min staining shown directly above. Scale bars, 20 µm in all images.

Figure 5. NFκB signaling corresponds to changes in IL-6 production in response to HGF.

(A) Confocal staining for the NFκB subunits p50 and p65 in serum-free hepatocyte cultures over 15 min, in the presence or absence of 20 or 500 ng HGF/ml. Actin staining (phalloidin) of the plasma membrane is shown in blue; p50, red; p65, green; with co-localization appearing as yellow. The control cells (no HGF) were used as the immunofluorescent gating control. (B) Summary graph of the percent nuclear staining for NFκB (co-localized staining for p50 and p65) in the experiment shown in A, with mean ± s.e.m. P = 0.0284, significant for one way ANOVA. * indicates statistical significance, P<0.05 from control, Newman-Keuls test. (C) Immunofluorescent staining for the NFκB inhibitor, IκB, in serum-free hepatocyte cultures from the experiment shown in A, at 15 min after stimulation with 0 (control), 20 or 500 ng HGF/ml. The cells without primary antibody were used as the immunofluorescent gating control. Scale bars, 20 µm in all images.

Figure 6. Hepatocyte-specific HGF-mediated effects on IL-6 production in mice after PHx.

(A) Western blot analysis of hepatic NFκB p50 protein expression at 6 h post-PHx in MET flox mice either lacking Cre recombinase (WT) or possessing Cre (KO) is presented, using nuclear-enriched lysates. A representative Ponceau Red-stained protein band, stained prior to antibody probing, is shown for loading comparison. Ponceau-normalized densitometry comparing WT to KO levels of p50 is shown at right. P<0.0001, significant, for two-tail Student's t-test. *** indicates statistical significance, n = 3–5. (B) Standard immunohistochemical staining using the same antibody against p50 as used for the western blot shown in A. Livers removed at 6 h post-hepatectomy from WT (Cre−) or KO (Cre+) animals are shown. Counterstain was deliberately omitted to enhance visualization. Insets represent magnifications of the figure in order to better see the hepatocellular nuclei. (C) Western blot analyses of hepatic IL-6 protein expression in cytoplasmic-enriched lysates for the 6 h PHx experiment described in subfigure A. Densitometric analysis of IL-6 protein expression was performed as described in subfigure A. P = 0.0002, significant, for two-tail Student's t-test. *** indicates statistical significance, n = 4–5. (D) Standard immunohistochemical staining using an antibody against IL-6. Livers extracted at 6 h post-hepatectomy from WT (Cre−) or KO (Cre+) animals are depicted. Insets represent magnifications of the figure in order to better see the hepatocytes and a presumptive macrophage. Scale bars, 20 µm in all images.

Figure 7. Effects of non-specific loss of MET on IL-6 production after PHx in rats.

Livers from rats pre-treated for 24 h with shRNAs to inactivate the HGF receptor (shMET) or control shRNAs (scrambled) were removed at time 0 (T0) or at 1 h after PHx and analyzed for p50 and IL-6. (A) Representative immunohistochemistry depicting nuclear p50 in the shRNA treated animals (scrambled and shMET) at 1 h after PHx. Arrows show nuclei stained with p50 (brown), arrowheads indicate unstained nuclei (counterstain, blue). Scale bars, 20 µm. (B) Densitometric analyses of percent IL-6 mRNA, compared to percent GAPDH control, with mean ± s.e.m., in livers from shRNA-treated animals. P = 0.0001, significant for one way ANOVA. *** indicates statistical significance, P<0.05, from all other samples, Newman-Keuls test. (C) Representative images of immunohistochemistry for IL-6 in the livers of shRNA treated animals at 1 h after PHx. Arrows show IL-6 positive immune cells (stained brown), arrowheads indicate unstained immune cells. Scale bars, 20 µm in all images.