Functional modification of CD11c+ liver dendritic cells during liver regeneration after partial hepatectomy in mice

Authors

  • Antonino Castellaneta,

    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    Search for more papers by this author
    • Antonino Castellaneta and Alfredo Di Leo contributed equally to this work and should be considered co-first authors.

  • Alfredo Di Leo,

    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    Search for more papers by this author
    • Antonino Castellaneta and Alfredo Di Leo contributed equally to this work and should be considered co-first authors.

  • Ruggiero Francavilla,

    1. Department of Biomedicine of Evolutive Age, Institute of Pediatrics, University of Bari, Italy
    Search for more papers by this author
  • Marcella Margiotta,

    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    Search for more papers by this author
  • Michele Barone,

    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    Search for more papers by this author
  • Annacinzia Amoruso,

    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    Search for more papers by this author
  • Laura Troiani,

    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    Search for more papers by this author
  • Angus W. Thomson,

    1. Departments of Surgery and Immunology, University of Pittsburgh, Pittsburgh, PA
    Search for more papers by this author
  • Antonio Francavilla

    Corresponding author
    1. Department of Emergency and Organ Transplantation, Section of Gastroenterology and Endoscopy, University of Bari, Italy
    • Department of Emergency and Organ Transplantation, Section of Gastroenterology, University of Bari, Padiglione Chini (4th floor), P.zza Giulio Cesare,11, 70124 Bari, Italy
    Search for more papers by this author
    • fax: (39)-080-5593251


  • Potential conflict of interest: Nothing to report.

Abstract

Local immunosuppression within the liver and sex steroid changes, in both blood and tissue during liver regeneration, are well-recognized events. Dendritic cells (DC) play pivotal roles in the induction and regulation of immune responses. Their numbers are expanded markedly in vivo by fms-like tyrosine kinase 3 ligand (Flt3L) administration, without modification of their maturation state. Recent evidence suggests that estrogen can modulate DC function and promote a Th2-type immune response. Few data are available concerning the role of DC in liver regeneration. After 75% partial hepatectomy (PH) in male C57BL/6 mice, CD11c+ liver (L)DC increased significantly within 6 hours and maintained an immature phenotype. Numbers returned to pre-hepatectomy levels by 24 hours. The expanded LDC population showed increased IL-10 and reduced IFN-γ gene transcription. Using these DC compared with control LDC as T cell stimulators in 72-hour mixed leukocyte cultures, IL-10 production was enhanced and IFN-γ production reduced. LDC isolated 6 hours after 75% PH exhibited enhanced estrogen receptor (ER) expression, concomitant with increased serum estrogen levels. By contrast, spleen (S)DC isolated before and after PH showed no significant changes in their function (maturation state, T cell stimulatory activity, cytokine production, and ER expression). Increased liver regeneration (more than 50%) was observed 48 hours after 40% PH in the Flt3L-pretreated compared with the PBS group. In conclusion, interstitial LDC may play a key role in local immune regulation during liver regeneration, possibly linking estrogen-mediated immune modulation and hepatocyte proliferation. (HEPATOLOGY 2006;43:807–816.)

Liver regeneration after reduction in liver mass due to viral, traumatic, or surgical causes is well recognized in rodents, dogs, and humans.1–3 Although various genetic and biological factors that play a role in this phenomenon have been extensively investigated,4–8 the mechanisms responsible for the initiation, maintenance, and suspension of hepatocyte proliferation have not been fully elucidated. Two aspects of liver regeneration that appear to be important are (1) strong inhibition of the cytotoxic activity of liver-resident natural killer (NK) cells, which becomes evident within a few hours after partial hepatectomy (PH) and terminates at the end of the regenerative process9, 10; and (2) a significant increase in circulating estrogen level and expression of estrogen receptors in tissue, regardless of gender or species. Partial hepatectomy in male animals is followed by an almost complete disappearance of serum androgens and their related tissue receptors in a process known as “feminization” of the male regenerating liver.11–14 Inhibition of NK cell activity involves only NK cells within the liver, leaving NK isolated from blood or spleen unaffected. This local protective effect against NK cell activity, which appears to be elicited by changes in the level of MHC class I antigen expression after PH, may be necessary for restoration of the original liver mass.9

In recent years, the role of dendritic cells ([DC] rare, bone marrow–derived, antigen-presenting cells) in the regulation of innate and adaptive immune responses has been well-documented.15–20 Recently, several groups21–23 have shown that estrogen administration leads to clinical improvement in experimental autoimmune encephalomyelitis, a model of multiple sclerosis, due to changes in DC function and the promotion of a Th2 immune response. In the current investigation, we have studied the role of murine DC in the early phase of liver regeneration after PH in normal animals, or those pretreated with the DC poietin Flt3L (fms-like tyrosine-3 ligand, a hematopoietic growth factor that expands dramatically the number of DC in lymphoid and non-lymphoid tissues, including the liver,24 without changing their maturation state).25 In addition, DC modifications were correlated with serum estrogen levels and the expression of estrogen receptors by liver DC (LDC).

We have examined (1) the role of DC in liver regeneration after PH and (2) whether the phenomenon of “feminization” of male regenerating liver also influences hepatic DC function.

Abbreviations

NK, natural killer; PH, partial hepatectomy; DC, dendritic cells; Flt3L, fms-like tyrosine kinase-3 ligand; LDC, Liver dendritic cell; PBS, phosphate-buffered saline; SDC, spleen dendritic cell; RT, room temperature; MAb, monoclonal antibody; PCNA, proliferating cell nuclear antigen; LI, labeling index; PE, phycoerythrin; ER, estrogen receptor; RT-PCR, reverse transcription polymerase chain reaction; IFN-γ, interferon gamma; MLC, mixed leukocyte culture; IL, interleukin; ELISA, enzyme-linked immunosorbent assay; EAE, experimental allergic encephalomyelitis.

Materials and Methods

Experimental Animals.

Male 10- to 12-week-old C57BL/6 JICO (B6) and C3H/HeJ (C3H) mice were purchased from the Charles River Laboratory (Calco, Italy) and maintained in the specific pathogen-free Central Animal Facility of the University of Bari. The animals received Purina rodent chow and tap water ad libitum. Investigations were conducted in accordance with the National Institutes of Health guide for use and care of laboratory animals. B6 mice were used to isolate DC, and C3H animals were used to isolate splenocytes and Th cells to set up mixed leukocyte cultures (see later discussion).

Treatment of Mice with Flt3L.

B6 mice were injected intraperitoneally once daily with Chinese hamster ovary cell-derived recombinant (r) human Flt3L (Amgen, Seattle, WA) in low endotoxin phosphate-buffered saline (PBS; 10 μg/mouse per day) for 10 consecutive days. Tissue was harvested or PH performed 24 hours after the final dose.

Partial Hepatectomy.

B6 mice were submitted to PH or sham operation. The animals were anesthetized with a single intraperitoneal injection of 0.02 mL Avertin (2,2,2-Tribromoethanol, 1.25%, Sigma-Aldrich, Milan, Italy). PH was performed by removing the left lobe (40%) or the left lobe plus the medial lobe, including the gallbladder (75% PH). Forty percent PH, which is characterized by moderate hepatocyte proliferation, was performed to better define the inhibitory or stimulatory effect of in vivo expansion of DC by Flt3L administration on hepatocyte proliferation. All operations were performed between 8:00 and 9:00 AM. In sham-operated animals (control), the liver was manipulated in the same manner as livers of PH-operated animals, then replaced in the abdomen. Animals were killed 2, 6, 12, 24, 48, 72, or 96 hours after the operation.

Isolation and Purification of CD11c+ Liver (L) and Spleen (S) DC.

CD11c+ LDC were isolated from PBS- or Flt3L-treated B6 mice 6, 24, or 48 hours after PH or sham operation, as described.26 Briefly, the liver was chopped into small pieces and digested in type IV collagenase solution (1 mg/mL: Sigma-Aldrich) for 30 minutes at 37°C. The digested tissue was then filtered through a 0.1-mm sterile steel mesh. The liver cell mixture was resuspended and hepatocytes sedimented by centrifugation at 40g. The non-parenchymal cells were recovered from the supernatant and further purified by density centrifugation using 30% v/v metrizamide (Sigma-Aldrich) at 1200g for 20 minutes at 4°C. The low-density interface cell fraction was collected and washed in PBS. The resulting cell suspension was incubated with anti-mouse CD11c-coated immunomagnetic beads (10 μL/107 cells; clone N418; Miltenyi Biotech, Bologna, Italy) for 15 minutes and then positively selected by passage through a paramagnetic LS+ column (Miltenyi Biotech), yielding a highly-enriched (≥95%) CD11c+ population.

CD11c+ SDC were isolated from Flt3L-treated B6 mice 6, 24, or 48 hours after PH or sham operation, as described.27 Briefly, B6 spleens were flushed with 100 U/mL collagenase (type IV, Sigma-Aldrich), teased apart with fine forceps, and digested with 400 U/mL collagenase (30 minutes at 37°C). After digestion, cells were resuspended in ice-cold Ca++-free 0.01 mol/L EDTA Hanks balanced salt solution. Splenic DC-enriched suspensions were obtained by gradient centrifugation (1,040g, 20 minutes at 4°C) over 16% (wt/vol) Histodenz (Sigma-Aldrich). Thereafter, DCs were washed with ice-cold Ca++-free 0.01 mol/L EDTA Hanks balanced salt solution, then labeled with bead-conjugated anti-CD11c monoclonal antibody (10 μL/107 cells; clone N418, Miltenyi Biotec), and immunobead-sorted by positive selection using LS+ separation columns (Miltenyi Biotec), yielding a highly enriched (≥95%) CD11c+ population.

Immunofluorescence Staining of Tissue Sections.

Livers were dissected from B6 mice and small pieces snap frozen in OCT medium (Bio-optica, Milan, Italy). Cryostat sections (7 μm) were fixed in 96% ethanol for 20 minutes at room temperature (RT). After washing and blocking with 5% (v/v) normal goat serum (Sigma-Aldrich) in PBS for 20 minutes, the slides were incubated overnight at 4°C with primary purified Armenian hamster anti-mouse CD11c monoclonal antibody (MAb) (BD PharMingen, Milan, Italy). The next day, after washing in PBS, sections were incubated for 30 minutes at RT with goat cyanine 3(Cy3)-conjugated anti-hamster IgG (Jackson Immuno Research Laboratories, Milan, Italy). The slides were then mounted in Gel/mount medium (Biomeda, Milan, Italy) and sealed. The number of CD11c+ cells per low-power microscopic field (×63) were counted at each time point and the mean value and standard deviation determined.

Proliferating Cell Nuclear Antigen Staining.

Paraffin sections of liver (6 μm) were treated with a methanol/3% H2O2 solution (2/1) for 15 minutes, followed by a brief wash in PBS to block endogenous peroxidase. The sections were preincubated with 1.5% v/v normal horse serum diluted in PBS for 20 minutes and thereafter incubated overnight at RT with anti-proliferating cell nuclear antigen (PCNA) MAb (Santa Cruz Biotechnologies, Santa Cruz, CA) diluted 1:50 in PBS. After draining, the sections were incubated for 60 minutes at RT with biotinylated horse anti-mouse IgG, diluted to 0.5% in PBS containing 1.5% normal horse serum. After incubation for 40 minutes with avidin-biotin-horseradish-peroxidase complex, the sections were washed and incubated with amino-ethyl-carbazole for 15 minutes, counterstained with aqueous hematoxylin, then embedded in Crystal gel mount. Cells that reacted positively with the mouse MAb against PCNA (positive staining) stained red.

Labeling Index for PCNA.

The labeling index (LI) for PCNA corresponded to the percentage of positive nuclei in at least 1,000 cells in randomly selected fields.

Flow Cytometric Analysis of Co-stimulatory Molecule and MHC Class II Antigen Expression.

LDC and SDC were blocked with 5% v/v normal goat serum and then incubated for 30 minutes on melting ice with phycoerythrin (PE)-conjugated anti-CD11c (HL3; all MAbs from BD PharMingen) and with one of the following FITC-conjugated MAbs: anti-IAb β chain (25-9-17), anti-CD40 (3/23), anti-CD80 (16-10A1), or anti-CD86 (GL1). Fluorochrome-conjugated species- and isotype-matched irrelevant MAb were used as negative controls. After washing, the cells were fixed in 4% paraformaldehyde and analyzed using a PARTEC PAS flow cytometer (Partec GmbH, Munster, Germany).

Intracellular Estrogen Receptor-α Staining.

Freshly immunobead-isolated CD11c+ LDC and SDC were blocked with normal goat serum (Sigma-Aldrich), incubated for 30 minutes on ice with PE-conjugated anti-CD11c (HL3) MAb (BD PharMingen), then treated for 30 minutes with 0.5% paraformaldehyde at RT. Cells were then permeabilized using PBS/0.05%Tween-20/0.5% bovine serum albumin (Sigma-Aldrich) for 30 minutes at RT. FITC–anti-estrogen receptor (ER)α (MC-20) polyclonal Ab (Santa Cruz Biotechnology, Santa Cruz, CA) was added and the cells incubated for 1 hour at RT. To verify the specificity of staining, cells were incubated with a known ER-α blocking peptide (Santa Cruz Biotechnology, Inc) for 30 minutes at RT before staining with FITC–anti ER-α and subsequent flow cytometric analysis.

Quantitation of Serum 17β-Estradiol Levels.

Serum 17β-estradiol levels were determined by specific radioimmunoassay, as described.11

Real-Time Polymerase Chain Reaction for Interleukin-10, Interferon-γ, and ER-α.

Real-time polymerase chain reaction (RT-PCR) was performed on an ABI Prism 7900 PCR cycler (Applied Biosystems, Foster City, CA) as described.28 The following validated PCR primers and TaqMan MGB probes (FAM labeled) were used: murine IL-10 (Assay ID: 4329593T), interferon gamma (IFN-γ) (Assay ID: 4339850T), ER-α (Assay ID: 4331182 Hs 00174860) and 18S ribosomal RNA (Assay ID: 4333760F) as endogenous control. PCR mix was prepared according to the manufacturer's instructions (Assay on Demand; Applied Biosystems), and thermal cycler conditions were as follows: 1 × 2 min 50°C, 1 × 10 min 95°C, 40 to 50 cycles denaturation (15 s, 95°C) and combined annealing/extension (1 minute, 60°C). Relative expression of the IL-10, IFN-γ, and ER-α gene were calculated using the ddct method (2-ddCT) after normalization to the endogenous control 18S RNA and calibration to Ftl3L-expanded CD11c+ LDCs as reference.

Mixed Leukocyte Culture.

Freshly immunobead-isolated bulk CD11c+ LDC and SDC, from B6 Flt3L pre-treated B6 mice 6 hours after 75% PH or sham operation (control), were used as stimulators in mixed leukocyte culture (MLC), as described.29 Briefly, antigen-presenting cells were γ-irradiated, then used as stimulators in 72-hour primary MLC. Nylon-wool column-purified allogeneic (C3H) T cells were used as responders. For control purposes, γ-irradiated allogeneic (B6) or syngeneic (C3H) splenocytes were used as stimulators. The supernatants of 72-hour cultures were collected and interleukin-10 (IL-10) and IFN-γ determined by enzyme-linked immunosorbent assay (ELISA) as described later.

ELISA.

IFN-γ and IL-10 were quantified by ELISA in 96-well, round-bottom plates (Bender MedSystems Bouty, Vienna, Austria) with a detection limit of 1.5 to 2.0 pg/mL. The intra- and interassay coefficients of variation of the kits were ≤10%, and every sample was assayed in quadruplicate. For each assay, a standard curve was determined using highly purified, recombinant cytokines calibrated versus the International Reference Standard. The results were calculated according to the manufacturer's instructions and expressed as means ± 1 SD.

Statistical Analyses.

Student's t test was used in the analysis of data. A P value of less than .05 was considered significant.

Results

CD11c+ LDC Are Markedly Increased After 75% PH in Both PBS- and Flt3L- Pretreated Mice.

We first examined the influence of PH on the number of interstitial LDC in either PBS- or Flt3L-injected B6 animals by using immunofluorescence microscopy. Cryostat sections of liver harvested from PBS- or Flt3L-pretreated mice 2 (b, h), 6 (c, i), 12 (d, l), 24 (e, m), and 48 (f, n) hours after 75% PH or sham operation (control: a, g), were stained for the DC marker CD11c, using anti-CD11c MAb. Fig. 1A shows the incidence of CD11c+ cells after PH both in PBS- (a-f) and in Flt3L-pretreated animals (g-n). When compared with controls, we observed a significant, 5-fold increase in mean DC number at 6, but not at 2, 12, 24, or 48 hours after PH in PBS-treated animals (Fig. 1B). Increased numbers of DC were evident both in the portal tract and throughout the parenchyma. As expected, under the same experimental conditions, the number of interstitial DC in the livers of Flt3L pretreated animals was much higher (20- to 75-fold) as compared with PBS values at each point after PH (Fig. 1B). As for the PBS-pretreated mice, a significant elevation was seen in interstitial DC 6 hours after PH in the Flt3L-pretreated group. These data confirm a significant, but transient, increase in the number of CD11c+ LDC in mice 6 hours after 75% PH, both in PBS- and Flt3L-pretreated animals.

Figure 1.

CD11c+ LDC increase significantly during the early stage of liver regeneration. (A) Immunofluorescence staining for the CD11c marker (red) of cryostat sections (A) and mean number ± SD of CD11c+ LDC in sections (B) of liver harvested from PBS- or Flt3L-pretreated mice 2 (b, h), 6 (c, i), 12 (d, l), 24 (e, m), and 48 hours (f, n) after 75% PH or sham operation (control: a, g) were obtained as described in the Materials and Methods. In PBS-treated mice, CD11c+ cells were markedly expanded 6 hours after PH compared with control tissue, whereas no differences were found at the other time points (we show sham-operated controls only at 6 hours, because there were no differences in CD11c+ cells between sham-operated controls at different time points). The number of DC in liver tissue of Flt3L-pretreated animals was much higher than that in PBS controls at each time point (Fig. 1A g-n), with a peak 6 hours after PH (Fig. 1A i). Original magnification ×63. LDC, liver dendritic cell; PBS, phosphate-buffered saline; Flt3L, fms-like tyrosine kinase-3 ligand; PH, partial hepatectomy.

LDC exhibit a stable, immature surface phenotype during liver regeneration after 75% PH in both PBS- and Flt3L-pretreated mice.

To characterize the phenotype of LDC during liver regeneration, cell surface co-stimulatory molecule and MHC class II (IAb) Ag expression were determined on freshly immunobead-isolated CD11c+ LDC from PBS- and Flt3L-pretreated mice 6, 24, and 48 hours after 75% PH or sham operation (control). As shown in Fig. 2, LDC isolated from Flt3L-treated mice exhibited low (CD40, CD86) or moderate (CD80) expression of co-stimulatory molecules, whereas comparatively high levels of surface IAb were determined at each time point examined (similar data were obtained for the PBS group; data not shown). These findings indicate that LDC maintained an immature phenotype during liver regeneration. Moreover, under similar experimental conditions, SDC harvested at each time showed low-moderate (CD40: 14% ± 4%; CD80: 28% ± 2%) or very low (CD86: 5% ± 1%) co-stimulatory molecule expression, and they uniformly exhibited moderate to high levels of IAb (80% ± 3%). These findings demonstrate that during liver regeneration after 75% PH, SDC maintained an immature phenotype, as shown for LDC.

Figure 2.

CD11c+ LDC exhibit a stable, immature phenotype during liver regeneration. LDC were expanded in vivo by Flt3L administration and isolated, as described in the Materials and Methods, from regenerating liver, 0, 6, 24, and 48 hours after 75% PH or sham operation (control). Cells were double-labeled with PE-anti-CD11c and one of the following FITC-conjugated MAbs: anti-CD40, anti-CD80, anti-CD86, or anti-IAb. The figure shows control values only at 6 hours because no difference was found between control values at different times. Results are representative of three separate experiments, with two mice in each experiment. LDC, liver dendritic cell; PBS, phosphate-buffered saline; Flt3L, fms-like tyrosine kinase-3 ligand; PH, partial hepatectomy; MAb, monoclonal antibodies.

IL-10 Gene Expression Is Markedly Upregulated, But IFN-γ Expression Down-regulated in LDC After 75% PH in Both PBS- and Flt3L-Pretreated Mice.

IL-10 and IFN-γ are well-recognized anti- and pro-inflammatory cytokines, respectively, that play important roles in immune regulation. We examined IL-10 and IFN-γ mRNA expression by freshly immunobead-isolated LDC from tissue harvested 6 hoursr after 75% PH or sham operation (control) from PBS- and Flt3L-pretreated mice, using RT-PCR. As shown in Fig. 3, the IL-10 mRNA:18s ratio was dramatically increased in LDC from regenerating tissue at 6 hours as compared with control livers (PBS: 9.2 ± 0.8 vs 0.50 ± 0.15, P < .004; Flt3L: 11.2 ± 0.9 versus 0.42 ± 0.2, P < .0035). By contrast, the IFN-γ mRNA:18s ratio in LDC from regenerating tissue was lower than in control liver (PBS: 0.017 ± 0.008 versus 0.234 ± 0.025, P < .008; Flt3L: 0.022 ± 0.007 versus 0.195 ± 0.03, P < .036). No significant differences were found in the levels of gene expression between Flt3L- and PBS-pretreated mice.

Figure 3.

CD11c+ LDC exhibit increased IL-10 and decreased IFN-γ mRNA transcription in the early stage of liver regeneration after 75% PH. LDC were freshly immunobead-isolated from PBS- or Flt3L pretreated mice 6 hours after 75% PH (6 hours) or sham operation (control). IL-10 and IFN-γ mRNA:18S ratios of CD11c+ LDC were determined by RT-PCR, as described in the Materials and Methods. DC isolated from regenerating liver 6 hours after 75% PH exhibited significantly up-regulated IL-10 (A) and down-regulated IFN-γ gene expression (B) compared with DC isolated from control tissue. No significant differences were found between PBS- and Flt3L-treated animals. Data are representative of three independent experiments, with two/four (Flt3L/PBS group) mice in each experiment. LDC, liver dendritic cell; IL, interleukin; IFN, interferon; PH, partial hepatectomy; PBS, phosphate-buffered saline; Flt3L, fms-like tyrosine kinase-3 ligand.

Naïve T Cell Stimulatory Activity of LDC From Regenerating Liver Is Associated With Enhanced IL-10 But Reduced IFN-γ Production.

In the next series of experiments, we used only DC isolated from Flt3L-pretreated mice, because they had shown the same phenotype and behavior as those isolated from PBS-treated animals, at the same times after PH. Freshly isolated CD11c+ LDC and SDC from B6 mice 6 hours after 75% PH or sham operation (control) were used as stimulators of naive C3H T cells in 72-hour MLC. IL-10 and IFN-γ production was then assessed in the culture supernatants (as described in Materials and Methods). DC isolated from mice undergoing liver regeneration exhibited either moderate (LDC) or comparatively high (SDC) T-cell stimulatory activity (Fig. 4A), similar to that of freshly isolated LDC and SDC from control tissue. Notably, T cell stimulation by LDC isolated from regenerating liver was associated with stronger IL-10 and lower IFN-γ production than LDC from control tissue (IL-10: 11.7 ± 1.5 versus 2.9 ± 0.6 pg/mL; P < .009; IFN-γ: 2.83 ± 0.09 versus 4.66 ± 0.73 pg/mL, P < .01) (Fig. 4B), consistent with polarization toward a Th2-type environment. By contrast, T cell stimulation by SDC isolated from control or 75% hepatectomized animals showed no significant change in IL-10 and IFN-γ production (IL-10: 5.40 ± 2.25 versus 7.30 ± 2.10 pg/mL; IFN-γ:1.57 ± 0.69 versus 1.49 ± 0.63 pg/mL).

Figure 4.

CD11c+ LDC from regenerating liver induce enhanced IL-10 production in 72-hour MLC. (A) γ-Irradiated, freshly immunobead-isolated bulk B6 CD11c+ LDC and SDC harvested 6 hours after 75% PH or sham operation (control) were co-cultured with naive C3H T cells for 72 hours, as described in the Materials and Methods. As negative controls, syngeneic (C3H) splenocytes were used as stimulators. LDC from regenerating or control tissue showed similar, moderate T-cell stimulatory activity. Similar results were obtained with SDC, although they induced a stronger T-cell proliferation compared with LDC. (B) IL-10 and IFN-γ production in the supernatants of MLC cultures set up with CD11c+ LDC or SDC, from either liver regenerating or control mice, and naïve allogeneic T cells, was assessed by ELISA, as described in Materials and Methods. T cell stimulation by LDC from regenerating tissue was associated with a striking increase in IL-10 production, and reduced production of IFN-γ compared with controls. SDC did not induce any significant changes in IL-10 and IFN-γ production in culture with naïve allogeneic T-cells. Data are representative of three independent experiments, with two mice in each experiment. LDC, liver dendritic cell; MLC, MLC: mixed leukocyte culture; IL, interleukin; IFN, interferon; PH, partial hepatectomy; SDC, spleen dendritic cell.

Estrogen Receptor (ER) Expression by LDC Is Enhanced After 75% PH, Concomitant With Elevated Serum 17-β-Estradiol Levels.

Estrogen levels are elevated after PH.11–13 Recent evidence indicates DC express estrogen receptors (ER)21 and estrogens can inhibit DC function and promote their tolerogenicity.21 ER expression was evaluated by flow cytometric analysis and RT-PCR in freshly isolated LDC and SDC, 6 hours after 75% PH or sham operation (control) in Flt3L-pretreated mice. Flow analysis (Fig. 5A) showed that, in the early stage of liver regeneration, a substantial proportion of LDC significantly up-regulate ER expression as compared with LDC from control tissue (25 ± 4 % versus 42 ± 5 %; P < .02). This finding is supported by RT-PCR data (Fig. 5B), which shows a significant increase in ER gene transcription in LDC isolated from regenerating but not from control liver (ER mRNA/18s: 0.84 ± 0.07 versus 0.47 ± 0.06; P < .01). A concomitant and significant increase in serum 17β-estradiol level was also detected 6 hours after PH as compared with basal levels (44 ± 6 vs. 19 ± 5 pg/mL; P < .015) (Fig. 5C). Interestingly, no difference was found between SDC isolated after 75% PH or sham operation in terms of ER gene transcription (ER/18S 1.4 ± 0.19 vs 1.25 ± 0.21) and ER expression (ER 38% ± 5% vs. 39% ± 4%). Thus, the increased number of LDC 6 hours after 75% PH is associated with a concomitant up-regulation of their ER; this effect is restricted to the liver and does not apply to secondary lymphoid tissue (i.e., spleen).

Figure 5.

Estrogen receptor expression by LDC is enhanced 6 hours after 75% PH, concomitant with increased serum 17β-estradiol levels . LDC were expanded in vivo with Flt3L and isolated as described in the Materials and Methods from regenerating liver 6 hours after 75% PH (75% PH; 6 hours) or from normal liver removed during PH (control). (A) After permeabilization using Tween-20, CD11c+ LDC were double-labeled with PE-anti-CD11c and FITC-anti-ER antibodies, as described in Materials and Methods. Flow cytometric analysis showed that DC isolated from regenerating liver display increased ER expression; (B) ER gene transcription was evaluated by RT-PCR, as described in Materials and Methods. Freshly immunobead-isolated LDC obtained 6 hours after 75% PH showed an ER mRNA:18S ratio equal to 0.83 ± 0.07 compared with 0.47 ± 0.06 for LDC isolated from control liver (P < .001); (C) serum 17β-estradiol levels determined by ELISA 6 hours after 75% PH (75% PH; 6 hours) and in control animals. Data are representative of three independent experiments, with two mice in each experiment. LDC, liver dendritic cell; IL, interleukin; IFN, interferon; PH, partial hepatectomy; ER, estrogen receptor; Flt3L, fms-like tyrosine kinase-3 ligand; RT-PCR, reverse transcription polymerase chain reaction.

In Vivo Expansion of LDC by Flt3L Administration Is Associated With Enhanced Hepatocyte Proliferation After PH. Finally, we examined whether enhanced numbers of intrahepatic interstitial LDC (in response to Flt3L administration) was associated with enhanced hepatocyte proliferation. B6 mice pretreated with Flt3L or PBS underwent 40% PH or sham operation. 40% PH, characterized by moderate hepatocyte proliferation, was preferred in these experiments, to better elucidate the inhibitory or stimulatory effect of the growth factor (Flt3L) on hepatocyte proliferation. Liver regeneration was determined at 0, 2, 6, 12, 24, 48, 72, and 96 hours after PH, by PCNA immunostaining. As shown in Fig. 6, liver proliferation peaked at 48 hours after PH, both in PBS- and Flt3L-pretreated animals, as described previously30–32; at 12, 24, 48, and 72 hours after PH the PCNA LI was 45% to 50% higher in Flt3L-pretreated animals than in PBS-injected mice (12 hours: 26 ± 6 vs. 13 ± 6, P < .006; 24 hours: 30 ± 6 vs. 17 ± 4, P < .007; 48 hours: 34 ± 2 vs. 22 ± 3, P < .005; 72 hours: 14 ± 2 vs. 8 ± 2, P < .009). The PCNA LI in liver tissue from Flt3L-treated, sham-operated, non-hepatectomized animals, at each time point, was very low, ruling out a direct effect of Flt3L on hepatocyte proliferation. Thus, expansion of intra-hepatic DC is associated with enhanced hepatocyte proliferation in response to PH.

Figure 6.

Flt3L administration promotes liver regeneration after PH. Liver sections from Flt3L- (Flt3L) or PBS-pretreated (PBS) B6 mice after 40% PH or sham operation (SO) were stained for proliferating cell nuclear antigen (PCNA) and the PCNA LI calculated as described in Materials and Methods. Flt3L-treated mice showed a significant increase in PCNA LI compared with PBS-treated mice 12, 24, 48, and 72 hours after PH, with a peak at 48 hours after PH. Flt3L-treated, sham-operated mice showed, as expected, a very low PCNA LI. Flt3L, fms-like tyrosine kinase-3 ligand; PH, partial hepatectomy; PBS, phosphate-buffered saline; LI, labeling index.

Discussion

The current investigations provide new insight into events associated with liver regeneration, suggesting a role for intrahepatic DC that undergo temporary specific modifications after PH. Our data show that LDC (a) increase in number after PH, peaking after 6 hours, without significant changes to their inherent, immature cell surface phenotype; (b) up-regulate IL-10 gene transcription and concomitantly down-modulate IFN-γ mRNA expression; (c) induce a corresponding (Th2) cytokine production profile in MLC when used as stimulators of naïve allogeneic T cells; (d) exhibit a significant increase in ER expression, parallel to the increase in circulating estrogen levels; and (e) are associated with increased liver regeneration capacity. Moreover, concomitant analysis of SDC showed that this effect is restricted to the liver.

Taken together, these observations suggest that DC may play a key role in events that lead to liver regeneration. An insightful finding is the modification of IL-10/IFN-γ gene transcription in LDC during the early phase of liver regeneration, and the ability of these cells to induce a similar cytokine pattern when co-cultured with naïve T cells. Notably, a decrease in IFN-γ has already been demonstrated in the liver tissue and is associated with inhibition of hepatic NK cell lytic activity,10—a “physiological” event in the process of hepatocyte proliferation after PH. Furthermore, the concomitant increase in expression of IL-10, a well-known anti-inflammatory cytokine, suggests that LDC may be actively involved in promoting a state of local immune suppression. Similar immune modifications (suppression of NK cell function and reduction in IFN-γ production) have been reported by Xu et al.33 in an experimental rat model of post-transplant liver regeneration. Authors have indeed demonstrated that infusion of nuclear factor-κB decoy oligodeoxynucleotide-modified DC (cells unable to undergo maturation and with tolerogenic potential) increases regeneration of the graft. These findings, obtained using an ex vivo modified DC, have been replicated by us under in situ physiological conditions (both in PBS- and Flt3L-treated animals). Results show that in the setting of liver regeneration the immaturity of DC should not be considered an inactive state. In fact, these potentially “tolerogenic” events (increase in IL-10 and decrease in IFN-γ gene transcription) occur in primary MLC culture in presence of LDC isolated from partially hepatectomized animals. According to these data, several authors show that immature DC can induce the development of T cell hyporesponsiveness or T regulatory cells that can promote tolerance,34–36 a finding in keeping with the polarization toward Th2 cytokine production by LDC observed in the current study. SDC isolated at the same time as LDC did not show such changes in their IL-10 and IFN-γ gene transcription or in their ability to modulate cytokine production by naïve T-cells. These findings indicate that functional modification of DC during the course of liver regeneration is restricted to the liver, as demonstrated previously for NK cells.9

DC also may represent a bridge between estrogens and regulation of intrahepatic immune reactivity. Indeed, estrogens have attracted significant interest as potential modulators of T cell–mediated autoimmune disease. Evidence shows that pregnancy is associated with reduced relapse rate of multiple sclerosis,37, 38 rheumatoid arthritis,39 and autoimmune thyroiditis.40 Moreover, in murine experimental allergic encephalomyelitis (EAE), an experimental model of multiple sclerosis, estrogen administration reduces disease severity.21 One interpretation of these results is that estrogen may drive Th2 and suppress Th1 responses, and that this is achieved by multiple interactions with cells of the immune system.21, 23, 41

Our data demonstrate a significant increase in ER expression by DC isolated from livers 6 hours after PH (both in PBS- and Flt3L-treated animals), associated with concomitant increases in serum estrogen. These events occur together with the increased numbers of immature LDC, and with upregulation of IL-10 and downregulation of IFN-γ gene expression by these cells. We speculate that in liver regeneration estrogen-exposed DC may play a role in local immunosuppression, by altering the balance toward a Th2-like microenvironment as shown in EAE.21 We have already shown a relation between immunosuppression and liver regeneration; indeed, the administration of the immune suppressants cyclosporine or tacrolimus (FK506) to 70% hepatectomized animals is associated with significant enhancement of liver regeneration.42–44 Support for this hypothesis comes from our finding that the hepatocyte proliferation index was significantly higher in Flt3L- as compared with PBS-treated animals (Fig. 6), possibly reflecting a state of stronger local immunosuppression in relation to the higher number of LDC.

In conclusion, our experiments provide new evidence of alterations in LDC kinetics and function that may predispose to local immune regulation during liver regeneration, and that may open up new fields of inquiry concerning the relation between estrogens and hepatocyte proliferation through modulation of hepatic DC function.

Acknowledgements

We thank Amgen (Seattle, WA) for providing Flt3L, Dr. Carmen Capobianco for assistance with immunofluorescence imaging analysis, and Dr. Adrian E. Morelli for helpful discussion and suggestions.

Ancillary