SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Clinical studies of bone marrow (BM) cell therapy for liver cirrhosis are under way but the mechanisms of benefit remain undefined. Cells of the monocyte-macrophage lineage have key roles in the development and resolution of liver fibrosis. Therefore, we tested the therapeutic effects of these cells on murine liver fibrosis. Advanced liver fibrosis was induced in female mice by chronic administration of carbon tetrachloride. Unmanipulated, syngeneic macrophages, their specific BM precursors, or unfractionated BM cells were delivered during liver injury. Mediators of inflammation, fibrosis, and regeneration were measured. Donor cells were tracked by sex-mismatch and green fluorescent protein expression. BM-derived macrophage (BMM) delivery resulted in early chemokine up-regulation with hepatic recruitment of endogenous macrophages and neutrophils. These cells delivered matrix metalloproteinases-13 and -9, respectively, into the hepatic scar. The effector cell infiltrate was accompanied by increased levels of the antiinflammatory cytokine interleukin 10. A reduction in hepatic myofibroblasts was followed by reduced fibrosis detected 4 weeks after macrophage infusion. Serum albumin levels were elevated at this time. Up- regulation of the liver progenitor cell mitogen tumor necrosis factor-like weak inducer of apoptosis (TWEAK) preceded expansion of the progenitor cell compartment. Increased expression of colony stimulating factor-1, insulin-like growth factor-1, and vascular endothelial growth factor also followed BMM delivery. In contrast to the effects of differentiated macrophages, liver fibrosis was not significantly altered by the application of macrophage precursors and was exacerbated by whole BM. Conclusion: Macrophage cell therapy improves clinically relevant parameters in experimental chronic liver injury. Paracrine signaling to endogenous cells amplifies the effect. The benefits from this single, defined cell type suggest clinical potential. (HEPATOLOGY 2011;)

Chronic liver injury results in scar deposition, hepatocyte loss, and ultimately cirrhosis. The only effective treatment for endstage liver disease is liver transplantation; however, organ demand exceeds available supply. There is, therefore, an urgent need to develop alternative therapies for cirrhosis. BM (bone marrow)-derived cell populations influence the progression and recovery phases of liver fibrosis.1-3 Clinical studies of BM cell therapy for cirrhosis are under way. However, the use of mixed cell populations limits the understanding of mechanisms of action.4 The identification of defined single cell types with beneficial effects will enable rational and predictable therapy.

Macrophages have a broad repertoire of context-dependent immune, inflammatory, trophic, and regulatory actions.5 We have previously shown that upon cessation of chronic liver injury, endogenous macrophages mediate hepatic scar remodeling through local matrix metalloproteinase (MMP) expression.2, 6 BM precursors differentiate into macrophages under the control of colony stimulating factor-1 (CSF-1) via its receptor (CSF-1R). CSF-1 also regulates macrophage proliferation, viability, and phenotypic fate.7 Furthermore, exogenous CSF-1 stimulates macrophage infiltration, improving fibrosis and function in models of renal8 and cardiac9 injury. Developing therapy using cells from the monocyte-macrophage lineage therefore holds promise. In chronic liver injury, hepatocyte proliferation is impaired and liver progenitor cells (LPCs) become activated to supply hepatocytes.10 LPCs are not of BM origin11, 12; however, their activation is influenced by a number of paracrine signals that represent potential targets for BM-derived cell therapy.10, 13

We examined the therapeutic potential of exogenous unmanipulated BM cells, in particular those of the monocyte-macrophage lineage, delivered during chronic liver injury. The intraportal application of differentiated BM-derived macrophages (BMMs) improved liver fibrosis, regeneration, and function. Distinct from our current understanding of endogenous macrophages in postinjury scar resolution, the application of these ex vivo cultured and expanded cells activates a wide range of reparative pathways during ongoing injury, with therapeutic benefit. Importantly, we observed paracrine signaling from the exogenous cells to larger populations of endogenous cells, which amplified their effects. This allowed comparatively modest numbers of donor BMMs to exert whole organ changes—encouraging from a translational perspective.

Materials and Methods

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Preparation and Characterization of Donor Cells.

Femurs and tibias were removed from age-matched, syngeneic male mice. BM cells were extracted and a single-cell suspension prepared by passing the cells through a 40-μm filter (BD Falcon). The Tg(Csf1r-Gfp)Hume (MacGreen) mouse has been characterized.14 Briefly, this transgenic model uses the promoter region of the CSF-1R gene to direct expression of an enhanced green fluorescent protein (EGFP). Flow cytometric analysis of MacGreen mouse BM shows that EGFP colocalizes with CD11b, indicating that transgene expression is confined to myeloid cells. Approximately 50% of EGFP+ BM cells express F4/80.14 EGFP+ BM cells expressing the Gr-1 antigen include Ly-6C+ monocytes and Ly-6G+ granulocytes. Monocytes are physiological precursors of macrophages. Culture with CSF-1 converts Ly-6G+ granulocytes into F4/80+ macrophages.15 Therefore, all macrophage precursor cells within the BM with the potential to respond to CSF-1 (and differentiate into macrophages) express the EGFP reporter, allowing their selection by fluorescence-activated cell sorting (FACS, FACSVantage, Becton and Dickinson). BM-derived macrophages were prepared as described16 by BM culture for 7 days in Teflon pots using Dulbecco's Modified Eagle Medium (DMEM)/F12 medium conditioned with CSF-1 from L929 cells. Diff-Quik staining was performed on cytospin samples. BMM marker expression was analyzed by flow cytometry (FACSCalibur, Becton and Dickinson). Cells were stained using the following preconjugated antibodies: F4/80, CD11b (eBiosciences), Ly-6G (Biolegend), Ly-6C, CD3 and CD19 (BD Pharmingen) with appropriate isotype controls. For phenotypic comparison, naïve BMMs were classically activated (M1) by overnight stimulation with lipopolysaccharide (Sigma, 50 ng/mL) and interferon-γ (Peprotech, 20 ng/mL) or alternatively activated (M2) with interleukin (IL)-4 and IL-13 (both Peprotech, 20 ng/mL).5

Disease Models and Cell Delivery.

Wildtype mice were supplied by Harlan (UK) and housed in a sterile animal facility with a 12-hour dark/light cycle and free access to food and water. All animal experiments were carried out under procedural and ethical guidelines of the British Home Office. Advanced liver fibrosis was induced in adult female mice over a 10- week period by twice weekly intraperitoneal (IP) injection of 0.75 mL/kg carbon tetrachloride (CCl4) dissolved in sterile olive oil. One day after the 12th CCl4 injection (6 weeks), mice from the same cohort were randomly allocated to receive either cell or control medium injections via the hepatic portal vein (HPV). Candidate cells from age- and strain-matched mice were suspended in 0.1 mL of DMEM. CCl4 administration continued for a further 4 weeks. The HPV was accessed by midline laparotomy using aseptic technique. Anesthesia was induced using 1 mg/kg medetomidine and 76 mg/kg ketamine intraperitoneally (IP) and reversed with 1 mg/kg atipamezole subcutaneously (SC). Then 22.5 μg/kg buprenorphine (SC) was given as analgesia.

The following candidate cell types were tested: (1) 1 × 106 unfractionated whole BM cells were given to syngeneic fibrotic C57Bl/6 mice (n = 6, control n = 6). (2) 1 × 106 differentiated BMMs physically disrupted by sonication were given to syngeneic fibrotic C57Bl/6 mice (n = 7, control n = 6) to test whether intact, live BMMs were required for therapeutic effect. BMMs were sonicated twice for 10 seconds at 50% power using a Bandelin sonicator (Bandelin). (3) 1 × 106 macrophage precursor cells sorted from the BM of MacGreen mice14 on a Balb-c background were given to fibrotic Balb-c mice (n = 7, control n = 6). (4) 1 × 106 differentiated wildtype BMMs were given to syngeneic fibrotic C57Bl/6 mice (n = 7, control n = 6). As no male donor BMMs were detected 4 weeks after BMM delivery, donor cells were also tracked by an independent method. BMMs were derived from the BM of constitutively GFP+ mice (TgTP6.3 tau-GFP mice on a CBA/Ca background17) using the same 7-day macrophage differentiation protocol as for wildtype BMMs. The 7 × 106 GFP+ BMMs were given to fibrotic wildtype CBA mice (n = 7, control n = 8).

BMM engraftment was transient; therefore, we examined the early effects of BMMs on fibrotic C57Bl/6 mice. 1 × 106 wildtype BMMs were given after 6 weeks of CCl4 (n = 17, control n = 17). These mice were euthanized 1, 3, or 7 days after BMM delivery.

Additionally, 1 × 106 differentiated BMMs were delivered to mice 8 weeks into a longer schedule of 12 weeks 0.4 mL/kg CCl4 (n = 8, control n = 8). Mice were venesected when euthanized. Harvested livers were split and pieces were snap-frozen in Tissue-Tek OCT Compound (Sakura Finetek) or fixed in formalin.

Immunohistochemistry.

Collagen (Sirius red) and immunostaining were carried out as described.1 Three-μm sections of formalin-fixed tissue were used for single immunostains. MMP-9, collagen 1, Dlk, and α-smooth muscle actin (α-SMA) detection required antigen retrieval with 0.01M sodium citrate pH 6.0; pancytokeratin (PCK) staining additionally required proteinase K solution (125 μg/mL). For Ki67, MMP-13, and GFP detection, slides were treated with Tris-EDTA pH 9.0. Primary antibodies were used at the following dilutions: 1:50 for F4/80 (Abcam), 1:100 for Ly-6G (BD Pharmingen) and collagen 1 (Southern Biotech), 1:150 for Dlk (Abcam), 1:200 for PCK (Dako), 1:500 for Ki67 (Novo Castro), GFP and MMP-9 (both Abcam), 1:800 for MMP-13 (Abcam), and 1:2,000 for α-SMA (Sigma). Secondary antibody was applied at a 1:400 dilution. Appropriate isotype controls were used for each primary antibody. Sections were developed using 3,3′-diaminobenzidine (Dako) then counterstained with Harris' hematoxylin. Frozen sections were used for dual staining with MMP-9 and F4/80 or Ly-6G. Detection was performed with Alexa Fluor 488, 546, and 555 (Invitrogen) followed by mounting using Vectashield with DAPI (Vector Laboratories). TUNEL staining (Promega) was performed on formalin-fixed tissue as per the manufacturer's instructions; dual staining with α-SMA was detected with streptavidin-Alexa Fluor 555 (Invitrogen). Male cells were detected by Y chromosome fluorescent in situ hybridization (FISH) using FITC-labeled Y-chromosome paint (Star-FISH; Cambio) as described.1

Assessment of Tissue Sections.

Stained slides were blinded and a minimum of 20 serial, nonoverlapping fields were photographed at ×200 magnification. Male donor BMMs were detected by Y chromosome FISH. Not all male BMMs in a tissue section will exhibit the nucleus, and therefore permit binding of the Y chromosome probe. Male liver was used to establish the proportion of nonparenchymal cells that bound the probe (54%) and adjust subsequent counts to determine the total number of male donor cells present. For assessment of F4/80, Ly-6G, MMP-9, and MMP-13 staining, positive cells were counted in each field. PCK is a sensitive and validated marker of murine LPCs.18 LPCs were defined as PCK+ cells with typical LPC morphology not directly abutting a lumen (thereby excluding biliary epithelia) as described.18 For α-SMA, collagen I and Sirius red assessment, the percentage staining of the total field was measured using image analysis software (Adobe Photoshop). Measurements are expressed relative to matched control recipient samples from the same timepoint.

Quantification of Protein Levels.

Whole liver protein extracts were quantified by Bradford assay. Samples were used at a concentration of 10 mg/mL. Cytokine concentrations were measured in duplicate using the Bioplex Protein Array System (Bio-Rad) according to the manufacturer's instructions. Data were analyzed using Bio-Plex Manager 3.0 software (Bio-Rad). Protein levels are expressed relative to matched control samples from the same timepoint. Commercial kits were used to measure serum albumin (Randox Laboratories) and alanine aminotransferase (ALT) (Alpha Laboratories).

Hydroxyproline Assay.

Snap-frozen liver samples (≈200 mg) were weighed, hydrolyzed in NaOH, and hydroxyproline content determined as described.19 Absorbance was measured at 550 nm and hydroxyproline content expressed as μg/g liver.

Quantification of Messenger RNA (mRNA) Levels by Real-Time Reverse-Transcription Polymerase Chain Reaction (PCR).

RNA was extracted from whole liver tissue using RNA extraction kits (Qiagen) according to the manufacturer's instructions. Complementary DNA was generated from 1 μg of RNA using the Superscript II kit (Invitrogen). Primers for MMPs-2, 9, 12, and 13, Fizz-1, IL-10, inducible nitric oxide synthase (iNOS), macrophage chemoattractant protein (MCP)-1, mannose receptor, tumor necrosis factor (TNF)-α, and Ym-1 were designed using primer express software (sequences supplied in the Supporting material). Predesigned, validated primer sets for macrophage inflammatory protein (MIP)-1α, MIP-2, KC, MMP-8, hepatocyte growth factor (HGF), insulin-like growth factor-1 (IGF-1), CK-19, and TNF-like weak inducer of apoptosis (TWEAK) were purchased from Qiagen (UK). A predesigned, validated eukaryotic 18S primer/probe set (Applied Biosystems) was used for internal control. Quantitative real-time PCR (qPCR) was performed using Express SYBR Green or TaqMan Express qPCR Supermix (Invitrogen). All reactions were performed in triplicate. Levels are expressed relative to matched control samples from the same timepoint.

Statistics.

Data are presented as mean ± standard error of the mean. Two-tailed Student's t and Mann-Whitney U tests were used to analyze parametric and nonparametric data, respectively using Prism (GraphPad Software) unless otherwise stated.

Results

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

BMM Cell Therapy Improves Murine Liver Fibrosis.

A hierarchical approach to candidate donor cell selection from the monocyte-macrophage lineage was taken. The effects of delivering differentiated macrophages (Fig. 1A-E), macrophage precursors from the BM (Fig. 1F), and unfractionated whole BM were tested. Macrophages were generated by 7 days of BM culture with CSF-1 conditioned medium. Diff-Quik staining confirmed that the injected cells were a morphologically homogenous population of macrophages (Fig. 1A). BMMs possessed the characteristic macrophage cell surface markers F4/80 and CD11b.20 Flow cytometric analysis demonstrated that markers of other leukocyte populations (monocytes, neutrophils, and T and B cells) were not present in significant numbers (Fig. 1B). Donor BMMs were not manipulated and did not conform to either the traditional classically (M1) or alternatively activated (M2) macrophage phenotype (Fig. 1C,D). BMMs expressed antiinflammatory (IL-10), antifibrotic (MMP-13), proregenerative (TWEAK), and chemotactic (MCP-1, MIP-1α, MIP-2) mediators (Fig. 1E) that were subsequently found to be elevated in BMM recipient livers (Figs. 5C, 6C, 7E,F). The 1 × 106 wildtype BMMs delivered to recipient mice resulted in a significant reduction in fibrosis measured by Sirius red quantification (66% of control, P < 0.05, Fig. 2A,B). This effect was confirmed by reduced hydroxyproline content (368.2 ± 41.0 versus 558.8 ± 94.6 μg/g liver, P = 0.05, Fig. 2C) and collagen I staining (73% of control, P < 0.01, Fig. 2D,E). Experiments with GFP+ donor BMMs in an independent strain of wildtype recipients also demonstrated this reduction in fibrosis (Sirius red staining 67% of control, P < 0.05, Fig. 2B, Supporting Fig. 1A). Furthermore, in a 12-week CCl4 injury model, BMMs injected at 8 weeks also reduced fibrosis to 69% of control (n = 8 versus n = 8 controls, P < 0.05).

thumbnail image

Figure 1. The derivation and characterization of candidate donor cells. Macrophages were differentiated from whole BM cultured for 1 week in CSF-1 conditioned media. (A) Diff-Quik staining of these BMMs demonstrates a uniform morphology and staining profile. Original magnification ×200. (B) Flow cytometric analysis revealed that BMMs possessed the typical macrophage markers F4/80 and CD11b. Markers of other leukocyte populations (monocytes, neutrophils, and T and B cells) were minimally detected. (C,D) Gene expression analysis of classical (M1) and alternative (M2) macrophage activation markers revealed that BMMs have a naive phenotype in this regard. (E) qPCR analysis showed that BMMs, prior to delivery, expressed specific antiinflammatory, antifibrotic, proregenerative, and chemotactic mediators compared to untreated fibrotic liver. (F) BM cells expressing the CSF-1 receptor/EGFP transgene in the MacGreen mouse14 represent a population of potential macrophage precursors. These cells were positively selected by FACS (denoted “EGFP+”).

Download figure to PowerPoint

thumbnail image

Figure 2. Cells of the monocyte-macrophage lineage have differential therapeutic effects on liver fibrosis. Advanced liver fibrosis was induced in adult female mice by chronic administration of IP CCl4. (A) Photomicrographs show Sirius red staining for hepatic collagens 4 weeks after wildtype BMMs, macrophage precursor cells, or unfractionated BM cells were delivered to syngeneic fibrotic mice (right-side column). Age- and strain-matched control mice within each cohort received an equal volume of control medium (left column). Original magnification, ×80. (B) Morphometric analysis of Sirius red staining revealed that the delivery of differentiated BMMs (both wildtype and GFP+) caused a reduction in the amount of fibrosis. BM macrophage precursor cells did not significantly reduce the amount of fibrosis. Unfractionated whole BM increased liver fibrosis, whereas dead BMMs lead to a trend towards this. The reduced fibrosis following BMM delivery was confirmed (1-tailed analysis) by (C) hydroxyproline assay and (D) collagen I immunostaining (original magnification ×80) with (E) morphometric analysis (*P ≤ 0.05, **P < 0.01 compared with control recipients; n = 6-8 per group).

Download figure to PowerPoint

In contrast to the effects of 7-day differentiated macrophages, injecting 1 × 106 BM macrophage precursor cells did not significantly reduce fibrosis (P = 0.21, Fig. 2A,B). The 1 × 106 unfractionated whole BM cells increased liver fibrosis to 161% of control (P < 0.05, Fig. 2A,B) and 1 × 106 sonically disrupted BMMs led to a trend of increased liver fibrosis (P = 0.08, Fig. 2B, Supporting Fig. 1B). Therefore, liver fibrosis was exacerbated by unfractionated BM and did not significantly improve following the delivery of BM macrophage precursors. Differentiated BMMs consistently reduced hepatic scar and cell viability was required; the underlying processes are examined in the following experiments.

Transient Engraftment of BMMs in the Fibrotic Liver.

Engraftment of donor BMMs was confirmed using two independent cell tracking techniques. GFP+ BMMs were located by immunostaining sections of wildtype recipient liver for GFP. Male donor BMMs in the female recipient liver were identified by Y chromosome FISH. The majority of identified donor BMMs were located within or closely apposed to the hepatic scar (Fig. 3A). One day after the delivery of 1 × 106 BMMs, the mean number of engrafted donor BMMs was 6.9 per ×200 magnification field by GFP immunostaining. Y chromosome FISH revealed 6.5 donor BMMs (per ×200 field) at day 1, which decreased to 5.3 within the first week. In keeping with the known rapid turnover of hepatic macrophages,21 donor BMMs were not detected 1 month after BMM delivery (Fig. 3B).

thumbnail image

Figure 3. Donor BMMs engraft transiently in the liver. (A) Donor cells were tracked by treating wildtype fibrotic mice with GFP+ BMMs (arrowed). Original magnification ×200. In addition, male donor cells (arrowed) were detected within injured female liver using FISH for the Y chromosome. Original magnification ×320. (B) Quantification of donor cell engraftment by Y chromosome FISH revealed a reduction in number during the first 7 days after BMM delivery. No donor cells were present 4 weeks after infusion (n = 3 per timepoint).

Download figure to PowerPoint

Early Reduction in Myofibroblasts Following BMM Delivery.

A reduction in the number of α-SMA+ myofibroblasts through apoptosis is a key early event during fibrosis resolution.22 The amount of α-SMA staining in the BMM treatment group decreased within the first week (Fig. 4A), falling to 40% of control 7 days after macrophage therapy (P < 0.05, Fig. 4B). Apoptotic myofibroblasts were detected during this reduction (Supporting Fig. 2). The decrease in myofibroblasts was no longer statistically significant 1 month after intervention (P = 0.29), suggesting that the peak antifibrotic effect on the myofibroblast population occurs soon after BMM delivery.

thumbnail image

Figure 4. BMM delivery causes a reduction in hepatic myofibroblasts. (A) Photomicrographs demonstrate the reduction in α-SMA+ myofibroblasts 7 days after BMM delivery. Original magnification ×80. (B) Quantification of α-SMA immunostaining revealed that myofibroblast numbers declined within 7 days of BMM treatment; this effect did not persist 1 month after infusion (*P < 0.05 compared with control recipients per timepoint; n = 5-8 per group).

Download figure to PowerPoint

Up-regulation of Hepatic MMP-Expressing Cells in BMM Recipients.

A critical component of fibrosis resolution is the degradation of extracellular matrix mediated by the MMP family of enzymes. Prior to the reduction in myofibroblasts 7 days after BMM delivery, there were increases in the numbers of cells producing MMP-13 and -9 protein (P < 0.01 and < 0.05, respectively, Fig. 5A,B). These MMP-expressing cells were predominantly located in the hepatic scar. Within 1 day of BMM therapy, whole liver gene expression of MMP-9 was markedly elevated (P < 0.05) alongside trends toward increases in MMP-13 (P = 0.21), MMP-8 (neutrophil collagenase, P = 0.17), and MMP-12 (macrophage metalloelastase, P = 0.08) (Fig. 5C). Serial section analysis indicated that a subset of predominantly scar associated macrophages (SAMs) produced MMP-13 (Fig. 5D). We have previously shown that SAMs are an important cellular source of MMP-13 contributing to scar resolution after liver injury.6 Dual immunostaining revealed the MMP-9 producing cells to be neither donor nor endogenous macrophages (Fig. 5Ei) but endogenous Ly-6G+ neutrophils (Fig. 5Eii). Therefore, the initial donor BMMs caused an increase in the numbers of MMP-producing leukocytes in the hepatic scar.

thumbnail image

Figure 5. BMM delivery causes the hepatic up-regulation of MMPs-13 and -9. (A) Immunohistochemical analysis of livers 1 day after BMM delivery revealed infiltration of MMP-13 and MMP-9+ cells (arrowed). Original magnification ×320 and ×200, respectively. (B) The numbers of MMP-13 and -9-expressing cells were greater than in control recipients. (C) Whole liver gene expression of MMP-9 was significantly elevated in BMM recipients at this timepoint. Gene expression levels of MMPs-8, -12, and -13 were not significantly increased. By day 3, these levels had decreased (*P < 0.05, **P < 0.01 compared with control recipients per timepoint; n = 5-6 per group). (D) Immunostaining of serial liver sections indicates that a subset of scar associated F4/80+ macrophages produced MMP-13 (arrowed). Original magnifications ×80 and ×320. (E i) Dual staining for MMP-9 (red) and F4/80 (green) demonstrates that hepatic macrophages (arrowheads) did not express MMP-9 (arrows) in BMM recipients. Original magnification ×1000. (E ii) Colocalization (arrows) of the neutrophil marker Ly-6G (red) with MMP-9 staining (green) indicates that scar associated hepatic neutrophils were expressing MMP-9 in BMM recipients. Original magnification ×200.

Download figure to PowerPoint

BMMs Initiate the Hepatic Recruitment of Circulating Macrophages and Neutrophils.

Within 1 day of BMM infusion, there was a marked change in the cellular composition of the fibrotic liver. F4/80 immunostaining demonstrated a 44% increase in macrophages (P < 0.05, Fig. 6A,B). The absolute increase in macrophage number in BMM-treated mice (from 53 to 76, i.e., an additional 23 per ×200 field) is greater than the number of donor BMMs (mean <7) in the same area of tissue, indicating that the majority of these macrophages were recruited. Ly-6G immunostaining revealed a 242% increase in hepatic neutrophils (P < 0.01, Fig. 6A,B).

thumbnail image

Figure 6. Host leukocytes are recruited to the liver following BMM delivery. (A) Immunohistochemical analysis of livers 1 day after BMM delivery revealed that macrophages and neutrophils were recruited to the injured liver. Original magnification ×200. (B) The degree of cell influx was greater than in control tissue. (C) Whole liver protein expression of macrophage and neutrophil chemokines (MCP-1, MIP-1α, MIP-2, and KC) was significantly up-regulated 1 day after BMM delivery. By day 3 these levels returned toward baseline (*P < 0.05, **P < 0.01 compared with control recipients per timepoint; n = 5-6 per group).

Download figure to PowerPoint

Analysis of whole liver protein from this timepoint revealed that BMM recipients had significantly higher levels of several chemokines expressed by the donor BMMs (Figs. 1E, 6C). The macrophage chemoattractant MCP-1 (CCL2) was increased to 160% (P < 0.001), whereas MIP-1α (CCL3) was 137% of control (P < 0.05). The neutrophil chemoattractants KC (CXCL1) and MIP-2 (CXCL2) were also strongly up-regulated (242%, P < 0.001 and 842%, P < 0.01, respectively). Whole liver protein levels of the antiinflammatory cytokine IL-10 were elevated to 346% in BMM recipients (P < 0.05), whereas proinflammatory mediators such as IL-6 and TNF-α were unchanged (Fig. 7F). Four weeks after BMM delivery, serum ALT levels were not significantly reduced in recipient mice (399.2 ± 120.7) compared to controls (505.7 ± 91.7 u/l, P = 0.5).

thumbnail image

Figure 7. BMM therapy activates regenerative pathways. (A) Serum albumin was increased 4 weeks after the delivery of wildtype or GFP+ BMMs to fibrotic mice. (B) Photomicrographs show Ki67, PCK, and Dlk staining of liver tissue 7 days after BMM therapy. Original magnification ×200. (C) The number of Ki67+ hepatocytes did not increase significantly after BMM therapy. (D) Increased numbers of pancytokeratin-positive liver progenitor cells were detected 7 days after BMM infusion. This expansion was not maintained at 4 weeks. (E) Gene expression of HGF was not significantly elevated in BMM recipients. The liver progenitor cell mitogen TWEAK was up-regulated 3 days after BMM delivery. IGF-1 mRNA was increased 3 and 7 days after BMM treatment. (F) Whole liver protein levels of IL-10 and CSF-1 were elevated 1 day after BMM delivery, whereas IL-6 and TNF-α were unchanged. VEGF protein levels increased after BMM therapy, reaching significance at day 7 (*P ≤ 0.05, **P < 0.01 compared with control recipients per timepoint; n = 5-8 per group).

Download figure to PowerPoint

Therefore, BMM therapy switches the hepatic milieu towards an antiinflammatory cytokine environment while recruiting host macrophages and neutrophils into this altered setting.

BMM Cell Therapy Stimulates Regeneration of the Injured Liver.

Serum albumin was increased in BMM recipients 4 weeks after cell delivery (46.0 ± 2.6 g/l versus 39.9 ± 0.9 g/l, P = 0.05, Fig. 7A). The elevated serum albumin was confirmed in mice receiving GFP+ BMMs (43.3 ± 0.6 g/l versus 40.4 ± 1.0 g/l, P < 0.05, Fig. 7A), suggesting improved regeneration. Hepatocyte proliferation (Ki67+) was not significantly increased after BMM therapy (P = 0.21, Fig. 7B,C). Expression of the hepatocyte mitogen HGF also did not change (Fig. 7E). In keeping with human chronic liver disease, increased numbers of LPCs were present in CCl4-injured mice. Three days after BMM delivery, whole tissue mRNA levels of the LPC marker CK-19 were increased by 55% over control recipients (1.55 ± 0.1 versus 1.00 ± 0.2, P = 0.05). By day 7, there was a periportal expansion of PCK and Dlk+ LPCs in BMM recipients. The number of LPCs increased by 40% over control (P < 0.05, Fig. 7B,D). There was no increase in the level of the cytokines IL-6 and TNF-α which are associated with LPC proliferation10 (Fig. 7F). Donor BMMs used here express high levels of the LPC mitogen TWEAK relative to recipient liver (Fig. 1E). Three days after BMM therapy, at a time when hepatic macrophage numbers were increased, whole liver TWEAK mRNA levels were significantly elevated to 216% of control (P < 0.05, Fig. 7E).

IGF-1 mRNA levels were increased 3 and 7 days after BMM delivery (P < 0.05 and 0.001, respectively, Fig. 7E). CSF-1 protein levels increased to 165% 1 day after BMM delivery (P < 0.01, Fig. 7F) before decreasing over the first week. Vascular endothelial growth factor (VEGF) protein levels increased over this period in BMM recipients, reaching 127% of control at day 7 (P < 0.05, Fig. 7F). In addition to the up-regulation of these reparative factors, the increased TWEAK expression and expanded LPC compartment are also implicated in the improved hepatic function in BMM-treated mice.

Discussion

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Cell therapy based on a defined, homogenous cell population adds clarity to the cause-effect relationship. Importantly for clinical translation, our data reveal that unfractionated BM had a deleterious effect on liver fibrosis. Interestingly, exogenous macrophage precursors did not significantly improve liver fibrosis. Of note, this population contains Gr-1hi (Ly-6Chi) monocytes15 that have profibrogenic actions during liver injury.23 Following culture in CSF-1 conditioned medium, CSF-1R+ macrophage precursors within BM differentiate into macrophages.15 The BMMs used here are a relatively homogenous population of cells without significant contamination from other cell types such as monocytes, granulocytes, and stem cells. The differentiated macrophages generated by this process are antifibrotic and proregenerative in this model. Unmanipulated BMMs cultured in these nonadherent conditions possess neither the typical classically (M1) nor alternatively activated (M2) profiles. Donor BMM engraftment was transient; however, their effects persisted and were amplified by paracrine signaling to host cell populations. The net effect was a reduction in fibrosis and improved regeneration of the injured liver.

BMM therapy caused the recruitment of MMP producing host cells into the hepatic scar. MCP-1 and MIP-1α are members of the CC chemokine subfamily that bind to the CCR2 and CCR1/5 receptors of monocytes, respectively. These interactions contribute to the navigation of monocytes into target tissues during their maturation into macrophages.5 The delivery of MCP-1 and MIP-1α-expressing BMMs to injured mice caused up-regulation of hepatic MCP-1 and MIP-1α and the recruitment of endogenous macrophages. These macrophages produced MMP-13, whose actions include the degradation of fibrillar collagens and gelatin as well activation of other MMPs (such as MMP-9).6 Donor BMMs also express MIP-2 and KC, which are examples of CXC chemokines that recruit neutrophils through the surface receptor CXCR2.5 One day after BMM delivery, hepatic expression of these neutrophil chemoattractants was markedly up-regulated, with elevated hepatic neutrophil numbers. This is in keeping with the role of macrophage-mediated neutrophil recruitment in fibrosis resolution following cessation of cholestatic injury.24 In our model, recruited neutrophils produce MMP-9. MMP-9 overexpression reduces myofibroblast number and inhibits fibrogenesis during experimental liver injury.25 The simultaneous trend of increased MMP-12 (macrophage metalloelastase) and MMP-8 (neutrophil collagenase) expression following BMM therapy reinforces the fibrolytic role of recruited leukocytes. The markedly elevated hepatic IL-10 levels in BMM recipients may modify the behavior of resident and incoming leukocytes and the degree of injury.26 Simultaneous up-regulation of IL-10 and MMPs following BMM therapy may reduce myofibroblast activation26 and promote apoptosis.27 The chemokine-mediated recruitment of host effector cells to the injured liver, importantly at a time when the prevailing hepatic environment is antiinflammatory, represents a novel and realistic mechanism for the therapeutic actions of comparatively few donor cells in the context of the whole organ.

The improved liver function following BMM therapy is multifactorial. There is a less fibrotic cellular milieu, a proregenerative stimulus to LPCs, and elevated levels of cytokines such as CSF-1, VEGF, and IGF-1 that are involved in reparative processes during tissue injury.9, 28, 29 Hepatocyte proliferation was not significantly increased following BMM therapy. There was significant activation of the LPC compartment, compatible with the recent observation that BM infusion transiently stimulated LPCs and improved serum albumin in a series of cirrhotic patients.30 We have previously noted the close spatial relationship between LPCs and endogenous macrophages in vivo.12 The cytokine TWEAK is a member of the TNF superfamily and is currently the only known mitogen that is selective for LPCs but not mature hepatocytes.13 TWEAK acts through its cognate receptor Fn14 to stimulate LPC proliferation. Interestingly, endogenous hepatic macrophages have recently been identified as a cellular source of TWEAK during chronic liver injury.31 Donor BMMs used in our studies expressed high levels of TWEAK and recruited additional host macrophages to the injured liver, supporting the paradigm of donor cell-derived paracrine signals having downstream actions on host cell populations. In addition, we have recently found that hepatic scar degradation promotes LPC activation,32 suggesting that LPC proliferation is also indirectly enhanced by the macrophage-mediated hepatic scar reduction.

In conclusion, we have demonstrated the benefit of BMM therapy upon structural and functional parameters of chronic liver injury. BMMs clearly have multiple actions, some direct and others mediated indirectly through recruitment of host effector cells with antiinflammatory, antifibrotic, and proregenerative results. A number of the mediators reported here have previously been shown to determine the course of experimental liver injury. When overexpressed in isolation, MMP-9,25 IGF-1,29 and IL-1026 have each been shown to reduce myofibroblast numbers and fibrosis in injured liver. CSF-1 also reduces organ fibrosis while improving function.8, 9 MMP-13 knockout impairs fibrosis resolution6 and MMPs-8 and -12 mediate hepatic scar degradation. Overexpression of TWEAK and IL-10 improve LPC proliferation13, 31 and hepatic regeneration,26 respectively. The simultaneous up-regulation of these factors demonstrates the multifaceted effects of cell therapy. This contrasts with studies of single molecules or genes where the effects of the single pathway can be shown. Future work will examine the cellular events underpinning leukocyte recruitment and also activation of progenitor cells within the injured liver following BMM therapy. With regard to clinical translation, the use of a differentiated, readily available, single cell type increases the predictability of effect. The data reported here will inform the rational design of clinical studies to determine the efficacy of autologous cell therapy in chronic liver disease.

References

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Materials and Methods
  4. Results
  5. Discussion
  6. References
  7. Supporting Information

Additional Supporting Information may be found in the online version of this article.

FilenameFormatSizeDescription
HEP_24315_sm_suppinfofig1.tif2803KSupporting Figure 1 Photomicrographs show sirius red staining for hepatic collagens 4 weeks after (A) GFP+ BMMs or (B) dead, sonicated BMMs were delivered to syngeneic fibrotic mice (right sided column). Age and strain-matched control mice within each cohort received an equal volume of control medium (left column). Original magnification, ×80.
HEP_24315_sm_suppinfo.doc33KSupp Info

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.