Hepatocyte growth factor ameliorates acute graft-versus-host disease and promotes hematopoietic function

Abstract

Acute graft-versus-host disease (GVHD) is a major complication of bone marrow transplantation (BMT) and is characterized by hematopoietic dysfunction, immunosuppression, and tissue injury in the skin, liver, and intestinal mucosa. Hepatocyte growth factor (HGF), originally identified and cloned as a potent mitogen for hepatocytes, induces mitogenic and antiapoptotic activity in various epithelial cells and promotes hematopoiesis. Working in a murine model of acute GVHD, we performed repeated transfection of the human HGF cDNA into skeletal muscle and showed that this treatment inhibited apoptosis of intestinal epithelial cells and donor T-cell infiltration into the liver, thereby ameliorating the enteropathy and liver injury caused by acute GVHD. HGF also markedly suppressed IFN-gamma and TNF-alpha expression in the intestine and liver and decreased the serum IL-12. Furthermore, extramedullary hematopoiesis by donor cells was increased, and the survival rate was improved. These results suggest that HGF may be useful for controlling acute GVHD after allogeneic BMT.

Figure 1

Experimental protocol and effect of HGF on intestinal injury. (a) Schedule for induction of GVHD and injection of HGF-HVJ liposomes or PBS. Top: Nonlethal GVHD model. GVHD was induced by injection of 10⁸ spleen cells harvested from B6 donors into nonirradiated BDF₁ mice. From the time of GVHD induction, Gluteal muscles of BDF₁ mice were injected either with HVJ liposomes containing 8 μg of human HGF expression vector (HGF-HVJ liposomes) or with PBS (GVHD control). Gene transfer was repeated once a week after GVHD induction. Bottom: Lethal GVHD model. For the induction of lethal GVHD, recipient BDF₁ mice were exposed to 900 cGy of TBI, and bone marrow cells (5 × 10⁶) plus spleen cells (2 × 10⁷) from B6 donors were injected. HGF gene transfer was repeated once a week for 3 weeks after induction of GVHD. i.m., intramuscularly; i.v., intravenously. (b) Nonirradiated BDF₁ mice were injected intravenously with 10⁸ spleen cells from B6 mice. HGF-HVJ liposomes (8 μg) were administered intramuscularly on day 0 and day 7 after induction of GVHD. Animals were sacrificed after 2 weeks of GVHD, and the intestines were removed. The length of villi in the small intestine was measured using a calibrated lens to study at least 15–20 complete and straight villi per slide in five untreated control mice, five GVHD control mice injected with PBS, and five GVHD mice injected with HGF-HVJ liposomes. Data represent the mean ± SD of five mice. ^AP < 0.05. (c) Hematoxylin and eosin staining of the small intestine in GVHD mice with or without HGF gene therapy. ×200. (d) Apoptosis of small intestinal epithelial cells. The TUNEL method was used to detect apoptotic cells. ×200. (e) Hematoxylin and eosin staining of the large intestine. ×200.

Figure 2

Effect of HGF on liver injury. (a) Hematoxylin and eosin staining of liver sections from GVHD mice with or without HGF gene transduction 2 weeks after induction of GVHD. Liver tissue from GVHD control mice showed cellular infiltration in the periportal area (arrow). ×100. (b) Hematoxylin and eosin staining of liver sections from GVHD mice without HGF gene transduction at 2 weeks after induction of GVHD. ×400. (c) Effect of HGF gene transduction on intrahepatic infiltration of mononuclear cells and donor T cells. Liver-infiltrating mononuclear cells were obtained from untreated mice and GVHD mice with or without HGF gene transduction. The number of donor-derived T cells was determined by multiplying the total mononuclear cell count by the H-2K^d–negative and CD3-positive populations. Data are the mean ± SD of five mice. ^AP < 0.05.

Figure 3

Effect of HGF on the inflammatory cytokine cascade. (a) IL-12 levels in serum from untreated and GVHD mice with or without HGF gene therapy. Serum was obtained 2 weeks after the induction of GVHD. Data represent the mean ± SD of five mice. ^AP < 0.05. (b) IFN-γ and TNF-α mRNA expression in the small intestine, liver, and thymus. Total RNA was isolated from tissue samples of the small intestine, liver, and thymus that were harvested from untreated mice and GVHD mice with or without HGF gene therapy 2 weeks after the induction of GVHD. The cDNAs were generated by RT-PCR and were amplified using primers specific for IFN-γ, TNF-α, and the housekeeping gene β-actin. The amounts of IFN-γ and TNF-α mRNA were normalized for that of β-actin. Expression of cytokine mRNA is shown relative to β-actin, and data represent the mean ± SD of 5 mice. ^AP < 0.05.

Figure 4

Extramedullary hematopoiesis in GVHD mice. (a–i) Histopathological examination was performed using hematoxylin and eosin–stained slides of tissue samples from the spleen (a–c), liver (d–f), and bone marrow (g–i) obtained from untreated mice (a, d, g), PBS-treated GVHD mice (b, e, h), and HGF-treated GVHD mice (c, f, i) 4 weeks after the induction of GVHD. Spleen tissue from HGF-treated GVHD mice showed marked extramedullary hematopoiesis along with numerous megakaryocytes. Liver tissue from HGF-treated GVHD mice showed numerous hematopoietic foci containing granulocyte precursor cells and erythroblasts. ×200; inset, ×400. (j) Spleen cells and liver-infiltrating mononuclear cells were obtained from untreated mice and HGF-treated GVHD mice. The number of donor-derived granulocytes was determined by multiplying the total mononuclear cell count by the H-2K^d–negative and Gr-1–positive populations. Data represent the mean ± SD of five mice. P < 0.05.

Figure 5

HGF caused a reduction of mortality (a) and weight loss (b) in a lethal GVHD model. Recipients were transplanted with 5 × 10⁶ bone marrow cells plus 2 × 10⁷ spleen cells from allogeneic (B6) donors after 9 Gy of TBI. Recipients transplanted with 5 × 10⁶ bone marrow cells from syngeneic (BDF₁) donors after TBI were used as BMT controls. HGF-HVJ liposomes (or PBS) were injected on day 0, 7, 14, and 21. PBS-treated syngeneic BMT (open squares; n = 4), HGF-treated syngeneic BMT (open circles; n = 4), PBS-treated GVHD (filled squares; n = 8), and HGF-treated GVHD (filled circles; n = 8) are indicated. P < 0.01 for survival after PBS injection versus HGF-HVJ liposomes injection and P < 0.05 for body weight from 4 weeks after BMT with PBS versus HGF-HVJ liposomes. Representative data from three similar experiments are shown.

Figure 6

Augmentation of hematopoietic function by HGF in a lethal GVHD model. Recipients were transplanted with 5 × 10⁶ bone marrow cells plus 2 × 10⁷ spleen cells from allogeneic (B6) donors after 9 Gy of TBI. Recipients transplanted with 5 × 10⁶ bone marrow cells from syngeneic (BDF₁) donors after TBI were used as BMT controls. HGF-HVJ liposomes (or PBS) were injected on day 0, 7, 14, and 21. Sixty days after the induction of lethal GVHD, the peripheral blood cell profile, as well as the number of spleen cells, bone marrow cells, and thymus cells, was determined. Data represent the mean ± SD of five mice. ^AP < 0.05. Similar results were obtained in two additional experiments.

Figure 7

Augmentation of hematopoietic function by HGF in a syngeneic BMT model. Recipients were transplanted with 5 × 10⁶ bone marrow cells from syngeneic (BDF₁) donors after 9 Gy of TBI. HGF-HVJ liposomes (or PBS) were injected on days 0 and 7. Ten days after BMT, histological examination was performed (a), and the peripheral blood cell profile, as well as the number of spleen cells, bone marrow cells, and thymus cells (b), was determined. (a) Hematoxylin and eosin staining of the liver and spleen in syngeneic BMT mice with or without HGF gene transduction. Liver tissue from HGF-treated GVHD mice showed hematopoietic foci containing granulocyte precursor cells and erythroblasts (arrow). Spleen tissue from HGF-treated GVHD mice showed marked extramedullary hematopoiesis along with numerous megakaryocytes (arrow). ×200. (b) The peripheral blood cell profile and the number of spleen cells, bone marrow cells, and thymus cells are shown as the mean ± SD of four mice. ^AP < 0.05.