Biophysical characterization of recombinant human Bcl-2 and its interactions with an inhibitory ligand, antimycin A.

Apoptosis is an essential physiological process, regulated by the family of Bcl-2-related proteins. However, the molecular mechanism by which Bcl-2 regulates apoptosis still remains elusive. Here we report the functional studies of recombinant human Bcl-2 with the deletion of 22 residues at the C-terminal membrane-anchoring region (rhBcl-2Delta22). Characterization of rhBcl-2Delta22 showed that the recombinant protein is homogeneous and monodisperse in nondenaturing solutions, stable at room temperature in the presence of a metal chelator, and an alpha-helical protein with unfolding of secondary structure at a T(m) of 62.8 degrees C. Optimal membrane pore formation by rhBcl-2Delta22 required negatively charged phospholipids. The existence of a hydrophobic groove in rhBcl-2Delta22 was demonstrated by the fluorescence enhancement of the hydrophobic ANS probe with which a pro-apoptotic Bak BH3 peptide competed. The respiratory inhibitor antimycin A also bound to the hydrophobic groove of rhBcl-2Delta22 with a K(d) of 0.82 microM. The optimal binding conformation of antimycin A was predicted from molecular docking of antimycin A with the hBcl-2 model created by homology modeling. Antimycin A selectively induces apoptosis in cells overexpressing Bcl-2, suggesting that hydrophobic groove-binding compounds may act as selective apoptotic triggers in tumor cells.

Antimycin A mimics a cell-death-inducing Bcl-2 homology domain 3.

The Bcl-2-related survival proteins confer cellular resistance to a wide range of agents. Bcl-xL-expressing hepatocyte cell lines are resistant to tumour necrosis factor and anti-cancer drugs, but are more sensitive than isogenic control cells to antimycin A, an inhibitor of mitochondrial electron transfer. Computational molecular docking analysis predicted that antimycin A interacts with the Bcl-2 homology domain 3 (BH3)-binding hydrophobic groove of Bcl-xL. We demonstrate that antimycin A and a Bak BH3 peptide bind competitively to recombinant Bcl-2. Antimycin A and BH3 peptide both induce mitochondrial swelling and loss of DeltaPsim on addition to mitochondria expressing Bcl-xL. The 2-methoxy derivative of antimycin A3 is inactive as an inhibitor of cellular respiration but still retains toxicity for Bcl-xL+ cells and mitochondria. Finally, antimycin A inhibits the pore-forming activity of Bcl-x L in synthetic liposomes, demonstrating that a small non-peptide ligand can directly inhibit the function of Bcl-2-related proteins.

Persistent nausea and anorexia after marrow transplantation: a prospective study of 78 patients.

BACKGROUND: Persistent nausea, vomiting, anorexia, and poor oral intake are common after hematopoietic cell transplantation. In the past, herpesvirus infections and acute intestinal graft-versus-host disease (GVHD) were the most common causes. METHODS: We studied 76 patients with 78 episodes of these symptoms to discern the causes. Diagnoses were based on histology of skin and intestinal biopsy specimens, viral cultures, and responses to therapy. RESULTS: The mean day of study entry was day 57+/-31.3 posttransplant. Acute GVHD was the most common cause of symptoms, affecting 63 patients (81%) as the sole cause of symptoms and an additional 4 patients (5%) who had other concurrent causes. Patients with GVHD had marrow donors who were unrelated or HLA-mismatched in 27/63 cases. Gastric edema, erythema, and apoptotic epithelial cells were the most useful findings for the diagnosis of GVHD. Prednisone therapy (1-2 mg/kg/day) was effective in 58 of 63 patients (92%). Infection by herpes simplex virus, cytomegalovirus, or Candida was found in six patients, three of whom had concurrent GVHD. Other causes of symptoms were medications (one patients), parenteral nutrition (one patient), and sagittal sinus thrombosis (one patient). CONCLUSIONS: Acute GVHD is now the dominant cause of persistent nausea and anorexia in marrow transplant patients who are beyond day 20 posttransplant. The diagnosis can be made clinically in most cases and confirmed by endoscopic biopsy of gastric mucosa. Infections, medications, and rare cases of central nervous system disease are much less common.

Lymphocytic gastritis resembling graft-vs.-host disease following autologous hematopoietic cell transplantation.

Although cutaneous graft-vs.-host disease (GVHD) has been noted after autologous hematopoietic cell transplantation, intestinal involvement has not been well documented. We evaluated 197 patients undergoing autologous transplantation for intestinal symptoms; the source for hematopoietic cells was marrow (n=32), peripheral blood stem cells (n=146), or both (n=19). Patients with persistent nausea, vomiting, and anorexia after day 20 underwent upper intestinal endoscopy and mucosal biopsy. Eight patients (4.1%) had diffuse edema, erythema of gastric mucosa, and histological evidence of lymphocytic gastritis with focal apoptosis of crypt epithelial cells-typical of the findings in acute GVHD. All studies for viral, fungal, or bacterial causes were negative. Two patients showed evidence of GVHD in skin and liver, respectively. All patients received 1 mg/kg/day of oral prednisone for 10 days; symptomatic improvement often occurred within days of therapy onset. At the end of corticosteroid treatment, complete resolution of symptoms was seen in all eight patients. In one patient, elevated serum alkaline phosphatase levels gradually normalized over the ensuing 3-4 weeks. When followed up 3 months after treatment, all patients remained symptom-free without evidence of recurrent intestinal symptoms. We concluded that recipients of autologous hematopoietic cells may develop intestinal symptoms caused by a lymphocytic gastritis that is typical of acute GVHD. Patients with this syndrome promptly responded to treatment of a short course of prednisone. The pathogenesis of gastric epithelial damage after autologous transplant is unknown.

Expression of Bcl-2 family during liver regeneration and identification of Bcl-x as a delayed early response gene.

Induction of Bcl-2 and Bcl-x has been demonstrated in mitogen-stimulated lymphocytes in vitro, suggesting that these two apoptosis modulators may also play a role during proliferation. To explore this possibility in a physiological setting, mRNA expression of various Bcl-2 family members was examined during liver regeneration induced by partial hepatectomy, a well characterized in vivo model of cell cycle progression. After a 60% partial hepatectomy in C3H/HeN mice, the steady-state levels of Bcl-x mRNA exhibited a cyclical pattern, with peaks at 4 hours (early G1) and 48 to 72 hours (G1 phase of the second hepatocyte cell cycle). A1 and Bcl-2 mRNA were not detected, and the levels of two Mcl-1 mRNA species remained low without significant changes. The three pro-apoptotic members of the family, Bak, Bad, and Bax, all showed an early decline in mRNA levels when Bcl-x transcripts increased, followed by later peaks at 12, 24, and 48 to 72 hours, respectively. Experiments were subsequently conducted in C3H/HeJ mice, an endotoxin-resistant strain with slower liver regeneration marked by a protracted G1 phase. Even though immediate-early gene responses measured by c-myc induction remained intact, the timing of Bcl-x mRNA expression was delayed in C3H/HeJ mice. When C3H/HeN mice were pretreated with cycloheximide before hepatectomy, the early peak of Bcl-x mRNA at 4 hours was essentially abrogated whereas the immediate-early gene c-myc was hyperinduced, thus implicating Bcl-x as a delayed early response gene during liver regeneration. Bcl-x was localized in hepatocytes and by both immunohistochemistry and Western blot analysis, Bcl-xL protein reached highest levels at 12 hours (mid-G1), consistent with the expression of a delayed early gene. In summary, the expression profiles of Bcl-2 family members during liver regeneration suggest a cell-cycle-dependent regulation as well as a physiological role for these apoptosis-modulating genes during growth and proliferation.

Hrp12, a novel heat-responsive, tissue-specific, phosphorylated protein isolated from mouse liver.

Previous studies from this laboratory identified a 28-kd nonreducible protein, liver-derived immunoinhibitory factor (LDIF) from the mouse liver. Isolation of this protein resulted in the co-purification of another unique protein called heat responsive protein 12 kd (Hrp12). In contrast to LDIF, Hrp12 was totally reducible to a protein of 12 kd suggesting a dimer. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) purification, followed by sequencing of an in situ cyanogen bromide digest of membrane bound Hrp12, yielded an internal 20-amino acid polypeptide. Degenerate oligonucleotides made from this peptide were used to screen a murine liver complementary DNA (cDNA) library. A 1240-bp cDNA clone was obtained with an internal 521-bp open reading frame (ORF). Sequence analysis of the 173-amino acid ORF of mouse Hrp12 showed a high degree of homology with a 99 amino acid rat liver-kidney perchloric acid-soluble protein (LKPS) and a 136-amino acid perchloric acid soluble rat protein (PSP). Transcripts for Hrp12 were mainly restricted to the liver and kidney in mouse and man. The protein was estimated to be approximately 0.8% of the total liver-soluble cytosolic protein. A zoo-blot probed at moderate stringency with labeled cDNA revealed a strong conservation of the gene in all of the mammalian species tested. Analysis of the protein structure of Hrp12 revealed motifs predicted to be targets for protein kinase C (PKC). More importantly, purified mouse Hrp12 could be phosphorylated in vitro with PKC. The protein had significant similarity to DnaK heat shock protein (Hsp)70 and contained a 54-amino acid stretch with sequence similarity to Hsp90. This prompted us to investigate the heat shock response of Hrp12. Isolated hepatocytes and hepatoma cells were exposed to different heat shock temperatures (39.5 degrees C, 42.5 degrees C, and 44.5 degrees C); and then total RNA was extracted and Northern analysis carried out. The message for this novel protein responded atypically to heat shock. Although the steady-state level of the message increased after heat shock, a marked oscillatory pattern was superimposed on it. In contrast, the steady-state levels of Hsp90 and Hsp70 messenger RNA (mRNA) were found to respond to heat shock in the expected manner. Finally, the amount of Hrp12 protein was also found to increase after heat shock in a manner that was consistent with heat-responsive proteins.

Interferon-alpha prevents endotoxin-induced mortality in mice.

Endotoxins, the lipopolysaccharide (LPS) moieties on the bacterial cell wall, cause many of the pathological features of Gram-negative septicemia. Tumor necrosis factor (TNF), primarily a product of monocyte/macrophages, has been shown to mediate many of the pathophysiological effects of endotoxin. Kupffer cells, the largest macrophage population in the body, release TNF when stimulated by LPS in vitro. A recombinant human hybrid interferon-alpha A/D (rIFN-alpha) markedly inhibited this LPS-elicited TNF production by Kupffer cells. The effects of rIFN-alpha were further tested in C57BL/6 mice receiving a lethal dose (400 micrograms/mouse) of LPS. All LPS-treated mice died within 2 days. Pretreatment with rIFN-alpha 1 h before LPS challenge improved the survival at 3 days to 22% (5/23, p < 0.04). In contrast, rIFN-alpha was more effective when administered 20 min after LPS injection, increasing the survival rate to 81% (13/16, p < 0.0001). TNF mRNA expression in the liver and spleen 50 min after LPS challenge, and plasma TNF 1.5 h after LPS were also reduced by either pretreatment or post-treatment with rIFN-alpha. Subsequently, experiments were carried out to test the efficacy of delayed rIFN-alpha treatment. A significant protective effect was still apparent when rIFN-alpha was administered 6, 10 and even 14 h (81%, 62% and 28% survival, respectively) after LPS challenge when serum TNF levels had already returned to near baseline. These experimental results suggest that rIFN-alpha might have a therapeutic potential for the prevention and treatment of the deleterious effects associated with endotoxemia besides mechanisms initially blocking TNF production.

Hepatic metastasis alters the immune function of murine liver nonparenchymal cells.

To examine the effect of a single hepatic focus of metastatic colon tumor on the immune function of liver non-parenchymal cells (NPCs) from C57Bl/6 mice, we injected 2.5 x 10(5) liver-derived murine colon adenocarcinoma (LD-MCA-38) cells beneath the liver capsule. Three weeks following injection of the tumor cells, the immune function of the NPCs was studied. The NPCs from tumor-bearing mice exhibited increased cytotoxic and proliferative activity. The NPCs from tumor-bearing mice also contained a greater percentage of CD8+ and T-cell receptor gamma/delta+ liver-associated T lymphocytes. Levels of interleukin 6 and tumor necrosis factor were increased in the NPC supernatant, and interleukin 6 levels were increased in serum from tumor-bearing mice. We conclude that the presence of a single hepatic focus of metastatic tumor results in augmented immune function of murine liver NPCs.

Isolation and characterization of a novel liver-derived immunoinhibitory factor.

Cytosolic extracts prepared from perfused whole liver or purified hepatocytes of C57BL/6 mice inhibited interleukin-2--and concanavalin A--induced spleen cell proliferation in vitro. In contrast, cytosolic extracts from purified nonparenchymal liver cells had no effect. Arginase and very-low-density lipoprotein were previously identified as two immuninhibitory substances present in liver cytosolic extracts. We demonstrated, however, that inhibitory activity remained after removal of very-low-density lipoprotein and arginase from liver cytosolic extract by repeated ultracentrifugation and gel filtration chromatography, respectively, suggesting the presence of another inhibitor. Further purification by anion-exchange chromatography and chromatofocusing led to the isolation of a novel liver-derived immunohibitory factor. This liver-derived immunoinhibitory factor is sensitive to pronase digestion and heat and acid treatment; it has an estimated isoelectric point of 8.25. The Mr of liver-derived immunoinhibitory factor is 28 kD as estimated from its migration on sodium dodecyl sulfate-polyacrylamide gel electrophoresis, which is identical under both reducing and nonreducing conditions, indicating a monomeric nature of this protein. Amino acid composition analysis discloses that liver-derived immunoinhibitory factor is relatively rich in glycine and proline residues. Interleukin-2--induced spleen cell proliferation in vitro is inhibited by ths liver-derived immunoinhibitory factor, with a 50% inhibitory dose of 1.4 nmol/L. Furthermore, the biological activity of the liver-derived immunoinhibitory factor is not confined to mouse spleen cells, since the growth of B16 mouse melanoma and H35 rat hepatoma cells is also inhibited. A comparison with other liver-derived immunoinhibitors reported previously supports our claim that the liver-derived immunoinhibitory factor is a novel inhibitory protein.

Endogenous interferon alpha/beta produced by Kupffer cells inhibits interleukin-1, tumor necrosis factor alpha production and interleukin-2-induced activation of nonparenchymal liver cells.

We have previously reported liver-specific interferon (IFN) alpha/beta production by murine Kupffer cells that was not observed with other tissue macrophages incubated in the absence of stimulators such as IFN gamma or lipopolysaccharide (LPS). Consequently, while interleukin-2 (IL-2) alone induced pronounced lymphokine-activated killer (LAK) activity from splenocytes, combination of anti-IFN alpha/beta antibody with IL-2 was required to generate significant LAK activity from nonparenchymal liver cells. This endogenous IFN alpha/beta production by Kupffer cells was not induced by LPS because (a) addition of polymyxin B did not abolish the positive effects of anti-IFN alpha/beta antibody on nonparenchymal liver cells, and (b) similar results were obtained when comparing the responses of LPS-responsive C3HeB/FeJ and LPS-hyporesponsive C3H/HeJ mice. The possibility of hepatotropic infection was also ruled out in that anti-IFN alpha/beta antibody enhanced hepatic but not splenic LAK cell induction in vitro in both conventional and germ-free C3H/HeN mice. IFN alpha/beta played an autoregulatory role by down-regulating the production of IL-1 and tumor necrosis factor alpha by Kupffer cells. However, the augmenting effect of anti-IFN alpha/beta antibody on LAK induction from non-parenchymal liver cells was not mediated through an increase in the level of either IL-1 or TNF alpha, as specific antisera against either cytokine did not abrogate this positive effect. Finally, flow-cytometry analysis showed that IFN alpha/beta significantly diminished the expression of IL-2 receptor alpha chain, indicating an inhibition of LAK cell generation at a relatively early stage of induction.

Suppression of hepatic lymphokine-activated killer cell induction by murine Kupffer cells and hepatocytes.

Murine lymphokine-activated-killer cell activity was readily induced by culturing spleen cells with 10 U/ml of interleukin-2 for 4 days. In contrast, very little activity was generated under the same culture conditions when nonparenchymal liver cells were used as the responding cells. It was concluded that Kupffer cells produced prostaglandin and interferon alpha/beta, which suppressed lymphokine-activated-killer induction because (a) induction of lymphokine-activated-killer activity from nonparenchymal liver cells was observed in the presence of indomethacin and anti-interferon alpha/beta antibody; (b) when adherent nonparenchymal liver cells, primarily Kupffer cells, were removed, lymphokine-activated-killer activity could be obtained with interleukin-2 alone; (c) coculture of Kupffer cells with nonadherent nonparenchymal liver cells in a two-chambered system inhibited lymphokine-activated killer cell induction in a dose-dependent manner; (d) exogenous prostaglandin E2 and interferon alpha/beta added at the start of culture inhibited interleukin-2-induced cytotoxicity and proliferation, whereas the other major prostaglandin species in the liver, prostaglandin D2, had little effect. These findings are distinctive with Kupffer cells because splenic macrophages did not exert such inhibition in parallel experiments. Moreover, the supernatant collected from the 24-hr culture of nonparenchymal liver cells contained greater than 20-fold more prostaglandin E2 and interferon alpha/beta than that from culture of spleen cells. In subsequent in vivo experiments, when interleukin-2 was given intraperitoneally to mice, the combination of indomethacin and anti-interferon alpha/beta antibody significantly enhanced lymphokine-activated-killer activity recovered from the liver.(ABSTRACT TRUNCATED AT 250 WORDS)

Role of asialo-GM1 positive liver cells from athymic nude or polyinosinic-polycytidylic acid-treated mice in suppressing colon-derived experimental hepatic metastasis.

Liver-derived (LD) murine colon adenocarcinoma MCA-38 cells injected into the ileocolic vein (ICV) of C57BL/6 mice developed distinct hepatic foci within 14-21 days and survived for an average of 19-35 days. In contrast, C57BL/6-nu/nu mice given injections of LD-MCA cells by the same route did not develop hepatic lesions. Furthermore, 111In-labeled LD-MCA-38 tumor cells were rapidly taken up by the liver of conventional mice within 1 h and 73% of the radioactivity remained after 24 h. However, about 60% of the 111In-labeled LD-MCA-38 tumor cells were cleared from the liver of nude mice after 24 h. Nonparenchymal liver cells isolated from untreated conventional mice displayed little cytotoxicity against freshly excised 51Cr-labeled LD-MCA-38 cells but did lyse the standard natural killer target, YAC-1 tumor cells, in 4 h chromium release assays. On the other hand, nonparenchymal liver cells but not spleen cells from nude mice were cytotoxic to 51CR-labeled LD-MCA-38 in vitro. The nonparenchymal liver cell population responsible for tumor killing was phenotypically nonadherent and asialo-GM1 (AsGM1) positive. C57BL/6 mice treated with polyinosinic-polycytidylic acid [poly(IC)] also displayed cytotoxic activity against LD-MCA-38 tumor cells in vitro. Furthermore, poly(IC) treatment of mice 1-8 days after tumor inoculation suppressed the number of hepatic foci and also significantly increased the life span of tumor-bearing mice. Treatment of athymic nude mice or poly(IC)-treated conventional mice with anti-AsGM1 induced significant numbers of foci and significantly decreased the life span of MCA-38-bearing mice suggesting that AsGM1-positive cells in the liver of these mice may inhibit tumor growth in vivo. In conclusion, the host defense system of the liver from athymic nude or poly(IC)-treated mice possess AsGM1-positive cells that can suppress tumor implantation or tumor growth in the early stages of metastasis in liver.