Antiangiogenic activity of prostate-specific antigen.

BACKGROUND: Measurement of serum levels of prostate-specific antigen (PSA) is widely used as a screening tool for prostate cancer. However, PSA is not prostate specific, having been detected in breast, lung, and uterine cancers. In one study, patients whose breast tumors had higher levels of PSA had a better prognosis than patients whose tumors had lower PSA levels. To test the hypothesis that PSA may have antiangiogenic properties, we evaluated the effects of PSA on endothelial cell proliferation, migration, and invasion, which are key steps in angiogenesis, the process by which tumors develop a blood supply. METHODS: To assess the antiproliferative effects of PSA, we treated bovine endothelial cells and human endothelial cell lines (HUVEC and HMVEC-d) with purified human PSA (0.1-10 microM) and then stimulated them with 10 ng/mL fibroblast growth factor-2 (FGF-2). Effects on FGF-2- or vascular endothelial growth factor (VEGF)-stimulated endothelial cell migration, invasion, and tube formation were measured by use of one cell line only (HUVEC). PSA was administered to mice at 9 microM for 11 consecutive days after intravenous inoculation of B16BL6 melanoma cells to assess its ability to inhibit the formation of lung colonies (i.e., metastatic tumors). RESULTS: PSA inhibited endothelial cell proliferation, migration, and invasion at IC(50) (i. e., the concentration at which inhibition was 50%) values ranging from 0.3-5 microM. In addition, PSA inhibited endothelial cell responses to both angiogenic stimulators tested, FGF-2 and VEGF. In a mouse model of metastatic disease, daily PSA treatment resulted in a 40% reduction in the mean number of lung tumor nodules compared with phosphate-buffered saline treatment (two-sided P =.003). CONCLUSION: To our knowledge, this is the first report that PSA may function in tumors as an endogenous antiangiogenic protein. This function may explain, in part, the naturally slow progression of prostate cancer. Our findings call into question various strategies to inhibit the expression of PSA in the treatment of prostate cancer.

A recombinant human angiostatin protein inhibits experimental primary and metastatic cancer.

Endogenous murine angiostatin, identified as an internal fragment of plasminogen, blocks neovascularization and growth of experimental primary and metastatic tumors in vivo. A recombinant protein comprising kringles 1-4 of human plasminogen (amino acids 93-470) expressed in Pichia pastoris had physical properties (molecular size, binding to lysine, reactivity with antibody to kringles 1-3) that mimicked native angiostatin. This recombinant Angiostatin protein inhibited the proliferation of bovine capillary endothelial cells in vitro. Systemic administration of recombinant Angiostatin protein at doses of 1.5 mg/kg suppressed the growth of Lewis lung carcinoma-low metastatic phenotype metastases in C57BL/6 mice by greater than 90%; administration of the recombinant protein at doses of 100 mg/kg also suppressed the growth of primary Lewis lung carcinoma-low metastatic phenotype tumors. These findings demonstrate unambiguously that the antiangiogenic and antitumor activity of endogenous angiostatin resides within kringles 1-4 of plasminogen.

Immunotherapy of tularemia: characterization of a monoclonal antibody reactive with Francisella tularensis.

An IgM monoclonal antibody (mAb) recognized surface antigens specific to Francisella tularensis wild-type (Schu4) and live vaccine strain (LVS), and reacted with both in ELISA and slide agglutination tests. This mAb also reacted with LVS microorganisms in tissues of infected mice as assessed by an indirect fluorescence technique. Western blot analysis showed the mAb to react with antigens associated with F. tularensis LPS.

Heat stress alters the virulence of a rifampin-resistant mutant of Francisella tularensis LVS.

We have studied the stress response of a rifampin-resistant mutant of Francisella tularensis LVS. This mutant, Rif 7, was avirulent with an intraperitoneally administered 50% lethal dose greater than 10(7) CFU in a murine model of infection. Exposure of Rif 7 to heat stress for 5 h in vitro resulted in a 2-log decrease in its 50% lethal dose (P < 0.02). The increase in virulence was dependent on the time of exposure to high temperature and was maximal at 5 h. Envelope preparations from heat-stressed cells showed increased levels of several proteins. Notable among these were polypeptides with approximate molecular masses of 16, 60, and 75 kDa. Increases in both virulence and envelope protein levels were reversed when heat-treated cells were subsequently grown at 37 degrees C. Inhibition of protein synthesis by actinomycin D during heat stress blocked the increase in virulence of Rif 7. Cell-free media from the heat-stressed Rif 7 reacted with the whole spectrum of bacterial proteins were not toxic to mice. Hyperimmune serum against Rif 7 reacted with the whole spectrum of bacterial proteins in Western blots (immunoblots), although its reaction with 34- and 45-kDa proteins and two 60- and 75-kDa proteins upregulated during heat stress was weak. Other stress conditions, low iron and low pH, caused similar increases in the virulence of Rif 7. However, examination of the protein profile did not reveal any major common polypeptides induced by different stresses. Heat-treated Rif 7 bacteria were fully able to replicate in macrophages in vitro and in the host tissues, even though heat treatment only partially restored virulence.

Nitric oxide-independent killing of Francisella tularensis by IFN-gamma-stimulated murine alveolar macrophages.

Alveolar macrophages (AMs) were analyzed for ability to support replication of the intracellular bacterium Francisella tularensis live vaccine strain (LVS). AM supported in vitro growth (2 to 3 logs over 5 days) of LVS with a doubling time of 8 +/- 0.8 h. AMs were analyzed for responsiveness to rIFN-gamma for destruction of this lung pathogen. AM treated with 50 U/ml rIFN-gamma allowed early growth of bacteria (six doublings over 48 h) but between 48 and 96 h rIFN-gamma-treated AM eliminated 1.5 logs of LVS. AMs were sensitive to effects of rIFN-gamma; as little as 5 U/ml rIFN-gamma stimulated AM antimicrobial activity, with half-maximal activity 0.3 U/ml. rIFN-gamma-induced antimicrobial effects in AM correlated with amount of nitrites produced, but nitric oxide played only a minimal role in antibacterial effects induced in AM, because NG-MMLA (specific inhibitor of L-arginine-dependent nitric oxide production) failed to block antimicrobial activity of IFN-gamma-stimulated AM. IL-10, TGF-beta 1, and IFN-alpha (cytokines known to regulate effector functions of activated macrophages) also did not block anti-F. tularensis activity of IFN-gamma-stimulated AM. Reactive oxygen metabolites, depletion of tryptophan, and sequestration of iron did not contribute to anti-F. tularensis activity because addition of superoxide dismutase or catalase, excess iron, or tryptophan to IFN-gamma-treated AM did not reverse the anti-F. tularensis activity observed in these cells. Regulation of AM effector activity differed from that of other macrophage populations, in that while rIFN-gamma-stimulated AM produced TNF-alpha (100 U/ml at 72 h), TNF-alpha was not required as a costimulator for induction of antimicrobial activities by rIFN-gamma because anti-TNF-alpha treatment of rIFN-gamma-stimulated AM blocked TNF-alpha but had no effect on either production of nitrites or anti-F. tularensis activity.

Similar molecular requirements for antigen receptor-triggered secretion of interferon and granule enzymes by cytolytic T lymphocytes.

At least two biologically significant responses are triggered by the crosslinking of the T-cell receptor (TcR) on the surface of cloned cytotoxic T lymphocytes (CTL): synthesis and secretion of macrophage-activating factor(s) (MAF) that can be attributed to interferon-gamma (IFN) and release of preformed cytolytic granules. We directly compared the molecular requirements for synthesis and secretion of IFN and secretion of granule enzymes triggered in the same cell by the same activating ligand (antigen or monoclonal antibody (mAb) to TcR). An increase in the surface density of activating ligand (immobilized anti-TcR mAb) enhanced both secretion of IFN and secretion of granules. Secretion of IFN occurred immediately after synthesis: only low (but detectable) levels of IFN were detected in cell cytosolic or particulate fractions isolated from Percoll gradients of lysed CTL, while very high levels of IFN were found in the stimulated CTL culture fluids. Inhibitors of RNA synthesis and protein synthesis blocked secretion of IFN, but did not inhibit release of preformed cytolytic granules. The requirement for TcR crosslinking in triggering both secretion of granules and secretion of IFN from CTL was pharmacologically reproduced by the synergistic action of PMA, a protein kinase C activator, and the Ca2+ ionophore A23187. Both secretion of IFN and secretion of granules were absolutely dependent upon extracellular Ca2+: EGTA completely blocked both TcR- and PMA/A23187-induced secretion of IFN and exocytosis of granules. These studies suggest that similar molecular mechanisms are involved in secretion of newly synthesized IFN and secretion of preformed cytolytic granules. One notable difference between the molecular requirements for the two secretory events was a much lower concentration requirement for PMA for IFN synthesis and secretion than for granule secretion in the synergistic interactions with A23187. Implications of these studies for the exocytosis model of cell-mediated cytotoxicity are discussed.

Susceptibility of inbred mice to Leishmania tropica infection: genetic control of the development of cutaneous lesions in P/J mice.

Leishmania tropica infections of P/J mice are characterized by the development of progressive nonhealing cutaneous lesions, followed by visceral metastases to liver and spleen. To analyze the genetic control of this disease, we produced F1, backcross (BX), and F2 progeny by breeding susceptible P/J mice with L. tropica-resistant C3H/HeN mice. Infections in these hybrid animals suggested that genetic control of the cutaneous lesion was by a single, autosomal, dominant gene. Resistance was the dominant trait. Analysis of liver and spleen impression smears in these animals, however, indicated that development of the cutaneous lesion segregates independently of the second component of L. tropica infections, systemic disease.

Macrophage activation to kill Leishmania tropica: characterization of P/J mouse macrophage defects for lymphokine-induced antimicrobial activities against Leishmania tropica amastigotes.

Macrophages from P/J mice demonstrated both quantitative and qualitative defects in lymphokine (LK)-induced activated macrophage antileishmanial effector reactions: a) these cells recognized the same LK signals that generated resistance to infection in responsive C3H/HeN macrophages, but more signal was required to observe maximal activity; b) LK-induced intracellular destruction of Leishmania tropica by P/J macrophages was minimal (less than 20%), and was induced by only one of three LK signals that regulate antimicrobial activities in C3H/HeN macrophages. The defective microbicidal activity of P/J macrophages observed with LK activation in vitro could also be demonstrated in vivo. Macrophages from P/J mice exposed to the macrophage-activating agent Mycobacterium bovis strain BCG in vivo were capable of restricting the intracellular replication of L. tropica but could not eliminate intracellular parasites, even with further incubation with LK during the 72-hr culture period. The defect of P/J macrophages for intracellular destruction of L. tropica, then, occurred in the activation sequence before the triggering stage that characterizes the macrophage defect of C3H/HeJ mice. Genetic regulation of the P/J macrophage defect appears to be by a single autosomal gene, with defective microbicidal activity as a recessive trait in these animals.

Intracellular replication of Leishmania tropica in mouse peritoneal macrophages: amastigote infection of resident cells and inflammatory exudate macrophages.

C3HeB/FeJ peritoneal exudate cells elicited by a variety of sterile inflammatory agents were exposed to Leishmania tropica amastigotes in vitro. Cytochemical characterization of cells that contained intracellular parasites suggested that young, peroxidase-positive macrophages were more susceptible to infection by amastigotes than more mature cells. Replication of the parasite in these younger cells, however, was similar to that observed in resident peritoneal macrophages.

Susceptibility of inbred mice to Leishmania major infection: genetic analysis of macrophage activation and innate resistance to disease in individual progeny of P/J (susceptible) and C3H/HeN (resistant) mice.

We tested the possibility that two phenotypic traits, defective activation of macrophage antileishmanial activities and susceptibility to infection with Leishmania major, were controlled by the same gene. We used P/J (susceptible) and C3H/HeN (resistant) mice to breed F1, backcross (Bx), and F2 mice that were tested individually for both traits, each of which is known to be controlled by a single autosomal gene. We found no correlation between the macrophage defect and cutaneous disease. There was a correlation between development of systemic disease and defective macrophage activation in Bx mice; this correlation, however, was not confirmed in the F2 population.

Transfer of immunity against lethal murine Francisella infection by specific antibody depends on host gamma interferon and T cells.

Both serum and spleen cells from mice immune to Francisella tularensis transfer protection to naive recipients. Here we characterize the mechanism of protection induced by transfer of immune mouse serum (IMS). IMS obtained 4 weeks after intradermal infection with 10(3) bacteria of the live vaccine strain (LVS) contained high levels of immunoglobulin G2 (IgG2a) and IgM (end point titers, 1:16,600 and 1:7,200, respectively) and little IgG1, IgG2b, or IgG3. LVS-specific antibodies were detected 5 days after intradermal infection, and reached peak levels by 2 weeks postinfection. Only sera obtained 10 days or more after sublethal infection, when IgG titers peaked, transferred protection against a challenge of 100 50% lethal doses (LD50s). Purified high-titer IgG anti-LVS antibody but not IgM anti-LVS antibody was responsible for transfer of protection against an intraperitoneal challenge of up to 3,000 LD50s. IMS had no direct toxic effects on LVS and did not affect uptake or growth of bacteria in association with peritoneal cells. One day after LVS infection, liver, spleen, and lung tissue from mice treated with IMS contained 1 to 2 log units fewer bacteria than did tissue from mice treated with normal mouse serum or phosphate-buffered saline. Between 2 and 4 days after infection, however, bacterial growth rates in tissues were similar in both serum-protected mice and unprotected mice. Bacterial burdens in IMS-treated, LVS-infected mice declined in infected tissues after day 5, whereas control animals died. This lag phase suggested that development of a host response was involved in complete bacterial clearance. In fact, transfer of IMS into normal recipients that were simultaneously treated with anti-gamma interferon and challenged with LVS did not protect mice from death. Further, transfer of IMS into athymic nu/nu mice did not protect against LVS challenge; protection was, however, reconstituted by transfer of normal T cells into nu/nu mice. Thus, "passive" transfer of protection against LVS with specific antibody is not passive but depends on a host T-cell response to promote clearance of systemic infection and protection against lethal disease.

Live vaccine strain of Francisella tularensis: infection and immunity in mice.

The live vaccine strain (LVS) of Francisella tularensis caused lethal disease in several mouse strains. Lethality depended upon the dose and route of inoculation. The lethal dose for 50% of the mice (LD50) in four of six mouse strains (A/J, BALB/cHSD, C3H/HeNHSD, and SWR/J) given an intraperitoneal (i.p.) inoculation was less than 10 CFU. For the other two strains tested, C3H/HeJ and C57BL/6J, the i.p. log LD50 was 1.5 and 2.7, respectively. Similar susceptibility was observed in mice inoculated by intravenous (i.v.) and intranasal (i.n.) routes: in all cases the LD50 was less than 1,000 CFU. Regardless of the inoculation route (i.p., i.v., or i.n.), bacteria were isolated from spleen, liver, and lungs within 3 days of introduction of bacteria; numbers of bacteria increased in these infected organs over 5 days. In contrast to the other routes of inoculation, mice injected with LVS intradermally (i.d.) survived infection: the LD50 of LVS by this route was much greater than 10(5) CFU. This difference in susceptibility was not due solely to local effects at the dermal site of inoculation, since bacteria were isolated from the spleen, liver, and lungs within 3 days by this route as well. The i.d.-infected mice were immune to an otherwise lethal i.p. challenge with as many as 10(4) CFU, and immunity could be transferred with either serum, whole spleen cells, or nonadherent spleen cells (but not Ig+ cells). A variety of infectious agents induce different disease syndromes depending on the route of entry. Francisella LVS infection in mice provides a model system for analysis of locally induced protective effector mechanisms.

Introduction of Francisella tularensis at skin sites induces resistance to infection and generation of protective immunity.

Mice are susceptible to systemic infection with Francisella tularensis strain LVS; thus, the intraperitoneal (i.p.) lethal dose at 50% (LD50) in C3H/HeN and C57BI/6J mice is only a single bacterium, while the intradermal (i.d.) LD50 is more than 10(4). Here we show that the LD50 when LVS is introduced via the skin, either i.d. or subcutaneously (s.c.), ranges from 7 x 10(4) to 2 x 10(6). Sublethal i.d. or s.c. infection (priming) invariably leads to the generation of systemic and specific protective immunity: primed mice survive lethal i.p., intravenous (i.v.), or i.d. challenges of LVS but not Salmonella typhimurium W118 or Escherichia coli 018:K1:H7 strain BORT.

Activation of macrophages for destruction of Francisella tularensis: identification of cytokines, effector cells, and effector molecules.

Francisella tularensis live vaccine strain (LVS) was grown in culture with nonadherent resident, starch-elicited, or Proteose Peptone-elicited peritoneal cells. Numbers of bacteria increased 4 logs over the input inoculum in 48 to 72 h. Growth rates were faster in inflammatory cells than in resident cells: generation times for the bacterium were 3 h in inflammatory cells and 6 h in resident macrophages. LVS-infected macrophage cultures treated with lymphokines did not support growth of the bacterium, although lymphokines alone had no inhibitory effects on replication of LVS in culture medium devoid of cells. Removal of gamma interferon (IFN-gamma) by immunoaffinity precipitation rendered lymphokines ineffective for induction of macrophage anti-LVS activity, and recombinant IFN-gamma stimulated both resident and inflammatory macrophage populations to inhibit LVS growth in vitro. Inflammatory macrophages were more sensitive to effects of IFN-gamma: half-maximal activity was achieved at 5 U/ml for inflammatory macrophages and 20 U/ml for resident macrophages. IFN-gamma-induced anti-LVS activity correlated with the production of nitrite (NO2-), an oxidative end product of L-arginine-derived nitric oxide (NO). Anti-LVS activity and nitrite production were both completely inhibited by the addition of either the L-arginine analog NG-monomethyl-L-arginine or anti-tumor necrosis factor antibodies to activated macrophage cultures. Thus, macrophages can be activated by IFN-gamma to suppress the growth of F. tularensis by generation of toxic levels of NO, and inflammatory macrophages are substantially more sensitive to activation activities of IFN-gamma for this effector reaction than are more differentiated resident cells.

Mycobacterium bovis BCG-induced protection against cutaneous and systemic Leishmania major infections of mice.

We examined the protective effects of Mycobacterium bovis bacillus Calmette-Guérin (BCG) administration on Leishmania major infections of BALB/c and P/J mice. There were two treatment protocols. In the first, the footpads of naive animals were inoculated with mixtures of L. major and BCG (viable or heat killed) or the soluble mycobacterial antigen, purified protein derivative. Viable BCG, but not heat-killed BCG or purified protein derivative, inoculated with L. major amastigotes into the footpads of naive BALB/c or P/J mice protected these animals from the metastatic spread of parasites to the viscera and from ensuing lethal systemic infection. This treatment also induced cures of the cutaneous lesions of P/J mice but not of BALB/c mice. In the second protocol, we induced an immune response to BCG before inoculation of L. major. BCG given intraperitoneally 10 days before infection of footpads with leishmania offered protection against the metastatic spread of amastigotes in both P/J and BALB/c mice, regardless of intralesional treatment, and modulated the severity of cutaneous infection by 30 to 50%. Inoculation of a mixture of viable BCG and L. major amastigotes into BCG-immune mice completely protected both BALB/c and P/J strains from cutaneous disease; we recovered no parasites from the inoculated footpads of these animals. Furthermore, each of the nonspecifically protected mice of both the BALB/c and P/J strains developed immunity to rechallenge with viable L. major. Injection of amastigotes at a site remote from the original lesion, the contralateral footpad, resulted in the complete clearance of parasites in the inoculum with no evidence of either cutaneous or systemic disease over an extended observation period.

Susceptibility of inbred mice to Leishmania tropica infection: correlation of susceptibility with in vitro defective macrophage microbicidal activities.

Eleven mouse strains were inoculated in footpads with amastigotes of Leishmania tropica and observed for 12 weeks. Liver and spleen impression smears from infected mice were examined for the presence of intracellular parasites. Four strains (BALB/cJ, C57L/J, NZW/N, and P/J) failed to heal the subcutaneous lesion and showed evidence of systemic infection; the remaining seven strains (A/J, C3H/HeJ, C3H/HeN, C3HeB/FeJ, C57BL/6J, C57BL/10J, and C57BL/10ScN) were each resistant to infection and resolved their lesions by Week 10. Macrophages from the four susceptible strains could not be activated to kill L. tropica amastigotes by treatment with soluble lymphocyte products in vitro. In contrast, macrophages from all seven resistant strains responded to lymphokine treatment and eliminated 80-90% of intracellular parasites. These results suggest that in vitro macrophage microbicidal activities predict the course of systemic leishmanial disease.

Genetic control of systemic Leishmania major infections: dissociation of intrahepatic amastigote replication from control by the Lsh gene.

Systemic disease induced by Leishmania major was estimated by microscopic examination of liver impression smears and determination of numbers of intrahepatic amastigotes in intravenously or subcutaneously infected inbred, hybrid, and congenic mice. The distribution of susceptible phenotypes among these mice, particularly the susceptibility of a strain congenic for Lshr, strongly suggested that Lsh, a gene which controls intrahepatic replication of Leishmania donovani, does not influence systemic disease by L. major.

Administration of a liposomal FGF-2 peptide vaccine leads to abrogation of FGF-2-mediated angiogenesis and tumor development.

Basic fibroblast growth factor (FGF-2) is an important stimulator of angiogenesis that has been implicated in neoplastic progression. Attempts to neutralize or modulate FGF-2 have met with some success in controlling neovascularity and tumor growth. In the present study, two peptides: one corresponding to the heparin binding domain and the other to the receptor binding domain of FGF-2, exerted dose-dependent inhibition of FGF-2-stimulated human umbilical vein endothelial cell proliferation (IC(50)=70 and 20 microg/ml, respectively). The identification of these functional regions suggested that targeting these domains might be an approach for the modulation of FGF-2 function. To investigate this possibility, we vaccinated mice with either the heparin binding domain peptide or the receptor binding domain peptide of FGF-2 in a liposome/adjuvant format, and analyzed the effect of vaccination on FGF-2-driven angiogenesis, tumor development and immune status. Mice vaccinated with the heparin binding domain peptide generated a specific antibody response to FGF-2, blocked neovascularization in a gelfoam sponge model of angiogenesis, and inhibited experimental metastasis by >90% in two tumor models: the B16BL6 melanoma and the Lewis lung carcinoma. These effects were not observed in mice treated with the receptor binding domain peptide conjugated to liposomes or liposomes lacking conjugated peptide. These data suggest that a heparin binding domain peptide of FGF-2, when presented to a host in a liposomal adjuvant formulation, can ultimately lead to inhibition of angiogenesis and tumor growth.