Antibody-directed coupling of endoglin and MMP-14 is a key mechanism for endoglin shedding and deregulation of TGF-ß signaling.

Endoglin is a transforming growth factor ß (TGF-ß) coreceptor that serves as a prognostic, diagnostic and therapeutic vascular target in human cancer. A number of endoglin ectodomain-targeting antibodies (Abs) can effectively suppress both normal and tumor-associated angiogenesis, but their molecular actions remain poorly characterized. Here we define a key mechanism for TRACON105 (TRC105), a humanized monoclonal Ab in clinical trials for treatment of advanced or metastatic tumors. TRC105, along with several other endoglin Abs tested, enhance endoglin shedding through direct coupling of endoglin and the membrane-type 1 matrix metalloproteinase (MMP)-14 at the cell surface to release the antiangiogenic factor, soluble endoglin (sEng). In addition to this coupling process, endoglin shedding is further amplified by increased MMP-14 expression that requires TRC105 concentration-dependent c-Jun N-terminal kinase (JNK) activation. There were also notable counterbalancing effects on canonical Smad signaling in which TRC105 abrogated both the steady-state and TGF-ß-induced Smad1/5/8 activation while augmenting Smad2/3 activation. Interestingly, TRC105-induced sEng and aberrant Smad signaling resulted in an excessive migratory response through enhanced stress fiber formation and disruption of endothelial cell-cell junctions. Collectively, our study defines endoglin shedding and deregulated TGF-ß signaling during migration as major mechanisms by which TRC105 inhibits angiogenesis.

Diabetic endothelial dysfunction: the role of poly(ADP-ribose) polymerase activation.

Diabetic patients frequently suffer from retinopathy, nephropathy, neuropathy and accelerated atherosclerosis. The loss of endothelial function precedes these vascular alterations. Here we report that activation of poly(ADP-ribose) polymerase (PARP) is an important factor in the pathogenesis of endothelial dysfunction in diabetes. Destruction of islet cells with streptozotocin in mice induced hyperglycemia, intravascular oxidant production, DNA strand breakage, PARP activation and a selective loss of endothelium-dependent vasodilation. Treatment with a novel potent PARP inhibitor, starting after the time of islet destruction, maintained normal vascular responsiveness, despite the persistence of severe hyperglycemia. Endothelial cells incubated in high glucose exhibited production of reactive nitrogen and oxygen species, consequent single-strand DNA breakage, PARP activation and associated metabolic and functional impairment. Basal and high-glucose-induced nuclear factor-kappaB activation were suppressed in the PARP-deficient cells. Our results indicate that PARP may be a novel drug target for the therapy of diabetic endothelial dysfunction.

Regulation of bleomycin-induced DNA breakage and chromatin structure in lung endothelial cells by integrins and poly(ADP-ribose) polymerase.

Activation of endothelial cell integrins inhibits DNA breakage by diverse agents, including the DNA-damaging agent bleomycin. DNA breaks activate nuclear poly(ADP-ribose) polymerase (PARP), which regulates chromatin structure and DNA repair. We determined the role of PARP in suppression of bleomycin genotoxicity by integrins using wild-type and PARP knockout mouse lung endothelial cells (MLEC), and the PARP inhibitor, 3-aminobenzamide (3AB). Activation of beta1 integrins by antibody clustering enhanced the sensitivity of wild-type nuclei to digestion with micrococcal nuclease and deoxyribonuclease I, indicating that chromatin structure was altered. 3AB blocked this effect. Knockout and 3AB-treated wild-type MLEC were hypersensitive to deoxyribonuclease I compared with wild-type cells, demonstrating that PARP regulates chromatin structure. Integrin clustering reduced the hypersensitivity of knockout cells, suggesting additional, PARP-independent mechanisms that inhibit nuclease interaction with chromatin. Bleomycin caused DNA breakage in wild-type and knockout MLEC. Breaks were eliminated after 60 min incubation of wild-type cells in drug-free medium, whereas 3AB or PARP knockout inhibited DNA repair. Integrin clustering protected wild-type cells from DNA breakage, and 3AB and PARP knockout inhibited this protection. Bleomycin caused large increases in PARP activity in wild-type but not knockout MLEC, and integrin clustering inhibited the activation of PARP. The results indicate that the antigenotoxic effects of integrin activation require PARP and that integrins alter chromatin structure by PARP-dependent and -independent mechanisms.

The neutral cysteine protease bleomycin hydrolase is essential for epidermal integrity and bleomycin resistance.

The papain superfamily member bleomycin hydrolase (Blmh) is a neutral cysteine protease with structural similarity to a 20S proteasome. Bleomycin (BLM), a clinically used glycopeptide anticancer agent, is deaminated in vitro by Blmh. We used gene targeting to generate mice that lack Blmh and demonstrated that Blmh is the sole enzyme required for BLM deamination. Although some Blmh null mice were viable and reproduced, only about 65% of the expected number survived the neonatal period, revealing an important role for Blmh in neonatal survival. Mice lacking Blmh exhibited variably penetrant tail dermatitis that resembled rodent ringtail. The histopathology of the tail dermatitis was similar to skin lesions in humans with pellagra, necrolytic migratory erythema, and acrodermatitis enteropathica. Compared with controls, Blmh null mice were more sensitive to acute BLM lethality and developed pulmonary fibrosis more readily following BLM treatment. Thus, we have established that Blmh is an essential protectant against BLM-induced death and has an important role in neonatal survival and in maintaining epidermal integrity.

Nitric oxide inhibits lipopolysaccharide-induced apoptosis in pulmonary artery endothelial cells.

Our group recently reported that cultured sheep pulmonary artery endothelial cells (SPAECs) became resistant to lipopolysaccharide (LPS)-induced apoptosis several days after constitutive synthesis of nitric oxide (NO) after adenoviral (Ad) transfer of inducible NO synthase (iNOS) or exposure to the NO donor S-nitroso-N-acetylpenicillamine (SNAP) (E. Tzeng, Y.-M. Kim, B. R. Pitt, A. Lizonova, I. Kovesdi, and T. R. Billiar. Surgery 122: 255-263, 1997). In the present study, we confirmed this observation by establishing stable transfectants after retroviral gene transfer [replication-deficient retrovirus (DFG)] of human iNOS (DFG-iNOS) SPAECs and then used all three approaches (Ad, DFG, and SNAP) to determine underlying mechanisms of this phenomenon. Continuous endogenous production of NO in itself did not cause apoptosis as assessed by phase-contrast microscopy, nuclear morphology, and internucleosomal DNA fragmentation. Prolonged (72-96 h) synthesis of NO, however, after DFG- or replication-deficient adenovirus (Ad. CMV)-iNOS or SNAP (100 microM, 96 h) inhibited LPS-induced apoptosis. The kinetics of such protection suggested that NO may be inducing other gene products. Ad-mediated transfer of manganese superoxide dismutase (MnSOD) decreased the sensitivity of wild-type SPAECs to LPS-induced apoptosis. MnSOD, however, was not induced in an NG-monomethyl-L-arginine (L-NMMA)-sensitive time-dependent fashion after Ad.CMV-iNOS. Other inducible genes that may be affected by NO and that may protect against potential oxidant-mediated LPS-induced apoptosis including 70-kDa heat shock protein, heme oxygenase-1, metallothionein, and Bcl-2 also were not elevated in an L-NMMA-sensitive, time-dependent fashion. Although the candidate gene product underlying NO-induced protection remains unclear, we did note that prolonged synthesis of NO inhibited LPS-induced activation of an interleukin-1beta-converting enzyme-like cysteine protease (cysteine protease protein-32-like) in a dithiothreitol-sensitive fashion, suggesting that S-nitrosylation of an important downstream target of convergence of apoptotic signals may contribute to the sensitivity of SPAECs to LPS.

Integrin activation protects pulmonary endothelial cells from the genotoxic effects of bleomycin.

Integrin activation promotes the survival of endothelial cells undergoing diverse forms of stress. Here we determined the ability of integrins to inhibit DNA strand breakage by bleomycin (BLM), a DNA-cleaving antitumor antibiotic that causes acute endothelial injury and subsequent pulmonary fibrosis. We found that BLM produced DNA breakage in cultured murine lung endothelial cells (MLEC) within 45 min of treatment as measured by DNA sedimentation and in situ labeling of 3'-OH by nick translation (ISNT). Two hours after the removal of BLM, we found a marked but incomplete reduction in DNA strand breakage as measured by ISNT, indicating that the damage was reversible. DNA sedimentation and ISNT demonstrated that strand breakage due to BLM was inhibited in MLEC cultured on fibronectin, and no evidence of breakage was found 2 h after removal of the drug in ISNT experiments. Gelatin, type IV collagen, laminin, and the integrin ligand peptide Gly-Arg-Gly-Asp-Ser-Pro, but not the inactive Gly-Arg-Ala-Asp-Ser-Pro peptide, also inhibited DNA strand breakage. Activation of integrins, either by coating surfaces with antibodies to alpha 5-, beta 1-, or beta 3-integrin subunits or by receptor clustering with the soluble antibodies, inhibited BLM-induced DNA breakage. Inhibition of BLM-induced DNA strand breakage by soluble beta 1-integrin antibody increased with increasing antibody concentration and duration of receptor clustering before BLM treatment. Thus integrin activation protects pulmonary endothelial cells from the genotoxic effects of BLM.

Enhanced apoptosis in metallothionein null cells.

Metallothioneins (MTs) are major intracellular, zinc-binding proteins with antioxidant properties. Mouse embryonic cells null for MT due to loss of functional MT I and II genes (MT-/-) were more susceptible to apoptotic death after exposure to tert-butyl hydroperoxide or the anti-cancer agents cytosine arabinoside, bleomycin, melphalan, and cis-dichlorodiammineplatinum(II) compared with wild-type mouse embryonic cells (MT+/+). We measured basal levels of the tumor suppressor protein p53 and the death effector protein Bax and found the basal levels of both proteins were higher in MT null cells compared with MT+/+ cells. After treatment with the DNA-damaging agent cis-dichlorodiammineplatinum(II), p53 protein levels were induced in both MT+/+ and MT-/- cells with MT null cells always maintaining the highest p53 levels. The elevated sensitivity to apoptosis was not restricted to embryonic cells. Primary pulmonary fibroblasts were isolated from distinct litters of MT null, heterozygous, and wild-type mice, and all had undetectable basal MT levels. Zinc exposure increased MT levels in the wild-type and heterozygous fibroblasts but not in the MT null fibroblasts. Consistent with the induced MT levels, we found MT+/+ and MT+/- embryonic cells were less sensitive to cis-dichlorodiammineplatinum(II)-induced apoptosis compared with MT-/- cells. Our results implicate MT as a stress-responsive factor that can regulate apoptotic engagement.

Integrin activation suppresses etoposide-induced DNA strand breakage in cultured murine tumor-derived endothelial cells.

Tumor endothelium is critical for solid tumor growth and is a potential site for anticancer drug action. Within 2 h, etoposide caused marked DNA strand breakage in xenograft tumor-derived endothelial cells (TDECs). Etoposide-induced DNA breakage was inhibited by culturing TDECs on gelatin, type IV collagen, laminin, fibronectin, and the integrin ligand hexapeptide, GRGDSP, but not the inactive peptide, GRADSP. It was also inhibited when TDECs were on surfaces coated with antibodies to alpha 5, beta 1, or beta 3 integrin subunits and by clustering integrins with soluble antibodies. After 8 h with etoposide, TDECs detached from the monolayer, and 50-kb DNA fragments were seen. Fibronectin inhibited both processes. Thus, integrins are survival factors for TDEC that inhibit the genotoxicity of etoposide and may influence the sensitivity of tumors to drugs.

Nucleophilic distribution of metallothionein in human tumor cells.

Metallothionein (MT), a major zinc-binding intracellular protein thiol, has been associated with cytoprotection from heavy metals, antineoplastic drugs, mutagens, and cellular oxidants. Despite its small mass (7 kDa), nuclear partitioning of MT has been observed in both normal and malignant tissues. The factors controlling MT sequestration are unknown. Thus, we examined the regulation of MT subcellular distribution in human cancer cell lines that exhibit prominent nuclear MT. The nuclear disposition of MT was unaltered during cell cycle passage in synchronized cells. MT redistributed to the cytoplasm when cells were exposed to reduced temperature. Cytoplasmic redistribution was also seen in DU-145 and HPC36M prostatic cancer cells after ATP depletion, but not in PC3-MA2 and SCC25/CP cells. Pretreatment with 10 microM CdCl2 did not significantly alter MT distribution but did render all cells sensitive to cytoplasmic redistribution after either reduced temperature or ATP depletion. Thus, nuclear retention of MT is energy requiring and this ability of MT to accumulate in subcellular compartments against its concentration gradient may be important in the capacity of MT to supply Zn or other metals to target sites within the cell.

Integrins inhibit LPS-induced DNA strand breakage in cultured lung endothelial cells.

Collagen inhibits acute DNA strand breakage and apoptosis in sheep pulmonary artery endothelial cells (SPAEC) treated with lipopolysaccharide (LPS). Here we tested the ability of major basement membrane components, type IV collagen, laminin and fibronectin, and integrin ligands and anti-integrin antibodies to inhibit DNA breakage caused by LPS in SPAEC and BALB/c murine lung endothelial cells (MLEC). In situ labeling of DNA strand breaks with terminal deoxynucleotidyl transferase revealed similar DNA breakage in attached SPAEC and MLEC within 2 h after incubation with 1 microgram LPS/ml. Acute DNA strand breakage was reduced in cells plated on gelatin, type IV collagen, laminin, cellular fibronectin, or plasma fibronectin. DNA breakage was also suppressed by plating cells on surfaces coated with the integrin ligand hexapeptide, GRGDSP (40 micrograms/cm2), but not with GRADSP. LPS-induced DNA strand breakage was inhibited in MLEC plated on surfaces coated with antibodies to murine alpha 5-, beta 1, or beta 3-integrin subunits. Addition of anti-integrin antibodies, but not GRGDSP, to the medium above cell monolayers inhibited strand breakage. Despite similar acute DNA breakage, MLEC exhibited less detachment and apoptosis than SPAEC, consistent with a difference in the sensing or processing systems for apoptosis in these two cell types. These results demonstrate that extracellular matrices and integrin activation can inhibit the genotoxicity of LPS.

Genesis of discrete higher order DNA fragments in apoptotic human prostatic carcinoma cells.

Higher order DNA fragmentation may be an essential signal in apoptosis. We found that etoposide (VP-16) induced apoptosis in human DU-145 prostatic carcinoma cells in a time- and concentration-dependent manner. Chromatin condensation was morphologically evident only when cells detached from the monolayer; untreated or VP-16-treated attached cells retained a normal morphology. We describe a radiolabeled alu-I sequence-based quantitative field inversion gel electrophoresis (QFIGE) method that permitted observation and quantification of discrete high molecular weight DNA fragments in detached (apoptotic) and attached (preapoptotic) DU-145 cells. The DNA fragments generated during the apoptotic death of these cells were > or = 1 (mega-base pairs) mbp, 450-600 (kilo-base pairs) kbp, and 30-50 kbp; we observed that these DNA fragments increased 9 +/- 2-, 8 +/- 2-, and 25 +/- 11-fold versus control, respectively, with a 24-hr exposure to 30 microM VP-16 in attached cell populations. In detached VP-16-treated cells, there was accrual of 30-50-kbp DNA fragments with a concomitant loss of the > or = 1-mbp and 450-600-kbp fragments; internucleosomal DNA cleavage was never observed. This pattern of high molecular weight DNA fragmentation was inhibited by cycloheximide treatment and was common to other apoptotic agents, including melphalan and bleomycin. These findings suggest that the > or = 1-mbp and 450-600-kbp DNA fragments are products of endonuclease activation and are not topoisomerase II/DNA interactions. Finally, the generation of the 30-50-kbp DNA fragments may mediate chromatin condensation, which characterizes apoptosis.

Collagen is a survival factor against LPS-induced apoptosis in cultured sheep pulmonary artery endothelial cells.

Lipopolysaccharide (LPS) causes direct pulmonary endothelial injury that can precipitate cell death. We investigated the ability of LPS to produce apoptosis in sheep pulmonary artery endothelial cells (SPAEC) grown in monolayer on plastic or collagen. When SPAEC were grown on plastic, LPS (100 ng/ml) caused internucleosomal DNA fragmentation (IDF) to 180- to 200-base pair ladders after 4 h. Higher-order chromatin damage, producing 50-kilobase DNA fragments, occurred within 2 h. Significant DNA strand breaks were seen in attached cells within 1 h incubation with > or = 1 ng LPS/ml, using in situ labeling by break extension (ISBE). DNA strand breakage in attached cells peaked after 2 h and remained elevated after 4 h. Detachment of SPAEC from the monolayer did not begin until 4 h. SPAEC cultured on collagen were protected from LPS-induced apoptosis; DNA damage measured by IDF, high-molecular-weight DNA fragmentation, and ISBE were suppressed. The protective effect of collagen was not due to inactivation of LPS. Thus LPS-induced apoptosis occurs in SPAEC after genotoxic damage and this process is suppressed by the extracellular matrix.

Acute pneumocyte injury, poly(ADP-ribose) polymerase activity, and pyridine nucleotide levels after in vitro exposure of murine lung slices to cyclophosphamide.

Cyclophosphamide (CYC) is a metabolically activated, DNA-alkylating, antitumor agent that causes pulmonary fibrosis. BALB/cN (B) mice are sensitive and C57Bl/6N (C) mice are resistant to CYC-induced fibrosis. Pulmonary bioactivation may contribute to strain sensitivity. Therefore, we tested the intrinsic susceptibility of murine lung slices to cell injury by direct exposure to CYC for 2-8 hr. Injury was measured by release of lactate dehydrogenase (LDH). DNA damage activates the nuclear enzyme poly(ADP-ribose) polymerase (PAP, EC 2.4.2.30), causing depletion of its substrate, NAD. NAD can also be decreased by phosphorylation to NADP, as seen with oxidative stress. Depletion of NAD can lead to loss of ATP. Thus, we measured LDH release, PAP activation, NAD, NADP and ATP in slices incubated with or without the PAP-inhibitor, 3-aminobenzamide (3-AB). CYC (0.1 to 1.0 mg/mL for 4-8 hr) caused LDH release in slices from both murine strains, but LDH release was significantly greater in B lung slices than in C slices. After an 8-hr incubation 63.9 +/- 3.7% (mean +/- SEM) of total LDH was released from B lung slices with 1.0 mg CYC/mL, whereas only 45.8 +/- 2.6% was released from C lung slices (P < 0.05). 3-AB reduced LDH release to 44.7 +/- 2.4% in B slices and 28.1 +/- 2.0% in C slices (P < 0.05 vs CYC only). PAP activity in nuclei isolated from CYC-treated B lung slices was increased 2- to 4-fold after 2 hr of incubation with 0.5 and 1.0 mg CYC/mL. PAP activation was delayed and reduced with incubation in 3-AB. PAP was activated 2-fold in nuclei from C slices treated with 0.5 mg CYC/mL for 2 hr. NAD was decreased at 2 and 4 hr in B slices treated with 0.5 and 1.0 mg CYC/mL, and at 4 hr with 0.1 mg CYC/mL. NAD depletion occurred only at 4 hr in the resistant C slices treated with 1.0 mg CYC/mL. CYC increased NADP by a similar extent in B and C lung slices. In B slices, NAD losses were approximately 4 times the increases in NADP. CYC did not decrease ATP in B slices and ATP dropped 25% only after 4 hr in the resistant C slices. We conclude that CYC is directly toxic to lung tissue and observe that strain sensitivity in vitro mirrors the sensitivity to fibrosis in vivo. PAP activation and oxidative stress may contribute to this toxicity.

NAD depletion after in vitro exposure of murine lung slices to bleomycin.

Bleomycin (BLM), a DNA-cleaving, antitumor antibiotic, causes pulmonary fibrosis. It also causes cell injury and activates the nuclear enzyme poly(ADP-ribose) polymerase (PAP; EC 2.4.2.30) in lung slices exposed to the drug in vitro. 3-Aminobenzamide (3-AB), a PAP inhibitor, prevents enzyme activation and cell injury. We have examined the potential role of ATP and NAD depletion in injury of BLM-sensitive C57B1/6N and -resistant BALB/cN murine lung slices treated with BLM or deprived of glucose, the major metabolic substrate of lung. Lung slices either were treated for 45 min with injurious concentrations of BLM (10-500 micrograms/mL) or were incubated without glucose, in the presence or absence of 2.5 mM 3-AB. Only the highest concentration of BLM, 500 micrograms/mL, caused any ATP depletion, and this 35% decrease was transient, occurring at 220 min in C57B1/6N slices. In contrast, glucose deprivation caused 50-70% ATP depletion in slices from both strains. BLM alone at 100 and 500 micrograms/mL caused a sustained 30-70% NAD depletion from 75 min through 400 min in C57B1/6N mouse lung slices. In the resistant BALB/cN lung slices, NAD depletion by BLM was only seen at 400 min. 3-AB almost completely antagonized NAD depletion in slices from both strains. In contrast to BLM, glucose deprivation did not decrease NAD levels unless 3-AB was present in C57B1/6N slices. Thus, ATP depletion may play a role in the injurious effects of glucose deprivation, but does not appear to be a major factor in pneumocyte injury caused by BLM. NAD depletion or other effects of PAP activation appear to account for the strain-selective, injurious effect of BLM on lung tissue.

Murine strain differences in acute lung injury and activation of poly(ADP-ribose) polymerase by in vitro exposure of lung slices to bleomycin.

The DNA-cleaving, antitumor antibiotic bleomycin (BLM) causes pulmonary fibrosis, but the essential early events initiating the fibrotic state have not been well characterized. Thus, we have directly examined BLM-mediated pulmonary cell injury by monitoring lactate dehydrogenase (LDH) release and nuclear poly(ADP-ribose) polymerase (PAP) activity, which is stimulated by DNA breakage, using lung slices isolated from BLM-sensitive (C57B1/6) and BLM-resistant (BALB/c) mice. Lung slices were incubated continuously with or without the PAP inhibitor, 3-aminobenzamide (3-AB), and exposed to BLM for 45 min. LDH release from C57B1/6 lung slices increased 2-fold by 8.5 h after treatment with BLM. In contrast, BLM failed to enhance cumulative LDH release by BALB/c mouse lung slices. Co-incubation of C57B1/6 lung slices with 3-AB prevented BLM-induced LDH release. Nuclear PAP was activated 3- to 4-fold 1.25 h after exposure of C57B1/6 lung slices to BLM but returned to control levels by 3.75 h. Nuclear PAP was only marginally affected at these times in BALB/c lung slices. Co-incubation of C57B1/6 slices with 3-AB prevented the early increases in PAP activity. These results demonstrate that murine strain sensitivity to acute cell injury and early PAP activation by BLM in lung slices parallels the in vivo sensitivity of lungs. In addition, 3-AB suppresses PAP activation and acute cell injury in lung slices. Differential activation of PAP appears to govern murine strain variation in response to BLM and is consistent with the hypothesis that activation of PAP participates in acute pneumocyte injury, initiating the process of BLM-induced fibrosis.

Bleomycin and cyclophosphamide increase pulmonary type IV procollagen mRNA in mice.

Constant 7-day subcutaneous infusion of bleomycin (100 mg/kg) induces pulmonary fibrosis in C57Bl/6N mice, whereas BALB/cN mice are relatively resistant. In contrast, cyclophosphamide (200 mg/kg, ip) induces fibrosis in BALB/cN mice, whereas C57Bl/6N mice are resistant. The effect of these drugs on the pulmonary levels of mRNA encoding the major basement membrane components, laminin and type IV collagen, relative to poly (A+)RNA was determined in both C57Bl/6N and BALB/cN mice. In the sensitive C57Bl/6N mice, bleomycin increased alpha 1IV and alpha 2IV procollagen mRNA/poly (A+)RNA twofold in the absence of increases in laminin A, B1, and B2 mRNA/poly (A+)RNA. In the relatively resistant BALB/cN mice, bleomycin did not alter alpha 1IV procollagen mRNA/poly (A+)RNA and only transiently increased laminin A, B1, B2, and alpha 2IV procollagen mRNA/poly (A+)RNA. Similarly, cyclophosphamide increased alpha 1IV and alpha 2IV procollagen mRNA/poly (A+)RNA twofold in the sensitive BALB/cN mice and not in C57Bl/6N mice. Laminin mRNAs/poly (A+)RNA were not increased by cyclophosphamide in either strain. Thus, in these models, pulmonary fibrosis is preceded by a coordinate increase in steady-state levels of mRNA encoding basement membrane procollagen but is not associated with an increase in laminin gene expression.

Bleomycin: a pharmacologic tool in the study of the pathogenesis of interstitial pulmonary fibrosis.

Bleomycin is a unique DNA-interactive antitumor agent that has become a popular tool in studies of the pathogenesis of interstitial pulmonary fibrosis. The biochemical and morphological changes seen in the lungs of many species after bleomycin simulate those seen in humans. The availability of these animal models of interstitial pulmonary fibrosis also provides the opportunity to investigate novel pharmacological approaches to preventing this disease.

Acute pulmonary toxicity of bleomycin: DNA scission and matrix protein mRNA levels in bleomycin-sensitive and -resistant strains of mice.

The severity of bleomycin (BLM)-induced pulmonary fibrosis in mice varies markedly among several different murine strains. We have examined the DNA from lungs of sensitive (i.e., C57BL/6N) and resistant (i.e., BALB/c) strains of mice using a nucleoid sedimentation technique to detect early in vivo changes in the integrity of DNA after intravenous BLM. Mice received intravenous injections of BLM (80 mg/kg) or vehicle; lung nucleoids were prepared 15 min to 6 hr later. BLM produced striking decreases in nucleoid sedimentation distance versus paired controls in both strains within 15 min after injection, indicating extensive DNA scission. Repair of DNA strand breaks was complete in the resistant (BALB/c) mice by 5 hr; in contrast, only partial repair occurred in the sensitive (C57BL/6N) strain during that time. We then examined lungs for subsequent changes in steady state poly-(A)+ RNA levels and mRNA levels for lung matrix proteins (type I procollagen, type III procollagen, and fibronectin). Steady state levels of poly-(A)+ RNA were depressed to 50% of control 1 through 6 days after BLM injection in the lungs of sensitive mice. Resistant mice had pulmonary poly-(A)+ RNA levels similar to those of C57BL/6N mice, except for a 2-fold elevation 1 day after BLM injection. BLM injection affected the steady state levels of mRNA encoding lung matrix proteins differently than total poly-(A)+ RNA. Fibronectin mRNA/poly(A)+ RNA was elevated 2-fold 1 day after BLM treatment only in the sensitive strain and remained elevated at 3 and 6 days. In contrast, alpha 2I procollagen mRNA increased in both murine strains and alpha 1III procollagen mRNA decreased in both strains. Thus, a 7-fold or greater increase in the type I: type III procollagen mRNA ratio was seen in both strains 3 to 6 days after BLM injection. These data demonstrate that BLM treatment rapidly produces extensive pulmonary DNA damage in vivo, that persistence of DNA damage rather than the initial level of strand scission is associated with sensitivity to BLM lung disease in these mice, and that changes in the levels of mRNA encoding pulmonary matrix proteins occur in vivo within 1 to 3 days after intravenous BLM treatment.

Early increases in pulmonary mRNA encoding procollagens and transforming growth factor-beta in mice sensitive to cyclophosphamide-induced pulmonary fibrosis.

Pulmonary fibrosis was induced 7 weeks after a single i.p. injection of cyclophosphamide (200 mg/kg b.wt.) in BALB/c mice; C57Bl/6 mice were unaffected. There was a corresponding strain variation in the effects of cyclophosphamide on levels of pulmonary mRNA encoding alpha 2I and alpha 1III procollagen, and transforming growth factor-beta. In BALB/c mice, the ratios of alpha 2I and alpha 1III procollagen mRNA to polyadenylated RNA were increased 1 week after cyclophosphamide injection. No increases in levels of either procollagen mRNA occurred in C57Bl/6 mice. The ratio of fibronectin mRNA to polyadenylated RNA was elevated to a similar extent in both murine strains during the 1st week after cyclophosphamide treatment. The pulmonary content of transforming growth factor-beta mRNA and its ratio to polyadenylated RNA increased 2-fold at 1 and 2 weeks in BALB/c but not C57Bl/6 mice. Thus, collagen accumulation in cyclophosphamide-sensitive mice is preceded by increased pulmonary alpha 2I and alpha 1III procollagen mRNA. The early strain selective elevation of transforming growth factor-beta mRNA in response to cyclophosphamide suggests a role, in vivo, for transforming growth factor-beta in drug-induced pulmonary fibrosis.