Infection of Cultured Human Endothelial Cells by Legionella pneumophila Authors: Lucius Chiaraviglio, Daniel A. Brown, James E. Kirby Published: 2008, PLOS ONE DOI: 10.1371/journal.pone.0002012 ABSTRACT Legionella pneumophila is a gram-negative pathogen that causes a severe pneumonia known as Legionnaires' disease. Here, we demonstrate for the first time that L. pneumophila infects and grows within cultured human endothelial cells. Endothelial infection may contribute to lung damage observed during Legionnaires' disease and to systemic spread of this organism. INTRODUCTION Legionella pneumophila causes an estimated 2–15% of community acquired pneumonia requiring hospitalization [1] . L. pneumophila is a gram-negative pathogen whose primary ecological niche appears to be as a parasite of protozoa. However, humans and other animals may become secondarily infected after inhaling or aspirating organisms. Fascinatingly, the cell biology of infection of protozoa and humans cells appears remarkably similar. On entering both macrophages [2] and protozoa [3] , [4] , L. pneumophila inhibits phagolysosome fusion [5] and multiplies in a compartment having properties of endoplasmic reticulum [6] . Escape from phagolysosome fusion, intracellular multiplication, and lysis of the host cell relies on a type IV secretion system encoded by the dot/icm gene complex [7] – [9] . Resident alveolar macrophages appear to be a Trojan horse for human infection [10] . L. pneumophila infects these cells and continues to grow within macrophages recruited during the subsequent inflammatory response. Infection may then spread to alveolar epithelial cells that line the airways, supported by observations of growth in primary [11] and transformed type I [12] and II pneumocytes [13] , and growth of the related L. dumoffii in alveolar epithelial cells in vivo [14] . Bacterial dissemination does not stop at the alveolar wall. In both human autopsy studies [15] and animal models [16] , L. pneumophila spreads to multiple organs. This suggests that the organism may infect and breach the remaining anatomic barrier in the lung — pulmonary blood vessels — as a potential initiating event in systemic spread of the organism via the blood stream. Here, we begin to address this hypothesis through in vitro study of the interaction of Legionella pneumophila with primary human endothelial cells. RESULTS We first tested whether Legionella pneumophila strain Lp01 [17] invades into primary human umbilical vein endothelial cells (HUVEC). Using a gentamicin protection assay, we assessed whether bacteria were protected from gentamicin [18] , a polar antibiotic that penetrates slowly into eukaryotic cells and does not kill internalized bacteria during the time frame of the assay. Using this procedure, we found that L. pneumophila , strain Lp01, reproducibly invades into endothelial cells (see Fig. 1 ) with approximately 0.05% of bacteria internalized during a 2 hour incubation. The ability of L. pneumophila to invade into endothelial cells was confirmed by fluorescent microscopy where internalized bacteria were distinguished by differential staining as described in detail below. L. pneumophila strains Lp01 (wt) and an isogenic dotI mutant were seeded onto HUVEC at a ratio of approximately 300∶1. After a two hour incubation in the presence or absence of 100 nM wortmannin, gentamicin treatment was used to kill extracellular bacteria, and then intracellular bacteria were enumerated. Intracellular bacteria are expressed as percentage of the original titered inoculum to determine the invasion efficiency. Results shown are the mean and standard deviation of sextuplicate assays and are representative of two independent experiments. Of note, the invasion efficiency was roughly 100 fold lower than what was previously observed for primary macrophages [19] , an expected result since endothelial cells are not professional phagocytes. We compared invasion efficiency to the L. pneumophila , isogenic dotI mutant, HL051C [20] , which was previously shown to be defective in inhibition of phagolysosome fusion and intracellular replication in macrophages. Although a trend towards slightly lower invasion efficiency by HL051C was noted, this did not reach statistical significance ( Fig. 1 ). Previously, invasion of a dot/icm mutant, but not wild type L. pneumophila was shown to be blocked by the PI-3 kinase inhibitor, wortmannin [21] , implying that invasion of wild type and mutant bacteria occur through distinct pathways. We found the same to be true in endothelial cells ( Fig. 1 ). Here, treatment with 100 nM wortmannin led to a 4.5 fold reduction in invasion of HL051C ( P  = 0.005) vs. an insignificant reduction of 1.5 fold for Lp01. Therefore, since wortmannin did not substantially block invasion of wild type bacteria, invasion of wild type and the dotI mutant appears to occur by distinct pathways as described previously for macrophages [21] , [22] . REPLICATION OF BACTERIA WITHIN ENDOTHELIAL CELLS We next tested whether L. pneumophila replicates inside of endothelial cells. As shown in Fig. 2 , we found that wild type Lp01 replicates to high levels in HUVEC with a two to three log growth increase, leading to lysis of host cells and release of bacteria (data not shown). In contrast, the dotI mutant, HL051C, failed to grow and was gradually eliminated. These results are consistent with previous observations in macrophages and suggest that growth is dependent on the dot/icm -dependent, type IV secretion system [17] , [23] . These observations were confirmed by fluorescent microscopy where large numbers of Lp01 were found intracellularly at later time points vs. the infrequent appearance of HL051C. The latter often appeared to have degraded nucleoids by DAPI staining. Growth curves of L. pneumophila Lp01 (wt) and an isogenic dotI mutant after infection of HUVEC. At indicated time points after infection, the numbers of intracellular and extracellular bacterial colony forming units (cfu) were determined and added together. As bacteria do not multiply extracellularly, the total number of bacteria reflects intracellular growth. Data points are the mean and standard deviation of three separate assays. Similar growth curves were observed for other endothelial cell types tested including primary human lung microvascular and umbilical artery endothelial cells (data not shown). The ability to grow in the diverse types of endothelial cells surveyed is especially interesting as they are known to have diverse physiological properties [24] . Bacteria also grew within primary pulmonary artery endothelial cells of bovine origin (a species that also develops Legionnaires' disease [25] ), suggesting that the ability to infect endothelial cells crosses species boundaries. ALTERED PHAGOSOME PROPERTIES We next examined whether L. pneumophila inhibits fusion of bacterial phagosomes with lysosomes, as observed previously during infection of macrophages [5] . Using an immunofluorescence staining technique, we assessed the colocalization of internalized bacteria with a lysosomal marker, LAMP-1. At 1.5 hours after infection, only 6% of Lp01 colocalized with LAMP-1 vs. 57% of the dotI mutant ( P <0.001, Fig. 3 ). Therefore, only the replication-competent Lp01, but not the replication-defective dotI mutant, efficiently escaped from lysosomal fusion. Representative images showing lysosomal targeting of bacteria 1.5 hours after infection of HUVEC at a bacterial to endothelial cell ratio of approximately 300∶1. The dotI mutant strain, but not Lp01 (wt), show extensive colocalization with LAMP-1, a lysosomal marker (bacteria stained blue with DAPI, LAMP-1 stained green). An example of a colocalizing bacteria is indicated with an arrow. Extracellular bacteria (not present in these images) were distinguished by immuno-labeling with a red fluorophore prior to permeabilization of endothelial cells. We next examined the morphogenesis of L. pneumophila containing vacuoles by transmission electron microscopy. Within 6 hours of infection by Lp01, most bacterial phagosomes were surrounded by ribosome studded endoplasmic reticulum (RER) ( Fig. 4A,B ), as observed previously in infected macrophages [26] and amoeba [27] . At 24 hours after infection, organisms had multiplied in vacuoles to largely fill the endothelial cell ( Fig. 4C ). In contrast, the dotI mutant was always contained in smooth vacuoles, usually as single bacteria, and without associated RER or ribosomes at all time points ( Fig. 4D, E ). Furthermore, mutant bacteria often appeared degraded and were occasionally found in vacuoles containing membranous debris ( Fig. 4F ). Both mutant and wild type organisms were found in membrane-bound phagosomes, never freely in the cytoplasm, although disruption of the vacuolar membrane was noted late in the infection cycle of wild type organisms. Taken together, intracellular parasitism of endothelial cells appears similar to observations in macrophages. (A, B) The Lp01strain is surrounded by ribosome studded endoplasmic reticulum (RER) at 6 hours after infection. (C) At 24 hours, many phagosomes are filled with large numbers of bacteria. In this image, RER association is still evident (left side of vacuole), however, at this late stage in infection, vacuolar membrane appears partially disrupted. In contrast, dotI mutants do not associate with RER and are surrounded by a simple, smooth membrane, at times also containing membranous debris, e.g., (D) 6 hours and (E) 24 hours. Intact bacteria were far less numerous, while electron dense structures suggestive of partially degraded bacteria were often present, (F) 6 hours. Magnifications (in figure): A, ×54,000; B, ×54,000; C ×19,000; D ×82,000; E ×119,000; F ×41,000. DISCUSSION Here, we for the first time identify the ability of Legionella pneumophila to invade and replicate inside diverse types of primary human endothelial cells, including endothelial cells obtained from the lung. We believe that this property may have special relevance for the pathogenesis of Legionnaires' disease. Previous studies have focused on the organism's ability to grow within macrophages and alveolar epithelial cells. However, the capillary network, formed by endothelial cells, also makes up a critical and potentially vulnerable part of the lung. The biology of endothelial infection is consistent with what has been described for infection of macrophages, alveolar epithelial cells, and protozoa. The organism inhibits phagolysosome fusion and grows intracellularly in a process dependent on the dot/icm type IV secretion machinery. However, contact-dependent cytotoxicity observed during infection of macrophages and red blood cells [19] , [28] does not occur (i.e., as determined by LDH release assay [19] after pelleting bacteria onto an endothelial monolayer for 5 minutes at 800 g , data not shown). This may reflect a greater resistance to osmotic lysis, differences in the interaction with the bacterial type IV secretion machinery, or decreased invasion efficiency. Nevertheless, endothelial cell death still occurs after intracellular multiplication of bacteria. It is possible that endothelial cell infection and death in vivo leads to damage of pulmonary capillaries and allows bacteria access to the systemic blood circulation, although lymphatic drainage of organisms might also contribute to systemic spread. Endothelial damage may also in part account for pathologies such as hemorrhage that have been observed during Legionnaires' disease [29] , [30] or adverse sequelae such as pulmonary fibrosis [31] . Here, the lung forms scar tissue rather than re-establishing surfaces capable of air exchange, leading to permanent disability after bacteriological cure [31] . Interestingly, patients with prosthetic heart valves appear to be at particular risk of developing endocarditis from Legionella [32] – [34] . Possibly, this relates to a susceptibility of endothelial cells overlying prosthetic material to infection with this organism. Clearly, the extent and consequences of endothelial infection by L. pneumophila need to be examined using in vivo models. However, the vascular endothelium should now be considered a potential participant in the pathogenesis of Legionnaires' disease. MATERIALS AND METHODS Legionella pneumophila strains were described previously [17] . Endothelial cells were purchased from Cambrex and passaged in EGM2-MV medium from the same manufacturer. To perform infections, bacteria were grow in AYE medium to early stationary phase, resuspended in EGM-2MV, and used to infect endothelial cells grown on collagen I (Vitrogen, Cohesion, Inc.) treated twenty-four well dishes. ASSESSMENT OF BACTERIAL INVASION After infection at a bacterial to endothelial cell ratio of approximately 300∶1, extracellular bacteria were killed by a 45 minute treatment with 100 µg/ml of gentamicin, and surviving intracellular bacteria enumerated after eukaryotic cell lysis with 0.1% saponin. Statistical comparisons were performed using JMP IN Vs. 5.1.2 (SAS Institute, Cary, NC) by the Wilcoxon Ranked Sum Test for continuous and Fisher's Exact Test for categorical variables with P <0.05 considered significant. ASSESSMENT OF BACTERIAL REPLICATION At multiple time points after infection, both extracellular and intracellular bacteria were enumerated by sampling extracellular medium and then intracellular compartments after saponin lysis. For bacterial replication experiments only, the infectious inoculum was harvested from αBCYE plates rather than from early stationary phase liquid cultures used for all other assays. ENDOCYTIC TRAFFICKING ASSAY To assess fusion of L. pneumophila containing phagosomes with lysosomes, endothelial cells were plated on collagen I-treated, German glass coverslips (Bellco Glass) and infected at a ratio of approximately 300 bacteria per endothelial cell. Extracellular bacteria were stained prior to permeabilization with a rabbit anti- Legionella primary antibody (gift of Ralph Isberg, Tufts Medical School) followed by a Alexa-594-labeled goat anti-rabbit secondary antibody. After permeabilization with 0.2% Triton®-X 100 for 15 minutes, lysosomal compartments were stained with anti-LAMP-1 mouse monoclonal antibody (Santa Cruz Biotechnology, clone H5G11) [35] followed by Alexa-488-labeled goat anti-mouse secondary antibody (Molecular Probes). Bacterial nucleoids were then stained with 4′,6-diamidino-2-phenylindole (DAPI). Intracellular bacteria were identified by the absence of extracellular antibody staining and positive staining with DAPI. Colocalization of intracellular bacteria (DAPI + , Alexa-594 - ) and LAMP-1 (Alexa-488 + ) was assessed and quantified using a Nikon Eclipse E300 epifluorescence microscope and images capture using a Retiga 2000RV digital camera (QImaging, Surrey, British Columbia) and IPLab software 3.9.4r5 (BD Biosciences, Rockville, MD) ELECTRON MICROSCOPY Bacteria at a multiplicity of infection of approximately 1000∶1 were centrifuged onto HUVEC monolayers at 580 g for five minutes. After incubation for the indicated time points, monolayers were washed with PBS, fixed for twenty minutes in Trump's fixative containing 4% formaldehyde and 1% glutaraldehyde, washed in 0.1 M cacodylate buffer, and post fixed in 1% OsO 4 . Monolayers were then washed in maleate buffer, incubated with uranyl acetate, washed again in maleate buffer, dehydrated in ascending alcohol washes, embedded as monolayers in Eponate using an inverted BEEM® capsule (Beem, Inc., Westchester, PA), and sectioned en face. Sections were stained with lead citrate and viewed with the digital image acquisition system on the Jeol (Tokyo, Japan) MEM-1011, transmission electron microscope. REFERENCES [1] Stout JE Yu VL 1997 Legionellosis. N Engl J Med 337 682 687 [2] Horwitz MA Silverstein SC 1980 Legionnaires' disease bacterium (Legionella pneumophila) multiples intracellularly in human monocytes. J Clin Invest 66 441 450 [3] King CH Fields BS Shotts EB Jr. White EH 1991 Effects of cytochalasin D and methylamine on intracellular growth of Legionella pneumophila in amoebae and human monocyte-like cells. Infect Immun 59 758 763 [4] Newsome AL Baker RL Miller RD Arnold RR 1985 Interactions between Naegleria fowleri and Legionella pneumophila. Infect Immun 50 449 452 [5] Horwitz MA 1983 The Legionnaires' disease bacterium (Legionella pneumophila) inhibits phagosome-lysosome fusion in human monocytes. J Exp Med 158 2108 2126 [6] Roy CR Tilney LG 2002 The road less traveled: transport of Legionella to the endoplasmic reticulum. J Cell Biol 158 415 419 [7] Segal G Purcell M Shuman HA 1998 Host cell killing and bacterial conjugation require overlapping sets of genes within a 22-kb region of the Legionella pneumophila genome. Proc Natl Acad Sci U S A 95 1669 1674 [8] Segal G Shuman HA 1997 Characterization of a new region required for macrophage killing by Legionella pneumophila. Infect Immun 65 5057 5066 [9] Vogel JP Andrews HL Wong SK Isberg RR 1998 Conjugative transfer by the virulence system of Legionella pneumophila. Science 279 873 876 [10] Nash TW Libby DM Horwitz MA 1984 Interaction between the legionnaires' disease bacterium (Legionella pneumophila) and human alveolar macrophages. Influence of antibody, lymphokines, and hydrocortisone. J Clin Invest 74 771 782 [11] Mody CH Paine R 3rd Shahrabadi MS Simon RH Pearlman E 1993 Legionella pneumophila replicates within rat alveolar epithelial cells. J Infect Dis 167 1138 1145 [12] Gao LY Stone BJ Brieland JK Abu Kwaik Y 1998 Different fates of Legionella pneumophila pmi and mil mutants within macrophages and alveolar epithelial cells. Microb Pathog 25 291 306 [13] Cianciotto NP Stamos JK Kamp DW 1995 Infectivity of Legionella pneumophila mip mutant for alveolar epithelial cells. Curr Microbiol 30 247 250 [14] Maruta K Miyamoto H Hamada T Ogawa M Taniguchi H 1998 Entry and intracellular growth of Legionella dumoffii in alveolar epithelial cells. Am J Respir Crit Care Med 157 1967 1974 [15] Evans CP Winn WC Jr. 1981 Extrathoracic localization of Legionella pneumophila in Legionnaires' pneumonia. Am J Clin Pathol 76 813 815 [16] Brieland J Freeman P Kunkel R Chrisp C Hurley M 1994 Replicative Legionella pneumophila lung infection in intratracheally inoculated A/J mice. A murine model of human Legionnaires' disease. Am J Pathol 145 1537 1546 [17] Berger KH Isberg RR 1993 Two distinct defects in intracellular growth complemented by a single genetic locus in Legionella pneumophila. Mol Microbiol 7 7 19 [18] Isberg RR 1991 Discrimination between intracellular uptake and surface adhesion of bacterial pathogens. Science 252 934 938 [19] Kirby JE Vogel JP Andrews HL Isberg RR 1998 Evidence for pore-forming ability by Legionella pneumophila. Mol Microbiol 27 323 336 [20] Andrews HL Vogel JP Isberg RR 1998 Identification of linked Legionella pneumophila genes essential for intracellular growth and evasion of the endocytic pathway. Infect Immun 66 950 958 [21] Khelef N Shuman HA Maxfield FR 2001 Phagocytosis of wild-type Legionella pneumophila occurs through a wortmannin-insensitive pathway. Infect Immun 69 5157 5161 [22] Watarai M Derre I Kirby J Growney JD Dietrich WF 2001 Legionella pneumophila is internalized by a macropinocytotic uptake pathway controlled by the Dot/Icm system and the mouse Lgn1 locus. J Exp Med 194 1081 1096 [23] Marra A Blander SJ Horwitz MA Shuman HA 1992 Identification of a Legionella pneumophila locus required for intracellular multiplication in human macrophages. Proc Natl Acad Sci U S A 89 9607 9611 [24] Chi JT Chang HY Haraldsen G Jahnsen FL Troyanskaya OG 2003 Endothelial cell diversity revealed by global expression profiling. Proc Natl Acad Sci U S A 100 10623 10628 [25] Fabbi M Pastoris MC Scanziani E Magnino S Di Matteo L 1998 Epidemiological and environmental investigations of Legionella pneumophila infection in cattle and case report of fatal pneumonia in a calf. J Clin Microbiol 36 1942 1947 [26] Horwitz MA 1983 Formation of a novel phagosome by the Legionnaires' disease bacterium (Legionella pneumophila) in human monocytes. J Exp Med 158 1319 1331 [27] Abu Kwaik Y 1996 The phagosome containing Legionella pneumophila within the protozoan Hartmannella vermiformis is surrounded by the rough endoplasmic reticulum. Appl Environ Microbiol 62 2022 2028 [28] Husmann LK Johnson W 1994 Cytotoxicity of extracellular Legionella pneumophila. Infect Immun 62 2111 2114 [29] Marruchella A Franco C 2003 Severe alveolar haemorrhage in legionella pneumonia. Sarcoidosis Vasc Diffuse Lung Dis 20 77 [30] Sundar KM Pearce MJ 2004 Diffuse alveolar hemorrhage due to Legionella pneumonia. Sarcoidosis Vasc Diffuse Lung Dis 21 158 159 [31] Chastre J Raghu G Soler P Brun P Basset F 1987 Pulmonary fibrosis following pneumonia due to acute Legionnaires' disease. Clinical, ultrastructural, and immunofluorescent study. Chest 91 57 62 [32] Massey R Kumar P Pepper JR 2003 Innocent victim of a localised outbreak: legionella endocarditis. Heart 89 16e- [33] Tompkins LS Roessler BJ Redd SC Markowitz LE Cohen ML 1988 Legionella prosthetic-valve endocarditis. N Engl J Med 318 530 535 [34] McCabe RE Baldwin JC McGregor CA Miller DC Vosti KL 1984 Prosthetic valve endocarditis caused by Legionella pneumophila. Ann Intern Med 100 525 527 [35] Swanson MS Isberg RR 1996 Identification of Legionella pneumophila mutants that have aberrant intracellular fates. Infect Immun 64 2585 2594