TRENDSin Microbiology Vol.9 No.11 November 2001 531 News& Comment http://tim.trends.com 0966-842X/01/$ – see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S0966-842X(01)02176-X endothelial cells (human umbilical vein endothelial cells) is highly variable. We have not observed proliferation of either bovine aortic endothelial cells or a human microvascular cell line (HMEC-1) with B. henselae or with subcellular fractions of the bacterium. However, both of these cell types have been used extensively to study endothelial cell mitogens. Based on these observations, the possibility should be considered that the B. henselae angiogenic factor might act upon a second (non- endothelial) cell type to activate an effector molecule. It is certainly possible that other cell types can contaminate primary human endothelial cells (particularly low-passage cells). Thus, it is possible that the angiogenic factor from B. henselae is an activator of a cell growth factor or cytokine such as vascular endothelial growth factor (VEGF), rather than having a direct role as an endothelial cell mitogen. It is important to note that the model advanced by Dehio precludes the possibility of the involvement of a second cell type; however, others have noted similarities between the B. henselae angiogenic factor and VEGF (Refs 5,6). More recently, a leading role for VEGF in B. henselae -induced angiogenesis has been pr oposed 7 . Dr Dehio describes the differential expr ession of B. henselaegenes during interaction with host cells. In fact, we have shown that B. henselaestrongly upregulates the expression of surface proteins upon exposure to human endothelial cells. Using flow cytometry we have shown a fivefold increase in mean fluorescence intensity with intracellular bacteria reacted with rabbit anti- B. hensel aeserum compared with non-cell- associated bacteria reacted with the same serum (B. Anderson et al ., unpublished). Thus, the level of expression and/or the composition of individual B. henselae surface proteins changes inside the human endothelial cell. We have shown that the virB oper on is one such set of genes and includes the surface-exposed 17-kDa antigen 8 . The identification of other differentially expressed genes encoding surface proteins will undoubtedly facilitate the study of Bartonella–host interactions. Finally, Dr Dehio recognizes the importance of the B. henselae genome sequence to our understanding of the molecular biology and pathogenesis of B. henselae . This is clearly a point of universal agreement among all investigators working with B. henselae . The B. henselae genome sequence is expected to be released during the winter or spring of 2002 (S.G.E. Andersson, pers. commun.). Regardless, the elegant work of Dr Dehio has raised several important scientific questions about the study of the interaction of B. henselae with its host cell. Burt Anderson Dept of Medical Microbiology and Immunology, College of Medicine, University of South Florida, Tampa, Florida, USA. e-mail: banderso@hsc.usf.edu References 1 Dehio, C. (2001) Bartonella interactions with endothelial cells and erythrocytes. Trends M i cr obi ol . 9, 279–285 2 Regnat h, T. et al. (1998) Murine model of Bartonella henselaeinfection in the immunocompetent host. Infect. Immun. 66, 5534–5536 3 Karem, K. et al. (1999) Characterization of Bartonella henselae -specific immunity in BALB/c mice. Immunology 97, 352–358 4 Conley, T. et al. (1994) Rochalimaea species stimulate human endothelial cell proliferation and migration in vitro. J. Lab. Clin. Med. 124, 521–528 5 Maeno, N. et al. (1999) Live Bartonella henselae enhances endothelial cell proliferation without direct contact. Microb. Pathog. 27, 419–427 6 Wong, A. et al . (2001) Excessive tumor-elaborated VEGF and its neutralization define a lethal paraneoplastic syndrome. Proc. Natl. Acad. Sci. U. S. A. 98, 7481–7486 7 Kempf, V.A.J. et al . Evidence of a leading role for VEGF in Bartonella henselae -induced endothelial cell proliferations. Cel l M i cr obi ol . (i n pr ess) 8 Schmiederer, M. et al . (2001) Intracellular induction of the Bartonella henselae virB oper on by human endothelial cells. Infect. Immun. 69, 6495–6502 The interactions of Bartonella with endothelial cells and erythrocytes Response from Dehio A common feature among Bartonella species is the ability to parasitize the erythrocytes of their mammalian reservoir hosts 1 . Anderson points out that the recently described Bartonella tribocorum–rat infection model, which shows bacterial persistence within erythrocytes 2 , is most appropriate for understanding haemotropic infection in animal reservoirs. From a medical point of view, this model might be relevant for the design of appropriate strategies to prevent infection of animal reservoirs of zoonotic bartonellae (i.e. for controlling infection in the cat reservoir of the emerging human pathogen Bartonella henselae ). M or eover, the B. tribocorum–rat model might gain additional value by modelling human disease caused by Bartonella quintana and Bartonella bacilliformis. These two Bartonella species cause haemotropic infections in humans, which are their only known mammalian reservoir hosts. B. quintana infection typically results in trench fever, which resembles the intraerythrocytic infection course of B.tribocorum in rats in several ways 2 . Remarkably, the periodicity of feverish relapses of trench fever (also known as five-day fever) coincides with the time pattern of erythrocyte infection waves in the B. tribocorum–rat model 2 . Clinical research on trench fever has paid little attention to the interaction of B. quintana with erythrocytes. However, at the recent American Society for Rickettsiology/Bartonella Joint Conference 2001 (Big Sky, MT, USA; 17–22 August, 2001), Didier Raoult gave the first report of direct evidence for an intraerythrocytic localization of B.quintana in the blood of trench fever patients. It is now of paramount importance to investigate to what extent B. quintana can invade human erythrocytes and how this process might relate to the feverish relapses of trench fever. B. bacilliformis invades human erythrocytes resulting in Oroya fever, a haemolytic anaemia clearly distinguishable in this aspect from the non-haemolytic intraerythrocytic infection of rats by B. tribocorum 2 . It has been suggested that a recently identified contact-dependent haemolytic activity of B. bacilliformis contributes to this haemolytic infection course 3 . Heterologous expression of this haemolysin (and eventually auxiliary factors from B. bacilliformis ) in B.tribocorum might provide a haemolytic anaemia model in rats. The B. tribocorum–rat model is unsuitable for studying pathological angiogenesis, the most remarkable disease manifestation in humans. Angiogenesis can result from infection