- Email: [email protected]
Inhibition of adherence of Mycobacterium avium complex and Mycobacterium tuberculosis to fibronectin on the respiratory mucosa
ARTICLE IN PRESS Respiratory Medicine (2004) 98, 1203–1206
Inhibition of adherence of Mycobacterium avium complex and Mycobacterium tuberculosis to fibronectin on the respiratory mucosa A.M. Middletona, M.V. Chadwicka, A.G. Nicholsona, A. Dewara, R.K. Grogerb, E.J. Brownc, T.L. Ratliffd, R. Wilsona,* a
Host Defence Unit, Royal Brompton Hospital, Sydney Street, London SW3 6NP, UK Washington University, St Louis, USA c University of California, San Francisco, USA d University of Iowa, IA, USA b
Received 27 May 2003; accepted 27 April 2004
KEYWORDS Mycobacterium; Tuberculosis; Avium complex; Adherence; Inhibition
Summary Mycobacterium species adhere to the respiratory mucosa via mucus and fibronectin of extracellular matrix exposed by damaged epithelium. We have investigated whether inhibiting adherence to fibronectin influences subsequent infection of human respiratory tissue by Mycobacterium avium complex and Mycobacterium tuberculosis. Human respiratory tissue was pretreated with mycobacterial fibronectin attachment proteins prior to infection with M. avium complex and M. tuberculosis and the number of recoverable bacteria over time was compared to untreated controls. Inhibition significantly reduced recovery of M. avium complex at 15 min (P ¼ 0:02), 7 days (P ¼ 0:04), and 14 days (P ¼ 0:03); whereas recovery of M. tuberculosis was only reduced at 15 min (P ¼ 0:01) and not at later timepoints. We conclude that M. avium complex and M. tuberculosis infection of the mucosa proceeds by different mechanisms, since M. tuberculosis infection is independent of fibronectin adherence. & 2004 Elsevier Ltd. All rights reserved.
Introduction Mycobacterium avium complex (MAC) and Mycobacterium tuberculosis (MTB) both cause pulmonary disease in humans. MAC is less virulent than MTB, and is generally referred to as an opportunistic or facultative pathogen. Although MAC infections do occur in patients without any respiratory history, most patients have prior lung damage. In these *Corresponding author. Tel.: þ 44-(207)-351-8337; fax: þ 44(207)-351-8338. E-mail address: [email protected] (R. Wilson).
cases, the primary site of infection may be the airway mucosa. In contrast, MTB is an obligate human pathogen transmitted from patient-to-patient in aerosols of respiratory secretions and the primary route of infection of MTB is regarded as via interaction with the alveolar macrophages. Little is known about the virulence factors of MAC and MTB and whether they possess any that are specifically relevant to mucosal infections. We have previously investigated the role of adherence to fibronectin in the interaction of mycobacteria with the respiratory mucosa.1 In these studies, mycobacteria adhered to fibronectin in extracellular matrix (ECM), as well as
0954-6111/$ - see front matter & 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.rmed.2004.04.013
ARTICLE IN PRESS 1204
mucus with a fibrous appearance. They also adhered rarely to healthy unciliated epithelium, degenerated extruded epithelial cells and open tight junctions between epithelial cells.2,3 There have been a number of reported mechanisms of mycobacterial attachment to host cells, including heparin-binding haemagglutinin interaction with epithelial cells and at least five mycobacterial components that bind with HEp-2 cell extracts.4,5 Two distinct mycobacterial proteins or protein complexes, fibronectin attachment protein (FAP) and proteins of the antigen 85 complex, are involved in mycobacterial attachment to fibronectin.3,6 The ability to bind fibronectin is highly conserved in mycobacteria.3,7 We have previously shown that pre-incubation of MAC and MTB with soluble fibronectin and pre-incubation of tissue with FAP and/or antigen 85B protein (a85B) resulted in significant inhibition of the bacterial adherence to fibronectin exposed in areas of damage on the mucosal surface of organ cultures of human respiratory tissue.1,2 Antigen 85B protein, was selected for these experiments as it has been shown to be simultaneously associated with the bacterial cell wall as well as being secreted.8 We have also previously studied the interaction of MAC and MTB with the respiratory mucosa over 14 days where we showed that MAC increased in numbers extracellularly on the mucosal surface whereas MTB was lost from the surface, but invaded the organ culture and multiplied within the tissue.9 We have now investigated the effect of inhibiting adherence to fibronectin with FAP and a85B combined on MAC and MTB infection of human respiratory mucosa in an organ culture model with an air interface incubated over 14 days. We hypothesized that the interaction of MAC with fibronectin may be a more important pathogenic mechanism for MAC than MTB, and that by inhibiting adherence we may be able to reduce or prevent infection of the organ culture.
Materials and methods The organ culture using human nasal turbinate and bronchial tissue has been described previously.1,10,11 Each experiment required 12 pieces of dissected tissue from the same donor. Six experiments were performed each with tissue from a different donor. Tissue was either pre-incubated with sterile PBS or with 100 mg/ml M. avium FAP (amino acids 269–292) and 100 mg/ml purified M. bovis BCG a85B together by gently pipetting onto the mucosal surface. Tissue was incubated at
A.M. Middleton et al.
371C for 15 min and then washed ( 3) by gently pipetting 100 ml of PBS onto the mucosal surface. A clinical strain of MAC1 and MTB H37RV (ATCC 27294) were used. Washed bacteria resuspended in 100 ml of sterile PBS of which 2 ml containing approximately 1 108 cfu were inoculated onto each organ culture. Viable counts were performed and there was no significant difference between the inocula of the different experiments. Tissues were incubated at 371C in 5% CO2 in air for 15 min, 7 days and 14 days. Organ cultures were then washed by pipetting 100 ml (X3) of sterile PBS onto the tissue in order to remove unattached bacteria. Then each of the four edges of the organ culture were touched with a sterile loop and plated onto Columbia agar base (Lab M, Bury, UK) containing 5% defibrinated horse blood and Lo. wenstein–Jensen (L–J) medium (Royal Brompton Hospital, Microbiology Department, London, UK), in order to assess the sterility of the control organ cultures and the purity of the mycobacterial growth in the infected organ cultures. Tissue was removed from the organ culture taking care to remove all surrounding agar. One hundred microlitre of sterile PBS was placed in a sterile-graduated centrifuge tube and the meniscus was marked on the side of the tube. Tissue was placed in the PBS and the displacement volume was estimated by the amount of PBS required to be removed to return the meniscus to the marked level. The tissue was then placed in 1 ml of sterile PBS in a homogenizer grinder tube and homogenized three times for 10 s, with a 10 s interval between each spin to prevent temperature buildup in the tissue/PBS due to friction. Ten-fold dilutions of the homogenate were performed down to 106. Viable counts were performed and the number of bacteria recovered per microlitre of tissue was calculated using the displacement volume of the tissue. The number of MAC and MTB recovered from tissue pre-incubated with PBS was compared to the number recovered from tissue pre-incubated with FAP and a85B at the three incubation times. The number of bacteria recovered at different time intervals were analysed using Wilcoxon’s matched signed rank test, P-values of p0:05 were judged to be significant. All experiments were coded and randomized prior to analysis making the observer blind to the identity of each individual piece of tissue.
Results The results at 15 min reflect surface adherence, whereas 7 and 14 days reflect a mixture of surface
ARTICLE IN PRESS Inhibition of Mycobacteria to fibronectin
infection and tissue invasion. The figure shows the mean number of bacteria recovered from macerated organ culture tissue (cfu/ml) incubated for 15 min, 7 and 14 days after pre-incubation of tissue with either PBS or FAP and a85B together. The numbers of MAC recovered from tissue pre-incubated with FAP and a85B were reduced by 89.1% (P ¼ 0:02), 88.8% (P ¼ 0:04) and 86.7% (P ¼ 0:03) after 15 min, 7 and 14 days incubation, respectively, compared to tissue pre-incubated with PBS. Therefore, inhibition of MAC adherence to fibronectin substantially reduced the level of MAC infection of the organ culture. The numbers of MTB recovered from tissue pre-incubated with FAP and a85B were reduced by 92.7% (P ¼ 0:01), but at later time points there was no significant differ-
Figure 1 The mean number of M. avium complex (MAC) and M. tuberculosis (TB) recovered from macerated tissue (cfu/ml). Left shaded columns, 15 min incubation; clear, 7 days; right shaded, 14 days. MAC&PBS and TB&PBS, pre-incubation of tissue with PBS; MAC&FAP&85B and TB&FAP&85B, pre-incubation of tissue with 100 mg/ml FAP and 100 mg/ml a85B together. *P ¼ 0:02; ’ P ¼ 0:04; K P ¼ 0:03 MAC&PBS versus MAC&FAP&85B at 15 min, 7 and 14 days, respectively; m P ¼ 0:01 TB&PBS versus TB&FAP&85B at 15 min. Results are expressed as the mean and standard error. n ¼ 6 (Turbinate ¼ 2, Bronchial ¼ 4).
1205
ence between organ cultures incubated with either PBS or FAP and a85B together. There was a 35.5% (P40:05) and 30.7% (P40:05) reduction in the FAP/ a85B incubated tissue at 7 and 14 days incubation, respectively (Fig. 1).
Discussion The significant reduction in MAC and MTB adhering to tissue pre-incubated with FAP and a85B after only 15 min incubation was the same as shown in our previous studies.1,2 In these studies, SEM showed significant inhibition of mycobacterial adherence to ECM, the major site of adherence. The results of the extended incubation times, 7 and 14 days, showed that the numbers of MAC recovered were still significantly reduced and the large rise in MAC numbers recovered at 14 days, as seen in the untreated control experiment, did not occur. It may be that a critical number of MAC are required before rapid multiplication can occur, similar to the principle of ‘quorum sensing’ reported for other bacteria.12,13 These results show that MAC infection of the organ culture is dependent upon initial adherence to fibronectin. Our previous studies show that bacteria replicate at these sites to form discrete colonies. Therefore, adherence to fibronectin is an important pathogenic mechanism for MAC and could possibly provide an opportunity to interrupt the infectious process of binding could be inhibited in vivo. However, the number of MTB recovered at 7 and 14 days were no different compared to untreated control tissue. Therefore, in contrast to MAC, inhibition of adherence to fibronectin does not reduce MTB infection of the organ culture. This is likely to be because MTB infection proceeds via invasion of the mucosa, using mechanisms independent of FAP and a85B, such as the heparinbinding haemagglutinin described by Menozzi et al. This study highlights differences in pathogenic mechanisms of MAC and MTB with the respiratory mucosa, despite similarities in their initial interaction. Inhibition of MAC adherence to fibronectin, for example by immunization with fibronectin attachment proteins, could reduce mucosal infection by MAC.
Acknowledgements This research was funded by a grant from the Hawley Legacy and was conducted while Richard Groger was a Pfizer Postdoctoral Fellow.
ARTICLE IN PRESS 1206
A.M. Middleton et al.
References 1. Middleton AM, Chadwick MV, Nicholson AG, Dewar A, Groger RK, Brown EJ, Wilson R. The role of Mycobacterium avium complex fibronectin attachment protein in adherence to the human respiratory mucosa. Mol Microbiol 2000;38:381–91. 2. Middleton AM, Chadwick MV, Nicholson AG, Dewar A, Groger RK, Brown EJ, Ratliff TL, Wilson R. Interaction of Mycobacterium tuberculosis with human respiratory mucosa. Tuberculosis 2002;82:69–78. 3. Abou-Zeid C, Ratliff TL, Wiker HG, Harboe M, Bennedsen J, Rook GAW. Characterization of fibronectin-binding antigens released by Mycobacterium tuberculosis and Mycobacterium bovis BCG. Infect Immun 1988;56:3046–51. 4. Menozzi FD, Rouse JH, Alavi M, Laude-Sharp M, Muller J, Bischoff R, Brennan MJ, Locht C. Identification of a heparinbinding hemagglutinin present in mycobacteria. J Exp Med 1996;184:993–1001. 5. Reddy VM, Hayworth DA. Interaction of Mycobacterium tuberculosis with human respiratory epithelial cells (Hep-2). Tuberculosis 2002;82:31–6. 6. Ratliff TL, McCarthy R, Telle WB, Brown EJ. Purification of a mycobacterial adhesin for fibronectin. Infect Immun 1993;61:1889–94. 7. Schorey JS, Li Q, McCourt DW, Bong-Mastek M, Clark-Curtiss JE, Ratliff TL, Brown EJ. A Mycobacterium leprae gene
8.
9.
10.
11.
12. 13.
encoding a fibronectin binding protein is used for efficient invasion of epithelial cells and schwann cells. Infect Immun 1995;63:2652–7. Rambukkana A, Das PK, Burggraaf JD, Yong S, Faber WR, Thole JER, Harboe M. Heterogenicity of monoclonal antibody-reactive epitopes on mycobacterial 30-kilodaltonregion proteins and the secreted antigen 85 complex and demonstration of antigen 85B on the Mycobacterium leprae cell wall surface. Infect Immun 1992;60:5172–81. Middleton AM, Chadwick MV, Nicholson AG, Dewar A, Feldman C, Wilson R. Investigation of mycobacterial colonisation and invasion of the respiratory mucosa. Thorax 2003;58:246–51. Dowling RB, Rayner CFJ, Rutman A, Jackson AD, Kanthakumar K, Dewar A, Taylor GW, Cole PJ, Johnson M, Wilson R. Effect of salmeterol on Pseudomonas aeruginosa infection of respiratory mucosa. Am J Respir Crit Care Med 1997;155: 327–36. Jackson AD, Rayner CF, Dewar A, Cole PJ, Wilson R. A human respiratory-tissue organ culture incorporating an air interface. Am J Respir Crit Care Med 1996;153:1130–5. Dunny GM, Leonard BA. Cell–cell communication in gram positive bacteria. Am Rev Microbiol 1997;51:527–64. Salmond GPC, Bycroft BW, Stewart GS, Williams P. The bacterial ‘enigma’: cracking the code of cell–cell communication. Mol Microbiol 1995;16:615–24.
Inhibition of adherence of Mycobacterium avium complex and Mycobacterium tuberculosis to fibronectin on the respiratory mucosa A.M. Middletona, M.V. Chadwicka, A.G. Nicholsona, A. Dewara, R.K. Grogerb, E.J. Brownc, T.L. Ratliffd, R. Wilsona,* a
Host Defence Unit, Royal Brompton Hospital, Sydney Street, London SW3 6NP, UK Washington University, St Louis, USA c University of California, San Francisco, USA d University of Iowa, IA, USA b
Received 27 May 2003; accepted 27 April 2004
KEYWORDS Mycobacterium; Tuberculosis; Avium complex; Adherence; Inhibition
Summary Mycobacterium species adhere to the respiratory mucosa via mucus and fibronectin of extracellular matrix exposed by damaged epithelium. We have investigated whether inhibiting adherence to fibronectin influences subsequent infection of human respiratory tissue by Mycobacterium avium complex and Mycobacterium tuberculosis. Human respiratory tissue was pretreated with mycobacterial fibronectin attachment proteins prior to infection with M. avium complex and M. tuberculosis and the number of recoverable bacteria over time was compared to untreated controls. Inhibition significantly reduced recovery of M. avium complex at 15 min (P ¼ 0:02), 7 days (P ¼ 0:04), and 14 days (P ¼ 0:03); whereas recovery of M. tuberculosis was only reduced at 15 min (P ¼ 0:01) and not at later timepoints. We conclude that M. avium complex and M. tuberculosis infection of the mucosa proceeds by different mechanisms, since M. tuberculosis infection is independent of fibronectin adherence. & 2004 Elsevier Ltd. All rights reserved.
Introduction Mycobacterium avium complex (MAC) and Mycobacterium tuberculosis (MTB) both cause pulmonary disease in humans. MAC is less virulent than MTB, and is generally referred to as an opportunistic or facultative pathogen. Although MAC infections do occur in patients without any respiratory history, most patients have prior lung damage. In these *Corresponding author. Tel.: þ 44-(207)-351-8337; fax: þ 44(207)-351-8338. E-mail address: [email protected] (R. Wilson).
cases, the primary site of infection may be the airway mucosa. In contrast, MTB is an obligate human pathogen transmitted from patient-to-patient in aerosols of respiratory secretions and the primary route of infection of MTB is regarded as via interaction with the alveolar macrophages. Little is known about the virulence factors of MAC and MTB and whether they possess any that are specifically relevant to mucosal infections. We have previously investigated the role of adherence to fibronectin in the interaction of mycobacteria with the respiratory mucosa.1 In these studies, mycobacteria adhered to fibronectin in extracellular matrix (ECM), as well as
0954-6111/$ - see front matter & 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.rmed.2004.04.013
ARTICLE IN PRESS 1204
mucus with a fibrous appearance. They also adhered rarely to healthy unciliated epithelium, degenerated extruded epithelial cells and open tight junctions between epithelial cells.2,3 There have been a number of reported mechanisms of mycobacterial attachment to host cells, including heparin-binding haemagglutinin interaction with epithelial cells and at least five mycobacterial components that bind with HEp-2 cell extracts.4,5 Two distinct mycobacterial proteins or protein complexes, fibronectin attachment protein (FAP) and proteins of the antigen 85 complex, are involved in mycobacterial attachment to fibronectin.3,6 The ability to bind fibronectin is highly conserved in mycobacteria.3,7 We have previously shown that pre-incubation of MAC and MTB with soluble fibronectin and pre-incubation of tissue with FAP and/or antigen 85B protein (a85B) resulted in significant inhibition of the bacterial adherence to fibronectin exposed in areas of damage on the mucosal surface of organ cultures of human respiratory tissue.1,2 Antigen 85B protein, was selected for these experiments as it has been shown to be simultaneously associated with the bacterial cell wall as well as being secreted.8 We have also previously studied the interaction of MAC and MTB with the respiratory mucosa over 14 days where we showed that MAC increased in numbers extracellularly on the mucosal surface whereas MTB was lost from the surface, but invaded the organ culture and multiplied within the tissue.9 We have now investigated the effect of inhibiting adherence to fibronectin with FAP and a85B combined on MAC and MTB infection of human respiratory mucosa in an organ culture model with an air interface incubated over 14 days. We hypothesized that the interaction of MAC with fibronectin may be a more important pathogenic mechanism for MAC than MTB, and that by inhibiting adherence we may be able to reduce or prevent infection of the organ culture.
Materials and methods The organ culture using human nasal turbinate and bronchial tissue has been described previously.1,10,11 Each experiment required 12 pieces of dissected tissue from the same donor. Six experiments were performed each with tissue from a different donor. Tissue was either pre-incubated with sterile PBS or with 100 mg/ml M. avium FAP (amino acids 269–292) and 100 mg/ml purified M. bovis BCG a85B together by gently pipetting onto the mucosal surface. Tissue was incubated at
A.M. Middleton et al.
371C for 15 min and then washed ( 3) by gently pipetting 100 ml of PBS onto the mucosal surface. A clinical strain of MAC1 and MTB H37RV (ATCC 27294) were used. Washed bacteria resuspended in 100 ml of sterile PBS of which 2 ml containing approximately 1 108 cfu were inoculated onto each organ culture. Viable counts were performed and there was no significant difference between the inocula of the different experiments. Tissues were incubated at 371C in 5% CO2 in air for 15 min, 7 days and 14 days. Organ cultures were then washed by pipetting 100 ml (X3) of sterile PBS onto the tissue in order to remove unattached bacteria. Then each of the four edges of the organ culture were touched with a sterile loop and plated onto Columbia agar base (Lab M, Bury, UK) containing 5% defibrinated horse blood and Lo. wenstein–Jensen (L–J) medium (Royal Brompton Hospital, Microbiology Department, London, UK), in order to assess the sterility of the control organ cultures and the purity of the mycobacterial growth in the infected organ cultures. Tissue was removed from the organ culture taking care to remove all surrounding agar. One hundred microlitre of sterile PBS was placed in a sterile-graduated centrifuge tube and the meniscus was marked on the side of the tube. Tissue was placed in the PBS and the displacement volume was estimated by the amount of PBS required to be removed to return the meniscus to the marked level. The tissue was then placed in 1 ml of sterile PBS in a homogenizer grinder tube and homogenized three times for 10 s, with a 10 s interval between each spin to prevent temperature buildup in the tissue/PBS due to friction. Ten-fold dilutions of the homogenate were performed down to 106. Viable counts were performed and the number of bacteria recovered per microlitre of tissue was calculated using the displacement volume of the tissue. The number of MAC and MTB recovered from tissue pre-incubated with PBS was compared to the number recovered from tissue pre-incubated with FAP and a85B at the three incubation times. The number of bacteria recovered at different time intervals were analysed using Wilcoxon’s matched signed rank test, P-values of p0:05 were judged to be significant. All experiments were coded and randomized prior to analysis making the observer blind to the identity of each individual piece of tissue.
Results The results at 15 min reflect surface adherence, whereas 7 and 14 days reflect a mixture of surface
ARTICLE IN PRESS Inhibition of Mycobacteria to fibronectin
infection and tissue invasion. The figure shows the mean number of bacteria recovered from macerated organ culture tissue (cfu/ml) incubated for 15 min, 7 and 14 days after pre-incubation of tissue with either PBS or FAP and a85B together. The numbers of MAC recovered from tissue pre-incubated with FAP and a85B were reduced by 89.1% (P ¼ 0:02), 88.8% (P ¼ 0:04) and 86.7% (P ¼ 0:03) after 15 min, 7 and 14 days incubation, respectively, compared to tissue pre-incubated with PBS. Therefore, inhibition of MAC adherence to fibronectin substantially reduced the level of MAC infection of the organ culture. The numbers of MTB recovered from tissue pre-incubated with FAP and a85B were reduced by 92.7% (P ¼ 0:01), but at later time points there was no significant differ-
Figure 1 The mean number of M. avium complex (MAC) and M. tuberculosis (TB) recovered from macerated tissue (cfu/ml). Left shaded columns, 15 min incubation; clear, 7 days; right shaded, 14 days. MAC&PBS and TB&PBS, pre-incubation of tissue with PBS; MAC&FAP&85B and TB&FAP&85B, pre-incubation of tissue with 100 mg/ml FAP and 100 mg/ml a85B together. *P ¼ 0:02; ’ P ¼ 0:04; K P ¼ 0:03 MAC&PBS versus MAC&FAP&85B at 15 min, 7 and 14 days, respectively; m P ¼ 0:01 TB&PBS versus TB&FAP&85B at 15 min. Results are expressed as the mean and standard error. n ¼ 6 (Turbinate ¼ 2, Bronchial ¼ 4).
1205
ence between organ cultures incubated with either PBS or FAP and a85B together. There was a 35.5% (P40:05) and 30.7% (P40:05) reduction in the FAP/ a85B incubated tissue at 7 and 14 days incubation, respectively (Fig. 1).
Discussion The significant reduction in MAC and MTB adhering to tissue pre-incubated with FAP and a85B after only 15 min incubation was the same as shown in our previous studies.1,2 In these studies, SEM showed significant inhibition of mycobacterial adherence to ECM, the major site of adherence. The results of the extended incubation times, 7 and 14 days, showed that the numbers of MAC recovered were still significantly reduced and the large rise in MAC numbers recovered at 14 days, as seen in the untreated control experiment, did not occur. It may be that a critical number of MAC are required before rapid multiplication can occur, similar to the principle of ‘quorum sensing’ reported for other bacteria.12,13 These results show that MAC infection of the organ culture is dependent upon initial adherence to fibronectin. Our previous studies show that bacteria replicate at these sites to form discrete colonies. Therefore, adherence to fibronectin is an important pathogenic mechanism for MAC and could possibly provide an opportunity to interrupt the infectious process of binding could be inhibited in vivo. However, the number of MTB recovered at 7 and 14 days were no different compared to untreated control tissue. Therefore, in contrast to MAC, inhibition of adherence to fibronectin does not reduce MTB infection of the organ culture. This is likely to be because MTB infection proceeds via invasion of the mucosa, using mechanisms independent of FAP and a85B, such as the heparinbinding haemagglutinin described by Menozzi et al. This study highlights differences in pathogenic mechanisms of MAC and MTB with the respiratory mucosa, despite similarities in their initial interaction. Inhibition of MAC adherence to fibronectin, for example by immunization with fibronectin attachment proteins, could reduce mucosal infection by MAC.
Acknowledgements This research was funded by a grant from the Hawley Legacy and was conducted while Richard Groger was a Pfizer Postdoctoral Fellow.
ARTICLE IN PRESS 1206
A.M. Middleton et al.
References 1. Middleton AM, Chadwick MV, Nicholson AG, Dewar A, Groger RK, Brown EJ, Wilson R. The role of Mycobacterium avium complex fibronectin attachment protein in adherence to the human respiratory mucosa. Mol Microbiol 2000;38:381–91. 2. Middleton AM, Chadwick MV, Nicholson AG, Dewar A, Groger RK, Brown EJ, Ratliff TL, Wilson R. Interaction of Mycobacterium tuberculosis with human respiratory mucosa. Tuberculosis 2002;82:69–78. 3. Abou-Zeid C, Ratliff TL, Wiker HG, Harboe M, Bennedsen J, Rook GAW. Characterization of fibronectin-binding antigens released by Mycobacterium tuberculosis and Mycobacterium bovis BCG. Infect Immun 1988;56:3046–51. 4. Menozzi FD, Rouse JH, Alavi M, Laude-Sharp M, Muller J, Bischoff R, Brennan MJ, Locht C. Identification of a heparinbinding hemagglutinin present in mycobacteria. J Exp Med 1996;184:993–1001. 5. Reddy VM, Hayworth DA. Interaction of Mycobacterium tuberculosis with human respiratory epithelial cells (Hep-2). Tuberculosis 2002;82:31–6. 6. Ratliff TL, McCarthy R, Telle WB, Brown EJ. Purification of a mycobacterial adhesin for fibronectin. Infect Immun 1993;61:1889–94. 7. Schorey JS, Li Q, McCourt DW, Bong-Mastek M, Clark-Curtiss JE, Ratliff TL, Brown EJ. A Mycobacterium leprae gene
8.
9.
10.
11.
12. 13.
encoding a fibronectin binding protein is used for efficient invasion of epithelial cells and schwann cells. Infect Immun 1995;63:2652–7. Rambukkana A, Das PK, Burggraaf JD, Yong S, Faber WR, Thole JER, Harboe M. Heterogenicity of monoclonal antibody-reactive epitopes on mycobacterial 30-kilodaltonregion proteins and the secreted antigen 85 complex and demonstration of antigen 85B on the Mycobacterium leprae cell wall surface. Infect Immun 1992;60:5172–81. Middleton AM, Chadwick MV, Nicholson AG, Dewar A, Feldman C, Wilson R. Investigation of mycobacterial colonisation and invasion of the respiratory mucosa. Thorax 2003;58:246–51. Dowling RB, Rayner CFJ, Rutman A, Jackson AD, Kanthakumar K, Dewar A, Taylor GW, Cole PJ, Johnson M, Wilson R. Effect of salmeterol on Pseudomonas aeruginosa infection of respiratory mucosa. Am J Respir Crit Care Med 1997;155: 327–36. Jackson AD, Rayner CF, Dewar A, Cole PJ, Wilson R. A human respiratory-tissue organ culture incorporating an air interface. Am J Respir Crit Care Med 1996;153:1130–5. Dunny GM, Leonard BA. Cell–cell communication in gram positive bacteria. Am Rev Microbiol 1997;51:527–64. Salmond GPC, Bycroft BW, Stewart GS, Williams P. The bacterial ‘enigma’: cracking the code of cell–cell communication. Mol Microbiol 1995;16:615–24.