Cytokine secretion in vivo and ex vivo following chemotherapy of Mycobacterium tuberculosis infection

TRANSACTIONS OFTHEROYALSOCIETYOFTROPICALMEDICINEANDHYGIENE(1996)90,199-203

Cytokine secretion in viva and ex viva following Mycobacterium tuberculosis infection

chemotherapy

199

of

Jon S. Friedland’, John C. Hartleyz, Caroline G. C. Hartley 2, Robin J. Shattock3 and George E. Griffin3 ‘Departmen t of In f ectious Diseases and Bacteriology, Royal Postgraduate Medical School, London, WI2 ONN, UK; 2Monze District Hospital, Monze, Zambia; 3Division of Infectious Diseases, St George’s Hospital Medical School, London, S W17 ORE, UK Abstract

The human immune responseto tuberculosis is partly mediated by the proinflammatory cytokines tumour necrosis factor (TNF), interleukin (IL)-6, and IL-8. We investigated plasma concentrations of these cytokines before and after maximal lipopolysaccharide stimulation ex vivo of whole blood leucocytes from Zambian patients. 32 patients with non-fatal tuberculosis, 25 of whom were seropositive for human immunodeficiency virus (HIV), were followed for 9 months. Patients were assessedat presentation to hospital (visit A), after 2 months’ antimycobacterial therapy (visit B), and when chemotherapy was completed (visit C). Between visits A and B, patients regained weight (P=O.O3) and becameless anaemic (P=O.OOOl). At visit B, haemoglobin concentration remained lower in HIV seropositive patients (P=O.OOl) and the erythrocyte sedimentation rate (ESR), initially elevated in all patients, was higher in HIV seropositive patients (loo+-6 mm vs. 43?11 mm in 1 h in seronegative patients; P=O.O02). Plasma IL-8 concentrations were increased at visit C as was IL-8 secretion ex vivo (P
tuberculosis, Mycobacterium tuberculosis, chemotherapy, cytokine secretion, human immunodeficiency

Introduction

The clinical and pathological manifestations of tuberculosis appear to depend in part upon the orchestrated secretion of a number of proinflammatory cytokines. Secretion of tumour necrosis factor (TNF), interleukin (IL)-6 and IL-8 has been demonstrated in cellular models of tuberculosis (ROOKet al., 1987; FRIEDLANDet al., 1992a; FRIEDLAND, 1993). There is good evidence that each of these cytokines is likely to have an important role in the human immune responsein tuberculosis. For example, TNF is involved in weight loss (BEUTLERet al., 1985) and pivotal in granuloma formation (KINDLER et al., 1989; AMIRI et al., 1992).IL-6 is involved in stimulation of the hepatic acute phase response(GAULDIEet al., 1990), a characteristic and persistent feature of tuberculosis. IL-6 will also stimulate T cell proliferation (HOLSTI & RAULET, 1989) and activation (LOTZ et al., 1988). In addition, both TNF and IL-6 may have direct antimicrobial actions on the causative organism, Mycobacterium tuberculosis (see FLESCH& KAUFMANN, 1990a, 1990b). Chemotactic cytokines (chemokines) such as IL-8 have a central role in leucocyte recruitment to areas of granuloma formation (FRIEDLAND, 1995). Cellular studies have demonstrated that phagocytosis ofM. tuberculosis by human monocytic cells is a particularly potent stimulus for IL-8 gene expression and secretion and recruitment of antigen-specific T lymphocytes (FRIEDLAND et al., 1992a; WILKINSON & NEWMAN. 1992: ZHANG et al.. 1995): An important factor in the re-emergenceof tuberculosis as a major pathogen world-wide is concurrent infection with human immunodeficiency virus (HIV). This has been clearly documented in both developed (BARNES et al., 1991) and developing nations (ELLIOTT et aZ., 1990). Mycobacterial granuloma formation is impaired during HIV infection (NAMBUYAet al., 1988). However, there are contradictory reports about the effects of concurrent HIV infection on proinflammatory cytokine production (reviewed by BUTERA, 1994). There are no specific human data on the effects of tuberculosis and its treatment on cytokine secretion in patients with concurrent HIV infection. Address for correspondence: Dr Jon S. Friedland, Department of Infectious Diseases and Bacteriology, Royal Postgraduate Medical School, Hammersmith Hospital, Du Cane Road, London, W12 ONN, UK; fax +44 (0)lgl 740 3394.

virus

The present study was designed to investigate prospectively secretion of the proinflammatory cytokines TNF, IL-6 and IL-8 both in plasma and from whole blood leucocytes of patients inf&ted by M. tuberculosis, the maioritv of whom were HIV seronositive, which had been iaxiially stimulated by lipopdlysaccharide (LPS) ex vivo. The data obtained from 41 patients at admission to this study have already been published (FRIEDLANDet al., 1995) and we now report the longitudinal data from surviving patients who have been followed for up to 9 months. Plasma cytokine concentrations, leucocyte messenger ribonucleic acid (mRNA) production in vivo and stimulation ex vivo of whole blood leucocytes from patients have been assessedat 2 and 9 months after initial presentation. Materials

and Methods

Aspects of the protocol for investigation following admission to hospital have previously been described (FRIEDLANDet al., 1995) and hence it will be outlined only in brief together with details of methods used during the longitudinal phase of study. Patients

Consecutive patients admitted to Monze District Hospital, Zambia, with tuberculosis diagnosed by detection of acid-alcohol fast bacilli were recruited to this study. All subjects were aged over 12 years and they (or their relatives if profoundly unwell) gave informed consent to blood samples, including HIV antibody testing, before entering the study, which did not affect their treatment. Patients with tuberculosis were studied on admission to hospital (visit A), after 2 months’ antimycobacterial chemotherapy (visit B), and again after completing 9 months’ chemotherapy (visit C). Healthy blood donors who had been routinely screened for HIV antibody status and shown to be HIV seronegative were control subjects. Such subjects were bled on-only one occasion. Ethical annroval for the nroiect was obtained from the Zambian’fiesearch and &hi& Committee on AIDS and from Monze Hospital Ethics Committee. At each assessment,patients had a full clinical history taken and a chest X-ray, and haemoglobin level, leucocyte count, platelet count and erythrocyte sedimentation rate (ESR) were measured from a venous blood sample. Patient serum was tested at presentation for antibodies

JONSFRIEDLAND ETAL.

200

Table. Clinical characteristics of HIV seropositive and seronegative patients at presentation months’ anti-mycobacterial therapy (visit B), and at the end of 9 months’ treatment (visit C)

Visit Controls HIV+ No. of patients 5 25 Male:female ratio 4:l 6:19 29.6 (2) 31.5 (2) Age (years)a Weight (kg)a 70.0 (3.9) 43.1 (1.6) Haemoglobin (g/dL)a 14.8 (0.5) 8.6 (0.5)C White cell count (x 10YmL)a 3.0 (0.7) 5.4 (0.6) Erythrocyte sedimentation rate 23(11) 104 (8) (mm/h)a 3.5 (8.5) TNF (pg/mL)b 0 (0) IL-6 (pg/mL)b 31(11) 25 (37) IL-8 (pg/mL)b 2 (2) 20 (39) aMean (standard error in parentheses). bMedian (interquartile range in parentheses). ‘Significant difference (P=O.OOS). dSignificant difference (P=O*OOl). YSignilicant difference (P=O*OO2).

Visit HIVHIV+ 9 22 6:3 6:16 35.2 (6) 31.9 (2) 47.8 (2.2) 47.8 (2.2) 11.1 (0.4)C 11.1 (0.4)d 3.9 (0.3) 6.5 (1.0)

to HIV with Wellcozyme@ kits (Wellcome, Crewe, UK). Specific antimycobacterial and supportive therapy, including rehydration, were administered as clinically indicated and previously described (FRIEDLAND et al., 1995). Study protocol

At each assessment,20 mL of venous blood were collected from patients and divided into 4 aliquots of 5 mL. Each aliquot was placed in a 15 mL polypropylene, endotoxin-free tube containing potassium ethylenediaminetetraacetate. One aliquot was immediately centrifuged at 1000g for 5 min and plasma stored at -30°C for later cytokine analysis. In addition, buffy coat was removed from this sample, contaminating etythrocytes were lysed and leucocytes-homogenized in RNA extraction buffer as described (FRIEDLAND et al.. 1995). Eschtichia coli LPS (serotype 0127, B8) was added to the remaining whole blood aliquots to a final concentration of 10pg/mL to produce maximal cytokine releaseduring incubation ex viva. These aliquots were then incubated at 37°C and after 2,8 and 24 h a 5 mL blood aliquot was centrifuged at 1000 g for 5 min and plasma stored at -30°C for subsequent analysis of cytokine secretion from the stimulated patient leucocytes in whole blood. We have previously shown no loss of cytokine bioactivity or of mRNA when samples are stored in this matter (FRIEDLAND~~U~., 1992b). Plasma cytokine measurement

Plasma TNF, IL-6 and IL-8 concentrations were measured in category III containment laboratories in the UK using enzyme-linked immunosorbent assay kits according to the manufacturer’s instructions (R & D Systems, Minneapolis, USA). mRNA extraction and Northern analysis

RNA frozen in extraction buffer was transported to the UK in dry ice and subsequently processed as described bv FRIEDLAND et al. (1995). Brieflv, RNA underwent double phenol-chloroform and<. chloroformisoamyl alcohol extraction, precipitation in isopropanol and then in ethanol before being susnended in DEPC (diethyl pyrocarbonate) water. After quantification by optical densitometry (PyeAInicam SP6-450), 15 wg aiiquots of RNA were eiectrophoresed on denaturing formaldehvde 1% aaarose gels. transferred bv canillarv blotting to HybondTN@ (Almersham, UK), and fixed by exposure to ultraviolet light.

A

91(6) 4 (4.8) 22 (30.5) l(8.3)

100(6)e 0 (0.5) 6.5 (31) 4.5 (23)

B HIV9 6:3 35.7 (6) 52.6 (3.0) 14.7(0.5)d 5.0 (0.7) 43 (1l)e 8 (44.8) 32 (74.3) 4 (10.5)

(visit A), after 2

Visit HIV+ 10 28 27.7 (2) 50.6 (2.6) 12.7(0.8) 3.0 (0.2) 98 (13) lO(13) 0 (0) 94 (24)

C HIV3 2:l 32 (9.5) 57 (3.6) 14.6(1.0) 2.9 (0.3) 14(16) 2 (4.5) 32 (39) 70 (48.8)

Oligonucleotide probe hybridization of Northern blots from study samples together with those from appropriate positive (LPS stimulated) and negative control experiments was performed in the UK as described (F~IEDLAND et al., 1992b). Autoradiographs were scanned (Hewlett Packard Scaniet Plus@. Colorado. USA) to a Power Macintosh 61608 computer (Apple Computer Inc., California, USA) and analysed with Image 1.52 (from the National Institutes of Health Research Services Branch, USA). Statistical analyses

Normally distributed data (clinical) are presented as means f standard error of the mean, and groups were compared using the unpaired t test. Data not normally distributed (plasma cytokine concentrations) are presented as medians with interquartile ranges (25th and 75th centiles), and groups were compared using the Mann-Whitney U test. Values of P
The clinical characteristics of patients with non-fatal tuberculosis throughout the study period and those of controls are presented in the Table. Clinical and cytokine characteristics of non-survivors have been analysed and reported in detail previously (FRIEDLAND et al., 1995).At visit A, the weight of patients infected with M. tuberculosis was much less than that of control subjects (P=O*OOOl)and this was partly, but never fully, regained during the study. Thus, the weights of HIV seropositive and negative patients at visit B were greater than at visit A (P=O.O3).There was a trend throughout for HIV seropositive patients to be thinner than seronegative individuals but this was not statistically significant in the study population. Similarly, at visit B patients were less anaemic than at visit A (P=O.OOOl).Bv visit C. natient haemoglobin levels were not signilicanily different from controls’. At both visits A and B, HIV seropositive individuals were more anaemic than seronegative ones (P=O*OOSand 0.001 respectively). By visit C, too few HIV seronegative patients remained in the study for meaningful comparison, which partly reflects the distance that patients had to travel to reach the hospital (up to a 6 h journey on foot). Overall, the ESR remained elevated in patients compared to controls, although by visit B the absolute level was significantly lower in HIV se-

CYTOKINESAND

M. TUBERCULOSIS

ronegative individuals than in HIV seropositive patients

Cytokine secretion from whole blood leucocytes stimulated ex

(P=O.O02).

vivo Secretion of TNF from whole blood leucocytes maximally stimulated by LPS occurred within 2 h and was detectable throughout the 24 h period of incubation (Fig. 1, A). There was no significant difference between TNF secretion by leucocytes from control subjects or from patients at visits A, B or C. IL-6 secretion from LPS-stimulated whole blood was detectable at 2 h and reached maximal concentration at 8 h, and levels were similar at 24 h (Fig. 1, B). There was also no difference in IL-6 secretion from control leucocytes compared to that from leucocytes of patients studied at visits A, B or C. The kinetics of IL-8 secretion were similar in controls and patients at visits A, B and C, with elevated IL-8 concentrations present at 2 h and increasing throughout the 24 h study. However, the magnitude of IL-8 secretion was much greater at visit C with IL-8 concentrations following LPS stimulation of leucocytes being significantly higher (P
Plasma cytokine concentrations and peripheral leucocyte cytokine mRhLA content

There was no significant difference in circulating TNF or IL-6 concentrations in plasma of patients at visits A, B or C and these were similar to those of the controls, who were studied on one occasion. There was a trend for plasma IL-8 concentrations to be elevated in patients at visit A compared to controls but this was not significant and they remained at a similar level at visit B. However, after completion of antimycobacterial therapy at visit C, plasma IL-8 concentrations had risen to levels significantly higher than at visit B (P
A 3

IL-6 cont. (PM-W 100000

10000

IOOOOO B I

1000

100

10

‘d

24

Time (h) Fig. 2. Plasma IL-6 concentrations in HIV seropositive (m) and seronegative (0) patients following 2 months’ therapy with antimycobacterial drugs (log scale). Median values ? interquartile ranges are shown. Values at time = 0 represent cytokine concentrations in plasma and subsequent time points indicate IL-6 secretion into plasma by whole blood leucocytes after ex viva stimulation by lipopolysaccharide at a final concentration of 10 &mL. The difference in IL-6 secretion between the 2 groups was significant at 8 and 24 h (P
2

8

24

Time (h) Fig. 1. Plasma concentrations of TNF (A), IL-6 (B) and IL-8 (C) in all surviving patients with tuberculosis before (i.e., concentrations in patient circulation) and 2,8 and 24 h after ex viva stimulation of whole blood leucocytes with lipopolysaccharide at a final concentration of 10 &rnL (log scale). Median values + interquartile ranges are shown. Visit A, open boxes; visit B, diagonally hatched boxes; visit C, shaded boxes; control patients (seenonce only), horizontally hatched boxes.

Discussion In this prospective study, the patients with non-fatal tuberculosis improved clinically with antimycobacterial treatment, as was expected. Patients gained weight and became less anaemic. HIV seropositive patients had lower haemoglobin concentrations than HIV seronegative individuals throughout the study, although the absolute value increased in both groups following antimycobacterial chemotherapy (Table). In contrast, there was no significant alteration in white cell count during

JONS.FRIEDLANDETAL.

202

the study in either group of patients. The ESR approximately halved in HIV seronegative patients, but not in those who were seropositive, during the 9 months of chemotherapy. Even in HIV seronegative patients, the ESR remained above the level found in control subjects. This is consistent with the prolonged acute phase response that is characteristic of the immune response to tuberculosis. In terms of circulating plasma cytokine concentrations, the most striking change was at visit C, when a marked elevation in plasma concentration of the leucocyte chemoattractant IL-8 was observed (Fig. 1, C). This was not associatedwith an alteration in peripheral white cell count. There was no concurrent rise in plasma TNF or IL-6 concentrations, indicating that this was a specific effect. The measured absolute plasma IL-8 concentrations were lower than those in patients who died with tuberculosis (FRIEDLAND et al., 1995). In addition, higher plasma IL-8 levels have been reported in severe sepsis (FRIEDLAND et al., 1992b; HACK et al., 1992) and in Plasmodium falciparum malaria (FRIEDLAND et al., 1993). Since IL-8 may inhibit leucocyte-endothelial interactions (GIMBRONE et al., 1989), it is possible to speculate that the late rise in IL-8 may result in leucocyte retention within the circulation and prevent further recruitment of such cells to tissues with residual antigen but no active infection. This would provide an additional role for IL-8 in tuberculosis, as well as its being involved in recruitment of leucocytes to developing antimycobacterial granulomas. However, measurement of plasma IL-8 concentrations does not give information about overall functioning of chemotactic or haptotactic gradients (ROT, 1993). Alternatively, elevated plasma IL-8 concentations may reflect a non-specific reactive process following clearance of mycobacterial lesions. Indeed, the secretion of this cytokine into plasma might be a bystander phenomenon, with IL-8 having its principal actions in the tissues where cytokine concentrations are likely to be different. However, there is little doubt from cellular and animal studies that IL-8 is involved in the immune response to mycobacteria. At visits A and B, there was no difference between plasma cytokine concentrations in patients and in control subjects, despite the probable importance of these mediators in immunity to tuberculosis. Furthermore, the acute phase response,as measured by ESR, remained elevated in all patients throughout the recovery period with the ESR in HIV seropositive patients remaining similar at visits A and B, presumably due to viral activity, whilst being more than halved in HIV seronegative patients. The decreasein ESR in seronegative patients reflects resolution of tuberculosis. However, IL-6, which mediates this response, was not especially elevated in HIV seropositive patients. Plasma IL-6 level may be only an indirect reflection of the concentration of this cytokine in liver, the principal site of acute phase protein synthesis. Whole blood leucocytes were shown not to contain IL-6 (or TNF or IL-8) mRNA by Northern analysis, indicating that there had been no recent major transcriptional activation of these genes in peripheral leucocytes, although low levels of proinflammatory cytokine mRNA may be detectable in only some patients with tuberculosis using the sensitive polymerase chain reaction (SCHAUF et al., 1993). Interesting data emerged from the ex viva whole blood leucocyte studies. These experiments generate useful information about control of cytokine secretion since they utilize the host environment (DEFORGE & REMICK, 1991; WILSONet al., 1991; DE GROOTEet al., 1992). Such ex vzvo experiments have previously provided insight into the immunopathology of viral and bacterial infections (BROWN et al., 1994; VAN DEURENet al., 1994). In addition, they have the practical advantage of removing the need for sampling large volumes of blood from anaemic patients, as would be required for monocyte isola-

tion. LPS stimulation was used in the majority of studies, although mycobacterial lipoarabinomannan is an additional stimulus to cytokine secretion which it would be interesting to investigate in the specific context of tuberculosis. The surprising finding was that the magnitude of stimulated IL-8 secretion at visit C was approximately one log order of magnitude greater than that observed at visits A or B (Fig. 1, C). No such change was observed in TNF or IL-6 secretion ex vivo, indicating that the increase in IL-8 secretion was not simply a reflection of altered culture conditions at visit C. Increased IL-8 secretion did not appear to be secondary to raised leucocyte count since this did not change significantly during the study. There was a tendency for leucocyte numbers to decreaseby visit C, but we have no data on the percentage of monocytes. The kinetics of IL-8 secretion were unchanged and similar to those previously reported (DEFORGE& REMICK, 1991).The reason why tuberculosis infection appears to prime cells of the immune system circulating in blood for enhanced IL-8 secretion requires investigation. In contrast, no difference was observed between results of experiments ex vivo on control blood and those performed on blood collected at visits A or B. Thus, the overall pattern is very different from that of decreasedIL-8 secretion ex vivo, which we have reported previously in patients who died (FRIEDLAND et al., 1995). These findings were independent of HIV serological status. Previously elevated concentrations of proinflammatory cytokines have been reported in tuberculosis following stimulation of adhesion-purified patient monocytes (TAKASHIMA et al., 1990; OGAWAet aZ.,1991). However, the processof adhesion is known to modulate cytokine secretion (HASKILL et al., 1988; KASAHARAet al., 1991) and another advantage of the whole blood technique is that adhesion purification techniques are not used. Finally, our initial data demonstrated that HIV seropositivity did not affect plasma cytokine concentrations in survivors. In addition, at visit A, a paradox was noted (FRIEDLANDet al., 1995). IL-6 secretion ex vivo was significantly decreased in HIV seropositive patients compared to those who were seronegative, yet these patients had an equivalent acute phase response as documented by raised ESR. This phenomenon has now also been observed at visit B (Fig. 2)! despite the fact that, at visit B, ESR remained elevated m HIV seropositive patients but was reduced in seronegative individuals (Table). The samepattern appeared in the smaller number of patients attending visit C. These experiments ex vivo suggest both that this is an additional area in which the immune response in HIV infection is abnormal and that other factors contribute significantly to the acute phase reaction. However, this study principally focused on tuberculosis and an investigation to dissect fully the effect of concurrent HIV infection should ideally incorporate HIV patients without tuberculosis. In summary, the principal finding from this study was that both in vivo and LPS-stimulated ex-vivo plasma IL8 concentrations are increased following antibacterial therauv of tuberculosis. There is a considerable bodv of data from cellular, animal and human studies that implicates IL-8 as being central in T lymphocyte recruitment during the acute stageof granuloma formation. The current finding stresses the importance of longitudinal studies and indicates that dysregulation of chemokine secretion may be a late consequence of tuberculosis. However, further investigation in experimental models is required to determine the functional significance of the current observation. Acknowledgements

J. S. F., R. J. S. and G. E. G. were supported by the Medical Research Council (UK), and G. E. G. also by the Wellcome Trust of Great Britain. References

Amiri, I?., Locksley, R. M., Parslow, T. G., Sadick, M., Rector,

CYTOKINESAND

203

M.TUBERCULOSI.5

E., Ritter, D. & McKerrow, J. H. (1992). Tumour necrosis factor CLrestores granulomas and induces parasite egg-laying in schistosome-infected SCID mice. Nature, 356,604-607. Barnes, P. F., Bloch, A. B., Davidson, P. T. & Snider, D. E. (1991). Tuberculosis in patients with human immunodeficiency virus infection. New England Journal of Medicine, 324, 1644-1650. Beutler, B., Mahoney, J., Le Trang, N., Pekala, P. & Cerami, A. (1985). Purification of cachectin, a lipoprotein lipase-suppressing hormone secreted by endotoxin-induced RAW 264.7 cells.yournal of Experimental Medicine, 161,984-985. Brown, C. C., Poli, G., Lubaki, N., St Louis, M., Davachi, F., Musey, L., Manzila, T., Kovacs, A., Quinn, T. C. & Fauci, A. S. (1994). Elevated levels of tumour necrosis factor-cc in Zairian neonate plasmas: implications for perinatal infection with the human immunodeficiency virus. Journal of Znfectious Diseases, 169,975-980.

Butera, S. T. (1994). Interrelationships between cytokines and the human immunodeficiencv virus. Baillitie’s Clinical Znfectious Diseases, 1, 109-125. ’ DeForge, L. E. & Remick, D. G. (1991). Kinetics of TNF, IL-6 and IL-8 gene expression in LPS-stimulated whole blood. Biochemical and Biophysical Research Communications, 174,

18-24. De Groote, D., Zangerle, P. F., Gevaert, Y., Fassotte, M. F., Beguin, Y., Noizat-Pirenne, F., Pirenne, J., Gathy, R., Lopez, M.. Dehart. I.. Iaot. D.. Baudrihave. M.. Delacroix. D. & Franchimont, 6. rl9$2). direct sti&uiation of cytokiies (ILI@,TNF-q, IL-6, IL-2, IFN-1/ and GM-CSF) in whole blood. 1. Comparison with isolated PBMC stimulation. Cytokine, 4, 239-248. Elliott, A. M., Luo, N., Tembo, G., Halwiindi, B., Steenbergen, G., Machiels, L., Pobee, J., Nunn, P., Hayes, R. J. & McAdam? K. P. W. J. (1990). Impact of HIV on tuberculosis in Zambia: a cross sectional study. British Medical Journal, 301,412-415.

Flesch, I. E. A. & Kaufmann, S. H. E. (1990a). Stimulation of antibacterial macrophage activities by B-cell stimulatory factor 2 (interleukin-6). Infection and Immunity, S&269-271. Flesch, I. E. A. & Kaufmann, S. H. E. (1990b). Activation of tuberculostatic macronhage functions bv gamma interferon. interleukin-4 and t&our-necrosis facto;. fife&on and Zmmunity, 58,2675-2677.

Friedland, J. S. (1993). Phagocytosis, cytokines and Mycobacterium tuberculosis. Lymphokine

and Cytokine Research, 12,

127-133. Friedland, J. S. (1995). Chemokines and human infection. Clinical Science,88,393-lOO. Friedland, J. S., Remick, D. G., Shattock, R. & Griffin, G. E. (1992a). Secretion of interleukin-8 following phagocytosis of Mycobacterium tuberculosis by human monocyte cell lines. EuroDean ‘fburnal of Zmmunoloav, 22.1373-I 378.

Fried&d, -J. S., -Suputtam&kdl, Y., Remick, D. G., Chaowagul, W., Strieter, R. M., Kunkel, S. L., White, N. J. & Griffin, G. E. (1992b). Prolonged elevation of interleukin-8 and interleukin-6 concentrations in ulasma and of leucocvte interleukin-8 mRNA levels during iepticemic and localized Pseudomonas pseudomallei infection. Infection and Immunity, 60,2402-2408. Friedland. I. S.. Ho. M.. Remick. D. G.. Bunnap. D.. White. N. J. & G&n, 6. E: (1493). Inte;leukin~l and 3asmidium fhlciparum malaria in Thailand. Transactions of the Royal Societyof Tropical Medicine and Hygiene, 87,54-55.

Friedland, J. S., Hartley, J. C., Hartley, C. G. C., Shattock, R. J. & Griffin, G. E. (1995). Inhibition of ex viva proinflammatory cytokine secretion in fatal Mycobacterium tuberculosis infection. Clinical and Experimental Immunology, 100,233-238. Gauldie, J., Richards, C., Harnish, D., Lansdorf, P. & Baumann, H. (1990). Interferon /32/B-cellstimulatory factory type 2 shares identity with monocyte-derived hepatocyte-stimulating factor and regulates the major acute phase protein response in liver cells. Proceedingsof the National Academy of

(1989). Endothelial interleukin-8: a novel inhibitor of leucocyte-endothelial interactions. Science, 246,1601-1603. Hack. C. E.. Hart. M.. van Schiindel. R. 1. M. S., Eerenberg, A. J. h., Nuijens, j. H:, Thijs, i. G. & A&den, i. A. (1992yInterleukin-8 in sepsis: relation to shock and Inflammatory mediators. Znfection and Immunity, 60,2835-2842. Haskill, S., Johnson, C., Eierman, D., Becker, S. & Warren, K. (1988). Adherence induces selective mRNA expression of monocyte mediators and proto-oncogenes. Journal of Zmmunology, 140,1690-1694.

Holsti, M. A. & Raulet, D. H. (1989). IL-6 and IL-1 synergise to stimulate IL-2 production and proliferation of peripheral T cells.Journal of Immunology, 143,2514-2519.

Kasahara, K., Strieter, R. M., Chensue, S. W., Standiford, T. J. & Kunkel, S. L. (1991). Mononuclear cell adherence induced neutrophil chemotactic factoriinterleukin-8 gene expression. Journal of LeukocyteBiology, 50,287-295. Kindler, V., Sappino, A., Grau, G. E., Piguet, P. & Vassalli, P. (1989). The inducing role of tumour necrosis factor in the development of bactericidal granulomas during BCG infection. Cell, 56,731-740.

Lotz, M., Jirki, F., Kabouridis, P., Tsoukas, C., Hirano, T., Kishimoto, T. & Carson, D. A. (1988). B cell stimulating factor 2iinterleukin 6 is a costimulant for human thymocytes and T lymphocytes. Journal of Experimental Medicine, 167, 1253-1258. Nambuya, A., Sewankambo, N:, Mugerwa, J., Goodgame, R. & Lucas, S. (1988). Tuberculosis lymphadenitis associatedwith human immunodeficiency virus (HIV) in Uganda. Journal of Clinical Pathology, 41,93-96.

Ogawa, T., Uchida, H., Kusumoto, Y., Mori, Y., Yamamura, Y. & Hamada! S. (1991). Increase in tumour necrosis factor alpha- and mterleukin-6-secreting cells in peripheral blood mononuclear cells from subjects infected with Mycobacterium tuberculosis. Infection and Immunity, 59,3021-3025. Rook, G. A. W., Taverne, J., Leverton, C. & Steele, J. (1987). The role of gamma-interferon, vitamin D3, metabolites and tumour necrosis factor in the pathogenesis of tuberculosis. Immunology, 62,229-234.

Rot, A. (1993). Neutrophil attractantiactivation protein-l (interleukin-8) induces in vitro neutrophil migration by haptotactic mechanism. EuropeanJournal of Immunology, 23,303-306. Schauf, V., Rom, W. N., Smith, K. A., Sampalo, E. P., Meyn, P. A., Tramontan?, J. M., Cohn, Z. A. & Kaplan, G. (1993). Cytokine gene activation and modified responsivenessto interleukin-2 in the blood of tuberculosis patients. Journal of Infectious Diseases, 168, 1056-1059.

Takashima, T., Ueta, C., Tsuyuguchi, I. & Kishimoto, S. (1990). Production of tumour necrosis factor alpha by monocytes from patients with pulmonary tuberculosis. Infection and Zmmunity, 58,3286-3292.

Van Deuren, M., van der Ven-Jongekrijg, J., Demacker, P. N. M., Bartelink, A. K. M., van Dalen, R., Sauerwein, R. W., Gallati, H., Vannice, J. L. & van der Meer, J. W. M. (1994). Differential expression of proinflammatory cytokines and their inhibitors during the course of meningococcal infections.Journal of Znfectious Diseases, 169, 157-161. Wilkinson, P. C. & Newman, I. (1992). Identification of IL-8 as a locomotor attractant for activated human lymphocytes in mononuclear cell cultures with anti-CD3 or purified protein derivative of Mycobacterium tuberculosis. 3oumd of Zmm&ology, 149.2689-2694.

Wilson, B. M. G., Severn, A., Rapson, N. T., Chana, J. & Hopkins, P. (1991). A convenient human whole blood culture system for studvine the rermlation of tumour necrosis factor release by baiter&l lipopaysaccharide. Journal of Zmmunological Methods, 139,233-240.

Zhang, Y., Broser, M., Cohen, H., Bodkin, M., Lay, K., Reibman, J. & Rom, W. N. (1995). Enhanced interleukm-8 release and gene expression in macrophages after exposure to Mycobacterium tuberculosis and its components. Yournal of Clinical Znvestigation,95,586-592.

_

-

1

Sciences of the USA, 84,7251-7255.

Gimbrone, M. A., Obin, M. S., Brock, A. F., Luis, E. A., Hass, P. E., Hebert, C. A., Yip, Y. K., Leung, D. W., Lowe, D. G., Kohr, W. J., Darbonne, W. C., Bechtol, K. B. & Baker, J. B.

Received 21 August 1995; revised 6 October 1995; accepted 31 October 1995

for publication