Group B Streptococcus Induces Tumor Necrosis Factor in Neonatal Piglets Effect of the Tumor Necrosis Factor Inhibitor Pentoxifylline on Hemodynamics and Gas Exchange1-3
RONALD L. GIBSON: GREGORY J. REDDING, WILLIAM R. HENDERSON, and WILLIAM E. TRUOG
Introduction
Group B streptococcus (GBS), a common neonatal gram-positive pathogen, causes similar pathophysiologic features in human newborns and neonatal animal models of sepsis(1-4). Animal models of GBS sepsis delineate an early and late phase response (3, 4). Early effects of GBS infusion include pulmonary hypertension, reduced cardiac output, and hypoxemia « 1h). Late GBS effects include the continuation of early features, with a progressive fall in cardiac output (CO), systemic hypotension, and lung injury (2 to 6 h). Increased blood levels of thromboxane B2 (lXB2 ) , a stable metabolite of the potent vasoconstrictor 1XA2 , occur during the early phase, whereas increased blood levels of both lXB2 and 6-keto-prostaglandin FlU (6keto-PGF IU), a stable metabolite of the vasodilator prostacyclin (PGI 2) occur during the late phase (4). Inhibition of lXB2 synthesis by indomethacin or dazmegrel completely reversesthe early features of GBS bacteremia in piglets (4, 5). Indomethacin also blocks the late increases in lXB2 and PGI 2 and attenuates the late features of systemichypotension, reduction in CO, and hypoxemia (4). However, indomethacin does not inhibit late-phase increasesin pulmonary artery pressure 2 to 4 h into GBS infusion (4). These combined observations suggest that 1XA2 is an important mediator of the acute GBS effects, but mediators in addition to 1XA2 and PGI 2 contribute to the late GBS effects. Tumor necrosis factor (TNF), a proinflammatory and vasoactive cytokine, mimics many of the late features of GBS sepsis, including pulmonary hypertension, systemic hypotension, hypoxemia, and lung injury upon infusion in animals (6, 7). TNF maycause these effectsdirectly or indirectly via the induction of ara-
SUMMARY Group B atreptococcus (GBS), a common neonatal gram-positive pathogen, causes similar pathophysiologic features In human n_borns Ind neonatal animal models of sepsis. Previous reports suggest that medlatora In addition to 1XA, and PGI, contribute to the late effects of GBS Infusion (2 to 4 h), which Include peralstent Increases In Ppa, hypoxemia, systemic hypotension, and a progressive fall In CO.1\Jmornecrosis factor (TNF) Infusion In animals produces several of the late GBS effects. We hypothesized that GBS causes Increased serum TNF levels 2 to 4 h Into Infusion In neonatal piglets. We also postUlated that the TNF Inhibitor, pentoxlfylllne (PTF), would attenuate both GBS-Induced TNF production and late GBS effects. In piglets Infused with 1.25 x 10' cfu/kg/h of GBS, serum TNF levels (pg/ml, ELISA assay) significantly Increased at 2 h (231 ± 41) and at 4 h (1,047 ± 290, n 9). In piglets Infused with concomitant GBS+PTF, serum TNF levels at 4 h (208 ± 39, n = 8) were reduced compared to GBS alone piglets (p < 0.02). Control piglets Infused with 0.9% sellne or PTF alone for 4 h had no detectable serum TNF « 35). GBS alone and GBS+PTF Infusion caused similar Increases In serum lXB,levels at 1, 2, and 4 h. Serum 6-keto-PGF,C2 levels (pg/0.1 ml) significantly Increased at 4 h (85 ± 18) with GBS alone, and were more elevated at 4 h (306 ± 75) with GBS+PTF Infusion. PVR significantly Increased at 1, 2, and 4 h with GBS alone; PVR was reduced ('" 20%) at 1, 2, and 4 h with GBS+PTF compared with GBS alone Infusion (p < 0.05). GBS+ PTF piglets had significantly Improved arterial po, values (mm Hg) at 4 h (72 ± 2) compared with GBS alone piglets (59 ± 3). GBS+PTF plglata showed grester systemic hypotension at 4 h compared with GBS alone piglets. Weconclude that circulating TNF Is a potential mediator of late GBS effects In neonatal piglets. PTF treatment attenuates GBSInduced TNF production and results In a mild Improvement In pulmonary hemodynamics and hypoxemia In a piglet model of GBS sepsis. AM REV RESPIR DIS 1991; 143:5IS-604
=
chidonic acid metabolites (7, 8). Serum TNF is increased 1 to 4 h after infusion of endotoxin or gram-negative bacteria into adult humans and animals (9-11), and this time course is concordant with the late features of both GBS and gramnegativebacteremiain animal models (4). There are limited data on TNF production in neonatal sepsis (12), but human neonatal mononuclear cells produce TNF in vitro in response to endotoxin and mitogens (13, 14).The ability of GBS to stimulate TNF production has not been tested. Pentoxifylline (PTF), a methylxanthine, inhibits endotoxin-induced TNF production in vitro, and blunts TNFinduced cytotoxicity (15, 16). Pretreatment of adult animals with PTF decreases both TNF- and endotoxininduced lung injury, lung neutrophil accumulation, and systemic hypotension
(16-18). Serum TNF levels of activity have not been measured during endotoxin + PTF infusion in animals (17, 18). However, PTF was recently shown to inhibit the mild rise in serum TNF induced by low dose endotoxin infusion in humans (19). There is no information on the effect of PTF on serum TNF levels (Received in original form February 12, 1990 and in revised form October 15, 1990) 1 From the Departments of Pediatrics and Medicine, University of Washington School of Medicine, Seattle, Washington. 2 Supported by Program Project Grant No. HL39157 and by Grant No. HL-30542 from the National Institutes of Health. 3 Correspondence should be addressed toRonald L. Gibson, M.D., Ph.D., Department of Pediatrics, RD-20, University of Washington School of Medicine, Seattle, WA 98195. 4 Recipient of an RJR Nabisco Pulmonary Research Scholar Award.
GROUP B STREPTOCOCCUS AND TUMOR NECROSIS FACTOR
or pulmonary hemodynamics and gas exchange in neonatal animal models of sepsis. We hypothesized that (1) GBS causes increased serum TNF levels 2 to 4 h into infusion in neonatal piglets and is a mediator of late GBS effects; (2) PTF infusion would attenuate late-phase GBSinduced TNF production, pulmonary hypertension, systemic hypotension, and hypoxemia; (3) PTF infusion would not alter GBS-induced thromboxane or prostacyclin synthesis in neonatal piglets. Methods
Animal Preparation Twenty-four piglets, 13 ± 3 days of age and weighing 3.6 ± 0.5 kg, were anesthetized intravenously (30 mg/kg pentobarbital), paralyzed intravenously (0.3 mg/kg pancuronium bromide), anticoagulated intravenously with heparin (1,000IV IV), and mechanically ventilated via a tracheostomy tube with a largeanimal Harvard ventilator (Harvard Apparatus Co., South Natick, MA) adjusted to deliver a tidal volume of 12 ± 2 ml/kg at a rate to maintain Paco, at 35 to 40 mm Hg during baseline conditions. All animals were ventilated with room air throughout each experiment. As previously described (20), catheters were placed in (1) the left external jugular vein for infusion of GBS and PTF (double lumen 5 Fr Swan-Ganz catheter), (2) the aorta to measure systemic arterial pressure (Psa) and sample arterial blood for pH, blood gas tensions, and eicosanoid and TNF measurements, and (3) a branch of the left pulmonary artery (5 Fr Swan-Ganz thermodilution catheter) to measure pulmonary artery and capillary wedge pressures (Ppa and Ppcw), cardiac output (CO) in triplicate by thermodilution using an Edwards 9520A cardiac output computer (Edwards Laboratories, Santa Ana, CAl and for sampling mixed venous blood. After instrumentation, anesthesia and muscle paralysisweremaintained intravenously with pentobarbital (3 mg/kg) and pancuronium bromide (0.3 mg/kg), respectively, for 1 h every day. The piglets receivedsigh breaths to 30 em H,O proximal airway pressure every 20 min to minimize spontaneous development of atelectasis. Vascular pressures were measured using Hewlett-Packard 1280transducers (Hewlett-Packard, Waltham, MA) referenced to midchest. Core temperature was maintained at 38.5 ± 0.5° C with an overhead radiant heat source. GBS Strain and Preparation The GBS strain is a type III clinical isolate made rifampin- and streptomycin-resistant (COH 31 r/s) (20). The culture conditions, mode of resuspension in sterile nonbacteriostatic saline, tests of culture purity, and measurement of bacterial concentrations wereperformed as described (20). The GBS suspensions for infusion into pigletscontained < 0.03 units (EU)/ml of endotoxin based on the
limulus amebocyte lysate assay (5 EU/ml = 1 ng/ml) (Associates of Cape Cod, Woods Hole, MA).
Eicosanoid Assays Two-milliliterarterial blood samples wereobtained under each of four experimental conditions. The blood samples were drawn into cold inhibitor solution containing indomethacin and sodium EDTA and centrifuged, and the decanted plasma was frozen at - 70° C until radioimmunoassays for lXB" a stable hydrolysis product of 'IXA" and 6-ketoPGF,a, a stable metabolite of PGI" were performed (21). lXB, and 6-keto-PGF,a were assayed by measuring competitive inhibition of ['H]lXB, to rabbit anti-Ixll, binding or [3H]6-keto-PGF,a to anti-6-keto-PGF,a, respectively (21).The average of duplicate assays was used to determine group means for the different experimental conditions. The limits of detection for both serum lXB, and 6-keto-PGF,a were c 10 pg/O.1 ml. Matrix effects caused by protein present in piglet plasma were measured in standard curves using eicosanoid-free piglet plasma prepared by charcoal stripping (21).
599
perature. The reaction was stopped with 100ul of 4.5 N H,S04' and the optical density was read at 490 nm. A relationshipbetween OD at 490 nm and TNF concentration is derived from the standard curvebylinearregression. The piglet serum TNF levels were derived by interpolation from the averageof the two assays. The limits of TNF detection by this assay were.". 35 pg/ml.
TNF L-929 Bioassay TNF activity was measured in a subset of piglet serum samples used in the ELISA assay by an adaptation of the previously described lytic assay of mouse L-929fibroblasts (22). Briefly, mouse L-929 fibroblasts were grown to f\J 75% confluence in RPMI media with 3% fetal calf serum cells and then trypsinized, and 3 X 104 cells were plated into each of 96-well microtiter plates in 100-~1 aliquots containing 0.1 ug/ml actinomycin D (CalB!ochem-Behring, San Diego, CAl. Appropnate duplicate dilutions of piglet serum samples and human TNF-alpha standards in RPMI media were then added (final volume/well = 200 ~1). To determine the specificity of the assay, appropriate dilutions of TNF ELISA Assay selectedserum samplesor TNF standards were 1\vo-milliliterarterial blood samples wereob- preincubated for 1h at room temperature with tained under each of four experimental con- 100 neutralizing units of antihuman TNF ditions for each piglet. The samples were monoclonal antibody (Lot 4707-25; Genenstored on ice during the protocol and centri- tech), the same antibody as used in the ELISA fuged, and the decanted serum was frozen and assay. The assay was incubated at 37° C in stored at -70° C until a TNF-alpha ELISA 5% CO, for 18 h. Samples were decanted, assay was performed. (A human TNF-alpha and the wells were washed three times with serum ELISA kit was kindly provided by 200 ul of PBS. Residual L-929 cells were Genentech Inc., San Francisco, CA.) Ninety- stained with crystalviolet (0.5070 in 20% methsix-well microtiter plates (Nunc-Immuno Plate anol) for 10min and washed three times with Maxisorp, certified; Nunc, Roskilde, Den- distilled water. The wellswere then eluted for mark) were coated with 100 ul of rabbit anti- 30 min with 100ul of extraction buffer (30.5 TNF-alpha coat antibody (Lot 4701-92) at ml 0.1 M Na citrate, 19.5 ml O.1N HCI, and 0.5 ug/ml in 0.05 M sodium carbonate buffer 50 ml 95% ethanol). The optical density at at pH 9.6 for 14 to 18 h at 4° C. Plates 570 nm was then read on a Dynatech Microwere washed three times with a minimum of elisa MR580 Auto Reader (Dynatech Labora200 ~l of PBS containing 0.05070 1\veen 20 tories, Alexandria, VA), and a relations~ip using the Nunc-Immunowash. The wellswere between OD 570nm and TNF concentratIOn treated with 200 ul of PBS containing 5 mg/rnl wasderived from the standard curve. Optical bovine serum albumin (BSA) and 0.05% density of L·929 cells incubated in media 1\veen 20 (Sigma Chemical, St. Louis, MO) alone represented 0% lysis; OD at 570 nm for 2 h at room temperature (23° C), and the for L-929 cells treated with 1% ltiton-X-loo plates werewashed as before.A human recom- in distilled water represented 100% lysis. A binant TNF-alpha (Lot THI51A; Genentech) concentration of 1 TNF unit/ml caused 50% standard curve was prepared over the range lysisof the L-929 cells. Serum TNF-alpha conof 35 to 2,000 pg/ml in PBS containing BSA centrations (units/ml) were then interpolat5 mg/ml, and duplicate lOO-~1 samples of ed from the average of duplicate samples using these standards were applied to the microti- the standard curve. The lower limit of detecter plates. Additional controls included PBS tion was f\J 50 pg/ml (5 U/ml). and PBS + BSA (5 mg/ml). Duplicate loo-~l samples of piglet serum were also applied to Experimental Design the wells, the microtiter plates were incubat- After instrumentation, all piglets were given ed for 14to 16 h at 4° C, and the plates were 10ml/kg 0.9% saline intravenously to ensure washed as before. Then 100ul of horseradish- a standardized euvolemic state. Four groups peroxidase-conjugated mouse monoclonal ofpiglets werestudied, Group 1 (GBSalone): anti-TNF-alpha (Lot 4707-25; Genentech) nine piglets received a 2-ml/kg bolus of sawere applied to the wells and incubated for line intravenouslyfollowed by saline infusion 2 h at room temperature on a rotary shaker. at 2 ml/kg/h, and 30 min later they received The plates were then washed as before and GBS intravenously at 1.25 x 109 colonyincubated with 100 ul of o-phenylenediamine forming units (cfu)/kg/h for 4 h. Group 2 substrate solution for 20 min at room tem- (GBS+ PTF): eight piglets received a 20-
600
GIBSON, REDDING, HENDERSON, AND TRUOG
1200 jG8S
"" 1000
~-
LL
Z
I-
E 2
til
70
Results
1400
~/#
•
/
800
//
600 /
400
/
* ,/' ::!* -
200
~ 1h
2h
,,*
-0 GBS + PTF /Saline
PTF
4h
Fig. 1. GBS infusion increases serum TNF; PTF attenuates GBS-induced TNF production. Serum TNF polypeptide levels are plotted versus time for the four groups of piglets. Values are expressedas mean ± SEM. Group mean values that were below the limits of detection are plotted as 35 pglml. Asterisks indicate p < 0.05 compared with the intragroup PRE value. Number sign indicates p < 0.05 for GBS+PTF compared with the GBS group value at the same time point.
mg/kg bolus of pentoxifylline (2 ml/kg dissolved in 0.9070 saline) intravenously followed by a PTF continuous infusion at 20 mg/kg/h, and 30 min later received intravenously GBS at 1.25 x 109 cfu/kg/h for 4 h. Group 3 (saline): three piglets received0.9% saline alone intravenously at 2 ml/kg/h for 4 h to control for the effects of anesthesia, surgery, and the duration of the protocol. Group 4 (PTF alone): four piglets received a 20-mg/kg bolus of PTF alone intravenously followed by 20 mg/kg/h infusion for 4 h to control for the effects of PTF alone. The piglets in each group were similar in age and weight. In each piglet, Ppa, Ppcw, Psa, CO, arterial and mixed venous blood gas tensions, and arterial blood samples for measurement of lXB2 , 6-keto-PGF ta , and TNF were obtained at baseline (PRE) and at 1, 2, and 4 h. Pulmonary vascular resistance was calculated (PVR == Ppa - Ppcw/CO) at the same time points for each piglet.
Statistical Analysis Analysis of variance with Student-NewmanKeul correction for multiple comparisons was used to compare values between the experimental groups; Student's paired t tests were used to compare intragroup values (SPSS-PC+). A Mann-Whitney U test was used for intragroup comparisons of data with a non-normative distribution. A p value of < 0.05 was considered significant.
Tumor Necrosis Factor Serum TNF polypeptide levels during the 4-h infusion for each group of piglets are depicted in figure 1. Serum TNF polypeptide levels significantly increased 2 h after the onset of continuous GBS infusion, and they increased further after 4 h of GBS alone infusion. PTF administration significantly reduced serum TNF polypeptide levelsafter 4 h of GBS infusion in GBS + PTF pigletscompared with GBS alone piglets. However, despite PTF treatment, serum TNF polypeptide levels remained significantly elevated from baseline at 2 and at 4 h in GBS + PTF piglets. No serum TNF was detected in the PRE or l-h serum samples for GBS alone and GBS + PTF piglets. Saline alone and PTF alone control piglets produced no detectable serum TNF polypeptide during the 4-h infusion. A subset of baseline and 4-h serum samples from saline (n = 1) and GBS alone (n = 4) infused piglets were also assayed for TNF bioactivity using the mouse fibroblast L-929 lytic assay. The saline-infused piglet had < 5 U TNF/ml at baseline and 8 U TNF/ml at 4 h; GBSinfused piglets had 7 ± 5 U TNF/ml at baseline and 849 ± 240 U TNF/ml at 4 h (p < 0.05). Preincubation of the 4-h GBS samples with antihuman TNF antibody reduced activity to < 10 U TNF/ml for all four samples. Eicosanoids The serum TxB2 and 6-keto-PGF ttt levels for each group of piglets during the 4-h protocol are shown in table 1. Serum ThB2 levels were significantly increased to a similar degree in both GBS alone and GBS+PTF piglets at 1,2, and 4 h after the onset of infusion. Serum 6-ketoPGF iu levels increased significantly by 4 h of infusion in both the GBS alone and GBS+ PTF groups. In addition, serum 6-keto-PGF t tt levelsat 4 h were significantly greater in GBS + PTF than in
~
~
~--.-i.---._·_·_·. GBS /0-. _..li GBS+ PTF / I. - -- --6- _ - ;
50
I
40
E
30
~
20
.s
.
~
II ' II
• '
; PTF rJ.::.=~==~====~ Satine
10~
O+--~--~--~-~-
PRE
2h
1h
Fig. 2. GBS infusion causes sustained increases in PVR; PTF mildly reduces GBS-induced increases in PVR. PVR (Ppa - Ppcw/CO) values are plotted versus time for the four groups of piglets. Values are expressed as mean ± SEM. Asterisks indicate p < 0.05 compared with the intragroup PRE value. Number signs indicate p < 0.05 for GBS+PTF compared with the GBS group value at the same time point.
GBS alone piglets. Significantly elevated serum levelsof Txll, or 6-keto-PGF t tt were not observed at I, 2, or 4 h in the control saline and PTF alone piglets.
Pulmonary Hemodynamics Changes in PVR and Ppa, respectively, during the 4-h infusion for each group of piglets are depicted in figures 2 and 3. PVR and Ppa were significantly increased at I, 2, and 4 h in GBS alone and GBS + PTF piglets. PTF treatment caused significant 21, 32, and 20% reductions in the mean PVR at 1, 2, and 4 h, respectively, in GBS+PTF piglets compared with GBS alone piglets. In GBS + PTF piglets, PTF also caused a significant reduction (22070) in the mean Ppa at 4 h of infusion compared with GBS alone piglets. Saline and PTF alone piglets had no significant changes in Ppa or PVR during the 4-h infusions. Systemic Hemodynamics Changes in CO and Psa, respectively, for each group of piglets are depicted in figures 4 and 5. CO was significantly reduced at I and 2 h, and was further reduced at 4 h in GBS alone and GBS+ PTF piglets. In GBS + PTF piglets, CO
TABLE 1 SERUM TxB 2 AND 6-KETO-PGF,. LEVELS IN PIGLETS· TxB 2 (pg/0.1 mil
6-keto-PGF,. (pg/0.1 ml)
Groups
PRE
1h
2h
4h
PRE
1h
2h
4h
Saline, n = 3 PTF, n = 4 GBS, n = 9 GBS + PTF, n = 8
< 10 < 10
< 10 14 ± 8 84 ± 26t 103 ± 12t
< 10 12 ::!: 6 104 ± 28t 101 ± 19t
< 10 12 ± 6 139 ± 33t 176 ± 45t
< 10 < 10 < 10 < 10
< 10 < 10 < 10
< 10 < 10
< 10 < 10
20 ± 15 37 ± 20
85 ± 18t 306 ::!: 75t:!:
12 ::!: 4 < 10
Definition of abbrevletlons: P~ • = pentoxilylline; GBS ~ Group B streptococcus . • Values are expressed as mean ± SEM. t p < 0.05 compared to the intragroup PRE value. p < 0.05 for GBS + PTF compared with the GBS alone value at the same time point.
*
4h
14 ± 7
601
GROUP B STREPTOCOCCUS AND TUMOR NECROSIS FACTOR
50
~
I-
40
~
l
a.
~Q)
30
wI )i E
~.s 20
~ o
~ * * .... -.;._.0-- _ . _ . - . -..... GBS *
-
-
-
/f
-0
GBS + PTF
*#
D
2 h
_
4 h
I
Systemic Blood Pressure (rnrn Hg)
50 6. Stephens KE, Ishizaka A, Larrick JW, Raffin TA. Tumor necrosis factor causes increased pul~M PTF to inhibit TNF-induced biomonary permeability and edema: comparison to activity in vitro (16). However, PTF blood septic acute lung injury. Am Rev Respir Dis 1988; levels have not been reported in PTF- 137:1364-70. treated animal models of sepsis nor in 7. Johnson J, Meyrick B, Jesmok G, Brigham KL. humans at the PTF dose used in this Human recombinant tumor necrosis factor alpha
603 infusion mimics endotoxemia in awake sheep. J Appl Physiol 1989; 66:1448-54. 8. Beutler BA. Orchestration of septic shock by cytokines: the role of cachectin (tumor necrosis factor). Prog Clin BioI Res 1989; 286:219-35. 9. Tracey KJ, Fong Y, Hesse DG, etat. Anticachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteremia. Nature 1987; 330:662-4. 10. Hesse DG, Tracey KJ, Fong Y, et at. Cytokine appearance in human endotoxemia and primate bacteremia. Surg Gynecol Obstet 1988; 166:147-53. ll. Michie HR, Manogue KR, Spriggs DR, et at. Detection of circulating tumor necrosis factor alpha after endotoxin administration. N Engl J Med 1988; 318:1481-6. 12. Girardin E, Berner M, Grau GE, Dayer J-M, Roux-Lombard P, Suter S. Tumour necrosis factor in neonatal listeriosis: a case report. Eur J Pediatr 1989; 148:644-5. 13. English BK, Burchett SK, English JD, Ammann AJ, Wara DW, Wilson CB. Production of Iymphotoxin and tumor necrosis factor by human neonatal mononuclear cells. Pediatr Res 1988; 24:717-22. 14. Weatherstone KB, Rich EA. Tumor necrosis factor/cachectin and interleukin-I secretion by cord blood rnonocytes from premature and term neonates. Pediatr Res 1989; 25:342-6. 15. Strieter RM, Remick DG, Ward PA, et at. Cellular and molecular regulation of tumor necrosis factor-alpha production by pentoxifylline. Biochem Biophys Res Commun 1988; 155:1230-6. 16. Lilly CM, Sandhu JS, Ishizaka A, et at. Pentoxifylline prevents tumor necrosis factor-induced lung injury. Am Rev Respir Dis 1989; 139:1361-8. 17. lshizaka A, Wu Z, Stephens KE, et at. Attenuation of acute lung injury in septic guinea pigs by pentoxifylline. Am Rev Respir Dis 1988; 138:376-82. 18. Welsh CH, Lien D, Worthen GS, Weil rv. Pentoxifylline decreases endotoxin-induced pulmonary neutrophil sequestration and extravascular protein accumulation in the dog. Am Rev Respir Dis 1988; 138:1106-14. 19. Zabel P, Schonharting MM, Wolter Dr, Schade UF. Oxpentifylline in endotoxemia. Lancet 1989; 2:1474-7. 20. Gibson RL, Redding GJ, Truog WE, Henderson WR, Rubens CEo Isogenic group B streptococci devoid of capsular polysaccharide or lJ-hemolysin: pulmonary hemodynamic and gas exchange effects during bacteremia in piglets. Pediatr Res 1989; 26:241-5. .: 21. Truog WE, Gibson RL, Juul SE, Henderson WR, Redding GJ. Neonatal group B streptococcal sepsis: effects of late treatment with dazmegrel. Pediatr Res 1988; 23:352-6. 22. Aggarwal BB, Kohr WJ, Hass PE, et at. Human tumor necrosis factor. J Bioi Chern 1985; 260:2345-54. 23. Fast DJ, Schlievert PM, Nelson RD. Toxic shock syndrome-associated staphylococcal and streptococcal pyrogenic toxins are potent inducers of tumor necrosis production. Infect Immun 1989; 57:291-4. 24. Glover DM, Brownstein D, Burchett SB, Larsen A, Wilson CB. Expression of HLA class II antigens and secretion of interleukin-I by monocytes and macrophages from adults and neonates. Immunology 1987; 61:195-201. 25. Waage A, Brandtzaeg P, Halstensen A, Kierulf P, Espevik T. The complex pattern of cytokines in serum from patients with meningococcal septic shock: association between interleukin 6 and interleukin I and fatal outcome. J Exp Med 1989; 169:333-8. 26. Beutler B, Cerami A. Cachectin and tumour necrosis factor as two sides of the same biological
604 coin. Nature 1986; 320:584-8. 27. Mathison JC, Wolfson E, Ulevitch RJ. Participation of tumor necrosis factor in the mediation of gram-negative bacterial lipopolysaccharideinduced injury in rabbits. J Clin Invest 1988; 81:1925-37. 28. Okusawa S, Gelfand JA, Ikejima T, Connol-
ly RJ, Dinarello CA. Interleukin-l induces a shocklike state in rabbits: synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition. J Clin Invest 1988; 81:1162-72. 29. Hakim TS, Petrella J. Attenuation of pulmonary and systemic vasoconstriction with pentoxifylline and aminophylline. Can J Physiol Pharmacol 1988; 66:396-401.
GIBSON, REDDING, HENDERSON, AND TRUOG
30. Chick TW, Scotto P, Icenogle MV, et al. Ef-
fects of pentoxifylline on pulmonary hemodynamics during acute hypoxia in anesthetized dogs. Am Rev Respir Dis 1988; 137:1099-1103. 31. Matzky R, Darius H, Schror K. The release of prostacyclin (PGI.) by pentoxifylline from human vascular tissue. Arzneimittelforschung 1982; 32:1315-8. 32. McFall TL, Zimmerman GA, Augustine NH,
Hill HR. Effect of group B streptococcal typespecific antigen on polymorphonuclear leukocyte function and polymorphonuclear leukocyte-endothelial cell interaction. Pediatr Res 1987; 21:517-23. 33. Engelhardt B, Sandberg K, Bratton D, et al. The role of granulocytes in the pulmonary response
to group B streptococcal toxin in young lambs. Pediatr Res 1987; 21:159-65. 34. Hammerschmidt DE, Kotasek D, McCarthy T, Huh P-W, Freyburger G, Vercellotti GM. Pentoxifylline inhibits granulocyte and platelet function. J Lab Clin Med 1988; 112:254-62. 35. Ings RMJ, Nudemberg F, Burrows JL, Bryce TA. The pharmacokinetics of oxpentifyllinein man when administered by constant intravenous infusion. Eur J Clin Pharmacol 1982; 23:539-43. 36. Beerman B, Ings R, Mansby J, Chamberlain J, McDonald A. Kinetics of intravenous and oral pentoxifylline in healthy subjects. Clin Pharmacol Ther 1985; 37:25-8.
RONALD L. GIBSON: GREGORY J. REDDING, WILLIAM R. HENDERSON, and WILLIAM E. TRUOG
Introduction
Group B streptococcus (GBS), a common neonatal gram-positive pathogen, causes similar pathophysiologic features in human newborns and neonatal animal models of sepsis(1-4). Animal models of GBS sepsis delineate an early and late phase response (3, 4). Early effects of GBS infusion include pulmonary hypertension, reduced cardiac output, and hypoxemia « 1h). Late GBS effects include the continuation of early features, with a progressive fall in cardiac output (CO), systemic hypotension, and lung injury (2 to 6 h). Increased blood levels of thromboxane B2 (lXB2 ) , a stable metabolite of the potent vasoconstrictor 1XA2 , occur during the early phase, whereas increased blood levels of both lXB2 and 6-keto-prostaglandin FlU (6keto-PGF IU), a stable metabolite of the vasodilator prostacyclin (PGI 2) occur during the late phase (4). Inhibition of lXB2 synthesis by indomethacin or dazmegrel completely reversesthe early features of GBS bacteremia in piglets (4, 5). Indomethacin also blocks the late increases in lXB2 and PGI 2 and attenuates the late features of systemichypotension, reduction in CO, and hypoxemia (4). However, indomethacin does not inhibit late-phase increasesin pulmonary artery pressure 2 to 4 h into GBS infusion (4). These combined observations suggest that 1XA2 is an important mediator of the acute GBS effects, but mediators in addition to 1XA2 and PGI 2 contribute to the late GBS effects. Tumor necrosis factor (TNF), a proinflammatory and vasoactive cytokine, mimics many of the late features of GBS sepsis, including pulmonary hypertension, systemic hypotension, hypoxemia, and lung injury upon infusion in animals (6, 7). TNF maycause these effectsdirectly or indirectly via the induction of ara-
SUMMARY Group B atreptococcus (GBS), a common neonatal gram-positive pathogen, causes similar pathophysiologic features In human n_borns Ind neonatal animal models of sepsis. Previous reports suggest that medlatora In addition to 1XA, and PGI, contribute to the late effects of GBS Infusion (2 to 4 h), which Include peralstent Increases In Ppa, hypoxemia, systemic hypotension, and a progressive fall In CO.1\Jmornecrosis factor (TNF) Infusion In animals produces several of the late GBS effects. We hypothesized that GBS causes Increased serum TNF levels 2 to 4 h Into Infusion In neonatal piglets. We also postUlated that the TNF Inhibitor, pentoxlfylllne (PTF), would attenuate both GBS-Induced TNF production and late GBS effects. In piglets Infused with 1.25 x 10' cfu/kg/h of GBS, serum TNF levels (pg/ml, ELISA assay) significantly Increased at 2 h (231 ± 41) and at 4 h (1,047 ± 290, n 9). In piglets Infused with concomitant GBS+PTF, serum TNF levels at 4 h (208 ± 39, n = 8) were reduced compared to GBS alone piglets (p < 0.02). Control piglets Infused with 0.9% sellne or PTF alone for 4 h had no detectable serum TNF « 35). GBS alone and GBS+PTF Infusion caused similar Increases In serum lXB,levels at 1, 2, and 4 h. Serum 6-keto-PGF,C2 levels (pg/0.1 ml) significantly Increased at 4 h (85 ± 18) with GBS alone, and were more elevated at 4 h (306 ± 75) with GBS+PTF Infusion. PVR significantly Increased at 1, 2, and 4 h with GBS alone; PVR was reduced ('" 20%) at 1, 2, and 4 h with GBS+PTF compared with GBS alone Infusion (p < 0.05). GBS+ PTF piglets had significantly Improved arterial po, values (mm Hg) at 4 h (72 ± 2) compared with GBS alone piglets (59 ± 3). GBS+PTF plglata showed grester systemic hypotension at 4 h compared with GBS alone piglets. Weconclude that circulating TNF Is a potential mediator of late GBS effects In neonatal piglets. PTF treatment attenuates GBSInduced TNF production and results In a mild Improvement In pulmonary hemodynamics and hypoxemia In a piglet model of GBS sepsis. AM REV RESPIR DIS 1991; 143:5IS-604
=
chidonic acid metabolites (7, 8). Serum TNF is increased 1 to 4 h after infusion of endotoxin or gram-negative bacteria into adult humans and animals (9-11), and this time course is concordant with the late features of both GBS and gramnegativebacteremiain animal models (4). There are limited data on TNF production in neonatal sepsis (12), but human neonatal mononuclear cells produce TNF in vitro in response to endotoxin and mitogens (13, 14).The ability of GBS to stimulate TNF production has not been tested. Pentoxifylline (PTF), a methylxanthine, inhibits endotoxin-induced TNF production in vitro, and blunts TNFinduced cytotoxicity (15, 16). Pretreatment of adult animals with PTF decreases both TNF- and endotoxininduced lung injury, lung neutrophil accumulation, and systemic hypotension
(16-18). Serum TNF levels of activity have not been measured during endotoxin + PTF infusion in animals (17, 18). However, PTF was recently shown to inhibit the mild rise in serum TNF induced by low dose endotoxin infusion in humans (19). There is no information on the effect of PTF on serum TNF levels (Received in original form February 12, 1990 and in revised form October 15, 1990) 1 From the Departments of Pediatrics and Medicine, University of Washington School of Medicine, Seattle, Washington. 2 Supported by Program Project Grant No. HL39157 and by Grant No. HL-30542 from the National Institutes of Health. 3 Correspondence should be addressed toRonald L. Gibson, M.D., Ph.D., Department of Pediatrics, RD-20, University of Washington School of Medicine, Seattle, WA 98195. 4 Recipient of an RJR Nabisco Pulmonary Research Scholar Award.
GROUP B STREPTOCOCCUS AND TUMOR NECROSIS FACTOR
or pulmonary hemodynamics and gas exchange in neonatal animal models of sepsis. We hypothesized that (1) GBS causes increased serum TNF levels 2 to 4 h into infusion in neonatal piglets and is a mediator of late GBS effects; (2) PTF infusion would attenuate late-phase GBSinduced TNF production, pulmonary hypertension, systemic hypotension, and hypoxemia; (3) PTF infusion would not alter GBS-induced thromboxane or prostacyclin synthesis in neonatal piglets. Methods
Animal Preparation Twenty-four piglets, 13 ± 3 days of age and weighing 3.6 ± 0.5 kg, were anesthetized intravenously (30 mg/kg pentobarbital), paralyzed intravenously (0.3 mg/kg pancuronium bromide), anticoagulated intravenously with heparin (1,000IV IV), and mechanically ventilated via a tracheostomy tube with a largeanimal Harvard ventilator (Harvard Apparatus Co., South Natick, MA) adjusted to deliver a tidal volume of 12 ± 2 ml/kg at a rate to maintain Paco, at 35 to 40 mm Hg during baseline conditions. All animals were ventilated with room air throughout each experiment. As previously described (20), catheters were placed in (1) the left external jugular vein for infusion of GBS and PTF (double lumen 5 Fr Swan-Ganz catheter), (2) the aorta to measure systemic arterial pressure (Psa) and sample arterial blood for pH, blood gas tensions, and eicosanoid and TNF measurements, and (3) a branch of the left pulmonary artery (5 Fr Swan-Ganz thermodilution catheter) to measure pulmonary artery and capillary wedge pressures (Ppa and Ppcw), cardiac output (CO) in triplicate by thermodilution using an Edwards 9520A cardiac output computer (Edwards Laboratories, Santa Ana, CAl and for sampling mixed venous blood. After instrumentation, anesthesia and muscle paralysisweremaintained intravenously with pentobarbital (3 mg/kg) and pancuronium bromide (0.3 mg/kg), respectively, for 1 h every day. The piglets receivedsigh breaths to 30 em H,O proximal airway pressure every 20 min to minimize spontaneous development of atelectasis. Vascular pressures were measured using Hewlett-Packard 1280transducers (Hewlett-Packard, Waltham, MA) referenced to midchest. Core temperature was maintained at 38.5 ± 0.5° C with an overhead radiant heat source. GBS Strain and Preparation The GBS strain is a type III clinical isolate made rifampin- and streptomycin-resistant (COH 31 r/s) (20). The culture conditions, mode of resuspension in sterile nonbacteriostatic saline, tests of culture purity, and measurement of bacterial concentrations wereperformed as described (20). The GBS suspensions for infusion into pigletscontained < 0.03 units (EU)/ml of endotoxin based on the
limulus amebocyte lysate assay (5 EU/ml = 1 ng/ml) (Associates of Cape Cod, Woods Hole, MA).
Eicosanoid Assays Two-milliliterarterial blood samples wereobtained under each of four experimental conditions. The blood samples were drawn into cold inhibitor solution containing indomethacin and sodium EDTA and centrifuged, and the decanted plasma was frozen at - 70° C until radioimmunoassays for lXB" a stable hydrolysis product of 'IXA" and 6-ketoPGF,a, a stable metabolite of PGI" were performed (21). lXB, and 6-keto-PGF,a were assayed by measuring competitive inhibition of ['H]lXB, to rabbit anti-Ixll, binding or [3H]6-keto-PGF,a to anti-6-keto-PGF,a, respectively (21).The average of duplicate assays was used to determine group means for the different experimental conditions. The limits of detection for both serum lXB, and 6-keto-PGF,a were c 10 pg/O.1 ml. Matrix effects caused by protein present in piglet plasma were measured in standard curves using eicosanoid-free piglet plasma prepared by charcoal stripping (21).
599
perature. The reaction was stopped with 100ul of 4.5 N H,S04' and the optical density was read at 490 nm. A relationshipbetween OD at 490 nm and TNF concentration is derived from the standard curvebylinearregression. The piglet serum TNF levels were derived by interpolation from the averageof the two assays. The limits of TNF detection by this assay were.". 35 pg/ml.
TNF L-929 Bioassay TNF activity was measured in a subset of piglet serum samples used in the ELISA assay by an adaptation of the previously described lytic assay of mouse L-929fibroblasts (22). Briefly, mouse L-929 fibroblasts were grown to f\J 75% confluence in RPMI media with 3% fetal calf serum cells and then trypsinized, and 3 X 104 cells were plated into each of 96-well microtiter plates in 100-~1 aliquots containing 0.1 ug/ml actinomycin D (CalB!ochem-Behring, San Diego, CAl. Appropnate duplicate dilutions of piglet serum samples and human TNF-alpha standards in RPMI media were then added (final volume/well = 200 ~1). To determine the specificity of the assay, appropriate dilutions of TNF ELISA Assay selectedserum samplesor TNF standards were 1\vo-milliliterarterial blood samples wereob- preincubated for 1h at room temperature with tained under each of four experimental con- 100 neutralizing units of antihuman TNF ditions for each piglet. The samples were monoclonal antibody (Lot 4707-25; Genenstored on ice during the protocol and centri- tech), the same antibody as used in the ELISA fuged, and the decanted serum was frozen and assay. The assay was incubated at 37° C in stored at -70° C until a TNF-alpha ELISA 5% CO, for 18 h. Samples were decanted, assay was performed. (A human TNF-alpha and the wells were washed three times with serum ELISA kit was kindly provided by 200 ul of PBS. Residual L-929 cells were Genentech Inc., San Francisco, CA.) Ninety- stained with crystalviolet (0.5070 in 20% methsix-well microtiter plates (Nunc-Immuno Plate anol) for 10min and washed three times with Maxisorp, certified; Nunc, Roskilde, Den- distilled water. The wellswere then eluted for mark) were coated with 100 ul of rabbit anti- 30 min with 100ul of extraction buffer (30.5 TNF-alpha coat antibody (Lot 4701-92) at ml 0.1 M Na citrate, 19.5 ml O.1N HCI, and 0.5 ug/ml in 0.05 M sodium carbonate buffer 50 ml 95% ethanol). The optical density at at pH 9.6 for 14 to 18 h at 4° C. Plates 570 nm was then read on a Dynatech Microwere washed three times with a minimum of elisa MR580 Auto Reader (Dynatech Labora200 ~l of PBS containing 0.05070 1\veen 20 tories, Alexandria, VA), and a relations~ip using the Nunc-Immunowash. The wellswere between OD 570nm and TNF concentratIOn treated with 200 ul of PBS containing 5 mg/rnl wasderived from the standard curve. Optical bovine serum albumin (BSA) and 0.05% density of L·929 cells incubated in media 1\veen 20 (Sigma Chemical, St. Louis, MO) alone represented 0% lysis; OD at 570 nm for 2 h at room temperature (23° C), and the for L-929 cells treated with 1% ltiton-X-loo plates werewashed as before.A human recom- in distilled water represented 100% lysis. A binant TNF-alpha (Lot THI51A; Genentech) concentration of 1 TNF unit/ml caused 50% standard curve was prepared over the range lysisof the L-929 cells. Serum TNF-alpha conof 35 to 2,000 pg/ml in PBS containing BSA centrations (units/ml) were then interpolat5 mg/ml, and duplicate lOO-~1 samples of ed from the average of duplicate samples using these standards were applied to the microti- the standard curve. The lower limit of detecter plates. Additional controls included PBS tion was f\J 50 pg/ml (5 U/ml). and PBS + BSA (5 mg/ml). Duplicate loo-~l samples of piglet serum were also applied to Experimental Design the wells, the microtiter plates were incubat- After instrumentation, all piglets were given ed for 14to 16 h at 4° C, and the plates were 10ml/kg 0.9% saline intravenously to ensure washed as before. Then 100ul of horseradish- a standardized euvolemic state. Four groups peroxidase-conjugated mouse monoclonal ofpiglets werestudied, Group 1 (GBSalone): anti-TNF-alpha (Lot 4707-25; Genentech) nine piglets received a 2-ml/kg bolus of sawere applied to the wells and incubated for line intravenouslyfollowed by saline infusion 2 h at room temperature on a rotary shaker. at 2 ml/kg/h, and 30 min later they received The plates were then washed as before and GBS intravenously at 1.25 x 109 colonyincubated with 100 ul of o-phenylenediamine forming units (cfu)/kg/h for 4 h. Group 2 substrate solution for 20 min at room tem- (GBS+ PTF): eight piglets received a 20-
600
GIBSON, REDDING, HENDERSON, AND TRUOG
1200 jG8S
"" 1000
~-
LL
Z
I-
E 2
til
70
Results
1400
~/#
•
/
800
//
600 /
400
/
* ,/' ::!* -
200
~ 1h
2h
,,*
-0 GBS + PTF /Saline
PTF
4h
Fig. 1. GBS infusion increases serum TNF; PTF attenuates GBS-induced TNF production. Serum TNF polypeptide levels are plotted versus time for the four groups of piglets. Values are expressedas mean ± SEM. Group mean values that were below the limits of detection are plotted as 35 pglml. Asterisks indicate p < 0.05 compared with the intragroup PRE value. Number sign indicates p < 0.05 for GBS+PTF compared with the GBS group value at the same time point.
mg/kg bolus of pentoxifylline (2 ml/kg dissolved in 0.9070 saline) intravenously followed by a PTF continuous infusion at 20 mg/kg/h, and 30 min later received intravenously GBS at 1.25 x 109 cfu/kg/h for 4 h. Group 3 (saline): three piglets received0.9% saline alone intravenously at 2 ml/kg/h for 4 h to control for the effects of anesthesia, surgery, and the duration of the protocol. Group 4 (PTF alone): four piglets received a 20-mg/kg bolus of PTF alone intravenously followed by 20 mg/kg/h infusion for 4 h to control for the effects of PTF alone. The piglets in each group were similar in age and weight. In each piglet, Ppa, Ppcw, Psa, CO, arterial and mixed venous blood gas tensions, and arterial blood samples for measurement of lXB2 , 6-keto-PGF ta , and TNF were obtained at baseline (PRE) and at 1, 2, and 4 h. Pulmonary vascular resistance was calculated (PVR == Ppa - Ppcw/CO) at the same time points for each piglet.
Statistical Analysis Analysis of variance with Student-NewmanKeul correction for multiple comparisons was used to compare values between the experimental groups; Student's paired t tests were used to compare intragroup values (SPSS-PC+). A Mann-Whitney U test was used for intragroup comparisons of data with a non-normative distribution. A p value of < 0.05 was considered significant.
Tumor Necrosis Factor Serum TNF polypeptide levels during the 4-h infusion for each group of piglets are depicted in figure 1. Serum TNF polypeptide levels significantly increased 2 h after the onset of continuous GBS infusion, and they increased further after 4 h of GBS alone infusion. PTF administration significantly reduced serum TNF polypeptide levelsafter 4 h of GBS infusion in GBS + PTF pigletscompared with GBS alone piglets. However, despite PTF treatment, serum TNF polypeptide levels remained significantly elevated from baseline at 2 and at 4 h in GBS + PTF piglets. No serum TNF was detected in the PRE or l-h serum samples for GBS alone and GBS + PTF piglets. Saline alone and PTF alone control piglets produced no detectable serum TNF polypeptide during the 4-h infusion. A subset of baseline and 4-h serum samples from saline (n = 1) and GBS alone (n = 4) infused piglets were also assayed for TNF bioactivity using the mouse fibroblast L-929 lytic assay. The saline-infused piglet had < 5 U TNF/ml at baseline and 8 U TNF/ml at 4 h; GBSinfused piglets had 7 ± 5 U TNF/ml at baseline and 849 ± 240 U TNF/ml at 4 h (p < 0.05). Preincubation of the 4-h GBS samples with antihuman TNF antibody reduced activity to < 10 U TNF/ml for all four samples. Eicosanoids The serum TxB2 and 6-keto-PGF ttt levels for each group of piglets during the 4-h protocol are shown in table 1. Serum ThB2 levels were significantly increased to a similar degree in both GBS alone and GBS+PTF piglets at 1,2, and 4 h after the onset of infusion. Serum 6-ketoPGF iu levels increased significantly by 4 h of infusion in both the GBS alone and GBS+ PTF groups. In addition, serum 6-keto-PGF t tt levelsat 4 h were significantly greater in GBS + PTF than in
~
~
~--.-i.---._·_·_·. GBS /0-. _..li GBS+ PTF / I. - -- --6- _ - ;
50
I
40
E
30
~
20
.s
.
~
II ' II
• '
; PTF rJ.::.=~==~====~ Satine
10~
O+--~--~--~-~-
PRE
2h
1h
Fig. 2. GBS infusion causes sustained increases in PVR; PTF mildly reduces GBS-induced increases in PVR. PVR (Ppa - Ppcw/CO) values are plotted versus time for the four groups of piglets. Values are expressed as mean ± SEM. Asterisks indicate p < 0.05 compared with the intragroup PRE value. Number signs indicate p < 0.05 for GBS+PTF compared with the GBS group value at the same time point.
GBS alone piglets. Significantly elevated serum levelsof Txll, or 6-keto-PGF t tt were not observed at I, 2, or 4 h in the control saline and PTF alone piglets.
Pulmonary Hemodynamics Changes in PVR and Ppa, respectively, during the 4-h infusion for each group of piglets are depicted in figures 2 and 3. PVR and Ppa were significantly increased at I, 2, and 4 h in GBS alone and GBS + PTF piglets. PTF treatment caused significant 21, 32, and 20% reductions in the mean PVR at 1, 2, and 4 h, respectively, in GBS+PTF piglets compared with GBS alone piglets. In GBS + PTF piglets, PTF also caused a significant reduction (22070) in the mean Ppa at 4 h of infusion compared with GBS alone piglets. Saline and PTF alone piglets had no significant changes in Ppa or PVR during the 4-h infusions. Systemic Hemodynamics Changes in CO and Psa, respectively, for each group of piglets are depicted in figures 4 and 5. CO was significantly reduced at I and 2 h, and was further reduced at 4 h in GBS alone and GBS+ PTF piglets. In GBS + PTF piglets, CO
TABLE 1 SERUM TxB 2 AND 6-KETO-PGF,. LEVELS IN PIGLETS· TxB 2 (pg/0.1 mil
6-keto-PGF,. (pg/0.1 ml)
Groups
PRE
1h
2h
4h
PRE
1h
2h
4h
Saline, n = 3 PTF, n = 4 GBS, n = 9 GBS + PTF, n = 8
< 10 < 10
< 10 14 ± 8 84 ± 26t 103 ± 12t
< 10 12 ::!: 6 104 ± 28t 101 ± 19t
< 10 12 ± 6 139 ± 33t 176 ± 45t
< 10 < 10 < 10 < 10
< 10 < 10 < 10
< 10 < 10
< 10 < 10
20 ± 15 37 ± 20
85 ± 18t 306 ::!: 75t:!:
12 ::!: 4 < 10
Definition of abbrevletlons: P~ • = pentoxilylline; GBS ~ Group B streptococcus . • Values are expressed as mean ± SEM. t p < 0.05 compared to the intragroup PRE value. p < 0.05 for GBS + PTF compared with the GBS alone value at the same time point.
*
4h
14 ± 7
601
GROUP B STREPTOCOCCUS AND TUMOR NECROSIS FACTOR
50
~
I-
40
~
l
a.
~Q)
30
wI )i E
~.s 20
~ o
~ * * .... -.;._.0-- _ . _ . - . -..... GBS *
-
-
-
/f
-0
GBS + PTF
*#
D
2 h
_
4 h
I
Systemic Blood Pressure (rnrn Hg)
50 6. Stephens KE, Ishizaka A, Larrick JW, Raffin TA. Tumor necrosis factor causes increased pul~M PTF to inhibit TNF-induced biomonary permeability and edema: comparison to activity in vitro (16). However, PTF blood septic acute lung injury. Am Rev Respir Dis 1988; levels have not been reported in PTF- 137:1364-70. treated animal models of sepsis nor in 7. Johnson J, Meyrick B, Jesmok G, Brigham KL. humans at the PTF dose used in this Human recombinant tumor necrosis factor alpha
603 infusion mimics endotoxemia in awake sheep. J Appl Physiol 1989; 66:1448-54. 8. Beutler BA. Orchestration of septic shock by cytokines: the role of cachectin (tumor necrosis factor). Prog Clin BioI Res 1989; 286:219-35. 9. Tracey KJ, Fong Y, Hesse DG, etat. Anticachectin/TNF monoclonal antibodies prevent septic shock during lethal bacteremia. Nature 1987; 330:662-4. 10. Hesse DG, Tracey KJ, Fong Y, et at. Cytokine appearance in human endotoxemia and primate bacteremia. Surg Gynecol Obstet 1988; 166:147-53. ll. Michie HR, Manogue KR, Spriggs DR, et at. Detection of circulating tumor necrosis factor alpha after endotoxin administration. N Engl J Med 1988; 318:1481-6. 12. Girardin E, Berner M, Grau GE, Dayer J-M, Roux-Lombard P, Suter S. Tumour necrosis factor in neonatal listeriosis: a case report. Eur J Pediatr 1989; 148:644-5. 13. English BK, Burchett SK, English JD, Ammann AJ, Wara DW, Wilson CB. Production of Iymphotoxin and tumor necrosis factor by human neonatal mononuclear cells. Pediatr Res 1988; 24:717-22. 14. Weatherstone KB, Rich EA. Tumor necrosis factor/cachectin and interleukin-I secretion by cord blood rnonocytes from premature and term neonates. Pediatr Res 1989; 25:342-6. 15. Strieter RM, Remick DG, Ward PA, et at. Cellular and molecular regulation of tumor necrosis factor-alpha production by pentoxifylline. Biochem Biophys Res Commun 1988; 155:1230-6. 16. Lilly CM, Sandhu JS, Ishizaka A, et at. Pentoxifylline prevents tumor necrosis factor-induced lung injury. Am Rev Respir Dis 1989; 139:1361-8. 17. lshizaka A, Wu Z, Stephens KE, et at. Attenuation of acute lung injury in septic guinea pigs by pentoxifylline. Am Rev Respir Dis 1988; 138:376-82. 18. Welsh CH, Lien D, Worthen GS, Weil rv. Pentoxifylline decreases endotoxin-induced pulmonary neutrophil sequestration and extravascular protein accumulation in the dog. Am Rev Respir Dis 1988; 138:1106-14. 19. Zabel P, Schonharting MM, Wolter Dr, Schade UF. Oxpentifylline in endotoxemia. Lancet 1989; 2:1474-7. 20. Gibson RL, Redding GJ, Truog WE, Henderson WR, Rubens CEo Isogenic group B streptococci devoid of capsular polysaccharide or lJ-hemolysin: pulmonary hemodynamic and gas exchange effects during bacteremia in piglets. Pediatr Res 1989; 26:241-5. .: 21. Truog WE, Gibson RL, Juul SE, Henderson WR, Redding GJ. Neonatal group B streptococcal sepsis: effects of late treatment with dazmegrel. Pediatr Res 1988; 23:352-6. 22. Aggarwal BB, Kohr WJ, Hass PE, et at. Human tumor necrosis factor. J Bioi Chern 1985; 260:2345-54. 23. Fast DJ, Schlievert PM, Nelson RD. Toxic shock syndrome-associated staphylococcal and streptococcal pyrogenic toxins are potent inducers of tumor necrosis production. Infect Immun 1989; 57:291-4. 24. Glover DM, Brownstein D, Burchett SB, Larsen A, Wilson CB. Expression of HLA class II antigens and secretion of interleukin-I by monocytes and macrophages from adults and neonates. Immunology 1987; 61:195-201. 25. Waage A, Brandtzaeg P, Halstensen A, Kierulf P, Espevik T. The complex pattern of cytokines in serum from patients with meningococcal septic shock: association between interleukin 6 and interleukin I and fatal outcome. J Exp Med 1989; 169:333-8. 26. Beutler B, Cerami A. Cachectin and tumour necrosis factor as two sides of the same biological
604 coin. Nature 1986; 320:584-8. 27. Mathison JC, Wolfson E, Ulevitch RJ. Participation of tumor necrosis factor in the mediation of gram-negative bacterial lipopolysaccharideinduced injury in rabbits. J Clin Invest 1988; 81:1925-37. 28. Okusawa S, Gelfand JA, Ikejima T, Connol-
ly RJ, Dinarello CA. Interleukin-l induces a shocklike state in rabbits: synergism with tumor necrosis factor and the effect of cyclooxygenase inhibition. J Clin Invest 1988; 81:1162-72. 29. Hakim TS, Petrella J. Attenuation of pulmonary and systemic vasoconstriction with pentoxifylline and aminophylline. Can J Physiol Pharmacol 1988; 66:396-401.
GIBSON, REDDING, HENDERSON, AND TRUOG
30. Chick TW, Scotto P, Icenogle MV, et al. Ef-
fects of pentoxifylline on pulmonary hemodynamics during acute hypoxia in anesthetized dogs. Am Rev Respir Dis 1988; 137:1099-1103. 31. Matzky R, Darius H, Schror K. The release of prostacyclin (PGI.) by pentoxifylline from human vascular tissue. Arzneimittelforschung 1982; 32:1315-8. 32. McFall TL, Zimmerman GA, Augustine NH,
Hill HR. Effect of group B streptococcal typespecific antigen on polymorphonuclear leukocyte function and polymorphonuclear leukocyte-endothelial cell interaction. Pediatr Res 1987; 21:517-23. 33. Engelhardt B, Sandberg K, Bratton D, et al. The role of granulocytes in the pulmonary response
to group B streptococcal toxin in young lambs. Pediatr Res 1987; 21:159-65. 34. Hammerschmidt DE, Kotasek D, McCarthy T, Huh P-W, Freyburger G, Vercellotti GM. Pentoxifylline inhibits granulocyte and platelet function. J Lab Clin Med 1988; 112:254-62. 35. Ings RMJ, Nudemberg F, Burrows JL, Bryce TA. The pharmacokinetics of oxpentifyllinein man when administered by constant intravenous infusion. Eur J Clin Pharmacol 1982; 23:539-43. 36. Beerman B, Ings R, Mansby J, Chamberlain J, McDonald A. Kinetics of intravenous and oral pentoxifylline in healthy subjects. Clin Pharmacol Ther 1985; 37:25-8.
