RESULTS: Full-length Mal with and without endotoxin stimula-tion (94, 61 RLU) resulted in an increased NF-kB response abovethat of vector only (77, 50 RLU). Transfection with dominant neg-ative constructs abrogates that response. Mal homeobox 1 and 2 withand without endotoxin stimulation (260, 219 RLU) produced amore vigorous NF-kB response than full-length Mal (221, 130RLU).

CONCLUSIONS: The TIR domain of Mal appears to be essentialfor the full response of TLR4/endotoxin signaling. Homeobox 2 ofthe TIR domain may contain an activity site important for MyD88-dependent TLR4 signaling.

Hemorrhagic shock recruits Toll-like receptor 4 tothe plasma membrane in macrophages: A novelmechanism of cell priming for LPS signalingKinga A Powers MD, Katalin Szaszi MD, PhD,Rachel Khadaroo MD, PhD, Patrick Tawadros MD,John C Marshall MD, Andras Kapus MD, PhD,Ori D Rotstein MD, FACSUniversity of Toronto, University Health Network, Toronto, ONCanada

INTRODUCTION: Oxidative stress (OS) generated by ischemia/reperfusion is known to prime inflammatory cells for increased re-sponsiveness to subsequent stimuli such as lipopolysaccharide (LPS).LPS signals via TLR4 in cluster with other molecules in lipid rafts.We hypothesized that OS primes macrophages by inducing migra-tion of the TLR4 to the lipid rafts and thus initiates NFkappaBsignaling.

METHODS: Rats bled to a MAP of 40mmHg were resuscitatedafter 1h with their shed blood plus Ringer’s Lactate. Alveolar macro-phages (AM) were recovered by bronchoalveolar lavage and treatedex vivo with LPS. Alternatively, RAW 264.7 cells were incubatedwith H2O2 100 mM �/- LPS. Flow Cytometry detected surfaceTLR4 and Fluorescence Resonance Energy Transfer (FRET) de-termined molecular proximity between TLR4 and a lipid raftmarker GM-1. NFkappaB translocation was measured by immuno-fluorescence.

RESULTS: AM from resuscitated rodents expressed increased sur-face TLR4, an effect inhibited by the antioxidant N-acetylcysteine.Consistent with a role for OS in this effect, in vitro H2O2 caused anincrease in surface TLR4 in RAW 264.7 cells which was prevented bydepleting intracellular calcium or disrupting the cytoskeleton, sug-gesting the involvement of receptor exocytosis. Further, FRET be-tween TLR4 and GM1 as well as biochemical analysis demonstratedthat OS redistributes TLR4 to lipid rafts. Preventing the oxidant-induced movement of TLR4 to lipid rafts using methyl-beta-cyclodextrin precluded the increased responsiveness of cells to LPS.

CONCLUSIONS: These studies suggest a novel mechanismwhereby OS might prime the innate immune system and contributeto the development of organ failure in trauma victims.

Insulin prevents oxidant-induced endothelial cell(EC) barrier dysfunction and actin cytoskeletalreorganization through a cGMP/ protein kinase Gsignaling pathwaySiddhartha Rath MD, Theodore Kalogeris PhD, Nicholas Mai,John Alexander PhD, Gazi Zibari MD,Richard Turnage MD, FACSLouisiana State University Health Sciences Center,Shreveport, LA

INTRODUCTION: Insulin induces endothelial nitric oxide (NO)release through a PI3Kinase / eNOS signaling cascade. We havepreviously related insulin-induced NO release to the prevention ofoxidant-induced barrier dysfunction in vitro. cGMP and PKG aredownstream effectors of many of the physiologic effects of NO. SincePKG inhibits rho mediated actin cytoskeletal reorganization, we pos-tulated that insulin’s effect on H2O2-mediated barrier dysfunction isdependent upon PKG activity.

METHODS: Pulmonary EC’s grown to confluence on 8 �M porefilters were exposed to media, H2O2(100�M), insulin(100nM) �H2O2, or insulin � H2O2 � the PKG inhibitor (DT-2, 1-4 �M).Monolayer permeability was quantitated by measuring the trans-endothelial electrical resistance (TEER) every 15 min. for 120 min.Other EC’s were exposed to media, H2O2 or H2O2 � dibutyrl-cGMP (1mM, cGMP analogue) after which TEER was measured.Fluorescent microscopy of AlexaFluor-488-phalloidin stained cellswas performed to relate changes in barrier dysfunction to H2O2-induced actin cytoskeletal reorganization.

RESULTS: Insulin prevented H2O2-induced barrier dysfunction,whereas inhibition of PKG abolished insulin’s protective effect.Dibutyrl cGMP mimicked insulin’s protective effect. Microscopydemonstrated actin stress fibers and intercellular gaps in cells exposedto H2O2 but not in cells treated with H2O2 � insulin or dibutyrlcGMP. In contrast, actin stress fibers and intercellular gaps werereadily apparent in cells exposed to H2O2 alone or H2O2 � insu-lin � DT-2.

CONCLUSIONS: Together with our previous work, these data sug-gest that insulin modulates H2O2-induced barrier dysfunction byactivating a signal transduction pathway involving IRS-1/2/PI3K/Akt/eNOS/guanylyl cyclase/PKG.

15 min 60 min 90 min 120 min

Control 100 � 1.2 94 � 1.1 92 � 0.2 92 � 0.6

H2O2 101 � 0.3 73 � 1.3** 66 � 1** 60 � 3.7***

Insulin � H2O2 97 � 1.9 94 � 2.7 90 � 3.1 87 � 2.6

KT-5823 � insulin � H2O2 87 � 0.4 80 � 0.1* 68 � 0.6** 59 � 0.3***

DT-2 � insulin � H2O2 98 � 0.8 83 � 3.4* 69 � 5.2** 60 � 4.6***

Dibutyrl cGMP � H2O2 99 � 0.6 90 � .03 88 � 0.2 85 � 0.5

SperNo � DT-2 � H2O2 97 � 1 81 � 0.8* 68 � 5** 63 � 4**

*p�0.05 vs. baseline; **p�0.01 vs. baseline; ***p�0.001 vs. baseline; mean � SEM;statistical analysis with repeated measures ANOVA; p�0.05 considered significant.

S36 Critical Care II J Am Coll Surg