Donor pretreatment with FK506reduces reperfusion injury andaccelerates intestinal graftrecovery in ratsMihai Oltean, MD,a Rille Pullerits, MD,b Changlian Zhu, MD,c Klas Blomgren, MD,c

Eva C. Hallberg, PhD,a and Michael Olausson, MD,a Göteborg, Sweden

Background. FK506 alleviates warm ischemia-reperfusion injury, but it remains unknown if suchprotection is manifest after cold storage and transplantation. We studied the early outcome aftertransplantation of intestines from donors pretreated with FK506 compared to grafts from controlstreated with saline (154 mM NaCl).Methods. Sprague-Dawley rats received 0.3 mg/kg FK506 or saline intravenously 6 hours before graftretrieval. The small bowel was harvested, stored for 3 hours, and then transplanted heterotopically.Samples were taken after preservation and at 20 minutes, 6 hours, 12 hours, and 24 hours afterreperfusion. Heat shock protein 72 (Hsp72) and iintercellular adhesion molecule (ICAM)-1 expressionand nuclear factor �B (NF-�B) activation were assessed via Western blots and eelectrophoretic mobilityshift assay (EMSA), respectively. Dissacharidase activity and enterocyte proliferation rate were alsostudied.Results. Preservation injury was similar between groups, but pretreated grafts had better morphologyalready 20 minutes after reperfusion. Control grafts always had thinner mucosa and more PMNinfiltration. Hsp72 expression was greater in pretreated grafts. ICAM-1 was absent after harvesting,preservation, and immediately after reperfusion but increased in control grafts at the later time points.Control grafts showed a biphasic NF-�B activation pattern, whereas NF-�B activation was inhibitedeffectively in pretreated grafts. Dissacharidase activity decreased during the first 6 hours afterreperfusion but recovered within 24 hours in pretreated grafts but not in control grafts. Earlierenterocyte proliferation was observed in pretreated grafts.Conclusions. FK506 donor pretreatment reduced graft proinflammatory activation and neutrophilinflammation. Pretreated groups revealed a milder reperfusion injury and accelerated morphologic andfunctional recovery. The mechanisms involved appear to involve Hsp72 upregulation and NF-�Binhibition. (Surgery 2007;141:667-77.)

From the Department of Surgerya and Rheumatology and Inflammation Researchb and Institute for Neurosciences

and Physiology,c Sahlgrenska Academy at Göteborg University, Sweden

Despite advances in organ preservation andsurgical techniques, advanced ischemic injuriesstill occur in the transplanted organs; these injuries

Supported by funds from Professor Lars Erik Gelin MemorialFoundation and Sahlgrenska University Hospital, Göteborg,Sweden (LUA).

Accepted for publication November 13, 2006.

Reprint requests: Mihai Oltean, MD, Sahlgrenska UniversityHospital, Department of Surgery and Transplantation, 431 45Göteborg, Sweden. E-mail: mihai.oltean@wlab.gu.se.

0039-6060/$ - see front matter

© 2007 Mosby, Inc. All rights reserved.

doi:10.1016/j.surg.2006.11.005

may lead to graft dysfunction or nonfunction andincreased graft immunogenicity, and may promotelate changes in the graft.1-3 Among transplantableorgans, the intestine has the highest susceptibilityto ischemic injury and, unlike in other organs,intestinal ischemia-reperfusion (IR) may be fol-lowed by bacterial translocation, endotoxemia, andmassive release of proinflammatory mediators sec-ondary to the impairment of the mucosal barrier.4

Initiating protective mechanisms before the on-set of ischemia (preconditioning) by brief cycles ofischemia-reperfusion or by various chemicals hasbeen shown to be effective in numerous experi-mental models of IR and transplantation and in

several clinical trials.5-7 Translation of these meth-

SURGERY 667

668 Oltean et al SurgeryMay 2007

ods into the clinical practice, however, faces nu-merous obstacles such as the narrow therapeuticwindow of some compounds and excessive toxicity,or practical impediments (complicated, long pro-cedures), whereas an ideal alternative would be ashort course of an established compound using asimple protocol. In addition, extrapolation of re-sults obtained in warm ischemia models into atransplantation setting faces two major differences.First, warm and cold IR injury have different mech-anisms8 and, second, pretreatment acts simulta-neously on leukocytes and other distant organswith significance in IR and therefore makes it dif-ficult to delineate the potential overlapping or di-vergent contributions of each to the overall results.Transplantation of organs from pretreated donorsdissociates the organ from the “pretreated milieu”and delineates only the intrinsic mechanisms oper-ating within the graft.

FK506, a macrolide with strong immunosuppres-sive activity, prevents activation of T-lymphocytes inresponse to antigenic or mitogenic stimulation. Inaddition to its immunosuppressive potential, thecompound reduces both the local and systemicconsequences of warm ischemia and subsequentreperfusion in various organs.9,10 The partly unrav-eled mechanisms seem both related and unrelatedto calcineurin.11-13 By altering the activity of nu-merous kinases and thus interfering with severalintracellular signaling pathways, FK506 may influ-ence various transcription factors including nu-clear factor �B (NF-�B), the hallmark transcriptionfactor of the innate immune response. Indeed, arecent study providing further insight into themechanisms of IL-1�–mediated NF-�B activationconfirmed the ability of FK506 to prevent NF-�Bactivation and modulate the endothelial inflamma-tory activation.14

The present study was conducted to establish ifpretreatment of the donor with a single, low doseof FK506 would reduce the preservation-reperfu-sion injury of the intestinal graft after transplanta-tion.

MATERIAL AND METHODSAnimals. Male Sprague-Dawley rats (190 to

250 g) were housed in the Experimental Biomedi-cine Department with 12-hour light/dark cycles,controlled temperature, and a pathogen-free envi-ronment. The rats received rat chow and water adlibitum. The study was approved by the local Ethi-cal Committee for animal experiments of GöteborgUniversity and followed the rules and regulationsoutlined by the National Institutes of Health and

the European Union.

Animal treatment. Food was redrawn from do-nors 6 hours before start of the experiment, butaccess to water was maintained. Recipients had un-restricted access to food and water. In the precon-ditioned animals, FK506 (kindly provided byFujisawa Europe, Munich, Germany) was injectedintravenously (0.3 mg/kg), while control animals(n � 30/group) received equivalent volumes ofsaline (154 mM NaCl). Six hours after treatment,intestinal graft harvesting was performed. The jeju-noileum from 5 untreated nonoperated animalsserved as reference for histologic and biochemicalanalyses.

Surgical procedures and sampling. Heterotopicintestinal transplantation was performed using pre-viously described techniques.15 The small-bowelgrafts were stored in ice-cold saline for 3 hours, andthen transplanted by end-to-side anastomoses be-tween the graft portal vein and superior mesentericartery (retaining an aortic patch) and the recipientinfrarenal vena cava and aorta, respectively. Prelim-inary experiments indicated that cold storage insaline for 3 hours generates a preservation injurysimilar to 6 to 8 hours of cold preservation inUniversity of Wisconsin solution (the average coldischemia in clinical setting) but without requiringflush-out before reperfusion.16,17 Graft ends wereexteriorized as stomas. Warm ischemia was keptbelow 25 minutes and was similar in both groups.Immediately after reperfusion, the animals re-ceived 4 ml of warm saline (37°C) intravenouslyand 4 ml saline intraperitoneally, and then theabdomen was temporary closed. Twenty minutesafter reperfusion, a 3-cm segment was sampledfrom the distal graft end, the wound was closed,and the animals were returned to their cages withfree access to food and water. At 6, 12, or 24 hoursafter reperfusion, animals were reanesthetized (n� 10/time point/group), graft segments were sam-pled and fixed in formalin or snap-frozen, and theanimal was killed.

Donor biochemistry. To investigate renal func-tion in donors, blood urea nitrogen (BUN) andserum creatinine were measured using a Modular Pautoanalyzer (Roche Diagnostics, Mannheim, Ger-many). FK506 level in blood was measured by en-zyme immunoassay (Abbott Scandinavia AB, Solna,Sweden) in donors at harvesting and in severalrecipients at 20 minutes and 6 hours postreperfu-sion.

Histology and morphometry. Five-micron paraf-fin sections were cut and stained with hematoxylin-eosin (H&E). Slides were examined blindly, and IRinjury was graded according to the Park/Chiu scor-

ing system18 (Table). Mucosal thickness (from the

Surgery Oltean et al 669Volume 141, Number 5

tip of the villi to the muscular layer) and cryptdepth (from the villus base to the muscular layer)were measured in 30 axially oriented villi on at least3 different sections. Villus height was considered asthe difference between mucosal thickness andcrypt depth. Morphometry was performed usingthe AxioCam image analyzer software (Zeiss, Jena,Germany).

Histochemical staining for polymorphonuclearneutrophil (PMN) infiltration. Intragraft PMNswere stained using the Naphtol AS-D chloroacetateesterase kit (Sigma Chemicals, St Louis, Mo). Stain-ing was carried out on paraffinized sections, ac-cording to manufacturers’ instructions (Sigma).The sections were counterstained with Meyers’ he-matoxylin. All the samples were coded, and posi-tively stained PMNs were identified and counted, ina blinded fashion, on at least 3 different sections in20 randomly selected fields at intermediate magni-fication (�200).

Immunohistochemical assessment of crypt cellproliferation. Ki-67 is a nonhistone nuclear proteinpresent in all phases but G0 (the resting state) ofthe cell cycle. The ratio between the positive andtotal number of cells is accepted generally as agood indicator of a tissue-proliferating activity.

Paraffin slides were deparaffinized and hydratedby successive soakings in xylene and ethanol, andantigen retrieval was performed by microwavingthe slides in 0.5 M Tris buffer (pH 10). The slideswere incubated with Ki67 antibody (Novocastra,Newcastle/Tyne, UK) in a wet chamber at 37°C,according to manufacturers’ instructions. After theapplication of the second antibody and color de-velopment, slides were counterstained with Meyers’hematoxylin and the Ki67-positive nuclei werecounted blindly in 15 randomly selected crypts sec-tioned transversally in 3 different sections under

Table. The Park/Chiu scoring system forintestinal ischemia reperfusion injury

Grade Description

0 Normal mucosa1 Subepithelial space (bleb) at villus tip2 More extended subepithelial space (upper

half of the villi)3 Epithelial lifting down the sides of the villus,

occasional epithelial breakdown4 Denuded villi5 Loss of villous tissue6 Crypt layer destruction7 Transmucosal infarction8 Transmural infarction

high magnification (�400). Proliferation rate was

defined as the ratio between the number of Ki67-positive nuclei and the total number of the cryptcells.

Western blot analysis for Hsp72 and ICAM-1.Heat shock protein (Hsp) 72 is an inducible stressprotein with essential roles in cytoprotection,housekeeping, and protein folding. Frozen intesti-nal tissue was homogenized in ice-cold lysis buffer(0.2 M Tris-HCl, 0.1 M NaCl, 0.1 M EDTA, 0.5 MDMSF, 1% Triton X-100, pH 7.4) containing pro-tease inhibitors and centrifuged at 10,000g for 10minutes at 4°C. Protein concentration in the super-natant was measured, and then proteins were sep-arated by sodium dodecyl-sulphate-polyacrylamidegel electrophoresis (SDS-PAGE) with 30 �g of totalprotein loaded per lane. In 1 lane, recombinantHsp72 (SPP-758E; Stressgen, Victoria, BC) servedas control. After SDS-PAGE, proteins were trans-ferred onto polyvinylidene fluoride membranes(Bio-Rad Laboratories, Hercules, Calif). The mem-branes were blocked (5% skim milk in PBS con-taining 0.05% Tween-20) and incubated for 1 hourat room temperature with anti–Hsp72 antibody (1:3000, SPA-812; StressGen). A corresponding anti-body was added, and then blots were developed bythe use of the Advanced ECL kit (Amersham Bio-sciences Ltd, Buckinghamshire, UK).

Intercellular adhesion molecule (ICAM)-1 is anadhesion molecule essential in neutrophil trans-capillary migration, which increases after IR and isregulated by NF-�B. For ICAM-1 determination,protein extraction, concentration measurements,and blotting were performed as described above.Membranes were incubated with a mouse monoclo-nal antibody against ICAM-1 (1:3000, MCA773; Se-rotec, Oxford, UK) and an adequate secondaryantibody. After development, the bands from im-munoblots were quantified using computerizeddensitometry (Quantity One; Bio-Rad). Resultswere normalized to �-tubulin and reported as op-tical density (OD) units.

Electrophoretic mobility shift assay (EMSA). Ac-tivated transcription factors translocate to the nu-cleus and bind to specific DNA sequences where itinitiates gene transcription. The complex DNAtranscription factor may be sought indirectly byusing EMSA.

Frozen intestinal tissue was homogenized gentlyin 2 ml of hypotonic buffer at 4°C. After centrifu-gation at 4°C 14000g for 10 minutes, the superna-tant was removed carefully; the pellet wasresuspended in ice-cold extraction buffer and ex-tracted 2 hours at 4°C on a rotator. Cell particleswere sedimented at 14000g at 4°C for 1 hour and

the supernatant was collected, aliquoted, and

670 Oltean et al SurgeryMay 2007

stored at �80°C until the analysis of NF-�B wasperformed. Before freezing, the protein concentra-tion was determined using the Bradford method.

EMSA was performed as described elsewhere19

with minor modifications. Briefly, the double-stranded consensus oligonucleotide (sc-2505; SantaCruz Biotechnology, San Diego, Calif) was labeledwith 32P (Amersham Pharmacia Biotech, Uppsala,Sweden) using T4-polynucleotide kinase (New En-gland Biolabs, Ipswich, Mass) and used in the bind-ing reaction with the nuclear extracts. The bindingreaction was carried out in a volume of 20 �l,containing equal amounts of nuclear protein(6.3 �g), 5 �g of poly(dIdC) poly(dI-dC) (Amer-sham Pharmacia Biotech), 1 ng (80,000 cpm) ofthe 32P-labeled oligonucleotide containing theNF-�B binding site, 1 mM DTT, 0.2 mg/ml bovineserum albumin, and binding buffer. In some bind-ing reactions, 1 � l of an antibody against the p65subunit of NF-�B (Santa Cruz Biotechnology) wasadded and incubated for another 20 minutes. Thesame mixture using unlabelled oligonucleotideswere used as negative control. The protein-DNAcomplexes were resolved on a native 5% polyacryl-amide gel containing 0.25� TRIS-borate-EDTAbuffer. The gel was vacuum-dried and exposed tox-ray film for 24 to 36 hours at �80°C. Band inten-sity was analyzed using computerized densitometry.

Intestinal disaccharidase activity. Intestinal dis-accharidase activity is an established marker of in-testinal functional integrity. The activity of maltaseand sucrase, 2 dissacharidases present on the en-terocyte brush border, was determined in wholetissue homogenates. Whole tissue homogenateswere incubated with substrate dissacharides (2%maltose or 2% sucrose in 0.1 M sodium maleatebuffer, pH 6.0) for 1 hour at 37°C, as describedpreviously.20 The reaction was stopped by submerg-ing the samples in boiling water for 5 minutes. Oneunit of dissacharidase was defined as the quantity ofenzyme hydrolyzing 1 �mol of substrate in 1minute. The resulting glucose was measured usingthe automated hexokinase assay. The experimentswere performed in triplicate, and the amount ofprotein in the homogenate was measured using theBio-Rad method. Results were expressed as activityunits/gram protein.

Statistical analysis. Nonparametric methodswere used for statistical comparisons. Statistical dif-ferences between independent groups were calcu-lated using the Kruskal- Wallis test followed by theMann-Whitney U test and the Fisher exact proba-bility test. Data are reported as means � standarderror of the mean (SEM). A P value � .05 was

considered significant.

RESULTSDonor biochemistry. In FK506-treated rats and

in untreated rats, the serum creatinine was 13.5 �3.53 �mol/L and 12.5 � 3.3, respectively (P � notsignificant [ns]), indicating that pretreatment didnot impair donor renal function. Measured BUNwas 3.06 � 0.64 �mol/L in FK506-treated rats and3.15 � 1.01 mmol/L in untreated rats. The level ofFK506 was 6.94 � 0.2 �g/L in the whole blood ofpretreated donors at harvesting. The FK506 wasundetectable in recipients (�3 ng/mL) at both 20minutes and 6 hours postreperfusion.

Histology and morphometry. Cold preservationinduced a similar ischemic lesion in both groups(2.4 � 0.1 in control vs 2.2 � 0.1 in treated grafts;ns). All grafts showed a moderate lifting of theepithelium, which sometimes extended down onthe sides of the villi. Nonetheless, epithelial conti-nuity was maintained (Fig 1, A, F). At 20 minutesafter reperfusion, pretreated grafts had an im-proved structure compared with grafts from un-treated animals (4.1 � 0.1 vs 5 � 0.1; P � .001).Grafts from untreated rats displayed extensive den-udations and disintegration of the lamina propriadown to the villus base (grade 5). Focal crypt de-struction was also present. In pretreated grafts,most of the villi were denuded but maintained anintact lamina propria (grade 4) (Fig 1, B).

At 6 hours postreperfusion, microscopic exami-nation of pretreated grafts showed near-normalvilli, completely covered with epithelium, andlonger than those of control grafts (329 � 68 vs240 � 60 �m; P � .01) (Fig 1, C). Control graftsalso revealed scattered subepithelial blebs andbleeding, ulcerations, and inflammation (Fig 1, G).At 12 hours after reperfusion, a significant differ-ence in mucosal appearance persisted between thegroups. At 24 hours postreperfusion, the differ-ences between the groups virtually subsided(Fig 1 D, H).

Enterocyte proliferation. The percentage ofKi67-positive cells in normal animals was 46.5% �4%. At 6 hours postreperfusion, pretreated graftsrevealed an increased proliferation rate comparedwith the control grafts (60.3 � 9.9 vs 41.8 � 7.9;P � .001). The difference between groups persistedand increased over the next 6 hours (72 � 9.6 vs44.9 � 8; P � .001). A proliferative response alsooccurred in the control grafts 24 hours postreper-fusion, and the difference between the 2 groupsdisappeared (62.2 � 4.14 vs 61.7 � 6.2; P � ns).Overall, the results indicated that significantlymore crypt cells were leaving the resting state (G0)

and entered cell division over the first 12 hours in

Surgery Oltean et al 671Volume 141, Number 5

the pretreated group (Fig 2). Many of the existingKi-67 positive crypt cells in the control grafts mayhave been arrested initially between the G1 and G2checkpoints of the cellular cycle, but they ulti-mately resumed division and partly regained thedelayed restoration in comparison with the pre-treated grafts.

Analysis of Hsp72 and ICAM-1 expression usingWestern blot analysis. The analysis of Hsp72 ex-pression in donor livers and kidneys obtained atthe time of graft harvesting showed a modest(140% � 30%) increase in the livers of treatedanimals compared with the untreated controls.

Fig 1. Preservation, early and late reperfusion injury inshowed variable degrees of epithelial detachment (supostreperfusion pretreated grafts (B) revealed extensivwhereas complete disappearance of the villi was seen in cpretreated grafts had slim villi and continuous epithelialvilli and many denuded areas. At 24 hours after reperwhereas control grafts had markedly improved structurepreservation injury (open symbols) and early reperfusion(squares), according to the Chiu/Park scale. The morphvillus length (J) and total mucosal thickness (K) were signcontrol grafts at both 6 hours and 12 hours postreperfusinormal values (black bar). Bar represents 100 �m; originacontrol; **P � .01 vs control.

Hsp72 expression in the kidney increased to

240% � 70% of that of controls. The upregulationwas similar between the left and right kidney ofeach animal, suggesting that the pretreatment in-duced a systemic upregulation of Hsp72.

Faint Hsp72 expression was found in the smallintestines of untreated animals. At 6 hours after theFK506 administration, Hsp72 expression was 240%increased in grafts from pretreated donors com-pared to reference animals, whereas control ani-mals receiving saline showed no increase in Hsp72.Hsp72 expression remained at similar levels afterpreservation and early postreperfusion. At 6 hoursposttransplantation, Hsp72 expression increased

testinal grafts. After preservation, pretreated grafts (A)elial edema), similar to the control grafts (B). Earlyelial denudations but maintained the lamina propria,grafts (F) (original magnification �200). After 6 hours,

age (C), whereas control grafts (G) had short, flattened, pretreated grafts revealed near-normal histology (D),me deformed villi (H). The diagram (I) summarizes the(closed symbols) in control (circles) and pretreated graftsic analysis of the intestinal mucosa indicated that bothtly superior in pretreated grafts (gray bar) compared with24 hours postreperfusion, both groups approached the

nification �100, unless otherwise indicated. *P � .05 vs

the inbepithe epithontrolcover

fusionand soinjuryometrificanon. Byl mag

more than 4-fold in the pretreated grafts and 3-fold

672 Oltean et al SurgeryMay 2007

in control grafts and remained increased through-out the study (Fig 3, A).

ICAM-1 is a major adhesion molecule whosetranscription is regulated by NF-�B. The endothe-lial ICAM-1 is used by the neutrophils to transmi-grate in the extravascular tissue. ICAM-1 wasundetectable in normal animals, as well as in con-trol and pretreated grafts before and after preser-vation and 20 minutes after reperfusion.

At 6 hours postreperfusion, immunoblots fromcontrol donors revealed increased expressions ofICAM-1 compared with the previous time points,whereas in grafts from pretreated donors, ICAM-1expression remained very low. At 12 hours and 24hours after reperfusion, ICAM-1 expression contin-ued to increase in control but not in pretreatedgrafts. In the latter, ICAM-1 expression returned tobaseline levels (Fig 3, B).

Graft PMN infiltration. After reperfusion, thenumber of infiltrating PMNs increased rapidly ingrafts from both groups. The highest tissue infiltra-tion by PMNs was found at 6 hours, with clusters ofPMNs found equally in the lamina propria andamong crypts. At the later time points, PMN countdecreased and infiltrating neutrophils were foundmostly in the lower half of the villi. Neutrophilrecruitment was less at all time points in the graftsfrom pretreated donors compared with controls(P � .05) (Fig 4).

Electrophoretic mobility shift assays. Weak con-stitutive DNA-binding activity of the NF-�B com-plex was seen in the reference animals, probablydue to the continuous gut stimulation by the con-tact with the intraluminal content (Fig 5). Coldischemia induced a modest NF-�B activation (nu-clear translocation of the p65/p50 heterodimer).

Fig 2. Proliferation rate (the ratio between Ki67-positivecrypt cells and the total number of crypt cells) in thecrypts of pretreated grafts (gray bars), control grafts (whitebars), and normal animals (black bar). #P � .001 vs con-trols; **P � .01 vs 6 hours.

NF-�B nuclear translocation and activation was in-

creased immediately after reperfusion in the con-trol grafts, a feature reduced in the pretreatedgrafts. DNA-binding activity, which remained highin the nuclear extracts from control grafts, had asecond peak of activation 12 hours after reperfu-sion whereas in pretreated grafts, the DNA-bindingactivity remained low and similar to the referenceanimals. As revealed by the supershift analysis, theNF-�B complex contained p65 subunits, indicatingthat the complexes detected by EMSA correspondto the p65/p50 heterodimer (data not shown). Insummary, pretreated grafts had a greatly reducedNF-�B activation compared with controls.

Intestinal disaccharidase activity. Pretreatmentdid not influence intestinal dissacharidase activity(data not shown). In grafts of either groups, theactivity of both enzymes decreased slightly afterpreservation and decreased further 20 minutes af-

Fig 3. A, Analysis of heat-shock protein 72 (Hsp72) ex-pression in the intestinal grafts using Western blot incontrol (white bars) and pretreated grafts (gray bars);Hsp72 was upregulated in the small intestine of pre-treated donors at harvesting, after cold preservation, andearly postreperfusion compared with the saline-treatedcontrols or normal animals (not shown). B, Western blotanalysis of intercellular adhesion molecule (ICAM)-1 ex-pression in the intestinal grafts. ICAM-1 was undetectablein pretreated (gray bars) and control grafts (white bars) atharvesting, after cold preservation, and early postreper-fusion, but increased in control grafts with a peak at 12hours postreperfusion. *P � .05 vs controls; **P � .01 vscontrols.

ter reperfusion (P � ns). At 6 hours after reperfu-

controls; **P � .01 vs controls.

Surgery Oltean et al 673Volume 141, Number 5

sion, both maltase and sucrase activity in theintestinal grafts were decreased to the same extent,irrespective of groups. Over the following hours,the dissacharidase activity evolved gradually towardrecovery; however, whereas pretreated graftsachieved faster and basically complete restorationof maltase activity by 24 hours postreperfusion,both dissacharidases in control grafts remainedlower and below the normal values (Fig 6). Overall,the functional recovery appeared accelerated inthe pretreated grafts with significant differencesas early as 12 hours after reperfusion. At 24 hourspostreperfusion in many of the pretreated grafts,maltase activity exceeded the values found in ref-erence animal, most likely due to the lack of the

Fig 4. Graft PMN infiltration at various time points afterreperfusion as assessed by naphtol AS-D chloroacetatehistochemical staining. Microphotographs of pretreatedgrafts (upper panel) and control grafts (lower panel) (orig-inal magnification �200). The number of infiltratingneutrophils (purple) was significantly less in the pre-treated grafts (closed bars) at all time points comparedwith the control grafts (open bars). The average PMNfrequency in reference animals was 6 � 3 cells/field (notshown). *P � .05 vs controls.

luminal pancreatic proteases, involved in the deg-

Fig 5. Electrophoretic mobility shift assay (EMSA) show-ing nuclear factor �B (NF-�B) activation in the intestinalgrafts in FK506 pretreated and control animals. Nuclearextracts of intestines from normal animals, FK506-pre-treated animals (indicated with a plus sign [�]), andcontrol animals (indicated with a minus sign [�]) atdifferent time points were prepared as described in Ma-terial and Methods, incubated with radioactive 32P-la-beled NF-�B specific oligonucleotides, and submitted toelectrophoresis. NF-�B DNA-binding activity in the nu-clear extracts of intestines from normal animals is shownin lane 1. After preservation (P), NF-�B had a modestincrease in DNA-binding activity in control grafts(lane 2), whereas no NF-�B activation was seen in FK506pretreated grafts (lane 3). At 20 minutes postreperfusion,an increase in NF-�B activation was recorded in controlgrafts (lane 4), and DNA binding activity remained highup to 24 hours (lanes 6, 8, and 10). In contrast, inFK506-pretreated grafts, the NF-�B activity in the earlypostreperfusion period was minimal (lane 5), and re-mained low throughout the study period (lanes 7, 9and 11). The arrow indicates the NF-�B band. The sam-ples were obtained from 3 different animals at each timepoint, and the experiment was repeated with similarresults. One representative experiment out of three isillustrated here. B, The densitometric analysis of theNF-�B electrophoretic band intensity. Data are mean �standard error of the mean (SEM) of optical density(OD) units obtained in 3 different experiments at eachtime point. The results show 2- to 3-fold increase inNF-�B activity in control grafts compared to FK506-treated animals. OD, Optical density units. #P � .001 vs

674 Oltean et al SurgeryMay 2007

radation and turnover of the brush-border en-zymes.

DISCUSSIONThe present study suggests that donor pretreat-

ment with a low-dose of FK506 inhibits intragraftNF-�B activation, reduces endothelial activationand neutrophil infiltration, and leads to an accel-erated morphologic and functional recovery com-pared with control grafts.

Extensive experimental evidence underscoredthe efficiency of preemptive anti-ischemic interven-tions in transplantation, while preconditioningstrategies have recently attracted great interest.8,21

From the myriad of compounds reducing warmischemic injury, only a few were tested ultimately intransplantation, with Hsp32 (heme oxygenase-1)and its byproducts and the nitric oxide donorsadvocated most recently advocated.22-24 The major-ity were deemed unsuitable because of short-livedeffects, elusive mechanisms, or toxicity.

Many studies have reported reductions of thepostischemic damage after induction of Hsp72, amajor cytoprotective and anti-apoptotic chaper-one.25-27 Nevertheless, the methods of Hsp72 in-duction were often impractical, thus limiting an

Fig 6. Intestinal brush-border dissacharidase activity inthe intestinal grafts at different time points. Maltase (A)and sucrase (B) activity were affected marginally by thecold storage and early after reperfusion, but decreasedsignificantly 6 hours after reperfusion. Enzyme activityrecovered faster in the pretreated grafts (gray bars) thanin control grafts (white bars). The black bars represent theactivity in normal animals. *P � .05 vs controls.

ultimate clinical use. In line with a previous report,

the current study found increased expression ofHsp72 after FK506 treatment,28 and the currentresults suggest that sizeable Hsp72 increases may beobtained specifically after FK506 treatment, notonly in intestine but also in other organs.

Reperfusion upregulated Hsp72 in controls,seemingly as part of the physiologic repair mecha-nism, whereas its late upregulation presumably as-sisted the morphologic restoration. Its timely,preemptive induction in pretreated grafts ap-peared also able to lessen the reperfusion injury.

Immediately after reperfusion, oxidative stresscauses local tissue damage and activates several keyproinflammatory transcription factors, such as AP-1and NF-�B.29-31 Their transcription products fur-ther enhance the local leukocyte recruitment, ini-tially prompted by cell debris and locally releasedmediators.32,33 Although the current study did notinvestigate oxidative stress, previous work sug-gested that FK506 can decrease free radical dam-age.13,34 Pretreatment clearly influenced anotherimportant mechanism triggered by oxidative stressand was certainly involved in the improvementsrecorded in the pretreated grafts, namely NF-�Bactivation. NF-�B activation normally occurs withinminutes after reperfusion and has a biphasic pat-tern. The first peak of activation consists of theactive heterodimer p65/p50 conferring transcrip-tional activity, whereas the second consists of p50/p50 homodimers (lacking the transcriptionaldomain) that are thought to competitively bindDNA, thus contributing to the resolution of inflam-mation.35 Pretreated grafts always had greatly re-duced NF-�B activation as shown by differences inDNA-binding activity and ICAM-1 (an NF-�B tran-scription product) expression and further reflectedin different neutrophil sequestration.

In addition to blunting the endothelial activa-tion, NF-�B inhibition may affect other proinflam-matory mediators synthesized locally and releasedon reperfusion, such as IL-1, IL-6, TNF-�, andiNOS, because several studies indicated that thepostischemic intestine produces a multitude ofproinflammatory mediators.4,36

Several reports suggested that heat shock pre-conditioning can reduce NF-�B activation.37,38 Themechanism of NF-�B modulation remains unclear,but it has been proposed that Hsp72 binds theNF-�B/I�B complex and prevents I�B conforma-tional changes and dissociation from NF-�B. Be-cause increased cytoplasmic I�B expression acts asan inhibitor of NF-�B activation, another hypothe-sis implies that Hsp72 prevents the ubiquitinationand degradation of free I�B after its phosphoryla-

tion by IKK and increasing its cytosolic concentra-

Surgery Oltean et al 675Volume 141, Number 5

tion.39,40 Our study, however, found significantNF-�B activation in control grafts in the presenceof Hsp72 levels far exceeding those found in pre-treated grafts at time points where NF-�B activationwas absent. Thus, the current results do not sup-port an inhibitory role for Hsp72 in NF-�B nucleartranslocation and activation.

Increased leukocyte-endothelial interactions werecommonly reported during reperfusion and wereconsidered key events for the microvascular perfusionfailure and neutrophil infiltration41,42; indeed, nu-merous strategies of inhibiting leukocyte-endothelialinteractions improved morphology and survival.43,44

After endothelial adhesion, leucocytes migrate acrosscapillaries into the injured tissue and may cause fur-ther tissue damage through enzyme secretion, ox-idative stress in the lamina propria, and additionalproinflammatory activation.32,42 Similar to otherstudies, the current study shows less tissue neutro-phil sequestration with reduced endothelial activa-tion.

Homeostasis of the intestinal mucosa is a tightequilibrium between enterocyte production andloss, as well as continuous epithelial cell interac-tions with the underlying lamina propria. Our re-sults suggest that more crypt cells in the pretreatedgrafts could pass beyond the G1-checkpoint of thecell cycle, the key step in cell division. The transi-tion from the resting state (G0) toward the G1 andS phase is dependent on growth and proliferationfactors, as well as a favorable cellular microenviron-ment. In the presence of unfavorable circum-stances (ie, cellular stress), the cells remainquiescent, and proliferation will be blocked or de-layed. There may be several, yet unidentified mech-anisms underlying our observations, but it isreasonable to assume that increased Hsp72 allowedcrypt stem cells to recover the reperfusion stressand resume proliferation faster. Considering theseobservations that a single crypt stem cell may giverise to 60 to 120 enterocytes,45 modest changesin the rate of crypt proliferation or apoptosis mayresult in dramatic changes in villus morphologythat may explain both the morphologic and funct-ional differences between the groups. We suggestthat increased intracellular Hsp72 and the dimin-ished neutrophil infiltrate in the neighboring lam-ina propria of pretreated grafts provided a morefavorable microenvironment for the enterocytes toproliferate, migrate, and maturate.

Numerous observations suggest that several daysare needed for the human intestine to recoverreperfusion injury after transplantation. While thisperiod overlaps with the induction therapy, initia-

tion of immunosuppression and frequent episodes

of hemodynamic instability, the addition of an in-fectious risk factor due to the breakdown of theintestinal barrier may have dramatic consequences.In addition, the current findings may also havebroader implications in other circumstances evolv-ing with intestinal ischemia, such as during opera-tions on the supraceliac aorta. Indeed, besides thegeneralized induction of a protective heat shockresponse in several abdominal organs, the prospectof modulating the NF-�B–driven inflammation,which is a major cause for the postoperative mor-tality in patients undergoing aortic surgery, seemsparticularly interesting.46,47,12

In summary, the current study suggests that do-nor pretreatment with a single, low dose of FK506reduces reperfusion injury and accelerates intesti-nal graft morphologic and functional recovery bymounting a systemic heat-shock response and alle-viating the NF-�B–dependent inflammation.

The authors gratefully acknowledge Associate Profes-sor Bertil Ohlsson for expert advice and comments onthe manuscript.

REFERENCES1. Ploeg RJ, D’Alessandro AM, Knechtle SJ, Stegall MD, Pirsch

JD, Hoffmann RM, et al. Risk factors for primary dysfunc-tion after liver transplantation—a multivariate analysis.Transplantation 1993;55:807-13.

2. Goes N, Urmson J, Ramassar V, Halloran PF. Ischemic acutetubular necrosis induces an extensive local cytokine re-sponse. Evidence for induction of interferon-gamma, trans-forming growth factor-beta 1, granulocyte-macrophagecolony-stimulating factor, interleukin-2, and interleukin-10.Transplantation 1995;59:565-72.

3. Tullius SG, Heemann U, Hancock WW, Azuma H, TilneyNL. Long-term kidney isografts develop functional and mor-phologic changes that mimic those of chronic allograftrejection. Ann Surg 1994;220:425-32.

4. Grotz MR, Deitch EA, Ding J, Xu D, Huang Q, Regel G.Intestinal cytokine response after gut ischemia: role of gutbarrier failure. Ann Surg 1999;229:478-86.

5. Tullius SG, Nieminen-Kelha M, Buelow R, Reutzel-Selke A,Martins PN, Pratschke J, et al. Inhibition of ischemia/reperfu-sion injury and chronic graft deterioration by a single-donortreatment with cobalt-protoporphyrin for the induction ofheme oxygenase-1. Transplantation 2002;74:591-8.

6. Azoulay D, Del Gaudio M, Andreani P, Ichai P, Sebag M,Adam R, et al. Effects of 10 minutes of ischemic precondi-tioning of the cadaveric liver on the graft’s preservation andfunction: the ying and the yang. Ann Surg 2005;242:133-9.

7. Redaelli CA, Wagner M, Kulli C, Tian YH, Kubulus D,Mazzucchelli L, et al. Hyperthermia-induced HSP expres-sion correlates with improved rat renal isograft viability andsurvival in kidneys harvested from non-heart-beating do-nors. Transpl Int 2001;4:351-60.

8. Selzner N, Rudiger H, Graf R, Clavien PA. Protective strat-egies against ischemic injury of the liver. Gastroenterology

2003;125:917-36.

676 Oltean et al SurgeryMay 2007

9. Sakr M, Zetti G, McClain C, Gavaler J, Nalesnik M, Todo S,et al. The protective effect of FK506 pretreatment againstrenal ischemia/reperfusion injury in rats. Transplantation1992;53:987-91.

10. Squadrito F, Altavilla D, Squadrito G, Saitta A, Deodato B,Arlotta M, et al. Tacrolimus limits polymorphonuclear leuco-cyte accumulation and protects against myocardial ischaemia-reperfusion injury. J Mol Cell Cardiol 2000;32:429-40.

11. St Peter SD, Moss AA, Mulligan DC. Effects of tacrolimus onischemia-reperfusion injury. Liver Transpl 2003;9:105-16.

12. Klettner A, Baumgrass R, Zhang Y, Fischer G, Burger E,Herdegen T, et al. The neuroprotective actions of FK506binding protein ligands: neuronal survival is triggered byde novo RNA synthesis, but is independent of inhibitionof JNK and calcineurin. Brain Res Mol Brain Res 2001;97:21-31.

13. Garcia-Criado FJ, Palma-Vargas JM, Valdunciel-Garcia JJ,Toledo AH, Misawa K, et al. Tacrolimus (FK506) down-regulates free radical tissue levels, serum cytokines, andneutrophil infiltration after severe liver ischemia. Trans-plantation 1997;64:594-8.

14. Brunn G, Saadi S, Platt J. Differential regulation of endothelialcell activation by complement and interleukin 1 alpha. CircRes 2006;98:793-800.

15. Kellersmann R, Kiyochi H, Black R, Grant D, Zhong R.Microsurgical techniques for small bowel transplantation inrats. In: Timmermann W, Gassel HJ, Ulrichs K, Zhong R,Thiede A, editors. Organ transplantation in rats and mice.Berlin Heidelberg: Springer, 1998:83.

16. Olofsson R, Oltean M, Olausson M. Intestinal preservationinjury: a comparison between different rat strains. Trans-plant Proc 2006;38:1789-91.

17. deRoover A, de Leval L, Gilmaire J, Detry O, Boniver J,Honore P, Meurisse M. A new model for human intestinalpreservation: comparison of University of Wisconsin and Cel-sior preservation solutions. Transplant Proc 2004;36:270-2.

18. Park PO, Haglund U, Bulkley GB, Falt K. The sequence ofdevelopment of intestinal tissue injury after strangulationischemia and reperfusion. Surgery 1990;107:574-80.

19. Bergquist J, Ohlsson B, Tarkowski A. Nuclear factor-kappa B is involved in the catecholaminergic suppressionof immunocompetent cells. Ann N Y Acad Sci 2000;917:281-9.

20. Dahlqvist A. Assay of intestinal dissacharidases. Scand J ClinLab Invest 1984;44:169-72.

21. Mallick I, Yang W, Winslet M, Seifalian A. Ischemia–Reperfusion injury of the intestine and protective strategiesagainst injury. Dig Dis Sci 2004;49:1359-77.

22. Man K, Lee TK, Liang TB, Lo CM, Fung PC, Tsui SH, et al.FK 409 ameliorates small-for-size liver graft injury by atten-uation of portal hypertension and down-regulation of Egr-1pathway. Ann Surg 2004;240:159-68.

23. Nakao A, Neto JS, Kanno S, Stolz DB, Kimizuka K, Liu F,et al. Protection against ischemia/reperfusion injury incardiac and renal transplantation with carbon monoxide,biliverdin and both. Am J Transplant 2005;5:282-91.

24. Nakao A, Otterbein LE, Overhaus M, Sarady JK, Tsung A,Kimizuka K, et al. Biliverdin protects the functional integrityof a transplanted syngeneic small bowel. Gastroenterology2004;127:595-606.

25. Jayakumar J, Suzuki K, Sammut IA, Smolenski RT, Khan M,Latif N, et al. Heat shock protein 70 gene transfectionprotects mitochondrial and ventricular function against

ischemia-reperfusion injury. Circulation 2001;104:I303-7.

26. Fleming SD, Starnes BW, Kiang JG, Stojadinovic A, TsokosGC, Shea-Donohue T. Heat stress protection against mes-enteric I/R-induced alterations in intestinal mucosa in rats.J Appl Physiol 2002;92:2600-7.

27. Kelly KJ, Baird NR, Greene AL. Induction of stress responseproteins and experimental renal ischemia/reperfusion. Kid-ney Int 2001;59:1798-802.

28. Yang CW, Ahn HJ, Han HJ, Kim WY, Li C, Shin MJ, et al.Pharmacological preconditioning with low-dose cyclospor-ine or FK506 reduces subsequent ischemia/reperfusion in-jury in rat kidney. Transplantation 2001;72:1753-9.

29. Parks DA, Bulkley GB, Granger DN. Role of oxygen freeradicals in shock, ischemia, and organ preservation. Surgery1983;94:428-32.

30. Parks DA, Bulkley GB, Granger DN, Hamilton SR, McCordJM. Ischemic injury in the cat small intestine: role of super-oxide radicals. Gastroenterology 1982;82:9-15.

31. Yeh K, Yeh M, Glass J, Granger DN. Rapid activation ofNF-kappaB and AP-1 and target gene expression in postisch-emic rat intestine. Gastroenterology 2000;118:525-34.

32. Anaya-Prado R, Toledo-Pereyra LH, Lentsch AB, Ward PA.Ischemia/reperfusion injury. J Surg Res 2002;105:248-58.

33. Jaeschke H, Smith C. Cell adhesion and migration. IIILeukocyte adhesion and transmigration in the liver vas-culature. Am J Physiol Gastrointest Liver Physiol 1997;273:G1169-73.

34. Kawano K, Kim YI, Goto S, Ono M, Kobayashi M. A protec-tive effect of FK506 in ischemically injured rat livers. Trans-plantation 1991;52:143-5.

35. Zou L, Attuwaybi B, Kone BC. Effects of NF-kappa B inhi-bition on mesenteric ischemia-reperfusion injury. Am JPhysiol Gastrointest Liver Physiol 2003;284:G713-21.

36. Tamion F, Richard V, Lyoumi S, Daveau M, Bonmarchand G,Leroy J, et al. Gut ischemia and mesenteric synthesis of inflam-matory cytokines after hemorrhagic or endotoxic shock. Am JPhysiol Gastrointest Liver Physiol 1997;273:G314-21.

37. Uchinami H, Yamamoto Y, Kume M, Yonezawa K, Ishikawa Y,Taura K, et al. Effect of heat shock preconditioning on NF-kappaB/I-kappaB pathway during I/R injury of the rat liver.Am J Physiol Gastrointest Liver Physiol 2002;282:G962-71.

38. Kume M, Yamamoto Y, Yamagami K, Ishikawa Y, Uchinami H,Yamaoka Y. Pharmacological hepatic preconditioning: involve-ment of 70-kDa heat shock proteins (HSP72 and HSP73) inischaemic tolerance after intravenous administration of doxo-rubicin. Br J Surg 2000;87:1168-75.

39. Shanley TP, Ryan MA, Eaves-Pyles T, Wong HR. Heat shockinhibits phosphorylation of I-kappaBalpha. Shock 2000;14:447.

40. Yoo CG, Lee S, Lee CT, Kim YW, Han SK, Shim YS. Anti-inflammatory effect of heat shock protein induction is re-lated to stabilization of I kappa B alpha through preventingI kappa B kinase activation in respiratory epithelial cells.J Immunol 2000;164:5416-23.

41. Beuk RJ, Heineman E, Tangelder GJ, Quaedackers JS,Marks WH, Lieberman JM, et al. Total warm ischemia andreperfusion impairs flow in all rat gut layers but increasesleukocyte-vessel wall interactions in the submucosa only.Ann Surg 2000;231:96-104.

42. Hayward R, Lefer AM. Time course of endothelial-neutro-phil interaction in splanchnic artery ischemia-reperfusion.Am J Physiol 1998;275:H2080-6.

43. Farmer DG, Shen XD, Amersi F, Anselmo D, Ma JP, Ke B,et al. CD62 blockade with P-Selectin glycoprotein ligand-immu-noglobulin fusion protein reduces ischemia-reperfusion injuryafter rat intestinal transplantation. Transplantation 2005;

79:44-51.

Surgery Oltean et al 677Volume 141, Number 5

44. Rentsch M, Post S, Palma P, Lang G, Menger MD, MessmerK. Anti-ICAM-1 blockade reduces postsinusoidal WBC ad-herence following cold ischemia and reperfusion, but doesnot improve early graft function in rat liver transplantation.J Hepatol 2000;32:821-8.

45. Potten C, Wilson J, Booth C. Regulation and significance ofapoptosis in the stem cells of the gastrointestinal epithe-

lium. Stem Cells 1997;15:82-93.

46. Welborn M, Oldenburg H, Hess P, Huber T, Martin T, RauwerdaJ, et al. The relationship between visceral ischemia, proinflamma-tory cytokines, and organ injury in patients undergoing thoraco-abdominal aortic aneurysm repair. Crit Care Med 2000;28:3191-7.

47. Norwood M, Bown M, Lloyd G, Bell P, Sayers R. The clinicalvalue of the systemic inflammatory response syndrome(SIRS) in abdominal aortic aneurysm repair. Eur J Vasc

Endovasc Surg 2004;27:292-8.