thermal facilitation of lymphocyte trafficking involves temporal induction of intravascular icam-1

Thermal Facilitation of Lymphocyte Trafficking InvolvesTemporal Induction of Intravascular ICAM-1

QING CHEN,* MICHELLE M. APPENHEIMER,* JASON B. MUHITCH,*DANIEL T. FISHER,* KRISTEN A. CLANCY,* JEFFERY C. MIECZNIKOWSKI,$

WAN-CHAO WANG,* AND SHARON S. EVANS*Department of *Immunology; $Biostatistics, Roswell Park Cancer Institute, Buffalo, New York, USA

ABSTRACT

Objective: Fever is associated with improved survival, although its beneficial mechanisms are poorlyunderstood. Previous studies indicate that the thermal element of fever augments lymphocytemigration across high endothelial venules (HEVs) of lymphoid organs by increasing the intravasculardisplay of a gatekeeper trafficking molecule, intercellular adhesion molecule-1 (ICAM-1). Here, weevaluated the spatio-temporal relationship between the thermal induction of intravascular ICAM-1and lymphocyte trafficking.

Methods: Intravascular ICAM-1 density was quantified by immunofluorescence staining in miceexposed to fever-range whole-body hyperthermia (39.590.58C). ICAM-1�dependent lymphocytetrafficking was measured in short-term homing assays.

Results: A linear relationship was observed between the duration of heat treatment and intravascularICAM-1 density in HEVs with maximal responses requiring sustained (i.e., five hours) thermal stress.Circulating lymphocytes were found to sense incremental changes in ICAM-1 on HEVs, such thattrafficking is proportional to the intravascular density of ICAM-1. We further identified ahydroxamate-sensitive shedding mechanism that restores ICAM-1 expression to homeostatic levelsfollowing the cessation of thermal stress.

Conclusions: The time-dependent response to thermal stress indicates that ICAM-1 density governsthe efficiency of lymphocyte interactions with HEVs in vivo. These studies highlight the dynamic roleof the microcirculation in promoting immune surveillance during febrile inflammatory responses.Microcirculation (2009) 16, 143�158. doi:10.1080/10739680802353850

KEY WORDS: fever-range thermal stress, high endothelial venules, lymphocyte trafficking, intercellularadhesion molecule-1

The adaptive immune response depends on thecontinuous migration of blood-borne lymphocytesinto secondary lymphoid organs where naı̈ve andcentral memory lymphocytes encounter antigen-presenting cells. The major site of entry of blood-borne lymphocytes into lymph nodes and Peyerpatches is across postcapillary high endothelialvenules (HEVs) [9, 66]. Cuboidal HEVs are dis-tinguished from squamous endothelium lining themajority of blood vessels throughout the body bytheir ability to support efficient trafficking under

homeostatic conditions [25]. Lymphocyte migrationacross HEVs depends on a highly ordered sequenceof adhesion events that includes: (1) primarytethering and rolling along vessel walls, (2) chemo-kine-dependent activation, (3) secondary firm ar-rest, and (4) transendothelial migration [9,47,66].

Recent studies have shown that lymphocyte traffick-ing across HEVs in secondary lymphoid organs isaugmented by the thermal element of fever [15�18,23]. Fever has been linked with improved survivalduring acute infection in vertebrate species,although its physiological benefit is poorly under-stood [3, 27]. The duration and temperature rangeof the hyperthermic component of physiologicalfever has been simulated in vivo by the administra-tion of whole-body hyperthermia (WBH) to raisecore temperatures to 39.590.58C [8, 16]. Fever-range WBH treatment of both mice and advanced

Address correspondence to Sharon S. Evans, Department ofImmunology, Roswell Park Cancer Institute, Carlton and ElmStreets, Buffalo, NY 14263, USA. E-mail: [email protected]*Q.C. and M.M.A. contributed equally to this work

Received 23 May 2008; accepted 19 July 2008.

Microcirculation, 16: 143�158, 2009Copyright # 2009 Informa UK Ltd.ISSN: 1073-9688 print / 1549-8719 onlineDOI: 10.1080/10739680802353850

cancer patients causes a transient decrease in thenumber of circulating lymphocytes [23, 29, 43]. Inmice, this corresponds with an �twofold increase inlymphocyte trafficking selectively across the specia-lized microvasculature of HEVs [16, 17, 23]. Thischange represents a profound effect on an alreadyefficient trafficking mechanism, whereby �105

cells are estimated to transit across HEVs hourlyin a single lymph node under normothermic condi-tions [65]. Heightened trafficking across HEVs ispredicted to improve immune surveillance by en-hancing the probability that rare antigen-restrictedlymphocytes will encounter cognate antigens inorgans strategically positioned to be the first lineof defense against pathogens that enter via the skinor the respiratory, urogenital, or gastrointestinaltracts.

Systemic thermal stress improves lymphocyte traf-ficking by acting at discrete steps in the adhesioncascade required for extravasation across HEVs.Intravital microscopy revealed that thermal stresscauses an �twofold increase in the ability of HEVsto support the firm arrest of lymphocytes on vesselwalls, which leads to transendothelial migration intothe underlying parenchyma [15,16,23]. Enhancedfirm arrest in response to fever-range thermal stressis coordinated by increases in the intravasculardensity of the chemokine CCL21 and intercellularadhesion molecule-1 (ICAM-1) on HEVs[16,54,64]. CCL21 contributes to lymphocyte traf-ficking through the engagement of CCR7 receptorson apposing lymphocytes, resulting in increasedaffinity and avidity of leukocyte function-associatedantigen-1 (LFA-1) for its endothelial counter-re-ceptors, ICAM-1 and ICAM-2 [9,16,33,66]. Inter-leukin-6 (IL-6) is the sole cytokine found to regulateHEV expression of ICAM-1 during thermal stress,whereas CCL21 is induced through an IL-6�inde-pendent mechanism [16,64]. Thus, ICAM-1 upre-gulation was shown to be crucial for the thermalresponse based on evidence that disruption of IL-6signaling prevents thermal stimulation of lympho-cyte trafficking across HEVs [16].

The effects of systemic thermal stress on ICAM-1density were previously examined at a fixed timepoint (i.e., six hours), at which a marked increase inICAM-1 density was detected, similar to levelsinduced by potent inflammatory stimuli includingtumor necrosis factor (TNF) or IL-6 [16,54,64].These observations raised fundamental questionsregarding the spatio-temporal relationship betweenthermally induced changes in the intravascular

density of ICAM-1 and the efficiency of lymphocytetrafficking, as well as the mechanisms leading to therestoration of homeostatic trafficking after thecessation of thermal stress. Here, we present evi-dence of a linear relationship between the durationof heat treatment and the density of ICAM-1expressed on the lumenal surface of HEVs. Further,circulating lymphocytes are capable of sensingincremental increases in ICAM-1 presented byHEVs, such that their trafficking response is pro-portional to the intravascular density of ICAM-1.Finally, the restoration of ICAM-1 to homeostaticlevels following the cessation of thermally-inducedinflammation was shown to require the activity of azinc-dependent metalloproteinase. These findingssupport the notion that dynamic changes in intra-vascular ICAM-1 density in HEVs act as a rheostatto regulate the immune surveillance of secondarylymphoid organs during inflammation and infection.

MATERIALS AND METHODS

Mice

Age-matched (8�12 weeks) BALB/c and C57BL/6female mice were from the National Cancer Institute(Frederick, Maryland, USA) and Taconic (Hudson,New York, USA) and were maintained underpathogen-free barrier conditions. All animal proto-cols were approved by the Roswell Park CancerInstitute Institutional Animal Care and Use Com-mittee (Buffalo, New York, USA). Since it wasessential to process all mice and tissue specimensidentically within a given time-course experiment(i.e., during heating, adoptive transfer of spleno-cytes, injection of ICAM-specific antibody [Ab],harvesting tissues, and staining/photographing tis-sue sections), each experiment included one mouseper time point (totaling four to seven mice perexperiment), and each independent experiment wasrepeated two to five times.

Treatment with Fever-Range WBH or TAPI-1Protease Inhibitor

Mice were treated with fever-range WBH (coretemperature of 39.590.58C) for the indicated timeperiods in an environmental chamber (modelBE5000; Memmert, Schwabach, Germany), as de-scribed [8,16,23,45]. Normothermic (NT) controlmice (core temperature, 36.890.28C) were main-tained at room temperature in a darkened cabinet

Thermal control of ICAM-1�dependent trafficking144 Q Chen et al.

for the experimental period. Heated and controlmice were injected intraperitoneally with 1 mL ofsterile saline prior to the initiation of experiments toavoid dehydration. Core temperatures of sentinelmice in all experimental groups were measured byusing a subcutaneously implanted microchip ther-motransponder and a programmable data-acquisi-tion system (Bio Medic Data Systems, Seaford,Delaware, USA). Heat treatment was maintainedfor the designated time period, once the micereached the target temperature of 39.590.58C[45]. The TAPI-1 protease inhibitor (Calbiochem,San Diego, California, USA) was injected via the tailvein at a dose reported to block the sheddase-dependent release of TNF in vivo by �50-fold(0.5 mg in 0.25 mL saline) [38].

Analysis of Intravascular Display of AdhesionMolecules and HEV Density

Quantitative analysis of intravascular staining wasperformed as described [16]. Primary monoclonalantibodies (mAbs) specific for ICAM-1 (3E2, 25 mg/mouse) and ICAM-2 (3C4, 50 mg/mouse) (BDBiosciences, San Jose, California, USA) or isotypecontrol Ab were injected via the tail vein ofexperimental mice 20 minutes after the cessationof heat treatment. Organs (pooled peripheral lymphnodes [paired inguinal, brachial, axillary, sciatic,superficial, and deep cervical nodes], spleen, pan-creas, all mesenteric lymph nodes, and Peyerpatches) were harvested 20 minutes later andembedded in an optimum cutting temperaturecompound (OCT; Sakura Finetek, Torrance, Cali-fornia, USA). Cryosections (9 mm) were fixed inmethanol:acetone (3:1), rehydrated in phosphate-buffered saline (PBS), and incubated with goatserum (1:10 dilution) to block Fc receptors. Sometissue cryosections were subsequently stained withprimary mAb specific for PNAd (MECA79, 20 mg/mL), CD31 (MEC13.3, 5 mg/mL), or isotype con-trols (BD Biosciences). After washing, primary Abswere detected by fluorochrome-conjugated (fluor-escein or rhodamine) secondary Ab (Jackson Im-munoResearch, West Grove, Pennsylvania, USA). Inselected experiments, frozen tissues were stainedwith primary anti-ICAM-1 Ab (3E2; 5 mg/mL) aftersectioning.

Digital images were captured with an OlympusBX50 upright fluorescence microscope equipped

with a SPOT RT camera (Spectra Services, Roche-ster, New York, USA), using identical exposuretimes and image settings within each experiment.Cuboidal HEV structures (in lymph nodes, Peyerpatches) or CD31� vessels (in pancreas or spleen)were encircled by using an Intuos3 Tablet (Wacom,Vancouver, Washington, USA) and fluorescenceanalyzed by using ImageJ software [1] was ex-pressed as fluorescence intensity/pixels (i.e., reflect-ing a fixed unit area) [16]. Histograms represent thedata from all pixels analyzed (ranging from 1 to1.7�106) for the total number of HEVs percryosection of pooled 14 peripheral lymph nodes(ranging from 100 to 175); similar numbers ofHEVs and pixels were analyzed for each conditionwithin individual experiments. Presentation of thedata as a distribution measures population shiftsand can detect subtle differences between condi-tions. HEV density in pooled peripheral lymph nodecryosections was determined by enumerating thetotal HEVs per field in ] 25 fields (unit area ofeach field�0.34 mm2).

Intravital Analysis of Vessel Diameter and Blood CellVelocity

Intravital microscopy of inguinal lymph nodes wasperformed as described [16,65]. C57BL/6 micewere anesthetized by an intraperitoneal injection of1 mg/mL of xylazine and 10 mg/mL of ketamine(10 mL/kg body weight). The postcapillary vascularstructures were observed under fluorescent micro-scopy, and each venule was assigned an order (fromI to V). High-order (III�V) venules within theparacortical region represent HEVs, whereas low-order (I and II) venules are flat-walled vessels thatdrain into the superficial epigastric vein [65]. Thediameters of order III venules were measured atorder II/III branching points. To measure blood cellvelocity, approximately 2.5�107 calcein-labeledsplenocytes were injected into a catheter insertedinto the right femoral artery and fluorescent micro-scopy was used to measure the velocity of noninter-acting (i.e., fast moving) cells in order III venules, asdescribed [16,57,65]. All observations were cap-tured by using an EB CCD camera (HamamatsuPhotonics, Bridgewater, New Jersy, &USA) andrecorded by using a digital videocassette recorder(DSR-11; Sony, New York, New York, USA) foranalysis.

Thermal control of ICAM-1�dependent trafficking145Q Chen et al.

Determination of Vascular Permeability

Quantitative analysis of vascular integrity/perme-ability was performed essentially as described[7,22]. For this analysis, 2.5 mg of TRITC-con-jugated bovine serum albumin (BSA; Invitrogen,Carlsbad, California, USA) was injected via the tailvein of experimental mice 20 minutes after cessationof WBH treatment. Organs were harvested threeminutes later and fixed for four hours in Carnoy’ssolution (ethanol:chloroform:acetic acid, 6:3:1), fol-lowed by graded ethanol (100%, 90%, and 70%)and PBS prior to embedding and freezing in OCT.HEVs of frozen cryosections (9 mm) were identifiedby MECA-79 Ab staining (25 mg/mL) [16]. Leakageof TRITC-BSA across HEVs was analyzed asdescribed [7], in paired regions of 11�12 mm(equivalent to 30�33 pixels) positioned 50 mmapart, one within the lumen of HEVs and the secondwithin the surrounding lymph node stroma. Eachpixel within these defined regions (total, �80HEVs) was assigned a fluorescence intensity valueon a scale from 0 to 255 (generated by ImageJ). Thepermeability index is the ratio of lumenal meanfluorescence intensity (MFI) to stromal MFI.

In Vivo Homing Assay

Lymphocyte homing to lymphoid and nonlymphoidorgans was assessed in short-term (i.e., one hour)homing assays [16,23]. Mouse splenocytes werelabeled with 4 mM TRITC (Molecular Probes, Carls-bad, California, USA) for 20 minutes at 378C and thelabeling was stopped by centrifugation through afetal calf serum cushion (Invitrogen). Within indivi-dual experiments, equivalent numbers of labeled cells(ranging from 1 to 2�107) were adoptively trans-ferred by intravenous injection into mice 20 minutesafter the cessation of WBH and organs were harvestedone hour after cell transfer. The number of fluor-escent-labeled cells was quantified (i.e., blinded) in]10 fields (unit area of each field�0.34 mm2) ofnonsequential 9-mm cryosections, using an OlympusBX50 (Spectra Services, Rochester, New York, USA)upright fluorescence microscope. In selected experi-ments, mice were treated 30 minutes before celltransfer with neutralizing mAb specific for ICAM-1(3E2, 50 mg/mouse; BD Biosciences), ICAM-2 (3C4,50 mg/mouse; BD Biosciences), or isotype control Ab.Alternatively, TRITC-labeled splenocytes were pre-treated with mAb specific for LFA-1 (M17/4, 50 mg/

mL; eBioscience, San Diego, California, USA) orisotype control Ab 30 minutes before transfer.

Analysis of Plasma-Soluble (s)ICAM-1

Concentrations of soluble ICAM-1 (sICAM-1) inheparinized plasma were quantified by enzyme-linked immunosorbent assay (ELISA; R&D Systems,Minneapolis, Minneasota, USA).

Statistical Analysis

Statistical analysis of short-term homing, vesseldiameter, blood cell velocity, vascular permeability,and soluble ICAM-1 was performed by using theunpaired, two-tailed Student’s t-test, where P50.05was significant [16]; data are expressed as themean9standard error of the mean (SEM).

For the analysis of ICAM-1 and ICAM-2 density,Kolmogorov Smirnov tests [19] were used to eval-uate the null hypothesis that paired datasets werefrom the same underlying distribution. A permuta-tion scheme was performed (1000 runs) to deter-mine the P-values for significance for theKolmogorov Smirnov statistic, D. To generate athreshold D-value that represents intermouse varia-bility, we compared ICAM-1 immunofluorescencedata from five nonheated control mice in a singleexperiment. We found that MFI values were highlyconsistent in these replicate mice analyzed underidentical Ab staining conditions and camera settings(MFIs: 45.6, 42.9, 42.1, 42.9, and 42.8; mean�43.3, SEM�0.06). The largest observed Kolmo-gorov Smirnov statistic (D�0.09) representsmouse-to-mouse variation and was applied to allpair-wise comparisons of immunofluorescence re-sults, such that D50.09 indicates that the twodatasets are not different from each other, whereasD�0.09 indicates that the datasets are different.The R-programming language was used to performall tests and all permutation simulations [59].

RESULTS

ICAM-1 Induction Depends on Prolonged andSustained Exposure to Fever-Range Thermal Stress

Kinetics studies were performed to identify spatio-temporal changes in the intravascular density ofICAM-1 in peripheral lymph node HEVs ofBALB/c mice after treatment with fever-range


WBH (Figure 1). ICAM-1 expression was detectedby in situ immunofluorescence analysis followinghyperthermia treatment for various time periods(two, four, and six hours of continuous elevation ofcore temperature to 39.590.58C). ICAM-1 wasdetected on the lumenal surface of HEVs (i.e., thesite of lymphocyte contact during extravasation) byinjecting ICAM-1�specific mAb into the vascularcompartment and then counter-staining frozen tis-sue sections with fluorescence-labeled detectionantibody.

Quantitative image analysis revealed a linear rela-tionship between ICAM-1 intravascular density andthe duration of inflammatory cues provided bythermal stress. Maximal ICAM-1 induction wasobserved following continuous heat treatment forsix hours, reflected by the �twofold increase in themean fluorescence intensity (MFI) of ICAM-1 de-tected in HEVs of heated mice, compared with thesteady-state levels of the normothermic control (NT)(Figure 1). Incremental increases observed in in-travascular ICAM-1 density after treatment with

WBH for each of the experimental intervals (two,four, and six hours) were all different from nor-mothermal control as well as from every other timepoint (D�0.09, by Kolmogorov Smirnov distribu-tion analysis). Similar kinetics of ICAM-1 inductionwere observed in Peyer patch and mesenteric lymphnode HEVs (data not shown). In contrast, thermalstress did not alter intravascular ICAM-1 density inCD31� vessels of extralymphoid organs that lackHEV structures, such as pancreas or spleen (i.e.,MFIs for NT and mice treated six hours with WBH,respectively, were 34.7 and 36.1 for pancreas and17.9 and 19.3 for spleen; D50.09). Intravasculardisplay of the functionally and structurally relatedadhesion molecule, ICAM-2 was measured in HEVsafter WBH treatment as an internal control, sinceICAM-2 is not inducible under inflammatory condi-tions [24]. No change was detected in ICAM-2staining (MFI: 67 [NT] and 69 [WBH]; D50.09),demonstrating that the observed increases forICAM-1 cannot be attributed to trivial mechanisms,such as improved access of Ab to adhesion moleculeson the surface of HEVs.

A possible explanation for the limited ICAM-1induction detected after short-term (i.e., two-hour)heat treatment was that insufficient time was allowedfor de novo synthesis and expression of ICAM-1 on thelumenal surface of HEVs. In this regard, studies inprimary endothelial cells in vitro have shown thatICAM-1 mRNA levels are detectable as early as twohours following stimulation by inflammatory cyto-kines (TNF, IL-1b), while expression of ICAM-1protein at high density on the endothelial surfacerequires six to eight hours [40]. To determine if short-term exposure to fever-range thermal stress wassufficient to initiate ICAM-1 induction on HEVs,mice were treated with continuous WBH for variableperiods, ranging from one to six hours. Mice were thenallowed to revert to normothermal temperature andintravascular ICAM-1 density was detected after atotal six-hour period (Figure 2). ICAM-1 stainingafter WBH treatment for one to two hours wasindistinguishable from baseline levels of ICAM-1(D50.09). ICAM-1 expression after either three orfour hours of heat exposure was moderately elevated,compared to control (D�0.09), but not differentfrom each other (D50.09). Maximal lumenal displayof ICAM-1 (�twofold upregulation) was observedafter WBH treatment for five or six hours; ICAM-1densities at these time points were significantlydifferent from all other conditions (D�0.09), butwere indistinguishable from each other (D50.09).

Figure 1. Thermal induction of intercellular adhesionmolecule-1 (ICAM-1) depends on the duration of heattreatment. Intravascular staining of ICAM-1 in peripherallymph nodes was performed by intravenous injection ofprimary monoclonal antibody (mAb) in normothermiccontrol (NT) mice and mice treated with whole-bodyhyperthermia (WBH) for various periods. Peripherallymph node cryosections were counter-stained with fluor-ochrome-conjugated secondary Ab. Immunofluorescencestaining was not observed using isotype-matched controlAb (not shown). Photomicrographs show typical imagesfrom NT mice and mice with WBH treatment for sixhours. Scale bar, 50 mm. Histograms indicate the im-munofluorescence intensity of ICAM-1 staining quantifiedin cuboidal high endothelial venules by image analysis.The x-axis indicates fluorescence intensity, and the y-axisindicates the number of pixels with each intensity; themean fluorescence intensity (MFI) is indicated in eachcondition. Data represent ]3 independent experiments.


Collectively, these results demonstrate that prolongedand sustained thermal stress is required for optimalexpression of ICAM-1 in HEVs.

Fever-Range Thermal Stress Enhances LymphocyteTrafficking Across HEVs in an ICAM-1-DependentManner

Since maximal induction of ICAM-1 expression inHEVs was detected after continuous WBH treatmentfor six hours (Figures 1 and 2), this time point wasused in short-term (i.e., one-hour) homing studies toinvestigate the contribution of ICAM-1 density tolymphocyte trafficking. For these experiments,fluorescent (TRITC)-labeled splenocytes were adop-tively transferred into mice 20 minutes after the

cessation of WBH treatment. This allowed mice torevert to normal core body temperature (Figure 3A),thereby avoiding the reported effects of direct heaton the homing potential of transferred lymphocytes[23,18].

An �twofold increase in trafficking of TRITC-labeled lymphocytes was detected in response tothermal stress exclusively in organs with HEVs,including peripheral lymph nodes (Figure 3B and3C), mesenteric lymph nodes, and Peyer patches(data not shown). Enumeration of fluorescent-la-beled transferred cells within or outside HEVs(demarked by staining for the HEV-specific mole-cule, PNAd [58]) revealed that thermal stress didnot alter the relative proportion (�70%) of cellsthat completed the adhesion cascade leading toextravasation into the lymph node stroma (Figure3B; P�0.67). Labeled cells within the vascularspace of HEVs (Figure 3B) appear to be activelyinteracting with vessel walls, rather than merelypassing through the local vasculature. This conclu-sion is based on comparisons with organs such as thepancreas (Figure 3C), in which adoptively trans-ferred cells are rarely detected, despite high vasculardensity (i.e., 131.9916.5 CD31� vessels per unitarea in pancreas versus 5.391.6 PNAd� HEVs perunit area in lymph nodes; PB0.0001). Consistentwith findings that thermal stress did not augmentICAM-1 expression on non-HEV vessels in spleen orpancreas, no increase in lymphocyte trafficking wasdetected in these organs following WBH (Figure 3C;P�0.4). The constant level of trafficking observedin these organs provided an internal control for thetransfer of equivalent number of lymphocytes inindividual mice within a given experiment.

The role of ICAM-1 in thermally regulated traffick-ing across HEVs was previously examined by usingfunction-blocking mAb directed against ICAM-1[16]. Here, we extended these findings by blockingthe ICAM-1 counter-receptor, LFA-1, on lympho-cytes prior to adoptive transfer into control or heat-treated mice. Under normothermic conditions, lym-phocyte homing across HEVs was substantiallyinhibited by LFA-1�blocking mAb, whereas mAbspecific for either ICAM-1 or -2 had little effect(Figure 3C). This result is consistent with priorfindings showing that ICAM-1 and -2 are function-ally redundant ligands for LFA-1 during homeo-static trafficking in lymphoid organs [16,31].Transferred cells detected within HEVs in thepresence of LFA-1�blocking mAb likely reflectL-selectin/PNAd�dependent tethering and rolling

Figure 2. Sustained heat treatment is required formaximal thermal induction of intercellular adhesionmolecule-1 (ICAM-1) expression. Experimental mice weretreated with whole-body hyperthermia (WBH) for varioustime periods and then allowed to revert to normothermaltemperature (NT). At the end of the six-hour period,intravascular density of ICAM-1 in peripheral lymphnodes was quantified by image analysis, as shown inhistograms. The x-axis indicates fluorescence intensityand the y-axis indicates the number of pixels with eachintensity; the mean fluorescence intensity (MFI) is in-dicated in each condition. Data represent ] 3 indepen-dent experiments.


Figure 3. Thermal enhancement of trafficking across high endothelial venules (HEVs) is dependent on leukocytefunction associated antigen-1/intercellular adhesion molecule-1 (LFA-1/ICAM-1) interactions. (A) Core temperaturesof sentinel mice were monitored every 30 minutes during 6 hour whole-body hyperthermia (WBH) and every minuteafter cessation of WBH. The dotted line represents baseline normothermic temperature (NT); measurements werestarted when the target temperature range was reached (39.590.58C). Arrow indicates the time point at whichmonoclonal antibodies (mAbs) or adoptive transfer cells were injected. Data represent ]3 independent experiments.(B, C) In vivo short-term (one-hour) homing of TRITC (red)-labeled splenocytes to peripheral lymph nodes (PLN),spleen (SPL), and pancreas (Panc) was assessed in NT control or WBH-treated (six-hour) BALB/c mice. (B) PLNcryosections were stained with MECA-79 mAb specific for PNAd to visualize HEVs (green). Photomicrographs showtypical images of TRITC-labeled cells either associated with HEVs or infiltrated into the tissue parenchyma. The totalnumber of TRITC-labeled cells and the percentage of cells within the lymph node parenchyma were quantified in tissuecryosections by fluorescence microscopy (n=10 fields; bar=50 mm). (C) Splenocytes were treated with LFA-1 function-blocking mAb or control mAb prior to transfer. Alternatively, 20 minutes prior to adoptive transfer, recipient mice weretreated with function-blocking mAb specific for ICAM-1 or -2 or with isotype-matched control Ab. The number ofTRITC-labeled cells in tissue cryosections was quantified by fluorescence microscopy (n�10 fields). P-values representthe difference between NT and WBH conditions in (B) and represent the difference between function-blocking reagentsand controls in (C). *PB0.0001. Data represent ]3 independent experiments.


interactions. Improved lymphocyte traffickingacross HEVs in response to thermal stress wasfurther shown to be dependent on cognate interac-tions between LFA-1 and ICAM-1 (Figure 3C). Inthis regard, thermal enhancement of lymphocytetrafficking to peripheral lymph nodes was fullyprevented by the mAb blockade of LFA-1 onlymphocytes prior to adoptive transfer as well asby neutralizing mAb specific for ICAM-1, but notICAM-2 (Figure 3C). Heat treatment did notinfluence LFA-1/ICAM-1�dependent lymphocytehoming to organs lacking HEV structures includingspleen and pancreas (Figure 3C; P�0.4).

We formally excluded the possibility that lympho-cyte homing detected following WBH was influenced

by changes in hemodynamic parameters by measur-ing vessel diameter, blood cell velocity, HEV den-sity, and vascular integrity/permeability at the timeof the initiation of homing assays (i.e., 20 minutesfollowing cessation of WBH). In these studies, vesseldiameter and blood cell velocity were determined tobe equivalent in normothermal controls and WBH-pretreated mice (Figure 4A; P�0.10 and 0.36,respectively) by intravital microscopic analysis ofHEVs in inguinal lymph nodes. Moreover, there wasno difference in the overall density of HEVs inheated versus nonheated lymph nodes (Figure 4B;P�0.77). To rule out the possibility that thermalstress enhanced trafficking by compromising theintegrity of the vascular barrier, we measured thepermeability of HEVs following the intravenous

Figure 4. Hemodynamic parameters are normal in high endothelial venules (HEVs) after whole-body hyperthermia(WBH) treatment. (A) Vascular diameter and blood cell velocity were quantified by intravital microscopy in level IIIvenules of inguinal lymph nodes. For vascular diameter (left), data were generated from 15 order III venules (i.e.,HEVs) of normothermic (NT) control mice and 11 order III venules of WBH-treated (six hours) mice (n�3 per group).For blood cell velocities (right), the velocity of�60 noninteracting (i.e., fast moving) cells in order III venules wasmeasured in NT control and WBH-treated (six hours) mice (n�3 per group). (B) HEV density was quantified inpooled peripheral lymph nodes from NT control and WBH-treated (six hours) mice. (C) TRITC-conjugated bovineserum albumin was introduced into the vascular compartment of mice 20 minutes after WBH treatment and allowed tocirculate for three minutes. Peripheral lymph node cryosections were counterstained for PNAd (green) to demarkHEVs. Permeability was determined by measuring the relative fluorescence within the vascular space of �80 PNAd�

HEVs versus fluorescence within the surrounding lymph node stroma. Permeability index�lumenal MFI/stromal MFI.Bar�50 mm. Data in (B) and (C) represent ] 2 independent experiments. Differences between NT control and WBH-treated mice are not significant (P ] 0.1).


injection of TRITC-conjugated BSA [7,22]. Quanti-tative image analysis of the relative fluorescencewithin the vascular space of PNAd� HEVs versusfluorescence within the surrounding lymph nodestroma revealed that there was no difference in thepermeability of HEVs from WBH-treated mice,compared to normothermal control mice (Figure4C; P�0.34). Taken together, these results estab-lish that changes in lymphocyte trafficking detectedin response to thermal stress in adoptive transferexperiments can be attributed to the induction ofICAM-1 expression in HEVs.

Maximal ICAM-1 Density is Required for OptimalLymphocyte Homing Across HEVs

Evidence that ICAM-1 density in HEVs is spatio-temporally regulated by thermal stress provided aunique in vivo model system to investigate therelationship between ICAM-1 density and the effi-ciency of lymphocyte homing to secondary lymphoidorgans. The contribution of ICAM-1 density tolymphocyte-endothelial adhesion has previouslybeen restricted mainly to in vitro systems, in whichICAM-1 was immobilized on artificial surfaces (e.g.,plastic, beads) or where ICAM-1 was induced incytokine (TNF)-stimulated cultured endothelialcells [20,37,46,50]. A key question in the currentstudy was whether moderate increases in ICAM-1density reached the threshold required for optimallymphocyte trafficking, or if homing is furtherimproved when ICAM-1 is maximally induced onHEVs.

This issue was addressed by performing short-termhoming assays in mice pretreated with WBH for two,four, or six hours to achieve variable levels of ICAM-1 on HEVs, following the heating schemas shown inFigures 2 and 5. Paralleling the lack of change inICAM-1 density after two hours of WBH treatment(Figure 2), there was no difference in lymphocytehoming to peripheral lymph nodes at this time point(Figure 5; P�0.34). Intermediate levels of homingwere observed following WBH pretreatment for fourhours, at a point when ICAM-1 density is moderatelyincreased, but suboptimal. Maximal homing wasobserved after six hours of continuous heat treat-ment when ICAM-1 levels were greatest (Figure 5).No change in lymphocyte trafficking was observedin spleen or pancreas at any of these time points(Figure 5). Thus, circulating lymphocytes senseincremental differences in ICAM-1 density presentedon the lumenal surface of HEVs, such that their

trafficking response is proportional to the intravas-cular density of ICAM-1.

Thermal Induction of ICAM-1 is Restored to NormalFollowing Cessation of Heat Treatment by a Zinc-Dependent Metalloproteinase

A hallmark of the febrile response is that it is self-limiting, thereby restoring homeostasis [28]. We,therefore, determined the duration of increasedICAM-1 intravascular density in HEVs followingthe cessation of WBH. Quantitative image analysisrevealed a linear decrease in ICAM-1 density, reach-ing homeostatic levels within six hours of cessationof WBH (Figure 6A). In this regard, statisticalanalysis showed that ICAM-1 densities at two,four, or six hours after the cessation of WBH wereall different from each other and, further, weredifferent from the highly induced levels detectedfollowing continuous WBH treatment for six hours(D�0.09). By six hours after the cessation ofWBH, ICAM-1 levels were indistinguishable from

Figure 5. Prolonged heat treatment is required foroptimal lymphocyte trafficking in peripheral lymphnodes. Experimental mice were treated with whole-bodyhyperthermia (WBH) for varying amounts of time andthen allowed to revert to normothermal temperature, asindicated. At the end of the six-hour period, in vivo short-term (one-hour) homing of fluorescent (TRITC)-labeledsplenocytes to peripheral lymph nodes (PLN), spleen(SPL), and pancreas (Panc) was assessed in normother-mic (NT) control (C) or WBH-treated mice. Organs wereharvested one hour after adoptive transfer, and thenumber of TRITC-labeled cells was quantified in tissuecryosections (n�10 fields) by fluorescence microscopy.P-values represent the difference between NT control andWBH treatment (*PB0.0001). Data represent ] 2independent experiments.


normothermal control (D50.09). Consistent withthe downregulation of ICAM-1 expression, lympho-cyte trafficking in peripheral lymph nodes returnedto normothermic levels at six hours after thecessation of heat treatment (Figure 6B).

These results raised the question of the fate ofICAM-1 following the cessation of WBH. Onepossibility was that ICAM-1 was internalized follow-ing ligation by LFA-1 on adherent cells, whichoccurs in cultured human endothelial cells only afterextensive ICAM-1 cross-linking. [30,39]. Alterna-tively, ICAM-1 could be actively shed from thelumenal surface of HEVs by proteases. The zinc-dependent proteolytic enzymes, ADAM-17/TACE (adisintegrin and metalloproteinase/TNF-alpha con-

verting enzyme) and MT1-MMP (membrane type 1-matrix metalloproteinase) have been implicated inthe cleavage of ICAM-1 from the surface of TNF-activated cultured endothelial cells in vitro [53,63],although the contribution of sheddases to theregulation of endothelial ICAM-1 in vivo has notbeen previously investigated. To determine if therestoration of ICAM-1 density to basal levels inHEVs involves a shedding mechanism, we usedTAPI-1, a peptidomimetic zinc-ligating metallopro-teinase inhibitor that has been shown to inhibitICAM-1 release by endothelial cells in vitro [53,63].In initial studies performed in normothermic mice,we determined that administration of TAPI-1 at aconcentration reported to block ADAM-17/TACE

Figure 6. Intercellular adhesion molecule-1 (ICAM-1) expression and lymphocyte trafficking is reversible afterwhole-body hyperthermia (WBH) treatment. (A) Intravascular staining of ICAM-1 in peripheral lymph nodes wasperformed in mice at the indicated time points after a six-hour heat treatment. The histograms indicate theimmunofluorescence intensity of ICAM-1 staining quantified in cuboidal high endothelial venules (HEVs) by imageanalysis. The x-axis indicates fluorescence intensity and the y-axis indicates the number of pixels with each intensity;the mean fluorescence intensity (MFI) is indicated in each condition. Data represent ] 3 independent experiments. (B)Short-term (one-hour) homing of fluorescent (TRITC)-labeled splenocytes to peripheral lymph nodes (PLN), spleen(SPL), and pancreas (Panc) was assessed in mice with different treatments: normothermic (NT) control, immediatelyafter six-hour WBH treatment, and six hours after cessation of heat treatment. Organs were harvested one hour afteradoptive transfer and the number of TRITC-labeled cells was quantified in tissue cryosections (n�10 fields) byfluorescence microscopy. P-values represent the difference between NT control and WBH treatment (*PB0.0001).Data represent ] 2 independent experiments.


activity in vivo [38] had no effect on basal levels ofICAM-1 density over a six-hour period, compared toDMSO vehicle controls (MFIs: 42 [NT], 43 [DMSO],and 39 [TAPI-1]; D50.09). Findings that the rateof constitutive TAPI-1�sensitive shedding of ICAM-1 by HEVs is limited under homeostatic conditionssuggest that the observed increase in ICAM-1density that occurs during fever-range thermal stresscannot be attributed solely to the inhibition ofconstitutive ICAM-1 shedding from high endothelialcells.

Treatment with TAPI-1 immediately after cessationof WBH inhibited the loss of ICAM-1 over a six-hourperiod by 89% (Figure 7). Statistical analysisindicated that ICAM-1 staining in HEVs of micetreated with TAPI-1 post-WBH was different frommice treated with DMSO post-WBH (D�0.09).Identical results were obtained when intravascularstaining was performed using higher amounts ofanti-ICAM-1 primary Ab (50 mg instead of 25 mg/mouse), or when frozen tissues were stained withprimary anti-ICAM-1 Ab after sectioning. Theseresults excluded the possibility that shed ICAM-1in the vascular compartment was interfering with

intravascular staining of ICAM-1 in HEVs. Inde-pendent confirmation that TAPI-1 affects cellularrelease of ICAM-1 in vivo was provided by findingsthat administration of TAPI-1 immediately aftercessation of thermal stress significantly decreasedcirculating levels of soluble (s)ICAM-1 over a six-hour period (i.e., plasma sICAM-1 concentrationswere 375.6924.2 ng/mL for TAPI-1 treatment vs.490.7944.8 ng/mL for non-TAPI-1�treated micewhen measured six hours post-WBH; PB0.05, dataare the mean9SEM; n�3 mice per group). More-over, a trend toward an increase in plasma sICAM-1was detected six hours after cessation of thermalstress, although the results did not reach statisticalsignificance (i.e., 490.7944.8 ng/mL for six hourspost-WBH vs. 448.2922.3 ng/mL for normother-mal controls, P�0.41; n�3 mice). The failure todetect an increase in sICAM-1 during the period thatICAM-1 was shed from HEVs may reflect the limitednumber of high endothelial cells contributing to thecirculating pool as well as clearance of sICAM-1 bythe liver and kidneys [6,61]. Collectively, these dataindicate that the rapid reduction in ICAM-1 displayon HEVs after cessation of WBH is due to active

Figure 7. Loss of intercellular adhesion molecule-1 (ICAM-1) after cessation of whole-body hyperthermia (WBH)treatment is due to proteolytic cleavage. Mice were subjected to WBH treatment for six hours to induce maximal ICAM-1 expression in high endothelial venules (HEVs). Mice were then treated with either the hydroxamic acid-basedmetalloproteinase inhibitor, TAPI-1, or with vehicle alone (dimethyl sulfoxide; DMSO). Photomicrographs show typicalimages of intravascular staining of ICAM-1 under the indicated conditions. Scale bar, 50 mm. Histograms indicate theimmunofluorescence intensity of ICAM-1 staining in cuboidal HEVs, as determined by quantitative image analysis. Thex-axis indicates fluorescence intensity and the y-axis indicates the number of pixels with each intensity; the meanfluorescence intensity (MFI) is indicated in each condition. Data represent ] 2 independent experiments.


shedding by ADAM-17/TACE or a related zinc-dependent metalloproteinase.

DISCUSSION

The continuous recirculation of naı̈ve lymphocytesacross HEVs of secondary lymphoid organs plays acritical role in immune surveillance. Under homeo-static conditions, the cellular content of a singlemouse lymph node (comprised of �106 cells) isreplaced �three times each day [65]. This high-throughput process favors the chance that rareantigen-specific T cells, present at a frequency of�1 in 105�106 cells [5,44], will encounter dendriticcells presenting cognate antigen from peripheraltissues. It has long been recognized that thedynamics of lymphocyte trafficking through regio-nal lymph nodes is dramatically altered during theclassical ‘‘shutdown’’ response that occurs 3�18hours after local infection or inflammation[26,10]. Exit from lymph nodes during the shut-down phase is prevented by transient downregula-tion of the sphingosine-1-phosphate receptor onantigen-stimulated lymphocytes, thereby allowingsufficient opportunity for these cells to undergoproliferation within the supportive environment ofthe lymph node [52]. A concomitant increase inentry of naı̈ve T cells across HEVs results fromdilation of the main arterioles that supply themicrocirculation in lymph nodes [55]. The totalnumber of HEVs is also elevated in inflamed lymphnodes, providing additional surface area for naı̈velymphocytes to initiate the multistep adhesioncascade leading to extravasation [55,68].

Although the adhesion mechanisms controllinghomeostatic lymphocyte trafficking in HEVs arewell understood, little is known about the molecularchanges in HEV adhesion that lead to increasedentry of naı̈ve T cells during inflammation. Here, weused thermal stress as a model of systemic febrileinflammation to investigate the relationship betweenICAM-1 density on HEVs and lymphocyte traffick-ing in lymphoid organs. The major findings of thisreport are: (1) The surface density of the gatekeeperhoming molecule ICAM-1 in HEVs is temporallycontrolled by the thermal element of the febrileinflammatory response; (2) ICAM-1 density onHEVs is a critical determinant of the efficiency oflymphocyte trafficking; and (3) Restoration ofICAM-1�dependent trafficking to homeostatic levels

following cessation of thermal stress requires theactivity of a zinc-dependent metalloproteinase.

Progressive increases in ICAM-1 density on HEVsfollowing heat treatment were linked to enhancedLFA-1/ICAM-1�dependent trafficking of lympho-cytes across this microvascular barrier. Hemody-namic parameters were normal at the time point atwhich homing assays were performed, stronglysuggesting that improved trafficking is attributedto increased availability of ICAM-1 to circulatinglymphocytes as they pass through HEVs. Theseobservations do not exclude the possibility thattransient hemodynamic changes occur during theactual heating period that contribute to increasedintravascular ICAM-1, particularly in view of re-ports that shear stress stimulates ICAM-1 synthesisand surface display on endothelial cells in vitro[41,62]. The temporal effects of heat on microvas-cular adhesion are consistent with the sustained timeinterval of physiological fever during inflammationand infection. Of particular note, we have shownthat strong HEV responses are triggered by inflam-matory cues provided by systemic thermal stress inthe absence of antigenic stimulation. Targetedregulation of trafficking across HEVs in lymphoidorgans throughout the body in the context of febriletemperatures, as shown in the present study, wouldbe immunologically advantageous, since immunesurveillance would be amplified locally in sentinellymph nodes (i.e., the site of antigen deposition) aswell as systemically.

The sensitivity of circulating lymphocytes to incre-mental changes in intravascular ICAM-1 is remark-able given that HEVs constitutively express ICAM-1at high density (i.e., at �13,80091500 sites/mm2

[20]). Elevation of ICAM-1 density on the lumenalsurface of HEVs by thermal stress is predicted toaugment trafficking by enhancing the probability ofencounters with a fixed number of LFA-1 moleculeson circulating lymphocytes. Our observations re-garding ICAM-1�dependent trafficking in HEVs invivo are in line with in vitro studies showing thatICAM-1 density affects lymphocyte adhesion[20,50]. In this regard, significantly greater lym-phocyte adhesion occurs on immobilized ICAM-1 athigher, compared to lower, density (6000 versus 500sites/mm2 [20] or 1900 versus 190 sites/mm2 [50]).Ligand interactions under conditions of high ICAM-1 density have been shown in vitro to stabilizeadhesion via outside-in signaling that converts LFA-1 to a high-affinity conformer [11,50]. Notably,ICAM-1, but not ICAM-2, forms dimers when


expressed at high density on cultured endothelialcells and bond lifetimes of dimeric ICAM-1 for LFA-1 are �10 times longer than the monomeric form[37,46,48]. High endothelial expression of ICAM-1could contribute to the preferential role demon-strated in vitro for this LFA-1 ligand in activation ofRho GTPase, which is required for cytoskeletalreorganization, generation of endothelial dockingstructures, and transendothelial migration of lym-phocytes [2,4,12,13,34,36,42,51,60].

The rapid reduction in ICAM-1 density detected inHEVs following cessation of thermal stress isreminiscent of the kinetics of the early phase ofresolution of inflammation, which aims for a returnto homeostasis. Normalization of ICAM-1 density inHEVs within six hours restores trafficking to con-stitutive levels, thus reducing the metabolic de-mands associated with new ICAM-1 synthesis andactive lymphocyte migration. Evidence that the lossof ICAM-1 from HEVs is inhibited by the hydro-xamic acid-based metalloproteinase inhibitor,TAPI-1, as reported for endothelial cells in vitro[53,63], strongly implicates ADAM-17/TACE and/or MT1-MMP in ICAM-1 ectodomain cleavageduring the return to vascular homeostasis followingfebrile episodes. Recent findings [35] that ICAM-1 isshed as functionally active dimers in vivo raise thepossibility that sICAM-1 could locally dampenlymphocyte-endothelial adhesion in HEVs duringthe resolving phase of inflammation.

Our data further demonstrate a differential effect ofTAPI-1 on ICAM-1 expression in HEVs of nor-mothermic mice and heat-treated mice. In thisregard, baseline steady-state levels of ICAM-1 onHEVs were not altered by TAPI-1 treatment innormothermic mice, whereas TAPI-1 had a pro-found effect when administered following WBHtreatment. These findings suggest that the rate ofICAM-1 shedding in high endothelial cells wasaccelerated during the resolving phase followingthe cessation of fever-range thermal stress. ADAM-17 activity in neutrophils or epithelial cells has beenshown in vitro to be induced in response to PMAactivation or phagocytosis of bacterial pathogenswithin a similar time frame as our studies (5 sixhours) by multiple mechanisms, including control ofmRNA levels, processing of the proenzyme to thecatalytically active mature form, and translocationfrom intracellular stores (endoplasmic reticulum) tothe cell surface [14,49,56,67,69]. It remains to bedetermined whether ICAM-1 cleavage in high en-dothelial cells following thermal stress is regulated

by these intrinsic mechanisms or if extrinsic me-chanisms are involved in which sheddases onadherent leukocytes act in trans to release ICAM-1substrate from HEVs.

CONCLUSIONS

In summary, these findings highlight the dynamicrole of the microcirculation in regulating immunesurveillance under acute inflammatory conditionsassociated with fever and the early phase of resolu-tion of inflammation. These results suggest that animportant function of fever is to act as a rheostat toaugment lymphocyte migration through secondarylymphoid organs by controlling ICAM-1�dependentadhesion in the lymphocyte�HEV axis throughoutthe body. The complex mechanisms regulatingICAM-1 density in vascular endothelium are likelyto also be relevant to inflammation associated withautoimmune disorders and in cancer patients duringlocal and systemic thermal therapy [21,32,47].

Declaration of interest: The authors report noconflicts of interest. The authors alone are respon-sible for the content and writing of the paper.

REFERENCES

1. Abramoff MD, Magelhaes PJ, Ram SJ. (2004). ImageProcessing with Image. J. Biophoton Int 11:36�42.

2. Adamson P, Etienne S, Couraud PO, Calder V,Greenwood J. (1999). Lymphocyte migration throughbrain endothelial cell monolayers involves signalingthrough endothelial ICAM-1 via a rho-dependentpathway. J Immunol 162:2964�2973.

3. Appenheimer MM, Chen Q, Girard RA, Wang WC,Evans SS. (2005). Impact of fever-range thermalstress on lymphocyte-endothelial adhesion and lym-phocyte trafficking. Immunol Invest 34:295�323.

4. Barreiro O, Yanez-Mo M, Serrador JM, Montoya MC,Vicente-Manzanares M, Tejedor R, Furthmayr H,Sanchez-Madrid F. (2002). Dynamic interaction ofVCAM-1 and ICAM-1 with moesin and ezrin in anovel endothelial docking structure for adherentleukocytes. J Cell Biol 157:1233�1245.

5. Blattman JN, Antia R, Sourdive DJ, Wang X, KaechSM, Murali-Krishna K, Altman JD, Ahmed R. (2002).Estimating the precursor frequency of naive antigen-specific CD8 T cells. J Exp Med 195:657�664.

6. Bonomini M, Reale M, Santarelli P, Stuard S,Settefrati N, Albertazzi A. (1998). Serum levels ofsoluble adhesion molecules in chronic renal failureand dialysis patients. Nephron 79:399�407.


7. Bossi F, Fischetti F, Pellis V, Bulla R, Ferrero E,Mollnes TE, Regoli D, Tedesco F. (2004). Platelet-activating factor and kinin-dependent vascular leak-age as a novel functional activity of the solubleterminal complement complex. J Immunol173:6921�6927.

8. Burd R, Dziedzic TS, Xu Y, Caligiuri MA, Subjeck JR,Repasky EA. (1998). Tumor cell apoptosis, lympho-cyte recruitment, and tumor vascular changes areinduced by low temperature, long duration (fever-like) whole body hyperthermia. J Cell Physiol177:137�147.

9. Butcher EC, Picker LJ. (1996). Lymphocyte homingand homeostasis. Science 272:60�6.

10. Cahill RN, Frost H, Trnka Z. (1976). The effects ofantigen on the migration of recirculating lymphocytesthrough single lymph nodes. J Exp Med 143:870�888.

11. Carman CV, Springer TA. (2003). Integrin avidityregulation: are changes in affinity and conformationunderemphasized? Curr Opin Cell Biol 15:547�556.

12. Carman CV, Springer TA. (2004). A transmigratorycup in leukocyte diapedesis both through individualvascular endothelial cells and between them. J CellBiol 167:377�388.

13. Cernuda-Morollon E, Ridley AJ. (2006). RhoGTPases and leukocyte adhesion receptor expressionand function in endothelial cells. Circ Res 98:757�767.

14. Chalaris A, Rabe B, Paliga K, Lange H, Laskay T,Fielding CA, Jones SA, Rose-John S, Scheller J.(2007). Apoptosis is a natural stimulus of IL6Rshedding and contributes to the proinflammatorytrans-signaling function of neutrophils. Blood110:1748�1755.

15. Chen Q, Clancy KA, Wang WC, Evans SS. Inflam-matory Cues Controlling Lymphocyte-EndothelialInteractions During Fever-Range Thermal Stress.In: Aird W, editor. Endothelial Biomedicine: Cam-bridge University Press; 2007. pp 471�479.

16. Chen Q, Fisher DT, Clancy KA, Gauguet JM, WangWC, Unger E, Rose-John S, von Andrian UH,Baumann H, Evans SS. (2006). Fever-range thermalstress promotes lymphocyte trafficking across highendothelial venules via an interleukin 6 trans-signal-ing mechanism. Nat Immunol 7:1299�1308.

17. Chen Q, Fisher DT, Kucinska SA, Wang WC, EvansSS. (2006). Dynamic control of lymphocyte traffick-ing by fever-range thermal stress. Cancer ImmunolImmunother 55:299�311.

18. Chen Q, Wang WC, Bruce R, Li H, Schleider DM,Mulbury MJ, Bain MD, Wallace PK, Baumann H,Evans SS. (2004). Central role of IL-6 receptorsignal-transducing chain gp130 in activation of L-selectin adhesion by fever-range thermal stress.Immunity 20:59�70.

19. Conver WJ. Practical Nonparametric Statistics (pp.309�314). New York: John Wiley & Sons, 1971.

20. Constantin G, Majeed M, Giagulli C, Piccio L, KimJY, Butcher EC, Laudanna C. (2000). Chemokinestrigger immediate beta2 integrin affinity and mobilitychanges: differential regulation and roles in lympho-cyte arrest under flow. Immunity 13:759�769.

21. Corry PM, Dewhirst MW. (2005). Thermal medicine,heat shock proteins, and cancer. Int J Hyperthermia21:675�677.

22. Dafni H, Landsman L, Schechter B, Kohen F, Nee-man M. (2002). MRI and fluorescence microscopy ofthe acute vascular response to VEGF165: vasodila-tion, hyper-permeability, and lymphatic uptake,followed by rapid inactivation of the growth factor.NMR Biomed 15:120�131.

23. Evans SS, Wang WC, Bain MD, Burd R, Ostberg JR,Repasky EA. (2001). Fever-range hyperthermiadynamically regulates lymphocyte delivery to highendothelial venules. Blood 97:2727�2733.

24. de Fougerolles AR, Stacker SA, Schwarting R,Springer TA. (1991). Characterization of ICAM-2and evidence for a third counter-receptor for LFA-1.J Exp Med 174:253�267.

25. Girard JP, Springer TA. (1995). High endothelialvenules (HEVs): specialized endothelium for lym-phocyte migration. Immunol Today 16:449�457.

26. Hall JG, Morris B. (1965). The immediate effect ofantigens on the cell output of a lymph node. Br J ExpPathol 46:450�454.

27. Hasday JD, Fairchild KD, Shanholtz C. (2000). Therole of fever in the infected host. Microbes Infect2:1891�1904.

28. Kluger MJ. (1991). Fever: role of pyrogens andcryogens. Physiol Rev 71:93�127.

29. Kraybill WG, Olenki T, Evans SS, Ostberg JR,O’Leary KA, Gibbs JF, Repasky EA. (2002). A phaseI study of fever-range whole body hyperthermia (FR-WBH) in patients with advanced solid tumours:correlation with mouse models. Int J Hyperthermia18:253�266.

30. Kuijpers TW, Raleigh M, Kavanagh T, Janssen H,Calafat J, Roos D, Harlan JM. (1994). Cytokine-activated endothelial cells internalize E-selectin intoa lysosomal compartment of vesiculotubular shape. Atubulin-driven process. J Immunol 152:5060�5069.

31. Lehmann JC, Jablonski-Westrich D, Haubold U,Gutierrez-Ramos JC, Springer T, Hamann A.(2003). Overlapping and selective roles of endothelialintercellular adhesion molecule-1 (ICAM-1) andICAM-2 in lymphocyte trafficking. J Immunol171:2588�2593.

32. Luster AD, Alon R, von Andrian UH. (2005).Immune cell migration in inflammation: present andfuture therapeutic targets. Nat Immunol 6:1182�1190.


33. Luo BH, Carman CV, Springer TA. (2007). Struc-tural basis of integrin regulation and signaling. AnnuRev Immunol 25:619�647.

34. Lyck R, Reiss Y, Gerwin N, Greenwood J, AdamsonP, Engelhardt B. (2003). T-cell interaction withICAM-1/ICAM-2 double-deficient brain endotheliumin vitro: the cytoplasmic tail of endothelial ICAM-1 isnecessary for transendothelial migration of T cells.Blood 102:3675�3683.

35. Melis M, Pace E, Siena L, Spatafora M, Tipa A,Profita M, Bonanno A, Vignola AM, Bonsignore G,Mody CH, Gjomarkaj M. (2003). Biologically activeintercellular adhesion molecule-1 is shed as dimers bya regulated mechanism in the inflamed pleural space.Am J Respir Crit Care Med 167:1131�1138.

36. Millan J, Hewlett L, Glyn M, Toomre D, Clark P,Ridley AJ. (2006). Lymphocyte transcellular migra-tion occurs through recruitment of endothelial ICAM-1 to caveola- and F-actin-rich domains. Nat Cell Biol8:113�123.

37. Miller J, Knorr R, Ferrone M, Houdei R, Carron CP,Dustin ML. (1995). Intercellular adhesion molecule-1 dimerization and its consequences for adhesionmediated by lymphocyte function associated-1. J ExpMed 182:1231�1241.

38. Mohler KM, Sleath PR, Fitzner JN, Cerretti DP,Alderson M, Kerwar SS, Torrance DS, Otten-EvansC, Greenstreet T, Weerawarna K, et al. (1994).Protection against a lethal dose of endotoxin by aninhibitor of tumour necrosis factor processing. Nature370:218�220.

39. Murciano JC, Muro S, Koniaris L, Christofidou-Solomidou M, Harshaw DW, Albelda SM, GrangerDN, Cines DB, Muzykantov VR. (2003). ICAM-directed vascular immunotargeting of antithromboticagents to the endothelial luminal surface. Blood101:3977�3984.

40. Myers CL, Wertheimer SJ, Schembri-King J, Parks T,Wallace RW. (1992). Induction of ICAM-1 by TNF-alpha, IL-1 beta, and LPS in human endothelial cellsafter downregulation of PKC. Am J Physiol 263:C767�C772.

41. Nagel T, Resnick N, Atkinson WJ, Dewey CF Jr.,Gimbrone MA Jr. (1994). Shear stress selectivelyupregulates intercellular adhesion molecule-1 expres-sion in cultured human vascular endothelial cells. JClin Invest 94:885�891.

42. Oh HM, Lee S, Na BR, Wee H, Kim SH, Choi SC, LeeKM, Jun CD. (2007). RKIKK motif in the intracel-lular domain is critical for spatial and dynamicorganization of ICAM-1: functional implication forthe leukocyte adhesion and transmigration. Mol BiolCell 18:2322�2335.

43. Ostberg JR, Repasky EA. (2000). Comparison of theeffects of two different whole body hyperthermiaprotocols on the distribution of murine leukocytepopulations. Int J Hyperthermia 16:29�43.

44. Pabst R, Binns RM. (1989). Heterogeneity of lym-phocyte homing physiology: several mechanismsoperate in the control of migration to lymphoid andnon-lymphoid organs in vivo. Immunol Rev 108:83�109.

45. Pritchard MT, Ostberg JR, Evans SS, Burd R, Kray-bill W, Bull JM, Repasky EA. (2004). Protocols forsimulating the thermal component of fever: preclini-cal and clinical experience. Methods 32:54�62.

46. Reilly PL, Woska JR Jr, Jeanfavre DD, McNally E,Rothlein R, Bormann BJ. (1995). The native struc-ture of intercellular adhesion molecule-1 (ICAM-1) isa dimer. Correlation with binding to LFA-1.J Immunol 155:529�532.

47. Rosen SD. (2004). Ligands for L-selectin: homing,inflammation, and beyond. Annu Rev Immunol22:129�156.

48. Sarantos MR, Raychaudhuri S, Lum AF, StauntonDE, Simon SI. (2005). Leukocyte function-associatedantigen 1-mediated adhesion stability is dynamicallyregulated through affinity and valency during bondformation with intercellular adhesion molecule-1.J Biol Chem 280:28290�28298.

49. Schlondorff J, Becherer JD, Blobel CP. (2000).Intracellular maturation and localization of thetumour necrosis factor alpha convertase (TACE).Biochem J 347(Pt 1):131�138.

50. Shamri R, Grabovsky V, Gauguet JM, Feigelson S,Manevich E, Kolanus W, Robinson MK, StauntonDE, von Andrian UH, Alon R. (2005). Lymphocytearrest requires instantaneous induction of an ex-tended LFA-1 conformation mediated by endothe-lium-bound chemokines. Nat Immunol 6:497�506.

51. Shaw SK, Ma S, Kim MB, Rao RM, Hartman CU,Froio RM, Yang L, Jones T, Liu Y, Nusrat A, ParkosCA, Luscinskas FW. (2004). Coordinated redistribu-tion of leukocyte LFA-1 and endothelial cell ICAM-1accompany neutrophil transmigration. J Exp Med200:1571�1580.

52. Shiow LR, Rosen DB, Brdickova N, Xu Y, An J,Lanier LL, Cyster JG, Matloubian M. (2006). CD69acts downstream of interferon-alpha/beta to inhibitS1P1 and lymphocyte egress from lymphoid organs.Nature 440:540�544.

53. Sithu SD, English WR, Olson P, Krubasik D, BakerAH, Murphy G, D’Souza SE. (2007). Membrane-type1-matrix metalloproteinase regulates intracellularadhesion molecule-1 (ICAM-1)-mediated monocytetransmigration. J Biol Chem 282:25010�25019.

54. Skitzki JJ, Chen Q, Wang WC, Evans SS. (2007).Primary immune surveillance: some like it hot. J MolMed 85:1361�1367.

55. Soderberg KA, Payne GW, Sato A, Medzhitov R,Segal SS, Iwasaki A. (2005). Innate control ofadaptive immunity via remodeling of lymph nodefeed arteriole. Proc Natl Acad Sci U S A 102:16315�16320.


56. Soond SM, Everson B, Riches DW, Murphy G.(2005). ERK-mediated phosphorylation of Thr735in TNFalpha-converting enzyme and its potential rolein TACE protein trafficking. J Cell Sci 118:2371�2380.

57. Stein JV, Cheng G, Stockton BM, Fors BP, ButcherEC, von Andrian UH. (1999). L-selectin-mediatedleukocyte adhesion in vivo: microvillous distributiondetermines tethering efficiency, but not rolling velo-city. J Exp Med 189:37�50.

58. Streeter PR, Berg EL, Rouse BT, Bargatze RF,Butcher EC. (1988). A tissue-specific endothelialcell molecule involved in lymphocyte homing. Nature331:41�46.

59. Team RDC. (2006). R: A Language and Environmentfor Statistical Computing. Vienna, Austria: R Foun-dation for Statistical Computing.

60. Thompson PW, Randi AM, Ridley AJ. (2002). Inter-cellular adhesion molecule (ICAM)-1, but not ICAM-2, activates RhoA and stimulates c-fos and rhoAtranscription in endothelial cells. J Immunol169:1007�1013.

61. Thomson AW, Satoh S, Nussler AK, Tamura K, WooJ, Gavaler J, van Thiel DH. (1994). Circulatingintercellular adhesion molecule-1 (ICAM-1) in auto-immune liver disease and evidence for the productionof ICAM-1 by cytokine-stimulated human hepato-cytes. Clin Exp Immunol 95:83�90.

62. Topper JN, Gimbrone MA Jr. (1999). Blood flow andvascular gene expression: fluid shear stress as a

modulator of endothelial phenotype. Mol Med Today5:40�46.

63. Tsakadze NL, Sithu SD, Sen U, English WR, MurphyG, D’Souza SE. (2006). Tumor necrosis factor-alpha-converting enzyme (TACE/ADAM-17) mediates theectodomain cleavage of intercellular adhesion mole-cule-1 (ICAM-1). J Biol Chem 281:3157�3164.

64. Vardam TD, Zhou L, Appenheimer MM, Chen Q,Wang WC, Baumann H, Evans SS. (2007). Regula-tion of a lymphocyte-endothelial-IL-6 trans-signalingaxis by fever-range thermal stress: hot spot ofimmune surveillance. Cytokine 39:84�96.

65. von Andrian UH. (1996). Intravital microscopy of theperipheral lymph node microcirculation in mice.Microcirculation 3:287�300.

66. von Andrian UH, Mempel TR. (2003). Homing andcellular traffic in lymph nodes. Nat Rev Immunol3:867�878.

67. Walcheck B, Herrera AH, St Hill C, Mattila PE,Whitney AR, Deleo FR. (2006). ADAM17 activityduring human neutrophil activation and apoptosis.Eur J Immunol 36:968�976.

68. Webster B, Ekland EH, Agle LM, Chyou S, RuggieriR, Lu TT. (2006). Regulation of lymph node vasculargrowth by dendritic cells. J Exp Med 203:1903�1913.

69. Zhang Q, Thomas SM, Lui VW, Xi S, Siegfried JM,Fan H, Smithgall TE, Mills GB, Grandis JR. (2006).Phosphorylation of TNF-alpha converting enzyme bygastrin-releasing peptide induces amphiregulin re-lease and EGF receptor activation. Proc Natl AcadSci U S A 103:6901�6906.

This paper was first published online on iFirst on 27 December

2008.


thermal facilitation of lymphocyte trafficking involves temporal induction of intravascular icam-1

Documents