JOURNAL OF CELLULAR PHYSIOLOGY 194:54–62 (2002)

Dynamic Deformation of Migratory EfferentLymph-Derived Cells ‘‘Trapped’’ in the Inflammatory

Microcirculation

MEI SU,1,2 CHARLES A. WEST,1,2 ALAN J. YOUNG,1,2 CHUFA HE,1,2 MORITZ A. KONERDING,3

AND STEVEN J. MENTZER1,2*1Harvard Surgical Research Laboratories, Harvard Medical School, Boston, Massachusetts

2Laboratory of Immunophysiology, Dana Farber Cancer Institute, Brigham & Women’s Hospital,Boston, MA

3Department of Anatomy, Johannes Gutenberg-University Mainz, Mainz, Germany

The cellular immune response depends on the delivery of lymphocytes fromthe lymph node to the peripheral site of antigenic challenge. During theirpassage through the inflammatory microcirculaton, the migratory cells can be-come transiently immobilized or ‘‘trapped’’ in small caliber vessels. In this report,we used intravital microscopy and temporal area mapping to define the dynamicdeformation of efferent lymph-derived mononuclear cells trapped in the systemicinflammatory microcirculation. Mononuclear cells obtained from the efferentlymph draining the oxazolone-stimulated microcirculation were labeled withfluorescent dye and reinjected into the feeding arterial circulation. Intravital videomicroscopy observed thousands of cells passing through the microcirculation;35 cells were ‘‘trapped’’ in the oxazolone-stimulated microcirculation. Temporalarea maps of the trapped cells demonstrated dramatic slowing and deformation.The cells were trapped in the microcirculation for a median of 8.90 sec (range 5–17 sec) prior to returning to the flow stream. During this period, the cells showedsustained movement associated with both antegrade locomotion (mean cellvelocity¼ 7.92 mm/sec; range 1.16–14.23 mm/sec) and dynamic elongation(median cell length¼ 73.8 mm; range 58–144 mm). In contrast, efferent lymph-derived cells passing unimpeded through the microcirculation demonstratedrapid velocity (median velocity¼ 216 mm/sec) and spherical geometry (mediandiameter¼ 14.6 mm). Further, the membrane surface area of the ‘‘trapped’’ cells,calculated based on digital image morphometry and corrosion cast scanningelectron microscopy, suggested that the fractional excess membrane of the cells inthe efferent lymph was significantly greater than previous estimates of membraneexcess. These data indicate that transient immobilization of efferent lymph-derivedmononuclear cells in the systemic inflammatory microcirculation is rare. When‘‘trapping’’ does occur, the shape changes and sustained cell movement facilitatedby excess cell membrane may contribute to the return of the ‘‘trapped cells’’ intothe flow stream. J. Cell. Physiol. 194: 54–62, 2002. � 2002 Wiley-Liss, Inc.

The immune response to cellular antigen is a coordi-nated process that involves the production of migratorylymphocytes by the regional lymph node and theirdelivery into the tissues by the peripheral microcircula-tion. In the first 48 h after antigenic challenge, there isan increase in lymph node size associated with thedramatic recruitment of lymphocytes into the paracor-tical areas (T-cell zones) of the lymph node (Oort 1965;Hay, 1980). Following nodal enlargement, the celloutput of the antigen-stimulated lymph node canincrease several-fold. The high cell output is associatedwith an increase in the size of a subset of lymphocytesin the efferent lymph. Increase in lymphocyte size de-fines the so-called ‘‘lymphoblast’’ or ‘‘activation’’ phase ofthe response (Su et al., 2001; West et al., 2002a). Theseblast cells were found to ‘‘home’’ preferentially to non-

inflamed gut and skin respectively, similar to smalllymphocytes (Hopkins and Hall, 1976; Hall et al., 1977).This homing, however, appeared to be independent of

� 2002 WILEY-LISS, INC.

Abbreviations: DME, Dulbecco’s minimal essential; FEM, frac-tional excess membrane.

Contract grant sponsor: NIH; Contract grant number: HL47078.

*Correspondence to: Steven J. Mentzer, Brigham & Women’sHospital, 75 Francis Street, Boston, MA 02115.E-mail: smentzer@partners.org

Received 31 July 2002; Accepted 2 August 2002

DOI: 10.1002/jcp.10190

antigenic specificity, since no significant differencecould be found in the ability of oxazolone-specificlymphoblasts to home to inflamed versus normal skin(Hall et al., 1980). The coincidence of the activatedlymphocytes leaving the lymph node and their appear-ance in the tissue has therefore led to the speculationthat physical characteristics may contribute to lympho-cyte localization or entrapment in the inflammatorymicrocirculation (Russell et al., 1975; Rose et al., 1976;Ratner and Heppner, 1985; Melder et al., 2000, 2001).

Attempts to clarify the potential physiologic role oflymphocyte entrapment in tissue localization havefocused on the biophysical properties of lymphoid cellsin vitro. Fisher et al. (1989) have suggested that theincreased rigidity of lymphokine-activated killer (LAK)cells may be responsible the localization of the cellsafter adoptive immunotherapy. Similarly, IL-2 activa-tion of rat and human large granular lymphocytesresults in a significant increase in rigidity as definedby micropipette aspiration (Melder et al., 1994). In adifferent experimental system, Bouwens et al. (1992)demonstrated that adherent IL-2 activated murinelymphoid cells were highly deformable migratingthrough 3-mm micropores. Further, the cells appearedto be highly motile and actively crawled across endothe-lial monolayers (Bouwens et al., 1992). These intriguingin vitro results suggest that in vivo studies mightprovide useful insights into the relative frequency andfunctional importance of activated lymphocyte ‘‘trap-ping’’ in the inflammatory microcirculation.

To investigate the entrapment of activated lymphoidcells in the inflammatory microcirculation, we stimu-lated the sheep ear with the epicutaneous antigenoxazolone. Lymphocytes leaving the draining regionallymph node at the peak of lymphocyte recruitment werefluorescently labeled and reinjected into the feedingmicrocirculation. The entrapment of injected cells wasexamined by fluorescence intravital microscopy andanalyzed by digital morphometry.

MATERIALS AND METHODSAnimals

Randomly bred sheep, ranging in weight from 25 to35 kg, were used in these studies. Sheep were excludedfrom the analysis if there was any gross or microscopicevidence of dermatitis. The sheep were given free accessto food and water. The care of the animals was consis-tent with guidelines of the American Association forAccreditation of Laboratory Animal Care (BethesdaMD).

Lymph duct cannulation

In these experiments, the prescapular lymph node(Grau, 1933) was used for all efferent lymph ductcannulations (Su et al., 2001; West et al., 2002a). Aftergeneral endotracheal anesthesia and sterile surgicalpreparation, an incision was placed in the jugularfurrow 5 cm cephalad to the suprasternal notch(Lascelles and Morris, 1961; Glover and Hall, 1976).The efferent lymph duct was cannulated with a heparin-bonded polyurethane catheter (Solo-Cath, CBAS-C35;Setters Life Sciences, San Antonio, TX). The cannulawas passed through a 5-cm subcutaneous tunnel and

secured at the skin. The lymph was collected in 50 ccsterile centrifuge tubes (Falcon, Franklin Lakes, NJ).Each collection tube contained 200 IU of heparin, 2,000IU of penicillin (Cellgro, Mediatech, Inc.; Herndon, VA),and 2,000 mg of streptomycin (Cellgro).

Antigen stimulation

The sheep ear and neck was sheared bilaterally andthe lanolin removed with an equal mixture of ether(JT Baker, Phillipsburg, NJ) and ethanol (AAPER,Shelbyville, KY). The antigen, a 5 or 7% solution of2-phenyl-4-ethoxymethylene-5-oxazolone (oxazolone)(Sigma, St. Louis, MO) (Gell et al., 1946), was sprayedonto the ear and a localized region of the neck as a 4:1oxazalone:olive oil mixture using a syringe and 23 gaugeneedle (West et al., 2001a). A vehicle only control wasapplied to the contralateral skin.

Electronic cell volume

The volume profile of the efferent lymph cells wereanalyzed by an electronic cell counter (Brecher et al.,1956; Ben-Sasson et al., 1974) (Coulter Counter ZMAnalyzer; aperture, 100 mm, 1/current, 1/4; 1/gain, 1/1)connected to a pulse height analyzer (Coulter Channe-lyzer Model 256; Beckman Coulter, Miami, CA). Cali-bration of the system to obtain absolute volumes wasperformed daily using 5 and 10 mm microspheres(Coulter). Channel numbers lower than 6 correspondedto cellular debris and electronic noise. Forward lightscatter by flow cytometry was used to confirm the re-lative cell volume distributions (Mullaney and Dean,1970).

Monoclonal antibodies

The monoclonal antibodies used in these experimentswere murine monoclonal antibodies specific for thefollowing specificities: CD4 (17D) (Mackay, 1988), CD8(A51) (Ellis et al., 1986), macrophages (clone VPM64;Biosource, Camarillo, CA), CD11b (Du-87; Young,manuscript in preparation), CD40 (M2/61) (Zhao et al.,2001).

Immunofluorescence and flow cytometry

Efferent lymph cells were washed twice with phos-phate buffered saline containing 2.5% fetal calf serumand 0.02% sodium azide. Approximately 106 cells wereincubated on ice for 30 min with an excess concentrationof monoclonal antibody. The cells were washed twice andstained with fluorescein-conjugated goat F(ab0)2-anti-mouse IgG antibody (Southern Biotechnology, Birming-ham, AL) diluted 1:10. The cells were incubated on icefor another 30 min. After three washes, the cells werefixed in 1% paraformaldehyde and analyzed by flowcytometry using a Coulter Epics XL flow cytometer withExpo 2.0 software (Miami, FL). Greater than 5�105

cells were examined to exclude the possibility of granu-locytes detected by side scatter. The flow cytometry datawas collected at room temperature and exported to theMicrosoft Excel (Redmond, WA) spreadsheet for dataanalysis using WinList 4.0 (Verity, Topsham, ME). Theflow cytometry experiments were calibrated daily usingSphero Rainbow Calibration Particles (SpheroTech,Libertyville, IL) (Su et al., 2001).

LYMPHOCYTE ENTRAPMENT IN MICROCIRCULATION 55

Fluorescent cell tracers

The lymph cells were labeled with succinimidylesters of the mixed isomer preparation of 5-(and-6)-carboxytetrmethylrhodamine (5(6)-TAMRA)(ex540nm/em 565 nm; Molecular Probes, Eugene, OR) (West et al.,2001b). The dye was aliquoted as a 0.1 M solution inanhydrous dimethylsulfoxide and stored at�708C. Priorto labeling, the lymph cells were washed three timesin Dulbecco’s Modified Eagle’s Medium (DME) with2,000 mg/L glucose (Sigma, St. Louis, MO) and resus-pended in phosphate buffered saline (PBS) containing25 ml of the stock 5(6)-TAMRA fluorescent dye. The cellswere incubated for 15 min at room temperature andwashed in cold DME medium. The cells were resus-pended in room temperature PBS at 0.7–5.0�107 cells/ml prior to injection of the 5 ml over a 12 sec injectionperiod.

Intravital microscopy system

The intravital microscopy system was describedelsewhere (West et al., 2001b). The basic microscopebody was the Nikon sub-stage metallurgical/industrialmicroscope fitted with a CF infinity corrected opticalsystem with custom extenders to accommodate theadditional length of the microscope window (MicroVideoInstruments, Arrow, MA). The microscopy system ismounted on floor-based adjustable height microscopestand with rotational axes in two directions to permitgentle apposition to the ear. The custom-designed epi-illumination system had a super high pressure mer-cury lamphouse for the delivery of light through theoptical system as bright-field, dark-field, or fluore-scence illumination. The fluorescent filter block usedin these experiments was an orange (DM 560 nm) filter.The Nikon epi-achromat objectives were 10 and 20�magnification.

The intravital microscopy window, machined fromtitanium (MicroSurg, Boston, MA), provided a vibra-tion-free stage for videomicroscopy. The windows 0.5900

internal diameter cylinder was designed to accommo-date the Nikon optical system. The tissue surface ofthe window was composed of a 0.63500 flange surround-ing the window lens. Within the flange, two concentric2.5-mm vacuum galleries provided tissue appositionto the lens surface without compression of the tissueand with minimal circulatory disturbances. Thewindow lens (1.27500 � 0.05900) was custom-designedwith ultra-violet grade fused silica (Kreischer Optics;McHenry, IL). The surfaces were optically polished toretical quality (20–10 scratch-dig per MIL-0-13830A).Mounted to the microscope window was a custom-designed machined acrylic stage that supported theventral surface of the ear and insured stable tissueapposition.

The camera was a Dage-MTI CCD-72 series highresolution CCD imager and high performance analogprocessor with 768� 493 active elements and 570 TVLresolution (Dage-MTI, Michigan City, IN). The micro-scope light level camera gain and offset were set to usethe full dynamic range of the video analog contrastsettings. The image was intensified using a GenIIsysoptically coupled image intensifier with >106 dynamicrange (Dage-MTI).

Image analysis

The intravital videomicroscopy images were recor-ded on a Panasonic model AG-675OA S-VHS videorecorder (Secaucus, NJ) (30 frame/sec) with horizontalresolution of 400 lines. Time base correction was per-formed using a TBC III board (VT2500, Digital Proces-sing Systems, Florence, KY). Video of the recordedimages was processed through a M-Vision 1000 PCI busframe grabber (Mutech, Waltham, MA) in a Pentium III(700 MHz, 256 MB RAM) computer running theMetaMorph Imaging System 4.0 (Universal Imaging,Brandywine, PA) under Microsoft Windows NT(Redmond, WA). Image stacks were routinely createdfrom 12 sec to 5 min video sequences. The imagestacks were processed with standard MetaMorphfilters. After routine distance calibration and thres-holding, the ‘‘stacked’’ image sequence was measuredusing the MetaMorph’s object tracking and integratedmorphometry applications.

Tracking cell movement

Fluorescence labeling permitted the identification ofthe relevant cells for tracking. After routine distancecalibration, Metamorph object tracking (UniversalImaging; Brandywine, PA) was used to determine theX-Y centroids of the lymphocytes and track theirdisplacement through the planes in the source imagestack (Li et al., 2001). For displacement reference, thealgorithm used the location of the cell at its first positionin the track. Each cell was imaged as a disc coveringmany pixels. The disc was imaged with high contrastand its position was determined with sub-pixel accu-racy. The image of the cell was tracked using a cross-correlation centroid-finding algorithm to determinethe best match of the cell position in successive images.The resulting measures included the X and Y coordi-nates, velocity of the cell.

Distance measurement

Distance measurements were performed using digi-tal images obtained from the fluorescence intravitalmicroscopy system. The 8-bit grayscale images werethresholded and standard distance calibration wasperformed. The MetaMorph Imaging System 4.0 (Uni-versal Imaging, Brandywine, PA) caliper applicationwas used to cell length and diameter. The data waslogged into Microsoft Excel 2000 (Redmond, WA) bydynamic data exchange. Cell width was arbitrarilydefined as the transverse diameter.

Temporal area map

The timelapse acquisition of time-corrected and dis-tance calibrated images was performed as previouslydescribed (Li et al., 1996, 2001). Briefly, the videosequences of cells ‘‘trapped’’ in the microcirculation wereacquired at a time-corrected video rate (30 fps) or bydigital streaming acquisition (30–60 images/sec). Thesubsequent images were analyzed at various timelapseintervals. The timelapse image ‘‘stacks’’ were thre-sholded using standard MetaMorph (Universal Ima-ging, Brandywine, PA) filters and binarized to facilitatetimebased color encoding. The selected pseudocoloredimages were overlaid and digitally combined using the

56 SU ET AL.

MetaMorph Image System (Brandywine, PA) to renderthe temporal area map.

Corrosion casting

After systemic heparinization with 750 Iu/kg intra-venous heparin, the external auricular arteries werebilaterally cannulated and perfused with approxi-mately 100 cc of 278C saline followed by a 2.5% bufferedglutaraldehyde solution (Sigma) at pH 7.40. The castswere made by perfusion of the ear arteries with 100 ccof a Mercox (SPI, West Chester, PA) diluted with20% methylmethacrylate monomers (Aldrich Chemical,Millwaukee, WI). After complete polymerization, theears were harvested and macerated in 5% potassiumhydroxide followed by drying and mounting for scanningelectron microscopy. The microvascular corrosion castswere imaged after coating with gold in Argon atmo-sphere with a Philips ESEM XL 30 scanning electronmicroscope (Eindhoven, The Netherlands).

Fractional excess membrane

The fractional excess membrane (FEM) is calculatedby the formula FEM¼ (Sc�Ss)/Ss, where Sc is thecalculated surface area of the trapped cell and Ss isthe surface area needed to cover a smooth sphere of thevarious efferent lymph cell diameters.

Statistical analysis

The migratory data was based on multiple compar-isons of paired data by Student-Newman-Keuls orMann–Whitney test for non-parametric analysis ofvariance. The data was expressed as meanþ one stan-dard deviation. The significance level for the sampledistribution was defined as P<0.05.

RESULTSEfferent lymph-derived migratory cells

The efferent lymph draining the oxazolon estimulatedmicrocirculation was the source of fluorescently labeledmigratory cells. The migratory cells were obtained atthe peak of lymphocyte recruitment: 96 h after oxazo-lone stimulation (West et al., 2001a). The size of thelymphocytes, measured by Coulter Counter impedence,demonstrated that lymphocytes obtained from theoxazolone-stimulated lymph were larger than cells ob-tained from the contralateral control (Fig. 1) (P< 0.05;Student’s t-test). The composition of the oxazolone-stimulated lymph-derived cells was 80% T lymphocytes,9% B lymphocytes, 11% monocytes/macrophages, andno detectable granulocytes (Fig. 2).

Movement of trapped cells

The efferent lymph-derived cells obtained 96 h afteroxazolone stimulation were labeled and sequentiallyinjected into the oxazolone and contralateral controlcirculation. Intravital videomicroscopy observed thou-sands of migratory cells passing through the micro-circulations; 35 cells were transiently immobilized inthe oxazolone-stimulated microcirculation. No entrap-ment was observed in the control microcirculation.Temporal area maps were used to characterize thepotential movement of the entrapped cells. The tem-poral area maps were composed of digital images com-

bined at 33 msec intervals and pseudocolored to reflecttime encoding (Fig. 3). Despite the appearance ofbeing immobilized, the entrapped cells demonstratedsustained antegrade locomotion (mean cell velocity¼7.92 mm/sec; range 1.16–14.23 mm/sec). The cells weretrapped in the microcirculation for a median of 8.90(range 5–17 sec) sec prior to returning to the flow stream(Fig. 4). All ‘‘trapped’’ cells returned to the flow stream.In contrast, efferent lymph-derived cells passing thro-ugh the microcirculation demonstrated rapid medianvelocity (median velocity¼ 216 mm/sec).

Shape change of trapped cells

To characterize the shape change of the ‘‘trapped’’cells, digital images were combined at 660 msec inter-vals and pseudocolored. These timelapse temporal area

Fig. 1. Size distribution of migratory efferent lymph-derived cellsleaving the oxazolone-stimulated and control lymph nodes 96 h afterstimulation (N¼8 sheep). A: The histogram represents the meanCoulter Counter electronic impedance measurement of cell volumesfrom the oxazolone-stimulated (solid line) and contralateral control(dashed line) efferent lymph (5–6�105 cells per measurement).B: Based on the measured cell volumes and the assumption thatthe efferent lymph cells were a polydispersed system of spheres, thediameters of the cells were calculated and represented as segmenteddata. The mean percentage of cells in the segmented histogram isshown for oxazolone-stimulated (solid bars) and control (white bars)lymph. Error bars reflect one standard deviation based on eight sheep.

LYMPHOCYTE ENTRAPMENT IN MICROCIRCULATION 57

maps of the ‘‘trapped’’ cells demonstrated active defor-mation and elongation (Fig. 5). The ‘‘trapped’’ cells hada median cell length of 73.8 mm (range 58–144 mm).In contrast, cells passing through the microcirculationdemonstrated a spherical geometry with a mediandiameter of 14.6 mm (Fig. 6).

Fig. 2. Cell surface phenotype of the efferent lymph-derived cells 96 h after oxazolone stimulation ina representative sheep. Flow cytometry histograms comparing the negative control (A) with themacrophage marker VPM64 (B), the T cell molecules CD4 (C), and CD8 (D), phagocyte and B cell markerCD11b (E) and the B cell activation marker CD40 (F). There were no detectable granulocytes (samplesize¼500,000 cells).

Fig. 3. Temporal area map demonstrating the velocity of a singlerepresentative cell ‘‘trapped’’ in the oxazolone-stimulated microcircu-lation 96 h after stimulation. A: The 9 sec image stack (271 frames)was digitally combined and pseudocolored to reflect the crawlingmovement of the cell during this time period. The mean instantaneousvelocity of the leading edge of this cell over the 9 sec interval was7.36 mm/sec. B: A single frame obtained 8.3 sec after the appearance ofthe cell shows that the apparent length of the cell exceeded 86 mm.

Fig. 4. The length of time migrating cells were tracked by intravitalmicroscopy is shown for migrating cells passing through the micro-circulation (circles) and cells ‘‘trapped’’ in the microcirculation(squares). The spherical migratory cells were tracked for a mediantime of 0.5 sec (trendline R2¼ 0.043). The ‘‘trapped’’ cells wereobserved for a median time of 7.9 sec (trendline R2¼ 0.032). Becausethe velocity calculations were obtained from paired measurements ofrapidly moving and trapped cells in the same temporal sequence, onlya subset of the elongate or trapped cells are shown.

58 SU ET AL.

Fractional excess membrane

The dynamic deformation of the ‘‘trapped’’ cells sug-gested an unexpected excess of cell membrane. To obtainan estimate of the cell membrane surface area (andfractional excess membrane) we measured: (1) thevolume of the injected cells, (2) the length of the trappedcells, and (3) the diameter of the trapping microvessels.Although the cell volume (Coulter Counter impedance)and cell length (intravital microscopy) were readilymeasured, the microvessel dimensions required corro-sion cast injections and examination by scanning andtransmission electron microscopy. The examination of58 afferent capillary segments in the oxazolone-stimu-lated microcirculation demonstrated a median diameterof 12 mm. Quantitative morphometry of both afferentand efferent microvascular segments in five sheep sug-gested a range of microvascular diameters from 10 to18 mm (Fig. 7).

Using these data, the predicted membrane surfacearea of cylindrical cells of varying lengths trapped inmicrovessels of 10, 15, and 25 mm diameter was calcu-lated. The corresponding calculated surface area of anentrapped cell of length 73.8 mm is shown by the dottedlines in Figure 8A. Assuming each portion of the cellvolume histogram (Fig. 1) is potentially the source of theelongate or entrapped cells, the necessary amount ofexcess membrane was calculated for efferent lymph cellswith original cell diameters of 5, 10, 15, 20, and 25 mm.The fractional excess membrane (FEM) was calculatedfor a cell 73.8 mm in length and trapped in a 15 mmmicrovessel ranged from 2- to 53-fold (Fig. 8B).

DISCUSSION

In this report, we directly observed the frequency andfunctional consequence of migratory cells ‘‘trapped’’ inthe inflammatory microcirculation. To provide anoperational definition of antigen-activated migratorycells, we obtained cells from the efferent lymphaticsdraining the tissue stimulated by the epicutaneousantigen oxazolone. These efferent lymph-derived cellswere fluorescently labeled and re-injected into the sameoxazolone-stimulated microcirculation. The movement

Fig. 5. Temporal area map showing the deformation of a singlerepresentative cell ‘‘trapped’’ in the oxazolone-stimulated microcircu-lation 96 h after stimulation. The cell movement appears to reflectdeformation in a microvascular loop. The shape change of the cell wasobserved with 216 frames over a 7.2 sec interval. The color-codedoverlays show the cell shape and location at arbitrary 660 msecintervals.

Fig. 6. Cell dimensions of migrating and ‘‘trapped’’ cells in theoxazolone microcirculation. Lymphocytes passing through the micro-circulation (solid circles) or transmigrating into extravascular tissue(open circles) are compared to ‘‘trapped’’ cells (squares). Thelymphocytes passing through the microcirculation were spherical(trendline R2¼0.811), whereas ‘‘trapped’’ cells were elongate (trend-line R2¼ 0.0013). To ensure comparable microhemodynamic andoptical conditions, a data is presented as paired measurements ofrapidly moving and trapped cells. Because the measurementsobtained in different regions of the superficial vascular plexus, butin the same temporal sequence, but the only a representative subset ofelongate or trapped cells are shown.

Fig. 7. Scanning electron microscopy of corrosion casts of theoxazolone-stimulated and control microcirculations. A: Photomicro-graph of microvascular loops obtained from the control microcircula-tion. B and C: Photomicrographs of microvascular loops obtained afterantigen stimulation with the epicutaneous antigen oxazolone. Afterarea calibration, the caliper application of the MetaMorph softwarewas used for microvessel diameter measurement (distance measuresin mm; Bar¼25 mm).

LYMPHOCYTE ENTRAPMENT IN MICROCIRCULATION 59

of the labeled cells was observed by intravital micro-scopy and analyzed by digital morphometry.

These studies suggest several conclusions. First, theimmobilization or entrapment of cells in our experimen-tal system was rare. Thousands of cells were observedin the inflammatory and control microcirculation, butonly 35 were entrapped in the inflammatory micro-circulation. No cells were entrapped in the contralateralcontrol microcirculation. Although physical entrapmentmay play a role in granulocyte sequestration (Hogget al., 1992; Doerschuk et al., 1993; Wiggs et al., 1994;Kitagawa et al., 1997) and lymphokine-activated lym-phocyte localization (Fisher et al., 1989) in the lung,

lymphocyte entrapment does not appear to play asignificant role in the delivery of lymphocytes to theoxazolone-stimulated skin.

Second, the phenomenon of cell ‘‘trapping’’ appearsto be less dependent on the migratory cells than onthe inflammatory microcirculation. Both the controland inflammatory microcirculation received identicalinjections of fluorescently labeled cells. The injectedlymphocyte populations had the same biophysicalcharacteristics (e.g., size and rigidity), yet entrapmentwas only observed in the inflammatory microcirculation.The explanations for the differential frequency of lym-phocytes ‘‘trapped’’ by the microcirculation may include:(1) inflammation-associated changes in the adhesivityof vascular lining cells (Li et al., 2001); (2) inflammation-dependent narrowing of the lumen of pre-existingmicrovessels (Doerschuk et al., 1993); and (3) theinflammation-induced recruitment of previously under-perfused microvessels (West et al., 2002b). It is possiblethat all of these mechanisms may participate in the in-creased frequency of cellular entrapment in the inflam-matory microcirculation.

Third, the entrapped migratory cells demonstratedsustained antegrade locomotion that appeared to facil-itate their release into the flow stream. Entrapmentoccurred whencells, apparentlymoving normally withinthe flow stream, entered a previously unappreciatedvascular segment and rapidly decelerated. The rapiddeceleration created the visual impression of mechan-ical impaction. Despite this abrupt deceleration, thecells continued antegrade locomotion at a medianvelocity of 7.9 mm/sec. The sustained movement of theentrapped cells was consistent with an intrinsic capacityof the cells to perform active locomotion (Keller et al.,1977). The velocity of 7.9 mm/sec was considerablygreater than the 5–30 mm/min leukocyte velocitiescommonly observed in vitro (Haston and Wilkinson,1988)—a difference that may reflect the locomotoryadvantages of a three-dimensional cellular substratum(Haston et al., 1982). In the setting of inflammation,active cellular locomotion may function to preventflow occlusion and maintain lymphocyte delivery to thetissues. It remains possible, however, that the ante-grade cell movement, observed in these studies, wasa result of local hydrostatic forces, and not intrinsiclocomotory capacity.

Fourth, the dynamic elongation of the entrapped cellsrevealed an extraordinarily large membrane surfacearea. Based on measured cell length and microvesseldiameters, the surface area of the entrapped cells wascalculated to be up to 60-fold greater than the predictedsurface area of the injected cells. Although the in vivodesign of the experiments precluded the phenotypicidentification of the individual entrapped cells, themembrane excess was noteworthy regardless of thesubpopulation represented by the elongated cells.Lymphocytes, the majority of the injected cells, havemicrovilli that provide a source of excess membrane.Attempts to quantify the surface area of microvilli usingelectron microscopy have been limited by the sensitivityof microvilli to temperature and activation state (Linet al., 1973; Alexander and Wetzel, 1975); nonetheless,most studies suggest that microvilli comprise only anadditional 20–40% of membrane surface area (Polliack

Fig. 8. Calculation of cell membrane surface area of the ‘‘trapped’’cells and the calculated FEM based on efferent lymph cell volumes.A: The predicted membrane surface area of cylindrical cells of varyinglengths trapped in microvessels of 10, 15, and 25 mm diameter isshown. The corresponding calculated surface area of the observed‘‘trapped’’ cells of length 73.8 mm is shown in dotted lines. B: Assumingeach segment of the cell volume histogram (Fig. 1) is potentially thesource of the elongate or ‘‘trapped’’ cells, the necessary amount ofexcess membrane is shown for efferent lymph cells with original celldiameters of 5, 10, 15, 20, and 25 mm. The fractional excess membrane(FEM) is shown for a cell 73.8 mm in length and trapped in amicrovessel of 10 mm (white bar), 15 mm (gray bar), and 18 mm (blackbar) diameter. FEM is calculated by the formula FEM¼ (Sc�Ss)/Ss,where Sc is the calculated surface area of the trapped cell and Ss is thesurface area needed to cover a smooth sphere of the various efferentlymph cell diameters.

60 SU ET AL.

et al., 1973, 1975; Newell et al., 1976). Using osmoticmanipulations, Schmid-Schonbein et al. (1980) esti-mated the average excess membrane of lymphocytes at1.3 and monocytes at 1.37.

The potential explanations for the discrepancy be-tween these earlier estimates of membrane excess andthe membrane area observed in our studies are twofold:(1) The migratory cell leaving the antigen-stimulatedlymph node may have a membrane excess that has beenconsiderably underestimated by studies of peripheralblood lymphocytes. The possibility of active membranesynthesis in the antigen-stimulated lymph node issupported by biochemical studies of the efferent lymph(West et al., 2002); (2) The entrapped cells mayrepresent a previously unappreciated and membrane-rich subpopulation in the efferent lymph. The dendrite-like veiled cells, previously documented in the afferentlymph, are an example of a cell population withmembrane excess (Knight, 1984; McKeever et al.,1992); however, these cells have not been convincinglydemonstrated in the efferent lymph (Haig et al., 1999).Nonetheless, several studies have suggested the exis-tence of a rare population of efferent lymph cells capableof differentiating into macrophage-like cells. The stud-ies in rodents have morphologically described theexistence of a ‘‘dendrite-like’’ population in thoracicduct lymph (Anderson et al., 1981; Demartini et al.,1983; Kuznetsov, 2000).

Finally, our observations support earlier speculationssuggesting a mutually exclusive relationship betweenleukocyte entrapment and transmigration. Gaehtgens(1984) has speculated, based entirely on theoreticalgrounds, that entrapment and deformation of leuko-cytes in the microcirculation would preclude their trans-migration. The reasoning was based on the observationthat transmigrating cells develop pseudopods that faci-litate their transendothelial movement. In contrast, the‘‘trapped’’ cell does not have sufficient excess membraneto facilitate transmigration. The Gaeghtgens hypothesisis intuitively plausible, yet difficult to directly test.Nonetheless, our data is entirely consistent with theGaeghtgens hypothesis: none of the 35 ‘‘trapped’’ cellsobserved in our experiments migrated into the extra-vascular tissue. All of the entrapped cells observed inour experiments returned to the flow stream.

LITERATURE CITED

Alexander EL, Wetzel B. 1975. Human lymphocytes: Similarity ofB and T cell surface morphology. Science 188:732–734.

Anderson AO, Warren JT, Gasser DL. 1981. Presence of lymphoiddendritic cells in thoracic duct lymph from lewis rats. TransplantProc 13:1460–1568.

Ben-Sasson S, Patinkin D, Grover NB, Doljanski F. 1974. Electricalsizing of particles in suspensions. IV. Lymphocytes. J Cell Physiol84:205–214.

Bouwens L, Narayani I, Wisse E. 1992. High deformability andmotility of lymphokineactivated killer cells in vitro and in vivo.J Leukoc Biol 51:214–219.

Brecher G, Schneiderman M, Williams GZ. 1956. Evaluation ofelectronic red blood cell counter. Am J Clin Pathol 26:1439–1449.

Demartini JC, Fiscus SA, Pearson LD. 1983. Macrophages in efferentlymph of sheep and their role in lectin-induced lymphocyteblastogenesis. Int Arch Aller A Immunol 72:110–115.

Doerschuk CM, Beyers N, Coxson HO, Wiggs B, Hogg JC. 1993.Comparison of neutrophil and capillary diameters and their relationto neutrophil sequestration in the lung. J Appl Physiol 74:3040–3045.

Ellis JA, Baldwin CL, MacHugh ND, Bensaid A, Teale AJ, GoddeerisBM, Morrison WI. 1986. Characterization by a monoclonal antibodyand functional analysis of a subset of bovine T lymphocytes thatexpress BoT8, a molecule analogous to human CD8. Immunology58:351–358.

Fisher B, Packard BS, Read EJ, Carrasquillo JA, Carter CS, TopalianSL, Yang JC, Yolles P, Larson SM, Rosenberg SA. 1989. Tumorlocalization of adoptively ransferred indium-111 labeled tumorinfiltrating lymphocytes in patients with metastatic melanoma.J Clin Oncol 7:250–261.

Gaehtgens P. 1984. Deformation and activation of leukocytes—Twocontradictory phenomena? In: Meiselman HJ, Lichtman MA,LaCelle PL, editors. White cell mechanics: Basic science and clincialaspects. Vol. 16. New York: Alan R. Liss, Inc. pp 159–165.

Gell PGH, Harington CR, Rivers RP. 1946. The antigenic function ofsimple chemical compounds: Production of precipitins in rabbits.Br J Exp Pathol 27:267–286.

Glover DJ, Hall JG. 1976. A method for the collection of lymph fromthe prescapular lymph node of unanaethetized sheep. Lab Anim10:403–408.

Grau H. 1933. Die lymphgefasse der haut des schafes (Ovis aries).Ztschr Anat Entw-Gesch 101:423–448.

Haig DM, Hopkins J, Miller HR. 1999. Local immune responses inafferent and efferent lymph. Immunology 96:155–163.

Hall JG, Hopkins J, Orlans E. 1977. Studies on the lymphocytes ofsheep. III. Dstination of lymph-borne immunoblasts in relation totheir tissue of origin. Eur J Immunol 7(1):30–37.

Hall JG, Hopkins J, Reynolds J. 1980. Studies of efferent lymphcells from nodes stimulated with oxazolone. Immunol 39:141–149.

Haston WS, Wilkinson PC. 1988. Visual methods for measuringleukocyte locomotion. Methods Enzymol 162:17–38.

Haston WS, Shields JM, Wilkinson PC. 1982. Lymphocyte locomotionand attachment on two-dimensional surfaces and in three-dimen-sional matrices. J Cell Biol 92:747–752.

Hay JB, Johnston MG, Vadas P, Chin W, Iasckutz T, Movai HZ. 1980.Relationships between changes in blood flow and lymphocytemigration induced by antigen. Monogr Allergy 16:112–125.

Hogg JC, Doerschuk CM, Wiggs B, Minshall D. 1992. Neutrophilretention during a single transit through the pulmonary circulation.J Appl Physiol 73:1683–1685.

Hopkins J, Hall JG. 1976. Selective entry of immunoblasts into gutfrom intestinal lymph. Nature 259(5564):308–309.

Keller HU, Wilkinson PC, Abercrombie M, Becker EL, Hirsch JG,Miller ME, Ramsey WS, Zigmond SH. 1977. A proposal for thedefinition of terms related to locomotion of leucocytes and othercells. Cell Biol Int Rep 1:391–397.

Kitagawa Y, VanEeden SF, Redenbach DM, Daya M, Walker BAM,Klut ME, Wiggs BR, Hogg JC. 1997. Effect of mechanical defor-mation on structure and function of polymorphonuclear leukocytes.J Appl Physiol 82:1397–1405.

Knight SC. 1984. Veiled cells–‘‘dendritic cells’’ of the peripherallymph. Immunobiology 168:349–361.

Kuznetsov AV. 2000. Dendritic (Langerhans) cells in central lymph ofrabbits. Lymphology 33:67–68.

Lascelles AK, Morris B. 1961. Surgical techniques for the collection oflymph from unanaesthetized sheep. Quart J Exp Physiol 46:199–205.

Li X, Abdi K, Rawn J, Mackay CR, Mentzer SJ. 1996. LFA-1 andL-selectin regulation of recirculating lymphocyte tethering androlling on lung microvascular endothelium. Am J Respir Cell MolBiol 14:398–406.

Li X, Su M, West CA, He C, Swanson SJ, Secomb TW, Mentzer SJ.2001. Effect of shear stress on efferent lymph-derived lymphocytesin contact with activated endothelial monolayers. In Vitro Cell DevBiol 37:599–605.

Lin PS, Wallach DF, Tsai S. 1973. Temperature-induced variations inthe surface topology of cultured lymphocytes are revealed byscanning electron microscopy. Proc Natl Acad Sci USA 70:2492–2496.

Mackay C. 1988. Sheep leukocyte molecules: A review of theirdistribution, structure, and possible function. Vet Immunol Immu-nopathol 19:1–20.

McKeever DJ, Awino E, Morrison WI. 1992. Afferent lymph veiledcells prime CD4þ T cell responses in vivo. Eur J Immunol 22:3057–3061.

Melder RJ, Jain RK. 1994. Reduction of rigidity in human activatednatural killer cells by thioglycollate treatment. J Immunol Methods175:69–77.

LYMPHOCYTE ENTRAPMENT IN MICROCIRCULATION 61

Melder RJ, Yuan J, Munn LL, Jain RKI. 2000. Erythrocytes enhancelymphocyte rolling and arrest in vivo. Microvasc Res 59:316–322.

Melder RJ, Kristensen CA, Munn LL, Jain RK. 2001. Modulation of A-NK cell rigidity: In vitro characterization and in vivo implicationsfor cell delivery. Biorheology 38:151–159.

Mullaney PF, Dean PN. 1970. The small angle light scattering ofbiological cells. Theoretical considerations. Biophys J 10:764–772.

Newell DG, Roath S, Smith JL. 1976. The scanning electronmicroscopy of normal human peripheral blood lymphocytes. Br JHaematol 32:309–316.

Oort J, Turk JL. 1965. A histological and autoradiographic study oflymph nodes during the development of contact sensitivity in theguinea-pig. Br J Exp Pathol 46:147–154.

Polliack A, Lampen N, Clarkson BD, De Harven E, Bentwich Z, SiegalFP, Kunkel HG. 1973. Identification of human B and T lymphocytesby scanning electron microscopy. J Exp Med 138:607–624.

Polliack A, Hammerling U, Lampen N, De Harven E. 1975. Surfacemorphology of murine B and T lymphocytes: A comparative study byscanning electron microscopy. Eur J Immunol 5:32–39.

Ratner S, Heppner GH. 1985. T cell locomotion in the tumor micro-environment. I. A collagen-matrix assay. J Immunol 135:2220–2227.

Rose ML, Parrott DM, Bruce RG. 1976. Migration of lymphoblasts tothe small intestine. II. Divergent migration of mesenteric andperipheral immunoblasts to sites of inflammation in the mouse. CellImmunol 27:36–46.

Russell RJ, Wilkinson PC, Sless F, Parrott DM. 1975. Chemotaxis oflymphoblasts. Nature 256:646–648.

Schmid-Schonbein GW, Shih YY, Chien S. 1980. Morphometry ofhuman leukocytes. Blood 56:866–875.

Su M, Young AJ, He C, West CA, Mentzer SJ. 2001. Biphasic responseof the regional lymphatics in the normal lymphocyte transfer (NLT)reaction. Transplantation 72:516–522.

West CA, He C, Su M, Rawn J, Swanson S, Hay JB, Mentzer SJ.2001a. Stochastic regulation of cell migration from the efferentlymph to oxazolone-stimulated skin. J Immunol 166:1517–1523.

West CA, He C, Su M, Secomb TW, Konerding MA, Young AJ, MentzerSJ. 2001b. Focal topographic changes in inflammatory microcircu-lation associated with lymphocyte slowing and transmigration. AmJ Physiol Heart Circ 281:H1742–H1750.

West CA, Fischer GA, Young AJ, He C, Su M, Mentzer SJ. 2002a.Biochemical changes in the efferent lymph plasma after oxazolonestimulation. Dev Comp Immunol 26:111–119.

West CA, He C, Su M, Young AJ, Swanson SJ, Mentzer SJ. 2002b.Spatial variation of plasma flow in the oxazolone-stimulatedmicrocirculation. Inflammation Res. In press.

Wiggs BR, English D, Quinlan WM, Doyle NA, Hogg JC, DoerschukCM. 1994. Contributions of capillary pathway size and neutrophildeformability to neutrophil transit through rabbit lungs. J ApplPhysiol 77:463–470.

Zhao T, He C, Su M, West CA, Swanson SJ, Young AJ, Mentzer SJ.2001. Cell adhesion molecule expression in the sheep thymus. DevComp Immunol 25:519–530.

62 SU ET AL.