Effects of phorbol esters on shape and locomotion of human blood

lymphocytes

P. C. WILKINSON1*, J. M. LACKIE2, WENDY S. HASTON3 and LAILA N. ISLAM1

1Department of Bacteriology and Immunology, University of Glasgow, Western Infirmary, Glasgow Gil 6NT, UK1Department of Cell Biology, University of Glasgow, Glasgow G12 8QQ, UK^Department of Bacteriology, University of Aberdeen, Foresterhill, Aberdeen AB9 2ZD, UK

* Author for correspondence

Summary

The effects of phorbol esters on shape change andlocomotion of human blood lymphocytes werestudied both immediately after separating thecells from blood and after overnight culture.Phorbol myristate acetate (PMA), phorbol dibu-tyrate (PDB) and related esters produced complexshape changes in lymphocytes at both times.These shapes were analysed quantitatively usingobjective measurements derived from the mo-ments of cell shapes. Immediately after removalfrom blood, many lymphocytes (20-60% of thetotal) protruded and retracted veils or spikes atmore than one point on the cell surface. Themorphology of these cells was not typical oflocomotor cells. Usually, formation of a veil wasnot followed by a contraction wave moving downthe cell, though some cells did show contractionwaves, and some moved into collagen gels orfilters. After overnight culture, a high proportion(70-80 %) of cells had changed shape in PMA andPDB. Although the shapes were still atypical, theyresembled classical locomotor morphology more

closely; veils formed at one point on the cellsurface tended to persist, and contraction wavesand constriction rings were seen in many cells.These cells moved in large numbers into collagengels or filters. Comparison of the paths traversedby PMA-treated lymphocytes in collagen gelssuggested that cells cultured in PMA for 24 hpursued more persistent paths than those in short-term culture, but the difference was not marked.We suggest that phorbol esters induce immediateshape change without inducing the complete se-quence of motor events necessary for efficientlocomotion, 'whereas after prolonged culture inphorbol esters, locomotion is more efficient, poss-ibly because phorbol esters, like other growthactivators, stimulate events during the Gi phaseof growth that are necessary for full expression oflocomotor capacity.

Key words: phorbol ester, lymphocyte locomotion,polarization.

Introduction

The earliest visible event in the initiation of leucocytelocomotion is a change from a spherical to a polarizedshape. Assays that measure the proportion of polarizedcells in a population in suspension, and the extent ofantero—posterior polarity in individual cells, have beenused to study the early events in locomotion of leuco-cytes including neutrophils (Smith et al. 1979; Kelleret al. 1983; Shields & Haston, 1985), monocytes(Cianciolo & Snyderman, 1981) and lymphocytes (Wil-kinson, 1986; Wilkinson & Higgins, 1987a,fc). Follow-Journal of Cell Science 90, 645-655 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

ing exposure to chemoattractants, the proportion ofpolarized cells in suspension and the proportion thatmove in locomotion assays arc similar. Thus shape-change assays may be useful for studying mechanismsof signal transduction in leucocyte locomotion. Phorbolesters, which activate protein kinase C, are obviously ofinterest in this context, and in this paper we describetheir effects on shape and movement of human bloodlymphocytes.

Although, using most locomotor stimulants, therelation between polarization and locomotion of leuco-cytes is clear, phorbol esters induce changes in cell

645

shape that are difficult to relate to locomotion anddiffer from one cell type to another. In neutrophils,phorbol myristate acetate (PMA) induces formationand retraction of projections all over the cell surfacewith increased pinocytosis, but with no clear antero-posterior polarity (Roos et al. 1987), and the cells showlittle displacement. In Walker carcinosarcoma cells,PMA inhibits polarization and locomotion (Keller etal. 1985). There have been several studies of thebehaviour of phorbol ester-treated human blood lym-phocytes. Patarroyo et al. (1982) observed non-spheri-cal morphologies in lymphocytes within 15 min ofadding phorbol dibutyrate and the cells rapidly formedhomotypic aggregates. They observed that after 24 and48 h of culture the cells had increased in size andcontinued to show non-spherical morphologies. Phor-bol esters also enhanced ligand-induced lateral distri-bution of membrane glycoproteins (Patarroyo &Gahmberg, 1984) and the adhesion of human Tlymphocytes to monolayers of human umbilical veinendothelium (Haskard et al. 1986).

Our interest in phorbol ester-induced shape changein lymphocytes arose during an earlier study (Wilkin-son & Higgins, 19876), in which we reported brieflythat phorbol ester-treated lymphocytes migrated intocollagen gels and showed unusual changes in shape inpolarization assays. These cells were studied bothfollowing a brief (30 min) exposure to PMA andfollowing overnight culture in its presence, the lattertreatment stimulating a higher proportion of locomotorcells than the former. We decided that a more detailedinvestigation was merited and might give insight intothe role of protein kinase C activation in the initiationof lymphocyte locomotion. Our findings are reportedhere. We began by using a polarization assay toquantify shape change in glutaraldehyde-fixed cells.However, the shape change seen following exposure toPMA proved to be more complex and variable than thetypical antero-posterior polarity seen with other loco-motor stimuli, and was difficult to score using thepolarization assay. We therefore used objectivemeasurements, derived from the moments of cellshape, and recently described in this journal (Dunn &Brown, 1986), to define this shape change more pre-cisely. Since these studies of the morphologies of fixedcells gave incomplete information about motile behav-iour, we then went on to examine the relation betweenshape change and locomotion in living cells using time-lapse videotaping of phorbol ester-treated cells insuspension and in collagen gels. The tracks of cellsmoving through collagen gels were analysed using aprocedure designed to quantify the true speed andpersistence time of moving cells (Dunn, 1983; Wilkin-son et al. 1984). This study showed that much of theshape change observed in lymphocytes immediatelyafter treatment with phorbol esters was not related to

locomotion, but that if lymphocytes were maintained inculture with phorbol ester for 24 h they showed more-typical locomotor morphologies and migrated in largenumbers into collagen gels or into micropore filters.This is consistent with our earlier finding that althoughonly a minority of lymphocytes direct from blood weremotile, activators of growth such as PHA, PPD, anti-CD3 antibody, or the Cowan strain of Staphylococcusaureus increase the proportion of cultured lympho-cytes with locomotor capacity, so that a majority of thecells show locomotor activity after 24-48 h (Wilkinson,1986; Wilkinson & Higgins, 1987a). Phorbol estershave also been reported to act as lymphocyte mitogens(Touraine et al. 1977; Abb et al. 1979; Kabelitz et al.1982) and, thus, like other growth activators, they maystimulate locomotion of cultured lymphocytes by acti-vating synthesis of proteins necessary for full ex-pression of locomotor capacity.

Materials and methods

Reagents and mediaHanks' balanced salt solution, Ca2+/Mg2+-free Hanks,RPMI-1640 (RPMI), and foetal calf serum (FCS) were fromFlow, Rickmansworth, Herts. Hanks' solution and RPMIwere buffered to p H 7 4 with morpholinopropane sulphonicacid (MOPS) (Sigma, Poole, Dorset) at 10 mM and sup-plemented with human serum albumin (lOmgml"1) (HSA,Behringwerke, Marburg, Germany) (Hanks'-HSA) asnecessary. The following phorbol esters were purchased fromSigma: phorbol 12,13-dibutyrate (PDB); phorbol 12-myris-tate 13-acetate (PMA); 4*-phorbol; 4a/-phorbol 12,13-di-decanoate; phorbol 12,13-didecanoate; phorbol 12,13,20-tri-acetate. All esters were made up as stock solutions in 10~2M-dimethyl sulphoxide (DMSO) (Sigma). H7 (1-(S isoquino-linyl sulphonyl)-2-methyl-piperazine) and W7 (/V-(6-amino-hexyl)-5-chloro-l-naphthalene sulphonamide) were fromSigma. H7 was dissolved in 2mM-HCl in water at 10~2M andW7 was dissolved in methanol at 10~ 2 M. Both were used fromfreshly made solutions. l-Oleoyl-2-acetyl glycerol (OAG) wasfrom Molecular Probes Inc., Eugene, Oregon, USA. B and Tcells were distinguished using either 'Simultest' (Becton-Dickinson, Mountain View, CA, USA; FITC anti-CD3 + phycoerythrin anti-CD19) or FITC-labelled anti-kappa and lambda light chains (combined) (Dakopatts,Glostrup, Denmark). Anti-CD3 (OKT3) was a monoclonalIgG2a antibody from Orthoklein, Raritan, NJ; it was dia-lysed before use to remove azide. PHA (purified) was fromWellcome, Dartford, Kent. Cellulose nitrate micropore fil-ters (8;Um pore size) were from Sartorius, Gottingen, Ger-many. Type I collagen was prepared freshly in the laboratoryfrom rat tail tendons as described (Elsdale & Bard, 1972;Hastoneia/. 1982).

Cell preparation and cultureLymphocytes were obtained from normal, heparinized, hu-man blood by separation on Lymphocyte Separation Medium(Flow). The cells were washed three times in Hanks'-Mops

646 P. C. Wilkinson et al.

and resuspended at 2X106 cells ml ' in Hanks'-HSA. Cellswere either used immediately far polarization and locomotionassays or cultured overnight in Hanks'-HSA in Linbro tissueculture dishes (wells, l-7cmXl>6cm; Flow). Cells werecultured at 2 x l 0 6 m r ' in lml samples for 24 h at 37°C.Control wells contained Hanks'-HSA alone. Appropriateconcentrations of phorbol esters were added, at concen-trations described in Results, to test wells. In control wellscontaining DMSO alone at appropriate dilutions, no effectwas seen on polarization or locomotion. Cells were alsocultured in anti-CD3 (aCD3; 25 ng, ml"1), a reagent that hasno short-term effect on lymphocyte locomotion, but inducesgrowth and morphological polarization of about 50-60 % oflymphocytes overnight (Wilkinson & Higgins, 1987a); or inPHA (1 ̂ gml~') , which, again, induces growth and polariz-ation only in overnight culture (Wilkinson, 1986).

Separation of T and B lymphocytesT lymphocytes were purified from the mononuclear cellfraction by passage through a nylon wool column at 37 °C.B lymphocytes and other cells adhere to the nylon wool, andT lymphocytes pass through. The B cells were then dis-lodged mechanically from the nylon wool as described byStewart (1981). The nylon wool non-adherent cells contained<1 % B cells; the adherent and dislodged cells contained65 % B cells.

Polarization assayThe proportion of polarized lymphocytes was scored either(1) immediately after purification from blood, or (2) after24 h of culture. In short-term polarization assays, lympho-cytes were incubated for 30min in plastic test tubes at 37 °Cwith the chosen reagents, then fixed by adding an equalvolume of glutaraldehyde (2-5%), centrifuged and washedwith phosphate-buffered saline. The supernatant fluid wasdecanted and the cell button (in about 100 /ul of fluid) wasused to make slide-and-coverslip preparations. A total of200-300 cells was counted and the percentage of polarizedcells estimated. FCS induced immediate shape change andwas used as a standard positive control for 30-min assays. Inovernight polarization assays, the non-adherent lymphocytesafter 24h of culture were removed, fixed and counted asabove. Note that a proportion of the lymphocytes thatremains adherent, particularly after treatment with phorbolesters, is lost from counting by this procedure. Since FCS haslittle effect on growth-induced shape change (Wilkinson,1986), aCD3 or PHA, which both activate lymphocytegrowth, were used as positive controls in 24 h assays. Scoringproportions of polarized cells was straightforward in the caseof FCS, aCD3 or PHA, all of which induced typicalfront-tail polarity, which was easily distinguishable from thenon-reacting spherical form. Shape change was more diffi-cult to score in phorbol ester-treated lymphocytes because ofthe variable morphologies discussed in detail in Results. Forthe polarization assays, cells were scored as 'spherical' or'non-spherical'. Among the spherical cells we included cellsthat were still obviously spherical with the exception of amicroprojection, 'spike', or other distortion, which did notoccupy more than about 10-15% of the cell circumference.All other deviations from sphericity were scored as non-spherical.

Locomotion assaysMicropore filter assays. Lymphocytes (2xl06ml~ ' in

Hanks'-HSA) were layered onto the upper surface of cellu-lose-ester filters of 8^m pore size (Sartorius, Gottingen,FRG) and incubated for 2h at 37°C to allow the lymphocytesto penetrate the filter. The assay has been described inextenso (Wilkinson, 1982). No gradients were set up: theconcentration of all reagents was uniform throughout thechamber. Filters were fixed and stained with haematoxylinand locomotion was scored as (1) the percentage of totaladded cells that had entered the filter, and (2) the distanceinto the filter attained by the leading front of cells.

Collagen gel assays. Collagen gels were formed by bring-ing the collagen solution back to isotonicity and neutral pHand casting gels (1-5 mg collagen ml"1) (Haston et al. 1982;Brown, 1982) either into Linbro tissue culture wells (forpopulation assays) or into metal filming chambers (Allan &Wilkinson, 1978) for analysis of the movement of individualcells. Once the gel had set, it was covered with a suspension oflymphocytes, which settled on its upper surface. For filmingindividual cells, the gels were cast into a shallow well made byattaching a glass coverslip to the lower surface of a metalfilming chamber. The lymphocyte suspension was addedwhen the gel had set, and the chamber was sealed with anupper coverslip. Lymphocytes that had penetrated and weremigrating within the gel matrix at 37 °C were filmed onvideotape using a time-lapse videorecorder system (Panasonic8050 at 12H time mode) and Nomarski optics (Olympus BH2microscope; X40 objective). Cell tracks were analysed bycovering the monitor screen with transparent paper anddotting in the position of the cell centre at intervals of 40 s.The transparent sheets were then transferred to a digitizingtablet (Summagraphics, Bit Pad 1) linked to a BBC micro-computer, and the coordinates of cell position at sequentialtimes were stored (Lackie & Burns, 1983). Values for speedand persistence were then computed for each cell using themethod proposed by Dunn (1983) and used by Wilkinson etal. (1984). The value for speed takes account of the under-estimation of the length of a smoothly curving track when it isapproximated by straight-line segments: the value for persist-ence (the tendency for a cell to continue in the direction ofprevious travel and to make only small-angle turns) is derivedfrom short-term data, and makes the assumption that the celltracks can be described as a random walk (Lackie, 1986). Thediffusion coefficient (R) for the cell (treating it as a randomlydiffusing particle) can be calculated from the speed (S) andpersistence (P) parameters, using the relationship:

R = 2.S2P.

Estimates of the mean speed and persistence for the popu-lation were calculated using the Jack-knife method of Mos-teller & Tukey (1977), which is robust when the distributionis skewed.

Shape analysisGlutaraldehyde-fixed cells were mounted on slides and theoutlines of cells were drawn using a camera lucida attachmenton a Wild M20 microscope with bright-field illumination anda X100 objective. The optical section that most nearlyrepresented the shape that could be deduced by focussing upand down was drawn. Outlines of the cells were then traced

Effects of phorbol esters on lymphocyte locomotion 647

on a digitizing tablet (as used in track analysis) with the stylusset in stream mode: approximately 50 coordinate points wereobtained for small spherical cells, rather more (up to 100) forcells with complex shapes. The coordinates of the outlinewere then used to calculate moments of zero-, first- andsecond-order, as described by Dunn & Brown (1986). Thezero-order moment is a measure of the area within theoutline, the elongation, extension and dispersion parametersare area-independent (the extension parameter is the sum ofthe elongation and dispersion parameters). Representativeshapes were kindly analysed by Dr A. F. Brown (MRCBiophysics, Kings College, London), and our values co-incided with his (within the limits of accuracy that can beachieved in tracing the outlines).

The three shape parameters (elongation, dispersion andextension) were stored for each cell, and at least 50 cells weremeasured from each experiment. Extension is a measure ofhow much the shape differs from a circle, and has a value ofzero if the shape is circular. Dispersion is a measure ofirregularity of contour and proved useful as an objectivecriterion to distinguish spherical from non-spherical cells. Byinspection it appeared that there was a bimodal distributionof shapes: some remained spherical, others did not. Thedispersion parameters for the spherical and non-sphericalcells were distinctly different; spherical cells had a dispersionof around O'Ol. Non-spherical cells usually gave a dispersionvalue of 10 times greater (or more). We used a dispersionvalue of 0'02 to separate the two classes. No outline that wejudged by eye to be non-spherical had a dispersion value aslow as this. On this basis the cells were subdivided, and theelongation and extension parameters of the spherical andnon-spherical cells were handled independently thereafter.Comparisons between values were made using Mann-Whit-ney U-tests, since the distribution of the shapes of non-spherical cells had significant skew and kurtosis, although itwas unimodal.

Results

Polarization assays: the morphological appearance offixed lymphocytes

PMA induced complex shape changes in lymphocytesin both short-term (0-30 min) and long-term (24 h)assays. Though a proportion of cells showed antero—posterior polarity, this was much more difficult todefine than with other locomotor stimuli, which inducethe classical polarized morphology (Haston & Shields,1984; Wilkinson, 1986). In short-term assays, lympho-cytes in PMA (Fig. 1A) frequently extended veils orruffles at one or more poles and these veils were oftenlarge, exceeding the projected area of the cell body onoccasion. Extensions of a variety of shapes includingspikes were seen. However, the cell body tended toremain spherical and constrictions were rare. The time-course of development of non-spherical morphology inblood lymphocytes in short-term assays is shown inFig. 2. Cells began to deviate from a spherical shapewithin 1-2 min of adding PMA and the response waswell developed within 5-15 min. This response was

temperature-dependent. At 30 min 49% of the cellswere non-spherical at 37°C; 35 % at 20°C and 0-5 % at4°C. Fig. 3 shows a comparison of dose-responses forblood lymphocytes after 30 min treatment and after24h culture in PMA. The cells responded to PMA atconcentrations down to 10~'°M (though storage of theester led to loss of activity; and one of four batches thatwe tested was only active to 3=10~ 8 M). The proportionof non-spherical forms induced by PMA in cells directfrom blood was 40-50%. After overnight culture, agreat majority (usually about 80 %) of the PMA-treatedlymphocytes were non-spherical. Many of these wereelongated and easier to score than at 30min (Fig. IB).Shape analysis (below) was used to quantify this.Constrictions were seen more commonly after 24 h inPMA than after 30 min. Overnight treatment withPMA also increased lymphocyte adhesiveness, and thiseffect was most marked in T cells. The proportion ofnon-adherent B cells in control overnight cultures was10-20%. This rose to 30-50% in PMA-treated cul-tures due to loss of adherent T cells.

The pattern of activity of various esters was studied:in both short-term and 24h cultures, PMA, phorbol12,13-didecanoate and phorbol 12,13-dibutyrate, allinduced shape change in similar proportions of cells.Phorbol triacetate, 4a'-phorbol and 4<r-phorbol dideca-noate, had no effect. This pattern of activity ofdifferent esters is similar to that seen in tumour-promoting and other systems.

Effect of phorbol esters on shape change in T-enrichedand B-enriched populations

T and B lymphocytes were separated by passagethrough nylon wool columns and tested for shapechange induced by PMA and phorbol dibutyrate. BothT and B cells changed shape in response to both estersin short-term assays (Table 1). The difference betweencontrol and test was more marked using T cells, but Bcells were tested in the presence of FCS, which itselfinduces some shape change and gives high controlvalues. In overnight assays, a majority of purified Tcells changed shape in the presence of both esters.

Analysis of the shapes of fixed lymphocytesThe shape analysis of Dunn & Brown (1986) was usedto try to overcome the difficulties in defining the shapesof glutaraldehyde-fixed lymphocytes that had beentreated in suspension with phorbol esters. For thisanalysis, we wished to select only those lymphocytesthat had responded to the phorbol ester, i.e. those thatwere non-spherical. As outlined above, the dispersionvalues proved a useful non-subjective means for mak-ing the distinction between cells with regular (e.g.spherical) and irregular outlines. All cells with adispersion greater than 0-02 were included as non-spherical. Fig. 4 shows examples of the shapes seen in

648 P. C. Wilkinson et al.

Fig. 1. Shape change inlymphocytes treated insuspension with PMA(10"9M). A. After 30min,the cells show peripheralveils and spikes, but thecell bodies remain rounded;B, after 24 h,antero-posterior polarity ismore evident andconstriction rings passingdown the cell body(arrowhead) are evident.Phase-contrast. Bar,

lymphocytes treated in different ways and illustratessome of the difficulties. Table 2 shows the measures'extension', 'dispersion' and 'elongation' for the non-spherical cells from the different populations studied.Elongation is probably the most useful of these todescribe the shape adopted by the cells that hadresponded; extension is simply the sum of elongationand dispersion. In the groups of lymphocytes treatedwith agents that induced conventional locomotor mor-phology (i.e. FCS in short-term assays; aCD3 andPHA at 24 h), the elongation value of the non-sphericalcells was around 0-9-1-0. In cells treated with phorbolesters for 30min, the values for elongation were low(0-35-0-49). However, when the same populationswere cultured overnight in PMA, the values forelongation (and extension) had increased considerably

and significantly (0-71-0-83). The low elongationvalues after 30 min exposure to PMA reflect the con-siderable deviation of these phorbol ester-treated cellsfrom the classical polarized form. After overnightculture the cells had elongated considerably and ap-proached this form more closely, though elongationvalues were still not as high as in the aCD3 and PHAcontrols. Constriction rings were frequently seen in thecell body after overnight culture.

Locomotion of PMA-treated lymphocytes: populationassays

Despite the unusual morphologies of lymphocytes inPMA, some of these cells were motile. This wasstudied both by filming and in a micropore filter assaythat measures locomotion of cell populations into a

Effects of phorbol esters on lymphocyte locomotion 649

60

50tft

8 40-

1jB 30

II 20'10

5 10 15 20 25 30Time (min)

Fig. 2. Time course of lymphocyte shape change during30min after exposure to PMA (10~9M).

1UU

80-

ical

eel

s

tI 40Z

20

/

/

4

0 10" 10-" 10"10 10"PMA(moir')

10"

Fig. 3. Dose-response curve: non-spherical cells in PMAafter 30min ( • • ) and 24h (X X). Mean±S.E.M.from five experiments.

three-dimensional matrix (Table 3). In filters, lympho-cytes cultured overnight in PMA invaded in greaternumbers and reached a greater distance than untreatedcontrols; exposure for a short period to PMA inducedlittle increase. Similar results were obtained when thenumbers of cells invading a collagen gel were counted(not shown). After short-term treatment with PMA,the number of cells migrating into filters or gels waslower (<20 %) than the number of non-spherical cellsassessed morphologically (Table 3, cf. Figs 2 and 3),suggesting that many of the latter were not, in fact,motile cells.

Shape change in living lymphocytes treated with PMAIt was evident from videotaping shape change in livinglymphocytes that the information given by studyingfixed cells was very limited. Lymphocytes, whenfreshly prepared from blood, and exposed in suspen-

sion to PMA, changed shape continuously and vigor-ously. They rapidly protruded broad veils from one ormore points, and these were often retracted equallyrapidly. In descriptions of the classical locomotorbehaviour of lymphocytes, it has been emphasized thatlamellipodium formation is usually followed by a waveof contraction that passes down the cell from front toback (Lewis, 1931; Haston & Shields, 1984). Contrac-tion waves were seen, but were uncommon, in lympho-cytes after short-term exposure to PMA (Table 4).Usually the cell body itself remained spherical withperipheral veils. Veils frequently broadened and sweptaround the cell in a circus movement. The position atwhich veils were formed was not random. Cells pro-truded veils at two poles on the cell surface, often atabout 180° from each other, more frequently than atother points (though this did not assume statisticalsignificance).

After 24h of culture in PMA, lymphocytes insuspension were behaving quite differently. The rapidprotrusion and retraction of veils was much less fre-quently seen. Rather, when a veil was protruded, ittended to persist for several minutes and contractionwaves were more often seen to spread from the point ofprotrusion than in the short-term cells (Table 4). Thiscorresponds to the 'constriction ring' frequently seen inmoving lymphocytes (Haston & Shields, 1984; andFig. IB). Occasionally, a lymphocyte would reverse itspolarity by simultaneously retracting a veil at one poleand protruding one at the other, but this behaviour wasless common after overnight culture than in cellsimmediately after treatment with PMA. In summary,as time in culture in PMA proceeded, morphologicalchanges in the cells approximated more closely toconventional locomotory behaviour (pseudopod exten-sion and formation of a contraction wave leading toforward propulsion) than immediately after exposureto this agent, at which time pseudopod extension andretraction without contraction waves were seen.

Time-lapse filming of PMA-treated lymphocytes incollagen gels

Lymphocytes that had invaded collagen gels weretracked both immediately following treatment withPMA and after 24h of culture in PMA. These cellswere moving in a three-dimensional matrix and track-ing was, of necessity, in only two dimensions, so someinformation about these tracks is lacking. Some of thePMA-treated cells invaded and moved within collagengels and many of the cells, judged by visual inspectionof the tracks, showed good persistence in a givendirection (no agent or attractant other than PMA waspresent in these gels). Immediately following treat-ment with PMA more cells in the population changeddirection sharply and frequently than in the overnightpopulation. These impressions were quantified by

650 P. C. Wilkinson et al.

Table 1. Shape change in purified T and B lymphocytes

A. Assay (30min): lymphocytes purified directly from blood

T cells in RPMI-HSAT cells in RPMI-HSAB cells in RPMI-FCSB cells in RPMI-FCS

B. Lymphocytes after overnight

T cells in RPMI-HSAT cells in RPMI-HSAUnseparated lymphocytes

Mediumalone

3-3 ±0-92,5

14-013-223

culture ± phorbol

Mediumalone

6-0 ±0-63

13, 9

% Non-spherical cells

Medium + PMA(2X1O"7M)

46-7 ±6-6

37, 25

ester

% Non-spherical cells

Medium + PMA(2X10"7M)

61-3 ±2-8

(meants.E.M.)

Medium + PDB(2xl(T8M)

52,49

36

(mean ±S.E.M.)

Medium + PDB(2xlO"8M)

7161, 58

n

3221

n

312

Fig. 4. Drawings of shapes of non-spherical cells treated: (A), for 24h with anti-CD3 antibody (control showing locomotorforms); B, for 30min with PMA ( 1 ( T 9 M ) ; and C, for 24h with PMA (1CT9M). These shapes were selected to illustrate thetypes of changes seen under the different experimental conditions, and may not be typical for the whole population. Themedian elongation values for the cells shown are: A, 114; B, 0-64; C, 0-85 (cf. values in Table 3 for the wholepopulation).

measuring the 5, P and R values (Dunn, 1983;Wilkinson et al. 1984; and Table 5). There was nosignificant difference in speed (S) between the twopopulations. Persistence (P) was shorter for short-termPMA cells than for the overnight culture group, butnot markedly so, and this was significant at the 5 %, butnot at the 1 %, level. R (diffusion rate) was notsignificantly different between the two groups. Thusafter short-term treatment with PMA, moving lympho-cytes show less persistence than when they have been inPMA overnight but the locomotion is otherwise notobviously different, though this difference might have

been expected from observations of shape change incells in suspension.

Divalent cation dependence of lymphocytepolarization

Table 6 shows that PMA-induced lymphocyte shapechange was not inhibited in either short-term orovernight assays by culture in divalent cation-freemedia+EDTA or EGTA. In the short-term assay therewas in fact some enhancement using 10~9M-PMA inthese media. In contrast, aCD3-induced polarizationwas dependent on divalent cations. Long-term culturein EDTA reduced lymphocyte viability, particularly in

Effects of phorbol esters on lymphocyte locomotion 651

Table 2.

*(<

Measures

7c)

of shapes

Extension

of lymphocytes

Median values

Dispersion Elongation

Lymphocytes suspended for 30 min in:

Positive controlFCS20%

Phorbol experiment IP M A ( 1 0 ~ 7 M )

PDB ( 1 0 ~ 7 M )

Phorbol experiment IIPMA(10~9M)PMA(10"'°M)

Lymphocytes cultured overnight in:

Positive controlsaCD3 (25ngmr')PHA (1 /igml"1)

Phorbol experiment IPMA ( 1 0 " 8 M )

Phorbol experiment IIP M A ( 1 0 " 9 M )

PMA ( 1 0 ~ ' ° M )

17

6663

3845

3272

114

5055

(29)

(57)(54)

(55)(83)

(43)(76)

(89)

(84)(88)

1-32

0-520-61

0-570-54

1-301-29

111

0-990-88

0-24

0-08

0-08

0100-09

0-230-15

0-12

0-130-15

0-93

0-410-50

0-490-35

0-991-09

0-83

0-760-71

Statistics: Mann-Whitney U-test. All 30min phorbol ester cells versus all overnight phorbol ester cells. Extension, U = 9-7; dispersion,U = 5-93; elongation, U = 9 1 ; P < 0 0 1 for all.

•Number of cells with dispersion > 0 0 2 only (i.e. non-spherical cells). The value in parenthesis is the % non-spherical cells in the totalscored population. The values for negative controls (HSA) are not included because there were too few non-spherical cells to givemeaningful results.

Table 3. Migration of PMA-treated lymphocytes into micropore filters

b Of cells withinfilter after 2h

Distance (^m) migratedby leading cell front

(2h)

Cells direct from bloodHSAPMA (10~7M)PMA (10~8M)

Cells after overnight cultureHSAPMA (10~7M)PMA (10"8M)

7-9 + 0-815-1 ± 1-311-3+1-0

13-4 ±3-038-1 ±4-841-4 ±0-9

30-1 ±2-736-9 ±3-232-9 ±3-4

32-8 ±5-479-6 ±2-879-6 ± 1 2 0

Cellulose ester filters (8/lm pore size): mean + S.E.M. for three filters.

the presence of PMA, and the percentages given in thetable are for viable cells only.

Agents other than phorbol esters that affect proteinkinase C

The response of lymphocytes to phorbol dibutyrate( 1 0 ~ 7 M ) was tested in the presence of the proteinkinase C inhibitors H7 and W7 in short-term assays.These agents at concentrations between 10~5 and10~7M had no inhibitory effect in a 30 min assay (notshown). We were unable to show shape change in

lymphocytes treated with OAG, an analogue of diacyl-glycerol, at 10-100jUgmP1 for 30min (5 experiments;details not shown), though in neutrophils it did causemorphological changes similar to those described byRoos et al. (1987). However, H. U. Keller (personalcommunication) has observed shape changes in OAG-treated human lymphocytes that are similar to thoseseen with PMA.

Discussion

Phorbol esters stimulate locomotion of lymphocytes

652 P. C. Wilkinson et al.

Table 4. Formation of contraction waves by lymphocytes exposed to PMA (10 M)

A. Number of motile cells (veils,ruffles, etc) studied

B. Number of these cells*showing contraction waves

C. Frequency of contractionwaves in B (Mean no. of wavesper cell per lOmin)

Cells immediately afterpurification

31

8(26%)

2-7

Cells after overnightculture

27

21 (78%)

3-7

•Any cell showing a contraction wave during the period of study was included. The data are derived from videotape sequences of5-30 min.

Table 5. Speed (S) persistence (P) and diffusion rate (R) of PMA-treated lymphocytes migrating withincollagen gels (mean ± S.D.)

Cells purified from blood andexposed immediately to PMA

Cells after 24 h culture in PMA

Mann-Whitney U-test: S, NS; P, < 0-05 > 0-01; R, NS.Mean (+S.D.) was calculated using the Jack-knife procedure (see Wilkinson el al. 1984, for details).

Table 6. Divalent cation dependence of lymphocyte shape change in response to PMA or anti-CD3

% Non-spherical cells in:

n

13

13

5(fimmin"1)

6-06 ±0-5

6-63 + 0-5

40-

49-

P(»)

2±5-8

3 + 8-9

R(f(m2min~'

0-82 ±0-2

1-19 + 0-3

30 min assayHSA alonePMA (10~9M)PMA (10~'°M)

24 h assayHSA alonePMA (10~9M)Anti-CD3 (25ngml"')

Hanks'-Mops( + Ca2+ and Mg2+)

5-54541

18-58754

Ca2+/Mg2+-free

+ EGTA (10"3M)

557-540

138517

Hanks'-Mops

+ EDTA (10~2M)

2-565-537-5

9-577-5*10-5

Results are from one of two similar experiments.* Viability reduced.

and also induce vigorous shape changes, many of whichare not typical of locomotor cells. In this paper we haveattempted to elucidate the relationship between shapechange and locomotion using an objective methodbased on the shape analysis described by Dunn &Brown (1986). The observations reported here suggestthat lymphocyte behaviour changed markedly withtime of exposure to phorbol esters. Immediately afterexposure, the cells were stimulated to vigorous forma-tion of pseudopodia and rapid protrusion and retrac-tion of veils was seen. However, formation of a veil wasoften not followed by the development of a contraction

wave moving down the cell body, as would be expectedfrom earlier descriptions of shape change in motilelymphocytes (Lewis, 1931; Haston & Shields, 1984).Rather, the body of the cell remained rounded, withperipheral veils. Some phorbol ester-treated cells didshow more-typical locomotor morphologies and behav-iour. Such cells were often sufficiently adhesive tocrawl across two-dimensional surfaces, which is un-usual in lymphocytes (Haston et al. 1982), and theyalso invaded collagen gels. It is not clear why phorbolester-treated lymphocytes tend to put out pseudopodiaat opposite poles. Possibly this is related to centrosoine

Effects of phorbol esters on lymphocyte locomotion 653

alignment.Inspection of the morphology of cells fixed after 24 h

in phorbol ester, did not reveal gross differences fromthe appearance at 30min. However, observation ofliving cells after 24 h at once showed that the cells werebehaving differently from cells immediately after ex-posure, and in a manner more typical of locomotorcells. Pseudopodia that protruded at one pole persisted,and were accompanied by the formation of contractionwaves. Constriction rings were commonly seen. Thesecells migrated into filters and into collagen gels in largernumbers than did those treated in the short term withphorbol ester. The elongation values suggest that, after24 h in phorbol ester, the morphology of the lympho-cytes was closer to the typical locomotor form, thoughstill more variable than that of locomotor cells stimu-lated with activators such as PHA or anti-CD3.

It would be expected that, since morphologicalchanges in the short-term phorbol ester-treated cellsare so different, the migration of these cells throughcollagen gels would also be different. In particular, theshort-term cells, which protrude and retract pseudo-pods frequently, might not persist in straight paths. Infact such differences were seen but were not marked.As expected long-term cells moved in straighter pathsthan short-term cells but the statistical significance ofthis was only marginal, possibly because the samplenumber was small. Both moved at about the samespeed. One possible explanation is that many of theshort-term cells did not display the necessary sequenceof morphological changes for locomotion, i.e. veilformation and a contraction wave, and therefore wereunable to enter the gel. Only those capable of a fulllocomotor reaction were able to penetrate the gel andwere analysed. Another possibility, which we have notexplored, is that exposure to a three-dimensional gelmatrix itself modifies cell behaviour and enhances theefficiency of locomotion.

It is not known how phorbol esters induce shapechange in lymphocytes. They have multiple effects, themost frequently studied of which involve activation ofprotein kinase C (PKC). Induction of shape changewas independent of external Ca2+ concentration, as isPKC activation, but was not detectably inhibited byPKC inhibitors. It seems probable that immediateexposure of lymphocytes to phorbol esters does notadequately stimulate the full sequence of events necess-ary for forward displacement of the cell. The situationmay be different after long-term exposure to phorbolesters, since these esters stimulate lymphocyte growth(Touraine et al. 1977; Abb et al. 1979; Kabelitz et al.1982). We have reported that activators that driveresting lymphocytes into the Gi phase of growth alsoactivate the locomotor capacity of the cells in Gi(Wilkinson, 1986; Wilkinson & Higgins, 1987a), poss-ibly by stimulating expression of new genes and

synthesis of new proteins necessary for locomotion.Thus the long-term effect of phorbol esters maylikewise be an effect on gene expression. CyclosporinA, which specifically inhibits expression of genes forcertain lymphokines in human lymphocytes, inhibitsexpression of locomotor capacity in lymphocytes cul-tured in anti-CD3 or in PHA, but not in those culturedin PMA (Wilkinson & Higgins, 19876), so the mechan-isms of locomotor activation are presumably differentin the two cases. We plan to study the relation betweenlocomotor capacity and increases in RNA and proteinsynthesis in phorbol ester-treated cells to obtain furtherinformation on this point.

This work was supported by an MRC project grant (Dr W.S. Haston and Professor P. C. Wilkinson). Dr L. N. Islam issupported by the Leukaemia Research Fund. We thank Dr A.F. Brown for help with checking the values of the shapeparameters.

References

ABB, J., BAYLISS, J. G. & DEINHARDT, F. (1979).

Lymphocyte activation by the tumor-promoting agent12-O-tetradecanoyl-phorbol-13-acetate (TPA)._7. Immun.122, 1639-1642.

ALLAN, R. B. & WILKINSON, P. C. (1978). A visualanalysis of chemotactic and chemokinetic locomotion ofhuman neutrophil leucocytes. Expl Cell Res. I l l ,191-203.

BROWN, A. F. (1982). Neutrophil granulocytes: adhesionand locomotion on collagen substrata and in collagenmatrices. J . Cell Sci. 58, 455-467.

CIANCIOLO, G. J. & SNYDERMAN, R. (1981). Monocyteresponsiveness to chemotactic stimuli is a property of asubpopulation of cells that can respond to multiplechemoattractants. jf. din. Invest. 67, 60-68.

DUNN, G. A. (1983). Characterizing a kinesis response:time-averaged measures of cell speed and directionalpersistence. Agents Actions Suppl. 12, 14-33.

DUNN, G. A. & BROWN, A. F. (1986). Alignment offibroblasts on grooved surfaces described by a simplegeometric transformation. J. Cell Sci. 83, 313-340.

ELSDALE, T. & BARD, J. (1972). Collagen substrata forstudies on cell behaviour. J. Cell Biol. 54, 626-637.

HASKARD, D., CAVENDER, D. & ZIFF, M. (1986). Phorbolester-stimulated T lymphocytes show enhanced adhesionto human endothelial cell monolayers. J. Immun. 137,1429-1434.

HASTON, W. S. & SHIELDS, J. M. (1984). Contractionwaves in lymphocyte locomotion. J . Cell Sci. 68,227-241.

HASTON, W. S., SHIELDS, J. M. & WILKINSON, P. C.

(1982). Lymphocyte locomotion and attachment on two-dimensional surfaces and in three-dimensional matrices.J. Cell Biol. 92, 747-752.

KABELITZ, D., TOTTERMAN, T. H., GIDLUND, M.,

NILSSON, K. & WIGZELL, H. (1982). Activation ofhuman T lymphocytes by 12-O-tetradecanoylphorbol-13-

654 P. C. Wilkinson et al.

acetate: role of accessory cells and interaction withlectins and allogenic cells. Cell Immun. 70, 277-286.

KELLER, H. U., ZIMMERMANN, A. & COTTIER, H. (1983).

Crawling-like movements, adhesion to solid substrataand chemokinesis of neutrophil granulocytes. J. Cell Sci.64, 89-106.

KELLER, H. U., ZIMMERMANN, A. & COTTIER, H. (1985).

Phorbol myristate acetate (PMA) suppresses polarizationand locomotion and alters F-actin content of Walkercarcinosarcoma cells. Int. J. Cancer 36, 495-501.

LACKIE, J. M. (1986). Cell Movement and Cell Behaviour.London: Allen & Unwin.

LACKIE, J. M. & BURNS, M. D. (1983). Leucocytelocomotion: comparison of random and directed pathsusing a modified time-lapse film analysis. J. Immun.Meth. 62, 109-122.

LEWIS, W. H. (1931). Locomotion of lymphocytes. Bull.Johns Hopkins Hosp. 49, 29-36.

MOSTELLER, F. & TUKEY, J. W. (1977). Data Analysis andRegression. Reading, MA. USA: Addison-Wesley.

PATARROYO, M. & GAHMBERG, C. G. (1984). Phorbol12,13-dibutyrate enhances lateral redistribution ofmembrane glycoproteins in human blood lymphocytes.Eur.J. Immun. 14, 781-787.

PATARROYO, M., YOGEESWARAN, G., BIBERFELD, P.,

KLEIN, E. & KLEIN, G. (1982). Morphological changes,cell aggregation and cell membrane alterations caused byphorbol 12,13-dibutyrate in human blood lymphocytes.Int.y. Cancer 30, 707-717.

Roos, F. J., ZIMMERMANN, A. & KELLER, H. U. (1987).Effect of phorbol myristate acetate and the chemotacticpeptide fNLPNTL on shape and movement of humanneutrophils. jf. Cell Sci. 88, 399-406.

SHIELDS, J. M. & HASTON, W. S. (1985). Behaviour ofneutrophil leucocytes in uniform concentrations of

chemotactic factors: contraction waves, cell polarity andpersistence. J . Cell Sci. 74, 75-93.

SMITH, C. W., HOLLERS, J. C , PATRICK, R. A. &

HASSETT, C. (1979). Motility and adhesiveness in humanneutrophils. Effects of chemotactic factors. J. din.Invest. 63, 221-229.

STEWART, K. N. (1981). Separation of T and Blymphocytes by nylon fibre columns. Med. Lab. Sci. 38,123-125.

TOURAINE, J.-L., HADDEN, J. W., TOURAINE, F.,HADDEN, R., ESTENSON, R. & GOOD, R. A. (1977).

Phorbol myristate acetate: a mitogen selective for a T-lymphocyte subpopulation. J. exp. Med. 145, 460-465.

WILKINSON, P. C. (1982). Chemotaxis and Inflammation,2nd edn, pp. 198-202. Edinburgh: Churchill-Livingstone.

WILKINSON, P. C. (1986). The locomotor capacity ofhuman lymphocytes and its enhancement by cell growth.Immunology 57, 281-289.

WILKINSON, P. C. & HIGGINS, A. (1987a). OKT3-activatedlocomotion of human blood lymphocytes: a phenomenonrequiring contact of T cells with Fc receptor-bearingcells. Immunology 60, 445-451.

WILKINSON, P. C. & HIGGINS, A. (19876). Cyclosporin Ainhibits mitogen-activated but not phorbol ester-activatedlocomotion of human lymphocytes. Immunology 61,311-316.

WILKINSON, P. C , LACKIE, J. M., FORRESTER, J. V. &

DUNN, G. A. (1984). Chemokinetic accumulation ofhuman neutrophils on immune complex-coatedsubstrata: analysis at a boundary. J. Cell Biol. 99,1761-1768.

{Received 12 February 1988 - Accepted 19 April 1988)

Effects of phorbol esters on lymphocyte locomotion 655