Time-lapse microphotography was used to film the locomotory behaviour of cockroach haemocytes in vitro, and the cell tracks were analysed for speed and persistence; the percentage mobilization and the diffusion rate of the population were calculated. Haemocytes are either fast locomotor or spread moving cells, or non-motile spread or rounded cells; the first three types are plasmatocytes and their behaviour is interchangeable. Approximately 20% of the cells are motile under control conditions and there is no correlation between orthokinesis and klinokinesis. If activated haemocyte lysate supernatant (HLS), a source of components of the prophenoloxidase enzyme sequence, is added to the cell monolayer, up to 80 % of the cells switch to fast locomotor behaviour, rounding up and moving faster and for longer in straight lines. Neither heat-inactivated HLS nor zymosan supernatant, used to activate HLS, had any effect. If the chemokinins present in activated HLS are also released in vivo on haemocyte activation or during cuticular wounding, then they and the induced changes in haemocyte adhesion could contribute to haemocyte recruitment to sites of infection.

The circulating leucocytes of insects, the haemocytes, are responsible for maintaining the integrity of the body and protecting it against invasion by microorganisms and metazoan parasites; haemocytes accumulate at the site of wounds and either phagocytose invading organisms or entrap them within multicellular nodules or capsules. Localized haemocyte recruitment is thus an important component of these responses and, as with the accumulation of mammalian leucocytes at sites of infection, may result from alterations in haemocyte locomotory behaviour.

Chemokinetic and chemotactic behaviour of mammalian neutrophil leucocytes is stimulated by molecules associated with microbial invasion, either directly by bacterial formyl-peptides, or indirectly by bacterial lipopolysaccharides (LPS) and fungal cell walls such as zymosan, both of which activate complement to produce chemotactic peptides. In arthropods, LPS and the β-(1 ⟶3)-glucans derived from zymosan activate the haemolymph prophenoloxidase activation (proPOA) pathway, a sequence of enzymic reactions that leads to production of a phenoloxidase (Ashida & Sôderhâll, 1984; Ashida et al. 1983). The phenoloxidase is responsible for converting tyrosine and DOPA to melanin, the black pigment that is deposited on the surface of an invading organism trapped within a haemocytic capsule or nodule. During activation of the proPOA sequence, peptides are released (Ashida & Dohke, 1980); if changes in haemocyte locomotory behaviour occur, leading to recruitment to the capsule, these peptides might be the most likely candidates for inducing such changes.

The aim of the work described here was to examine the kinetic behaviour of haemocytes from the cockroach Periplaneta americana and to determine the effect upon this behaviour of components of the activated prophenoloxidase pathway.

Insects

Cockroaches were reared and maintained at 25 (± 1) °C on a diet of rat cake and water.

Reagents

Haemocytes were maintained in vitro in Mark’s M20 medium (K. C. Biological Ltd) supplemented with 5% (v/v) heat-inactivated foetal calf serum (FCS, Flow Laboratories) (control medium, CMS), to which a small amount of bovine serum albumin (BSA, Sigma) was also added as a control for the protein added to the cultures in the form of haemocyte lysate supernatant (HLS).

The supernatant of a 1 % (w/v) suspension of zymosan A (Sigma) in CMS, prepared by agitation and centrifugation of the mixture (Dularay & Lackie, 1985) (ZS), was filter-sterilized before use.

Preparation of cockroach haemocyte lysate supernatant and assay for phenoloxidase activity

A 200ftl sample of citrate/EDTA anticoagulant solution, pH6’5 (Lackie et al. 1985), was injected into each of 30–35 CO2-anaesthetized adult female Periplaneta and the diluted haemolymph was pooled into 6 ml of cacodylate buffer (10mM-sodium cacodylate, 0·25 M-sucrose, 0·1 M-sodium citrate, pH 7·0). The haemocyte suspension was centrifuged at 700μ for 10min, washed with 6 ml cacodylate buffer and the pellet was homogenized in 300 μA Hill’s saline containing 5 mM-calcium chloride (Ca/Hill’s) (Dularay & Lackie, 1985). The homogenate was centrifuged at 70 000 g for 30 min and the supernatant was used as the source of the components of the proPOA pathway. All HLS preparations were carried out at 4°C.

The proPOA pathway in the HLS was activated by addition of supernatant from 2 % zymosan in Ca/Hill’s in the ratio 1 vol. ZS/4 vol. HLS, for 30 min at room temperature. Phenoloxidase (PO) activity in the HLS was determined spectrophotometrically (Pye Unicam SP6) at 490 nm using L-DOPA (4 mg ml-1 in distilled water; Sôderhâll, 1981) as substrate, and the protein content of the HLS was determined by the method of Lowry et al. (1951). HLS without ZS was heat-inactivated at 56°C for 30 min.

Different dilutions of ZS-activated HLS (see below) were added to the haemocyte monolayers in CMS; concentrations of activated HLS are expressed in units of PO activity, merely to provide an indication of the relative concentrations of putative chemokinins produced during the proPOA sequence.

Time-lapse filming

Haemocytes. To prevent haemolymph coagulation, 100μl of the anticoagulant solution was injected into ethanol-swabbed, CO2-anaesthetized adult male cockroaches. Haemolymph was collected into 5·0 ml CMS or into CMS containing 1% ZS. Haemocytes were not washed by centrifugation in these experiments since they tended to clump in protein-containing medium. The cell concentration was adjusted to 2-5×105 cells ml-1, and approximately 600/11 of the cell suspension was pipetted onto the lower glass surface of a filming chamber (Wilkinson et al. 1982) ; the cells were allowed to settle and adhere for 10 min prior to filming.

Filming

Haemocytes were filmed under phase-contrast optics on a Wild M40 inverted microscope with a Bolex cine camera set to take an exposure every 40 s. Filming was carried out for approximately 3 h at 21 (± 1)°C. All cells were examined from each film sequence and the tracks from those cells moving more than 1 cell diameter were analysed. Each experiment was carried out at least twice.

Analysis

The film was analysed using an L and W Photo-Optical Data analysing projector, and the position of the nucleus of each moving cell was plotted at 10-frame intervals (i.e. at 400-s intervals in real time) to give a cell track. Only the tracks of cells moving greater than one cell diameter in 250 frames were plotted. Cell tracks with more than five but less than 25 steps were entered into a computer programme using a Summagraphics Bit Pad and digitizer (J. Lackie & Burns, 1983), and the values for cell speed (S), cell persistence in direction (P) and population diffusion (P) were calculated using a computer programme devised by J. Lackie (Wilkinson et al. 1984) based on the statistical treatment of Dunn (1983). Briefly, the square of the haemocyte displacement was calculated for single, and for overlapping double, triple, quadruple and pentuple steps for each cell track. The root mean square displacement (r.m.s.) was determined for the calculated data corresponding to each step size. The reciprocal of the r.m.s. in pm was plotted against the reciprocal of the step size in seconds and a least-squares fit of a straight line was drawn between the plotted points. An estimate of 1/S could be obtained from the slope of the line and the intercept on the horizontal axis gave an estimate of — 1/(6P). The diffusion coefficient (P) was obtained for each cell by calculation using the formula P = 2S2P. Data from the film analysis were tested using the test for skewness (Snedecor & Cochran, 1967) to determine if they were normally distributed. Since, in all cases, the data were non-parametric, the results are expressed as medians and compared using the Mann-Whitney U-test.

From these values, kinetic indices (KIs) (Dunn, 1983) that allow direct comparison between control and experimental values were calculated: orthokinetic index (OI) = log2 (Si/S); klino-kinetic index (KI) = log2(Pi/P); mobilization index (MI) = \og2(Mi/M) and the general kinetic index (GKI) = log2(P1/P), where S1, P1, M1 and P1 are the median experimental cell speed, persistence, percentage of cells moving more than one cell diameter and the population diffusion coefficient, respectively, and S, P, M and P are the median values of these parameters for the controls. Logarithms to the base 2 were used to facilitate interpretation of the indices; a doubling over control thus gave a K1 of +1, a halving gave a value of — 1, and so on (Wilkinson et al. 1984).

Motile haemocyte types

On the basis of their locomotory behaviour, Periplaneta haemocytes can be classified into two groups: (1) the motile cells (cells moving more than one cell diameter in 250 frames (167 min)) and (2) the non-motile cells. Each type can be further subdivided: the motile cells comprise spread moving cells (SM cells, moving approximately 0·87–1·5 μm min-1) and fast locomotor cells (FL cells, moving more than 1·5 μmmin-1), whilst the non-motile cells are either spread (SNM) or rounded (RNM).

FL cells are phase-bright and rounded, extending a short broad lamellipodium at the anterior region (Fig. 1). Individual FL cells may, however, spread and become SNM cells by anchoring to the substratum and, conversely, individual SNM cells may round up, become phase-bright and start moving rapidly. Thus the proportions of the locomotory cell types in the haemocyte population of Periplaneta in vitro are constantly shifting. The FL, SM and SNM cells all correspond morphologically to plasmatocytes (Lackie et al. 1985); in addition, the phase-bright FL cells that were previously classified separately as granular cells are clearly identifiable as plasmatocytes, since they are also capable of extensive spreading. The RNM cells, which are phase-dark with distinctive perinuclear vacuoles, correspond to the coagulocytes (Fig. 1).

Fig. 1.

Phase-contrast photographs from cine film of haemocytes in CMS. la, spread non-moving cells; lb, rounded non-moving cells; 2a, spread moving cells; 2b, fast locomotor cells, la, 2a and 2b are plasmatocytes. Bar, 50μm.

Fig. 1.

Phase-contrast photographs from cine film of haemocytes in CMS. la, spread non-moving cells; lb, rounded non-moving cells; 2a, spread moving cells; 2b, fast locomotor cells, la, 2a and 2b are plasmatocytes. Bar, 50μm.

Kinetic behaviour in control medium

Locomoting haemocytes move slowly on protein-coated glass (that is, in medium containing FCS and BSA); the period during which locomotion in one direction is maintained is fairly short (approximately 6 min) and the net result is that the rate at which the population diffuses through a unit area (the population diffusion coefficient, R) is low.

Since approximately 20% of the cell population comprises coagulocytes (Lackie et al. 1985), which are non-motile (RNM cells), the maximum possible percentage mobilization is 80 % ; however, under control conditions, only 24 % of the cells move more than one cell diameter during the filming period, the other cells remaining spread and non-motile.

Effect of haemocyte lysate supernatant on kinetic behaviour

Heat-inactivated HLS, prepared in the absence of zymosan, slightly reduced the speed but had no significant effect on the persistence of haemocyte locomotion compared with the control (Table 1) ; the GKI (Table 2) was identical to that of the control.

Table 1.

Kinetic behaviour of cockroach haemocytes in vitro

Kinetic behaviour of cockroach haemocytes in vitro
Kinetic behaviour of cockroach haemocytes in vitro

Zymosan supernatant in CMS plus BSA was tested to determine whether the ZS used to activate HLS was responsible for the stimulatory effect of the HLS; no significant difference from control values was found (Table 1).

However, incubation of haemocytes with different dilutions of ZS-activated HLS stimulates positive chemokinesis at every activity tested. With the higher PO activities (0-007 and 0-016 unit) both the speed and the period of persistence in one direction are increased significantly, and this has a net effect of at least quadrupling/? (Table 2; Fig. 2).

Fig. 2.

Cell tracks from cine film of cockroach haemocytes in: A, 0·007 unit of PO; and B, CMS. Ten fastest cells from each. Each dot represents the position of the nucleus at 400-s intervals of real time, and the cell outline represents the position of each haemocyte at the start of filming. Note greater step lengths in A. Bar, 50 μm.

Fig. 2.

Cell tracks from cine film of cockroach haemocytes in: A, 0·007 unit of PO; and B, CMS. Ten fastest cells from each. Each dot represents the position of the nucleus at 400-s intervals of real time, and the cell outline represents the position of each haemocyte at the start of filming. Note greater step lengths in A. Bar, 50 μm.

When comparing the kinesis parameters it must be borne in mind that these data were collected from a motile subpopulation of haemocytes; the complete effect of the activated HLS on the whole population can only be appreciated by also considering the proportion of the cells being stimulated to move. The speed distribution histogram (Fig. 3) for the HLS concentration equivalent to 0·007 unit of PO shows that there is both an overall shift towards higher speeds and an increase in the proportion of fast locomotor cells; the maximum possible percentage mobilization is almost achieved (Table 1).

Fig. 3.

Speed distribution histogram of Periplaneta haemocytes in CMS (open bars –––and), or 0·007 unit PO (stippled bars and⃜)

Fig. 3.

Speed distribution histogram of Periplaneta haemocytes in CMS (open bars –––and), or 0·007 unit PO (stippled bars and⃜)

Throughout this series of experiments the activity of the proPOA pathway in HLS has been expressed in terms of the activity of the end product, the enzyme phenoloxidase. Since PO itself is unlikely to stimulate haemocyte locomotion and the putative chemokinins are probably small molecules released during activation of proPO, the final PO activity must be taken merely as an estimate of the concentration of possible chemokinetic factors.

The two types of motile haemocyte found in this investigation (FLs and SMs) apparently correspond to the motile and sedentary cells described by Baerwald & Bousch (1970) for Periplaneta. In addition, it has been observed that the morphology and behaviour of the two types of moving haemocyte are interchangeable within the two types and also with the highly spread attached cells (SNMs). These cells all correspond to the plasmatocyte type of cell (Lackie et al. 1985), since they are granular and can form pseudopodial extensions. Furthermore, when maximally stimulated approximately 75–80% of haemocytes become motile, this proportion corresponding closely with the proportion of plasmatocytes present in the total haemocyte population of a cockroach.

The mechanics of locomotion of haemocytes from other insect species (Arnold, 1959; Davies & Preston, 1985) and from molluscs (Partridge & Davies, 1974) have been examined, and locomotion has been shown to be dependent on the polarized extension of pseudopodia. Armstrong (1979) found that the amoebocytes of the horseshoe crab Limulus polyphemus moved rapidly when rounded or ‘contracted’ but remained stationary when spread and adherent. More recently, we have found that haemocytes of Periplaneta and of the locust Schistocerca gregaria remain rounded and move very rapidly, at speeds greater than 5μm min-1, upon a relatively nonadhesive substratum such as siliconized glass (Huxham & Lackie, 1986; and unpublished data), indicating that, as with the locomotion of mammalian leucocytes in vitro (J. Lackie & Wilkinson, 1984; Wilkinson et al. 1984), there is an inverse relationship between the degree of haemocyte adhesion and spreading, and the rate of locomotion.

Comparison of the median values obtained under different experimental conditions (Table 1) reveals that there is no correlation between directional persistence and speed: higher median persistence times do not correlate with faster median speeds (Spearman’s rank correlation test: rs = 0·771, P>0·05, n = 6). However, correlation analysis of the data for individual cells within an experiment is more representative, since it avoids variations due to different experimental conditions. Examination of the correlation between speed and persistence time for individual cells in control medium indicates that orthokinesis and klinokinesis are also not associated (rs = 0·113, P> 0·05, n = 93). Thus the fastest cells do not always move in the straightest paths.

However, when activated HLS, a crude mixture containing enzymes and peptides from the proPOA pathway, is added to the haemocyte cultures, positive chemokinetic behaviour is stimulated. A greater proportion of haemocytes are stimulated to move; these cells move faster and with greater persistence time, and the increased speed and persistence are correlated for individual cells in activated HLS (rs = 0·335, P< 0 05, n = 84). This behaviour ensures that an individual cell scans a larger area in a shorter period of time and, if this behaviour were to occur in vivo, would increase the chance of a cell contacting a site of invasion.

Intracellular production of PO occurs when haemocytes contact a foreign surface in vivo, indicated by the deposition of melanin at the core of capsules and nodules (Grimstone et al. 1967; Lackie et al. 1985; Gagen & Ratcliffe, 1976), and occurs in vitro when haemocytes are stimulated by the addition of β-(l⟶ 3)-glucans (Leonard et al. 1984; Huxham & Lackie, 1986) or lipopolysaccharides (Smith & Sôderhâll, 1983). The cuticle of arthropods also contains a proPOA, which is activated when the cuticle is damaged (Unestam & Ajaxon, 1976). If the chemokinins resulting from activation of the proPOA pathway in HLS are also released from the wounded cuticle or activated haemocytes in vivo they would be released at an appropriate time and place to recruit haemocytes.

Since haemocytes circulating in the haemolymph cannot show chemokinetic or chemotactic behaviour and chemotactic gradients cannot be maintained, recruitment must also depend on changes in haemocyte adhesiveness that would permit cellsubstratum and cell—cell contact. Once contact is established, changes in locomotory behaviour, leading to haemocyte accumulation at the site of invasion, can occur. A process comparable to the margination of mammalian leucocytes (Forrester & Lackie, 1984; Brown, 1982) has been observed in locusts, in which haemocytes adhere to and accumulate on the subepithelial ‘basement membrane’ below the site of cuticular invasion by the fungus Metarhizium anisopliae, even though the fungus has not yet penetrated the epithelium (S. Gunnarsson, personal communication); whether or not these changes in adhesion are accompanied by changes in kinetic behaviour is at present unknown.

In the work described in this paper, a mixed population of haemocytes comprising both coagulocytes and plasmatocytes was used, and thus effects of one cell type upon the behaviour of the other are unknown. Now that techniques are available for separating cell types by density centrifugation (Sôderhâll & Smith, 1983 ; Huxham & Lackie, 1986; Cook et al. 1985; Mead et al. 1985), it is possible to investigate the behaviour of the different cell types. Results from this laboratory show that HLS from the locust S. gregaria induces positive chemokinesis in a heavily granulated, PO-containing subpopulation of plasmatocytes but has no effect on a less granular subpopulation (Huxham & Lackie, 1986; and unpublished data). Thus, it is becoming possible to investigate possible cooperative effects between the haemocyte subpopulations in the process of encapsulation.

This work was carried out when G.B.T. was in receipt of an MRC studentship. We are grateful to John Lackie for useful discussions and for advice on the analysis of cell movement.

Armstrong
,
P. B.
(
1979
).
Motility of ÛieLimulus blood cell
.
J. Cell Sci
.
37
,
169
180
.
Arnold
,
J. W.
(
1959
).
Observations on the amoeboid motion of living haemocytes in the wing veins of Blaberusgiganteus (Orthoptera: Blattidae)
.
Can. J. Zool
.
37
,
371
375
.
Ashida
,
M.
&
Dohke
,
K.
(
1980
).
Activation of prophenoloxidase by the activating enzyme of the silkworm Bombyx mori
.
Insect Biochem
.
10
,
37
47
.
Ashida
,
M.
,
Ishizaki
,
Y.
&
Iwahana
,
H.
(
1983
).
Activation of pro-phenoloxidase by bacterial cell walls or /3-1,3-glucans in plasma of the silkworm, Bombyx mori. Biochem
.
biophys. Res. Commun
.
113
,
562
568
.
Ashida
,
M.
&
Soderhâll
,
K.
(
1984
).
The prophenoloxidase activating system in crayfish
.
Comp. Biochem. Physiol
.
77B
,
21
-
26
.
Baerward
,
R. J.
&
Bousch
,
G. M.
(
1970
).
Timelapse photographic studies of cockroach haemocyte migrations in vitro
.
Expl Cell Res
.
63
,
208
213
.
Brown
,
A. F.
(
1982
).
Neutrophil granulocytes: adhesion and locomotion on collagen substrata and in collagen matrices
.
J. Cell Sci
.
58
,
455
467
.
Cook
,
D.
,
Stoltz
,
D. B.
&
Pauley
,
C.
(
1985
).
Purification and preliminary characterisation of insect spherulocytes
.
Insect Biochem
.
15
,
419
426
.
Davies
,
H.
&
Preston
,
T.
(
1985
).
Behaviour of insect plasmatocytes in vitro’, changes in morphology, settling, spreading and locomotion. J’
,
exp. Zool
.
236
,
71
82
.
Dularay
,
B.
&
Lackie
,
A. M.
(
1985
).
Haemocytic encapsulation and the prophenoloxidase- activation pathway in the locust Schistocerca gregaria
.
Insect Biochem
.
15
,
827
834
.
Dunn
,
G. A.
(
1983
).
Characterising a kinesis response: time averaged measures of cell speed and directional persistence
.
Agents Actions Suppl
.
12
,
14
33
.
Forrester
,
J. V.
&
Lackie
,
J. M.
(
1984
).
Adhesion of neutrophil leucocytes under conditions of flow
.
J. Cell Sci
.
58
,
455
467
.
Gagen
,
S. J.
&
Ratcliffe
,
N. A.
(
1976
).
Studies on the in vivo cellular reactions and fate of injected bacteria in Galleria mellonella and Pieris brassicae larvae
.
J. Invertebr. Path
.
28
,
17
24
.
Grimstone
,
A. V.
,
Rotheram
,
S.
&
Salt
,
G.
(
1967
).
An electron microscope study of capsule formation by insect blood cells
.
J. Cell Sci
.
2
,
281
292
.
Huxham
,
I. M.
&
Lackie
,
A. M.
(
1986
).
Separation and partial characterisation of haemocytes from the locust Schistocerca gregaria
.
Devi Comp. Immunol
.
10
, (abstract, in press).
Lackie
,
A. M.
,
Takle
,
G. B.
&
Tetley
,
L.
(
1985
).
Haemocytic encapsulation in the locust Schistocerca gregaria (Orthoptera) and in the cockroach Periplaneta americana (Dictyoptera)
.
Cell Tiss. Res
.
240
,
343
351
.
Lackie
,
J. M.
&
Burns
,
M. D.
(
1983
).
Leucocyte locomotion: comparison of random and directed paths using modified timelapse film analysis
.
J. Immun. Meth
.
62
,
109
122
.
Lackie
,
J. M.
&
Wilkinson
,
P. C.
(
1984
).
Adhesion and locomotion of neutrophil leucocytes on 2-D substrata and in 3-D matrices
. In
White Cell Mechanics: Basic Science and Clinical Aspects
(ed.
H. J.
Meiselman
,
M. A.
Lichtman
&
P. L.
Lacelle
), pp.
237
254
.
New York
:
Alan R. Liss
.
Lowry
,
O. H.
,
Rosebrough
,
N. J.
,
Farr
,
A. L.
&
Randall
,
R. J.
(
1951
).
Protein measurement with the Folin phenol reagent
.
J. biol. Chem
.
193
,
265
275
.
Mead
,
G.
,
Ratcliffe
,
N. A.
&
Renwrantz
,
L.
(
1985
).
Separation of insect haemocyte types on Percoll gradients: methodology and problems
.
J. Insect Physiol
.
32
,
167
177
.
Partridge
,
T.
&
Davies
,
P. S.
(
1974
).
Limpet haemocytes. II. The role of spikes in locomotion and spreading
.
J. Cell Sci
.
14
,
319
330
.
Smith
,
V. J.
&
Soderhàll
,
K.
(
1983
).
)3-l,3glucan activation of crustacean haemocytes in vitro and in vivo
.
Biol. Bull. mar. Biol. Lab., Woods Hole
164
,
299
308
.
Snedecor
,
G. W.
&
Cochran
,
W. G.
(
1967
).
Statistical Methods, 6th edn. Iowa: Iowa State University Press
.
SoderhÀll
,
K.
(
1981
).
Fungal cell wall /3-l,3 glucans induce clotting and phenoloxidase attachment to foreign surfaces of crayfish haemocyte lysate
.
Devi Comp. Immun
.
5
,
565
573
.
Soderhàll
,
K.
&
Smith
,
V. J.
(
1983
).
Separation of the haemocyte population of Carcinus maenas and other marine decapods, and prophenoloxidase distribution
.
Devi Comp. Immun
.
7
,
229
239
.
Unestam
,
T.
&
Ajaxon
,
R.
(
1976
).
Phenoloxidation in soft cuticle and blood of crayfish
.
J. Invertebr. Path
.
27
,
287
295
.
Wilkinson
,
P. C.
,
Lackie
,
J. M.
&
Allan
,
R. B.
(
1982
).
Methods for measuring leucocyte locomotion
. In
Cell Analysis
, vol.
I
(ed.
N.
Catsimpoolas
), pp.
145
193
.
New York
:
Plenum
.
Wilkinson
,
P. C.
,
Lackie
,
J. M.
,
Forrester
,
J. V.
&
Dunn
,
G. A.
(
1984
).
Chemokinetic accumulation of human neutrophils on immune complex-coated substrata: analysis at a boundary. J’
.
Cell Biol
.
99
,
1761
1768
.