J. Cell Sci. i8, 375-384 (i975)
Printed in Great Britain
375
CELL SURFACE LIPIDS AND ADHESION
III. THE EFFECTS ON CELL ADHESION OF
CHANGES IN PLASMALEMMAL LIPIDS
A. S. G. CURTIS, C. CHANDLER AND N. PICTON
Department of Cell Biology, University of Glasgow, Glasgow G n 6NU, Scotland
SUMMARY
The two preceding papers of this series suggest that the state of the plasmalemmal lipids
affects cell adhesion. Plasmalemmal composition was altered by the experimental incorporation
of fatty acids into R, and Rj positions in the phosphatidyl components of the cell surface. In
this paper we report that:
(1) If the incorporation is of long chain length fatty acids (saturated) cell adhesion rises.
(2) If the incorporation is of unsaturated fatty acids cell adhesion falls as the unsaturation
increases.
(3) Incorporation has to be extensive to produce a large change in adhesion.
(4) Changes in adhesion parallel the plasmalemmal incorporation but do not follow the total
cell incorporation.
Item (4) argues that it is plasmalemmal and not other membrane lipids that are involved in
cell adhesion. Item (3) suggests that bulk membrane properties and not some very specific
grouping are involved in the effects of lipids on adhesion. The similar extents of incorporation
of the various different fatty acids and the negligible amounts of lysophospholipids in the
membranes of cells that have incorporated fatty acids argue that the effects are not due to
differential accumulations of these lysolipids when incubations are done with different fatty
acids. The changes in adhesion cannot be accounted for by changes in surface charge density
since the electrophoretic mobility of the cells is unchanged by these incubations.
It is suggested that these effects on adhesion due to changes in plasmalemmal lipids can be
explained either in terms of the action of intermembrane van der Waals-London (electrodynamic) forces in cell adhesion or of changes in surfacefluidity.These alternatives are discussed.
INTRODUCTION
The experimental results obtained in the 2 preceding papers (Curtis, Campbell &
Shaw, 1975a; Curtis, Shaw & Spires, 19756) suggest strongly that the state of the
plasmalemmal phospholipids can control the degree of adhesion shown by cells. We
showed that conditions that cause accumulation of lysophospholipids in the plasmalemma are accompanied by a diminution in adhesion, while those that stimulate
reacylation of lysophospholipids lead to the maintenance or an increase in adhesion
of the cells. We suggested that these correlations might be due either to a requirement
for membrane turnover before adhesion can be established, or to a relation between
plasmalemmal fluidity and adhesion, or to a correlation between intermembrane
forces of attraction and lipid composition. The purpose of the present paper is to
describe experiments designed to test between these different hypotheses.
In our second paper of this series (Curtis et al. 19756) we report that extensive
incorporation of a chosen fatty acid into plasmalemmal phospholipids could be
376
A. S. G. Curtis, C. Chandler and N. Picton
effected by making use of the plasmalemmal acyl transferase system. We have taken
advantage of this system in work described in the present paper to alter plasmalemmal
composition in a chosen manner and to discover whether such changes effect concomitant changes in adhesion. Chapman (1968), Demel, Van Deenen & Pethica (1967)
and Demel, Van Kessel & Van Deenen (1972) amongst others, have shown that the
lateral van der Waals interactions of molecules in artificial membranes are affected by
the chain length of the acyl moieties of phospholipids and by the degree of unsaturation
of these molecules. Curtis (1962, 1973) has suggested that intermembrane van der
Waals forces may act in cell adhesion. If changes in composition affect intermembrane
forces in the same manner as lateral forces, and if van der Waals forces act in adhesion,
making the appropriate changes in plasmalemmal composition will produce changes
in adhesion. Such a finding would argue strongly against the hypothesis that adhesion
just requires turnover but would leave the question of whether adhesion depends on
membrane fluidity or on intermembrane van der Waals forces unresolved. The reason
for this is that these changes will at least in theory affect both fluidity and intermembrane forces. Preliminary experiments to test between these two explanations
are also described.
METHODS
Cells
Chick (De Kalb strain) embryos incubated for 7 days were dissected to provide neural retinae.
The retinae were dispersed into single-cell suspensions by the techniques described by Curtis
(1969). The cell suspensions were stored in calcium- and magnesium-free Hanks' medium
(CMF-medium) for up to 5 min at 2 CC before the medium for stimulating reacylation was
added. Hanks' medium was used with the addition of Tris at 2-5 x io 1 M dm"3 at pH 7-4.
Measurements of cell adhesion
Measurements of adhesion were made by following the kinetics of adhesion of a suspension
of cells. Collision efficiency measurements (Curtis, 1969) provide an effective means of following
aggregation kinetics. Collision efficiency measurements were made on the cells in suspension,
in various media described below, in a Couette viscometer at a shear rate of ca. 10 s"1 and at 37 °C.
Measurements of surf ace charge density of cells
A Rank (Rank Bros. Ltd, Bottisham, Cambridgeshire, U.K.) cylindrical cell microelectrophoresis apparatus was used to measure the electrophoretic mobilities of the cells. Zeta potentials
and surface charge density values were calculated in the standard manner. Measurements were
made at 2 °C and at a potential gradient of 6 V cm"1. Measurements were made in Hanks'
medium; 40 measurements were made for each treatment of the cells. The reason for using
this low temperature for measurement was because there is evidence (Curtis et al. 1975 a) that
lysolecithins and other lysophosphatidyl compounds accumulate in the cell surface when the
cells are incubated at 37 °C, while they do not appear at 2 °C. Hampton & Bolton (1969) show
that lysolecithins may affect cell surface charge density.
Reagents
ATP, Coenzyme A (free acid) and unlabelled fatty acids were obtained from Sigma. "Clabelled fatty acids, Radiochemical Centre, Amersham, U.K. Other reagents, BDH.
Cell surface lipids and adhesion. Ill
377
Substitution of acyl groups in plasmalcmmal phospholipids and other components
Curtis et al. (19756) showed that incubation of neural retinal cells in a medium containing
ATP, CoA and a long chain fatty acid in the range C 12 to C 20 lead to substantial incorporation
of the fatty acid into the Rx and R a acyl chains of plasmalemmal phospholipids as well as into
some other plasmalemmal components. The techniques described in that paper were used in
this work, namely incubation of cell suspensions in 1-25 x io~ 5 M dm" 8 ATP, 5 x io"* M dm" 1
CoA and 1 x io~ a g dm~3 fatty acid dissolved in Hanks' saline at 37 °C for 20 min. The extent
of incorporation was measured by using "C-labelled fatty acids in the incubation medium,
followed by isolation of the plasmalemmal fractions and T L C separation of the lipids, using
the methods described by Curtis et al. (1975b).
RESULTS
We showed, in the first paper in this series (Curtis et al. 1975 a) that incubation of
neural retinal cells in a medium containing CoA, ATP and oleate leads to the cells
maintaining an adhesiveness that they would have otherwise lost on incubation in
Table 1. Effect on cell adhesion of incorporation of various fatty acids
Collision efficiency,
Fatty acid
%
S.D.
Myristate
Palmitate
Stearate
Oleate
Linoleate
Linolenate
Arachidate
Eicosedienoate
Arachidonate
Untreated
Incubated 20 s in Hanks'
61
06
77
1-9
119
2-8
13
8-o
5'5
044
18-4
5-9
050
1-2
o-i
16
o-6
0-2
14-6
2-O
3'i
06
Neural retinal cells. Measurement of adhesion at 37 °C, after 20 min incorporation. Adhesion
measured as collision efficiency (percentage).
Incorporation conditions fully defined in the previous paper (Curtis et al. 1975 ft). s.D.,
standard deviations of 10-14 measurements.
Hanks' medium. In the second paper (Curtis et al. 19756) we also showed that incubation in such media leads to incorporation of the oleate or other fatty acids into the
plasmalemmal phospholipids and other components of the cell surface. Thus will
incubation of these cells in the presence of other fatty acids and CoA + ATP have
any effect on adhesion ? The results of such an experiment in which cells were incubated in fatty acid-CoA-ATP medium for 20 min at 37 °C prior to measurement of
adhesion are shown in Table 1.
This Table shows that adhesiveness of the cells after this incubation is greater the
longer the chain length of the fatty acid used or the greater the saturation of the
molecule. The first question that arises from these results is whether low adhesiveness
is associated with a small amount of incorporation of the fatty acid in question and
378
A. S. G. Curtis, C. Chandler and N. Picton
high adhesiveness with extensive incorporation, or whether all the fatty acids used are
incorporated to roughly similar extents. Incorporation of the respective fatty acids
into plasmalemmal lipids was measured. The results are shown in Table 2. These
incorporations were measured in the presence of large excesses of fatty acid (10 /*g/ml:
in the cases where this exceeds the solubility limit of the acid the undissolved acid
was finely dispersed by ultrasonication).
Table 2. Extent of incorporation offatty acids into the plasmalemmae of
neural retinal cells
Acid
Myristate
Palmitate
Stearate
Oleate
Linoleate
Linolenate
Arachidate
Arachidonate
Incorporation per
1x10' cells, /ig
O-2O
080
o-8o
0-50
S.D.
005
0-04
0-03
0 0 2
050
O-O2
o-6o
0-03
0-03
0-05
075
0-47
External concentration of fatty acid 1 x io~* g dm"3, 1-25 x io" 5 M dm"3 ATP, 5 x io~° M
dm"3 CoA in Hanks' medium.
Incorporations expressed as weights of free fatty acid incorporated in 20 min incubation.
Incorporation into phospholipids and neutral lipid; identified and separated by TLC, for
detailed method and example see Curtis et al. (19756). Approximately 70% of incorporation
into phospholipids, remainder into neutral lipid. Standard deviations, S.D., of 3 replicates.
Incorporation of the fatty acids into the plasmalemma was roughly identical for all
the acids used with the exception of myristic acid which was incorporated to a lesser
extent. Thus it seems impossible to account for the results in terms of the extent of
incorporation of the different fatty acids. However, despite the equality of incorporation it might be that greater amounts of lysophosphatidyl compounds, which are
probably associated with diminished adhesion, accumulate on incubation in unsaturated
or shorter chain fatty acids. No evidence for this was found, since lysophosphatidyl
compounds could not be detected in TLC runs (Methods, A, B, C and D, Curtis et al.
19756) of the plasmalemmal lipids, after fatty acid incorporation, for any of the acids
used in this work.
It has been suggested (see Curtis, 1973, for review) that cell surface charge density
is associated with the adhesiveness of a cell, high surface charge densities being
associated with low adhesion and vice versa. Thus it is possible that the effects of fatty
acid incorporation on adhesion act through a change in surface charge density
Electrophoretic measurements of the neural retina cells before and after incorporation
of various fatty acids (see Table 3) show that there is no change in surface charge
density when any of the fatty acids used in this work are incorporated into these cells.
A further question which requires answer is about the relationship between the
extent of incorporation and the change in adhesion. The relatively large amounts of
Cell surface lipids and adhesion.
Ill
379
Table 3. Electrophoretic mobilities of cells after incorporation offatty acids
at 2 °C, pun s- 1 V-1 cm
3r
Acid incorporated
Mobility ± s.D.
Myristic
Palmitic
Stearic
Oleic
Linoleic
Linolenic
Untreated
59
062 ±009
0-5810-08
0-59 ±0-07
0-57 ±0-07
o-6i ±0-05
0-5810-05
A
1 JO
8.
0-1
- c
- D
-I 20
„
ft
005
-
- 10
7
100
200
0
External FFA,
i
i
100
200
Fig. 1. Incorporation of free fatty acids (FFA) into plasmalemmal and whole cell
phospholipid and effects on cell adhesion. Whole cell incorporation of stearate (A)
and of arachidate (B) respectively, O
O; see left-hand ordinate for scale. Adhesiveness (collision efficiency), H
h ; see right-hand ordinate for scale. Abscissae,
external free fatty acid levels in incorporation system. Plasmalemmal incorporations
of stearate (c) and arachidate (n), O
O; see left-hand ordinate for scale. Adhesiveness H
\- ; right-hand ordinates and abscissae as in (A) and (B). Note parallelism
of plasmalemmal incorporation and effects on adhesiveness. Incorporations uncorrected
for recovery.
CEL
l8
380
A. S.G. Curtis, C. Chandler and N. Picton
incorporation shown in Table 2 obtained with high amounts of fatty acid in the
incorporation medium are associated (Table 1) with considerable changes in adhesion.
Would smaller incorporations produce the same change in adhesion ? In other words
we need to make measurements of the dose-response curves in order to discover
whether adhesion is controlled by bulk changes in the membrane or whether very
small changes are sufficient to produce a new value of adhesiveness for the cells.
This question can only be effectively answered by obtaining the dose-response
curves for those saturated acids that produce an increase in adhesion. Since the highly
unsaturated fatty acids produce a decrease in adhesion, any effects they may have at
low dose levels are hard to distinguish from the low values of adhesion that would
result from the accumulation of lysophosphatidyl compounds under these conditions.
Consequently the effect on adhesion of incubating the cells in a range of levels of
stearic and arachidic acids was studied and plasmalemmal and total cell incorporations
of these acids were measured using 14C-labelled acids.
Results are shown in Fig. 1 A-D. Clearly a large amount of substitution is required
to produce a maximal effect. Thus it appears that adhesion reflects bulk membrane
properties in respect of fatty acid composition. It is also of interest that adhesion
increases as plasmalemmal fatty acid content increases up to the point at which the
membrane appears to have become saturated. If higher fatty acid levels are used in
the incorporation medium the only consequence is that further incorporation into the
remainder of the cell takes place but both plasmalemmal content and adhesion remain
at a plateau level. This suggests strongly that adhesion is only affected by the plasmalemmal lipid nature and content and that the remainder of the cell is unimportant in
relation to adhesion. If some internal section of the cell is of importance in relation to
lipid content and adhesion this must be at most a relatively small proportion of the
cell membranous component.
DISCUSSION
The main finding in this work is that alterations in the fatty acid moiety of the
plasmalemmal lipids produce substantial changes in adhesion. Incorporation of
unsaturated fatty acids or those of chain length less than 18 carbon atoms leads to a
diminution of adhesion compared with the condition in freshly isolated cells. Incorporation of saturated fatty acids of chain length of 18 carbon atoms or longer leads to
an increase in cell adhesion. A fairly extensive incorporation is necessary to produce
a large change in adhesion. This argues that adhesion is controlled by the bulk state
of the lipids and not by changes in a very small proportion of lipids.
It should however be remembered that the method we use to substitute fatty acids
into the plasmalemma also leads to extensive incorporation in the remainder of the
cell (see also Curtis et al. 19756). Do the effects on adhesion result from changes in
the inner cell components rather from changes in the plasmalemma ? The findings
that there is a saturation level for incorporation into the plasmalemma corresponding
to the maximal change in adhesion, while further increases in total cell incorporation
are not accompanied by any increase in adhesion, suggest that the plasmalemma is
Cell surface lipids and adhesion. Ill
381
the important site for adhesion. Nevertheless it is not impossible to exclude the
hypothesis that a small undetectable inner cytoplasmic pool of lipid might play a role
in adhesion. Warren (1969) suggested that increased plasmalemmal turnover, presumably in association with internal pools, might be associated with a decrease in
adhesion. Waddell, Robson & Edwards (1974) put forward the hypothesis that
plasmalemmal turnover may be necessary to maintain an adhesive state in the cells.
Such explanations, namely that turnover per se affects adhesion, seem unlikely to us.
Our reason is that the various different fatty acids used in this work are incorporated
to similar extents, with presumably similar turnover rates even though they can have
very varying effects on adhesiveness. If turnover is needed in order to repair plasmalemma which has been damaged by some event such as trypsinization of the cells, it
would be expected perhaps that initially turnover would be associated with an increase
in adhesion. Partial evidence for this in trypsinized neural retinal cells has been
described (Curtis, 1970). Again, the similar extent of incorporation of fatty acids that
have very different effects on adhesion defeats this second argument.
Another explanation of our results could be based on the assertion that lysolecithin accumulates to differing extents when different fatty acids are incorporated.
Plasmalemmal lysolecithin (Curtis et al. 1975 a) appears to affect cell adhesion, so that
alterations in lysolecithin level would be expected to affect adhesion. However this
argument is improbable because lysophosphatides do not appear at any appreciable
level in cells where fatty acids have been incorporated (Curtis et al. 19756).
Differences in the surface charge density of cells (Curtis, 1973) might affect cell
adhesion. However this explanation cannot account for the differences observed
because the fatty acid substitutions were without effect on surface charge density
(see Table 3).
We are left with two general types of explanation. The first is that adhesion is due
to some very specialized molecular grouping on cell surfaces present in only relatively
small amounts. It has been pointed out (Curtis, 1967) that on theoretical grounds only
a very few groupings would account for the measured strength of adhesion. The
second is that cell adhesion reflects large-scale averaged properties of the membrane
such as surface charge density or electrodynamic forces. Fatty acid incorporation (see
Table 2 and Fig. 1) has to be extensive before marked effects on adhesion appear.
This implies that changes in the bulk properties of the surface are required to alter
cell adhesion. Cell adhesion might, of course, still be effected by localized and specialized groupings, few in number, with bulk phase properties of the membrane playing
a secondary role in the display or perhaps the aggregation of such groupings. Changes
in membrane fluidity, a bulk property, might control the display of a small number
of groupings. Experimental tests between the two theories will be presented in a
future paper.
One of the few explanations that can be put forward to explain the results, in terms
of either direct or indirect effects of the bulk-phase properties of the membrane, is
that changes in chain length and unsaturation resulting from fatty acid incorporation
affect the electrodynamic interactions of the plasmalemmal lipids. Unsaturation and
reduction in chain length would be expected to reduce electrodynamic forces and
24-2
382
A.S.G.
Curtis, C. Chandler and N. Picton
these might reduce intermembrane forces of attraction (and thus cell adhesion) and
intramolecular forces (affecting plasmalemmal fluidity) (Curtis, 1972). Chapman
(1968) and Demel et al. (1972) have produced evidence that such changes in composition affect membrane electrodynamic (van der Waals) forces. Engelman (1971) showed
that growth of Mycoplasma laidlawii in cultures with various different fatty acids led
to changes in the transition temperatures for the membranes and probably also to
changes in membrane thickness. James & Branton (1971) found that growth of Mycoplasma in linolenate-enriched media led to changes that might be attributable to
changes in membrane van der Waals forces. Thus extensive incorporation of different
fatty acids would lead to changes both in membrane fluidity and perhaps to intermembrane forces.
Changes in membrane fluidity might affect adhesion by altering the aggregability
of any membrane sites that might be involved in adhesion. It is conceivable if a small
number of molecules are involved in adhesion that a certain population density of
them is required to initiate an adhesion, possibly for example a gap junction. It is,
however, difficult to predict what effect very fluid or very viscous membranes would
have on the aggregation of such sites. For example, a very fluid membrane would
allow the rapid migration of sites but would also tend to aid their rapid lateral diffusion
from sites of high concentration. Similarly a very rigid membrane would tend to
prevent the aggregation of such molecules but would maintain aggregates once formed.
Thus it is hard to predict the effects of changes in membrane fluidity on adhesion on
this theory.
The alternative explanation is that electrodynamic forces between membranes are
altered when changes in fatty acid composition of the plasmalemma are made. Little
evidence other than that of Jones (1974) has yet been advanced to prove the operation
of these forces in cell adhesion but there is much circumstantial evidence that they
act (see Curtis, 1967, 1973). If the site aggregation theory were disproven there would
be strong reason to think in view of our present results that electrodynamic forces act
in cell adhesion. The restrictions of molecular motion as double bonds are introduced
into a molecule should lead to a reduction in electrodynamic interactions. As the
length of the hydrocarbon chain of fatty acids is reduced the thickness of material
in which these electrodynamic forces arise is reduced, thus diminishing them. Another
way of stating this is to say that these changes would increase the dielectric constants
of the membrane and the mass of material in which they arise, thus reducing electrodynamic interaction (Parsegian & Gingell, 1972).
It is of particular interest that Ninham & Parsegian (1970) calculate that the
electrodynamic interaction of two membranes should increase very rapidly when the
thickness of each membrane approaches half the dimensions of the gap between the
membranes. Thus increase in thickness of the plasmalemma due to the replacement
of shorter chain fatty acids by longer ones should at some point lead to a sudden
increase in adhesion if van der Waals forces play a role. We find that collision efficiencies
increase rapidly over the range C 16-C 20, which might correspond to membrane
thickness of about 5-6 nm approaching half the value of the gap width. Collision efficiencies give a logarithmic measure of adhesion energy. Since there is some degree of
Cell surface lipids and adhesion. HI
383
uncertainty in the evaluation of adhesive energies from the measured collision
efficiencies we have preferred to state the results as efficiencies. However it is clear
from the results that the energy of adhesion is increasing exponentially with increasing
chain length of fatty acids incorporated over the C 16-C 20 region. Thus these fatty
acid substitutions might be expected to alter electrodynamic interactions of cells in
part because of thickness changes in the membrane and in part because of vibrational
changes as double bonds are introduced. No comparable investigation of the effects
of altering thickness on intra-membrane forces appears to have been carried out but
it seems improbable that such a marked effect as appears in intermembrane forces
would occur.
At least one other system is known in which saturated fatty acids probably lead to
an increase and unsaturated fatty acids to a decrease in adhesion. Stoltz et al. (1973)
report that platelets aggregate more rapidly in the presence of saturated fatty acids as
chain length is increased and progressively less rapidly with an increase in the number
of double bonds per molecule.
We thank Science Research Council for a grant (B/SR49099). We should like to express our
appreciation of the skilled technical assistance we have received from Miss Rose McKinney.
REFERENCES
D. (1968). Biological Membranes: Physical Fact and Function. New York and London:
Academic Press.
CURTIS, A. S. G. (1962). Cell contact and adhesion. Biol. Rev. 37, 82-129.
CURTIS, A. S. G. (1967). The Cell Surface. London: Logos Press, Academic Press.
CURTIS, A. S. G. (1969). The measurement of cell adhesiveness by an absolute method.
J. Embryol. exp. Morph. 22, 305-325.
CURTIS, A. S. G. (1970). On the occurrence of specific adhesion between cells. J. Embryol. exp.
Morph. 23, 253-272.
CURTIS, A. S. G. (1973). Cell adhesion. Prog. Biophys. molec. Biol. 27, 315-386.
CURTIS, A. S. G., CAMPBELL, J. & SHAW, F. M. (1975a). Cell surface lipids and adhesion.
I. The effects of lysophosphatidyl compounds, phospholipase As and aggregation-inhibiting
protein. J. Cell Sci. 18, 347-356.
CURTIS, A. S. G., SHAW, F. M. & SPIRES, V. M. C. (19756). Cell surface lipids and adhesion.
II. The turnover of lipid components of the plasmalemma in relation to cell adhesion. J. Cell
Sci. 18, 357-373DEMEL, R. A., VAN DEENEN, L. L. M. & PETHICA, B. A. (1967). Monolayer interactions of
phospholipids and cholesterol. Biochint. biophys. Acta 135, 11-19.
DEMEL, R. A., VAN KESSEL, W. S. M. G. & VAN DEENEN, L. L. M. (1972). The properties of
polyunsaturated lecithins in monolayers and liposomes and the interactions of these lecithins
with cholesterol. Biochint. biophys. Acta 266, 26-40.
ENCELMAN, D. M. (1971). Lipid bilayer in the membrane of Mycoplasma laidlawii. J. molec.
Biol. 58, 153-165.
HAMPTON, J. R. & BOLTON, C. M. (1969). Effects of phospholipids on platelet electrophoretic
mobility. J. Atherosclerosis Res. 9, 131-139.
JAMES, R. & BRANTON, R. D. (1971). The correlation between the saturation of membrane
fatty acids and the presence of membrane fracture faces after osmium fixation. Biochint.
biophys. Acta 233, 804-812.
JONES, G. (1974). Intercellular adhesion: modification by dielectric properties of the medium.
J. Membrane Biol. 16, 297-312.
NINHAM, B. W. & PARSEGIAN, V. A. (1970). van der Waals forces across triple-layer films.
J. Chem. Phys. 52, 4578-4587.
CHAPMAN,
384
A. S. G. Curtis, C. Chandler and N. Picton
V. A. & GINGELL, D. (1972). Some features of physical forces between biological
cell membranes..7. Adhesion 4, 283-306.
STOLTZ, J. F., PROST-DVOJABOVIC, R. J., STREIFF, F., SAMANA, M. & LARCAN, A. (1973).
Influence d'acides gras satur^s et non satur£s sur la pression de filtration de plasma riche en
plaquettes. Correlation avec l'aggregation plaquettaire. Ada liaemat. 49, 306-313.
WARREN, L. (1969). The biological significance of turnover of the surface membrane of animal
cells. Curr. Topics devl Biol. 4, 197-222.
WADDELL, A., ROBSON, R. T. & EDWARDS, J. G. (1974). Colchicine and vinblastine inhibit fibroblast aggregation. Nature, Lond. 248, 239-241.
PARSEGIAN,
[Received 9 January 1975)