J. Cell Sci. 41, 125-133 (1980)
Printed in Great Britain © Company of Biologists Limited 1980
RED BLOOD CELL ADHESION
I. DETERMINATION OF THE IONIC CONDITIONS FOR
ADHESION TO AN OIL-WATER INTERFACE
IAN TODD AND DAVID GINGELL
Department of Biology as Applied to Medicine, The Middlesex Hospital
Medical School, Cleveland Street, London W\P 6DB, England
SUMMARY
We have examined the adhesion of glutaraldehyde-treated human red blood cells to a clean
liquid hydrocarbon interface as a function of sodium chloride concentration. Cells adhere
reversibly to the interface over a wide range of concentrations but fail to do so below about
o-i mM. Adherent cells do not alter the tension of the oil/water interface.
These results show that cells can adhere by physical forces in special situations without
biochemical interactions. The data provide a basis for calculating the size of the physical
forces.
INTRODUCTION
The mode of adhesion of cells to other cells and to non-cellular surface under
physiological conditions is an unresolved problem. The interpretative difficulties are
compounded by the complexity of the components in physiological media. While
analyses in such media will ultimately prove essential, it is initially easier and more
instructive to use greatly simplified systems. In such systems it is possible to investigate the adhesion of cells to well characterized surfaces under strictly denned conditions (Gingell & Todd, 1975; Gingell & Fornes, 1975, 1976). In these systems
where the possibility of chemically specific interactions between cell and substratum
can be excluded, we have tried to define the role of electrostatic and electrodynamic
forces in the adhesive process. Our previous results have shown that electrostatic
forces can prevent cell adhesion to clean surfaces and that when electrostatic repulsion
is sufficiently reduced, adhesion takes place. We have described the construction
of a chamber for making a flat interface between oil and saline which can be used as
a substratum for cell adhesion (Gingell & Todd, 1975) and in the work described
here the use of this chamber has been extended to obtain data capable of yielding
quantitative estimates of attractive and repulsive forces experienced by cells interacting with a liquid hydrocarbon.
METHODS
Hexadecane (99%) was filtered through a 10-cm column of chromatographic alumina
(BDH, Poole, Dorset; Brockman Activity 11) before use. Twice-distilled water from a commercial all-glass apparatus was redistilled from alkaline potassium permanganate -and again
from orthophosphoric acid. Sodium chloride roasted at 700 °C was a gift of Dr J. Mingins.
126
/. ToddandD. Gingell
Glassware was cleaned in chromic acid rinsed in 4-times distilled water, then treated briefly
with a mixture containing 4 % hydrofluoric acid (40 %) in 40 % nitric acid. Finally, it was
rinsed in 4-times distilled water until the pH rose to 5-3—5-5 which was the pH range of our
experimental solutions. The stainless steel and PTFE chamber were cleaned by repeated
sonication in methanol, then ether, then dried in a warm oven. Analytical grade chemicals
were used where possible.
Human red cells obtained by venipuncture were treated with glutaraldehyde using 2 methods.
1
Normal' cells were fixed in 3-3 % glutaraldehyde in 145 HIM NaCl at 4 °C for 18 h. The cells
were then washed exhaustively in distilled water by centrifugation. ' Long fixation' cells had
similar treatment but were treated for 12 months in glutaraldehyde.
The oil/water inversion chamber (Gingell & Todd, 1975) was modified by addition of
inlet and outlet cocks so that a continuous flow could be passed through the aqueous compartment. The lid seal was improved with a PTFE ' O '-ring, and the spring-ball assemblies were
replaced with nylon screws for smoother operation. For the same reason, the stainless steel
torus was fitted into a PTFE carrier. Great care was necessary while exchanging the aqueous
phase; inlet and outlet were adjusted to precisely equal flow rates from a reservoir before
adding oil. This ensured that the meniscus did not bulge, displacing oil past the screw thread
into the aqueous compartment. Some adjustment during operation was necessary. After
sweeping the interface with a fine pipette and then flattening it in the closed chamber, the
cell suspension was admitted to the chamber via the reservoir feeding the inflow. After
allowing 20 min for cells to settle down on to the oil/water interface the chamber was inverted
and a field photographed before any cells fell out of focus. Photographs taken at intervals up
to 10 min with Zeiss Ukatron flash equipment recorded changes in the number of cells at
the interface. When reversibility of adhesion was tested the aqueous phase in the chamber
was exchanged at about 1-25 cc/min. This gave a sufficient rate of exchange without disturbing
cells at the interface.
Elecrrophoresis was performed in a Zeiss cytopherometer. The power pack was modified
to supply low constant currents necessary for measuring mobilities at low ionic strengths.
The glass measuring chamber was not exposed to hydrofluoric acid cleaning solution. In order
to prevent pH changes at the platinum electrode from affecting the working electrolyte the
electrode chamber was buffered with dialysed bovine serum albumin, isolated from the
working electrolyte by dialysis membranes. The pH of effluent from the measuring chamber
was measured after each set of mobility determinations. Emulsions of oil droplets in saline
for electrophoresis were prepared by sonicating the phases together using a Dawe Soniprobe.
Dilute emulsions were stable for several hours. Charge densities were calculated from mobilities
using the Gouy-Chapman equation
q — charge density, e = dielectric constant of water (taken as 8o), n0 = number of monovalent
ions of either sign/cc, e is electronic charge and i/r0 is surface potential, identified with the
zeta potential (£) calculated from electrophoretic mobility by the Smoluchowski equation
for large bodies of easy shape £ = 4rn)u/e where u is electrophoretic mobility and ?/ is bulk
viscosity of water. In these equations k, T and n have their usual significance. We have used
the ordinary relation between £ and u for large spheres, since correction is unnecessary under
the conditions employed (Carroll & Haydon, 1975).
The interfacial tension of the oil/water interface was measured by the hanging drop method
(Alexander & Hayter, 1971). Drops of water were allowed to fall downwards into hexadecane
from a tip machined from ' Kel F ' (poly(chlorotrifluoroethylene)) fitted to an Agla micrometer
syringe. The syringe barrel and drop-receiving vessel were thermostatted to 25 ± o-1 °C and
the apparatus was secured firmly to avoid inadvertent movement. All parts contacting clean
liquids were cleaned by immersion in chromic acid solution, followed by exhaustive rinsing
in 4-times distilled water.
Three types of measurement were made. First, the interfacial tension of hexadecane against
aqueous solutions of NaCl from o to 145 mM; this was then repeated using normal red cells
suspended in the saline solutions. The suspension density was similar to that used in inversion
chamber experiments. Lastly, drops of cell suspension in the saline solutions were left hanging
for 20 min just below the critical size for detachment, before allowing them to fall. This
Adhesion to hydrocarbon in salt solutions
127
allowed cells to sediment on to the oil/water interface, exactly paralleling the situation in the
inversion chamber, and providing a measurement of the maximal possible interfacial contamination which could have occurred in the chamber.
Interfacial tensions were calculated according to Alexander & Hayter (1971). The mass
of NaCl in the aqueous drops was included in calculations of drop relative density.
RESULTS
Adhesion of cells to the interface
Fig. 1 shows the percentage of red cells remaining attached to the interface after
inversion of the chamber. All cells were adherent from 100 mil NaCl down to
1 mM. Below i-o mM normally fixed cells began to fall on inversion and by 0-3 mM
50% remained, while in distilled water less than 10% were left. Long-fixation cells
required a slightly greater dilution before falling from the interface. Unfortunately,
loss of these cells prevented a complete run, but the concentration corresponding
to 50% adhesion was near 0-2 mM. Cells which were able to fall had gone out of
focus by less than 3 min and observations at 10 and 15 min did not show any further
loss from the interface.
100
90
80
70
CO
50
40
30
20
10
j
I
1
I
I
I
0-1 0-2 0 3 0-4 0-5 0-6 0-7 0-8 0-9 1 0
NaCI, mM
Fig. 1. Percentage of cells adherent to hexadecane saline interface at various NaCl
concentrations. Red blood cells treated with glutaraldehyde for 1 year ( • ) or 18 h ( • ) .
Observations on the reversibility of adhesion were conducted with long-fixation
cells as follows. After unattached cells had fallen away, electrolyte was allowed to
flow through the chamber at 1-25 cc/min. 50 cc of 0-4 mM produced no significant
change in the number of cells attached, showing that the flow alone was insufficient
to remove cells from the interface (Table 1). However, when a similar volume of
distilled water was passed at the same rate through the chamber, after inversion,
cells fell off progressively (Table 1). In 2 experiments 83 and 87% of cells initially
stuck to the surface became detached after approximately 50 cc of distilled water had
flowed through the chamber. This experiment was repeated in 10 mM NaCl. At
128
/ . Todd and D. Gingeli
this concentration, observation is complicated by strong lateral aggregation of cells
at the interface - rapid lateral motions occurring over several cell diameters. This
process is practically completed in the 20-min settling time allowed: during subsequent flow hardly any single cells were recruited into clusters. The cell counts for
10 mM were therefore recorded for non-aggregated cells only, and it was found that
Table 1. Reversibility of adhesion
No.
of random
fields
Electrolyte
concentration, photographed
mM NaCl
Experiment
i (control)
2
3
4 (control)
5
Initial
Final
Initial
Final
Initial
Final
Initial
Final
Initial
Final
Total
no. of cells
counted
Mean cell
no. per field,
is.D.
339 ±85 I
396145]
11
1696
1189
1999
005
04
6
223
1
0-05
1
197
26
—
—
932
0-4
o-4
o-4
5
3
10
—
10
—
10
003
5
9
502
222 ± 79 1
37 ±5 /
197 \
26 /
= }
i86i 17)
56114/
%
detached
~
0
83
87
O
7O
70% of initially adherent single cells could be detached by dilution from 10 to
0-03 mM NaCl. In a control experiment continuous microscopic observation of cells
attached to the interface in 10 mM NaCl showed that none became detached when
25 cc of 10 mM NaCl were passed through the chamber at 2 cc/min.
Electrophoresis
Electrophoretic mobilities of oil droplets and red cells fixed by both methods
were measured at pH 5-4 + 0-1 without buffer at a series of NaCl concentrations
from o-i to 20 mM (Fig. 2).
Cells subjected to normal fixation had significantly higher mobilities than those
with long fixation, involving a much longer aldehyde treatment. Fig. 3 shows red
cell charge density calculated according to equation 1. The dissociation characteristics
of normally fixed cells (Fig. 4) were investigated by electrophoresis as a function
of pH at constant ionic strength (145 mM NaCl). Acidity was adjusted with Analar
NaOH and HC1 solutions. The mobility curve shown in Fig. 4 indicates 2 major
ionogenic groups, one with a pK a near 2-7, which is sialic acid, and another of
higher pK a , near 6-5, which may be due to carboxyl groups of partially oxidized
glutaraldehyde, though such groups would be expected to have a pK a near 4-5.
About 10% of the total negative charge of the cell is due to the presence of this
group. The continuous curve in Fig. 4 shows the very similar results of Vassar et al.
(1972).
Adhesion to hydrocarbon in salt solutions
129
Interfacial tension
The tension of the oil/water interface rose by a small but significant amount as
the concentration of NaCl rose to 145 mM, though the salt effect is very small at
20 mM. Compared with these control values (Table 2) the introduction of fixed red
cells into the aqueous phase caused no significant change in interfacial tension, even
when the drop was left hanging for 20 min. We confirmed that contamination of the
interface with unfixed cells in 145 mM NaCl caused a progressive fall in interfacial
tension (Table 2). Microscopic observation showed that unfixed cells change their
refractive index when they contact the interface, indicating lysis.
r
5
3
3o
TS-...
j
I
1 1111
10
I
I
I I III
10
"T
1
11 n
100
[NaCll.triM
Fig. 2. Electrophoretic mobility of red blood cells and a hexadecane emulsion in
various concentrations of NaCl. Red blood cells fixed in glutaraldehyde for 1 year
(O) or 18 h ( • ) .
, droplets of hexadecane.
x
I
0
1
2
I
3
4
5
[NaCI],mM
Fig. 3. Charge density on fixed red blood cells at various NaCl concentrations
calculated from the electrophoretic mobility using equation 1. Red blood cells
fixed in glutaraldehyde for 1 year (Q) or 18 h (O). Bars show standard error.
O
O, charge density, calculated from equations 3 and 4, of a surface bearing
7-1 x io12 ionizable groups/cm with a pK. of 3-3 and 8-3 x io 11 ionizable groups/cma
with a pKj of 6-o.
/. Todd and D. Gingeli
130
DISCUSSION
Our experiments show that red blood cells, fixed in glutaraldehyde, adhere to
an oil/water interface at electrolyte concentrations from 20 mM down to below
1 mM NaCl. Further dilution prevents adhesion. The surface tension results show
that fixed cells do not contaminate the interface with surface active materials, even
under conditions where cells adhere to the interface, so that contamination is not
7
1-2
E
_
0-8
0-6
5
o
0-4
0-2
5
6
10
11
pH
Fig. 4. Electrophoretic mobility of red blood cells fixed in glutaraldehyde for 18 h.
The points were experimentally determined while the continuous line is from the
results of Vassar et al. (1972).
Table 2. Effect of red cells on interfacial tension
Interfacial tension of aqueous NaCl/hexadecane interfaces
Cell state
Molarity of
NaCl, mM
Fixed
Fixed
Fixed
Fixed
Unfixed
0
0-4
25
J
45
145
Without cells
Immediate value
with cells
With cells
after 20 min
delay
51-9910-03
51-92 ±006
5208 ±0-03
5299 ±006
5299 ±0-06
52-00 ±0-05
51-91 ±005
5201 ±0-03
52-8710-04
52-9710-05
5r-97
5^74
52-12
52-94
43-84
responsible for the adhesive behaviour we have observed. These results follow
earlier findings (Gingeli & Todd, 1975) that fixed red cells in 145 mM NaCl adhered
to an oil/water interface in the presence of a monolayer of weakly ionized or unionized
carboxylic acid at neutral pH. At higher pH values, where ionization occurred, cells
were unable to adhere to the interface.
The electrophoretic results from the hexadecane/saline emulsion shows that the
interfacial charge density remains constant in the range 20-0-5 mM NaCl. This is
conveniently shown by the double log plot in Fig. 5 derived from equation (1).
, sinh2
TT
\2kT
= <z 2 -zekT
= const if q = const.
Adhesion to hydrocarbon in salt solutions
131
Since the oil/water interface has a constant charge density over this extended range
of NaCl concentrations the charge is unlikely to be due to Cl~ ion adsorption; it may
depend on OH~ ion concentration but we have not investigated this possibility. We
have not investigated the reason for the apparent reduction in interfacial charge
density below 0-5 mM.
The charge density on fixed red cells of both types falls with dilution (Fig. 3), due
to the association of carboxyl groups of sialic acid caused by decreasing surface pH
10
p
0-1
I
0-01
I I I I
10
III
I
100
I
I I I
1000
(1/mM)X100
Fig. 5. Comparison of experimental data with predictions of the Gouy-Chapman
equation (equation 1). The continuous line was calculated from equation 2. The
points were determined experimentally.
as the negative surface potential increases. In order to calculate this association we
have used the high ionic strength data from Fig. 4 to assign the total dissociable
charge densities (qlyq2) present as groups of each pKa value (pK^, pKaj,). The
equation in Ht, the surface pH which must be solved for this case, is
where H+ is bulk hydrogen ion concentration and other symbols are as in equation i.
This was conveniently solved for H£ on a computer and the value of H+ inserted
into the mass action equations
^ ..
v
Itxr
i TJ+\ "\
(4)
Values of a represent degrees of dissociation so that qxa^ and <72a2 are the required
charge densities present at a given salt concentration n0. The result of this calculation
for normally fixed cells is shown in Fig. 3 (broken curve) and the fact that we can
get a reasonable fit over most of the curve suggests that the form of the experimental
curve is largely due to carboxyl reassociation. To obtain a perfect fit it is necessary
132
/. Todd and D. Gingell
to postulate a pK a value of about 3-1 for sialic acid, rather than 2-7. It is possible
that there may be a common mechanism which depresses both cell and oil/water
charges below expected values at very low ionic strengths, which remains to be
explained.
The measurements described provide all the necessary ingredients for calculating
the repulsive force between cells and the oil/water interface. From the charge
density on the cell and on the oil/water interface at the 'critical' salt concentration
where half the cells are able to adhere to the interface, it is possible to calculate
electrostatic repulsion as a function of separation. A preliminary account of the
result of this calculation has appeared (Gingell, Todd & Parsegian, 1977) and will
be discussed fully in an accompanying paper (Parsegian & Gingell, 1980). The
conclusion is that cells settling onto the oil/water interface under gravity should
not be able to approach closer than about 100 nm unless a long-range attractive
force is present which counteracts the electrostatic repulsion. The improbably large
attraction that would be required to pull red cells into molecular contact with the
interface against electrostatic repulsion makes it likely that in our experiment a
long-range force-balance situation exists, where attraction and repulsion are equal
at some finite cell-substratum separation. In order to directly ascertain whether this
is so, we have made optical interferometric measurements which are reported in a
following paper (Gingell & Todd, 1980).
Without further observations or calculations it is, however, clear that an attractive
force is capable of overcoming electrostatic repulsion between red cells and a clean
protein-free hydrocarbon water interface, where no 'biochemical' interactions can
occur. We are not aware that this has been demonstrated in any other cell adhesion
system and it shows that purely physical forces can be responsible for adhesion.
While such forces are demonstrable under idealized conditions, their role in physiological situations is much less clear (Gingell & Vince, 1979).
REFERENCES
A. E. & HAYTER, J. N. (1971). Determination of surface and interfacial tension.
In Techniques of Chemistry, vol. i (ed. Weissberger & Rossiter), part v, chapter 9. New
York: Academic Press.
CARROLL, B. J. & HAYDON, D. A. (1975). J. Chem. Soc. (Faraday Trans. 1) 71, 361-377.
GINGELL, D. & FORNES, J. A. (1975). Demonstration of intermolecular forces in cell adhesion
using a new electrochemical method. Nature, Lond. 256, 210-211.
GINGELL, D. & FORNES, J. A. (1976). Interaction of red blood cells with a polarized electrode.
Evidence of long-range intermolecular forces. Biophys. J. 16, 1131-1153.
GINGELL, D. & TODD, I. (1975). Adhesion of red blood cells to charged interfaces between
immiscible liquids. A new method. J. Cell Set. 18, 227-239.
GINGELL, D. & TODD, I. (1980). Red blood cell adhesion. II. Interferometric examination
of the interaction with hydrocarbon oil and glass. J. Cell Sci. 41, 135-149.
GINGELL, D., TODD, I. & PARSEGIAN, V. A. (1977). Long-range attraction between red cells
and a hydrocarbon surface. Nature, Lond. 268, 767-769.
GINGELL, D. & VINCE, S. (1979). Long-range forces and adhesion: an analysis of cellsubstratum studies. Br. Soc. Cell Biol. Symp., Cell Motility and Adhesion (ed. A. S. G.
Curtis & J. Pitts), Cambridge University Press (in Press).
ALEXANDER,
Adhesion to hydrocarbon in salt solutions
133
PARSEGIAN, V. A. & GINGELL, D. (1980). Red blood cell adhesion. I I I . Analysis of forces.
J. Cell Set. 41, 151-157.
VASSAR, P. S., HARDS, J. M., BROOKS, D. E., HACENBERGER, B. & SEAMAN, G. V. F. (1972).
Physico-chemical effects of aldehydes on the human erythrocytes. J. Cell Biol. 53, 809818.
[Received 8 May 1979)