Naturally occurring cationic proteins secreted by human granulocytes have pro-inflammatory effects including induction of increased vascular permeability and oedema, which are likely to be mediated by damage to vascular endothelium. Synthetic cationic polyamino acids have been shown to exert similar inflammatory effects in vivo. We have therefore used a range of synthetic polycationic amino acids to investigate the characteristics required to cause endothelial cell damage, assessed by in vitro inhibition of leucine incorporation into macromolecules by human umbilical vein endothelial cells (HUVEC) in culture. Exposure of HUVEC to 20 nM–2 μM cationic polypeptides of similar Mr(av) (≈40000) in the presence of 20% serum produced a dose-dependent inhibition of [3H] leucine incorporation by polymers of ornithine, arginine or lysine. Similar results were obtained using [3H]thymidine. Neutral or anionic polypeptides of similar Mr were without effect. The molar potency of polylysines increased over the range Mr 40000–120000, while polylysines of Mr(av; <25 000 had no effect. In the absence of serum, inhibition occurred more rapidly and at lower doses. Inhibition of leucine and thymidine incorporation was time-dependent, e.g. exposure to 800 nM-polylysine, Mr (av) 90 000, led to progressively increasing inhibition that was complete after 24 h exposure, and was irreversible. The effects of polycations could not be blocked by pretreatment of the cells with polyanions. Precoating of the culture surface with polylysines had no effect on leucine incorporation by HUVEC or their subsequent response to polylysines in solution. Exposure to the peptide Arg-Gly-Asp-Ser inhibited incorporation by 30% but did not increase susceptibility to polylysine. The extent of inhibition of radiolabel incorporation was correlated with changes in cell morphology and release of a cytoplasmic enzyme (1actate dehydrogenase). We conclude that cationic proteins can exert significant, essentially irreversible, cytotoxic effects on endothelium, the magnitude of which depends on the density of cationic residues interacting with the cell and the time of exposure to the polycation. The use of synthetic amino acid polymers in such a model system provides a means of investigating the characteristics of cationic polypeptides that affect endothelial cells and the nature of the responses induced.

Naturally occurring cationic proteins secreted by human granulocytes are cytotoxic to microorganisms (Zeya & Spitznagel, 1968; Odeberg & Olsson, 1975; Elsbach & Weiss, 1983), and can inhibit tumour cell growth (Clark et al. 1976). In addition they have pro-inflammatory effects, including induction of increased vascular permeability and oedema, which are likely to be mediated by damage to vascular endothelium (Janoff & Zweifach, 1964; Janoff et al. 1965; for a review, see Henson & Johnston, 1987), although the mechanisms involved have not been studied.

The overall positive charge on these macromolecules is due to the presence of a high proportion of lysine and arginine residues (Zeya & Spitznagel, 1968), and other basic proteins and synthetic cationic polyamino acids have been shown to exert similar pro-inflammatory effects in vivo (Stein et al. 1956; Vehaskari et al. 1984). We have taken advantage of the availability of a wide range of synthetic polycationic amino acids to investigate the characteristics required to cause endothelial damage, assessed by in vitro inhibition of thymidine or leucine incorporation into macromolecules by human umbilical vein endothelial cells. We found that cytotoxicity induced by polycations depended not only on concentration but also on molecular size; the effects were irreversible but not progressive after exposure to polycations, and the damage could not be inhibited by pretreatment of the cells with polyanions. A preliminary account of some of this work has been published as an abstract (Morgan et al. 1987).

Materials

The following chemicals were obtained from Sigma Ltd, Poole, Dorset, UK; polymers of L-lysine (as the hydrobromides), with mean molecular masses (Mr(av)) of 4000 (range 1000— 5000), 14000 (4000–15000), 25000 (15 000–30000), 40000 (30000–70000), 60000 (30000–70000), 90000 (70000–150000), 260000 (150000–300000), and 500000 (>300000); poly-L-glutamic acid (as the sodium salt), Mr(av) 32000 (20000–70000), poly-L-histidine (as the hydrochloride), Mr(av) 40000 (15000–70000); heparin (from porcine intestinal mucosa, sodium salt; Mr approx. 15 000); protamine sulphate (grade X, from salmon, Mr 10000–15000); Arg-Gly-Asp-Ser (fibronectin tetrapeptide); collagenase (type II), DL-lactic acid, 85% syrup, and reduced (1ow u.v.-absorbing) Triton X-100. Other chemicals were from BDH Chemicals, Poole, Dorset, UK. Pentosan polysulphate (Mr approx. 6000) was a gift from Sanofi et Cie, Toulouse, France. Calf serum and tissue culture media (in powder form) were from Flow Laboratories, Irvine Scotland. L-[4,5,-[3H]leucine (ICimmol−1) and [rnethyl-3H]thymidine (5 Ci mmol−1) were obtained from Amersham International, Amersham, UK. Scintillation fluid NE265 was from Nuclear Enterprises, Edinburgh, Scotland.

Cell culture

Human endothelial cells were isolated and cultured as described (Jaffe et al. 1973). Confluent first passage cells from different cords were pooled, resuspended in growth medium (medium 199 supplemented with 10% foetal calf serum plus 10% newborn calf serum) and dispensed into 96-well tissue culture plates. The initial cell concentration (1×104 cells per well) was chosen because we had previously established that under these conditions the cells were in continuous growth for the duration of the experiment and had not reached confluence when harvested (Morgan, 1987). In some experiments the culture plates were pretreated with polylysines (10μgml−1), or the fibronectin tetrapeptide Arg-Gly-Asp-Ser (25 pg ml−1) in serum-free medium for 30min, then rinsed three times before use.

Incorporation of radiolabel into macromolecules

At 24 h after plating the cells the medium was removed, the cells were washed with serum-free medium and fresh growth medium (200 pl) was applied. Polycations or other agents were then added in 10 pl of serum-free medium; controls received 10 pl medium only. After defined incubation periods, 1 μCi (1 μl) of [3H]thymidine or [3H]leucine was added to each well. In certain experiments, the medium was removed, the cells rinsed, and fresh medium applied before adding labelled precursor. Experiments were terminated 24 h later as follows: after removal of the medium, the cell monolayer in each well was washed twice with phosphate-buffered saline (PBS; 200 μl), exposed to 5% trichloroacetic acid (200pl) for 5 min, then washed with methanol (200μl). Finally, the cells were digested by the addition of 200 pl of 25 M-formic acid. Each formic acid digest was transferred, with one rinse of PBS, to a vial and 3–5 ml of NE265 scintillation fluid were added. Radioactivity was determined using a Wallac/Rackbeta scintillation counter. Control cultures incorporated 1000–4000 cts min−1 [3H]leucine or 5000–10000ctsmin−1 [3H]thymidine in 24 h.

Lactate dehydrogenase activity

Lactate dehydrogenase activity was assayed fluorimetrically using a modification of the method of Morgenstern et al. (1965). The substrate solution was lactic acid (10ml of an 85% solution) in 11 of 0·67 M-2-amino-2-methyl-l-propanol adjusted to pH9·0 by addition of 5M-HC1. NAD (4mgml−1) was added to ice-cold buffered substrate immediately before use. Portions of medium (50 μl) were removed from experimental cultures and placed in wells of a microtitre plate. Following addition of substrate (50μl), plates were incubated for 30 min at 37°C. The reaction was stopped by addition of glycine-sodium hydroxide buffer, pH 10·5 (0·5M; 50μl/well). Cells were lysed with reduced Triton X-100 (0·1%; 50μl) before assay. The amount of NADH formed was quantified using a fluorimetric microplate reader (‘Fluoroskan’, Flow Laboratories Ltd).

Analysis of data

Each test condition was replicated in 10–12 wells, and data were analysed by a multiple means comparison test (Peritz ‘F’ test; Harper, 1984; validated by comparison with the Newman-Keuls multiple range test in the statistical package SPP; Royston, 1984) with the overall confidence level set at 99% (P< 0·01).

Concentration-dependent effects of polycations

Human umbilical vein endothelial cells were exposed to increasing doses of cationic polypeptides of similar molecular masses (Mr approx. 40000) for 48 h and the effects on [3H] leucine or [3H]thymidine incorporation were determined. Polyornithine, polyarginine and polylysine (at concentrations up to 72 μg ml ; approx. 2μM) produced a dose-dependent inhibition of leucine incorporation (Table 1). Protamine (Mr 10000–15000), the cationic nature of which is due to a high proportion of arginine residues, had no effect. Polyhistidine (Mr(av) 26000), with an overall pK near neutrality, also had no effect. Similar results were obtained for thymidine incorporation (data not shown).

Dependence of inhibitory effects on polymer size

At concentrations of 10-800 nM, polylysines with Mr(aV) up to 25 000 were not inhibitory. Inhibition became evident at 800 nM with polymers of Mr (av) >25000, and at lower concentrations as the polymer size was increased (Table 2). A similar relationship between polymer size and the extent of inhibition was observed for thymidine incorporation (data not shown). At the highest molecular weights used (Mr(av) 260000 and 500000) maximum inhibition was reached at 200 nM and declined at 800 nM.

Modulation of inhibitory effects by serum

The aforementioned experiments were carried out in the presence of 20% serum in the culture medium. When the serum concentration was reduced, the extent of inhibition of leucine incorporation by polylysine (Mr 90000; 200 nM) increased from about 10% (relative to the control value in the absence of agonist) when the serum concentration was 20% to about 85% when the serum concentration was only 5% (Fig. 1). When serum levels were further reduced, inhibition by polycations was less apparent because cell growth in the controls was markedly reduced.

Ability of polyanions to block or reverse effects of polylysine

When cultures were incubated with polyglutamine for 1 h, and then polylysine was added without removing the polyglutamine, the effect of the polylysine was abolished (Fig. 2). Heparin or pentosan polysulphate also greatly reduced the extent of inhibition of leucine incorporation caused by the addition of polylysine (Fig. 2). Complete reversal was not achieved with these two polyanions, even at higher doses, because while polyglutamine (800 nM) alone had no effect on endothelial cell leucine incorporation, heparin (25 μM) or pentosan polysulphate (8 μM) each inhibited leucine incorporation to 82% or 77% of controls, respectively (heparin, although itself inhibitory, can act synergistically with endothelial growth factors to increase cell growth rate; Thornton et al. 1983). Although polyanions, when present in the medium, effectively blocked the action of polylysine, pretreatment of the cells for 2 h with polyanion followed by removal of the supernatant before addition of polylysine had no detectable effect on the subsequent inhibition by polylysine of leucine incorporation (90–95% in each case).

Time course of the inhibitory effect

The extent of inhibition by a given dose of a particular polylysine was related to the time of exposure: almost complete inhibition was achieved after 24 h exposure to a polymer of Mr 90000 at 800 nM in the presence of 20% serum, whereas only 50% inhibition occurred after 5h (Fig. 3). We next tested whether the cells could recover from the effects of polylysine, by incubating them for 24 h following removal of the agonist before assessing their ability to incorporate [3H] leucine. The results from such an experiment (Table 3, experiment B) demonstrate that the extent of inhibition under these conditions was similar to that found when no recovery period intervened (experiment A). Thus neither did the cells recover from the effects of polylysine nor did the effects progress after the removal of the agonist. When polylysine in the bulk phase was neutralized by addition of a polyanion, the results were similar to those found when the agonist was removed (experiment C).

Effects of modulating cell attachment

In contrast to the effects of polylysines in the bulk phase, leucine incorporation by cells grown in wells pretreated with polylysines of .’17,62000, 90000 or 500000 did not differ significantly from that of cells grown in untreated wells (data not shown). In addition, pretreatment of the culture surface had no effect on subsequent exposure to polylysine. When cells grown in untreated wells were exposed to the fibronectin tetrapeptide Arg-Gly-Asp-Ser 24h after plating, they became retracted, showed morphological changes similar to those observed in rat kidney cells by Hayman et al. (1985), and incorporated less leucine than controls (Table 4). Under these conditions there was no evidence that the cells were more susceptible to polylysine; the effects were simply additive (Table 4).

Polylysine-induced release of lactate dehydrogenase

Experiments were carried out to discover whether the 120–1 polylysine-induced inhibition of leucine or thymidine incorporation reflected cytostasis or cytotoxicity. Fig. 4 demonstrates that exposure to polylysine for periods that led to increasing inhibition of leucine incorporation (measured over a subsequent 24 h period) was accompanied by an increase in the amount of lactate dehydrogenase in the medium and a concomitant decrease in the cellular levels of this cytoplasmic enzyme, indicating that the effects of polylysine were cytolytic. In each case the reduced incorporation, which was not due to cell loss (Fig. 5), correlated with a difference in appearance of the cells, which became extremely granular in contrast to the normal cobblestone appearance exhibited by the controls (Fig. 5). These experiments were carried out in serum-free medium because there is significant lactate dehydrogenase activity in calf serum. It is notable that under these serum-free conditions the effects of polylysine, although still related to polymer size, were much more rapid. In separate experiments we found that the partial release of lactate dehydrogenase caused by exposing the cells to polylysine (Mr(av) 90000, 800 nM) for 15 or 30 min did not increase further over a period of 4 h after the polylysine was removed (data not shown).

Our results demonstrate that exposure to synthetic polycations produced Mr-, concentration- and time-dependent inhibition of the subsequent incorporation of [3H]leucine or [3H]thymidine into endothelial macromolecules. The range of polymers chosen encompasses the reported molecular weights of naturally occurring granulocytic cationic proteins (21000–76000; Nachman et al. 1972; Ohlsson & Olsson, 1973; Olsson & Venge, 1974). Although Vehaskari and co-workers (1984) demonstrated in vivo effects on vascular permeability using a polylysine of Mr 3400, the concentration used was more than 100-fold greater than the maximum used in the experiments reported here. Exposure to polycations was accompanied by progressive release of a cytoplasmic marker enzyme (1actate dehydrogenase), indicating that under the conditions of our experiments the effects were cytotoxic, and not merely cytostatic.

The observation that inhibition was more evident after exposure to larger polymers (at a fixed molar concentration of polylysine), together with the lack of detectable effect of short polymers even at high concentrations, suggests that co-operative interactions leading to damage are enhanced by the close apposition of multiple positively charged residues.

The degree of inhibition after a given time appears to relate to the effective concentration of free polycation in the medium. Neutralization of charge in the bulk phase (when the polyanion was added before the polylysine) was sufficient to abolish the effect of polylysine. Similarly, when the polyanion was added after polylysine, further action of the polycation was prevented. The implication that polyanions act by binding to polylysine in solution is supported by the inability of pretreatment of the cells with polyanion and then rinsing to protect them from subsequent polycation exposure; and by finding that after a defined exposure time the effects of polylysine were greater when the concentration of serum, and hence of competing anionic sites, was lowered.

Inhibition increased with time of exposure to a single concentration of polylysine of fixed Mr, and occurred much more rapidly if the polycation was added in the absence of serum. The extent of inhibition was neither reduced nor augmented if cells were left for 24 h in fresh medium after exposure to polylysine, before measuring [3H]leucine incorporation. These results imply that continued presence of active agonist in the bulk phase is necessary to induce progressive inhibition, yet cells once affected cannot recover from exposure to polylysine. It seems unlikely that this reflects heterogeneous behaviour within the cell population, since morphological evidence (Fig. 5) indicated homogeneous responses to polylysine. Modification of the culture surface by treatment with polylysine (a technique previously used to increase cell attachment; Freshney, 1987) had no effect on cell growth although the polylysines used were cytotoxic when added in solution. Thus the cytotoxic effect cannot be due to binding of added polylysine to the substratum. In addition, precoating plates did not confer protection from subsequently added polylysines.

We also found that induction of cytoplasmic enzyme release was time-dependent and did not progress further when polylysine was removed from the cells. The simplest model to account for this is one in which interaction of the polycation molecule with the cell leads to a transient local disruption of membrane permeability that is terminated by an event such as internalization of the polycation. It is well documented that basic polymers enhance endocytosis in an Mr-dependent manner (Ryser, 1967; Pratten et al. 1978). The subsequent, apparently permanent, effects on leucine incorporation are more difficult to understand, but could result from inhibition of leucine uptake, because the bound polycation blocks the appropriate recycling of membrane components including transport molecules, or additionally associates with the cell surface in a way that affects a variety of functions dependent on adequate membrane fluidity. Evidence in support of the idea that polylysines interfere with cell mobility is provided by the fact that similar and additive effects were elicited when cells were treated with the fibronectin tetrapeptide. In addition, as would be predicted, preliminary experiments indicate that confluent monolayers of endothelial cells are less susceptible to the effects of polylysines than actively growing cells.

Previous studies of the effects of polycations on endothelial cells have emphasized the overall negative charge provided by glycosaminoglycans such as heparan sulphate and sialic acid-containing proteoglycans at their surface (Simionescu & Simionescu, 1986). In the special case of the glomerulus, intercellular permeability in this fenestrated endothelial bed is enhanced by neutralization of the anionic sites of the basement membrane with polycations (Kanwar et al. 1980; Barnes et al. 1984). It is less clear whether increased permeability in other beds in response to locally applied or circulating polycations (Stein et al. 1956; Janoff et al. 1965; Nachman et al. 1970; Vehaskari et al. 1984; Needham et al. 1988) can be attributed to similar effects. Whilst treatment of endothelial cells with highly cationized ferritin leads to neutralization of negative charge and aggregation and detachment of glycocalyx components (Skutelsky & Danon, 1976), it has been shown that neutralization with moderately cationic ferritin, which does not lead to redistribution of sites, substantially decreases fluid permeability in capillaries (Turner et al. 1983).

Polycations are known to interact with purified phospholipid vesicles, in a process with several characteristics similar to those found for their effects on endothelial cells. Interaction with phospholipids is apparently a two-step event, first induced by electrostatic attraction to negative charges and then strengthened co-operatively by conformational changes in the polymer that permit integration of cationic side-chains into the lipid bilayer (Hammes & Schullery, 1970; Schafer, 1974; Hartman & Galla, 1978; Carrier et al. 1985). As in our experiments, very short polymers were ineffective, while longer polymers were more potent, and the interaction led to increased vesicular permeability (Hammes & Schullery, 1970; Gad et al. 1982). At very high concentrations there was evidence of reduced effectiveness, as we found, perhaps as a result of steric hindrance. Cytotoxic effects of polylysine, again proportional to molecular size, have been reported in neutrophils, and were not affected by pretreatment of the cell surface to remove sialyl residues (Elferink, 1985). Thus, although proof requires further experiments, we favour the concept that the endothelial responses to polycations that we have measured are due to interactions with membrane phospholipids.

In conclusion, we have demonstrated that the cytotoxic effects of synthetic polycations on endothelial cells depend on incubation time, concentration and charge density. In serum-free conditions, complete release of cytoplasmic enzyme occurred after 1–2 h exposure to polylysine, whereas inhibition of leucine incorporation was substantial within 15 min and complete in 30-60 min. This raises the possibility that brief exposure to polycations may alter biologically important functional responses of the endothelium without inducing cytotoxicity. This concept is supported by the observation that exposure of endothelial cells to polycations for as little as 2 min increases prostacyclin synthesis and release of purines without inducing release of lactate dehydrogenase or altered morphology (Needham et al. 1988). The use of synthetic amino acid polymers in such model systems provides a means of investigating the characteristics of cationic polypeptides that affect endothelial cells and the nature of the responses induced. Further work is needed to determine the cell surface components and cellular mechanisms involved, and to establish the relationship between these findings and the effects of endogenous cationic proteins such as those released during degranulation of platelets or leucocytes.

We thank the staff of the Delivery Suite, Northwick Park Hospital, for their continued expert collection of umbilical cords, and Drs Lindsey Needham and Paul Hellewell for stimulating our interest in this work.

Barnes
,
J. L.
,
Radnik
,
R. A.
,
Gilchrist
,
E. P.
&
Venkatachalam
,
M. A.
(
1984
).
Size and charge selective permeability defects induced in glomerular basement membrane by a polycation
.
Kidney Int
.
25
,
11
19
.
Carrier
,
D.
,
Dufourcq
,
J.
,
Fancon
,
J.-F.
&
Pezolet
,
K.
(
1985
).
A fluorescence investigation of the effects of polylysine on dipalmitoylphosphatidylglycerol bilayers
.
Biochim. biophvs. Acta
820
,
131
139
.
Clark
,
R. A.
,
Olsson
,
I.
&
Klebanoff
,
S. J.
(
1976
).
Cytotoxicity for tumour cells of cationic proteins from human neutrophil granules
.
J. Cell Biol
.
70
,
719
723
.
Elferink
,
J. G. R.
(
1985
).
Cytolytic effects of polylysine on rabbit polymorphonuclear leukocytes
.
Inflammation
9
,
321
331
.
Elsbach
,
P.
&
Weiss
,
J.
(
1983
).
A réévaluation of the roles of the Oz-dependent and Oz-independent microbicidal systems of phagocytes
.
Rev. infect. Dis
.
5
,
843
853
.
Freshney
,
R. I.
(
1987
).
Culture of Animal Cells
, 2nd edn.
New York
:
Alan R. Liss
.
Gad
,
A. E.
,
Silver
,
B. L.
&
Eytan
,
G. D.
(
1982
).
Polycation-induced fusion of negatively-charged vesicles
.
Biochim. biophvs. Acta
690
,
124
132
.
Hammes
,
E.
&
Schullery
,
S. E.
(
1970
).
Structure of macromolecular aggregates. II. Construction of model membranes from phospholipids and polypeptides
.
Biochemistry
9
,
2555
2563
.
Harper
,
J. F.
(
1984
).
Peritz’ F test: BASIC program of a robust multiple comparison test for statistical analysis of all differences among group means
.
Comput. Biol. Med
.
14
,
437
445
.
Hartmann
,
W.
&
Galla
,
H.-J.
(
1978
).
Binding of polylysine to charged bilayer membranes: molecular organisation of a lipid-peptide complex
.
Biochim. biophvs. Acta
509
,
474
490
.
Hayman
,
E. G.
,
Pierschbacher
,
M. D.
&
Ruoslahti
,
E.
(
1985
).
Detachment of cells from culture substrate by soluble fibronectin peptides../
.
Cell Biol
.
100
,
1948
1954
.
Henson
,
P. M.
&
Johnston
,
R. B.
(
1987
).
Tissue injury in inflammation: oxidants, proteinases, and cationic proteins
.
J. clin. Invest
.
79
,
669
674
.
Jaffe
,
E. A.
,
Nachman
,
R. L.
,
Becker
,
C. G.
&
Minick
,
C. R.
(
1973
).
Culture of human endothelial cells derived from umbilical vein. Identification by morphologic and immunologic criteria
.
J. din. Invest
.
52
,
2743
2756
.
Janoff
,
A.
,
Schaefer
,
S.
,
Scherer
,
J.
&
bean
,
M. A.
(
1965
).
Mediators of inflammation in leukocyte lysosomes. II. Mechanisms of action of lysosomal cationic protein upon vascular permeability in the rat
.
J, exp. Med
.
122
,
841
851
.
Janoff
,
A.
&
Zweifach
,
B. W.
(
1964
).
Production of inflammatory changes in the microcirculation by cationic proteins extracted from lysosomes
.
J. exp. Med
.
120
,
747
762
.
Kanwar
,
Y. S.
,
Limker
,
A.
&
Farquhar
,
M. G.
(
1980
).
Increased permeability of the glomerular basement membrane to ferritin after removal of glycosaminoglycans (heparan sulfate) by enzyme digestion./
.
Cell Biol
.
86
,
688
693
.
Morgan
,
D. M. L.
(
1987
).
Oxidised polyamines and the growth of human vascular endothelial cells: prevention of cytotoxic effects by selective acetylation
.
Biochem. J
.
242
,
347
352
.
Morgan
,
D. M. L.
,
Clover
,
J.
&
Pearson
,
J. D.
(
1987
).
Characteristics of polycation-induced cytotoxicity in human endothelial cells
.
Int.J. Microcirc. clin. Exp
.
6
,
289
(Abstr).
Morgenstern
,
S.
,
Flor
,
R.
,
Kessler
,
G.
&
Klein
,
B.
(
1965
).
Automated determination of NAD-coupled enzymes. Determination of lactate dehydrogenase
.
Analvt. Biochem
.
13
,
149
161
.
Nachman
,
R. L.
,
Weksler
,
B.
&
Ferris
,
B.
(
1970
).
Increased vascular permeability produced by human platelet granule cationic extract
J, clin. Invest
.
49
,
274
281
.
Nachman
,
R. L.
,
Weksler
,
B.
&
Ferris
,
B.
(
1972
).
Characterisation of human platelet vascular permeability-enhancing activity
.
J, clin. Invest
.
51
,
549
556
.
Needham
,
L. A.
,
Hellewell
,
P. G.
,
Williams
,
T. J.
&
Gordon
,
J. L.
(
1988
).
Endothelial functional responses and increased vascular permeability induced by polycations
,
Lab. Invest, (in press)
.
Odeberg
,
H.
&
Olsson
,
I.
(
1975
).
Antibacterial activity of cationic proteins from human granulocytes
.
J, din. Invest
.
56
,
1118
1124
.
Ohlsson
,
K.
&
Olsson
,
I.
(
1973
).
Neutral proteases of human granulocytes. Isolation and partial characterisation of two granulocyte collagenases
.
Eur.J. Biochem
.
36
,
473
481
.
Olsson
,
I.
&
Venge
,
P.
(
1974
).
Cationic proteins of human granulocytes. II. Separation of the granules of leukemic myeloid cells
.
Blood
44
,
235
246
.
Pratten
,
M. K.
,
Duncan
,
R.
&
Lloyd
,
J. B.
(
1978
).
A comparative study of the effects of polyamino acids and dextran derivatives on pinocytosis in the rat yolk sac and the rat peritoneal macrophage
.
Biochim. biophys. Acta
540
,
455
462
.
Royston
,
J. P.
(
1984
).
SPP - Statistical Package for Personal Computers. TCL., 40B Royal Hill, London SE10 8RT, UK
.
Ryser
,
H. J.-P.
(
1967
).
A membrane effect of basic polymers dependent on molecular size
.
Nature, Loud
.
215
,
934
936
.
Schafer
,
P. T.
(
1974
).
The interaction of polyamino acids with lipid monolayers
.
Biochim. biophys. Acta
373
,
423
435
.
Simionescu
,
M.
&
Simionescu
,
N.
(
1986
).
Functions of the endothelial cell surface. A
.
Rev. Physiol
.
48
,
279
293
.
Skutelsky
,
E.
&
Danon
,
D.
(
1976
).
Redistribution of surface anionic sites on the luminal front of blood vessel endothelium after interaction with polycationic ligands
.
J. Cell Biol
.
71
,
232
241
.
Stein
,
O.
,
De Vries
,
A.
&
Katchalski
,
E.
(
1956
).
The effects of polyamino acids on the blood vessels of the rat
.
Archs hit. Phannacodyn. Ther
.
107
,
243
253
.
Thornton
,
S. C.
,
Mueller
,
S. N.
&
Levine
,
E. M.
(
1983
).
Human endothelial cells: use of heparin in cloning and long-term serial cultivation
.
Science
222
,
623
625
.
Turner
,
M. R.
,
Clough
,
G.
&
Michel
,
C. C.
(
1983
).
The effect of cationised ferritin upon the filtration coefficient in single frog capillaries. Evidence that proteins in the endothelial cell coat influence permeabilit
J. Microvasc. Res
.
25
,
205
222
.
Vehaskari
,
V. M.
,
Chang
,
C. T.-C.
,
Stevens
,
J. K.
&
Robson
,
A. M.
(
1984
).
The effects of polycations on vascular permeability in the rat
.
J. din. Invest
.
73
,
1053
1061
.
Zeya
,
H. I.
&
Spitznagel
,
J. K.
(
1968
).
Arginine rich proteins of polymorphonuclear lysosomes. Antimicrobial specificity and biochemical heterogeneity
.
J. exp. Med
.
127
,
927
941
.