- Email: [email protected]
macrophages: a co-culture study
Atherosclerosis 145 (1999) 157 – 165
Inhibition of human arterial smooth muscle cell growth by human monocyte/macrophages: a co-culture study D. Proudfoot a, C. Fitzsimmons b, J. Torzewski b, D.E. Bowyer b,* a
Department of Medicine, Uni6ersity of Cambridge, Addenbrooke’s Hospital, Le6el 5, Hills Road, Cambridge CB2 2QQ, UK b Department of Pathology, Uni6ersity of Cambridge, Tennis Court Road, Cambridge CB2 1QP, UK Received 5 June 1998; received in revised form 10 November 1998; accepted 21 January 1999
Abstract Monocyte/macrophages produce a variety of substances which may influence the function of smooth muscle cells (SMC). During atherogenesis, macrophages are thought to modulate SMC migration, proliferation and synthesis of extracellular matrix. Such modulation is the balance between stimulatory and inhibitory influences. Thus, for example, our earlier studies have shown that macrophages not only secrete mitogens, but also produce small molecular weight inhibitors of SMC proliferation. In the present study, we have used a co-culture system in which human monocyte/macrophages were separated from human arterial SMC (hSMC) by a filter with the optional addition of a 12 kDa cut-off dialysis membrane, in order to assess their effect on hSMC growth. We have found that human peripheral blood-derived monocytes produced a substance of B 12 kDa that inhibited hSMC growth in the co-culture system. The monocyte-derived factor causing this effect was completely blocked by indomethacin, indicating that growth-inhibitory factors produced by the monocytes were cyclooxygenase products. We have shown that PGE1 and PGE2 inhibit hSMC growth, making them likely candidates for the effector molecules released from monocytes in our co-culture system. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Monocyte/macrophages; Smooth muscle; Co-culture
1. Introduction Accumulation of smooth muscle cells (SMC) in the arterial intima is found in arterial hyperplastic diseases, including atherosclerosis. The stimulus for the phenotypic modulation of the normally contractile SMC in the arterial media to a synthetic, secretory phenotype which can migrate, proliferate and secrete extracellular matrix in the arterial intima is largely unknown. Since monocyte infiltration appears to precede SMC proliferation in developing atherosclerotic lesions, factors released by monocyte/macrophages are thought to be involved [1–4]. Macrophages are known to release a wide range of substances, some of which are growth
* Corresponding author. Tel.: +44-1223-333722; fax: +44-1223333346.
regulatory factors and some which may affect matrix production [5]. The processes in atherogenesis are complex and are controlled by both stimulatory and inhibitory influences. Although macrophages secrete mitogens such as platelet-derived growth factor (PDGF) [6]; fibroblast growth factor (FGF) [7]; interleukin-1 (IL-1) [8], transforming growth factor a [9] and a heparin-binding epidermal growth factor-like molecule [10], there are also a number of macrophage products which have inhibitory effects on SMC growth. Transforming growth factor b (TGFb) has been reported to cause stimulation or inhibition depending on cell culture conditions [11,12] and on the concentration of TGFb [13]. Prostaglandins can inhibit SMC proliferation [14–16] and certain inhibitory prostanoids can be induced in SMC by IL-1 [17]. Macrophages also express an in-
0021-9150/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 2 1 - 9 1 5 0 ( 9 9 ) 0 0 0 2 8 - 3
158
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
ducible form of nitricoxide synthase which produces the nitric oxide radical which can inhibit SMC proliferation [18]. Previous studies in our laboratory have shown that a small molecular weight substance of less than 10 kDa is released by murine macrophages which affects porcine SMC growth [19]. More recently, the production of the growth-inhibitory material was shown to be regulated by macrophage activating substances, for example, PMA [20]. It seems clear that growth-inhibitory factors produced by macrophages may play an important role in regulating the proliferation of arterial SMC and that the balance between the release of growth-inhibitors and mitogens may influence the development of atherosclerotic lesions. In order to investigate the balance of production of human monocyte/macrophage growth regulators, we have explored the net effects of freshly isolated peripheral blood-derived monocytes on human arterial SMC growth in a co-culture system.
2. Materials and methods
2.1. Materials The culture medium used was Dulbecco’s Modified Eagle’s Medium (DMEM) (Sigma) buffered with 3.7 g/l NaHCO3 and 5% CO2, and supplemented with 100 IU/ml penicillin (Sigma PEN-NA), 100 mg/ml streptomycin (Sigma S-6501) and 4 mM L-glutamine (Sigma G-3126). Heat-inactivated foetal calf serum (FCS) was purchased from Imperial Laboratories. Heat-inactivated human AB serum was purchased from the Blood Transfusion Service, Cambridge, UK. Indomethacin, PGE1 and PGE2 were purchased from Sigma, recombinant human IL-1b was purchased from R and D systems and Dazmegrel was a gift from Pfizer.
2.2. Cells 2.2.1. Human peripheral blood monocyte (hPBM) isolation Human monocytes were isolated from preparations of buffy coats purchased from the Blood Transfusion Service, Cambridge. Mononuclear cells were first isolated by Ficoll–Hypaque sedimentation [21] and subsequently further purified by Percoll density fractionation [22]. Monocytes and lymphocytes were identified by flow cytometry analysis according to their characteristic forward and side scatter on a Becton Dickinson FACScan flow cytometer. Purity was further checked by expression of CD14 (for monocytes) and CD45 (for leukocytes) using monoclonal antibodies (LeucoGATE CD45/ CD14, and its control, Becton Dickinson). The purity of monocytes ranged between 80 and 88%. In order to
attempt to obtain monocyte preparations of higher purity, magnetic beads with a monoclonal antibody to CD2 were used to remove most of the lymphocytes (Dynabeads M-450 Pan-T, CD2, Dynal). Percoll-separated monocytes were resuspended in PBS containing 2% FCS to a concentration of 1× 107 cells/ml and incubated with 4×107 dynabeads for 30 min at 4°C. The beads were removed by placing the suspensions in a magnetic field and the fluid containing cells was pipetted off very gently. Monocytes further purified in this way resulted in small increases in purity i.e. ranging from 89 to 93% monocytes. In some experiments lymphocytes were also isolated after the Ficoll–Hypaque separation by plating out cells in NUNC 24-well plates at a density of 2×106 cells/well and then removing non-adherent cells after a 1 hour incubation at 37°C. The purity of lymphocytes isolated in this way ranged between 95 and 99%.
2.2.2. Human smooth muscle cells Human smooth muscle cells (hSMC) were a gift from Professor P.L. Weissberg, Department of Medicine, Addenbrooke’s Hospital, Cambridge. hSMC used in this paper were obtained from aortae of individuals ranging from 2 to 65 years old. The cells were prepared from explants of tissue and were confirmed as smooth muscle cells by positive staining with monoclonal antibodies against aSM-actin (Sigma). Cells were maintained in DMEM containing 20% FCS and were used between passages 5 and 15. 2.2.3. HepG2 cells HepG2 (human hepatic cell line) were purchased from the ECACC and were maintained in DMEM containing 10% FCS. 2.3. Preparation of cells and dialysis membrane for co-culture 2.3.1. Macrophages hPBM (1× 106 cells/well) were plated into cell culture filter-inserts (Falcon, code. 3095) in 0.5 ml medium using a 24-well plate with medium alone in the lower well. This concentration of hPBM resulted in an almost confluent monolayer of adherent cells in the filter-insert. Filter inserts containing hPBM were then transferred into wells containing hSMC. This type of co-culture is referred to as ‘filter-separated co-culture’. In some experiments, a 12 kDa cut-off membrane was placed around the culture-insert. This is referred to as ‘dialysis membrane separated co-culture’. 2.3.2. Smooth muscle cells hSMC were plated in 24-well plates (NUNC) at a density of 20 000 cells/well in 1 ml DMEM containing 20% FCS. After an overnight incubation the medium was changed to 0.1% FCS and the incubation was continued for a further 72 h in order to induce quiescence.
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
2.3.3. Preparation of 12 kDa cut-off dialysis membrane Dialysis membrane of 12 kDa cut-off (Spectra-Por) was boiled in 2% Na2CO3 and 1 mM EDTA for 10 min followed by boiling in distilled water for 10 min. The tubing was then cut into 2 cm2 pieces and placed in distilled water and autoclaved at 110°C/10 psi for 10 min. Dialysis membrane was bathed in DMEM prior to placing over a well of a 24-well plate and inserting a filter-insert. 2.4. Measurement of 3H-thymidine incorporation into hSMC Following a 40-h co-culture of hSMC with monocyte/macrophages, the filter-inserts containing the monocyte/macrophages were removed and the hSMC in the bottom of the culture wells were pulsed with 3 H-thymidine (2 mCi/well, NEN specific activity 6.7 Ci/mmol) for 8 h. Controls for these experiments included: (a) hSMC exposed to medium containing FCS or AB serum; (b) hSMC exposed to serum-free DMEM (to demonstrate that the cells were quiescent); (c) hSMC exposed to medium containing FCS or AB serum and a filter-insert with no cells (referred to as ‘no cell control’). The time points chosen for these experiments were used as the hSMC were still in S-phase during 40–48 h after serum-stimulation. Other studies on human vascular SMC proliferation have demonstrated a cell doubling time of \ 40 h [23,24]. Also, hPBM were removed before pulsing hSMC with 3Hthymidine in order to eliminate any uptake of thymidine by the hPBM. Cells were harvested by placing plates on ice, washing with PBS, fixing with methanol (10 min), precipitating with 10% TCA (20 min) and digesting in 200 ml of 0.3 M NaOH for 30 min. The cell digest was then placed in 3 ml HiSafe scintillation fluid and radioactivity was measured.
159
above and after 70 h in quiescing medium, the cells were loaded with 1 mCi/ml 3H-adenine (Amersham) in 0.1% FCS in DMEM and the incubation was continued for 2 h. This medium was then removed, the cells were washed four times in PBS and then co-cultured with macrophages as above. After 40 h, the medium from each well was removed and placed into scintillation vials containing 3 ml HiSafe scintillation fluid. The wells were then washed in PBS and 100 ml aliquots of washings were added to separate scintillation vials. Finally, the cells remaining in each well were treated with 0.1% Triton X-100 in PBS for 5 min in order to lyse the cells. 100 ml of lysate was also added to scintillation vials and counted for 3H on a scintillation counter. Streptolysin S (Sigma, 100 units/well) was included as a positive control. Percentage release of 3 H-adenine was calculated as follows: %3H-adenine release =
supernatant+ washings ×100 supernatant+ washings+lysate
2.6.2. Trypan blue Trypan blue exclusion was also used to estimate toxicity. Cells were washed three times in phosphate buffered saline (PBS) and exposed to trypan blue (0.25% in PBS, Sigma). 2.7. Statistical analysis Results are expressed as the mean9the standard error of the means (S.E.M.). Each individual experiment was carried out in triplicate wells. The results of three individual experiments gave rise to the means and standard errors in each graph. Sample means were compared using a Student’s t-test. PB 0.05 was considered statistically significant.
2.5. Measurement of change in hSMC number In order to measure change in hSMC number, hSMC and monocytes or macrophages were incubated in coculture as for the 3H-thymidine assay for periods up to 144 h. At intervals of 48 h, filter-inserts were removed and the hSMC in triplicate wells were trypsinised and counted on a haemocytometer. When cells were incubated for these longer periods, medium was changed for fresh medium at each further 48-h time point.
2.6. Toxicity assays 2.6.1. 3H-adenine release In order to determine whether the hSMC had been damaged by co-incubation with macrophages, a 3Hadenine release assay was performed using a modified method by Andreoli et al. [25]. hSMC were prepared as
3. Results
3.1. Effect of freshly isolated human monocytes on hSMC growth in co-culture To find out the net effect of human peripheral blood monocyte/macrophages (hPBM) on hSMC growth, hPBM were placed in filter-inserts and co-cultured with hSMC. Fig. 1 shows that freshly isolated human monocytes inhibit human SMC proliferation when grown in filter-separated co-culture, as measured by 3Hthymidine incorporation into hSMC. Human monocytes also had a growth-inhibitory effect when a fractionating, 12 kDa membrane was present during co-culture, indicating that the activity was less than 12 kDa in molecular mass (Fig. 1). In these co-culture
160
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
experiments a control cell type, HepG2, had no effect on 3H-thymidine incorporation into hSMC (Fig. 1). HPBM also significantly inhibited hSMC growth as assessed by counting cell numbers over time (Fig. 2). Freshly-isolated hPBM were added in the co-culture system at 48-h intervals.
3.2. Assessment of potential toxicity of hPBM to hSMC To investigate whether the growth-inhibitory effect of hPBM was due to toxicity, 3H-adenine release was measured from hSMC as an estimate of cell damage. 3 H-adenine release from hSMC was 21.259 0.25% for control cells, 16.65 92.35% for hSMC exposed to hPBM, and 39.09 1.5% for hSMC exposed to a positive control (Streptolysin S). The growth-inhibitor(s) released by hPBM were therefore not toxic towards hSMC. Also, greater than 99% of hSMC exposed to monocytes excluded trypan blue in either filter-separated or dialysis membrane-separated co-culture.
3.3. Influence of lymphocytes Since monocyte isolates contain significant numbers of lymphocytes, it was possible that the growth-in-
Fig. 1. Effect of hPBM on 3H-thymidine incorporation into hSMC in filter- and dialysis membrane separated co-culture. Freshly isolated human peripheral blood-derived monocytes, isolated by the Percoll method (83% monocytes/17% lymphocytes, by FACS analysis), or HepG2 cells were incubated with quiescent human SMC in 10% AB serum for 40 h in the presence or absence of a 12 kDa cut-off membrane. hSMC were pulsed with 3H-thymidine for 8 h then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 2 year old male. Similar results were observed with other isolates from older males and females. * Indicates significant difference from no cell control values (PB0.05).
Fig. 2. Effect of hPBM on hSMC growth. hPBM were incubated in filter-separated co-culture (as in Fig. 1) with quiescent hSMC (72 h in 0.1% FCS) for periods up to 6 days. At 48-h intervals, fresh hPBM were added to the co-culture and hSMC were also counted at 48-h intervals. Cells used in this experiment were from a 22 year old male and similar results were obtained with other isolates. * Indicates significant difference from no cell control values (PB0.05).
hibitory effects observed were due to lymphocytes in the co-culture. The effect of varying amounts of monocytes and lymphocytes on hSMC proliferation was therefore investigated. Fig. 3 shows that when the number of lymphoctyes:monocytes was increased, the inhibitory effect was lost. This experiment also revealed that relatively large numbers of monocytes were required to have a growth-inhibitory effect on the underlying hSMC. In experiments with lower numbers of monocytes (less than 0.5×106) and not supplemented with lymphocytes, the inhibitory effect was also lost (data not shown). An attempt was also made to obtain monocyte preparations of higher purity using magnetic beads with a monoclonal antibody to CD2 (to remove some of the lymphocytes). Monocytes further purified in this way resulted in small increases in purity i.e. ranging from 89–93% monocytes compared with lymphocytes. When these cells were incubated in co-culture with hSMC, inhibition of 3H-thymidine incorporation was again observed (control 104 8719 10 120 cpm/well, hPBM 62 9869 6980 cpm/well, PB 0.05).
3.4. Effect of differentiated human monocyte/macrophages on 3H-thymidine incorporation into hSMC Many studies have shown that incubation of freshly isolated monocytes with serum induces maturation of monocytes into macrophages [26,27]. It was found that monocytes which had been incubated in filter-inserts in 10% human AB serum for 6 days did not produce the
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
161
inhibitory effect seen with freshly isolated cultures (control 96 9209 9035 cpm/well, day 6 macrophages 104 6009 12 930 cpm/well, P \ 0.05). This indicates that the production of growth-inhibitory molecules by hPBM is reduced as monocytes mature in culture or that production of growth-stimulators overcomes the effect of growth-inhibitors.
3.5. Effect of indomethacin in the co-culture system To investigate the nature of the small molecular weight substance(s) produced by hPBM, an attempt was made to block candidate molecules, such as cyclooxygenase products. When indomethacin (10 mmol/l)was added during the co-culture of hPBM and hSMC, the growth-inhibitory effect was abolished (Fig. 4). We therefore concluded that cyclooxygenase products were involved in the growth-inhibition of hSMC. Since hSMC themselves have been shown to produce inhibitory prostaglandins [17], we investigated whether cyclooxygenase products from hPBM or hSMC were causing the growth-inhibitory effect. hPBM were pre-treated with indomethacin for 1 h and the monocytes were washed three times with serum-free medium before incubating in co-culture with hSMC. Pre-treatment of monocytes with indomethacin
Fig. 3. Effect of lymphocytes on 3H-thymidine incorporation into hSMC in filter-separated co-culture. Freshly isolated monocytes (86% monocytes/14% lymphocytes, by FACS analysis) were cultured together with varying amounts of lymphocytes (isolated after Ficollseparation and removal of monocytes by adherence, 95% by FACS analysis). The total number of leukocytes remained the same. These cells were incubated in 10% AB serum in co-culture with quiescent human SMC for 40 h. hSMC were pulsed with 3H-thymidine for 8 h and then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 2 year old male. Similar results were observed with other isolates from older males and females. * Indicates significant difference from no cell control and all other conditions in the above experiment (PB 0.05). M: hPBM, L: lymphocytes.
Fig. 4. Effect of indomethacin in the hPBM/hSMC co-culture. hPBM were incubated with quiescent human SMC in 10% AB serum for 40 h in the presence or absence of 10 mmol/l indomethacin. hSMC were pulsed with 3H-thymidine for 8 h then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 22 year old male and similar results were observed with other isolates. * Indicates significant difference from no cell control values (PB0.05).
prior to co-culture also abolished the growth-inhibitory effect on hSMC (Fig. 5), confirming that cyclooxygenase products from hPBM were causing the inhibitory effect. However, it has been shown that IL-1 can induce growth-inhibitory prostaglandins to be produced by hSMC in culture [17] and IL-1 has been shown to be a product of human macrophages in culture [28]. We therefore investigated whether IL-1 produced this effect
Fig. 5. Effect of pretreating hPBM with indomethacin before incubating hPBM in co-culture with hSMC. hPBM were incubated for 1 h in filter inserts with or without 10 mmol/l indomethacin and washed three times in serum-free medium before incubating with quiescent human SMC in 10% AB serum for 40 h. hSMC were pulsed with 3 H-thymidine for 8 h then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 42 year old male and similar results were observed with other isolates. * Indicates significant difference from no cell control values and from hPBM treated with indomethacin (P B0.05)
162
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
lipid metabolism, thrombosis, growth control, matrix accumulation and remodelling, and atherosclerotic plaque stability [29]. Since macrophages can produce both stimulators and inhibitors of SMC growth, these cells may take part in the control of SMC growth in atherogenesis. It is clear that macrophages secrete substances which have opposing effects on the function of target cells and that the net effect will depend, inter alia, on the proportion of these substances. In our previous studies, it was demonstrated that the conditioned medium from murine macrophages had either a modest stimulatory effect on SMC growth because it contained mitogens (\ 12 kDa) or an inhibitory effect
Fig. 6. Effect of IL-1b with or without indomethacin (10 mmol/l) on 3 H-thymidine incorporation into hSMC. Quiescent hSMC were incubated with 10% AB serum for 40 h in the presence or absence of IL-1b and/or indomethacin. hSMC were pulsed with 3H-thymidine for 8 h then TCA precipitated and radioactivity was counted. Cells used in this experiment were isolated from a 61 year old female and similar results were observed with other isolates.
in our culture system and if so, could it be blocked by indomethacin? Fig. 6 shows that IL-1b had no effect on hSMC 3H-thymidine incorporation, either in the presence or absence of indomethacin. It was therefore unlikely that IL-1 was the effector molecule released by hPBM and it was confirmed that the cyclooxygenase products causing the effect were produced from the monocytes in the co-culture. An inhibitor of thromboxane synthetase, Dazmegrel, was also included in the co-culture system (at a concentration of 50 mM) but had no effect on the growth-inhibition produced by hPBM (data not shown).
3.6. Effect of prostaglandins on hSMC growth To find out the effect of prostaglandins on hSMC growth, PGE1 and PGE2 were incubated with hSMC for 40 h and 3H-thymidine incorporation was measured (i.e. the same protocol for co-culture studies). Fig. 7 shows that PGE1 and PGE2 both inhibited 3Hthymidine incorporation into hSMC, and PGE1 was slightly more potent. To confirm whether these effects were on cell proliferation, PGE1 (1 mM) and PGE2 (3 mM) were added to hSMC at 48 hour intervals and cells were counted at these intervals (Fig. 8). Both PGE1 (1 mM) and PGE2 (3 mM) inhibited hSMC cell growth but only PGE1 showed significant inhibitory activity after 48 h. 4. Discussion Macrophages are believed to be involved in many processes including immune and inflammatory responses,
Fig. 7. Effect of prostaglandins E1 and E2 on hSMC 3H-thymidine incorporation. Quiescent hSMC were incubated with 10% AB serum for 40 h in the presence or absence of PGE1 (A) or PGE2 (B). hSMC were pulsed with 3H-thymidine for 8 hours then TCA precipitated and radioactivity was counted. Cells used in this experiment were isolated from a 28 year old female for (a) and a 61 year old female for (b) and similar results were observed with other isolates. * Indicates significant difference from no additions values (PB 0.05).
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
Fig. 8. Effect of prostaglandins E1 and E2 on hSMC growth. PGE1 (1 mM) or PGE2 (3 mM) was incubated with quiescent hSMC (72 h in 0.1% FCS) for periods up to 6 days. At 48 h intervals, fresh prostanoids were added to the co-culture and hSMC were also counted at 48 h intervals. Cells used in this experiment were from an 11 year old male and similar results were obtained with other isolates. * Indicates significant difference from no cell control values (PB 0.05).
because of the presence of small molecular weight inhibitors (B 12 kDa) [19,20]. In order to obtain insight into the balance of production of human macrophage secretory products, in the present studies we investigated the effect of human monocyte/macrophages on SMC proliferation. Freshly isolated human peripheral blood-derived monocytes (hPBM) produced a net inhibitory effect on hSMC proliferation as measured by 3H-thymidine incorporation and hSMC growth. This effect was still observed when monocytes were incubated in the dialysis-membrane separated co-culture, which indicated that the factor(s) responsible for the growth-inhibition is of less than 12 kDa in molecular mass. Contaminating lymphocytes in the monocyte preparations were not responsible for the inhibitory effect seen in freshly isolated cultures. The ratio of hPBM:hSMC necessary to achieve the inhibitory effect was greater than 25:1. This ratio of monocyte/macrophages:smooth muscle cells is very likely to occur in areas of macrophage accumulation such as in shoulder regions of the atherosclerotic plaque. However, it is difficult to draw too close a comparison between in vitro conditions and the artery wall. For example, the volume of fluid surrounding the cells is likely to be very different in the two situations and would therefore affect the concentration of any effector molecules. Other studies have demonstrated the release of growth-inhibitors from macrophages in co-culture. Rennick et al. [30] found that when mouse peritoneal macrophages were seeded at a high density, activated with Con A and grown in co-culture with rabbit SMC,
163
the overall effect on SMC proliferation was inhibitory. In the absence of Con A or when the macrophages were plated sparsely, SMC growth was stimulated. Fan et al. [31] found that macrophages isolated from mice fed with g-linolenic acid produced a growth-inhibitory effect on mouse SMC when grown in co-culture. The inhibitory effect was found to be partly due to cyclooxygenase metabolites but other uncharacterised factors released by the mouse macrophages were thought to be involved. Human monocytes have also been reported to produce uncharacterised growth-inhibitory activities specific for endothelial cells [32–34]. Halloran et al. [35], recently reported that the supernatant from human macrophages, both activated and non-activated, exhibited a growth-inhibitory effect on human aortic SMC growth. However, the growth-inhibitory activity was not characterised. In our previous studies with mouse macrophages, cyclooxygenase inhibitors had no effect on production of macrophage growth-inhibitory mediators. In the present study, further investigation of the nature of the growth-inhibitory effect produced by hPBM revealed that the activity was due to the release of cyclooxygenase products, since the growth-inhibitory activity was abolished in the presence of indomethacin. The principal product of the cyclooxygenase pathway in monocytes/macrophages is thromboxane A2 [36], but the thromboxane synthase inhibitor, Dazmegrel, had no effect on the production of growth-inhibitory activity by hPBM which rules out thromboxanes as effector molecules in this co-culture study. Pretreatment of hPBM with the cyclooxygenase inhibitor indomethacin confirmed that cyclooxygenase products from hPBM were involved in the growth inhibition, rather than products released from hPBM affecting hSMC eicosanoid production. This was further examined by treating hSMC with IL-1 which has been shown by Libby et al. [17] to induce growth-inhibitory prostanoid production in SMC. However, in our study, IL-1 had no effect on hSMC growth in the presence or absence of indomethacin. The differences observed in the two studies may be due to different culture conditions. In addition, evidence that hPBM prostaglandins inhibit hSMC growth was shown in experiments where prostaglandins were added to hSMC. Concentrations of prostaglandins which affected hSMC growth (PGE1 0.3-1 mM, PGE2 3 mM) would be expected to be produced by human monocytes/macrophages. Pawlowski and colleagues [36] summarised data from several different studies and estimated that freshly isolated monocytes synthesised 0.1 mM PGE2 in 24 h and that mature macrophages could produce in excess of 10 mM PGE2. The monocytes used in our studies were not deliberately activated but it would be expected that some activation towards a macrophage phenotype will have occurred with 48 hours of culture. This could be
164
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
brought about by the adherence to plastic or even in response to the presence of hSMC in the culture system. The mechanism of prostaglandin release from hPBM may involve either of the different forms of cyclooxygenase (COX-1 or COX-2). COX-2 is induced in monocytes when they are stimulated [37] and this may occur as a result of co-culture or early differentiation. In future studies, specific inhibitors could be used to determine the role of different COX isoforms in the hPBM effects on hSMC growth. Monocyte/macrophages are more commonly thought to produce growth factors only when appropriately stimulated. For example, Shimokado et al. [6], showed that alveolar macrophages adhered to plastic produced a mitogen for mesenchymal cells which could be augmented by the addition of Con A. As mentioned above, the monocytes used in the present study were freshly isolated and would therefore be expected initially to be in a non-activated state. However, it has been reported that monocytes become activated within 30 min of adherence when many of the genes for inflammatory regulators are induced [38]. Other studies suggest that monocytes develop into macrophages after about 1 week in culture when several differentiated features such as the acetyl low density lipoprotein receptor are present [27]. Since our co-culture experiments were conducted during the first 48 h of isolation of monocytes, it is likely that these cells were in the early stages of differentiation, which may mimic the state of a monocyte as it is entering the artery wall. Although little is known at present of the differentiation of a monocyte into a tissue macrophage, it does not seem unreasonable that the effect seen in our culture system may have a counterpart in the arterial wall in vivo during the early phase of entry of monocytes into the tissue. Fully mature macrophages and lipid-laden foam cells may secrete a different pattern of growth-regulating substances in vivo or in vitro. Indeed, foam cells have been reported to exhibit a marked reduction in eicosanoid production [39]. In our study, we found that incubation of hPBM for greater than 48 h lead to a loss of the growth-inhibitory activity. This indicates that there is either a net loss in production of inhibitors as the monocytes mature, or that production of growthstimulatory molecules is increased and thereby overcome the inhibitory effect. Measuring the balance between inhibitory and stimulatory influences in vivo would be a very difficult task but using our co-culture system we were able to investigate the balance of production of growth mediators from hPBM. Together, the results presented in this paper show that human monocyte/macrophages have the potential to inhibit human SMC growth. Our findings support the hypothesis that macrophages have an influence on the stability of the atherosclerotic plaque and also imply that cyclooxygenase inhibitors such as in-
domethacin and asprin may remove growth-inhibitory influences in the artery wall, promoting a more SMCrich environment which is favourable for plaque stability [40].
Acknowledgements The authors wish to thank Cheryl Smith for excellent technical assistance. We would like to thank RhoˆnePoulenc Rorer, Dagenham, UK, The Isaac Newton Trust and the British Heart Foundation for funding.
References [1] Gerrity RG. The role of the monocyte in atherogenesis 1. Transition of blood-borne monocytes into foam cells in fatty lesions. Am J Pathol 1981;103:181 – 90. [2] Gerrity RG. The role of the monocyte in atherogenesis 2. Migration of foam cells from atherosclerotic lesions. Am J Pathol 1981;103:191 – 200. [3] Faggiotto A, Ross R. Studies of hypercholesterolemia in the non human primate: II. Fatty streak conversion to fibrous plaque. Arteriosclerosis 1984;4:341 – 56. [4] Walker LN, Bowyer DE. Endothelial healing in the rabbit aorta and the effect of risk factors for atherosclerosis. Arteriosclerosis 1984;4:479 – 88. [5] Nathan CF. Secretory products of macrophages. J Clin Invest 1987;79:319 – 26. [6] Shimokado K, Raines EW, Madtes DK, Barrett TB, Benditt EP, Ross R. A significant part of macrophage-derived growth factor consists of at least two forms of platelet-derived growth factor. Cell 1985;43:277 – 86. [7] Baird A, Mormede P, Bo¨hlen P. Immunoreactive fibroblast growth factor in cells of peritoneal exudate suggests its identity with macrophage – derived growth factor. Biochem Biophys Res Commun 1985;126:358 – 64. [8] Metcalf D. Haematopoietic growth factors 1. Lancet 1989;i:825– 7. [9] Madtes D, Raines EW, Sakariasse K, Assoian RK, Sporn MB, Bell GI, Ross R. Induction of transforming growth factor-a in activated human alveolar macrophages. Cell 1988;53:285–93. [10] Higashiyama S, Abraham JA, Miller J, Fiddes JC, Klagsbrun M. A heparin-binding growth factor secreted by macrophage-like cells that is related to EGF. Science 1991;251:936 – 9. [11] Assoian RK, Sporn MB. Type b transforming growth factor in human platelets: Release during platelet degranulation and action on vascular smooth muscle cells. J Cell Biol 1986;102:1217– 23. [12] Roberts AB, Anzano MA, Wakefield LM, Roche NS, Stern DF. Type b transforming growth factor: A bifunctional regulator of cell growth. Proc Natl Acad Sci USA 1985;82:119 – 23. [13] Battegay EJ, Raines EW, Seifert RA, Bowen-Pope DF, Ross R. TGFb induces bimodal proliferation of connective tissue cells via complex control of an autocrine PDGF loop. Cell 1990;63:515– 24. [14] Sjo¨lund M, Nilsson J, Palmberg L, Thyberg J. Phenotype modulation in primary cultures of arterial smooth muscle cells: dual effect of prostaglandin E1. Differentiation 1984;27:158 –62. [15] Uehara Y, Ishimitsu T, Kimura K, Ikeda T, Sugimoto T. Regulatory effects of eicosanoids on thymidine uptake by vascular smooth muscle cells of rats. Prostaglandins 1988;36:847–57.
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165 [16] Nilsson J, Olsson AG. Prostaglandin E1 inhibits DNA synthesis in arterial smooth muscle cells stimulated with platelet-derived growth factor. Atherosclerosis 1984;53:77–82. [17] Libby P, Warner SJC, Friedman GB. Interleukin 1: a mitogen for human vascular smooth muscle cells that induces the release of growth-inhibitory prostanoids. J Clin Invest 1988;81:487 – 98. [18] Lyons CR, Orloff GJ, Cunningham JM. Molecular cloning and functional expression of an inducible nitric oxide synthase from a murine macrophage cell line. J Biol Chem 1992;267:6370– 4. [19] McMurray HF, Proudfoot D, Davis JB, Parrott DP, Bowyer DE. A small molecular weight inhibitor of growth of 3T3 cells and porcine aortic smooth muscle cells released from the macrophage cell line P388D1. J Cell Sci 1993;106:1301–11. [20] Proudfoot D, Parrott DP, Bowyer DE. A dialysis culture system for the study of the production and modulation of growth-regulatory molecules: studies using the P388D1 cell line. J Cell Sci 1995;108:379 – 86. [21] Bo¨yum A. Isolation of mononuclear cells and granulocytes from human blood: Isolation of mononuclear cells by one centrifugation by combining centrifugation and sedimentation at 1g. Scand J Clin Lab Invest 1986;21:77. [22] Denholm EM, Wolber FM. A simple method for the purification of human peripheral blood monocytes. a substitute for SeprecellMN. J Immunol Methods 1991;144:247–51. [23] Fujita H, Shimokado K, Yutani C, Takaichi S, Masuda J, Ogata J. Human neonatal and adult vascular smooth muscle cells in culture. Exp Mol Path 1993;58:25–39. [24] Grainger DJ, Kirschenlohr HL, Metcalf JC, Weissberg PL, Wade DP, Lawn RM. Proliferation of human smooth muscle cells promoted by lipoprotein(a). Science 1993;260:1655– 8. [25] Andreoli SP, Baechner RL, Bergstein JM. In vitro detection of endothelial cell damage using 2-deoxy-D-[3H]-glucose: Comparison with 51Cr, [3H]-leucine, [3H]-adenine and lactate dehydrogenase. J Lab Clin Med 1985;106:253–61. [26] Prieto J, Eklund A, Patarroyo M. Regulated expression of integrins and other adhesion molecules during differentiation of monocytes into macrophages. Cell Immunol 1994;156:191 – 211. [27] Geng Y-J, Kodama T, Hansson GK. Differential expression of scavenger receptor isoforms during monocyte-macrophage differentiation and foam cell formation. Arterioscler Thromb 1994;14:798 – 806.
.
165
[28] Dinarello CA. An update on human interleukin-1: from molecular biology to clinical relevance. J Clin Immunol 1985;5:287–97. [29] Libby P, Clinton SK. The role of macrophages in atherosclerosis. Curr Opin Lipidol 1993;4:355 – 63. [30] Rennick RE, Campbell JH, Campbell GR. Vascular smooth muscle phenotype and growth behavior can be influenced by macrophages in vitro. Atherosclerosis 1988;71:35 – 43. [31] Fan Y-Y, Ramos KS, Chapkin RS. Dietary g-linolenic acid modulates macrophage-vascular smooth muscle cell interactions: Evidence for a macrophage-derived soluble factor that down-regulates DNA synthesis in smooth muscle cells. Arterioscler Thromb Vasc Biol 1995;15:1397 – 403. [32] Vilette D, Detiadi H, Wautier MP, Caen J, Wautier JL. Identification of an endothelial cell growth-inhibitory activity produced by human monocytes. Exp Cell Res 1990;188:219 – 25. [33] Besner GE, Klagsbrun M. Macrophages secrete a heparin–binding inhibitor of endothelial cell growth. Microvascular Res 1991;42:187 – 97. [34] Loite F, Wautier MP, Kuntz D, Waultier JL. Inhibition of human endothelial cell growth by human monocytes in co-culture. Hematologie 1992;34:183 – 9. [35] Halloran BG, Grange JJ, Byung JS, Baxter BT. Macrophage products inhibit human aortic smooth muscle cell proliferation and alter 1a(I) procollagen expression. Ann Vasc Surg 1997;11:80 – 4. [36] Pawlowski NA. Arachadonic acid metabolism by human monocytes. In: Zembala M, Asherson GL, editors. Human monocytes. London: Academic Press, 1989:273 – 89. [37] Barrios-Rodiles M, Chadee K. Novel regulation of cyclooxygenase-2 expression and prostaglandin E-2 production by IFNgamma in human macrophages. J Immunol 1998;161:2441–8. [38] Sporn SA, Eierman DF, Johnston CE, Morris J, Martin G, Ladner M, Haskill S. Monocyte adherence results in selective induction of novel genes sharing homology with mediators of inflammation and tissue repair. J Immunol 1990;144:4434–41. [39] Arai H, Nagano Y, Narumiya S, Kita T. Decreased arachadonate metabolism in mouse peritoneal macrophages after foam cell transformation with oxidised low-density lipoproteins. J Biochem Tokyo 1992;112:482 – 7. [40] Lendon CL, Davies MJ, Born GVR, Richardson PD. Atherosclerotic plaque caps are locally weakened when macrophage density is increased. Atherosclerosis 1991;87:87–90.
Inhibition of human arterial smooth muscle cell growth by human monocyte/macrophages: a co-culture study D. Proudfoot a, C. Fitzsimmons b, J. Torzewski b, D.E. Bowyer b,* a
Department of Medicine, Uni6ersity of Cambridge, Addenbrooke’s Hospital, Le6el 5, Hills Road, Cambridge CB2 2QQ, UK b Department of Pathology, Uni6ersity of Cambridge, Tennis Court Road, Cambridge CB2 1QP, UK Received 5 June 1998; received in revised form 10 November 1998; accepted 21 January 1999
Abstract Monocyte/macrophages produce a variety of substances which may influence the function of smooth muscle cells (SMC). During atherogenesis, macrophages are thought to modulate SMC migration, proliferation and synthesis of extracellular matrix. Such modulation is the balance between stimulatory and inhibitory influences. Thus, for example, our earlier studies have shown that macrophages not only secrete mitogens, but also produce small molecular weight inhibitors of SMC proliferation. In the present study, we have used a co-culture system in which human monocyte/macrophages were separated from human arterial SMC (hSMC) by a filter with the optional addition of a 12 kDa cut-off dialysis membrane, in order to assess their effect on hSMC growth. We have found that human peripheral blood-derived monocytes produced a substance of B 12 kDa that inhibited hSMC growth in the co-culture system. The monocyte-derived factor causing this effect was completely blocked by indomethacin, indicating that growth-inhibitory factors produced by the monocytes were cyclooxygenase products. We have shown that PGE1 and PGE2 inhibit hSMC growth, making them likely candidates for the effector molecules released from monocytes in our co-culture system. © 1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Monocyte/macrophages; Smooth muscle; Co-culture
1. Introduction Accumulation of smooth muscle cells (SMC) in the arterial intima is found in arterial hyperplastic diseases, including atherosclerosis. The stimulus for the phenotypic modulation of the normally contractile SMC in the arterial media to a synthetic, secretory phenotype which can migrate, proliferate and secrete extracellular matrix in the arterial intima is largely unknown. Since monocyte infiltration appears to precede SMC proliferation in developing atherosclerotic lesions, factors released by monocyte/macrophages are thought to be involved [1–4]. Macrophages are known to release a wide range of substances, some of which are growth
* Corresponding author. Tel.: +44-1223-333722; fax: +44-1223333346.
regulatory factors and some which may affect matrix production [5]. The processes in atherogenesis are complex and are controlled by both stimulatory and inhibitory influences. Although macrophages secrete mitogens such as platelet-derived growth factor (PDGF) [6]; fibroblast growth factor (FGF) [7]; interleukin-1 (IL-1) [8], transforming growth factor a [9] and a heparin-binding epidermal growth factor-like molecule [10], there are also a number of macrophage products which have inhibitory effects on SMC growth. Transforming growth factor b (TGFb) has been reported to cause stimulation or inhibition depending on cell culture conditions [11,12] and on the concentration of TGFb [13]. Prostaglandins can inhibit SMC proliferation [14–16] and certain inhibitory prostanoids can be induced in SMC by IL-1 [17]. Macrophages also express an in-
0021-9150/99/$ - see front matter © 1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0 0 2 1 - 9 1 5 0 ( 9 9 ) 0 0 0 2 8 - 3
158
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
ducible form of nitricoxide synthase which produces the nitric oxide radical which can inhibit SMC proliferation [18]. Previous studies in our laboratory have shown that a small molecular weight substance of less than 10 kDa is released by murine macrophages which affects porcine SMC growth [19]. More recently, the production of the growth-inhibitory material was shown to be regulated by macrophage activating substances, for example, PMA [20]. It seems clear that growth-inhibitory factors produced by macrophages may play an important role in regulating the proliferation of arterial SMC and that the balance between the release of growth-inhibitors and mitogens may influence the development of atherosclerotic lesions. In order to investigate the balance of production of human monocyte/macrophage growth regulators, we have explored the net effects of freshly isolated peripheral blood-derived monocytes on human arterial SMC growth in a co-culture system.
2. Materials and methods
2.1. Materials The culture medium used was Dulbecco’s Modified Eagle’s Medium (DMEM) (Sigma) buffered with 3.7 g/l NaHCO3 and 5% CO2, and supplemented with 100 IU/ml penicillin (Sigma PEN-NA), 100 mg/ml streptomycin (Sigma S-6501) and 4 mM L-glutamine (Sigma G-3126). Heat-inactivated foetal calf serum (FCS) was purchased from Imperial Laboratories. Heat-inactivated human AB serum was purchased from the Blood Transfusion Service, Cambridge, UK. Indomethacin, PGE1 and PGE2 were purchased from Sigma, recombinant human IL-1b was purchased from R and D systems and Dazmegrel was a gift from Pfizer.
2.2. Cells 2.2.1. Human peripheral blood monocyte (hPBM) isolation Human monocytes were isolated from preparations of buffy coats purchased from the Blood Transfusion Service, Cambridge. Mononuclear cells were first isolated by Ficoll–Hypaque sedimentation [21] and subsequently further purified by Percoll density fractionation [22]. Monocytes and lymphocytes were identified by flow cytometry analysis according to their characteristic forward and side scatter on a Becton Dickinson FACScan flow cytometer. Purity was further checked by expression of CD14 (for monocytes) and CD45 (for leukocytes) using monoclonal antibodies (LeucoGATE CD45/ CD14, and its control, Becton Dickinson). The purity of monocytes ranged between 80 and 88%. In order to
attempt to obtain monocyte preparations of higher purity, magnetic beads with a monoclonal antibody to CD2 were used to remove most of the lymphocytes (Dynabeads M-450 Pan-T, CD2, Dynal). Percoll-separated monocytes were resuspended in PBS containing 2% FCS to a concentration of 1× 107 cells/ml and incubated with 4×107 dynabeads for 30 min at 4°C. The beads were removed by placing the suspensions in a magnetic field and the fluid containing cells was pipetted off very gently. Monocytes further purified in this way resulted in small increases in purity i.e. ranging from 89 to 93% monocytes. In some experiments lymphocytes were also isolated after the Ficoll–Hypaque separation by plating out cells in NUNC 24-well plates at a density of 2×106 cells/well and then removing non-adherent cells after a 1 hour incubation at 37°C. The purity of lymphocytes isolated in this way ranged between 95 and 99%.
2.2.2. Human smooth muscle cells Human smooth muscle cells (hSMC) were a gift from Professor P.L. Weissberg, Department of Medicine, Addenbrooke’s Hospital, Cambridge. hSMC used in this paper were obtained from aortae of individuals ranging from 2 to 65 years old. The cells were prepared from explants of tissue and were confirmed as smooth muscle cells by positive staining with monoclonal antibodies against aSM-actin (Sigma). Cells were maintained in DMEM containing 20% FCS and were used between passages 5 and 15. 2.2.3. HepG2 cells HepG2 (human hepatic cell line) were purchased from the ECACC and were maintained in DMEM containing 10% FCS. 2.3. Preparation of cells and dialysis membrane for co-culture 2.3.1. Macrophages hPBM (1× 106 cells/well) were plated into cell culture filter-inserts (Falcon, code. 3095) in 0.5 ml medium using a 24-well plate with medium alone in the lower well. This concentration of hPBM resulted in an almost confluent monolayer of adherent cells in the filter-insert. Filter inserts containing hPBM were then transferred into wells containing hSMC. This type of co-culture is referred to as ‘filter-separated co-culture’. In some experiments, a 12 kDa cut-off membrane was placed around the culture-insert. This is referred to as ‘dialysis membrane separated co-culture’. 2.3.2. Smooth muscle cells hSMC were plated in 24-well plates (NUNC) at a density of 20 000 cells/well in 1 ml DMEM containing 20% FCS. After an overnight incubation the medium was changed to 0.1% FCS and the incubation was continued for a further 72 h in order to induce quiescence.
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
2.3.3. Preparation of 12 kDa cut-off dialysis membrane Dialysis membrane of 12 kDa cut-off (Spectra-Por) was boiled in 2% Na2CO3 and 1 mM EDTA for 10 min followed by boiling in distilled water for 10 min. The tubing was then cut into 2 cm2 pieces and placed in distilled water and autoclaved at 110°C/10 psi for 10 min. Dialysis membrane was bathed in DMEM prior to placing over a well of a 24-well plate and inserting a filter-insert. 2.4. Measurement of 3H-thymidine incorporation into hSMC Following a 40-h co-culture of hSMC with monocyte/macrophages, the filter-inserts containing the monocyte/macrophages were removed and the hSMC in the bottom of the culture wells were pulsed with 3 H-thymidine (2 mCi/well, NEN specific activity 6.7 Ci/mmol) for 8 h. Controls for these experiments included: (a) hSMC exposed to medium containing FCS or AB serum; (b) hSMC exposed to serum-free DMEM (to demonstrate that the cells were quiescent); (c) hSMC exposed to medium containing FCS or AB serum and a filter-insert with no cells (referred to as ‘no cell control’). The time points chosen for these experiments were used as the hSMC were still in S-phase during 40–48 h after serum-stimulation. Other studies on human vascular SMC proliferation have demonstrated a cell doubling time of \ 40 h [23,24]. Also, hPBM were removed before pulsing hSMC with 3Hthymidine in order to eliminate any uptake of thymidine by the hPBM. Cells were harvested by placing plates on ice, washing with PBS, fixing with methanol (10 min), precipitating with 10% TCA (20 min) and digesting in 200 ml of 0.3 M NaOH for 30 min. The cell digest was then placed in 3 ml HiSafe scintillation fluid and radioactivity was measured.
159
above and after 70 h in quiescing medium, the cells were loaded with 1 mCi/ml 3H-adenine (Amersham) in 0.1% FCS in DMEM and the incubation was continued for 2 h. This medium was then removed, the cells were washed four times in PBS and then co-cultured with macrophages as above. After 40 h, the medium from each well was removed and placed into scintillation vials containing 3 ml HiSafe scintillation fluid. The wells were then washed in PBS and 100 ml aliquots of washings were added to separate scintillation vials. Finally, the cells remaining in each well were treated with 0.1% Triton X-100 in PBS for 5 min in order to lyse the cells. 100 ml of lysate was also added to scintillation vials and counted for 3H on a scintillation counter. Streptolysin S (Sigma, 100 units/well) was included as a positive control. Percentage release of 3 H-adenine was calculated as follows: %3H-adenine release =
supernatant+ washings ×100 supernatant+ washings+lysate
2.6.2. Trypan blue Trypan blue exclusion was also used to estimate toxicity. Cells were washed three times in phosphate buffered saline (PBS) and exposed to trypan blue (0.25% in PBS, Sigma). 2.7. Statistical analysis Results are expressed as the mean9the standard error of the means (S.E.M.). Each individual experiment was carried out in triplicate wells. The results of three individual experiments gave rise to the means and standard errors in each graph. Sample means were compared using a Student’s t-test. PB 0.05 was considered statistically significant.
2.5. Measurement of change in hSMC number In order to measure change in hSMC number, hSMC and monocytes or macrophages were incubated in coculture as for the 3H-thymidine assay for periods up to 144 h. At intervals of 48 h, filter-inserts were removed and the hSMC in triplicate wells were trypsinised and counted on a haemocytometer. When cells were incubated for these longer periods, medium was changed for fresh medium at each further 48-h time point.
2.6. Toxicity assays 2.6.1. 3H-adenine release In order to determine whether the hSMC had been damaged by co-incubation with macrophages, a 3Hadenine release assay was performed using a modified method by Andreoli et al. [25]. hSMC were prepared as
3. Results
3.1. Effect of freshly isolated human monocytes on hSMC growth in co-culture To find out the net effect of human peripheral blood monocyte/macrophages (hPBM) on hSMC growth, hPBM were placed in filter-inserts and co-cultured with hSMC. Fig. 1 shows that freshly isolated human monocytes inhibit human SMC proliferation when grown in filter-separated co-culture, as measured by 3Hthymidine incorporation into hSMC. Human monocytes also had a growth-inhibitory effect when a fractionating, 12 kDa membrane was present during co-culture, indicating that the activity was less than 12 kDa in molecular mass (Fig. 1). In these co-culture
160
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
experiments a control cell type, HepG2, had no effect on 3H-thymidine incorporation into hSMC (Fig. 1). HPBM also significantly inhibited hSMC growth as assessed by counting cell numbers over time (Fig. 2). Freshly-isolated hPBM were added in the co-culture system at 48-h intervals.
3.2. Assessment of potential toxicity of hPBM to hSMC To investigate whether the growth-inhibitory effect of hPBM was due to toxicity, 3H-adenine release was measured from hSMC as an estimate of cell damage. 3 H-adenine release from hSMC was 21.259 0.25% for control cells, 16.65 92.35% for hSMC exposed to hPBM, and 39.09 1.5% for hSMC exposed to a positive control (Streptolysin S). The growth-inhibitor(s) released by hPBM were therefore not toxic towards hSMC. Also, greater than 99% of hSMC exposed to monocytes excluded trypan blue in either filter-separated or dialysis membrane-separated co-culture.
3.3. Influence of lymphocytes Since monocyte isolates contain significant numbers of lymphocytes, it was possible that the growth-in-
Fig. 1. Effect of hPBM on 3H-thymidine incorporation into hSMC in filter- and dialysis membrane separated co-culture. Freshly isolated human peripheral blood-derived monocytes, isolated by the Percoll method (83% monocytes/17% lymphocytes, by FACS analysis), or HepG2 cells were incubated with quiescent human SMC in 10% AB serum for 40 h in the presence or absence of a 12 kDa cut-off membrane. hSMC were pulsed with 3H-thymidine for 8 h then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 2 year old male. Similar results were observed with other isolates from older males and females. * Indicates significant difference from no cell control values (PB0.05).
Fig. 2. Effect of hPBM on hSMC growth. hPBM were incubated in filter-separated co-culture (as in Fig. 1) with quiescent hSMC (72 h in 0.1% FCS) for periods up to 6 days. At 48-h intervals, fresh hPBM were added to the co-culture and hSMC were also counted at 48-h intervals. Cells used in this experiment were from a 22 year old male and similar results were obtained with other isolates. * Indicates significant difference from no cell control values (PB0.05).
hibitory effects observed were due to lymphocytes in the co-culture. The effect of varying amounts of monocytes and lymphocytes on hSMC proliferation was therefore investigated. Fig. 3 shows that when the number of lymphoctyes:monocytes was increased, the inhibitory effect was lost. This experiment also revealed that relatively large numbers of monocytes were required to have a growth-inhibitory effect on the underlying hSMC. In experiments with lower numbers of monocytes (less than 0.5×106) and not supplemented with lymphocytes, the inhibitory effect was also lost (data not shown). An attempt was also made to obtain monocyte preparations of higher purity using magnetic beads with a monoclonal antibody to CD2 (to remove some of the lymphocytes). Monocytes further purified in this way resulted in small increases in purity i.e. ranging from 89–93% monocytes compared with lymphocytes. When these cells were incubated in co-culture with hSMC, inhibition of 3H-thymidine incorporation was again observed (control 104 8719 10 120 cpm/well, hPBM 62 9869 6980 cpm/well, PB 0.05).
3.4. Effect of differentiated human monocyte/macrophages on 3H-thymidine incorporation into hSMC Many studies have shown that incubation of freshly isolated monocytes with serum induces maturation of monocytes into macrophages [26,27]. It was found that monocytes which had been incubated in filter-inserts in 10% human AB serum for 6 days did not produce the
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
161
inhibitory effect seen with freshly isolated cultures (control 96 9209 9035 cpm/well, day 6 macrophages 104 6009 12 930 cpm/well, P \ 0.05). This indicates that the production of growth-inhibitory molecules by hPBM is reduced as monocytes mature in culture or that production of growth-stimulators overcomes the effect of growth-inhibitors.
3.5. Effect of indomethacin in the co-culture system To investigate the nature of the small molecular weight substance(s) produced by hPBM, an attempt was made to block candidate molecules, such as cyclooxygenase products. When indomethacin (10 mmol/l)was added during the co-culture of hPBM and hSMC, the growth-inhibitory effect was abolished (Fig. 4). We therefore concluded that cyclooxygenase products were involved in the growth-inhibition of hSMC. Since hSMC themselves have been shown to produce inhibitory prostaglandins [17], we investigated whether cyclooxygenase products from hPBM or hSMC were causing the growth-inhibitory effect. hPBM were pre-treated with indomethacin for 1 h and the monocytes were washed three times with serum-free medium before incubating in co-culture with hSMC. Pre-treatment of monocytes with indomethacin
Fig. 3. Effect of lymphocytes on 3H-thymidine incorporation into hSMC in filter-separated co-culture. Freshly isolated monocytes (86% monocytes/14% lymphocytes, by FACS analysis) were cultured together with varying amounts of lymphocytes (isolated after Ficollseparation and removal of monocytes by adherence, 95% by FACS analysis). The total number of leukocytes remained the same. These cells were incubated in 10% AB serum in co-culture with quiescent human SMC for 40 h. hSMC were pulsed with 3H-thymidine for 8 h and then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 2 year old male. Similar results were observed with other isolates from older males and females. * Indicates significant difference from no cell control and all other conditions in the above experiment (PB 0.05). M: hPBM, L: lymphocytes.
Fig. 4. Effect of indomethacin in the hPBM/hSMC co-culture. hPBM were incubated with quiescent human SMC in 10% AB serum for 40 h in the presence or absence of 10 mmol/l indomethacin. hSMC were pulsed with 3H-thymidine for 8 h then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 22 year old male and similar results were observed with other isolates. * Indicates significant difference from no cell control values (PB0.05).
prior to co-culture also abolished the growth-inhibitory effect on hSMC (Fig. 5), confirming that cyclooxygenase products from hPBM were causing the inhibitory effect. However, it has been shown that IL-1 can induce growth-inhibitory prostaglandins to be produced by hSMC in culture [17] and IL-1 has been shown to be a product of human macrophages in culture [28]. We therefore investigated whether IL-1 produced this effect
Fig. 5. Effect of pretreating hPBM with indomethacin before incubating hPBM in co-culture with hSMC. hPBM were incubated for 1 h in filter inserts with or without 10 mmol/l indomethacin and washed three times in serum-free medium before incubating with quiescent human SMC in 10% AB serum for 40 h. hSMC were pulsed with 3 H-thymidine for 8 h then TCA precipitated and radioactivity was counted. The ‘no cell control’ wells contained hSMC and a filter insert with no cells. Cells used in this experiment were isolated from a 42 year old male and similar results were observed with other isolates. * Indicates significant difference from no cell control values and from hPBM treated with indomethacin (P B0.05)
162
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
lipid metabolism, thrombosis, growth control, matrix accumulation and remodelling, and atherosclerotic plaque stability [29]. Since macrophages can produce both stimulators and inhibitors of SMC growth, these cells may take part in the control of SMC growth in atherogenesis. It is clear that macrophages secrete substances which have opposing effects on the function of target cells and that the net effect will depend, inter alia, on the proportion of these substances. In our previous studies, it was demonstrated that the conditioned medium from murine macrophages had either a modest stimulatory effect on SMC growth because it contained mitogens (\ 12 kDa) or an inhibitory effect
Fig. 6. Effect of IL-1b with or without indomethacin (10 mmol/l) on 3 H-thymidine incorporation into hSMC. Quiescent hSMC were incubated with 10% AB serum for 40 h in the presence or absence of IL-1b and/or indomethacin. hSMC were pulsed with 3H-thymidine for 8 h then TCA precipitated and radioactivity was counted. Cells used in this experiment were isolated from a 61 year old female and similar results were observed with other isolates.
in our culture system and if so, could it be blocked by indomethacin? Fig. 6 shows that IL-1b had no effect on hSMC 3H-thymidine incorporation, either in the presence or absence of indomethacin. It was therefore unlikely that IL-1 was the effector molecule released by hPBM and it was confirmed that the cyclooxygenase products causing the effect were produced from the monocytes in the co-culture. An inhibitor of thromboxane synthetase, Dazmegrel, was also included in the co-culture system (at a concentration of 50 mM) but had no effect on the growth-inhibition produced by hPBM (data not shown).
3.6. Effect of prostaglandins on hSMC growth To find out the effect of prostaglandins on hSMC growth, PGE1 and PGE2 were incubated with hSMC for 40 h and 3H-thymidine incorporation was measured (i.e. the same protocol for co-culture studies). Fig. 7 shows that PGE1 and PGE2 both inhibited 3Hthymidine incorporation into hSMC, and PGE1 was slightly more potent. To confirm whether these effects were on cell proliferation, PGE1 (1 mM) and PGE2 (3 mM) were added to hSMC at 48 hour intervals and cells were counted at these intervals (Fig. 8). Both PGE1 (1 mM) and PGE2 (3 mM) inhibited hSMC cell growth but only PGE1 showed significant inhibitory activity after 48 h. 4. Discussion Macrophages are believed to be involved in many processes including immune and inflammatory responses,
Fig. 7. Effect of prostaglandins E1 and E2 on hSMC 3H-thymidine incorporation. Quiescent hSMC were incubated with 10% AB serum for 40 h in the presence or absence of PGE1 (A) or PGE2 (B). hSMC were pulsed with 3H-thymidine for 8 hours then TCA precipitated and radioactivity was counted. Cells used in this experiment were isolated from a 28 year old female for (a) and a 61 year old female for (b) and similar results were observed with other isolates. * Indicates significant difference from no additions values (PB 0.05).
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
Fig. 8. Effect of prostaglandins E1 and E2 on hSMC growth. PGE1 (1 mM) or PGE2 (3 mM) was incubated with quiescent hSMC (72 h in 0.1% FCS) for periods up to 6 days. At 48 h intervals, fresh prostanoids were added to the co-culture and hSMC were also counted at 48 h intervals. Cells used in this experiment were from an 11 year old male and similar results were obtained with other isolates. * Indicates significant difference from no cell control values (PB 0.05).
because of the presence of small molecular weight inhibitors (B 12 kDa) [19,20]. In order to obtain insight into the balance of production of human macrophage secretory products, in the present studies we investigated the effect of human monocyte/macrophages on SMC proliferation. Freshly isolated human peripheral blood-derived monocytes (hPBM) produced a net inhibitory effect on hSMC proliferation as measured by 3H-thymidine incorporation and hSMC growth. This effect was still observed when monocytes were incubated in the dialysis-membrane separated co-culture, which indicated that the factor(s) responsible for the growth-inhibition is of less than 12 kDa in molecular mass. Contaminating lymphocytes in the monocyte preparations were not responsible for the inhibitory effect seen in freshly isolated cultures. The ratio of hPBM:hSMC necessary to achieve the inhibitory effect was greater than 25:1. This ratio of monocyte/macrophages:smooth muscle cells is very likely to occur in areas of macrophage accumulation such as in shoulder regions of the atherosclerotic plaque. However, it is difficult to draw too close a comparison between in vitro conditions and the artery wall. For example, the volume of fluid surrounding the cells is likely to be very different in the two situations and would therefore affect the concentration of any effector molecules. Other studies have demonstrated the release of growth-inhibitors from macrophages in co-culture. Rennick et al. [30] found that when mouse peritoneal macrophages were seeded at a high density, activated with Con A and grown in co-culture with rabbit SMC,
163
the overall effect on SMC proliferation was inhibitory. In the absence of Con A or when the macrophages were plated sparsely, SMC growth was stimulated. Fan et al. [31] found that macrophages isolated from mice fed with g-linolenic acid produced a growth-inhibitory effect on mouse SMC when grown in co-culture. The inhibitory effect was found to be partly due to cyclooxygenase metabolites but other uncharacterised factors released by the mouse macrophages were thought to be involved. Human monocytes have also been reported to produce uncharacterised growth-inhibitory activities specific for endothelial cells [32–34]. Halloran et al. [35], recently reported that the supernatant from human macrophages, both activated and non-activated, exhibited a growth-inhibitory effect on human aortic SMC growth. However, the growth-inhibitory activity was not characterised. In our previous studies with mouse macrophages, cyclooxygenase inhibitors had no effect on production of macrophage growth-inhibitory mediators. In the present study, further investigation of the nature of the growth-inhibitory effect produced by hPBM revealed that the activity was due to the release of cyclooxygenase products, since the growth-inhibitory activity was abolished in the presence of indomethacin. The principal product of the cyclooxygenase pathway in monocytes/macrophages is thromboxane A2 [36], but the thromboxane synthase inhibitor, Dazmegrel, had no effect on the production of growth-inhibitory activity by hPBM which rules out thromboxanes as effector molecules in this co-culture study. Pretreatment of hPBM with the cyclooxygenase inhibitor indomethacin confirmed that cyclooxygenase products from hPBM were involved in the growth inhibition, rather than products released from hPBM affecting hSMC eicosanoid production. This was further examined by treating hSMC with IL-1 which has been shown by Libby et al. [17] to induce growth-inhibitory prostanoid production in SMC. However, in our study, IL-1 had no effect on hSMC growth in the presence or absence of indomethacin. The differences observed in the two studies may be due to different culture conditions. In addition, evidence that hPBM prostaglandins inhibit hSMC growth was shown in experiments where prostaglandins were added to hSMC. Concentrations of prostaglandins which affected hSMC growth (PGE1 0.3-1 mM, PGE2 3 mM) would be expected to be produced by human monocytes/macrophages. Pawlowski and colleagues [36] summarised data from several different studies and estimated that freshly isolated monocytes synthesised 0.1 mM PGE2 in 24 h and that mature macrophages could produce in excess of 10 mM PGE2. The monocytes used in our studies were not deliberately activated but it would be expected that some activation towards a macrophage phenotype will have occurred with 48 hours of culture. This could be
164
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165
brought about by the adherence to plastic or even in response to the presence of hSMC in the culture system. The mechanism of prostaglandin release from hPBM may involve either of the different forms of cyclooxygenase (COX-1 or COX-2). COX-2 is induced in monocytes when they are stimulated [37] and this may occur as a result of co-culture or early differentiation. In future studies, specific inhibitors could be used to determine the role of different COX isoforms in the hPBM effects on hSMC growth. Monocyte/macrophages are more commonly thought to produce growth factors only when appropriately stimulated. For example, Shimokado et al. [6], showed that alveolar macrophages adhered to plastic produced a mitogen for mesenchymal cells which could be augmented by the addition of Con A. As mentioned above, the monocytes used in the present study were freshly isolated and would therefore be expected initially to be in a non-activated state. However, it has been reported that monocytes become activated within 30 min of adherence when many of the genes for inflammatory regulators are induced [38]. Other studies suggest that monocytes develop into macrophages after about 1 week in culture when several differentiated features such as the acetyl low density lipoprotein receptor are present [27]. Since our co-culture experiments were conducted during the first 48 h of isolation of monocytes, it is likely that these cells were in the early stages of differentiation, which may mimic the state of a monocyte as it is entering the artery wall. Although little is known at present of the differentiation of a monocyte into a tissue macrophage, it does not seem unreasonable that the effect seen in our culture system may have a counterpart in the arterial wall in vivo during the early phase of entry of monocytes into the tissue. Fully mature macrophages and lipid-laden foam cells may secrete a different pattern of growth-regulating substances in vivo or in vitro. Indeed, foam cells have been reported to exhibit a marked reduction in eicosanoid production [39]. In our study, we found that incubation of hPBM for greater than 48 h lead to a loss of the growth-inhibitory activity. This indicates that there is either a net loss in production of inhibitors as the monocytes mature, or that production of growthstimulatory molecules is increased and thereby overcome the inhibitory effect. Measuring the balance between inhibitory and stimulatory influences in vivo would be a very difficult task but using our co-culture system we were able to investigate the balance of production of growth mediators from hPBM. Together, the results presented in this paper show that human monocyte/macrophages have the potential to inhibit human SMC growth. Our findings support the hypothesis that macrophages have an influence on the stability of the atherosclerotic plaque and also imply that cyclooxygenase inhibitors such as in-
domethacin and asprin may remove growth-inhibitory influences in the artery wall, promoting a more SMCrich environment which is favourable for plaque stability [40].
Acknowledgements The authors wish to thank Cheryl Smith for excellent technical assistance. We would like to thank RhoˆnePoulenc Rorer, Dagenham, UK, The Isaac Newton Trust and the British Heart Foundation for funding.
References [1] Gerrity RG. The role of the monocyte in atherogenesis 1. Transition of blood-borne monocytes into foam cells in fatty lesions. Am J Pathol 1981;103:181 – 90. [2] Gerrity RG. The role of the monocyte in atherogenesis 2. Migration of foam cells from atherosclerotic lesions. Am J Pathol 1981;103:191 – 200. [3] Faggiotto A, Ross R. Studies of hypercholesterolemia in the non human primate: II. Fatty streak conversion to fibrous plaque. Arteriosclerosis 1984;4:341 – 56. [4] Walker LN, Bowyer DE. Endothelial healing in the rabbit aorta and the effect of risk factors for atherosclerosis. Arteriosclerosis 1984;4:479 – 88. [5] Nathan CF. Secretory products of macrophages. J Clin Invest 1987;79:319 – 26. [6] Shimokado K, Raines EW, Madtes DK, Barrett TB, Benditt EP, Ross R. A significant part of macrophage-derived growth factor consists of at least two forms of platelet-derived growth factor. Cell 1985;43:277 – 86. [7] Baird A, Mormede P, Bo¨hlen P. Immunoreactive fibroblast growth factor in cells of peritoneal exudate suggests its identity with macrophage – derived growth factor. Biochem Biophys Res Commun 1985;126:358 – 64. [8] Metcalf D. Haematopoietic growth factors 1. Lancet 1989;i:825– 7. [9] Madtes D, Raines EW, Sakariasse K, Assoian RK, Sporn MB, Bell GI, Ross R. Induction of transforming growth factor-a in activated human alveolar macrophages. Cell 1988;53:285–93. [10] Higashiyama S, Abraham JA, Miller J, Fiddes JC, Klagsbrun M. A heparin-binding growth factor secreted by macrophage-like cells that is related to EGF. Science 1991;251:936 – 9. [11] Assoian RK, Sporn MB. Type b transforming growth factor in human platelets: Release during platelet degranulation and action on vascular smooth muscle cells. J Cell Biol 1986;102:1217– 23. [12] Roberts AB, Anzano MA, Wakefield LM, Roche NS, Stern DF. Type b transforming growth factor: A bifunctional regulator of cell growth. Proc Natl Acad Sci USA 1985;82:119 – 23. [13] Battegay EJ, Raines EW, Seifert RA, Bowen-Pope DF, Ross R. TGFb induces bimodal proliferation of connective tissue cells via complex control of an autocrine PDGF loop. Cell 1990;63:515– 24. [14] Sjo¨lund M, Nilsson J, Palmberg L, Thyberg J. Phenotype modulation in primary cultures of arterial smooth muscle cells: dual effect of prostaglandin E1. Differentiation 1984;27:158 –62. [15] Uehara Y, Ishimitsu T, Kimura K, Ikeda T, Sugimoto T. Regulatory effects of eicosanoids on thymidine uptake by vascular smooth muscle cells of rats. Prostaglandins 1988;36:847–57.
D. Proudfoot et al. / Atherosclerosis 145 (1999) 157–165 [16] Nilsson J, Olsson AG. Prostaglandin E1 inhibits DNA synthesis in arterial smooth muscle cells stimulated with platelet-derived growth factor. Atherosclerosis 1984;53:77–82. [17] Libby P, Warner SJC, Friedman GB. Interleukin 1: a mitogen for human vascular smooth muscle cells that induces the release of growth-inhibitory prostanoids. J Clin Invest 1988;81:487 – 98. [18] Lyons CR, Orloff GJ, Cunningham JM. Molecular cloning and functional expression of an inducible nitric oxide synthase from a murine macrophage cell line. J Biol Chem 1992;267:6370– 4. [19] McMurray HF, Proudfoot D, Davis JB, Parrott DP, Bowyer DE. A small molecular weight inhibitor of growth of 3T3 cells and porcine aortic smooth muscle cells released from the macrophage cell line P388D1. J Cell Sci 1993;106:1301–11. [20] Proudfoot D, Parrott DP, Bowyer DE. A dialysis culture system for the study of the production and modulation of growth-regulatory molecules: studies using the P388D1 cell line. J Cell Sci 1995;108:379 – 86. [21] Bo¨yum A. Isolation of mononuclear cells and granulocytes from human blood: Isolation of mononuclear cells by one centrifugation by combining centrifugation and sedimentation at 1g. Scand J Clin Lab Invest 1986;21:77. [22] Denholm EM, Wolber FM. A simple method for the purification of human peripheral blood monocytes. a substitute for SeprecellMN. J Immunol Methods 1991;144:247–51. [23] Fujita H, Shimokado K, Yutani C, Takaichi S, Masuda J, Ogata J. Human neonatal and adult vascular smooth muscle cells in culture. Exp Mol Path 1993;58:25–39. [24] Grainger DJ, Kirschenlohr HL, Metcalf JC, Weissberg PL, Wade DP, Lawn RM. Proliferation of human smooth muscle cells promoted by lipoprotein(a). Science 1993;260:1655– 8. [25] Andreoli SP, Baechner RL, Bergstein JM. In vitro detection of endothelial cell damage using 2-deoxy-D-[3H]-glucose: Comparison with 51Cr, [3H]-leucine, [3H]-adenine and lactate dehydrogenase. J Lab Clin Med 1985;106:253–61. [26] Prieto J, Eklund A, Patarroyo M. Regulated expression of integrins and other adhesion molecules during differentiation of monocytes into macrophages. Cell Immunol 1994;156:191 – 211. [27] Geng Y-J, Kodama T, Hansson GK. Differential expression of scavenger receptor isoforms during monocyte-macrophage differentiation and foam cell formation. Arterioscler Thromb 1994;14:798 – 806.
.
165
[28] Dinarello CA. An update on human interleukin-1: from molecular biology to clinical relevance. J Clin Immunol 1985;5:287–97. [29] Libby P, Clinton SK. The role of macrophages in atherosclerosis. Curr Opin Lipidol 1993;4:355 – 63. [30] Rennick RE, Campbell JH, Campbell GR. Vascular smooth muscle phenotype and growth behavior can be influenced by macrophages in vitro. Atherosclerosis 1988;71:35 – 43. [31] Fan Y-Y, Ramos KS, Chapkin RS. Dietary g-linolenic acid modulates macrophage-vascular smooth muscle cell interactions: Evidence for a macrophage-derived soluble factor that down-regulates DNA synthesis in smooth muscle cells. Arterioscler Thromb Vasc Biol 1995;15:1397 – 403. [32] Vilette D, Detiadi H, Wautier MP, Caen J, Wautier JL. Identification of an endothelial cell growth-inhibitory activity produced by human monocytes. Exp Cell Res 1990;188:219 – 25. [33] Besner GE, Klagsbrun M. Macrophages secrete a heparin–binding inhibitor of endothelial cell growth. Microvascular Res 1991;42:187 – 97. [34] Loite F, Wautier MP, Kuntz D, Waultier JL. Inhibition of human endothelial cell growth by human monocytes in co-culture. Hematologie 1992;34:183 – 9. [35] Halloran BG, Grange JJ, Byung JS, Baxter BT. Macrophage products inhibit human aortic smooth muscle cell proliferation and alter 1a(I) procollagen expression. Ann Vasc Surg 1997;11:80 – 4. [36] Pawlowski NA. Arachadonic acid metabolism by human monocytes. In: Zembala M, Asherson GL, editors. Human monocytes. London: Academic Press, 1989:273 – 89. [37] Barrios-Rodiles M, Chadee K. Novel regulation of cyclooxygenase-2 expression and prostaglandin E-2 production by IFNgamma in human macrophages. J Immunol 1998;161:2441–8. [38] Sporn SA, Eierman DF, Johnston CE, Morris J, Martin G, Ladner M, Haskill S. Monocyte adherence results in selective induction of novel genes sharing homology with mediators of inflammation and tissue repair. J Immunol 1990;144:4434–41. [39] Arai H, Nagano Y, Narumiya S, Kita T. Decreased arachadonate metabolism in mouse peritoneal macrophages after foam cell transformation with oxidised low-density lipoproteins. J Biochem Tokyo 1992;112:482 – 7. [40] Lendon CL, Davies MJ, Born GVR, Richardson PD. Atherosclerotic plaque caps are locally weakened when macrophage density is increased. Atherosclerosis 1991;87:87–90.