The effects of an atherogenic diet on macrophage/biomaterial interactions H o w a r d P. Greisler, M D , Joan Ellinger, Scott C. H e n d e r s o n , Anne M. Shaheen, Wilson H. Burgess, PhD, Dae U n Kim, M D , and Tina M. Lam, ME), Hines and
Maywood, Ill., Rockville, Aid., and Livingston, N.J. We previously reported that biomaterials differentially induce macrophages to secrete growth factors that mediate reendothelialization. The present study evaluates the effect of an atherogenic diet on macrophage/biomaterial interactions. Female New Zealand white rabbits were fed an atherogenic diet. Peritoneal macrophages were harvested from these as well as rabbits fed a normal diet and culntred in Minimum Essential Medium with platelet-poor sernm. Dacron or polyglactin 910 were added to two of three conditions of both cell groups in passage 2. Conditioned media were collected weekly through week 15. Mitogenicity assays were performed with quiescent mouse embryonal (BALB/cST3) fibroblasts, atherosclerotic rabbit aortic smooth muscle cells, and murine capillary lung (LE-II) endothelial cells. Mitogenic activity was assayed by scintillation counting of tritiated thymidine incorporation into deoxyribonucleic acid (DNA). Results showed increased mitogenic activity released by macrophages from atherosclerotic rabbits, in the absence of prosthetic material, when assayed against every cell line. In normal diet macrophages, polyglactin 910 stimulated mitogen release for every cell line, and Dacron yielded minimal mitogen release. In lipid diet macrophages polyglactin 910 slightly increased mitogen release for all three cell lines, whereas Dacron resulted in stimulation of DNA synthesis in smooth muscle cells and BALB/c3T3 cells but less DNA synthesis in LE-II cells than in control, no graft material, media. Western blotting demonstrated immunoreactivity to basic fibroblast growth factor in media from normal diet macrophages but only in the presence of polyglactin 910 or Dacron. Radioimmunoassay for platelet-derived growth factor B chain was negative in all groups, and polymerase chain reaction techniques to amplify transforming growth factor-beta messenger ribonucleic was negative. These data demonstrate the effect of in vivo dietary manipulation on macrophage activation as well as the effect of an atherogenic diet in modulating macrophage/biomaterial interactions. Additionally, different biomaterials differentially induce macrophages to release factors that stimulate and inhibit growth. (J Vxsc StYRG 1991;14:10-23.)
Small diameter vascular grafts fail not because of the biomaterials themselves but because of the biologic reactions occurring at the blood/material and tissue/material interfaces. Prominent among the foreign body reactions is monocyte recruitment and activation and differentiation into the activated macrophage. Macrophages when appropriately activated are capable of synthesizing a variety of growth From Loyola University Medical Center, Maywood, Hines Veterans AdministrationHospital, Hines, The American Red Cross (Ms. Shaheen,Dr. Burgess),Rockville,and St. Barnabas Medical Center (Dr. Fdm), Livingston. Supported by a grant from the NationalInstitutesof Health, No. RO1 HL41272. Reprint requests: Howard P. Greisler, MD, Loyola University Medical Center, Department of Surgery, 2160 S. First Ave., Maywood,IL 60153. 24/1/274~8 10
regulatory factors that modulate migration and proliferation of endothelial cells, smooth muscle cells, and fibroblastsY s Our laboratory has previously shown that a variety of bioresorbable materials including polyglactin 910 (PG910) (Ethicon, Inc., Somerville, N.J.), when woven into vascular graft configurations and implanted into animal models, elicit extensive transinterstitial ingrowth of capillaries, endothelial cells, and myofibroblasts to a significantly greater extent than does similarly woven Dacron. 6-1° The tissue ingrowth begins shortly after macrophage infiltration and phagocytosis occurs and terminates when macrophage-mediated prosthetic resorption is complete. We have also shown in past studies that macrophages cultured in vitro in the presence of PG910 release into their media factor(s) capable of stimulating endothelial cell proliferation to
Volume 14 Number ] July 1991
a significantly greater extent than do macrophages cultured with Dacron. ~'5 However, when PG910 prostheses are implanted into rabbits previously made atherosclerotic by feeding of a 2% cholesterol, 6% peanut oil diet, the tissue ingrowth is markedly attenuated, resulting in thin walls containing lipidladen macrophages bull relatively few myofibroblasts and endothelial cells. The current study evaluates the effect of an atherogenic diet on macrophage production of growth factor activity and further examines the effect of such a diet on macrophage/biomaterial interactions. These interactions are likely central to the tissue reactions that occur after implantation of any blood-contacting biomaterial, and the modulation of these interactions by atherosclerosis must be examined. METHODS Dietary regimens
Adult female New Zealand white rabbits weighing 3 to 4 kg were housed in a facility accredited by the American Association for Accreditation of Laboratory Animal Care (AAALAC) and fed either standard Purina Rabbit Chow or Purina Rabbit Chow (Ralston Purina Co., St. Louis, Mo.) specially formulated to provide a 2% cholesterol, 6% peanut oil diet. Before institution of the diet, baseline serum cholesterol and triglyceride levels were measured. Macrophages were harvested after 2 months at which time serum cholesterol and triglyceride levels were again determined. Macrophage cuhxtre Peritoneal macrophages were harvested from eight rabbits of each group via peritoneal lavage with Hank's balanced salt solution containing calcium and magnesium and with 10 units/ml heparin by our previously published technique. 4 Collected cells were centrifuged at 500g for 10 minutes and washed three times with phosphate-buffered saline (PBS). Cells from the six rabbits per group were pooled and suspended in Minimum Essential medium (MEM) with 10% equine platelet-poor serum. Collected cells, 1 × 106 cells/ml, were seeded into three T-25 Falcon (Becton-Dickinson Labware, Lincoln Park, N.J.) tissue culture flasks containing 3 ml of MEM. Cell viability and[ cell counts were determined by trypan blue dye exclusion and a hemocytometer. After a 2-hour adherence separation, adherent cells were washed with MEM and fed with 5 ml of complete media. Ambient conditions were humidified with 5% CO2,. Macrophages in each group were
Atherogenic diet and macvophage biomaterial interaction 11
then serially passaged twice resulting in nine T-75 flasks of macrophages for both dietary groups. Finely shredded 17G910 or Dacron particles were prepared by scalpel technique from the vascular prostheses used for previously published in vivo studies, s These materials were then added each to one-third of the nine T-75 flasks of both of the two groups of macrophages at a concentration of 1.6 mg/cm 2. Macrophage cultlare purity was determined by three methods: morphologic characteristics under phase contrast microscopy, nonspecific esterase staining by the method of'Yam et al.,11 and identification of Fc (f?agment crystallizable immunoglobulin G) receptors on the cell membrane by the method of Bianco and Pytowski. 12 Flasks were routinely fed three times weekly and examined for macrophage/biomaterial interactions under phase contrast microscopy. For overall flow chart see Fig. 1. Collection of conditioned media The cells remained viable in all three conditions of both groups of macrophages at subconfluent levels, with relatively small :numbers of mitotic figures and minimal polyploidy observed. Macrophages harvested from the group fed an atherogenic diet were noticeably more lipid filled. After 5 weeks in culture, intracytoplasmic PG910 inclusions were seen within the macrophages of both dietary groups (Fig. 2). By contrast, the cells adhered to but did not phagocytose the Dacron. Beginning at the time of addition of the prosthetic materials, conditioned medium from each flask was collected at the time ofrefeeding three times weekly. These three media collections were pooled and frozen to - 7 0 ° C yielding three groups of conditioned media for each of the two groups of harvested macrophages for each week in culture. Media were later thawed for analysis of growth promoting activity against test quiescent cells. BALB/c3T3 bioassay BALB/c3T3 mouse embryonic fibroblasts, kindly provided by Patricia D'Amore, MD, were cultured in complete media containing Dulbecco's modified Eagle's medium (DMEM) (glucose 4.5 gm/1) supplemented with 10% Colorado calf serum. The BALB/c3T3 bioassay is a well-established quantitative bioassay responsive to many growth factors and is therefore a good initial screen. Cells were removed with trypsin from a T-75 culture flask at 75% confluence and were counted and assessed for viability by use of trypan blue dye exclusion. Cells were
Journal of VASCULAR SURGERY
12 Greisler ctal.
A
~
N
Z
W
R
A
~
I MACROPHAGE
MACROPHAGE
HARVEST
cno. I
HARVEST
;TION
B
]
I M~a cou~c~oNs
BALB/c3T3 Mo0,Eni. o:II B o~ss,v
/
I
LE-II I T.EL~L CELL
SMC ( , ~ E . o -
SCLEROTICRABBIT AORTICB )IOASSAY
/
I
(6)
I /
/
/
RADIOIMM UNOASSAY - PDGFBCHAIN
/ \\ \
\
WESTERNBLOTS - HBGF-1 - HBGF-2 - PDGFBCHAIN
Fig. 1. Flow chart of experimental design. then seeded into 96-well microtiter plates at a density of 7000 cells/well. These cells were fed complete media and were then incubated at 37 ° C in humidified 10% CO2 for 5 days without media change, which resulted in quiescence. On the sixth day after seeding, 50 m m 3 of PBS were added to the negative control wells, 50 m m 3 of Colorado calf serum were added to the positive control wells, and 50 m m 3 of the undiluted or 1:10 diluted conditioned media were added to the test wells, each in replicates of four. 3H thymidine (3H TdR), 1 txCi/well, was then added to each well and allowed to incubate for 48 hours. The wells were then processed for scintillation
counting of tritiated thymidine incorporation into newly synthesized deoxyribonucleic acid (DNA). Cells were washed vigorously with normal saline three times and then fixed twice with 250 m m 3 of 100% methanol at 4 ° C for 5 minutes followed by vigorous washing with distilled water three times. Wells were then incubated twice in 5% trichloroacetic acid (TCA) for 10 minutes, again vigorously washed, and the acid precipitable material was then solubilized with 150 mm 3 0.3 N N a O H at room temperature for 20 minutes. Each 150 m m 3 sample was then transferred to 10 ml of ready protein scintillation fluid and counted in a Beckman LS6800
Volume 14 Ntunber 1 July 1991
Atherogenic diet and macrophage biomaterial interaction 13
Fig. 2. A, Phase-contrast photomicrograph of peritoneal macrophages heLrvestedfrom normal New 2"ealand white rabbits after 6 weeks in culture with PG910. Intracytoplasmic PG910 inclusions can be seen. (Original magnification × 500.) B, Phase-contrast photomicrograph of peritoneal macrophages harvested from atherosclerotic New Zealand white rabbits cultured for 8 weeks in the presence of PG910. Intracytoplasmic PG910 inclusions can be seen. No gross morphologic differences are seen comparing the macrophages from the two groups of rabbits. (Original magnification × 500.)
(Beckman Instruments, Inc., Fullerton, Calif.) scintillation counter. Atherosclerotic rabbh: aortic smooth muscle
cell bioassay One adult New Zealand white rabbit fed an atherogenic diet as above for 3 months was anesthetized with intravenous 2% methohexital, and the thoracic and abdominal aorta were sterilely resected and washed vigorously with Hank's balanced salt solution to remove blood. The aorta was then opened longitudinally and the endothelium removed by
gentle passage of a :scalpel blade. At the time of harvest the serum cholesterol and triglyceride levels in this rabbit measured 2130 mg/dl and 110 mg/dl, respectively. The adventitia was removed by careful dissection, and the remaining m e d i u m was minced into 1 mm 3 fragments and suspended in i ml of complete media. Complete media consisted of D M E M (glucose 4.5 gm/1) in 10% fetal bovine serum (FBS) supplemented with nonessential amino acids, sodium pyruvate, penicillin/streptomycin, and gentamicin. The suspension was transferred into a T-25 culture flask and incubated for 4 hours at 37 ° C
14
Greisler et al.
in humidified 5 % CO2 to allow adherence, followed by the addition of 1 ml of complete media. Cells were refed three times weekly and after their firm adherence, 5 ml of complete media were used per T-25 flask. The aortic smooth muscle cells migrated off the explants after 1 week and after confluence were subcultured twice. Smooth muscle cell identification was documented by immunofluorescent staining by use of a smooth muscle specific anti-alpha actin monoclonal antibody (Sigma Chemical Company, St. Louis, Mo.). Cells were seeded into 96-well microtiter plates at a density of 5000 cells/well and grown in complete media at 37 ° C in humidified 5% CO2 until confluent (3 days). The medium was then aspirated and the cells washed and refed with 200 mm 3 of serum-free media for 48 hours. Serum-free media consisted of a 1:1 mixture of DMEM and Hams F-12 media supplemented with transferrin (5 ~xg/ml), penicillin (100 units/ml), streptomycin (100 txg/ml), insulin (1 ~xmol/L [6 /xg/ml]), and ascorbate (0.2 ixmol/L [35.2 ~xg/ml]).13After 48 hours in serum-free media, 50 mm ~ of PBS were added to the negative control wells, 50 mm 3 of FBS to the positive control wells, and 50 mm 3 of either undiluted or 1:10 diluted conditioned media to the test wells, each in replicates of four. 3H-TdR, 1 ~xCi/well,was then added to each well and allowed to incubate for 48 hours. The wells were then processed for TCA precipitable material for scintillation counting of 3H-TdR incorporation into newly synthesized DNA as described above under the BALB/c3T3 bioassay.
LE-II bioassay Murine capillary endothelial (LE-II) cells, kindly provided by Thomas Maciag, MD, from an established cell line were identified by anti-factor VIIIrelated antibody staining and were grown to confluence in T-25 culture flasks in medium 199 with 10% fetal calf serum as previously published? This particular cell line was selected because its use has been well established for an endothelial cell mitogenesis assay. Cells were resuspended in complete media, and aliquots of 500 mm 3 were added to 48-well microtiter dishes. After 48 hours, quiescence of preconfluent wells was achieved by a 48-hour exposure to medium 199 with 2% serum. Quiescence was documented in replicate wells by scintillation counting of 3H-TdR incorporation into DNA 3 hours after a tritium pulse of 0.5 ixCi/ml. Thawed conditioned media were then added to the quiescing media in 1:10 or 1:20 dilutions in
Journal of VASCULAR SURGERY
replicate wells. These conditioned media included collections during even numbered weeks from the atherosclerotic diet group of macrophages from weeks 2 to 12. These results were compared with those we previously published using the weeks 5 to 10 collections of media from the normal diet group of macrophages. After 16 hours 3H-TdR incorporation was measured in triplicate wells for each group at each week by use of the method described above for scintillation counting of TCA precipitable material. Growth factor identification methods Conditioned media collected from all six macrophage groups during the tenth week of culture were analyzed by Western blotting techniques. These conditioned media were subjected to polyacrylamide gel electrophoresis (15% w/v) in the presence of sodium dodecyl sulfate (SDS-PAGE) using the Mighty Small gel apparatus (Hoeffer Scientific Instruments, San Francisco, Calif.). After electrophoresis, gels were stained with Coumassie brilliant blue R-250, dried, and exposed to x-ray film by our published techniques. 14'15 Gels were analyzed by Western blotting to determine immunoreactivity to polypeptides of appropriate apparent molecular weights by use of antibodies against platelet-derived growth factor (PDGF)-B chain, type 1 heparinbinding growth factor endothelial cell growth factorECGF and type 2 heparin binding growth factor. In addition, all groups of collected media from each week were assayed for the presence of PDGF-B chain by means of a highly specific radioimmunoassay (Amersham International, Arlington Heights, ILL). This assay used an iodine 125-labeled goat antihuman PDGF antibody sensitive to the range of 1.56 to 200 fmol/ml. We have demonstrated in other studies the cross-immunoreactivity of this antibody with PDGF-B chain produced by rabbits (Greisler H.P., unpublished observations).
Statistics Tritiated thymidine incorporation (counts per minute [cpm]) into newly synthesized DNA was calculated as a percent of maximum response (FBS positive control) and expressed as mean + standard deviation (SD), and the groups were statistically analyzed by the Student's t test. RESULTS The 2-month feeding of the a'cherogenic diet resulted in a significant increase (p < 0.001) in serum cholesterol levels with cholesterol and triglyc-
Volume 14 Number 1 luly 1991
Atherogenic diet and macrophage biomaterial interaction 15
Fig. 3. A, Photomicrograph from thoracic aorta of an atherosclerotic New Zealand white rabbit showing greatly thickened intima containing abundant foam cells. The elastic laminae of the media can be seen in the lower portion; the media contains smooth muscle cells with vacuolated cytoplasm. (Hematoxylin-eosin stain; original magnification ×400.) B, Higher magniification of intima outlined within box in A. (Hematox-ylin-eosin stain; original magniification × 400.) eride levels measuring 2840 _+ 339 and 167.5 _+ 86.1, respectively compared to 42.3 + 23.9 and 67.3 + 29.5 before institution of the diet. Gross atherosclerotic plaques in the aortas were evident, primarily in the aortic arch and in periosteal areas throughout the aortas. Histologically (Fig. 3, A and B) these areas showed extensive intimal thickening with fbam cells and cholesterol clefts. The macrophages harvested from the atherosclerotic rabbits contained greater numbers of lipid vacuoles when viewed under phase-contrast microscopy. BALB/c3T3 bioassay The BALB/c3T3 cells at the time of addition of test media were quiescent, documented by the PBS negative controls, 203.55 _+ 3455 cpm, and capable of DNA synthesis when stimulated, documented by the FBS positive controls, 237870 _+ 29122 cpm. Macrophages cultured in the absence of graft materials were mildly activated by the polystyrene culture flasks. Fig. 4, A and B shows the data as percent of maximal activity in 3H-TdR incorporation, with maximal activity being that in the FBS positive controls. Bo~h the normal and the athero-
sclerotic rabbit macrophages' media at each week produced small increases in 3H-TdR incorporation ranging from 11% to 26% of maximal activity. These data can also be expressed as an increase in 3H-TdR incorporation of 161% to 406% above quiescent (PBS) levels. In the absence of graft materials, the atherosclerotic group's media collected in weeks i to 7 contained significantly more mitogenic activity at each week (p < 0.001 in weeks 3 to 6) than the macrophages from normal rabbits. In later weeks these differences were abolished, largely because of increasing mitogenic activity in late weeks in the media of macrophages from normal rabbits. PG910 stimulated normal rabbits' macrophages to release significantly more mitogen into the media in weeks 1 to 7. To eliminate the effect of macrophage activation induced by polystyrene, the data are normalized to the samples at each week from macrophages cultured in the absence of graft materials. Thus Fig. 4 shows this normalized data with the mean percent increase above quiescence produced by media in the presence of a graft material divided by the mean increase above quiescence produced by the media collected that same week but in the absence of graft material. The increase in macrophage release of
Journal of VASCULAR SURGERY
16 Greisler et aL
~
7O m
60
GES
C
NO~.4~J7 ~ ~TA - ~
~
BIOMA'~L.~tK
5O
4
d 40 ~o
2o~
o "All'IEROSCLEROTIC" MACROPHAGES
0 n
c)
60 8ORilL'NUZI~ DATA-- ~
~
BIOMAIi'~4~AL~fK
b_ 50 0 40 F-Z W 30 C) rY 20 t.LI Cl 10
5 ~ J
1 ¢.ON~Ot_
o
2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
CULTURE DURATION (WEEKS) QUIESCE]WCE (PIBS) - 8.~1; OF PO~TIVE CONTROL F'B5 CPM: 237870 *: 29122 PBS CPM: 20355 :k 34,55
CONTROL (NO GRAFT MATERIA0 m PG910 ~DACRON
Fig. 4. BALB/c3T3 fibroblast bioassay results. Studies with conditioned media from macrophages harvested from normal New Zealand white rabbits are in the top two panels and those studies with conditioned media from macrophages harvested from atherosclerotic New Zealand white rabbits in the lower two panels. The panels on the left show the results of each group of conditioned media in each successive week, expressed as mean counts per minute sample + mean counts per minute fetal bovine serum positive control × 100. The right sided panels show the normalized data in which each week's media from the groups with prosthetic material is normalized by dividing its mean counts per minute by that of the media from the macrophages grown in the absence of prosthetic material, thus normalizing to polystyrene control values. Values > 1 indicate additional stimulation of mitogen release by the presence of the biomaterial, whereas values < 1 indicate less mitogen release than is seen in the absence of that biomaterial. (mean -+ SD)
mitogens caused by the interaction with PG910 was statistically significant (p - 0.013) each of the first 4 weeks. When expressed as percent of FBS positive control, the week 4 collection resulted in 42% of maximal stimulation. Dacron also induced a smaller increase in mitogen release in weeks 1 to 8, again statistically significant in weeks 1 to 4 (p - 0.01). However, at later time points, weeks 10 to 15, the media from macrophages cultured with Dacron resulted in significantly less (p < 0.001) mitogenic activity than in the absence of Dacron. This suggests either a decrease in mitogen release or a production of an inhibitor to fibroblast proliferation, or both. This late decrease in activity was not present in the bioassay on smooth muscle cells, which suggests that the macrophages in these late weeks remained competent to release mitogenic activity.
The 3H-TdR incorporation caused by PG910/macrophage media was significantly greater than that caused by Dacron/macrophage media in weeks4 (p < 0.05) andweeks 12to 15 (p < 0.005). Much less macrophage activation by either material is seen in the macrophages harvested from atherosclerotic rabbits. This increase in mitogenic activity above the polystyrene controls was significant (p < 0.001) only in week 1. No significant differences were found between the two materials on macrophages from atherosclerotic rabbits. The media from week 8 of the PG910 group was lost, and consequently this single data point is omitted in Fig. 4. In summary, these data show stimulation of fibroblast DNA synthesis by media from both groups ofmacrophages, but in the absence of graft materials (baseline conditions) significantly more mitogen
Volume 14 Number 1 July 1991
Atherogenic diet and macrophage biomaterial interaction 17
14
NORMALI~I~CROPH/~GES
A NORMALMACROPHAGE~
C
9 8
~'12
7 L,~IO
6 5
4 ~ ~_
3 IM 2
->o
1.1_10 0 F- 8 7 W 6
4 5 2
Q_ 2 0
0 2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
CULTURE DURATION (WEEKS) QUIESCENCE(PB$) - 0.6~ OF POSfIWE CONTROL F'BS CPM: 2 6 9 4 0 i 4.186 20
PB$ CPM: 165 i
[--'7 CONTROL (NO GRAFT MATERIAL) m PG910 DACRON
Fig. 5. Atherosclerotic rabbit aortic smooth muscle cell bioassay results. Studies that used conditioned media from macrophages harvested from normal New Zealand white rabbits are in the top two panels, and those studies that used conditioned media from macrophages harvested from atherosclerotic New Zealand white rabbits in the lower two panels. The panels on the left show the results of each group of conditioned media in each successive week, expressed as mean counts per minute sample + mean counts per minute fetal bovine serum positive control × 100. The right sided panels show the normalized data in which each week's media from the groups with prosthetic material is normalized by dividing its mean counts per minute by that of the media from the macrophages grown in the absence of prosthetic material, thus normalizing to polystyrene control values. Values > 1 indicate additional stimulation of mitogen release by the presence of the biomaterial while values < 1 indicate less mitogen release than is seen in the absence of that biomaterial. (mean +- SD) from those harvested from atherosclerotic rabbits. The a d d i f o n of either graft material to the mcdia increased mitogen release in early weeks but significantly more so by PG910. In late weeks Dacron appeared[ to inhibit release of mitogenic activity by macrophages harvested from normal rabbits. Atherosclerotic rabbit s m o o t h muscle
cell bioassay Quiescence of the smooth muscle cells was documented by the PBS negative controls, 165 --_ 20 cpm, and the capacity of quiescent cells to be stimulated to synthesize D N A documented by the FBS positive controls 26940 ___ 4186 cpm. Fig. 5 shows the mild effect of polystyrene on both groups of rnacrophages in this bioassay. There is a small but persistent mitogenic activity in the
media of macrophages harvested from both normal and atherosclerotic rabbits and cultured in the absence of graft materials. These increases in ~H-TdR incorporation are smaller (1% to 4% of maximal activity) than those !Found with the more sensitive BALB/c3T3 cell line, but still represent 164% to 601% increases above quiescent levels. As was true also in the BALB/c3T3 bioassay, the atherosclerotic group's media possessed more mitogenic activity, but these differences were significant only in week 1 (p = 0.009). The week 2 difference approached significance (p = 0.059). PG910 stimulated normal rabbits' macrophages to release significantly more mitogen into the media (Fig. 5) in weeks 1 to 5, reaching significance in weeks 3 to 5 (p -< 0.,005). In week 4 this increase in smooth muscle cell mitogen release induced by PG910 was 5.68 times that of the polystyrene
18
jrournal of VASCULAR SURGERY
Gre#ler et al.
4
o
NORI,W~JZ~D~TA -/~avrrY ~ ~ouA~ ,~'m'm¢ oF coNmoq,'~
Maywood, Ill., Rockville, Aid., and Livingston, N.J. We previously reported that biomaterials differentially induce macrophages to secrete growth factors that mediate reendothelialization. The present study evaluates the effect of an atherogenic diet on macrophage/biomaterial interactions. Female New Zealand white rabbits were fed an atherogenic diet. Peritoneal macrophages were harvested from these as well as rabbits fed a normal diet and culntred in Minimum Essential Medium with platelet-poor sernm. Dacron or polyglactin 910 were added to two of three conditions of both cell groups in passage 2. Conditioned media were collected weekly through week 15. Mitogenicity assays were performed with quiescent mouse embryonal (BALB/cST3) fibroblasts, atherosclerotic rabbit aortic smooth muscle cells, and murine capillary lung (LE-II) endothelial cells. Mitogenic activity was assayed by scintillation counting of tritiated thymidine incorporation into deoxyribonucleic acid (DNA). Results showed increased mitogenic activity released by macrophages from atherosclerotic rabbits, in the absence of prosthetic material, when assayed against every cell line. In normal diet macrophages, polyglactin 910 stimulated mitogen release for every cell line, and Dacron yielded minimal mitogen release. In lipid diet macrophages polyglactin 910 slightly increased mitogen release for all three cell lines, whereas Dacron resulted in stimulation of DNA synthesis in smooth muscle cells and BALB/c3T3 cells but less DNA synthesis in LE-II cells than in control, no graft material, media. Western blotting demonstrated immunoreactivity to basic fibroblast growth factor in media from normal diet macrophages but only in the presence of polyglactin 910 or Dacron. Radioimmunoassay for platelet-derived growth factor B chain was negative in all groups, and polymerase chain reaction techniques to amplify transforming growth factor-beta messenger ribonucleic was negative. These data demonstrate the effect of in vivo dietary manipulation on macrophage activation as well as the effect of an atherogenic diet in modulating macrophage/biomaterial interactions. Additionally, different biomaterials differentially induce macrophages to release factors that stimulate and inhibit growth. (J Vxsc StYRG 1991;14:10-23.)
Small diameter vascular grafts fail not because of the biomaterials themselves but because of the biologic reactions occurring at the blood/material and tissue/material interfaces. Prominent among the foreign body reactions is monocyte recruitment and activation and differentiation into the activated macrophage. Macrophages when appropriately activated are capable of synthesizing a variety of growth From Loyola University Medical Center, Maywood, Hines Veterans AdministrationHospital, Hines, The American Red Cross (Ms. Shaheen,Dr. Burgess),Rockville,and St. Barnabas Medical Center (Dr. Fdm), Livingston. Supported by a grant from the NationalInstitutesof Health, No. RO1 HL41272. Reprint requests: Howard P. Greisler, MD, Loyola University Medical Center, Department of Surgery, 2160 S. First Ave., Maywood,IL 60153. 24/1/274~8 10
regulatory factors that modulate migration and proliferation of endothelial cells, smooth muscle cells, and fibroblastsY s Our laboratory has previously shown that a variety of bioresorbable materials including polyglactin 910 (PG910) (Ethicon, Inc., Somerville, N.J.), when woven into vascular graft configurations and implanted into animal models, elicit extensive transinterstitial ingrowth of capillaries, endothelial cells, and myofibroblasts to a significantly greater extent than does similarly woven Dacron. 6-1° The tissue ingrowth begins shortly after macrophage infiltration and phagocytosis occurs and terminates when macrophage-mediated prosthetic resorption is complete. We have also shown in past studies that macrophages cultured in vitro in the presence of PG910 release into their media factor(s) capable of stimulating endothelial cell proliferation to
Volume 14 Number ] July 1991
a significantly greater extent than do macrophages cultured with Dacron. ~'5 However, when PG910 prostheses are implanted into rabbits previously made atherosclerotic by feeding of a 2% cholesterol, 6% peanut oil diet, the tissue ingrowth is markedly attenuated, resulting in thin walls containing lipidladen macrophages bull relatively few myofibroblasts and endothelial cells. The current study evaluates the effect of an atherogenic diet on macrophage production of growth factor activity and further examines the effect of such a diet on macrophage/biomaterial interactions. These interactions are likely central to the tissue reactions that occur after implantation of any blood-contacting biomaterial, and the modulation of these interactions by atherosclerosis must be examined. METHODS Dietary regimens
Adult female New Zealand white rabbits weighing 3 to 4 kg were housed in a facility accredited by the American Association for Accreditation of Laboratory Animal Care (AAALAC) and fed either standard Purina Rabbit Chow or Purina Rabbit Chow (Ralston Purina Co., St. Louis, Mo.) specially formulated to provide a 2% cholesterol, 6% peanut oil diet. Before institution of the diet, baseline serum cholesterol and triglyceride levels were measured. Macrophages were harvested after 2 months at which time serum cholesterol and triglyceride levels were again determined. Macrophage cuhxtre Peritoneal macrophages were harvested from eight rabbits of each group via peritoneal lavage with Hank's balanced salt solution containing calcium and magnesium and with 10 units/ml heparin by our previously published technique. 4 Collected cells were centrifuged at 500g for 10 minutes and washed three times with phosphate-buffered saline (PBS). Cells from the six rabbits per group were pooled and suspended in Minimum Essential medium (MEM) with 10% equine platelet-poor serum. Collected cells, 1 × 106 cells/ml, were seeded into three T-25 Falcon (Becton-Dickinson Labware, Lincoln Park, N.J.) tissue culture flasks containing 3 ml of MEM. Cell viability and[ cell counts were determined by trypan blue dye exclusion and a hemocytometer. After a 2-hour adherence separation, adherent cells were washed with MEM and fed with 5 ml of complete media. Ambient conditions were humidified with 5% CO2,. Macrophages in each group were
Atherogenic diet and macvophage biomaterial interaction 11
then serially passaged twice resulting in nine T-75 flasks of macrophages for both dietary groups. Finely shredded 17G910 or Dacron particles were prepared by scalpel technique from the vascular prostheses used for previously published in vivo studies, s These materials were then added each to one-third of the nine T-75 flasks of both of the two groups of macrophages at a concentration of 1.6 mg/cm 2. Macrophage cultlare purity was determined by three methods: morphologic characteristics under phase contrast microscopy, nonspecific esterase staining by the method of'Yam et al.,11 and identification of Fc (f?agment crystallizable immunoglobulin G) receptors on the cell membrane by the method of Bianco and Pytowski. 12 Flasks were routinely fed three times weekly and examined for macrophage/biomaterial interactions under phase contrast microscopy. For overall flow chart see Fig. 1. Collection of conditioned media The cells remained viable in all three conditions of both groups of macrophages at subconfluent levels, with relatively small :numbers of mitotic figures and minimal polyploidy observed. Macrophages harvested from the group fed an atherogenic diet were noticeably more lipid filled. After 5 weeks in culture, intracytoplasmic PG910 inclusions were seen within the macrophages of both dietary groups (Fig. 2). By contrast, the cells adhered to but did not phagocytose the Dacron. Beginning at the time of addition of the prosthetic materials, conditioned medium from each flask was collected at the time ofrefeeding three times weekly. These three media collections were pooled and frozen to - 7 0 ° C yielding three groups of conditioned media for each of the two groups of harvested macrophages for each week in culture. Media were later thawed for analysis of growth promoting activity against test quiescent cells. BALB/c3T3 bioassay BALB/c3T3 mouse embryonic fibroblasts, kindly provided by Patricia D'Amore, MD, were cultured in complete media containing Dulbecco's modified Eagle's medium (DMEM) (glucose 4.5 gm/1) supplemented with 10% Colorado calf serum. The BALB/c3T3 bioassay is a well-established quantitative bioassay responsive to many growth factors and is therefore a good initial screen. Cells were removed with trypsin from a T-75 culture flask at 75% confluence and were counted and assessed for viability by use of trypan blue dye exclusion. Cells were
Journal of VASCULAR SURGERY
12 Greisler ctal.
A
~
N
Z
W
R
A
~
I MACROPHAGE
MACROPHAGE
HARVEST
cno. I
HARVEST
;TION
B
]
I M~a cou~c~oNs
BALB/c3T3 Mo0,Eni. o:II B o~ss,v
/
I
LE-II I T.EL~L CELL
SMC ( , ~ E . o -
SCLEROTICRABBIT AORTICB )IOASSAY
/
I
(6)
I /
/
/
RADIOIMM UNOASSAY - PDGFBCHAIN
/ \\ \
\
WESTERNBLOTS - HBGF-1 - HBGF-2 - PDGFBCHAIN
Fig. 1. Flow chart of experimental design. then seeded into 96-well microtiter plates at a density of 7000 cells/well. These cells were fed complete media and were then incubated at 37 ° C in humidified 10% CO2 for 5 days without media change, which resulted in quiescence. On the sixth day after seeding, 50 m m 3 of PBS were added to the negative control wells, 50 m m 3 of Colorado calf serum were added to the positive control wells, and 50 m m 3 of the undiluted or 1:10 diluted conditioned media were added to the test wells, each in replicates of four. 3H thymidine (3H TdR), 1 txCi/well, was then added to each well and allowed to incubate for 48 hours. The wells were then processed for scintillation
counting of tritiated thymidine incorporation into newly synthesized deoxyribonucleic acid (DNA). Cells were washed vigorously with normal saline three times and then fixed twice with 250 m m 3 of 100% methanol at 4 ° C for 5 minutes followed by vigorous washing with distilled water three times. Wells were then incubated twice in 5% trichloroacetic acid (TCA) for 10 minutes, again vigorously washed, and the acid precipitable material was then solubilized with 150 mm 3 0.3 N N a O H at room temperature for 20 minutes. Each 150 m m 3 sample was then transferred to 10 ml of ready protein scintillation fluid and counted in a Beckman LS6800
Volume 14 Ntunber 1 July 1991
Atherogenic diet and macrophage biomaterial interaction 13
Fig. 2. A, Phase-contrast photomicrograph of peritoneal macrophages heLrvestedfrom normal New 2"ealand white rabbits after 6 weeks in culture with PG910. Intracytoplasmic PG910 inclusions can be seen. (Original magnification × 500.) B, Phase-contrast photomicrograph of peritoneal macrophages harvested from atherosclerotic New Zealand white rabbits cultured for 8 weeks in the presence of PG910. Intracytoplasmic PG910 inclusions can be seen. No gross morphologic differences are seen comparing the macrophages from the two groups of rabbits. (Original magnification × 500.)
(Beckman Instruments, Inc., Fullerton, Calif.) scintillation counter. Atherosclerotic rabbh: aortic smooth muscle
cell bioassay One adult New Zealand white rabbit fed an atherogenic diet as above for 3 months was anesthetized with intravenous 2% methohexital, and the thoracic and abdominal aorta were sterilely resected and washed vigorously with Hank's balanced salt solution to remove blood. The aorta was then opened longitudinally and the endothelium removed by
gentle passage of a :scalpel blade. At the time of harvest the serum cholesterol and triglyceride levels in this rabbit measured 2130 mg/dl and 110 mg/dl, respectively. The adventitia was removed by careful dissection, and the remaining m e d i u m was minced into 1 mm 3 fragments and suspended in i ml of complete media. Complete media consisted of D M E M (glucose 4.5 gm/1) in 10% fetal bovine serum (FBS) supplemented with nonessential amino acids, sodium pyruvate, penicillin/streptomycin, and gentamicin. The suspension was transferred into a T-25 culture flask and incubated for 4 hours at 37 ° C
14
Greisler et al.
in humidified 5 % CO2 to allow adherence, followed by the addition of 1 ml of complete media. Cells were refed three times weekly and after their firm adherence, 5 ml of complete media were used per T-25 flask. The aortic smooth muscle cells migrated off the explants after 1 week and after confluence were subcultured twice. Smooth muscle cell identification was documented by immunofluorescent staining by use of a smooth muscle specific anti-alpha actin monoclonal antibody (Sigma Chemical Company, St. Louis, Mo.). Cells were seeded into 96-well microtiter plates at a density of 5000 cells/well and grown in complete media at 37 ° C in humidified 5% CO2 until confluent (3 days). The medium was then aspirated and the cells washed and refed with 200 mm 3 of serum-free media for 48 hours. Serum-free media consisted of a 1:1 mixture of DMEM and Hams F-12 media supplemented with transferrin (5 ~xg/ml), penicillin (100 units/ml), streptomycin (100 txg/ml), insulin (1 ~xmol/L [6 /xg/ml]), and ascorbate (0.2 ixmol/L [35.2 ~xg/ml]).13After 48 hours in serum-free media, 50 mm ~ of PBS were added to the negative control wells, 50 mm 3 of FBS to the positive control wells, and 50 mm 3 of either undiluted or 1:10 diluted conditioned media to the test wells, each in replicates of four. 3H-TdR, 1 ~xCi/well,was then added to each well and allowed to incubate for 48 hours. The wells were then processed for TCA precipitable material for scintillation counting of 3H-TdR incorporation into newly synthesized DNA as described above under the BALB/c3T3 bioassay.
LE-II bioassay Murine capillary endothelial (LE-II) cells, kindly provided by Thomas Maciag, MD, from an established cell line were identified by anti-factor VIIIrelated antibody staining and were grown to confluence in T-25 culture flasks in medium 199 with 10% fetal calf serum as previously published? This particular cell line was selected because its use has been well established for an endothelial cell mitogenesis assay. Cells were resuspended in complete media, and aliquots of 500 mm 3 were added to 48-well microtiter dishes. After 48 hours, quiescence of preconfluent wells was achieved by a 48-hour exposure to medium 199 with 2% serum. Quiescence was documented in replicate wells by scintillation counting of 3H-TdR incorporation into DNA 3 hours after a tritium pulse of 0.5 ixCi/ml. Thawed conditioned media were then added to the quiescing media in 1:10 or 1:20 dilutions in
Journal of VASCULAR SURGERY
replicate wells. These conditioned media included collections during even numbered weeks from the atherosclerotic diet group of macrophages from weeks 2 to 12. These results were compared with those we previously published using the weeks 5 to 10 collections of media from the normal diet group of macrophages. After 16 hours 3H-TdR incorporation was measured in triplicate wells for each group at each week by use of the method described above for scintillation counting of TCA precipitable material. Growth factor identification methods Conditioned media collected from all six macrophage groups during the tenth week of culture were analyzed by Western blotting techniques. These conditioned media were subjected to polyacrylamide gel electrophoresis (15% w/v) in the presence of sodium dodecyl sulfate (SDS-PAGE) using the Mighty Small gel apparatus (Hoeffer Scientific Instruments, San Francisco, Calif.). After electrophoresis, gels were stained with Coumassie brilliant blue R-250, dried, and exposed to x-ray film by our published techniques. 14'15 Gels were analyzed by Western blotting to determine immunoreactivity to polypeptides of appropriate apparent molecular weights by use of antibodies against platelet-derived growth factor (PDGF)-B chain, type 1 heparinbinding growth factor endothelial cell growth factorECGF and type 2 heparin binding growth factor. In addition, all groups of collected media from each week were assayed for the presence of PDGF-B chain by means of a highly specific radioimmunoassay (Amersham International, Arlington Heights, ILL). This assay used an iodine 125-labeled goat antihuman PDGF antibody sensitive to the range of 1.56 to 200 fmol/ml. We have demonstrated in other studies the cross-immunoreactivity of this antibody with PDGF-B chain produced by rabbits (Greisler H.P., unpublished observations).
Statistics Tritiated thymidine incorporation (counts per minute [cpm]) into newly synthesized DNA was calculated as a percent of maximum response (FBS positive control) and expressed as mean + standard deviation (SD), and the groups were statistically analyzed by the Student's t test. RESULTS The 2-month feeding of the a'cherogenic diet resulted in a significant increase (p < 0.001) in serum cholesterol levels with cholesterol and triglyc-
Volume 14 Number 1 luly 1991
Atherogenic diet and macrophage biomaterial interaction 15
Fig. 3. A, Photomicrograph from thoracic aorta of an atherosclerotic New Zealand white rabbit showing greatly thickened intima containing abundant foam cells. The elastic laminae of the media can be seen in the lower portion; the media contains smooth muscle cells with vacuolated cytoplasm. (Hematoxylin-eosin stain; original magnification ×400.) B, Higher magniification of intima outlined within box in A. (Hematox-ylin-eosin stain; original magniification × 400.) eride levels measuring 2840 _+ 339 and 167.5 _+ 86.1, respectively compared to 42.3 + 23.9 and 67.3 + 29.5 before institution of the diet. Gross atherosclerotic plaques in the aortas were evident, primarily in the aortic arch and in periosteal areas throughout the aortas. Histologically (Fig. 3, A and B) these areas showed extensive intimal thickening with fbam cells and cholesterol clefts. The macrophages harvested from the atherosclerotic rabbits contained greater numbers of lipid vacuoles when viewed under phase-contrast microscopy. BALB/c3T3 bioassay The BALB/c3T3 cells at the time of addition of test media were quiescent, documented by the PBS negative controls, 203.55 _+ 3455 cpm, and capable of DNA synthesis when stimulated, documented by the FBS positive controls, 237870 _+ 29122 cpm. Macrophages cultured in the absence of graft materials were mildly activated by the polystyrene culture flasks. Fig. 4, A and B shows the data as percent of maximal activity in 3H-TdR incorporation, with maximal activity being that in the FBS positive controls. Bo~h the normal and the athero-
sclerotic rabbit macrophages' media at each week produced small increases in 3H-TdR incorporation ranging from 11% to 26% of maximal activity. These data can also be expressed as an increase in 3H-TdR incorporation of 161% to 406% above quiescent (PBS) levels. In the absence of graft materials, the atherosclerotic group's media collected in weeks i to 7 contained significantly more mitogenic activity at each week (p < 0.001 in weeks 3 to 6) than the macrophages from normal rabbits. In later weeks these differences were abolished, largely because of increasing mitogenic activity in late weeks in the media of macrophages from normal rabbits. PG910 stimulated normal rabbits' macrophages to release significantly more mitogen into the media in weeks 1 to 7. To eliminate the effect of macrophage activation induced by polystyrene, the data are normalized to the samples at each week from macrophages cultured in the absence of graft materials. Thus Fig. 4 shows this normalized data with the mean percent increase above quiescence produced by media in the presence of a graft material divided by the mean increase above quiescence produced by the media collected that same week but in the absence of graft material. The increase in macrophage release of
Journal of VASCULAR SURGERY
16 Greisler et aL
~
7O m
60
GES
C
NO~.4~J7 ~ ~TA - ~
~
BIOMA'~L.~tK
5O
4
d 40 ~o
2o~
o "All'IEROSCLEROTIC" MACROPHAGES
0 n
c)
60 8ORilL'NUZI~ DATA-- ~
~
BIOMAIi'~4~AL~fK
b_ 50 0 40 F-Z W 30 C) rY 20 t.LI Cl 10
5 ~ J
1 ¢.ON~Ot_
o
2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
CULTURE DURATION (WEEKS) QUIESCE]WCE (PIBS) - 8.~1; OF PO~TIVE CONTROL F'B5 CPM: 237870 *: 29122 PBS CPM: 20355 :k 34,55
CONTROL (NO GRAFT MATERIA0 m PG910 ~DACRON
Fig. 4. BALB/c3T3 fibroblast bioassay results. Studies with conditioned media from macrophages harvested from normal New Zealand white rabbits are in the top two panels and those studies with conditioned media from macrophages harvested from atherosclerotic New Zealand white rabbits in the lower two panels. The panels on the left show the results of each group of conditioned media in each successive week, expressed as mean counts per minute sample + mean counts per minute fetal bovine serum positive control × 100. The right sided panels show the normalized data in which each week's media from the groups with prosthetic material is normalized by dividing its mean counts per minute by that of the media from the macrophages grown in the absence of prosthetic material, thus normalizing to polystyrene control values. Values > 1 indicate additional stimulation of mitogen release by the presence of the biomaterial, whereas values < 1 indicate less mitogen release than is seen in the absence of that biomaterial. (mean -+ SD)
mitogens caused by the interaction with PG910 was statistically significant (p - 0.013) each of the first 4 weeks. When expressed as percent of FBS positive control, the week 4 collection resulted in 42% of maximal stimulation. Dacron also induced a smaller increase in mitogen release in weeks 1 to 8, again statistically significant in weeks 1 to 4 (p - 0.01). However, at later time points, weeks 10 to 15, the media from macrophages cultured with Dacron resulted in significantly less (p < 0.001) mitogenic activity than in the absence of Dacron. This suggests either a decrease in mitogen release or a production of an inhibitor to fibroblast proliferation, or both. This late decrease in activity was not present in the bioassay on smooth muscle cells, which suggests that the macrophages in these late weeks remained competent to release mitogenic activity.
The 3H-TdR incorporation caused by PG910/macrophage media was significantly greater than that caused by Dacron/macrophage media in weeks4 (p < 0.05) andweeks 12to 15 (p < 0.005). Much less macrophage activation by either material is seen in the macrophages harvested from atherosclerotic rabbits. This increase in mitogenic activity above the polystyrene controls was significant (p < 0.001) only in week 1. No significant differences were found between the two materials on macrophages from atherosclerotic rabbits. The media from week 8 of the PG910 group was lost, and consequently this single data point is omitted in Fig. 4. In summary, these data show stimulation of fibroblast DNA synthesis by media from both groups ofmacrophages, but in the absence of graft materials (baseline conditions) significantly more mitogen
Volume 14 Number 1 July 1991
Atherogenic diet and macrophage biomaterial interaction 17
14
NORMALI~I~CROPH/~GES
A NORMALMACROPHAGE~
C
9 8
~'12
7 L,~IO
6 5
4 ~ ~_
3 IM 2
->o
1.1_10 0 F- 8 7 W 6
4 5 2
Q_ 2 0
0 2
4
6
8
10
12
14
0
2
4
6
8
10
12
14
CULTURE DURATION (WEEKS) QUIESCENCE(PB$) - 0.6~ OF POSfIWE CONTROL F'BS CPM: 2 6 9 4 0 i 4.186 20
PB$ CPM: 165 i
[--'7 CONTROL (NO GRAFT MATERIAL) m PG910 DACRON
Fig. 5. Atherosclerotic rabbit aortic smooth muscle cell bioassay results. Studies that used conditioned media from macrophages harvested from normal New Zealand white rabbits are in the top two panels, and those studies that used conditioned media from macrophages harvested from atherosclerotic New Zealand white rabbits in the lower two panels. The panels on the left show the results of each group of conditioned media in each successive week, expressed as mean counts per minute sample + mean counts per minute fetal bovine serum positive control × 100. The right sided panels show the normalized data in which each week's media from the groups with prosthetic material is normalized by dividing its mean counts per minute by that of the media from the macrophages grown in the absence of prosthetic material, thus normalizing to polystyrene control values. Values > 1 indicate additional stimulation of mitogen release by the presence of the biomaterial while values < 1 indicate less mitogen release than is seen in the absence of that biomaterial. (mean +- SD) from those harvested from atherosclerotic rabbits. The a d d i f o n of either graft material to the mcdia increased mitogen release in early weeks but significantly more so by PG910. In late weeks Dacron appeared[ to inhibit release of mitogenic activity by macrophages harvested from normal rabbits. Atherosclerotic rabbit s m o o t h muscle
cell bioassay Quiescence of the smooth muscle cells was documented by the PBS negative controls, 165 --_ 20 cpm, and the capacity of quiescent cells to be stimulated to synthesize D N A documented by the FBS positive controls 26940 ___ 4186 cpm. Fig. 5 shows the mild effect of polystyrene on both groups of rnacrophages in this bioassay. There is a small but persistent mitogenic activity in the
media of macrophages harvested from both normal and atherosclerotic rabbits and cultured in the absence of graft materials. These increases in ~H-TdR incorporation are smaller (1% to 4% of maximal activity) than those !Found with the more sensitive BALB/c3T3 cell line, but still represent 164% to 601% increases above quiescent levels. As was true also in the BALB/c3T3 bioassay, the atherosclerotic group's media possessed more mitogenic activity, but these differences were significant only in week 1 (p = 0.009). The week 2 difference approached significance (p = 0.059). PG910 stimulated normal rabbits' macrophages to release significantly more mitogen into the media (Fig. 5) in weeks 1 to 5, reaching significance in weeks 3 to 5 (p -< 0.,005). In week 4 this increase in smooth muscle cell mitogen release induced by PG910 was 5.68 times that of the polystyrene
18
jrournal of VASCULAR SURGERY
Gre#ler et al.
4
o
NORI,W~JZ~D~TA -/~avrrY ~ ~ouA~ ,~'m'm¢ oF coNmoq,'~
