Culture of retinal capillary cells using selective growth media

MICROVASCULAR

RESEARCH

26, 74-80 (1983)

Culture JONATHAN

of Retinal Capillary Cells Using Selective Growth Media D.

GITLIN*

AND PATRICIA

A. D’AMoRE’+

*Department of Medicine, Division of Newborn Medicine and tDepartment of Surgery, Children’s Hospital Medical Center, and *Department of Pediatrics, and tDepartments of Surgery and Pathology, Harvard Medical School, Boston, Massachusetts 02115 Received

October

4,

1982

Capillary fragments are isolated from the microvasculature of the bovine retina using limited collagenase digestion and sieving. The endothelial cells obtained from these capillary fragments are cultured in media containing platelet-poor plasma supplemented with retinal extract, which contains an endothelial cell mitogen. These cells form typical monolayers in vitro, contain Factor VIII antigen, and can be passaged serially. The pericytes obtained from these capillary fragments are cultured in media containing calf serum. These cells can also be passaged serially, are not contact-inhibited, and do not stain for Factor VIII antigen. Cultures of capillary cells such as these will allow the study of the individual cells from specific microvascular beds involved in various normal and pathological processes.

INTRODUCTION Since the first cultures of endothelial cells were established from human umbilical vein in the early 1970s (1,2) researchers interested in the endothelium have questioned whether endothelial cells from various levels of the vascular tree differ structurally or functionally. Some differences have recently been reported among endothelial cells from different sites in the vascular system. These include density of insulin receptors (3), types of intercellular junctions (4), production of angiotensin-converting enzyme (5), prostacyclin synthesis (6), growth requirements (7,8), and type and quantity of collagen produced (9). The comparison of large vessel and microvessel endothelium has been impeded by the difficulties encountered in culturing capillary endothelium. Most attempts at culturing endothelial cells from capillaries have met with limited success. The harsh enzyme treatment and severe mechanical manipulations used to isolate the capillaries have resulted in both poor viability and low yield. Further, contamination of the cultures by pericytes, the other cellular component of the capillary, has interfered with efforts to obtain pure capillary endothelial cell cultures. In 1978 Folkman and his co-workers (8), with modifications of the procedures of Del Vecchio et al. (10) reported that large numbers of viable capillary fragments could be isolated and the endothelial cells cloned and carried in long-term culture on gelatin-coated substrates in the presence of media conditioned by mouse sarcoma 180 cells. ’ To whom correspondence should be mailed. 74 0026.2862/83 $3.00 Copyright Q 1983 by Academic Press, Inc. All rights of reproduction in any form reserved. Printed in U.S.A.

CULTURE

OF RETINAL

CAPILLARY

CELLS

75

Here we report a method that utilizes selective growth conditions for obtaining and maintaining cultures of capillary endothelial cells and pericytes. This study reports on cultures derived from bovine retinal tissue; we have observed similar results for the culture of capillary cells from bovine brain and adrenal gland. MATERIALS Culture of the Microvascular

AND METHODS

Cells

Adult bovine eyes were transported from the slaughterhouse on ice. Excess connective tissue was removed and the eye was immersed in Betadine for 10 min followed by thorough rinsing with phosphate-buffered saline (PBS). The eye was then bisected approximately 5 mm posterior to the limbus and the retina was removed. The retina was minced thoroughly using scalpel blades and washed once by suspension in PBS followed by centrifugation. The minced tissue preparation was then suspended in 5 ml of 0.375% collagenase (Type II, Millipore) with 0.25% bovine serum albumin in PBS with gentle aggitation for 20-30 min at 37°C. The digested preparation was then passed over a nylon sieve (110 pm mesh) and rinsed with 10 ml Dulbecco’s Modified Eagle’s medium (DMEM) with 10% calf serum. The fragments in the filtrate were then pelleted and washed twice with the same media. The vessel fragments that are collected by this process contain two to five cells and are referred to as capillary fragments. Although these microvessel fragments may also be derived from arterioles or venules, it is impossible to distinguish between these and true capillaries. With this caveat, these vessel fragments will be referred to as capillary fragments. The washed capillary fragments were plated in DMEM with 10% calf serum into gelatinized 35-mm dishes (Falcon) and allowed to attach for 1 to 3 hr. At the end of the attachment period, the dishes that were intended to select for endothelial cell growth were rinsed gently with DMEM with 5% human platelet-poor plasma (PPP) [prepared as described by Pledger et al. (1 l)‘] and refed with the same media containing 20 @ml retinal extract (12). Cultures that were intended to select for pericyte growth were rinsed gently with DMEM with 10% calf serum and refed with the same media. Between 2 and 8 days after the fragments were plated, the cultures were observed and colonies of capillary endothelial cells and pericytes selected by marking the bottom of the culture dish. When a single colony contained between 50 and 200 cells, the remainder of the dish was cleared with a rubber policeman and that colony was either allowed to proliferate to Eover a majority of the surface area or was trypsinized and transferred to another 35-mm dish (which was gelatinized for endothelial cell cultures but not for pericyte cultures). Factor

VZZZStaining

For Factor VIII staining, cells were grown to confluence on 12-mm gelatincoated coverslips. At confluence, the cells were rinsed in PBS, fixed with 100% methanol at -20°C for 5 min, washed in PBS, and then incubated for 45 min at 2 Human plasma, utilized as starting material, was obtained from New England Regional Red Cross Blood Services. In all instances, bovine and human platelet-poor plasma, similarly prepared, yielded similar results. Human PPP was used because of greater availability of starting material.

76

GITLIN

AND

D’AMORE

room temperature with rabbit antiserum to bovine Factor VIII (1:150) (Kindly provided by Dr. E. Kirby, Temple University). The cells were then washed again in PBS and incubated for 45 min at room temperature with fluorescein-conjugated anti-rabbit IgG (Miles; 1:50) and washed again in PBS. The coverslips were then mounted on slides, visualized, and photographed utilizing a Leitz fluorescent microscope. Proliferation

Studies

The endothelial cells and pericytes (passage 2 through 6) were plated into 24well dishes (Falcon; gelatin-coated for the endothelial cells) at 10,000 cells per well in growth media, DMEM with 5% PPP, and DMEM with 10% calf serum, respectively. The cells were allowed to attach overnight before being rinsed with Hanks’ balanced salt solution (HBSS) and changed to fresh DMEM with varying concentrations of PPP or serum. For each PPP or serum concentration, retinal extract (20 PYml; 120 pg of Lowry protein) was added to half of the wells and an equivalent volume of HBSS was added to the controls. After 4 days in culture, the cells were trypsinized and counted using a Coulter counter. The cell number at each data point is expressed as a function of the number of cells plated. RESULTS Limited collagenase digestion of bovine retinal tissue followed by sieving results in the isolation of a large number of capillary fragments with good viability (greater than 80% by trypan blue exclusion). These fragments which contain two to five cells attach and spread on gelatin-coated substrate. Although the capillary cells will also attach to uncoated tissue-culture plastic, they do not spread as well on this surface. The endothelial cells cultured using DMEM with 5% PPP supplemented with retinal extract have a doubling time of 24 hr. In contrast to large vessel endothelium, the retinal capillary endothelial cells grow as contiguous “islands” of cells (Fig. la,b). At confluence the cells form a contact-inhibited monolayer (Fig. lc) and stain with antisera to Factor VIII antigen, a definitive endothelial cell marker (Fig. Id). The pericytes obtained by these techniques and cultured in DMEM with 10% calf serum lack Factor VIII antigen. There is no unique marker that can be used to identify the pericyte. In this work, pericytes are identified by two indirect criteria: (i) In primary cultures pericytes are observed in the vicinity of capillary fragments from which they may have migrated (see Fig. la, lower right); (ii) both in primary cultures and subpassages the pericytes are characterized by highly irregular peripheries, phase-dense filaments, and non-contact-inhibited, overlapping pattern of growth at confluence (Fig. le,f). The effects of PPP, calf serum, and retinal extract on the growth of capillary endothelial cells and pericytes was determined in a series of proliferation studies. Concentrations of PPP and serum from 1 to 10% were tested for their ability to support cell growth over a 4 day period (Fig. 2). Medium containing PPP alone at concentrations over this range allows a two- to threefold increase in the number of endothelial cells while equivalent concentrations of serum stimulates a fourto eightfold increase in cell number. Addition of 20 PI/ml (10 pg Lory protein) of retinal extract, which has been shown to contain an endothelial cell

FIG. 1. The culture of retinal capillary cells. (a) Retinal capillary fragment after two days in culture. Note the pericytes in the lower right hand corner of the photograph. (b) Retinal capillary endothelium after 5 days in culture. (c) Confluent culture of retinal capillary endothelium. (d) Indirect immunofluorescent staining of Factor VIII antigen in a subconfluent culture of retinal capillary endothelium. (e) Subconfluent culture of retinal capillary pericytes. (f) Confluent culture of retinal capillary pericytes.

mitogen (1 l), results in increased proliferation in the presence of both PPP and serum (Fig. 2). Pericytes, on the other hand, are unable to proliferate in the presence of plasma at any concentration and the addition of retinal extract does not significantly alter this (Fig. 2). After 4 days of culture in the presence of PPP, pericyte cell

78

0

2

I

4

I

6

I

8

PLASMA OR SERUM CCJVCENTRATx3N. X

2. The effect of platelet-poor plasma, serum, and retinal extract on the growth of retinal capillary endothelial cells and pericytes in culture. Closed figures indicate culture in media containing plasma (0) or serum (W) and the open figures indicate culture in media containing plasma plus retinal extract (20 PYml; 120 pg protein) (0) or serum plus retinal extract (0). Each data point represents cell number expressed as a function of the number of cells plated. FIG.

number is below control, indicating cell death; the cells which lifted in the presence of PPP were unable to replate and did not exclude trypan blue. DISCUSSION The growth of endothelial cells isolated from large vessels has been shown to be supported equally well by growth media plus serum or PPP (13-15). This is not true for other mammalian cells in tissue culture. Smooth muscle cells, for example, have been shown to require the serum component, platelet-derived growth factor for growth in tissue culture (16). Knowledge of this led us to use PPP for the culture of capillary endothelium. Pericyte contamination had always been a major obstacle to obtaining pure cultures of capillary endothelial cells. Since it was probable that pericytes, like most other nontransformed mammalian cells would require serum, we reasoned that the use of PPP would allow endothelial cell proliferation while depressing pericyte growth. Thus, we have utilized PPP and retinal extract to culture two to four cell fragments obtained from the retina. Both the endothelial cells and pericytes can

CULTURE

OF RETINAL

CAPILLARY

CELLS

79

be grown to confluence and passaged serially. The endothelial cells have been maintained through passage 18 at 1:3 split ratio. The pericytes have been maintained through passage 12 at a 1:3 split ratio. At confluence, the endothelial cells are contact-inhibited whereas pericytes are not and the endothelial cells stain positively for Factor VIII antigen and the pericytes do not. Platelet-poor plasma (PPP), although it does not support endothelial cell growth to the same extent as an equivalent concentration of serum, can be utilized to select against the growth of pericytes. Growth of the primary capillary fragments in PPP does not result in immediate disappearance of all of the pericyte population. In fact, cells resembling pericytes do survive in the presence of PPP. However, these cells proliferate slowly, if at all, thereby providing a growth advantage for the endothelial cells. Finally, crude retinal extract which contains a growth stimulant for endothelial cells, but not for pericytes, is utilized as a supplement. Addition of the extract to cultures of retinal capillary cells in the presence of 5% PPP stimulates the proliferation of the endothelium to approximately the same extent as 5% serum. These culture conditions are not sufficient to support the proliferation of pericytes. Bowman and his co-workers (17) have recently reported similar success using platelet-poor plasma in the culture of bovine retinal microvessels. The major differences between the two techniques include their use of fibronectin as a substrate (versus our use of gelatin) and our use of crude retinal extract as a supplement to the culture media. The data shown here indicate the application of these culture techniques to the study of the retinal microvasculature. These retinal capillary endothelial cells are currently being used to investigate the cellular aspects of blood vessel involvement in ocular pathologies. Furthermore, the availability of capillary endothelial cell cultures from other systems will allow the investigation of a number of pathologic processes involving the microvasculature, and permit researchers to begin to study the normal biology of endothelial structure and function throughout the vascular tree. ACKNOWLEDGMENTS The authors thank Dr. Judah Folkman for his continued encouragement and support, Dr. Bruce Zetter for his critical review of the manuscript, and Mary Jo Canavan for preparing this manuscript. The work reported in this study was supported by a grant to Harvard University from the Monsanto Company, St. Louis, Missouri, and partial support from the Department of Ophthalmology, Johns Hopkins University School of Medicine. REFERENCES E. A., NACHMAN, R. L., BECKER,C. G., AND MINICK, C. R. (1973). Culture of human endothelial cells derived from umbilical veins. J. C/in. Znvesr. 52, 2745-2756. 2. GIMBRONE, M. A., COTRAN, R. S., AND FOLKMAN, J. (1974). Human vascular endothelial cells in culture: Growth and DNA synthesis. J. Cell Biol. 60, 673-684. 3. BAR, R. S., PEACOCK, M. L., SPANHEIMER, R. G., VEENSTRA, R., AND HOAK,J. C. (1980). Differential binding of insulin to human arterial and venous endothelial cells in primary culture. Diabetes 29. 991-995. 1. JAFFE,

80

GITLIN

AND

D’AMORE

4. SIMIONESCU, M., SIMIONESCU, N., AND PALADE, G. E. (1976). Segmental junctions in vascular endothelium. Arteries and veins. J. Cell Bjo/. 68, 5. JOHNSON, A. R. (1980). Human pulmonary endothelial cells in culture: arteries and cells from veins. J. C/in. Invest. 65, 841-849. 6. COU~HI.IN, S. R., MOSKOWITZ, M. A., ZETTER, B. R., ANTONIADES, H. N., Platelet-dependent stimulation of prostacyclin synthesis by platelet-derived

differentiations of cell 705-723. Activities of cells from AND LEVINE, L. (1980). growth factor. Nature

(London)288, 600-602. 7. DAVISON, P. M., AND KARASEK, M. A. (1981). Human dermal microvascular vitro:

Effect of cyclic AMP on cellular

morphology

and proliferation

endothelial cells in rate. J. Cell. Physio[. 106,

253-258. 8. FOLKMAN, J., HAUDENSCHILD, C. C., AND ZF.TTBR, endothelial cells. Proc. Nat. Acad. Sri. USA 76, 9. SAGE, H., PRITZL, P., AND BORNSTEIN, P. (1981). culture: Comparison of aortic, venous, capillary,

B. R. (1979). Long-term culture of capillary 5217-5221. Secretory phenotypes of endothelial cells in and cornea1 endothelium. Arteriosclerosis 1,

427-442. 10. DEL VECCHIO, P., RYAN, U. S., AND RYAN, J. W. (1977). Isolation of capillary segments from rat adrenal gland. J. Cell Biol. 75, 73a. 11. PLEIX~EH, W. J., STILES, C. D., ANTONAIDES, H. N., AND SHER, C. D. (1977). Induction of DNA synthesis in Balbk 3T3 cells by serum components: Re-evaluation of the commitment process. Proc. Nat. Acad. Sci. USA 74, 4481-448s. 12. D’AMORE, P., GLASER, B. M., BRUNSON, S. K., AND FENSELAU, H. J. (1981). Angiogenic activity from bovine retina: Partial purification and characterization. Proc. Nai. Acud. Sci. USA 78,

3068-3072. 13. WALL, R. T., HARKER, L. A., QUADRACCI, L. J., AND STRIKER, G. E. (1978). Factors influencing endothelial cell proliferation in vitro. J. Cell Physiol. 96, 203-215. 14. THORGEIRSSON, G., AND LAZZARINI-ROBERTSON. A. (1978). Platelet factors and the human vascular wall. Atherosclerosis 30, 67-78. 15. DICKINSON, E. S., AND SLAKEY, L. L. (1982). Plasma-derived serum as a selective agent to obtain endothelial cultures from swine aorta. In Vitro 18, 63-70. 16. Ross, R., NIST, C., KARIYA, B., RIVEST, M. J., RAINES, E., AND CALLIS, J. (1978). Physiological quiescence in plasma-derived serum: Influence of platelet-derived growth factor on cell growth in culture. J. Cell Physiol. 97, 497-508. 17. BOWMAN, P. D., BETZ, A. L., AND GOLDSTEIN, G. W. (1982). Primary culture of microvascular endothelial cells from bovine retina. In Vitro 18, 626-632.