Venous Ulcer Fibroblasts Respond to Basic Fibroblast Growth Factor at the Cell Cycle Protein Level

Venous Ulcer Fibroblasts Respond to Basic Fibroblast Growth Factor at the Cell Cycle Protein Level Craig Seidman, MD,1,2 Joseph D. Raffetto, MD,2,3,4 Kristen C. Overman, BS,2,3 and James O. Menzoian, MD,2,3,5 Eugene, Oregon, Boston and West Roxbury, Massachusetts, and Farmington, Connecticut

Fibroblasts cultured from venous ulcers demonstrate phenotypic characteristics of cellular senescence including slow growth, altered morphology, upregulation of fibronectin, and increased senescence-associated b-galactosidase activity. In senescent cells, arrest of cell replication is related to overexpression of p21 and underexpression of phosphorylated tumor-suppressor protein retinoblastoma (ppRb). The regulatory mechanisms for cell proliferation in venous ulcer fibroblasts are unknown. In this study, venous ulcer fibroblasts are examined for cell cycle protein expression and modulation by basic fibroblast growth factor (bFGF). Fibroblasts were isolated from the venous ulcer of the distal lower extremity (fb-D) of patients with chronic venous insufficiency. A control biopsy was obtained from the proximal ipsilateral thigh (fb-P). Paired cultures were plated at 100,000 cells/plate and the cells synchronized. After 24 hr, one culture set was treated with bFGF (20 ng/mL) and the other was kept in culture medium only (untreated). All cultures, treated and untreated, were lysed following 24 hr of incubation, and the lysate was used to perform immunoblot analysis for p21, ppRb, and cyclin D1. Immunoblot samples were standardized to protein content. In all patients analyzed (n = 4), at basal levels (untreated) fb-D demonstrated significant overexpression of p21 versus fb-P (p = 0.016). Treatment with bFGF resulted in significant downregulation of p21 levels for fb-D (p = 0.008) and fb-P (p = 0.037) compared to untreated fibroblasts. ppRb was underexpressed in fb-D versus fb-P (p = 0.069). Treatment with bFGF increased ppRb significantly in fb-D (p = 0.030) and in fb-P (p = 0.027) compared to untreated fibroblasts. No differences were observed in cyclin D1 with respect to basal levels in fb-P versus fb-D or in treated versus untreated groups. Venous ulcer fibroblasts show phenotypic similarity to senescent cells, with overexpression of p21 as well as down regulation of phosphorylated pRb. The aberrations seen in the cell cycle proteins in fb-D are similar to those seen in senescent cells; however, bFGF can modulate important cell cycle regulatory proteins, promoting a proliferative environment in fb-D that is not possible in a senescent cell. The role of bFGF may be useful in the clinical treatment of venous ulcer pathology. 1

Eugene Surgical Associates Peacehealth, Eugene, OR, USA.

2

Department of Vascular Surgery, Boston Medical Center, Boston, MA, USA. 3

Boston University School of Medicine, Boston, MA, USA. 4 Department of Vascular Surgery, Boston Veterans Administration Healthcare System, West Roxbury, MA, USA. 5

University of Connecticut, Farmington, Connecticut, USA. Correspondence to: Joseph D. Raffetto, MD, Section of Vascular Surgery, Department of Surgery, Boston Veterans Administration Healthcare System, 1400 VFW Parkway, West Roxbury, MA 02132, USA, E-mail: [email protected] Ann Vasc Surg 2006; 20: 376-380 DOI: 10.1007/s10016-006-9036-9 Ó Annals of Vascular Surgery Inc. Published online: April 12, 2006 376

INTRODUCTION Chronic venous ulceration represents a formidable problem in terms of patient morbidity and ongoing financial burden to society. With a prevalence of 0.2%, the cost of treatment is estimated at $1 billion per year.1,2 In the United States, 54% of venous ulcers require treatment for more than 1 year and have a high recurrence rate.3 Treated conservatively with dressings and compression, ulcers show decreased ability to heal for reasons that are poorly understood. Previous work has demonstrated that fibroblasts cultured from venous ulcer of the distal lower

Vol. 20, No. 3, 2006

extremity (fb-D) of patients with venous disease display cellular characteristics of senescence compared to control fibroblasts cultured from the skin of the ipsilateral proximal thigh (fb-P).4,5 Like senescent cells, ulcer fibroblasts show decreased proliferation, altered morphology, increased levels of fibronectin, and increased senescence-associated b-galactosidase activity (a specific senescence marker).4,6-8 Similarly, senescent characteristics have also been identified in fibroblasts cultured from the distal lower extremity in patients with chronic venous insufficiency (CVI) lacking ulcers.9 Previous studies have postulated that fb-D have exhausted their replicative potential and that a significant proportion of these cells have reached senescence, contributing to poor healing in chronic venous ulcer wounds.4 Prior studies have determined that the irreversible block to cellular proliferation in senescent cells is dependent on important cell cycle regulatory proteins.10-12 In senescent human diploid fibroblasts, the cyclin-dependent kinase inhibitory protein p21 is upregulated, phosphorylated retinoblastoma protein (ppRb) is downregulated, and cyclin D1, although present in variable amounts, is likely inactive.10 These cell cycle protein abnormalities keep the cells arrested in G1 unable to pass the restriction point, termed the R point, to progress to the S phase and synthesize DNA for cell proliferation.10,13,14 In order for the cell to progress to the DNA synthesis phase of the cell cycle (S phase), p21 must become downregulated and the retinoblastoma protein needs to become phosphorylated to release and allow activation of transcription factors.10 Cellular senescence and increased replicative age are thought to be responsible for the diminished growth of fb-D.5 In the present study, the regulation of cell proliferation was studied by examining the expression of the cell cycle proteins p21, ppRb, and cyclin D1 in fb-D compared to fb-P. In addition, the response and modulation of venous ulcer fibroblast cell cycle proteins to basic fibroblast growth factor (bFGF) was investigated.

METHODS Patients Patients with venous insufficiency and chronic venous ulcers were interviewed and consented for outpatient biopsy. Diagnosis was based on history and clinical exam and verified with duplex ultrasonography demonstrating venous reflux. All patients had active venous ulcers and were Clinical,

Venous ulcer fibroblast response to bFGF 377

Etiological, Anatomical, and Pathophysiological (CEAP) clinical class 6.15 Patients with diabetes, arterial insufficiency, positive human immunodeficiency virus (HIV) serology, or active infection were excluded from the study. This study was approved by the institutional review board at Boston Medical Center. Materials DulbeccoÕs modified Eagle medium (DMEM) and trypsin were purchased from GIBCO BRL Life Technologies (Gaithersburg, MD). Bovine calf serum was obtained from Hyclone Laboratories (Logan, UT). Polyvinylidene difluoride (PVDF) nylon membrane was purchased from Life Science Products (Boston, MA, USA). An enhanced chemiluminescence (ECL) Western blot detection system kit was purchased from Amersham Life Science (Arlington Heights, IL). Monoclonal antibody to p21 (anti-cip 1) was purchased from BD Transduction Laboratories (San Diego, CA). Polyclonal antibody to pRB and ppRb and monoclonal antibody to cyclin D1 were purchased from Calbiochem (San Diego, CA). Recombinant human bFGF was obtained from R&D Systems (Minneapolis, MN). Fibroblast Isolation and Cell Culture From each patient enrolled, a 4 mm punch biopsy was obtained from the normal dermal tissue of the proximal ipsilateral thigh (fb-P). A second biopsy was obtained from the margin of an active venous ulcer in the distal leg (fb-D). Fibroblasts were cultured as previously described.4 Biopsy specimens were treated in a providone-iodine bath, followed by an ethanol bath and washed in phosphate-buffered saline (PBS). The specimens were then incubated overnight in 1 lg/mL trypsin solution to separate the dermal tissue from the epidermis and adipose tissue. Using a sterile technique, the dermal tissue was cut into 1 mm segments and cultured in DMEM with 10% calf serum (DMEM 10% CS). During explantation, penicillin, methicillin, fungizone, and streptomycin were added to the medium. Following establishment of fibroblasts, cultures were expanded as needed for experimentation. Early-passage (1-3) fibroblasts were used in all experiments. Cell Cycle Proteins Parallel cultures of second-passage fb-P and fb-D cells from all patients were established in duplicate on day 0 by plating 100,000 cells in DMEM 10% CS

378 Seidman et al.

to allow for cell attachment. On day 1, the fibroblasts were placed in medium containing 0.1% CS for 24 hr to allow for cell cycle synchronization. On day 2, half of the plates were fed with 20 ng/mL bFGF in DMEM 10% CS and the remaining half were fed with DMEM 10% CS alone. This established four treatment groups per patient: (1) fb-P, (2) fb-P with bFGF, (3) fb-D, and (4) fb-D with bFGF. On day 3, all groups were lysed for immunoblot analysis using specific antibodies against p21, ppRb, and cyclin D1. For each patient, all experiments were performed to evaluate for basal and bFGF-stimulated cell cycle proteins and in duplicate for accuracy. Immunoblot Analysis All cells were harvested in radioimmunoprecipitation assay (RIPA) lysis buffer [0.25 M Tris/HCl, 0.75 N NaCl, 2.5% sodium dodecyl sulfate (SDS), 0.1% Triton], and a Bio-Rad (Richmond, CA) assay was used to determine protein concentration.16 For each gel column, 20 lg of cellular protein was subjected to polyacrylamide gel electrophoresis and transferred to a Polyscreen PVDF Transfer Membrane. Following transfer, the membrane was blocked with 5% dry milk in PBS with 0.1% Tween (Life Science Products, Boston, MA). The membrane was exposed to specific antibody. Antibodies were diluted in 0.1% PBS Tween with 5% dry milk. Membranes were incubated for 2 hr at 37°C. Membranes were then washed with 0.1% PBS Tween and incubated with the appropriate secondary antibody for 1 hr at room temperature. Secondary antibodies were diluted 1:2,000 with 0.1% Tween. Following incubation, the membranes were washed with 0.1% Tween. Signals were then visualized using the ECL system according to the manufacturerÕs instructions. Membranes were exposed to Kodak (Rochester, NY) X-OMAT film. Immunoblot analysis for p21 used a 1:200 dilution of the primary antibody and a 1:2,000 dilution of the secondary antibody. A 12% polyacrylamide gel was loaded with 20 lg protein in each well, and bands were sought at 21 kDa. Immunoblot analysis for pRb and ppRb used a 1:500 dilution for the primary antibody and a 1:5,000 dilution for the secondary antibody. A 5% polyacrylamide gel was loaded with 20 lg protein in each well, and bands were sought at 105-112 kDa. Immunoblot analysis for cyclin D1 used a 1:200 dilution for the primary antibody and a 1:2,000 dilution for the secondary antibody. A 12% poly-

Annals of Vascular Surgery

Patient 1 P

P+

D

Patient 2 D+

P

P+

D

D+

p21

Fig. 1. Representative immunoblot of p21 from cultured fibroblasts of two venous ulcer patients. P, proximal control fibroblasts, fb-P, D, distal venous ulcer fibroblasts, fb-D; +, treatment with 20 ng/mL bFGF. All samples were loaded at 20 lg of protein.

acrylamide gel was loaded with 20 lg protein in each well, and bands were sought at 21 kDa. Analysis Photoshop 5.5 Image analysis software (Adobe, Mountain View, CA) was used to calculate optical density (OD), a unitless integer, of the bands present on immunoblot autoradiographs. Image analysis software was licensed to Boston University. OD values are given as the mean ± standard error of the mean (SEM). Statistical analysis utilized StudentÕs t-test to compare fb-D and fb-P in untreated and treated groups.

RESULTS Patients A total of four patients (three male and one female) with CVI and active venous ulceration were enrolled in this study. No patient had a documented history of lower extremity trauma or venous thrombosis. The average age of the patients was 50.6 years (range 36-62). p21 In all four patients analyzed, basal levels of p21 were significantly overexpressed in fb-D (OD = 108 ± 10) compared to fb-P (OD = 85 ± 10), p = 0.016. In all four patients, bFGF caused a significant downregulation of p21 in both fb-D (OD = 56 ± 7, p = 0.008) and in fb-P (OD = 38 ± 10, p = 0.037). Figure 1 is a representative immunoblot of p21 in two patients. The results were consistent in duplicate analyses (this was also true for ppRb and cyclin D1). ppRb In all four patients analyzed, basal levels of ppRb were underexpressed in fb-D (OD = 4 ± 4) compared to fb-P (OD = 38 ± 19), p = 0.069. In all four patients, bFGF led to a significant upregulation of

Vol. 20, No. 3, 2006

Venous ulcer fibroblast response to bFGF 379

Patient 1 P

P+

Patient 2 D

D+

P

P+

Patient 1 D

D+

ppRb

P

P+

Patient 2 D

D+

P

P+ D

D+

Cyclin D1

pRb

Fig. 2. Representative immunoblot of pRb/ppRb from cultured fibroblasts of two venous ulcer patients. P, proximal control fibroblasts, fb-P; D, distal venous ulcer fibroblasts, fb-D; +, treated with 20 ng/mL bFGF. All samples were loaded at 20 lg of protein.

Fig. 3. Representative immunoblot of cyclin D1 from cultured fibroblasts of two venous ulcer patients. P, proximal control fibroblasts, fb-P; D, distal venous ulcer fibroblasts, fb-D; +, treated with 20 ng/mL bFGF. All samples were loaded at 20 lg of protein.

ppRb in both fb-D (OD = 47 ± 17, p = 0.030) and in fb-P (OD = 68 ± 20, p = 0.027). Figure 2 is a representative immunoblot of pRb and ppRb in two patients.

are responsive to bFGF with increased proliferation, indicating a reversal of the senescent phenotypic characteristic.4,9 In the current work, important cell cycle proteins were examined in bFGF-treated fibroblast cultures, to determine if bFGF induces favorable proliferative changes in cell cycle protein expression. The current data demonstrated that bFGF caused upregulation of ppRB concurrent with an equally dramatic downregulation of p21. We examined this response at 24 hr after bFGF treatment to establish that the cell cycle protein changes precede and are therefore responsible for the accelerated growth brought about by bFGF treatment. In addition a prior study noted that the qualitative response of fb-D cells was as vigorous as that of fb-P cells in the presence of bFGF treatment.19 This observation implies that it is the slowgrowing fb-D culture as a whole which responds to bFGF as opposed to a small, more replicatively competent subpopulation stimulated by bFGF. An important finding from this study is why venous ulcer fibroblast cell cycle protein alterations, which are similar to senescent cells, are reversible in the presence of bFGF. It has been proposed that senescence and aging are two wellcoordinated but separate processes.10,13 Aging, in the form of a replicometer, is a means of keeping track of cellular time and estimating the presence of genetic misinformation, which accumulates at a constant rate in a constant environment. Senescence is a means of halting DNA production and consequently replication.10,13 The latter occurs to prevent cells with a critical number of defects from replicating, thus decreasing rates of neoplastic transformation. Several studies have shown that endogenous p21 levels are increased by harmful agents (e.g., ionizing radiation, ultraviolet light, toxins, oxygen reactive species, disease) which damage DNA, by DNA autodamage, and by stabilization of p21 mRNA.6,20 The increase in p21 levels is likely mediated through oxidative damage to DNA and its endogenous by-products. Such

Cyclin D1 In the four patients analyzed for cyclin D1, the basal level expression in fb-D (OD = 154 ± 21) was similar to that in fb-P (OD = 155 ± 22), p = 0.55; there was no change in expression with bFGF in either fb-D (OD = 149 ± 26) or fb-P (OD = 152 ± 22). Figure 3 is a representative immunoblot of cyclin D1 in two patients.

DISCUSSION Previous studies have demonstrated that distal fibroblasts (fb-D) from patients with CVI with or without ulceration possess many pathological qualities consistent with the senescent phenotype.4,5,17 The possibility that cellular characteristics of senescence in CVI fibroblasts precede venous ulceration implicates cellular senescence as a causative mechanism in wound formation and poor healing. Phenotypic characteristics of senescence include diminished proliferation, large size and polygonal shape, overexpression of fibronectin, overexpression of matrix metalloproteinases, and increased senescence-associated b-galactosidase activity. Most of these characteristics have been demonstrated in fb-D.4,9,17-19 In this study, we investigated expression of the cell cycle proteins p21, ppRb, and cyclin D1 in venous ulcer cells to determine if fb-D had similar aberrations in cell cycle protein expression as senescent cells. The results clearly demonstrated that, like senescent cells, fb-D overexpress p21 and have decreased levels of ppRb, which are important regulating proteins of the R point and, hence, DNA synthesis. Prior work has demonstrated that although venous ulcer fibroblasts have attenuated growth, fb-D

380 Seidman et al.

by-products of DNA damage are certainly present in the chronic wound environment of a venous ulcer and likely lead to the proliferative block through endogenous DNA breakdown in senescent cells. When fibroblasts are cultured in vitro and removed from the hostile wound environment containing DNA breakdown products, the proliferative block that was present in vivo becomes reversible in vitro. It is at this level of the pathway that bFGF appears to intervene and cause downregulation of p21; where the stimulus is removable, i.e., not endogenous, bFGF is able to reverse the increase in p21 levels. This study was directed at fibroblasts cultured from venous ulcers (C6). Of interest, a study to evaluate fibroblasts from patients with healed venous ulcer (CEAP clinical class 5) would distinguish if the local environment plays a critical role in the proliferative block. Theoretically, the open wound environment would be eliminated in C5 patients; therefore, such a study would distinguish if the open wound environment is an important factor to the proliferative block (normal levels of cell cycle proteins in C5 fibroblasts would indicate that the open wound environment is influential, while altered levels of cell cycle proteins in C5 fibroblasts, as in C6 fibroblasts, would suggest that other cellular factors influence the proliferative block seen in vivo). The fact that bFGF is able to overcome the reversible block to cellular proliferation seen in wound fibroblasts suggests a role for the use of bFGF in the clinical treatment of venous ulcer pathology. Understanding the reversible block of proliferation in pathological fibroblasts from venous ulcers may ultimately elucidate our understanding of cellular senescence, the process of aging, and potential applications toward venous ulcer wound healing.

CONCLUSIONS Venous ulcer fibroblasts demonstrate similar expression of cell cycle proteins as senescent cells. The overexpression of p21 and underexpression of ppRb in venous ulcer fibroblasts are key features in regulating cell growth and explain attenuated growth. However, the proliferation block in venous ulcer fibroblasts is overcome by treatment with bFGF, suggesting that these pathological cultured fibroblasts with senescent characteristics are responsive to growth factors by modulating important cell cycle proteins.

Annals of Vascular Surgery

REFERENCES 1. Callam MH, Harper D. Chronic ulcer of the leg: clinical history. Br J Med 1987;294:1389-1391. 2. Pierce GF, Mustoe TA. Pharmacologic enhancement of wound healing. Annu Rev Med 1995;46:467-481. 3. Nelzen O, Bergquist D, Lindhagen A. Venous and non-venous leg ulcers: clinical history and appearance in a population study:. Br J Surg 1991;78:864-866. 4. Stanley AC, Park HY, Phillips TJ, et al. Reduced growth of dermal fibroblasts from chronic venous ulcers can be stimulated with growth factors. J Vasc Surg 1997;26:994-1001. 5. Mendez MV, Stanley AC, Park HY, et al. Fibroblasts cultured from venous ulcers display cellular characteristics of senescence:. J Vasc Surg 1998;28:876-883. 6. Campisi J. The biology of replicative senescence. Eur J Cancer 1997;33:703-709. 7. Cristofalo VJ, Pignolo RJ. Molecular markers of senescence in fibroblast-like cultures. Exp Gerontol 1996;31:111-123. 8. Dimri GP, Lee X, Basile G, et al. A biomarker that identifies senescent human cells in culture and in aging skin in vivo. Proc Natl Acad Sci USA 1995;92:9363-9367. 9. Mendez MV, Stanley AC, Phillips TJ, et al. Fibroblasts cultured from distal lower extremities in patients with venous reflux display cellular characteristics of senescence. J Vasc Surg 1998;28:1040-1050. 10. Stein GH, Dulic V. Molecular mechanisms for the senescent cell cycle arrest. J Inves Dermatol Symp Proc 1998;3:14-18. 11. Stein GH, Beeson M, Gordon L. Failure to phosphorylate the retinoblastoma gene product in senescent human fibroblasts. Science 1990;249:666-669. 12. Noda A, Ning Y, Venable SF, et al. Cloning of senescent cellderived inhibitors of DNA synthesis using an expression screen. Exp Cell Res 1994;211:90-98. 13. Sherr CJ, DePinho RA. Cellular senesence: mitotic clock or culture shock. Cell 2000;102:407-410. 14. Cristofalo VJ, Phillips PD, Sorger T, et al. Alterations in the responsiveness of Senescent Cells to Growth Factors. J Gerentol 1989;44:55-62. 15. Bergan JJ, Eklof B, Kistner RL, et al. Classification and grading of chronic venous disease in the lower limbs: a consensus statement In: Gloviczki, P, Yao, JST, eds. Handbook of Venous Disorders: Guidelines of the American Venous Forum. New York: Chapman and Hall, 1996, pp 652660. 16. Simpson IA, Soanne O. A simple, rapid, and sensitive method for measuring protein concentration in subcellular membrane fractions prepared by sucrose density ultracentrifugation. Anal Biochem 1982;119:424-427. 17. Raffeto JD, Mendez MV, Phillips TJ, et al. The effect of passage number on fibroblast cellular senescence in patients with chronic venous insufficiencey with and without ulcers. Am J Surg 1999;178:107-112. 18. West MD, Pereira-Smith OM, Smith JR. Replicative senescence of human skin fibroblasts correlates with a loss of regulation and overexpression of collagenase activity. Exp Cell Res 1989;184:138-147. 19. Seidman CS, Raffetto JD, Marien BJ, et al. bFGF induced alterations in cellular markers of senescence in growth rescued fibroblasts from chronic venous ulcer and venous reflux patients. Ann Vasc Surg 2003;17:239-44. 20. Chen Q, Fischer A, Reagan JD, et al. Oxidative DNA damage and senescence of human diploid fibroblast cells. Proc Natl Acad Sci USA 1995;92:4337-4341.