Effect of Estrogen on Cellular Metabolism of the Human Anterior Cruciate Ligament

Abstract

Previous studies from this laboratory have established the presence of estrogen receptors in the human anterior cruciate ligament. The purpose of this study was to investigate the effects of 17β-estradiol on cell proliferation and procollagen levels, as an indicator of collagen synthesis, in the human anterior cruciate ligament fibroblasts. Fibroblast proliferation and procollagen synthesis in response to near log concentrations of 17β-estradiol (at 0.0029 ng/mL, 0.025 ng/mL, 0.25 ng/mL, 2.5 ng/mL, and 25 ng/mL) were assessed with the measurement of ³H-thymidine incorporation and Types 1 and 3 procollagen specific equilibrium radioimmunoassays. On Days 1 and 3, there was a dose dependent decrease in the proliferation of anterior cruciate ligament fibroblasts with increasing estradiol concentrations. This dose dependent effect of decreased fibroblast proliferation with increasing estradiol concentrations became less apparent at 7, 10, and 14 days. On Days 1 and 3, procollagen synthesis decreased in a dose dependent manner with increasing estradiol concentrations. On Days 7, 10, and 14, this dose dependent effect was attenuated. No significant differences in Type 3 procollagen synthesis by anterior cruciate ligament fibroblasts were observed with varying estradiol concentrations at any of the designated points. These early physiologic changes in fibroblast proliferation and Type I procollagen synthesis may provide a biologic explanation for the increased anterior cruciate ligament injury rate observed in female athletes, suggesting that it is the acute cyclic variations in the female athlete who is menstruating that predisposes her to ligamentous injury.

Recent epidemiologic studies have recognized within various sports a significantly higher rate of anterior cruciate ligament injury in female athletes than in their male counterparts within various sports.^12,27,41,44 Although the etiology of this phenomena remains unclear, possible explanations include gender differences in ligament or muscle strength, conditioning and endurance, anatomy, and training techniques.^7,23,30,33 Unique to the female athlete is her exposure to constantly changing hormonal milieu throughout her reproductive years. For most of her life, the female athlete is exposed to a rhythmic variation in either endogenous hormones during a regular menstrual cycle or exogenous hormones via oral contraceptives. The subsequent fluctuations in the serum estrogen levels may in turn translate to changes in metabolism of cells in the anterior cruciate ligament. Realizing that estrogen can have individual or interactive effects on various metabolic processes, it is possible that estrogen also may have an influence on ligament structure, composition, and integrity.

A significant amount of data exist concerning the effects of estrogen on bone density and osteogenic cells^15,37; however, few studies have explored the effect of the female sex hormones on ligamentous tissue. The identification of estrogen receptor positive fibroblasts in the human anterior cruciate ligament strongly suggests that female sex hormones may have an effect on the structure and composition of this ligament.²⁰ The purpose of this study was to characterize the effects of estrogen on human anterior cruciate ligament fibroblast proliferation and procollagen synthesis within an in vitro system. It is hypothesized that there exists an inverse, dose dependent relationship between cellular metabolism (fibroblast proliferation and procollagen synthesis) and estrogen concentrations.

MATERIALS AND METHODS

Study Design

To compare the effects of exogenously applied estrogen on human anterior cruciate ligament fibroblasts, a primary cell culture was established using a modification of the explant culture technique by Nagineni et al.²⁶ With exposure to various concentrations of estrogen, fibroblast proliferation and rate of collagen synthesis were monitored concurrently using ³H-thymidine incorporation and Types 1 and 3 procollagen specific equilibrium radioimmunoassays.

The cell line was obtained from an otherwise healthy, 32-year-old multiparous woman undergoing total knee replacement secondary to traumatic arthritis. The patient was a nonsmoker, and conservative management had failed for her arthritis. This investigation was approved by the institutional human subjects protection committee, and informed consent was obtained before enrollment.

Western Blotting

To test the efficacy and reliability of the procollagen monoclonal antibodies within a tissue culture system, western blot analysis was performed. Logarithmic dilutions of certified fetal calf serum and cell culture homogenates were compared with the radioimmunoassay standards for Types 1 and Type 3 procollagens. Specifically, media and cell lysates from Day 3 samples dosed with 0.025 ng/mL of estradiol were homogenized and centrifuged for 5 minutes at 15,000 g. Thereafter, the radioimmunoassay standards, logarithmic dilutions of certified 10% fetal calf serum (manufacturer specifications indicate 29 pg/mL of estradiol in stock fetal calf serum; Gibco Technologies Inc, Grand Island, NY), and the cell culture supernatant were run in a 10% gradient polyacrylimide gel for 2 hours. After transfer to nitrocellulose paper and blocking of nitrocellulose with nonfat milk, the primary antibody followed by the secondary immunofluorescent tagged antibody were applied separately for 1 hour each at room temperature. The relative banding patterns and intensities for Type 1 procollagen (Fig 1) and Type 3 procollagen were compared.

Fig 1: Western blot analysis comparing antiprocollagen Type 1 antibody reactivity to radioimmunoassay standards for Type 1 procollagen, fetal calf serum, and Day 3 samples treated with 0.025 ng/mL of estradiol (E₂). Antiprocollagen Type 1 antibody reactivity to fetal calf serum (arrows) is absent. RIA = radioimmunoassay.

Cell Culture Protocol

The anterior cruciate ligament of the subject was obtained under aseptic conditions in the operating room. The tissue was placed immediately into sterile tissue culture media. All additional manipulations were conducted under a sterile biologic hood. Four millimeters of the tissue adjacent to the bony insertion were discarded to avoid cartilaginous and periosteal contamination. The synovial sheath was excised. The remaining specimen was minced finely by transverse sections with a surgical blade and divided equally into four 75-cm² tissue culture flasks (Corning Glass Works Inc, Corning, NY) containing 15 mL growth medium, which consisted of Minimal Essential Medium alpha (MEM-α) supplemented with certified 10% fetal calf serum (manufacturer specifications indicate 29 pg/mL of estradiol in stock fetal calf serum; Gibco Technologies Inc), gentamycin (5 μg/mL), and streptomycin (100 μg/mL).

Currently, estradiol free fetal calf serum is not available commercially. However, charcoal stripped fetal calf serum contains the lowest known concentration of estradiol available in the biotechnology market (at 0.0016 ng/mL of estradiol in 10% charcoal stripped fetal calf serum). The concentration of estradiol in charcoal stripped media was not significantly different than the 0.0029 ng/mL of estradiol present in 10% certified fetal calf serum. As such, it was opted to supplement the media with 10% certified fetal calf serum, which has been shown to contain a sufficient amount of growth factors to stimulate fibroblast proliferation optimally.²⁷

Primary cell cultures were maintained in a humidified atmosphere of 5% CO₂ and 95% air at 37° C without a change in media until 1 week after harvest. Thereafter, culture media was changed every 3 days. The growth of cells was monitored with an inverted phase contrast microscope. Once cells grew to 90% confluence, the explants were detached with 0.25% trypsin in calcium and magnesium free phosphate buffered saline (pH = 7.4) with 0.1 mmol ethylenediametetraacetic acid (Gibco Technologies Inc) for 5 minutes at 37°C. At this and subsequent subculturing, the cells were split at a 1:5 ratio.

Estrogen Administration

Fibroblasts cultured in the third passage to confluence were released with trypsin, washed in growth medium supplemented with 10% certified fetal calf serum, and cultured in 12-well Corning cluster plates at a density of approximately 5 × 10⁴ cells per well. To show accurate experimental techniques, Group 1 was raised in serum free media. To simulate the physiologic and supraphysiologic levels of estradiol present in the human menstrual cycle at 0.025 to 0.3 ng/mL,³¹ experimental Groups 2 through 6 were supplemented with 10% certified fetal calf serum and received a near final logarithmic concentration of 17β-estradiol for 1, 3, 7, 10, and 14 days, as outlined in Table 1. The media with the appropriate concentrations of estrogen were replaced every 3 days.

TABLE 1: Tabulation of Final Logarithmic Concentrations of Estradiol Administered per Group

³H-Thymidine Incorporation

Deoxyribonucleic acid (DNA) content and DNA synthesis of control and experimental groups were monitored using ³H-thymidine incorporation of cultures.²⁶ Confluent cultures at Passage 3, which had been exposed to the respective concentrations of estradiol for 1, 3, 7, 10, and 14 days, were incubated with 10 μL ³H-thymidine (1 μC/mL, New England Nuclear, Boston, MA) for 24 hours at 37°C in a humidity chamber at 5% CO₂ and 95% air. Subsequently, agarose was precipitated and the free ³H-thymidine was removed. The remaining contents were transferred into glass test tubes containing 3 mL of 10% trichloroacetic acid. The contents of the tubes were vortexed and left on ice for 30 minutes to precipitate any macromolecules. Trichloroacetic acid perceptible radioactivity was collected by filtration through Whatman glass microfiber filters using a millipore filtration unit. Filters were air dried and placed into scintillation vials containing 10 mL of scintillation fluid (Liquiscint, National Diagnostics Inc, Manville, NJ, with 10% acetic acid), and the radioactivity was measured in a scintillation counter (Beckman 1801 Beta Counter, Beckman Inc, Pasadena, CA). Cell numbers in replicate of six wells were determined, and the thymidine incorporation was expressed in counts per minute/10³ cells.

Quantification of Collagen Synthesis

The effects of varying concentrations of estradiol on procollagen levels, as an indicator of collagen synthesis, were determined using commercially available equilibrium radioimmunoassays. Levels of procollagen Type 1 were measured with the use of a monoclonal antibody directed against the trimeric carboxyterminal of procollagen 1 (PICP, Farmos Diagnostica, Turku, Finland),^5,25 and procollagen Type 3 levels were measured by a monoclonal antibody specific for the aminoterminal of procollagen Type 3 (PIIINP, Farmos Diagnostica).³²

Confluent cultures at Passage 3, which had been exposed to the respective concentrations of estradiol for 1, 3, 7, 10, and 14 days, were incubated with the respective ¹²⁵I tracer and antibody for 2 hours at 37°C. Thereafter, the precipitating second antibody complex was added to separate the bound from the free tracer element. The tube was centrifuged and decanted after a 30-minute incubation at room temperature. The bound tracer in the pellet was placed into scintillation vials containing 10 mL of scintillation fluid (Liquiscint, National Diagnostics Inc, with 10% acetic acid), and the gamma counts were measured in a scintillation counter (Micromedic Systems 4/600 Automatic Gamma Counter, Micromedic Systems Inc, Horsham, PA). The resulting counts are inversely proportional to the amount of procollagen present in each sample. Cell numbers in replicate were determined, and thymidine incorporation was expressed as count per minute/10³.

Statistical Analysis

Radiolabeled substrate incorporation was expressed as the mean ± standard error mean. Thereafter, data were analyzed using linear regression analysis. Statistical analysis of cellular proliferation, ³H-thymidine incorporation, and levels of procollagen Types 1 and 3 were performed by matched sign rank test. A significant level of 0.05 was selected and presented in the form of p values.

RESULTS

Western Blot

Western blot analysis of Type 1 and Type 3 procollagen radioimmunoassays showed numerous significant preliminary results that establish the scientific foundations of this study. The increased intensity of the banding pattern of the positive control radioimmunoassay standards for procollagens 1 and 3 proved the efficiency of the kits in recognizing increasing concentrations of the respective procollagens. In addition, as shown by the absence of any banding pattern when fetal calf serum was incubated with the antiType 1 procollagen antibodies (Fig 1, arrows), it was concluded that serum contains undetectable levels of procollagen. Similarly, Type 3 procollagen western blot analysis showed a banding pattern identical to that of the Type 1 procollagen. As a result, the quantitative measurements obtained through this assay reflect the actual level of procollagen produced by the human fibroblasts of the tissue culture system.

Fibroblast Proliferation

³H-thymidine incorporation is shown in Figure 2. On Days 1 and 3, there was a dose dependent decrease in the proliferation of anterior cruciate ligament fibroblasts with increasing estradiol concentrations (p < 0.01, R² = 0.56; p < 0.01, R² = 0.23 respectively; linear regression analysis). This dose dependent decrease in fibroblast proliferation with increasing estradiol concentrations became less apparent with time. By Days 7, 10, and 14, no statistically significant correlation was found between estrogen concentration and fibroblast proliferation.

Fig 2: Histogram describing anterior cruciate ligament fibroblast proliferation as function of estradiol concentration and time.

Procollagen Synthesis

The effects of estradiol on Type 1 procollagen synthesis are shown in Figure 3. On Days 1 and 3, procollagen synthesis decreased in a dose dependent manner with increasing estradiol concentrations (p < 0.05, R² = 0.33; p < 0.5, R² = 0.13, respectively; linear regression analysis). Again, on Days 7, 10, and 14, this dose dependent decrease in Type 1 procollagen synthesis was attenuated, and no statistically significant relationship between estrogen concentrations and procollagen 1 synthesis was seen at these points.

Fig 3: Histogram describing anterior cruciate ligament fibroblast Type 1 procollagen synthesis as a function of estradiol concentration and time.

The effects of estradiol on Type 3 procollagen synthesis are shown in Figure 4. No statistically significant differences in Type 3 procollagen synthesis were observed with varying estradiol concentrations at any time.

Fig 4: Histogram describing anterior cruciate ligament fibroblast Type 3 procollagen synthesis as a function of estradiol concentration and time.

DISCUSSION

Collagen is produced by fibroblasts and organized in a hierarchical manner from fascicles to microfibrils, whereby it subsequently performs the major loadbearing function of the anterior cruciate ligament.^13,36 Alterations in the metabolism of fibroblasts, by local stimuli with growth factors¹⁹ or systemic hormonal stimuli,¹⁶ have an influence on the quantity, type, and stability of the collagen in the anterior cruciate ligament.^2,4,22,34

Collagen is synthesized as a larger molecule with propeptide extensions at both ends that are cleaved stoichiometrically during fibrinogenesis and released into the extracellular fluid. The sequence removed from the carboxyterminal end of the molecule, known as the carboxyterminal propeptide, can be found in the extracellular matrix. As a result, the extracellular levels of these cleaved extensions, which have been named as procollagens by the manufacturer of the equilibrium radioimmunoassays, directly reflect the extracellular levels of collagen. This stoichiometric proportionality has been supported by numerous studies in which levels of Type 1 and Type 3 procollagen were increased when collagen synthesis was stimulated by growth hormone therapy.^4,38,40 Similarly, two studies of bone biopsy specimens from patients with various forms of metabolic bone disease²⁹ and in postmenopausal women¹⁵ have reported that serum procollagen Type 1 levels correlate significantly with histomorphometric measure of bone formation and density.

It is known that female sex hormones have a widespread effect on the growth and development of bone, muscle, and connective tissues.^18,35,43 Although numerous animal studies have documented the effect of estrogen on collagen metabolism in various tissues, ligaments have not been the subject of much investigation.^{3,7,10,11,13,35,43} Several in vivo studies have shown that the quantity of collagen in certain tissue is influenced by estrogen. Long term (in terms of weeks) estrogen administration causes a decrease in the total amount of collagen in the rat hip joint capsule, skin, aorta, and tail tendon.^11,13,35 In contrast, two separate studies have documented an estrogen dose dependent increase in the production of Type 1 collagen messenger ribonucleic acid (mRNA) in osteoblasts.^9,24

Many studies have shown how local estrogen concentrations may influence connective tissue metabolism. Recently, it has been reported that oral estrogen and progesterone therapy significantly reduced Type 1 procollagen in postmenopausal women.¹⁵ Equally, within a similar group of patients, there is an increase in total body Type 3 collagen content with daily administration of estrogen.¹⁴ Estrogen administration is known to decrease total collagen acutely in rat tendon and fascia and to decrease collagen synthesis acutely in rat periodontal tissue.^8,10 In the rat uterus and primate sex skin, estrogen acutely increases newly synthesized collagen, mostly Types 1 and 3, while concurrently increasing collagen degradation, with the net result being a dynamic flux between the synthesis and the degradation of collagens.^3,8 Although estrogen administration reduces stiffness in periarticular connective tissue of rabbit knees by approximately 50% compared with untreated animals,¹ it also has been shown to decrease hexosamine and soluble collagen content.⁴³ Thus, it stands to reason that acute fluctuations in the serum estrogen concentration during the menstrual cycle may induce changes in the metabolism, amount, and type of the collagen in the anterior cruciate ligament.

Previous work done in this laboratory has established the presence of estrogen receptors in the human anterior cruciate ligament.²⁰ The identification of estrogen receptor positive fibroblasts in the human anterior cruciate ligament strongly suggests that female sex hormones may have an effect on the structure and composition of this ligament. In a recent study, a dose dependent decrease in cellular proliferation and total collagen synthesis was seen when rabbit anterior cruciate ligament fibroblasts were treated with estradiol for 2 weeks.²¹

Estrogen exerts a dose dependent effect on cellular metabolism. Similar to the mentioned preliminary results seen in the rabbit fibroblast, the results of the current in vitro study indicate that at physiologic levels of estradiol (0.025-0.25 ng/mL), there was a significant decrease in human fibroblast proliferation and specifically Type 1 procollagen levels. Eventually, the inhibitory effect of estradiol became asymptotic to the horizontal at the pharmacologic estradiol concentrations (2.5 and 25 ng/mL), indicative of maximal estradiol effect with the saturation of the estrogen receptor or downregulation of the receptor protein.

Regarding a temporal relationship, the effects of estradiol were most pronounced in the early periods of hormone exposure, Days 1 and 3. By Day 7, the dose dependent effects of estradiol on fibroblast proliferation (Fig 2) and Type 1 procollagen levels became attenuated. Again, this phenomenon with time may be attributable to normal cellular physiology of receptor saturation, cell contact inhibition, or more commonly, because of downregulation of the receptor protein.

In contrast to Type 1 procollagen, no significant changes were observed in Type 3 procollagen levels with varying doses of estradiol or with time. These results clearly show that Type 3 procollagen synthesis remains unaffected with estrogen exposure, indicating that presumed differences in fibroblast metabolism are manifested largely by changes in Type 1 procollagen synthesis, rather than Type 3.

The proper function of a ligament depends on the appropriate type, synthesis, assembly, crosslinking, and remodeling of collagen.⁶ This complex interplay between synthesis and remodeling of collagen is influenced by hormones, exercise, and immobilization.^{18,28,31,36,39,42} The amount of collagen bundles and the individual types of collagen influence the ability of the tendon to withstand loading. It also is known that Type 1 collagen imparts a great mechanical strength to connective tissues, whereas Type 3 collagen has been correlated directly with tissue elasticity.²² Although the relative strength of Type 1 collagen to Type 3 collagen remains unknown, it generally is agreed that a larger ratio of Type 1 to Type 3 collagen in the ligament is indicative of greater strength, whereas a lower ratio may be characteristic of tissue laxity.¹⁷ In applying this understanding to the outlined results, it follows that the relative decrease in Type 1 procollagen synthesis with increasing estradiol concentrations may translate to ligament weakening.

The observed causal interrelationship between acute changes in estradiol concentrations and procollagen synthesis continues to complement the clinical significance of these findings when one considers that female athletes are exposed to rhythmic variations in either endogenous hormones during a menstrual cycle or exogenous, nonphysiologic levels of hormones via the use of oral contraceptives. Taken together, these acute fluctuations in the serum estrogen levels translate to acute changes in fibroblast metabolism. The net effect of these cellular events can result in reduced strength of the ligament and subsequently lead to a higher risk of anterior cruciate ligament injury in female athletes.

The results clearly show that physiologic and nonphysiologic levels of estradiol have a significant dose dependent effect on the fibroblasts of the anterior cruciate ligament in the early period of hormone exposure. Fibroblast proliferation and Type 1 procollagen synthesis are reduced significantly with increasing estradiol concentrations. Thus, cumulative or acute fluctuations in serum estrogen concentrations, such as those occurring during the menstrual cycle or via the additive effect of exogenous estrogen in oral contraceptives, may induce changes in the metabolism of anterior cruciate ligament fibroblasts within this in vitro system. The resulting structural and compositional changes on the molecular level could result in decreased strength of the anterior cruciate ligament and predispose female athletes to ligament injury. Future studies in animal models will be needed to show estrogen induced mechanical weakening of the anterior cruciate ligament.

The rate of anterior cruciate ligament injury and causes of such injuries in women are distinct from those in men and deserve equal attention. Clearly, more research is needed to address the associative causative factors and the role of prevention in curtailing the rate of ligamentous injury in female athletes. The results of this study recognize the importance of female sex hormones, specifically estrogen, on collagen metabolism and provide a basis for future in vivo studies investigating the effects of estrogen on the biomechanical properties of the anterior cruciate ligament.