Effects of Intranasal 17β-Estradiol on Bone Turnover and Serum Insulin-Like Growth Factor I in Postmenopausal Women

Estrogen therapy, using either oral or transdermal routes, decreases bone turnover and prevents postmenopausal bone loss. It has been suggested that oral and transdermal 17β-estradiol (E₂) may have different effects on serum insulin-like growth factor I (IGF-I), a potent bone-forming growth factor. In this study we investigated the effects of a new route of administration, the intranasal E₂ spray (S21400), on bone turnover and circulating IGF-I and IGF-binding protein-3 (IGFBP-3). Four hundred and twenty early postmenopausal women (<5 yr since menopause; mean age, 52 yr) were enrolled in a 3-month, double blind, placebo-controlled study of four doses of intranasal E₂ (100, 200, 300, and 400μ g/day), two doses of oral E₂ valerate (1 or 2 mg/day), and placebo. One hundred and twelve women were further treated for 12 months with intranasal E₂ (300 μg/day, i.e. the dose that has been shown to be adequate for the majority of postmenopausal women). Markers of bone resorption (urinary type I collagen C telopeptides) and formation [serum osteocalcin, serum type I collagen N-terminal extension propeptide (PINP), and serum bone alkaline phosphatase (BAP)] were measured at baseline, 1 month, 3 months, and 15 months. Serum IGF-I and IGFBP-3 were measured at baseline, 1 month, and 3 months. Urinary type I collagen C telopeptides decreased significantly in all active treatment groups as soon as 1 month (P < 0.001 vs. placebo) and continued to decrease at 3 months with a dose effect for intranasal E₂. Serum osteocalcin and PINP did not change at 1 month for oral E₂ (1 and 2 mg), but decreased significantly at 3 months. In contrast, formation markers increased significantly at 1 month for the two highest doses of intranasal E₂ (P < 0.01 vs. placebo for osteocalcin and BAP) and did not decrease at 3 months. Oral E₂ induced a marked decrease in circulating IGF-I as early as 1 month, which was amplified at 3 months (−29% and −32% for 1 and 2 mg, respectively), whereas no significant change from placebo was observed for intranasal E₂ during the 3-month period. Changes in circulating IGF-I correlated significantly (P < 0.01) with changes in osteocalcin, PINP, and BAP at 3 months. Oral and intranasal E₂ did not induce any significant change from placebo in serum IGFBP-3 at both 1 and 3 months. After 1 yr of treatment with intranasal E₂ (300μ g/day), both resorption and formation markers decreased, reaching the levels in premenopausal women, regardless of the type of treatment during the first 3 months.

We conclude that E₂ administered by this new nasal route normalizes bone turnover to premenopausal levels. The delayed decrease in bone formation observed with intranasal E₂ compared to oral E₂ may be related to different effects on serum IGF-I levels.

ESTROGEN supplementation, given mainly as conjugated equine extracts or 17β-estradiol (E₂), has been shown to decrease bone turnover, prevent postmenopausal bone loss, and significantly reduce fracture risk in both early and late postmenopausal women (1–5). The two most widely used routes of estrogen therapy are oral formulations and transdermal patches. It has been suggested that the short term effect (within 3 months) of estrogen on biochemical markers of bone turnover may differ according to the route of administration. Johansen et al. (6) have shown that osteocalcin did not change after 3 months of treatment with percutaneous estrogen at effective dose, whereas during the same time there was a 30% decrease in those women treated with oral estrogen. However, after 12 months of treatment, the decrease in osteocalcin was similar in both treatment groups, and levels returned to within the normal premenopausal range. Similar findings have been reported by Ho and Weissberger (7) in a small nonrandomized and nonplacebo-controlled study that showed an increase in osteocalcin and C-terminal type I collagen extension propeptide after 3 months of transdermal estrogen therapy, whereas the same markers decreased with oral estrogen treatment. The discrepancy between the two routes of administration regarding bone formation markers has been suggested to be related to the GH-insulin-like growth factor (IGF) axis. Although oral estrogens suppress plasma IGF-I levels despite increasing basal and stimulated GH secretion (8–10), transdermal administration has been found to have no effect on serum GH and plasma IGF-I levels (11) or to increase peripheral IGF-I concentrations (12). The different effect of oral and transdermal estrogen on serum IGF-I may be due to the first pass effect of oral estrogen on liver metabolism (13–15).

Although oral and transdermal estrogen have demonstrated their efficacy in preventing postmenopausal bone loss (4, 16), both of them have some disadvantages: the significant intestinal and hepatic first pass effects for the oral formulation as outlined above and the large interindividual variation in bioavailability, variable adhesion, and local dermatological reactions of the transdermal route (17–19).

A new E₂ spray formulation for nasal administration has been developed (S 21400, Aerodiol, Institut de Recherches Internationales Servier, France), avoiding the limitations of oral and transdermal estrogen. Intranasal absorption is facilitated by the highly vascularized, microvillous nature of the nasal mucosa (20). Intranasally absorbed estradiol has a very rapid uptake, achieving maximal plasma levels within 10–30 min. The plasma concentration returns rapidly to 10% of the maximal value approximately 2 h after administration and to the levels found in untreated postmenopausal women within 12 h (21). As a consequence, daily intranasal administration results in a pulse-like estrogen profile, rather than the relatively sustained serum levels attained with both oral and transdermal administration (22). However, no study to date has ascertained whether a sustained profile is a requirement of efficacy, and it has recently been shown that intranasal E₂ increases serum estradiol exposure to levels similar to those achieved with oral E₂ (1–2 mg) and with an efficacy similar to that of oral estrogens in reducing menopausal symptoms (23).

The aims of this double blind, placebo-controlled clinical study were to investigate the short and long term effects of intranasal E₂ on bone turnover and serum IGF-I and to compare its effects to those of oral E₂.

Subjects and Methods

Subjects

Four hundred and twenty Caucasian postmenopausal women, aged 45–60 yr (6 months to 5 yr postmenopause), with climacteric symptoms were included in a multinational European, double blind, placebo-controlled study. These women were randomly allocated to one of seven treatment groups: intranasal E₂ (100, 200, 300, or 400 μg/day; delivered by one spray per nostril), each with a placebo capsule, or oral E₂ valerate (1 or 2 mg/day), each with a placebo spray, or double placebo. E₂ was administered double blind once daily throughout the study period of 0–12 weeks and then for a further 2 weeks supplemented each evening by medroxyprogesterone acetate (MPA; 5 mg) to promote endometrial shedding. Treatment was administered between 0600–1000 h, and women were asked to take treatment at the same time every day. No calcium supplement was administered. For all women a blood sample and a 2-h urine sample were collected fasting before nasal administration at baseline and 1 and 3 months before any MPA supplementation. After the 3-month, double blind, placebo-controlled trial, 112 of these women regardless of the treatment allocation during the first 3 months were further treated with intranasal E₂ (300 μg/day), i.e. the dose that has been found to be optimal to reduce menopausal symptoms, as assessed by Kupperman index and incidence of hot flushes (23). During this second phase of the study, nonhysterectomized women received various cyclic progestative treatments, including MPA (5 mg/day), dydrogesterone (20 mg/day), chlormadinone acetate (10 mg/day), and promegestone (0.5 mg/day).

All subjects gave informed consent for their participation in the trial, which was conducted in accordance with the Declaration of Helsinki and local ethical committee approval.

Markers of bone turnover

Serum osteocalcin was measured with a human specific two-site immunoradiometric assay (ELSA-OSTEO, CIS Bio International, Gif-sur-Yvette, France), which recognizes a large N-terminal midfragment in addition to the intact molecule (24). The intra- and interassay coefficients of variation (CVs) are less than 4% and 6%, respectively, and the sensitivity is 0.4 ng/mL (26). Serum N-terminal extension propeptide of type I collagen (PINP) was measured by a new RIA that recognizes the intact circulating form of PINP (Intact PINP, Farmos Diagnostica, Uppsala, Sweden) (25). The intra- and interassay CVs are below 5% and 8%, respectively, and the sensitivity is 1 ng/mL (25). Serum bone alkaline phosphatase (BAP) was measured with a human specific immunoradiometric assay using two monoclonal antibodies directed against the human bone isoenzyme and BAP purified from human SAOS-2 osteosarcoma cells as a standard (Ostase, Hybritech, Inc., San Diego, CA). This assay cross-reacts only 16% with the circulating liver isoenzyme. The intra- and interassay CVs are less than 10% (26). Urinary type I collagen C-telopeptide breakdown products (CTX) were measured by an enzyme-linked immunosorbent assay (CrossLaps ELISA, CIS Bio International) based on an immobilized synthetic peptide with an amino acid sequence specific for a part of the C-telopeptide of theα ₁-chain of type I collagen (Glu-Lys-Ala-His-Asp-Gly-Gly-Arg; CrossLaps antigen) (27). The intra- and interassay CVs are below 5% and 8%, respectively, and the sensitivity is 0.05 μg/mL (28).

The premenopausal range for biochemical markers of bone turnover was determined in 134 healthy premenopausal women from 35–55 yr of age (mean, 40.6 ± 5.3 yr) belonging to a prospective population-based study (OFELY study) (29). Mean (±1 sd) values of premenopausal women were 18.3 ± 4.7 ng/mL, 42 ± 15 ng/mL, 8.7 ± 2.7 ng/mL, and 189 ± 88 μg/mmol creatinine for serum osteocalcin, PINP, BAP, and urinary CTX, respectively.

Serum IGF-I and IGFBP-3

Serum IGF-I was measured by a competitive RIA (IGF-I by extraction, Nichols Institute Diagnostics, San Juan Capistranao, CA) after acid-ethanol extraction. The intra- and interassay CVs are below 3% and 8%, respectively, and the sensitivity is 0.06 ng/mL.

Serum IGFBP-3 was measured by a two-site immunoradiometric assay (Diagnostic Systems Laboratories, Inc., Webster, TX). The intra- and interassay CVs are below 5%, and the detection limit is 0.5 ng/mL.

All biochemical measurements for both the postmenopausal women involved in the clinical trial and the premenopausal women used to establish normal ranges were performed in a central laboratory (Prof. Delmas, Synarc, Lyon, France).

Statistical analysis

All data are shown as the mean ± 1 sd. Measurements of serum osteocalcin, serum BAP, and urinary CTX were performed in all randomized women, whereas serum PINP, IGF-I, and IGFBP-3 were assessed only in the per protocol population, including the 326 women defined as those who received at least 56 days of study treatment, who had confirmed postmenopausal status with centrally verified baseline levels of serum FSH above 30 mIU/mL and serum estradiol below 30 pg/mL and no major protocol deviations. Statistical analysis was performed on the per protocol population. Comparison of bone markers, serum IGF-I, and serum IGFBP-3 between treatment groups was performed using a covariance analysis adjusted on baseline values. In case of a significant group effect, pairwise comparison were performed vs. placebo, where the α risk is corrected according to Bonferroni’s procedure. The within-group change with time was studied using a two-way ANOVA (patient × time) followed in case of significant time effect by a Newman-Keuls test. The correlation coefficients (Pearson) among the percent change in serum IGF-I, serum IGFBP-3, and biochemical markers at 3 months were obtained from linear regression analysis.

Results

Baseline relevant clinical characteristics are shown in Table 1. Women in the different treatment groups did not differ for history of previous hormone replacement therapy and percentage with surgical menopause. Women in the oral E₂ (1 mg/day) group were slightly older than those in the other groups.

Short term effects of oral and intranasal E₂ on bone turnover

At baseline, there was no significant difference between groups for biochemical markers of bone turnover (Table 2). Bone resorption, assessed by urinary CTX, was significantly decreased compared to that after placebo (P < 0.001) both by oral (−20% and −40% for 1 and 2 mg, respectively) and intranasal (−13% to −27%) E₂ within 1 month of treatment (Table 2 and Fig. 1). Urinary CTX continued to decrease at 3 months for both routes of administration. At 3 months, a dose effect was observed for intranasal E₂, with a larger decrease for the dose of 200 μg compared to 100 μg, but no additional effect with doses of 300 and 400 μg.

Contrasting with the similar time and magnitude of effect of oral and intranasal E₂ on bone resorption, bone formation markers responded differently for the two routes of administration. After 3 months of treatment with oral E₂, serum osteocalcin and PINP, but not serum BAP, significantly decreased compared to placebo (−10% and −22% for osteocalcin and PINP, respectively, for the 2-mg dose; Table 2 and Fig. 1). In contrast, neither osteocalcin, PINP, or BAP levels changed significantly after 3 months of treatment with intranasal E₂. Interestingly, the higher doses of intranasal E₂ induced a significant increase in serum osteocalcin (14% and 9% for 300 and 400 μg, respectively; P < 0.05) and serum BAP (10%, 13%, and 7% for 200, 300, and 400 μg, respectively; P < 0.05) at 1 month, an effect that was not observed with oral administration (Table 2 and Fig. 1). Similar results at 1 and 3 months were observed in the intention to treat population for urinary CTX and serum osteocalcin (not shown).

Short term effects of oral and intranasal E₂ on serum IGF-I and IGFBP-3 levels: relationships with short term changes in bone turnover

To explain the different short term effects of oral and intranasal E₂ on bone formation markers, serum IGF-I and IGFBP-3 were measured. Oral E₂ induced a rapid and marked decrease in serum IGF-I with declines of −23% and −28% at 1 month for the 1- and 2-mg doses, respectively (P < 0.0001 vs. placebo). At 3 months, the decrease was slightly larger, reaching −29% and −32% for 1 and 2 mg, respectively. In contrast, with intranasal E₂ at any dose, there was no significant change in serum IGF-I at 1 month (Table 3 and Fig. 1) and only a slight decrease at 3 months (∼−15% for all doses), which was, however, not significantly different from the effect of placebo. No significant change from the placebo group of serum IGFBP-3 levels was observed for both oral and intranasal E₂ during the 3-month period (Table 3). Baseline levels of serum IGF-I correlated with serum IGFBP-3 (r = 0.45; P < 0.001). No significant correlation was observed between changes in IGF-I and changes in IGFBP-3 at both 1 and 3 months. Changes in serum IGF-I at 3 months, but not changes in serum IGFBP-3, correlated significantly with changes in serum osteocalcin, serum PINP (Fig. 2), and serum BAP (r = 0.18; P = 0.006).

Long term effects of intranasal E₂ on bone turnover

After the first 3 months of treatment, 112 women continued treatment with intranasal E₂ (300 μg/day). As shown in Table 4 and Fig. 3, regardless of the type of therapeutic regimens and doses used during the first 3 months, 12 months of treatment with intranasal E₂ induced significant decreases in urinary CTX (−45%), serum osteocalcin (−26%), serum PINP (−31%), and serum BAP (−29%). Levels of bone markers returned to the premenopausal range in 96%, 88%, 97%, and 91% of the patients for urinary CTX, serum osteocalcin, serum PINP, and serum BAP, respectively.

Discussion

This large study demonstrates for the first time the efficacy, over 1 yr of treatment, of a nasal spray of E₂ in normalizing the bone turnover rate to levels in premenopausal women. Interestingly, intranasal E₂ exhibited short term effects on bone turnover markers that are different from those of oral E₂. This study also provides evidence that these differences could be mediated in part by effects on serum IGF-I, which was markedly suppressed by oral, but not intranasal, estrogen.

Both intranasal and oral E₂ induced a significant and dose-dependent decrease in urinary CTX, a specific marker of bone resorption. The significant decrease in urinary CTX as soon as 1 month induced by intranasal E₂ is similar to that observed in this study with oral E₂ and with that previously reported for oral conjugated equine estrogen (30), in agreement with a direct inhibition of the osteoclastic activity. The 30–35% decrease in urinary CTX after 3 months of treatment observed with intranasal E₂ at doses of 200, 300, and 400 μg was similar to that obtained with oral 1 mg E₂ and slightly less than that with the 2-mg dose. This magnitude was also in the range of that previously reported for transdermal E₂ (50 or 100 μg) (31, 32) and 0.625 mg oral conjugated equine estrogen (30). Thus, the short term antiresorptive effects of intranasal E₂ are similar to those of other estrogen regimens that have been proven to prevent postmenopausal osteoporosis.

The decrease in bone formation observed after 3–6 months of estrogen therapy has been generally ascribed to a decrease in bone turnover and a loss of remodeling sites, rather than to any direct effect on osteoblasts. Accordingly, for oral E₂ we found no significant changes of serum osteocalcin, serum PINP, and serum BAP after 1 month of treatment, but there was a significant decrease in the two former markers after 3 months. In contrast, intranasal E₂ at any dose did not decrease bone formation during the first 3 months. More importantly, 1 month of treatment with the two highest doses of intranasal E₂ induced a slight (9–22%), but significant, increase in formation markers, suggesting that in addition to its antiresorptive effect, intranasal E₂ could directly stimulate bone formation. Several in vitro studies have reported a direct stimulatory effect of E₂ on rodent and human osteoblastic cells, including an increase in α1 type I collagen messenger ribonucleic acid (mRNA) and osteocalcin secretion (33–35). However, the effects of E₂ on bone formation have been difficult to demonstrate in vivo, although some histomorphometric studies in rodents and humans have reported an increase in trabecular bone formation with high doses of estrogen (36–38). This large study using specific and sensitive indexes of bone turnover suggests for the first time that conventional doses of E₂ may induce a transient direct stimulation of bone formation in humans when given by the intranasal route. Whether this effect is specific for intranasal E₂ or general to the nonoral route of administration, such as transdermal patches, remains to be investigated, as most previous randomized placebo-controlled studies looking at the effects of transdermal E₂ have not assessed bone formation at 1 month.

To explain the different short term effects of oral and intranasal E₂ on bone formation markers, we measured circulating IGF-I. More than 75% of plasma IGF-I circulates in a 150-kDa ternary complex with IGFBP-3 and an acid-labile subunit, whereas the remaining peptides are found in a 50-kDa complex with IGFBP-1 to -6 (39). Thus, to evaluate the effects of estrogens on bioavailable IGF-I, alterations in the concentrations of IGFBPs need to be taken into consideration. In that study, we found no significant change from placebo in serum IGFBP-3 levels over the 3-month period in women treated with either oral or transdermal E₂ in agreement with previous smaller nonplacebo-controlled studies (40–42). Thus, the contrasting effects of oral and intranasal E₂ on serum IGF-I levels, i.e. a marked decrease for the oral formulation and a nonsignificant change for intranasal administration, are likely to reflect differences in bioavailable serum IGF-I. However, we cannot exclude different effects of oral and intranasal E₂ on the other IGFBPs. As E₂ has been shown to inhibit hepatic IGF-I mRNA expression (43), the decrease in serum IGF-I during oral E₂ treatment probably results from exposure of the liver, the major source of circulating IGF-I, to the high estrogen concentration that follows intestinal absorption. The link between circulating IGF-I levels and bone turnover, however, is unclear. Indeed, IGF-I in bone originates not only from the circulation, via trapping by hydroxyapatite (44), but also from the osteoblasts and the bone marrow stromal cells. Thus, whether serum concentration of IGF-I reflects skeletal tissue levels remains an open question. It has been shown, however, that the proportions of IGF-II/IGF-I in human skeletal tissue are nearly identical to the serum ratio (45) and that serum IGF-I correlated with spine, femoral neck, and radius BMD in women and men (46, 47), suggesting that circulating levels may be an adequate index of skeletal status. Whatever the major sources of IGF-I in bone, it is well recognized that this growth factor is a potent mitogen for osteoblastic cells in vitro (48, 49), stimulating alkaline phosphatase activity and increasing osteocalcin in rat calvaria (50) and human bone cells (51). Interestingly, E₂ stimulates the expression of IGF-I mRNA in several rat osteoblastic cell lines, and its bone-forming effects are blocked by anti-IGF-I antibodies, suggesting that, at least in part, the stimulatory effects of E₂ on osteoblastic activity in vitro could be mediated by IGF-I (52). We found that the changes in circulating IGF-I correlated with the changes in serum osteocalcin, serum PINP, and serum BAP, suggesting that the stimulatory effect of estrogen on bone formation may be mediated by IGF-I in vivo, although other mechanisms are likely to be involved. Taking into account these data, we can speculate that the increase in bone formation markers at 1 month with intranasal administration is likely to reflect the direct effect of E₂ in stimulating osteoblastic activity. At 3 months, this effect could be balanced by the decrease in the overall bone turnover rate, resulting in an apparent lack of change in bone formation.

After 1 yr of treatment with intranasal E₂ (300 μg), both bone resorption and bone formation markers decreased and returned to the premenopausal range, i.e. the levels in estrogen-repleted women, as previously reported after 1-yr treatment with 0.625 mg conjugated equine estrogen (30, 17), oral E₂ (2 mg) (6), and transdermal E₂ (50 μg) (17). These data indicate that although the decrease in bone formation is delayed for intranasal compared to oral E₂, their long term effects on the overall rate of bone turnover are likely to be similar. However, as the main increase in BMD with antiresorptive therapy occurs during the first year, a sustained bone formation with a decreased bone resorption in the initial phase of treatment may result in a higher bone gain, an interesting hypothesis that needs further investigation.

In conclusion, this large study shows that intranasal administration of E₂ normalizes bone turnover to premenopausal levels after 1 yr of treatment. Intranasal E₂ does not alter serum IGF-I levels, contrasting with the dramatic decrease in this growth factor induced by oral administration. This results in a sustained bone formation during the first 3 months of treatment with intranasal, but not oral, E₂ despite a decrease in bone resorption, which was similar for both routes of administration. Studies with BMD measurement are required to assess the long term effect of intranasal E₂ on bone mass.

Acknowledgements

The authors thank the “other members of the Scientific Committee:” Pr. A. Basdevant, Pr. J. Bringer, Pr. C. Christiansen, Pr. J. Gehanno, Pr. A. Gompel, Pr. P. Kenemans, Dr. J. Studd, Pr. C. Lauritzen, Pr. P. Lopez and the Aerodiol Study Group: Dr. J. M. Allouch, Dr. E. Amadio, Dr. K. Ardaens, Dr. N. Bessieres, Dr. J. P. Blanchere, Dr. J. M. Bazin-Borloo, Dr. A. Bloch, Prof. P. Buytaert, Dr. G. Chaix-Durand, Dr. O. Chevallier, Dr. D. Coliche, Dr. D. Collin, Dr. J. Coudray, Dr. D. Dalbos, Dr. M. N. de Granvilliers, Prof. P. Dellenbach, Prof. J. Donnez, Dr. R. Druckman, Dr. M. Fari, Dr. G. Favre-Eloff, Prof. P. Fenichel, Dr. J. L. Fieffe, Prof J. M. Foidart, Dr. F. Girard-Gueneron, Dr. A. M. Grange, Dr. A. Grimard, Dr. N. Gros, Dr. C. Grosjean, Dr. M. C. Guiral, Dr. S. Hughes, Dr. M. Kermanach, Dr. M. Klein, Prof. J. Laszlo, Dr. P. Lecocq, Dr. M. Levrier, Dr. J. Levy-Frebault, Dr. B. Martel, Dr. H. Martens, Dr. E. Maschy-Remy, Dr. M. Maupin, Dr. F. Mezzana, Dr. G. Moniot, Dr. F. Mousteou, Dr. N. N’Guyen, Dr. A. M. Neu, Dr. A. Odier, Dr. E. Paganelli, Dr. Z. Papp, Dr. R. Payan, Dr. D. Pierre, Dr. B. Pornel, Dr. F. Ragazzoni, Dr. M. Renaud, Dr. B. Rime, Prof. K. Rothe, Dr. J. C. Sage, Dr. A. Samier, Dr. J. Sentenac, Dr. I. Smith, Prof. K. Toth, Dr. M. T. Verdys, Dr. B. Vincent-Larrieu, Dr. D. Weil Masson, and Prof. L. Wildt.