HIV-infected children exhibit low bone mass measurements compared with healthy children [1–9]. Although the cohorts of patients studied so far differ in antiretroviral treatment regimens, immunological response, and adherence to treatment, the vast majority of data indicate that young patients have low bone mineral density or altered bone geometry measurements. The alterations in bone mass seem to be the result of defective bone metabolism. Bone cells are responsible for the changes in shape and dimensions occurring during growth, and for the maintenance of bone tissue during the entire lifespan . The coordinated activity of osteoblasts (bone-forming cells) and osteoclasts (bone-resorbing cells) is necessary to provide the correct amount of bone tissue in every step of growth and adult life. The activity of bone cells can be monitored by measuring the bone metabolism rate, which can be accurately studied by histomorphometry of the iliac crest [11,12]; however, bone biopsy is an invasive procedure that is not feasible for standard use in the evaluation of bone metabolism. An accepted surrogate is represented by the measurements in the serum and urine of specific markers of bone formation and bone resorption . Recent data indicate that HIV-infected children and adolescents have increased concentrations of bone formation markers, coupled with increased levels of bone resorption indices [2,5,14,15]. A high bone metabolism rate leads to a net decrease in bone formation, because the process of apposition lasts longer than that of resorption.
The recruitment of osteoclasts is made possible by the fine tuning of a specific mechanism [16,17]; osteoclast progenitors, present in the bone marrow, express on their membrane a specific receptor (receptor activator of nuclear factor kappa B, or RANK). Stromal cells and osteoblasts, after mechanical or humoral stimuli, produce a cytokine (RANK ligand; RANKL), which by binding on RANK induces fusion and the differentiation, and promotes the activity of mature osteoclasts . Stromal cells and osteoblasts also produce a soluble decoy receptor, named osteoprotegerin that binds to RANKL, regulating its activity. The production of RANKL and osteoprotegerin is stimulated by several different cytokines and hormones acting in concert to control osteoclasts . Circulating concentrations of osteoprotegerin and RANKL can be accurately measured, and may play a role in monitoring metabolic bone diseases .
The aim of the current study was to quantify the serum concentrations of osteoprotegerin and RANKL in a cohort of HIV-infected children on long-term HAART, and to compare the results with those obtained in healthy children. The changes in bone metabolism indices, osteoprotegerin, and RANKL have been studied before and after replacing a protease inhibitor (PI) with efavirenz and stavudine with tenofovir.
Materials and methods
Eligible for the study were vertically HIV-infected children and adolescents on long-term and successful HAART. We excluded patients treated with antiretroviral regimens other than HAART, and those of non-Caucasian ancestry. Twenty-seven patients (age range 4.99–17.3 years, 14 girls and 13 boys) receiving antiretroviral therapy containing stavudine plus lamivudine plus one PI with a long-lasting (at least 48 weeks) undetectable plasma HIV-RNA level were enrolled in the study. Their mean CD4 cell count was 841 (59) cells/μl and the mean CD4 cell percentage was 34 (1)%. The characteristics of the patients are shown in Table 1. All patients were maintained on lamivudine (4 mg/kg twice a day) throughout the duration of the study. During follow-up the PI was replaced with efavirenz and stavudine with tenofovir. The switch of antiretroviral therapy was carried out to reduce toxicity and to maintain adherence. Efavirenz was administered once a day, at weight-dependent doses, as recommended by the manufacturer. Tenofovir was administrated once a day at a body surface area-dependent dose: 150 mg for 0.5–0.84 m2, 225 mg for 0.85–1.29 m2 and 300 mg for 1.3 m2 and greater.
As a control group we studied 336 healthy children and adolescents of Caucasian origin (162 girls and 173 boys). None had a history of endocrine, nutritional, growth or renal problems. Their ages ranged from 4.8 to 17.9 years, weight ranged from 15.6 to 100 kg, height from 107 to 188 cm, and body mass index (BMI) from 10.9 to 29.2 kg/m2.
Informed consent was obtained from each patient's and healthy child's legal guardian and from the patients and healthy children when appropriate, before enrollment. This study was approved by the ethical committee of the L. Sacco Hospital.
Measurements of biochemical markers of bone metabolism were obtained before and approximately 6 months (6 ± 0.5 months) after switching antiretroviral regimen. Bone mineral measurements were performed at baseline. Control subjects were studied only once for ethical reasons.
Clinical and anthropometric assessment
All subjects enrolled in the study underwent physical examination to obtain anthropometric measures. Body weight was measured to the nearest 0.1 kg on a balance beam scale (Seca, Hamburg, Germany), and height was measured to the nearest millimeter using a wall-mounted stadiometer (Holtain Ltd., Crosswell, UK). BMI was then calculated as weight over height2 (kg/m2). Standard deviation scores of anthropometric measurements were calculated using specific Italian standards . The pubertal stage of HIV-infected and healthy children was defined according to the criteria of Tanner and Whitehouse .
Blood was drawn in the morning after an overnight fast. Blood was allowed to clot immediately after venipuncture; serum was separated by centrifugation and was stored at −80°C until analysis. Urine specimens were collected between 10:00 and 12:00 hours as the second voiding of the day, to minimize the effect of the circadian rhythm of excretion of collagen degradation products . Samples were aliquoted immediately and stored at −80°C until analysis.
Bone-specific alkaline phosphatase (BALP) was measured in serum as a bone formation marker, using a commercial immunoassay (Metra BAP enzyme immunoassay kit; Quidel Corp., San Diego, California, USA). Intraassay reproducibility was less than 4%, and interassay variation was less than 7%. Sensitivity was 0.7 U/l.
We measured the urine concentration of the N-terminal telopeptide of type I collagen (NTx) as a bone resorption index. NTx was measured using an enzyme-immunosorbent assay (Osteomark NTx Urine; Wampole Laboratories, Princeton, New Jersey, USA). Assay values were standardized to an equivalent amount of bone collagen, and were expressed in nanomoles bone collagen equivalent (BCE) per litre. The sample results from a single urine collection were normalized for urine dilution by urine creatinine analysis, and were reported as nmol BCE/mmol creatinine. In our laboratory the intra-assay variation was less than 10%. The interassay precision was less than 9%, and sensitivity was 20 nmol BCE/l. Urine creatinine was measured by a standard automated method.
The serum concentration of RANKL was determined by an enzyme immunoassay (sRANKL; Biomedica Medizinprodukte GmbH, Vienna, Austria). In our laboratory the intra-assay variation was less than 7%. The interassay precision was less than 9%, and sensitivity was 0.08 pmol/l.
Osteoprotegerin was measured in serum samples using a commercially available enzyme immunoassay (human osteoprotegerin enzyme-linked immunosorbent assay; BioVendor Laboratory Medicine, Inc., Brno, Czech Republic). In our laboratory the intraassay variation was less than 10%. The interassay precision was less than 7%, and sensitivity was 0.4 pmol/l.
Bone mineral measurements
Bone mineral density (BMD) was measured at the L2–L4 vertebrae level and in the whole skeleton. The data were analysed using proper pediatric software (version 1.5h). BMD measurements were made with a dual-energy X-ray absorptiometer (DPX-L, GE-Lunar Radiation Corp., Madison, Wisconsin, USA). The instrument was calibrated on a daily basis according to the manufacturer's instructions. Reproducibility was calculated as a coefficient of variation obtained by weekly measurements of a standard phantom on the instrument, and by repeated measurements obtained in children of different ages. The coefficient of variation of our instrument was 0.6 % with the standard phantom; in vivo we calculated a coefficient of variation of 1.4% for the lumbar spine and 1.5% for the whole skeleton.
Descriptive statistics were calculated for all the variables, and data are expressed as the mean (SE). All statistical analyses were conducted at the α = 0.05 level, and were two-tailed. The distribution of the variables was checked using the Shapiro–Wilk W test. The statistical software JMP IN (SAS Institute, Inc., Cary, North Carolina, USA) was used for the analyses.
Multivariate analyses were performed to evaluate the differences between HIV-infected patients and control subjects, after controlling for confounding variables. Bone metabolism indices and bone mineral measurements were the dependent variables, whereas sex, age, and anthropometric measurements were the confounding variables, and the presence of HIV infection was the independent dichotomous variable. All anthropometric measurements were initially included, and the backward procedure was used to build the best model.
The unpaired Student's t-test was employed to explore the differences between the RANKL/osteoprotegerin ratios observed in the two groups of subjects. The paired t-test was used to compare the anthropometric and biochemical variables measured at baseline and after 6 months.
The influence of the type of PI on the concentration of osteoimmune factors was explored by one-way analysis of variance.
Six months after the switch all patients maintained a good control of viral replication, and their mean CD4 cell counts and percentages were 776 cells/μl (55) and 36% (1), respectively. As expected, height and weight increased significantly during the 6 months of follow-up of HIV-infected patients (P < 0.0001). Interestingly, we observed a significant increase in weight z-scores (t = 4.8; P < 0.0001) and height z-scores (t = 3.7; P < 0.0009), but not of BMI (t = 1.4; P = 0.17). None of the HIV-infected patients showed signs of delayed puberty.
The bone mineral measurements of HIV-infected patients were 0.856 (0.040), and 0.978 (0.027) g/cm2 for the lumbar spine and whole body, respectively. Healthy controls had significantly (P = 0.049) higher BMD values at the lumbar spine [0.893 (0.018) g/cm2] compared with HIV-infected patients, after correcting for the effect of confounding variables. Similarly, the whole-body BMD measurements of healthy subjects [0.988 (0.012) g/cm2] were significantly higher than those of patients (P = 0.040).
The biochemical measurements of HIV-infected patients, obtained at baseline when they were receiving lamivudine plus stavudine plus one PI, and 6 months after they replaced stavudine with tenofovir and the PI with efavirenz, are shown in Table 2.
The mean BALP serum concentration of healthy children and adolescents was 104.4 (3.9) U/l. HIV-infected patients had a significantly higher concentration (P < 0.0001) of BALP at baseline compared with healthy subjects, after correcting for the confounding effect of age, sex, and anthropometric measurements. After 6 months the BALP values of patients were no different from baseline (t = 1.4; P = 0.15), and still significantly higher compared with healthy controls (P < 0.0001).
The bone resorption rate was markedly higher at baseline in HIV-infected patients than in healthy subjects [253.5 (11.5) nmol BCE/mmol creatinine]. The difference between the two groups after correction for differences in age, sex, and height was highly significant (P = 0.0001). Although the NTx values decreased after 6 months (t = 1.9; P = 0.05), they remained significantly different compared with healthy controls (P = 0.0081). The urine creatinine concentration did not change significantly during the 6 months of follow-up. Therefore, the urinary decrements of NTx concentration were not the result of increased urinary creatinine.
The mean osteoprotegerin serum concentration of healthy subjects was 5.3 (0.1) U/l, and the mean concentration of RANKL was 0.39 (0.05) pmol/l. The osteoprotegerin measurements of HIV-infected patients were significantly higher at baseline compared with those of healthy children and adolescents (P < 0.0001). Similarly, the RANKL serum concentration of patients was significantly higher than that of the control group (P < 0.0001). After 6 months of the new antiretroviral regimen the serum concentrations of both compounds were significantly lower compared with baseline (osteoprotegerin t = 2.7; P = 0.01; RANKL t = 3.3; P = 0.0026). Nevertheless, the osteoprotegerin concentration of the patients was still significantly higher than that of the controls (P < 0.0001). On the contrary, the RANKL concentration was no longer significantly different (P = 0.07).
The ratio between RANKL and osteoprotegerin concentrations indicates which of the two factors is prevailing, thus giving some information about osteoclast recruitment and function. The RANKL/osteoprotegerin ratio calculated in healthy children and adolescents was 0.078 (0.01). At baseline, the mean RANKL/osteoprotegerin ratio of HIV-infected patients was significantly higher (Fig. 1) than that of healthy controls (t = 2.4; P = 0.02). After 6 months of the new antiretroviral regimen, the RANKL/osteoprotegerin ratio of the patients did not longer differ from that of the control group (t = 1.1; P = 0.27; Fig. 1).
The evidence for an altered bone metabolism in HIV-infected children and adolescents, and particularly the marked increase in the bone resorption rate, prompted us to investigate the role of osteoimmune factors implicated in the recruitment, differentiation, activation, and survival of osteoclasts. RANKL is a member of the tumour necrosis factor superfamily, and is expressed by osteoblasts and their mature precursors . Osteoprotegerin is a glycoprotein belonging to the tumour necrosis factor receptor family that acts as an antagonist of RANKL, modulating its activity [16,18]. We found markedly elevated serum concentrations of RANKL and osteoprotegerin, but more importantly, we observed an increased RANKL/osteoprotegerin ratio in HIV-infected youths compared with healthy subjects. The imbalance between the two factors indicates a prevalence of RANKL, which in turn may be responsible for an increased osteoclast activation, leading to a decreased bone mineral mass. Alterations of the RANKL/osteoprotegerin ratio are critical in the pathogenesis of bone diseases that result from increased bone resorption activity . The results of previous studies conducted in HIV-infected adult patients are inconsistent [24,25]. HIV-seropositive men on HAART exhibited elevated RANKL serum concentrations compared with non-HAART-treated and untreated patients . HAART-treated adult patients were grouped according to the presence/absence of osteopenia or osteoporosis in another study . Although no differences in RANKL serum concentrations and the RANKL/osteoprotegerin ratio were observed between the two groups, increased osteoprotegerin was found in the osteopenic, osteoporotic group. The elevated osteoprotegerin concentratrion was interpreted as a possible reaction to protect against the high bone resorption rate. The main limitation of both studies is the lack of an appropriate healthy control population. In our study we compared the biochemical results of HIV-infected children and adolescents with those obtained in a large and well-selected cohort of healthy children and adolescents. The concentration of biochemical markers of bone metabolism are not constant, but change as a function of sex, age, puberty, and growth parameters , and all the comparisons have thus been performed correcting for the confounding effect of such variables.
Serum measurements of RANKL and osteoprotegerin are not bone specific, because other tissues can produces these factors. In particular, RANKL/RANK signaling enhances dendritic cell survival, and is indispensable for lymph node organogenesis, whereas activated T cells produce soluble molecules of RANKL [26,27]. To date it is not possible to separate the molecules produced by bone cells from those of other origins. Nevertheless, the action of RANKL and osteoprotegerin on bone is the same regardless of the site of their production. Our patients were in excellent control of viral replication and immune response, which were maintained even after switching antiretroviral regimen. The contribution of the immune system, mainly immune activation, on the production of RANKL and osteoprotegerin could thus be considered constant or even absent.
There is in-vitro evidence that some PI may interfere with the production of RANKL [28,29]. In particular, ritonavir seems to block RANKL activation, and therefore has a positive influence on bone metabolism . We categorized our patients according the PI they were receiving: we observed lower RANKL serum concentration and RANKL/osteoprotegerin ratio in patients receiving ritonavir, but the difference with the other patients receiving different PI was not significant. The small number of subjects in each group certainly limited the power of the statistical test.
The efficacy of a PI-based HAART regimen in HIV-infected children has been proved. Nevertheless, PI-based therapies usually involve many pills and dosing schedule constraints; factors known to decrease adherence to therapy and to interfere negatively with patients’ quality of life. The availability of new drugs that have similar antiretroviral efficacy, but have a simpler dosing schedule, prompted us to replace PI with efavirenz and stavudine with tenofovir. We have already reported on the safety of the new antiretroviral regimen , but some controversies persist on the effect of tenofovir on bone density in HIV-infected youths [7,9]. A reduction in BMD was reported after 24 and 48 weeks of treatment in one study , but no changes in the accrual of bone mineral content over 12 months treatment in the other . Nevertheless, little is known about the effect of tenofovir-containing treatment on bone metabolism in children. The present study shows that HIV-infected children on long-term HAART have increased circulating markers of bone metabolism compared with healthy children and adolescents. Replacing a PI with efavirenz and stavudine with tenofovir did not induce significant changes in the bone formation rate, but reduced slightly but significantly the urine concentration of the bone resorption index NTx. Bone metabolism alterations in HAART-treated HIV-infected children have already been reported by us and others [2,5,14,15]. Young patients receiving a PI showed consistently higher values of bone formation markers (osteocalcin, BALP) compared with patients not receiving a PI-containing regimen  and healthy controls [2,5,14]. A higher bone resorption rate was found in HIV-infected children on PI-based HAART [2,5,15]. The reduction in the bone resorption rate observed in our study after switching antiretroviral treatment is interesting and promising. Even though growing experimental evidence suggests that antiretroviral therapy (expecially PI-containing regimens) affect bone metabolism, the exact nature and mechanism of decreased bone density need to be defined further. Moreover, as the PI and stavudine were replaced with efavirenz and tenofovir at the same time, it is not possible to determine which of the switches was driving the change behind the improvement in bone resorption. Finally, our results need to be confirmed by a larger number of observations, and by a longer longitudinal follow-up period.
In conclusion, our study confirms the presence of bone metabolism derangement in HAART-treated HIV-infected children and adolescents. A marked alteration in the RANKL/osteoprotegerin system is also detectable in patients receiving PI-based HAART. Our short-term data indicate that switching to an antiretroviral regimen containing tenofovir and efavirenz, and excluding PI and stavudine restores the RANKL/osteoprotegerin equilibrium, and thus may lead to a reduction in the bone resorption rate.
The authors acknowledge the precious technical skills of Maria Puzzovio.
Sponsorship: This study was partly supported by grant no 30G.31 from Istituto Superiore di Sanità, VI Programma Nazionale di Ricerca sull’AIDS, 2006.
1. O’Brien KO, Razavi M, Henderson RA, Caballero B, Ellis KJ. Bone mineral content in girls perinatally infected with HIV. Am J Clin Nutr 2001; 73:821–826.
2. Mora S, Sala N, Bricalli D, Zuin G, Chiumello G, Viganò A. Bone mineral loss through increased bone turnover in HIV-infected children treated with highly active antiretroviral therapy. AIDS 2001; 15:1823–1829.
3. Arpadi SM, Horlick M, Thornton J, Cuff PA, Wang J, Kotler DP. Bone mineral content is lower in prepubertal HIV-infected children. J Acquir Immune Defic Syndr 2002; 29:450–454.
4. Arpadi S, Horlick M, Shane E. Metabolic bone disease in human immunodeficiency virus-infected children. J Clin Endocrinol Metab 2004; 89:21–23.
5. Mora S, Zamproni I, Beccio S, Bianchi R, Giacomet V, Viganò A. Longitudinal changes of bone mineral density and metabolism in antiretroviral-treated human immunodeficiency virus-infected children. J Clin Endocrinol Metab 2004; 89:24–28.
6. Jacobson DL, Spiegelman D, Duggan C, Weinberg GA, Bechard L, Furuta L, et al
. Predictors of bone mineral density in human immunodeficiency virus-1 infected children. J Pediatr Gastroenterol Nutr 2005; 41:339–346.
7. Giacomet V, Mora S, Martelli L, Merlo M, Sciannamblo M, Viganò A. A 12-month treatment with tenofovir does not impair bone mineral accrual in HIV-infected children. J Acquir Immune Defic Syndr 2005; 40:448–450.
8. Rosso R, Vignolo M, Parodi A, Di Biagio A, Sormani MP, Bassetti M, et al
. Bone quality in perinatally HIV-infected children: role of age, sex, growth, HIV infection, and antiretroviral therapy. AIDS Res Hum Retroviruses 2005; 21:927–932.
9. Gafni RI, Hazra R, Reynolds JC, Maldarelli F, Tullio AN, DeCarlo E, et al
. Tenofovir disoproxil fumarate and an optimized background regimen of antiretroviral agents as salvage therapy: impact on bone mineral density in HIV-infected children. Pediatrics 2006; 118:e711–e718.
10. Aubin JE, Heersche JNM. Bone cell biology: osteoblasts, osteocytes, and osteoclasts. In: Glorieux FH, Pettifor JM, Jüppner H, editors. Pediatric bone. Biology and diseases. San Diego: Academic Press; 2003. pp. 43–75.
11. Parfitt AM, Travers R, Rauch F, Glorieux FH. Structural and cellular changes during bone growth in healthy children. Bone 2000; 27:487–494.
12. Glorieux FH, Travers R, Taylor A, Bowen JR, Rauch F, Norman M, et al
. Normative data for iliac bone histomorphometry in growing children. Bone 2000; 26:103–109.
13. Schönau E, Rauch F. Biochemical markers of bone metabolism. In: Glorieux FH, Pettifor JM, Jüppner H, editors. Pediatric bone. Biology and diseases. San Diego: Academic Press; 2003. pp. 339–357.
14. Tan MB, Nelson RP, James-Yarish M, Emmanuel PJ, Schurman SJ. Bone metabolism in children with human immunodeficiency virus infection receiving highly active anti-retroviral therapy including a protease inhibitor. J Pediatr 2001; 139:447–451.
15. Zamboni G, Antoniazzi F, Bertoldo F, Lauriola S, Antozzi L, Tatò L. Altered bone metabolism in children infected with human immunodeficiency virus. Acta Paediatr 2003; 92:12–16.
16. Simonet WS, Lacey DL, Dunstan CR, Kelley M, Chang MS, Luthy R, et al
. Osteoprotegerin: a novel secreted protein involved in the regulation of bone density. Cell 1997; 89:309–319.
17. Hsu H, Lacey DL, Dunstan CR, Solovyev I, Colombero A, Timms E, et al
. Tumor necrosis factor receptor family member RANK mediates osteoclast differentiation and activation induced by osteoprotegerin ligand. Proc Natl Acad Sci U S A 1999; 96:3540–3545.
18. Lacey DL, Timms E, Tan H-L, Kelly MJ, Dunstan CR, Burgess T, et al
. Osteoprotegerin (OPG) ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 1998; 93:165–176.
19. Teitelbaum SL. Bone resorption by osteoclasts. Science 2000; 289:1504–1508.
20. Rogers, Eastell R. Circulating osteoprotegerin and receptor activator for nuclear factor kappa B ligand: clinical utility in metabolic bone disease assessment
. J Clin Endocrinol Metab
21. Cacciari E, Milani S, Balsamo A, Spada E, Bona G, Cavallo L, et al
. Italian cross-sectional growth charts for height, weight and BMI (2 to 20 yr). J Endocrinol Invest 2006; 29:581–593.
22. Tanner JM, Whitehouse RH. Clinical longitudinal standards for height, weight, height velocity and weight velocity and stages of puberty. Arch Dis Child 1976; 51:170–179.
23. Mora S, Prinster C, Proverbio MC, Bellini A, de Poli SC, Weber G, et al
. Urinary markers of bone turnover in healthy children and adolescents: age-related changes and effect of puberty. Calcif Tissue Int 1998; 63:369–374.
24. Seminari E, Castagna A, Soldarini A, Galli L, Fusetti G, Dorigatti F, et al
. Osteoprotegerin and bone turnover markers in heavily pretreated HIV-infected patients. HIV Med 2005; 6:145–150.
25. Konishi M, Takahashi K, Yoshimoto E, Uno K, Kasahara K, Mikasa K. Association between osteopenia/osteoporosis and the serum RANKL in HIV-infected patients. AIDS 2005; 19:1240–1241.
26. Andreson DM, Maraskovsky E, Billingsley WL, Dougall WC, Tometsko M, Roux ER, et al
. A homologue of the TNF receptor and its ligand enhance T cell growth and dendritic cell function. Nature 1997; 390:175–179.
27. Dougall WC, Glaccum M, Charrier K, Rohrbach K, Brasel K, De Smedt T, et al
. RANK is essential for osteoclast and lymph node development. Genes Dev 1999; 18:2412–2424.
28. Fakruddin JM, Laurence J. HIV envelope gp120-mediated regulation of osteoclastogenesis via receptor activator of nuclear faktor κB ligand (RANKL) secretion and its modulation by certain HIV protease inhibitors through interferon-γ/RANKL cross-talk. J Biol Chem 2003; 278:48251–48258.
29. Wang MW-H, Wei S, Faccio R, Takeshita S, Tebas P, Powderly WG, et al
. The HIV protease inhibitor ritonavir blocks osteoclastogenesis and function by impairing RANKL-induced signaling. J Clin Invest 2004; 114:206–213.
30. Viganò A, Aldrovandi GM, Giacomet V, Merlo M, Martelli L, Beretta S, et al
. Improvement in dyslipidaemia after switching stavudine to tenofovir and replacing protease inhibitors with efavirenz in HIV-infected children. Antiviral Ther 2005; 10:917–924.