- Top of page
- Materials and Methods
The pathogenesis of osteoporosis (OP) is complex. Lifestyle, environmental, nutritional, and genetic factors[1-4] impact the skeleton independently and synergistically throughout life. Heredity seems to account for 60% to 80% of peak bone mineral density (BMD), which together with bone structure and quality are the key components of fracture resistance.
Over the past 15 years especially, numerous local and systemic regulators of bone formation and resorption have been identified, considerably improving our molecular and biochemical understanding of skeletal remodeling and bone loss. Insight from the genetics of OP and high bone mass disorders has stimulated development of novel, targeted treatments for OP, perhaps enabling early intervention or prevention. In the special circumstance of premature OP, genetic evaluation could help prognostication and guide treatment.
In health, pregnancy is associated with increased bone turnover (as evidenced by bone histomorphometry and biochemical markers). Increased 1,25-dihydroxyvitamin D synthesis in the kidney leads to enhanced intestinal calcium absorption to compensate for the calcium transfer from mother to fetus.[8, 9] Urinary calcium is increased because of an increased filtered load of calcium and increased glomerular filtration rate. During a healthy pregnancy, a reversible loss of 1% to 5% of maternal BMD has been reported.[7, 10] Nevertheless, studies are few and confounded by gestational changes in maternal body composition and weight. Rarely, pregnancy-associated osteoporosis (PAO), an idiopathic condition, develops in late pregnancy or early postpartum, most commonly with the first pregnancy. Most patients with PAO (>60%) present with a vertebral compression fracture (VCF) and have low BMD. The pathogenesis of PAO may combine preexisting low BMD (possibly related to genetic factors) and increased bone turnover, leading to accumulation of immature bone. Fortunately, most women with PAO show improvement of postpartum symptoms within several weeks and spontaneous improvement of BMD.
OPPG and LRP5
Osteoporosis-pseudoglioma syndrome (OPPG) (OMIM #259770) is the rare autosomal recessive (AR) disorder that features infantile blindness and OP with childhood fractures owing to loss-of-function mutations in the gene that encodes low-density lipoprotein receptor-related protein 5 (LRP5).[11, 12] OPPG is characterized by severely decreased bone formation together with disruption of ocular structures (phthisis bulbi) and/or persistent hyperplasia of primary vitreous leading to blindness. Autosomal dominant (AD) transmission of loss-of-function LRP5 mutations (with variable penetrance) can be associated with low BMD[12-16] and/or familial exudative vitreoretinopathy (FEVR, OMIM #601813). Additionally, genome-wide association studies have found linkage between OP and LRP5.[17, 18] In contrast, heterozygous gain-of-function mutations in LRP5 cause high bone mass (OMIM #607634), which can be complicated by cranial nerve compression. These low- and high-bone mass Mendelian disorders embody the importance of LRP5 in bone formation and remodeling.
Methylenetetrahydrofolate reductase (MTHFR) catalyzes the conversion of 5,10-methylenetetrahydrofolate to 5-methyltetrahydrofolate, a cosubstrate for homocysteine remethylation to methionine. Homocystinuria from AR deactivation of MTHFR (OMIM #236250) is a rare disorder characterized by high endogenous homocysteine levels, central nervous system and vascular disease, and early-onset OP. In fact, prospective, population-based studies have documented increased plasma homocysteine levels to be a strong, independent risk factor for OP fractures.[21-23] Homozygosity for a specific MTHFR polymorphism (C677T) is associated with elevated plasma and urine homocysteine levels and also an increased risk of several disorders including fetal loss, thrombosis, fractures,[24-29] and/or decreased BMD.[29-31] The prevalence of homozygous C677T (“TT genotype”) ranges from 2.7% to 32.2% in various populations worldwide.
We report a now-37-year-old woman who manifested PAO at age 25 years and is heterozygous for a unique loss-of-function mutation in LRP5 and homozygous for the C677T MTHFR polymorphism, and our investigation of her family.
- Top of page
- Materials and Methods
Physical examination of the family (Fig. 1) revealed white individuals of normal stature (heights: proposita 63.5 inches, mother 64 inches, father 64 inches, and brother 69 inches) with no dysmorphic features, and was remarkable only for hypermobility in the proposita (II.2), her two children (III.1 and III.2), and her mother (I.4), but not in her father.
Low bone turnover in the proposita was suggested by a urine N-telopeptide (NTX) level that was subnormal at 8 BCE/mmol creatinine (11 to 48 normal), but without suppressed serum osteocalcin or alkaline phosphatase levels. She had mild vitamin D insufficiency and a mildly elevated serum HDL cholesterol level (Table 1). No other aberrations of lipids, glycosylated hemoglobin, or homocysteine were detected. Biochemical studies and markers of bone turnover were normal in all family members studied, except for mildly decreased serum osteocalcin of 2.3 ng/mL (3.1 to 13.7) in her father (I.3).
Table 1. Proposita: Biochemical Findings
|Calcium||9.7, 9.0 mg/dL||8.5–10.1|
|Osteocalcin||15.1, 5.0 ng/mL||5.8–41.0|
|Intact PTH||59 pg/mL||10–69|
|25 hydroxy vitamin D||23, 29 ng/mL||Sufficient >30|
|Protein electrophoresis||Normal|| |
|HDL cholesterol||61 mg/dL||40–59|
|LDL cholesterol||93 mg/dL||0–99|
|24-hour urine-free cortisol||30 mcg||2.0–42|
|24-hour urine N-telopeptides||8 BCE/mmol creatinine||11–48|
|24-hour urine calcium||223, 139 mg||100–300|
|Protein electrophoresis||Normal|| |
DXA results are summarized in Table 2 and Fig. 2. During the 4 years before her second successful pregnancy, while receiving calcium and vitamin D alone, the proposita's BMD remained low but stable in the spine and total hip, and improved in the femoral neck (Fig. 2). A repeat DXA study 3 months after her second successful pregnancy showed decreases in the spine and femoral neck BMDs (5% and 2.4%, respectively) but a 7.2% increase in total hip BMD. Subsequent DXA studies, two and four years later, showed an 18% increase at the spine but no significant change at the hip (Fig. 2) or the wrist (data not shown). Of note, the proposita's BMD was particularly low in areas rich in trabecular bone (ultradistal radius versus 33% radius [Table 2] and spine versus hip). BMD for spine and hip were low in her father and brother but normal in her mother.
Table 2. Family Demographic Findings and DXA Bone Mineral Density Z-Scores
|Person||Age (years)||Height||AP Spine (L1 to L4)||Femoral neck||33% radius||Ultradistal radius|
The proposita (II.2), her mother (I.4), and her two children (III.1 and III.2) all showed a unique, heterozygous, in-frame deletion mutation (c.4454_4465del, p.1485_1488del SSSS) in exon 21 of LRP5 (Fig. 1). This 12-bp in-frame deletion would remove four serines from a seven-serine stretch in the cytoplasmic domain of LRP5. This mutation was not found in the proposita's father or the brother who was studied.
The proposita and her mother were both homozygous for the C677T MTHFR polymorphism (Fig. 1). Her brother, father, and two children were heterozygous for the C677T MTHFR polymorphism (Fig. 1).
In addition, the proposita's maternal cousin (not shown in the pedigree) with significant, progressive postmenopausal OP (hip and spine T-scores –4.3), numerous atraumatic fractures, and short stature (<5 feet) had neither the LRP5 mutation nor the C677T MTHFR polymorphism. Despite an intensive evaluation, the cause of her osteoporosis remains unknown.
Genomic copy number
Affymetrix microarray-based genomic copy number analysis of LRP4, LRP5, and LRP6 in the proposita showed no gross genomic disruptions.
- Top of page
- Materials and Methods
PAO presents as fragility fractures in the third trimester or postpartum, generally in a primagravida.[34, 35] Insufficient dietary calcium and vitamin D levels to compensate for the calcium transfer from mother to fetus during pregnancy has been proposed as the pathogenesis of PAO. Yet, the rarity of this disorder (exact prevalence is unknown) and bone histomorphometry compatible with osteoblast (OB) failure35 suggest other contributory factors. Although low-normal BMD persists postpartum in these women,[36, 37] most do not have recurrent fractures and BMD improves for at least the next 2 to 4 years.[35, 36]
Although our patient's clinical presentation with postpartum fractures only after a first pregnancy and subsequent improvement in BMD with minimal pharmacologic intervention is consistent with PAO, her constellation of other medical problems (unilateral blindness from hyperplastic primary vitreous in infancy and two fetal losses, along with her family history of OP and blindness) prompted our molecular investigation for an LRP5 loss-of-function mutation and for the C677T MTHFR polymorphism. A heterozygous LRP5 mutation consistent with ophthalmologic and skeletal penetrance was found rather than two (homozygous or compound heterozygous) mutations, which cause OPPG. In fact, OPPG usually presents as severe disease with bilateral blindness in infancy followed by early childhood fragility fractures. Her 12-bp in-frame LRP5 deletion would remove four serines from a seven-serine stretch in the cytoplasmic domain of LRP5. The 207 aa cytoplasmic domain is both serine (15%) and proline (16%) rich, and represents a small portion of the large 1615 aa LRP5 protein. Because there are no more than two adjacent serines throughout the remainder of LRP5, this serine “stretch” appears to be a unique motif within the protein. However, we are unaware of any specific function for this seven-serine stretch.
LRP5 activates the WNT signaling pathway and thereby stimulates bone accrual important for acquisition of peak bone mass.[12, 38-40] The skeletal phenotype of OPPG features decreased diaphyseal diameter and gracile-appearing bones reminiscent of osteogenesis imperfecta as well as insufficient OB function with low BMD.[11, 12] Furthermore, single (heterozygous) LRP5-inactivating mutations[12, 13, 18, 41, 42] and some polymorphisms[14-17, 43] have been associated with idiopathic juvenile osteoporosis, low BMD, and/or retinal vascular disease, each with intrafamily variability.[12, 44] However, the impact of pregnancy or diet on BMD and fractures in individuals with heterozygous LRP5-inactivating mutations is not known. Although the bone histology of both OPPG and PAO[34, 36] are consistent with OB inactivity, family studies of OPPG have not reported PAO in carriers of loss-of-function LRP5 mutations.
Notably, OP did not cosegregate with the heterozygous loss-of-function LRP5 mutation in our kindred. The mutation instead paralleled hypermobility, a feature of multiple bone fragility syndromes, including OPPG. In fact, the variable penetrance observed in this kindred is consistent with that of the literature for familial exudative vitreoretinopathy and for OP associated with a single deactivating LRP5 mutation. The proposita's mother and children had normal vision. Her children had no fragility fractures, and her mother had a normal BMD and no history of fractures (even postpartum) despite all carrying the proposita's LRP5 mutation. No gross copy number abnormalities were found by microarray in LRP5 or in the structurally similar OB-expressed genes LRP4 and LRP6 to explain the exclusivity of the proposita's OP and visual difficulties.
Accordingly, we considered that homozygosity for the C677T MTHFR polymorphism contributed to the proposita's skeletal disease. Yet, her mother, who had normal BMD and no fracture history (and was heterozygous for the LRP5 mutation), was also homozygous for C677T MTHFR. The low BMD in her father and brother remain unexplained but may be related, in part, to their heterozygosity for the MTHFR polymorphism. Homozygosity for the MTHFR C677T polymorphism causes a 40% to 50% reduction of MTHFR function, and has been related to OP and fracture.[24-29] Two hypotheses explain this: interference of collagen cross linking from the elevated homocysteine levels and DNA hypomethylation affecting the structure and expression of other skeletally important genes. Folate, which may stabilize the C677T MTHFR thermolabile product, impacts DNA hypomethylation, decreases homocysteine levels, and increases bone density. Hence, both hypotheses are interesting for our patient. Perhaps noncompliance with folate supplementation (now recommended by the American Congress of Obstetricians and Gynecologists for all pregnancies) in the setting of the increased folate requirement of pregnancy leads to higher circulating homocysteine levels in C677T MTHFR homozygous women, explaining the development of PAO.
Inadequate peak bone mass before pregnancy has been proposed to increase the risk of PAO, but prepregnancy BMDs have generally not been reported for these patients. Furthermore, genetic predisposition to PAO is suggested by an increased family history of postmenopausal OP. Our PAO patient is homozygous for the C677T-MTHFR polymorphism (likely explaining her fetal losses) and had a forme fruste of OPPG, including low peak bone mass from a heterozygous deactivating LRP5 mutation. Surprisingly, her osteoporotic father, with a family history of blindness (which on closer questioning of her father was thought to be attributable to retinitis pigmentosa) and OP, did not share her MTHFR and LRP5 genotypes nor did her brother or her osteoporotic maternal cousin. Furthermore, neither her children nor her mother, who share her heterozygous LRP5 mutation, have ophthalmologic problems. Before clarifications concerning the family history and genetic analyses of family members, assumptions that her osteoporotic father and brother might carry her LRP5 mutation could have led to erroneous concerns for her brother's children. The discordance between genotype and phenotype suggests the impact of additional factors, such as pregnancy, vitamin intake, or other genes, when both the LRP5 mutation and C677T-MTHFR polymorphism are present. Although in 1995 Smith and colleagues demonstrated normal dermal fibroblast Type I collagen in 16 of 17 PAO patients (one previously known to have mild OI). Other investigations for genetic causes of PAO have not been undertaken.
In summary, our results are instructive for the following reasons: 1) They suggest that the clinical phenotype of heterozygous loss-of-function LRP5 mutations might include PAO. 2) They demonstrate that our proposita's putative LRP5-associated PAO occurred with variable penetrance (as is true for single LRP5 mutation-associated OP and retinopathy). 3) A unique LRP5 mutation was identified. 4) They suggest that PAO may occur in those with a genetically determined low peak bone density. Our results also serve as a reminder that risk assessment for associated conditions (eg, in this case, retinopathy in those with OP) should be based on the genetic study of family members, not just the index case.
- Top of page
- Materials and Methods
This study was supported by Shriners Hospitals for Children, the Clark and Mildred Cox Inherited Metabolic Bone Disease Research Fund, the Hypophosphatasia Research Fund, and the Barnes-Jewish Hospital Foundation.
Xiafang Zhang performed the LRP5 and MTHFR sequencing. Sharon McKenzie and Vivienne McKenzie helped to prepare the manuscript. The authors thank the index patient and her family members for agreeing to participate in this study.
Authors' roles: Study design: FJC and DW. Study conduct: FJC and DW. Data collection: FJC, DW, and SM. Data interpretation: FJC, DW, SM, and MPW. Drafting manuscript: FJC and DW. Revising manuscript content: FJC, DW, SM, and MPW. Approving final version of manuscript: FJC, DW, SM, and MPW.