|Home | About | Journals | Submit | Contact Us | Français|
Osteoporosis-pseudoglioma syndrome (OPPG) is a rare autosomal recessive disorder of severe juvenile osteoporosis and congenital blindness, due to mutations in the low-density lipoprotein receptor-related protein 5 (LRP5) gene. Approximately fifty cases of OPPG have been reported. We report 9 new cases of OPPG, in three related nuclear families of Conservative Mennonites in Pennsylvania. All 9 children with OPPG were blind and had osteoporosis. Four of six parents had low bone mineral density (BMD) or osteoporosis; 2 were normal. Sequence analysis from genomic DNA revealed homozygosity for a nonsense mutation of exon 6 of LRP5 (W425X) in four OPPG cases tested in families A and C. In family B, OPPG cases were compound heterozygotes for the exon 6 W425X LRP5 mutation and a second exon 6 mutation (T409A); bone phenotype was milder than in family A. Neither of these mutations was present in an unrelated normal. The four treated OPPG patients all responded to bisphosphonates (duration 1.5–6.5 years) with improvement in Z-scores. One patient had a negligible response to teriparatide. In summary, we report 9 new cases of OPPG due to two novel LRP5 mutations, note a milder bone phenotype but similar ocular phenotype in LRP5 W425X/T409A compound heterozygotes than in W425X homozygotes and describe positive response to bisphosphonate treatment in four cases.
Osteoporosis-pseudoglioma syndrome (OPPG) is a rare autosomal recessive syndrome of juvenile-onset osteoporosis and ocular abnormalities described first by Saraux in 1967 . To date, approximately 50 cases have been reported [1-14], most with onset of fractures after age 2 years , and congenital blindness, the latter due to persistence of the fetal ocular fibrovascular system and failure of retinal development [2,3]. OPPG is caused by inactivating mutations in the low-density lipoprotein receptor-related protein 5 gene (LRP5) . LRP5 is a membrane co-receptor in the canonical Wnt signaling pathway. LRP5 mutations that prevent Wnt from binding to LRP5 lead to pathway inactivation and OPPG, while mutations that prevent LRP5 from binding to co-factor Dikkopf lead to pathway activation and excessive bone formation . At least 12 different homozygous [2,4] and 15 compound heterozygous [2,5] LRP5 mutations have been described in OPPG, most in the second or third of its four beta-propeller domains, which are high affinity Wnt binding domains that are highly conserved . No phenotypic distinctions have been reported in compound heterozygotes compared to homozygotes . Obligate heterozygotes have been found to have modest degrees of low bone mass but no eye pathology [4,16].
Inactivating mutations in LRP5 can also cause familial exudative vitreoretinopathy (FEVR), a disease characterized by premature arrest of the retinal vasculature . The resulting retinal avascularity leads to complications including folding and neovascularization of the retina, which in turn leads to retinal detachment and blindness in most [18,19]. FEVR has been classically described without an osteoporotic phenotype (fractures, bone deformity) [18,20]. However, when bone mineral density testing has been routinely performed, most with FEVR have reduced bone mass or osteoporosis, leading to the suggestion that OPPG and FEVR are in a single phenotypic spectrum . FEVR is heterogeneous, with 20% having LRP5 mutations, 20–40% with mutations in the LRP5 co-receptor Frizzled 4 (FZD4) and 40–75% idiopathic but reduced bone mass has been seen only in those with LRP5 mutations .
Activating mutations in LRP5 cause Familial High Bone Mass Syndrome (HBM), an autosomal dominant syndrome associated with non-pathologic high bone mass and have also been detected in some pathological bone sclerosing disorders [21-23], providing evidence that this gene is an important regulator of bone production. In familial HBM, activating LRP5 mutations have been confined to the first beta-propeller domain [20-23]. No obvious ocular pathology has been reported in association with activating LRP5 mutations.
Scanty information is available on the treatment of osteoporosis in OPPG. We are aware of only two reports of bisphosphonate treatment in OPPG, a 21 year-old woman treated for 3 years with pamidronate  and 3 children aged 9–11 , treated with pamidronate for 2 years; all had improvement in bone mineral density. We report 9 new cases of OPPG in related members of 3 nuclear families from a Conservative Mennonite population in Pennsylvania. We describe 2 novel mutations in LRP5, note a milder bone phenotype in W425X/T409A compound heterozygotes than in W425X homozygotes and report response to treatment. Four patients were treated with bisphosphonates (1.5–6.5 years in duration); in one of these cases, bisphosphonate treatment was followed by teriparatide (total treatment duration 7.5 years).
Patients were evaluated at the Amish Research Clinic in Strasburg, PA. The protocol was approved by the University of Maryland Institutional Review Board and signed informed consent from adults and assent from children was obtained. A history including fractures was obtained. A general physical examination included height, weight, and musculoskeletal exam in those with OPPG (shown in Fig. 1 as shaded symbols). Ophthalmology exams were conducted on all with OPPG and in their parents.
Bone mineral density (BMD) was measured by dual energy X-ray absorptiometry (DXA), using a Hologic 4500W (Bedford, MA), at the lumbar spine and hip in children and adults by a registered nurse certified in bone densitometry. In children, premenopausal women and men under age 50, the Z-score, standard deviation (SD) from age and gender-matched controls, were used to assess normalcy of BMD. A Z-score above −2.0 was considered to be normal. For individuals under age 20 years, Hologic pediatric software was used to analyze scans. Pediatric Z-scores are not available for the spine under age 3 years or for the hip under age 5 years. For DXA scans done under age 3 years, we have used estimated Z-scores extrapolated from Hologic adult data. Because of the lack of pediatric Z-scores for the hip under age 3 years, we do not report the hip results here. In postmenopausal women and men over age 50, the T-score, (standard deviation from young adults at peak BMD) was used to assess BMD. According to the World Health Organization criteria , a T-score of −1.0 to −2.5 was considered “low bone mass” or “osteopenia” and lower than −2.5, osteoporosis. The coefficient of variation for BMD measurements, determined annually by 3 sequential measures on 1 day for each of 15 individuals, was 0.90% for total hip and 0.71% for the spine (L1–L4). The student’s Student’s t test was used to calculate the difference between Z-scores in W425X homozygotes and W425X/T409A compound heterozygotes.
Genomic DNA was extracted from peripheral blood leukocytes from OPPG cases and their parents. Reference genomic DNA sequence for LRP5 (accession ID NM_002335) and exon information (accession ID-s AF283320 for exons 1–9 and AF283321 for exons 10–23) were obtained from GenBank. This information was used to design intronic PCR primers that flanked each of the 23 exons in LRP5. PCR products were generated from genomic DNA obtained from the OPPG patients and their parents and an unrelated individual and sequenced on an ABI3730 (Applied Biosystems, Foster City, CA). PCR products were sequenced in both directions. The obtained sequences were analyzed using Sequencher (GeneCodes, Ann Arbor, MI).
The pedigree is shown in Fig. 1. Individuals who were evaluated are noted by an asterisk. In this 6-generation Conservative Mennonite family, there were at least 3 consanguineous marriages, the closest of which were second-degree relatives. Clinical characteristics and DXA Z-scores are shown in Table 1 (one child with OPPG did not have a DXA due to age under 1 year). In families A and C, all OPPG patients were blind since birth (minimal light perception at birth in some that was lost after the first year); in family B light perception was present at birth and was lost by age 10 years. Eye pain was noted in A1 in early childhood but resolved; in C1 and C2 mild eye pain was still present at the time of this report (ages 4 years and 6 months). All had grossly normal intelligence, white sclerae and normal joint mobility. Height was normal in all except for A1 and A2, both below 5th percentile. One of nine with OPPG, C1, had a behavioral disorder characterized by excessive aggression (e.g. biting others) and he was noted to be mildly hypotonic as an infant. C1 was seen by a child psychiatrist (Kennedy Institute at Johns Hopkins) who noted aggressive behavior, expressive language delay and social delay and diagnosed “autism spectrum disorder”. He had no improvement in behavior to risperdone but benadryl (used as needed) has had a beneficial effect on his aggressive behavior. The mother in family C (age 22 years) had a Z-score in the osteoporosis range (because she was premenopausal, the T-score is not appropriate to use for her) but had no history of fracture. Her vision and eye exam were normal. Affecteds in families A and C had typical changes seen in OPPG; in family B, eye findings were more typical of FEVR. Pathology findings from an eye removed (because of pain) in OPPG B1 are shown in Table 1. Biochemical data on patients with OPPG is shown in Table 2.
We sequenced each of the 23 exons of LRP5 in affected subjects and their parents. Homozygosity for a novel nonsense mutation in exon 6 (substitution of a stop codon for tryptophan, W425X) was found in affected individuals in families A and C. In both families A and C, both parents were heterozygous for the mutation. This mutation would be expected to lead to a truncated protein of 425 amino acids as compared to the normal protein of 1615 amino acids. In family B, the mother was heterozygous for the same W425X mutation found in families A and C while the father was heterozygous for a different mutation in exon 6, a novel missense mutation leading to the substitution of alanine for threonine (T409A). The children with OPPG in family B were compound heterozygotes for the two exon 6 mutations (W425X, T409A).
OPPG A1, A2, A3 and C1 were treated with bisphosphonates for 6, 4.5, 1.5, and 1.5 years respectively. The BMD treatment responses are shown in Fig. 2; biochemical responses in Table 2. In addition, all were treated with calcium 1000 mg daily and vitamin D3 400–800 IU daily, with vitamin D dose titrated to a target 25-hydroxyvitamin D level of >30 ng/ml. Initially, risedronate was used in A1 and A2 for cost reasons (no have health insurance). However, the BMD response to ridedronate was minimal, so both were changed to IV pamidronate after 2 (A1) and 1 (A2) years, using the dosing used in children with osteogenesis imperfecta , 1 mg/kg over 3 h on 3 successive weeks, repeated every 3 months. As shown in Fig. 2, both A1 and A2 had good responses to pamidronate, with improved Z-scores. Pamidronate was stopped after 3.5 years in A1 and changed to teriparatide because he had a femur fracture while on treatment with pamidronate, and his Z-score after an initial good response to pamidronate (Z-score −4.0 improved to −2.2), did not improve from age 13 (Z-score −2.2) to 14 (Z-score −2.2). After a 6 month period of bisphosphonate treatment, he was started on teriparatide, initially at 20 mg subcutaneously every other day. After 4 months to confirm lack of hypercalcemia, the dose was increased to 20 mg daily for the next 11 months. After 15 months of teriparatide, the Z-score did not change significantly from −2.5 to −2.6, so it was stopped. After discontinuing teriparatide, alendronate 1 mg/kg/day (70 mg 4 times a week) was started and after 6 months, his Z-score improved from −2.6 to −2.4 and his urine N-telopeptide decreased. In A1, his initial spine Z-score improved from −6.4 (age 8 years) to −2.4 (age 16 years). In A2, after 1.5 years of pamidronate, he was changed to oral alendronate 1 mg/kg/day ; his initial Z-score of −4.7 (age 4 years) improved to −0.4 (age 11.5 years).
Patient A3 was started on oral alendronate 1 mg/kg/day at age 4 years and after 1.5 years, her Z-score improved from an estimated −2.9 (age 2 years) to −0.5 (age 5.5 years). She has had no fractures. Patient C1 was started on the same alendronate regimen at age 2 8/12 years after a femur fracture. After 17 months of treatment, his Z-score improved from an estimated −4.2 (age 2 8/12 years) to −0.3 (age 4 years) and he had no further fractures. In family B, bisphosphonate therapy and follow up DXA were declined.
We report 9 new cases of osteoporosis-pseudoglioma (OPPG) syndrome in three related nuclear families. All OPPG patients had a severe ocular phenotype, characterized by congenital blindness, regardless of genotype. By contrast, the bone phenotype was more severe in those with a homozygous W425X genotype than in W425X/ T409A compound heterozygotes. The two novel mutations in LRP5 described here were both in the second beta-propeller structure. Four of the six adult obligate heterozygotes studied (parents of affected individuals) had mild osteopenia or osteoporosis, as has been shown in other families [4,16], and another one was normal. Improvement was seen in bone mineral density Z-score from treatment with bisphosphonates in the four treated OPPG patients (from baseline to most recent: A1 −6.4 to −2.4, A2 −4.7 to −0.4, A3 −2.9 to −0.5, C2 −4.2 to −0.3).
The clinical features described in our OPPG patients A1–3 and C1–2 were similar to those described previously, with congenital blindness, severe osteoporosis, short stature (in 2/9) and fractures occurring in early childhood . Our patient A1 had the most severe bone disease with multiple fractures and bone deformity leading to wheelchair confinement since age 7. Although he also had propionic academia, this condition has not been associated with bony fragility . As found in previous reports [2,16], most of our obligate heterozygotes had reduced bone mineral density, but none had eye pathology.
The homozygous W425X mutation predicts a stop codon and a severely truncated form of the LRP5 protein containing 425 instead of the normal 1615 amino acids, likely resulting in complete absence of functional protein. The clinical result was classic OPPG syndrome in the 4 affected children homozygous for this genotype, with mean spine Z-scores of −4.6, similar to the reported mean in OPPG of −4.7 ± 0.9 . The homozygous W425X phenotype was similar to a previously reported case of OPPG caused by a homozygous R428X mutation . It is noteworthy that we found a milder bone phenotype in compound heterozygotes (cases B1–4, W425X/T409A) consisting of normal stature and a mean Z-score of −1.9 vs −4.6 for W425X homozygotes (p = 0.02). These compound heterozygotes, however, had similarly severe blindness to W425X homozygotes.
We are unaware of any previous reports of a mild bone phenotype with a compound heterozygous genotype that combines a stop codon and a missense mutation. We speculate that the missense T409A mutation might result in partially functional LRP5, adequate to stimulate slightly reduced bone production through the Wnt pathway, yet inadequate for normal eye development. LRP5 expression constructs studies  designed to assess Wnt signal transduction have demonstrated variable β-catenin-mediated signal transduction. Interestingly, mutant constructs G404R and D434N that flank T409A and R425X had <50% of wild type activity, while the majority of other constructs were unable to transduce Wnt1 or Wnt10b signal. Despite this in vitro result demonstrating some Wnt signal transduction capability, the clinical result of the construct mutations in patients was OPPG. The function of LRP5 in eye development is complex and involves varied mechanisms, including direct stabilization of b-catenin (through Axin binding)  and transduction of Norrin signaling . We speculate that our T409A mutation, as found in other OPPG missense mutations , resulted in reduced Norrin transduction, disrupting eye development but resulted in enough Wnt signaling to disrupt bone formation less severely. However, the mechanism of mutational effect in the compound heterozygotes must await functional studies of the mutant proteins.
Bone turnover markers, urine N-telopeptide (NTX) and alkaline phosphatase were normal for age in our OPPG patients. In the two previous reports of treatment in OPPG, one did not report markers of bone resorption ; in the other, one of three patients had minimally elevated urinary deoxypyridinoline . We also found that vitamin D deficiency [25(OH)D<20] and insufficiency [25(OH)D<30 ng/ml] were common in our OPPG patients. Because of the importance of vitamin D to bone health , we recommend following 25(OH)D levels in OPPG patients.
The most robust treatment responses were seen in the youngest patients treated with alendronate (1 mg/kg/day). Possible explanations include a reduced response with age or more likely a better response with alendronate because it is a more potent antiresorptive agent than pamidonate and risedronate used previously. In our patient with the most severe osteoporosis and the oldest of those treated, OPPG A1, bisphosphonates appeared to lose efficacy over time. Because of ongoing fractures, he was given a trial of teriparatide, synthetic parathyroid hormone (PTH), the only anabolic bone agent available for treatment of osteoporosis [30-32]. PTH treatment is known to stimulate the Wnt signaling pathway, by decreasing DKK-1  and sclerostin , both of which bind to LRP5 and decrease Wnt signaling. In studies on Lrp5 knock out mice, PTH treatment increased bone formation similarly to that seen in wild type mice , with the predominant increase seen in cortical bone [30,35]. Therefore, studies in mice have predicted a beneficial response to PTH in human OPPG. However, the human response to PTH is somewhat different from that in mice in that the BMD increase is seen in trabecular bone in humans  and in cortical bone in mice . This raises the possibility that the mouse response to PTH might not be a perfect predictor of the human response.
Our OPPG patient treated with PTH (teriparatide) did not appear to have a definitive response to it in that his Z-score did not improve after 15 months of treatment. This could be explained by his prior longterm treatment with bisphosphonates, which are known to blunt the effect of teriparatide . The 6 month gap between bisphosphonate and teriparatide treatment may have been inadequate to reduce the bisphosphonate effect of suppressing bone turnover. In this patient, we observed what appeared to be a paradoxical decrease in urine NTX (bone resorption marker) while on teriparatide, opposite to the increase in NTX expected [24,31]. We do not have an explanation for this. After discontinuing teriparatide, this patient was started on alendronate, with subsequent improvement in Z-score and further reduction in NTX. A controlled trial of teriparatide as the first line treatment in humans with OPPG could help to determine whether this drug is effective in this condition.
OPPG is a rare genetic syndrome. However, LRP5 has been shown to be important in the attainment of peak bone mass and skeletal response to loading  and to be a determinant for normal bone density [38,39]. LRP5 mutations have been reported in several cases of idiopathic juvenile osteoporosis . In addition, common polymorphisms of LRP5 have been associated with variations in BMD, fracture risk and height in youth and old age ( for review). Therefore, learning more about OPPG could potentially help our understanding of the pathophysiology of common osteoporosis.
In summary, we report 9 new cases of OPPG and describe a milder bone phenotype in individuals who were W425X/T409A compound heterozygotes than in W425X homozygotes. We report a beneficial response to bisphosphonates in four patients and a lack of definitive response to teriparatide in one. We recommend starting treatment with bisphosphonates in OPPG patients within the first few years of life.