Introduction

Low-density lipoprotein receptor-related protein 5 (LRP5) is a type I transmembrane receptor belonging to the LDL receptor family, composed of an extracellular domain, a membrane-spanning domain, and an intracellular domain [1]. The extracellular domain is a binding region that has extremely high affinity for the ligand, composed of four YWTD β-propellers, four EGF-like domains, and LDLR type A domains [2]. It acts as a coreceptor involved in Wnt signal transduction and consequential normal osteogenic activity of osteoblasts, reported to be one of the regulators of peak bone mass in vertebrates already [3]. It is the coreceptor of Frizzled and can be activated by proteins of the Wnt family, and then through a series of intracellular reactions, β-catenin was allowed to enter the nucleus and to modulate relative gene-transcription regulators, which is the Wnt canonical pathway [4]. In the aspect of osteogenesis, it is shown to strongly activates bone formation by enhancing cell commitment toward the bone lineage and the proliferation and differentiation of osteoblasts [2]. On the other hand, Dickkopf (DKK)1 and sclerostin (SOST), the natural antagonists of the Wnt pathway, can bind Kremen and LRP5 and block this canonical pathway, leading to the inhibition of bone formation moderately and maintenance of bone homeostasis finally [5].

LRP5 has already been identified to be associated with both osteoporosis-pseudoglioma syndrome (OPPG, OMIM 259,770) and LRP5 high bone mass (LRP5-HBM, OMIM 144,750), due to LRP5 loss-of-function mutations and gain-of-function mutations, respectively [6, 7]. The failure of properly activating Wnt signaling in osteoblasts due to loss-of-function mutations may impaired bone formation. However, the weaken binding to DKK1 and SOST because of LRP5 gain-of-function mutations reduces inhibitory effects on Wnt signaling, where osteoblast activity is increased and therefore bone formation is enhanced, which may be the cause of the high-bone-mass phenotype [7, 8].

OPPG is an autosomal recessive disorder characterized by severe juvenile-onset osteoporosis and congenital or juvenile-onset blindness [6]. Low bone mass is often found in early childhood and is usually accompanied by fragility fractures or malformations of the long bones and spine. However, some patients suffer from fractures as early as infancy [9, 10]. And patients suffer from blindness after birth [6]. Eye problems such as phthisis bulbi, retinal detachments, and falciform folds usually appear within a few days or 1–3 years after birth [11, 12].

In contrast, LRP5-HBM is a type of autosomal dominant bone disease, and its typical clinical characteristics of this disease include generalized high bone density; torus palatinus; and a wide, deep mandible [7]. Patients often show radiographically cortical thickening of diaphyses of long bones but usually little change of bone outer shape and dimensions, and osteosclerosis of ribs, clavicles, vertebrae, hip, metacarpals, and calvaria [13]. The most common facial changes are elongated mandible and torus palatinus, accounting for 61% and 41%, respectively [14]. Fortunately, this disease seems benign for no danger of fragile fractures and anemia, and sometimes, it is totally asymptomatic except for the imaging abnormalities [7]. However, there are some cases reported symptoms of neurological involvement such as trigeminal neuralgia, sensorineural hearing loss, partial visual field defects, mild facial paralysis, chronic occipital headache, and type I Chiari [15, 16], speculated as the related brain tissue and nerve compression by the hyperostotic bone [17]. It is estimated that cranial nerve deficits and/or other neurological complications affect 19.4% of the patients [14]. The biochemical parameters are usually normal, including serum calcium, phosphonium, and bone turnover markers. We previously reported that two patients had elevated serum SOST levels, but the mechanism is unclear [18].

Herein, the clinical manifestations, disease-causing mutations, treatment, and follow-up of several patients were analyzed and summarized to improve the understanding of these two diseases.

Materials and methods

Subjects

This protocol was approved by the Ethics Committee of the Shanghai Sixth People’s Hospital Affiliated to Shanghai Jiao Tong University School of Medicine. Written informed consents were obtained from all the subjects or their guardians. Four female and two male patients aged 6 to 64 from five families were enrolled in the present study (Fig. 1).

Fig. 1
figure 1

Pedigree of the two OPPG and three LRP5-HBM affected families in this study

Full size image

Clinical features and biochemical and radiographic examination evaluation

Physical examination of the patients was performed, and detailed medical histories were recorded.

Biochemical parameters, including routine blood test, serum levels of calcium (Ca), phosphonium (P), alkaline phosphatase (ALP), liver and kidney functions, creatine kinase (CK) and its MB isoenzyme (CK-MB), 25-hydroxyvitamin D [25(OH)D], the intact parathyroid hormone (PTH), the bone formation marker serum osteocalcin (OC), and the bone resorption marker serum beta cross-linked C-terminal telopeptide of type 1 collagen (β-CTX), were measured individually. Normal reference of ALP, Ca, and P [19] and normal reference of OC and β-CTX [20] for children involved in the study are listed separately.

The radiographic examinations, including X-rays of the skull, thoracolumbar, and pelvic, were performed in six patients.

Bone mineral density (BMD) was measured by a lunar prodigy dual energy X-ray absorptiometry (DXA) densitometer (Lunar Corp, Madison, WI, USA), including the anteroposterior lumbar spine 1–4 (L1-4), the left proximal femur, and the femoral neck and total hip. These data were analyzed using Prodigy enCORE software (ver. 6.70 standard-array mode; GE Healthcare, Madison, WI, USA). The Z-scores (the number of standard deviations from the mean value for persons in the general population matched for age, sex, and race) were used for children, premenopausal women, and men under age 50.

Genetic analysis and variant prediction

Sanger sequencing was used to detect LRP5 mutations. Genomic DNA was extracted from the 3-mL peripheral blood sample of each participant using phenol extraction and isopropanol precipitation. Sanger sequencing was performed on patients and their family members by BigDye Terminator Cycle Sequencing Ready Reaction Kit v. 3.1 (Applied Biosystems, Foster City, CA, USA) on an ABI 3730XL automated sequencer (Thermo Fisher Scientific, Waltham, MA, USA). The sequencing files were analyzed by PolyPhred software for checking single-nucleotide polymorphisms (SNPs). The results were obtained after manual proofreading. All variants were mapped on transcript NM_002335.2 and protein NP_002326.2.

The ExAC and 1000 Genomes Project databases were used to identify the novel variant. Its amino acid conservation and pathogenicity were analyzed by UniProt databases (https://www.uniprot.org/), PolyPhen-2 (http://provean.jcvi.org), and MutationTaster (https://www.mutationtaster.org/).

Structural analysis of LRP5 protein

The protein sequence of LRP5 was obtained from the UniProt database (https://www.uniprot.org/) as a FASTA file. Three-dimensional structure homology modeling and visualization of the native and mutant proteins were performed using AlphaFold software (https://alphafold.ebi.ac.uk/) and Rosetta.

Results

OPPG

Clinical manifestation

Patient 1

Patient 1 (P1) was a 9-year-old girl born at full-term with healthy non-consanguineous parents. Her psychomotor development was normal. She was congenitally blind. The visual problems were noted at 3 months of age, showing right cataract and left microphthalmia in the ophthalmologic findings (Fig. 2a). At the age of 9, she had back pain and developed a kyphosis after falling. Radiologic examination showed significant systemic osteoporosis, increased dorsal kyphosis, and multiple vertebral compression fractures (Fig. 3a). BMD value of L1-4 was 0.363 g/cm2, which was significantly lower than the normal range for children of the same age and gender [21]. Results of biochemical parameters were mainly within the normal range, except for the deficiency of 25(OH)D (Table 1). Concurrently, ophthalmic reexamination showed that there was an absence of bilateral optic nerve.

Fig. 2
figure 2

Ocular clinical features of patients with OPPG. a Cataract (right) and microphthalmia(left). b Enucleation of eyeball (right) and cataract (left)

Full size image
Fig. 3
figure 3

X-rays of the spine and pelvic of OPPG. a Generalized osteoporosis with fish mouth vertebra of patient 1 before treatment. b Improvement of lumbar vertebral body morphology and increased BMD of patient 1 after 3-year alendronate treatment. c Generalized osteoporosis with lumbar scoliosis of patient 2 before treatment and d after 6-month zoledronate treatment

Full size image
Table 1 Baseline clinical features and bone mineral density in affected patients
Full size table

Patient 2

Patient 2 (P2) was the first child born to non-consanguineous parents. She was born blind and was diagnosed with right retinoblastoma and left retinal atrophy. At the age of 13, she underwent enucleation of her left eye due to tumor swelling and retinal perforation (Fig. 2b). She experienced chronic systemic bony pain and had her first nontraumatic fracture at the manubrium sterni at the age of 9. In the following years, many fractures occurred, including the ribs, humerus, thoracolumbar spine, hip, femur, tibia, and ankle, mainly involving long bones and vertebrae. When she was 16, she could not keep her upper body upright due to multiple vertebral compression fractures. She could stand unsupported but unable to walk independently and must rely on a wheelchair. In addition, she had torn ligaments in her knees and ankles and was also troubled by recurrent aseptic inflammation of her sacroiliac joint.

Besides, she has had frequent premature beats since childhood and developed into paroxysmal atrial fibrillation a few years later, and once had a cardiac arrest. However, no organic lesions were found by cardiac examination.

The results of biochemical parameters including serum calcium, phosphonium, ALP, PTH, and bone turnover markers were summarized in Table 1. Except vitamin D deficiency, all indicators were within the normal range. Radiologic examination showed thoracolumbar scoliosis, mild compression fractures of the local thoracic spine, and low pelvic BMD (Fig. 3c). The BMD of L1-4 was 0.630 g/cm2, and the Z-score value was − 4. Her parents and brother were healthy and had no vision or skeletal problems.

Identification of LRP5 variants

We verified the mutations in the patients and their family members if they agreed, and the results are presented in Figs. 1a, b and 6a. Sanger sequencing confirmed that P1 had a compound heterozygous mutation in exon 7 (c.1455G > T, p.Glu485Asp) and exon 8 (c.1708C > T, p.Arg570Trp) and P2 had a compound heterozygous mutation in exon 6 (c.1145C > T, p.Pro382Leu) and exon 23 (c.4830dupC, p.cys1611leufsx33). Meanwhile, some healthy relatives of the two patients carried a heterozygous mutation from their corresponding family.

Variant prediction and structural analysis

Multiple sequence alignment reveals that Glu485 is highly conserved in different species (Fig. 6a), and Glu485Asp is predicted as “disease-causing” by PolyPhen-2, MutationTaster, SIFT, and PROVEAN (Supplementary table). Furthermore, the three-dimensional structure homology modeling and visualization of the native and the three missense mutant proteins show that the three mutations have no apparent effect on protein structure. As for Pro382Leu, since no hydrogen bonds are formed before and after the mutation, there are no changes at least in the hydrogen bonds formed with neighboring atoms at the mutation site. However, as for Glu485Asp, our AlphaFold result shows there are changes in the length of the hydrogen bonds, while Rosetta result shows the number of hydrogen bonds changed from four in the native type to two in the mutant type. It also forms fewer hydrogen bonds at the site of the Arg570Trp in both modeling results (Fig. 6b and SI).

Treatment and follow-up

P1 received oral alendronate treatment at a dose of 70 mg/week from the age of 9 to 12. During the 3-year treatment, bone pain was significantly relieved. The baseline and follow-up serum levels of calcium, phosphorus, bone turnover markers, alkaline phosphatase, and PTH were summarized in Table 2. All of them were within in the normal level and showed no significant difference. Oral administration of 600 IU of vitamin D3 per day did not improve the state of vitamin D deficiency. The spinal X-ray after treatment showed significant improvement in the shape of the thoracolumbar spine (Fig. 3b), and there were no new fractures during the treatment. Furthermore, the BMD of L1-4, femur neck, and total hip was significantly increased by 107.2%, 70.0%, and 65.5%, respectively (Table 2).

Table 2 Pre- and post-treatment biochemical data and bone mineral density of two OPPG patients
Full size table

As for P2, she was diagnosed as “osteoporosis” or “osteogenesis imperfecta” in other hospitals and took alendronate 70 mg per week at the age of 13. Six months later, she quit because of recurrence of fractures and no significant improvement in BMD. Last year, at the age of 27, she received 5 mg of zoledronic acid intravenously. Simultaneously, she took vitamin D3 800 IU and calcium 1000 mg per day. The clinical symptoms improved within half a year after injection. She could walk longer without a wheelchair. Although she fell down twice, she did not have a fracture, which was likely to happen in the past. BMD of L1-4, femoral neck, and total hip increased by 5.87%, 6.76%, and 4.15%, and β-CTX and osteocalcin decreased slightly after 6 months of treatment. Serum calcium, phosphorus, and parathyroid hormone levels were almost unchanged. Although the level of 25(OH)D increased, it was still below the normal range (Table 2).

No serious adverse drug reactions occurred in both patients during the treatment.

LRP5-HBM

Clinical manifestations

Patient 3

Patient 3 (P3), a 32-year-old male, came to our department because of lumbodorsal pain. A torus palatinus in the center of the oral hard palate was detected by physical examination (Fig. 4a). X-ray examination revealed that the skull, thoracolumbar vertebrae, and pelvis bones were dense, and the cranial plate was significantly thickened (Fig. 5a). The Z-scores of BMD of the L1-4, femur neck, and total hip were + 10.5, + 11.1, and + 12.3, respectively (Table 1). The patients have no history of fractures, sensorimotor neuropathy, and visual or hearing impairment. The level of serum calcium, phosphonium, ALP, PTH, 25(OH)D, and bone turnover markers was all within the normal range (Table 1). His family members were all healthy.

Fig. 4
figure 4

Clinical manifestations of patients with LRP5-HBM. a, b, c The torus palatinus in the center of the hard palate of 3 patients. d A slight depression in the right chest of patient 4

Full size image
Fig. 5
figure 5

X-rays of the skull, spine, humerus, and femur of two LRP5-HBM patients (a: patient 3; b: patient 4). Osteosclerosis of the skull with an enlarged sella turcica. Generalized osteosclerosis. Cortical thickening of the humerus and femur with a normal external shape

Full size image

Patient 4

Patient 4 (P4) was a 6-year-old boy born at full-term who was found to have thoracic deformity at the age of 3. He was diagnosed as “Rickets” in the local hospital, but the symptoms did not improve after calcium and vitamin D supplementation. At the age of 6, the abnormal increased bone mineral density of ribs, sternum, thoracic vertebrae, and bilateral humeral metaphysis was identified by occasionally chest X-ray examination due to cough. In order to confirm the diagnosis, the child underwent further radiological examination. The examination found that the patient’s whole-body BMD increased and the skull plate thickened (Fig. 5b). The patient had normal tooth development and maxillary morphology. No jaw enlargement or palatal ring was found. The patient had a slight depression on the right side of the chest (Fig. 4d) and had no history of fracture.

Moreover, the value of BMD of L1-4, femoral neck, and total hip was 1.149 g/cm2, 1.327 g/cm2, and 1.321 g/cm2, respectively (Table 1). The biochemical indices including serum calcium and phosphonium, ALP, PTH, 25(OH)D, β-CTX, and OC were within the normal range (Table 1).

His parents were healthy and non-consanguineous.

Patients 5 and 6

Patients 5 (P5) and 6 (P6) were a mother-daughter pair, as reported in our previous study [18]. In brief, both of them had chronic lumbodorsal pain, an elongated mandible and torus palatinus (Fig. 4b and c), high bone mass, but no history of fractures. X-ray radiographs showed thickening of the cranial plates, an elongated mandible, cortical thickening of the long bones, and degenerative changes. The biochemical indices showed that they both had vitamin D deficiency. The serum ALP level of patient 6 was slightly higher than normal range, while β-CTX was much higher than the normal range (Table 1).

Identification of LRP5 variants

The LRP5 gene detection results of patients and their family members were shown in Figs. 1c, d, and e and 6a. By Sanger sequencing, the heterozygous c.724G > A (p.Ala242Thr) mutation was detected in P3, P5, and P6. And the heterozygous c.640G > A (p.Ala214Thr) mutation was detected in P4. We did not identify any pathogenic variants in family members other than the patients.

Fig. 6
figure 6

Sanger sequencing and the three-dimensional modeling. a Multiple sequence alignments revealed that the glu485 residue in the LRP5 protein was highly conserved among species; Sanger sequencing of the LRP5 mutations: p.Glu485Asp (E485D, c.1455G > T) and p.Arg570Trp (R570W, c.1708C > T) in patient 1; p.Pro382Leu (P382L, c.1145C > T) and p.Cys1611LeufsX33 (c.4830dupC) in patient 2; p.Ala242Thr (A242T, c.724G > A) in patients 3, 5, and 6; p.Ala214Thr (A214T, c.640G > A) in patient 4. b Comparison of the three-dimensional modeling of wild and missense mutations’ local structures by AlphaFold; distribution of the six mutation sites on the LRP5 protein

Full size image

Structural analysis

As predicted by AlphaFold and Rosetta, the conversion of alanine to threonine in two mutation sites has no apparent effect on protein structure. But in the modeling results of both software, the number of hydrogen bonds increases after the mutations (Fig. 6b and SI).

Follow-up

P5 underwent 12 years of follow-up. During the follow-up period, pain in the thoracolumbar spine, cervical spine, and shoulders was the main issue, but there were no fractures, no dental problems, and no significant changes in height. In addition, no neurological symptoms such as trigeminal neuralgia, neurogenic hearing loss, visual field deficits, or chronic headaches occurred during the follow-up period. She only takes painkillers when necessary to relieve her lower back pain. In the past year, she has experienced menstrual cycle disorders. After 12 years of diagnosis, the patient underwent a reexamination of BMD, bone turnover markers, and biochemical parameters. The detailed results were summarized in Table 3. Compared with the results 12 years ago, the absolute value of BMD of L1-4, femoral neck, and total hip was decreased by 9.62%, 12.81%, and 11.03%, respectively, but it was still higher than the reference value of the same age and gender. The serum levels of PTH, Ca, and P were normal, but 25(OH)D is still below the normal range. Bone turnover markers, including β-CTX and OC, were higher than the normal range.

Table 3 Baseline and follow-up data of biochemical data and bone mineral density of one LRP5-HBM patient (Patient 5)
Full size table

Discussion

In the current study, we reported five mutations and a variant, four from OPPG patients (Glu485Asp, Arg570Trp, Pro382Leu and Cys1611LeufsX33) and two from LRP5-HBM patients (Ala242Thr and Ala214Thr). Glu485Asp, Arg570Trp, and Pro382Leu all lie in the exons collectively coding for the second β-propeller domain, consistent with that most OPPG mutations reported have been described in the second and third b-propeller domains [22]. However, Cys1611LeufsX33 lies in the intracellular domain, which is supposed to be phosphorylated and therefore inhibit GSK3b when Wnts bind to its coreceptors Frizzled and LRP6/5 [23]. In addition, the novel variant Glu485Asp is considered to be pathogenic after conservation and pathogenicity prediction, which broadens the variation spectrum of pathogenic genes. Ala242Thr and Ala214Thr are the hot spot mutations in patients with LRP5-HBM [3, 14, 18], both located in the amino terminal part of the gene, before the first EGF-like domain, like the previously reported mutations that cause the high-bone-mass phenotype. Besides, Ala242Thr has been predicted to incline to disrupt the core packing of the LRP5 protein structure and affect the first β-propeller domain, thereby destabilizing the binding site of SOST and LRP5 [24]. We performed the three-dimensional structure homology modeling and visualization of the native and five missense mutant proteins using the online AlphaFold and Rosetta software. Glu485Asp, Arg570Trp, Ala214Thr, and Ala242Thr all had variations in the number or length of hydrogen bonds. Interestingly, we found that the number of hydrogen bonds at the mutation site decreased in the loss-of-function mutation (Glu485Asp and Arg570Trp), while they increased in the gain-of-function mutation (Ala214Thr and Ala242Thr), which may be associated with the loss and gain of LRP5 function. Modeling of Pro382Leu by two software showed no significant changes in structure and hydrogen bonding before and after mutation, but case reports of both homozygous and compound heterozygous mutations of Pro382Leu have been reported [25, 26]. Its effect on the local structure of LRP5 protein and the intensity of Wnt signaling pathway remained unclear. Cys1611LeufsX33 was a frameshift mutation where termination codon was shifted back. And the mutation was located at the site where LRP5 intracellular structure binds to Wnts and Frizzled, which might have a great influence on the transmission of the intracellular part of Wnt signal.

OPPG children may sometimes be misdiagnosed as osteogenesis imperfecta, but they have no identifiable defects in collagen synthesis and metabolism, with their calcium homeostasis, endochondral growth, and bone turnover index all normal [27, 28]. It is worth mentioning that P2 has developed nonorganic paroxysmal atrial fibrillation of unknown cause in recent years, which is not found in other OPPG patients reported. Besides, neurological diseases such as seizures, autism spectrum disorder, and intellectual deficit in individual OPPG cases are reported abroad [29,30,31], but there are not such problems in OPPG patients reported in our country [32], including the two patients in present study. It is still unclear whether the clinical variability of neurological phenotypes can be attributed to mutational effects or a specific protein domain [10, 33].

At present, there are few treatment options for OPPG children. Bisphosphonates are the most commonly used and rational drug for them, and its effect on improving the bone condition has been confirmed [9, 30]. In the present study, the beneficial effects have been observed in the 3-year alendronate sodium for P1 and half-year intravenous zoledronic acid for P2. During and after bisphosphonate treatment, for both the biochemical indexes and hormone values were basically within the normal range, the symptoms were improved, and the risk of fracture was reduced. However, P2 experienced a period of failed oral alendronate treatment before the zoledronic acid, manifested by the appearance of new fractures and unrelieved bone pain. Actually, in this case, the drug adjustment should be determined according to a combination of factors, including fracture history, bone pain, growth, and bone mineral density.

In the present study, high bone mass, torus palatinus, and an elongated mandible were the most common clinical symptoms for our LRP5-HBM patients. The serum ALP of all patients was within the normal range, and no symptoms of nervous system were involved. Interestingly, the 6-year-old patient (P4) did not have torus palatinus but had thoracic deformity, which had never been reported in cases of LRP5-HBM, and it was uncertain whether it was caused by LRP5 gene mutation.

In the current study, we found that P5 has entered a period of rapid bone turnover; β-CTX and OC were out of normal range. The BMD of L1-4 decreased by 9.6% within 12 years, with an average annual decrease of 0.8%. Considering that our patient has been in the late perimenopause stage (no menstrual bleeding in the last 3 months but some menstrual bleeding during the last 11 months), the menstrual disorder in the past year may partly be the reason for the decrease of bone mass and the acceleration of bone turnover. Since there were no long-term longitudinal studies on changes in bone mineral density during the perimenopausal transition in mainland Chinese women, we used data collected from women in Chinese Hong Kong [34] as a reference, from which we estimated that the lumbar spine bone loss rate of healthy women population from age 40 to 52 is about 7.94%. Anyhow, it seemed that our patient did not seem to lose bone at a slower rate than healthy controls, and the gain-of-function mutation Ala242Thr could not protect her from age-related and estrogen-related bone loss. Similarly, another study found that individuals with LRP5T253I-HBM were not protected from age-related bone loss, at least at the hip and femoral neck [35]. Whether LRP5 gain-of-function mutations can actually help carriers resist bone loss due to bone wasting factors, and whether this effect is related to the mutation site needs more follow-up data to be confirmed.

Conclusion

In the current study, two cases of OPPG caused by LRP5 loss-of-function mutations and four cases of LRP5-HBM caused by LRP5 gain-of-function mutations were reported. For OPPG, a novel heterozygous missense mutation was reported to expand the genotypic spectrum of OPPG. The bisphosphonates treatment can increase the BMD and decrease the risk of fracture for patients with OPPG. On the other hand, LRP5-HBM is a relatively benign disease. LRP5 gene mutation, however, may not slow down the rate of bone loss, including postmenopausal bone loss. This phenomenon deserves further observation and study.