Characteristics of early Paget’s disease in SQSTM1 mutation carriers: baseline analysis of the ZiPP study cohort

Objectives: Mutations in SQSTM1 are strongly associated with Paget’s disease of bone (PDB) but little is known about the clinical characteristics of those with early disease. Methods: Radionuclide bone scans, biochemical markers of bone turnover and clinical characteristics were analysed in SQSTM1 mutation carriers who took part in the Zoledronic acid in the Prevention of Paget’s disease (ZiPP) study. Results: We studied 222 individuals of whom 54.9% were female with average (± sem) age of 50.1± 0.6 years. Twelve SQSTM1 mutations were observed, including p.Pro392Leu which was present in 141/222 (63.5%) subjects. Bone scan examination revealed evidence of PDB in 20 subjects (9.0%) of which 9 (50%) had a single affected site. Participants with lesions were older than those without lesions but the difference was not significant (53.6 ± 9.1 vs. 49.8 ± 8.9, p=0.07). The mean age of participants with lesions was not significantly different from the age at which their parents were diagnosed with PDB (55 years vs. 59 years, p=0.25). All individuals with lesions were asymptomatic. Serum concentrations of total alkaline phosphatase (ALP) normalised to the upper limit of normal in each centre were higher in those with lesions (0.75 ± 0.69 vs 0.42 ± 0.29; p<0.0001). Similar findings were observed for other biochemical markers of bone turnover but the sensitivity of ALP and other markers in detecting lesions was poor. Conclusions: Asymptomatic PDB is present in about 9% of SQSTM1 mutation carriers by the fifth decade. Further follow up of this cohort will provide important information on the natural history of early PDB and its response to treatment.


Introduction
Paget's disease of bone (PDB) is a skeletal disorder characterised by focal increases in bone remodelling at one or more sites throughout the skeleton (1) . Genetic factors play a key role in PDB and the most important susceptibility gene is SQSTM1 (2,3) . Protein coding mutations of SQSTM1, which predominantly affect the ubiquitin-associated (UBA) domain, have been detected in between 40% and 50% of people with a family history of PDB and up to 15% of those who are unaware of a family history (4) . Cross-sectional studies suggest that SQSTM1 mutations are highly penetrant such that up to 80% of carriers develop the disease by the seventh decade (4) . Furthermore, people with PDB that carry SQSTM1 mutations have more severe disease with an earlier age at onset than those that do not have SQSTM1 mutations (5) . Although the causal link between SQSTM1 mutations and PDB is beyond doubt, there is limited information on the natural history of PDB in people who carry SQSTM1 mutations and the acceptability of genetic testing for PDB in routine clinical practice. In order to gain an insight into these issues, we investigated the clinical and biochemical characteristics of SQSTM1 mutation carriers with evidence of PDB-like bone lesions who were enrolled into the Zoledronate in the Prevention of Paget's disease (ZiPP) study which is a large scale prospective study which aims to investigate the acceptability of genetic testing for PDB, coupled with offer of targeted intervention with zoledronic acid (ZA) or placebo in SQSTM1 mutation carriers.

Study design
The ZiPP study (http://www.isrctn.com; Registration number: ISRCTN11616770) is a multi-centre, multinational randomized, placebo-controlled trial involving a large scale programme of genetic testing for SQSTM1 mutations coupled with the offer of targeted intervention in the form of a randomised trial, to determine whether treatment with a single infusion of intravenous zoledronic acid can modify the development of PDB-related lesions in asymptomatic carriers of SQSTM1 mutations (6) . The study commenced in 2009 and is still ongoing. The data presented here are derived from untreated individuals who consented to take part in the study and completed the baseline assessment. The ZiPP study involved an initial phase of genetic testing for SQSTM1 mutations in probands known to have PDB followed by testing of first degree relatives (who in almost all cases were children of probands). If the relative tested positive for SQSTM1 mutations they were then invited to participate in the ZiPP study. All individuals gave written informed consent and the study received ethical approval in all participating centres.

Participants
The selection process for inclusion was broadly similar in each centre. In each participating centre, the local investigator invited patients with PDB attending outpatient clinics to undergo genetic testing for SQSTM1 mutations and if the result was positive to pass on the information to first degree relatives about the study. The relatives who consented to undergo testing and were found to be positive for SQSTM1 mutations were then invited to participate in the interventional phase of the ZiPP study.
There were two exceptions; in Auckland and in Wrexham, relatives underwent genetic testing directly without their parents having undergone testing as part of this study. An overview of disposition of individuals who took part in the genetic testing phase is provided in Figure 1. We invited 2394 probands to undergo testing and of the 1428 (59.5%) that agreed, 247 (17.2%) were found to be carriers of SQSTM1 mutations. We then invited 1307 relatives of SQSTM1 mutation carriers to undergo testing and 750 (57.3%) agreed. Of these individuals, 350 (46.6%) tested positive for SQSTM1 mutations and 222 (63.4%) agreed to take part in the randomised trial. Supplementary table S1 provides full details of the numbers of probands and relatives screened in each study centre.
The participants were children of 129 probands diagnosed with PDB, giving an average recruitment rate of 1.72 subjects per family.
The inclusion and exclusion criteria for participation in the trial have been published previously but in essence, relatives who carried SQSTM1 mutations were eligible provided that they were age >30; had not been diagnosed with PDB and had no contraindication to receiving zoledronic acid. Recruitment to the trial commenced in May 2009 and closed in April 2015; the end of study visits are currently in progress and the results of the study are expected to be available in 2021.

Genetic testing
Testing for SQSTM1 mutations was carried out by Sanger sequencing of exons 7 and 8 and the introexon boundaries of the SQSTM1 gene according to standard techniques. If a previously unidentified mutation was detected, bioinformatics and functional studies were conducted to determine if it was likely to be pathogenic.

Biochemical studies
Samples were sent for routine biochemistry at participating centres and measurements of serum creatinine, serum total alkaline phosphatase (ALP), calcium, albumin, liver function tests and 25(OH)D were performed by the local chemical pathology laboratory according to standard techniques.
Measured ALP values were adjusted to the local reference range (RR) to give an adjusted ALP value which was used in statistical analysis. The formula used for calculation was as follows: Adjusted ALP = (measured ALP-ALP lower limit of RR) / (ALP upper limit of RR -ALP lower limit of RR). According to this methodology, a value of 1.0 would equate to an ALP at the upper limit of the reference range, a value of 2.0 to twice the upper limit; a value of 0.5 to the mid-point of the reference range and a value of 0 the lower limit of the reference range.

Radionuclide Bone Scans
Radionuclide scintigraphy was performed to detect evidence of PDB-like bone lesions according to standard techniques in all subjects following an injection of 99 m Tc labelled bisphosphonate at the baseline visit. Anonymised images were uploaded onto the study database and assessed independently by two imaging experts (DS and SHR) for evidence of PDB-like bone lesions. Lesions were coded as being present only when they were independently identified by both observers. The location and number of confirmed lesions was recorded in the study database. The reason for using radionuclide bone scans as the method of assessment is that this technique is more sensitive than plain x-rays at detecting PDB-like bone lesions (7)(8)(9) . Although many conditions can cause focal increases in tracer uptake on bone scan, the appearances in PDB almost always allow differentiation from other conditions (10) . Nonetheless, further imaging could be performed at the discretion of the local investigator with plain X-ray, CT or MRI if clinically indicated or if the pattern of uptake on the bone scan was thought not to be typical of PDB.

Statistical analysis
Descriptive and inferential statistics were generated using Minitab Express version 1.5.1. Differences between groups for continuous variables were assessed by one-way analysis of variance (ANOVA) and for categorical variables by chi-squared testing. Kaplan-Meier analysis and sensitivity and specificity testing, negative predictive (NPV) and positive predictive values (PPV) were performed using SPSS version 24. The significance level was set at p<0.05 and all tests were two-tailed.

Clinical characteristics, demographics and spectrum of SQSTM1 mutations
Of the 222 SQSTM1 positive individuals enrolled into the study, 101 (45%) were male with a mean ± sd age of 50.2 ± 9.1 years. The demographics of these subjects were similar to the 128 relatives who tested positive and who decided not to enrol in the study. The average age of these individuals was 53.1 ± 13.3 years and 49 (38.3%) were male. The spectrum of mutations in subjects who were enrolled into the trial is shown in Table 1. Twelve mutations were identified, the most common of which was p.Pro392Leu along with a variety of other missense and truncation mutations all of which that have previously been reported. The average adjusted ALP value was 0.44 ± 0.35, and in 10/222 individuals (4.5%) the ALP was above the upper limit of the local reference range. None of the participants had symptoms or signs of PDB at the time of recruitment or the baseline visit.
The characteristics of participants with lesions compared to those without lesions are shown in Table 2. The most commonly involved anatomical sites were the pelvis and spine. Most participants with lesions had monostotic disease (9/18; 60%). Representative examples of lesions are shown in Figure 2. Participants with lesions were on average about five years older than those without lesions but this difference was not significant (p=0.07). Lesions were equally common in men and women.
There was no difference in the proportion of subjects with lesions depending on whether the mutation was missense or truncating. All of the biochemical markers of bone turnover tested were significantly higher in those with lesions as compared with those who did not have lesions ( Table 2).
The sensitivity of ALP and other biochemical markers in detecting individuals lesions was generally poor, although specificity was above 80% for all of the markers except uNTX/Creatinine (Table 3).
Information on age at clinical diagnosis of PDB in the probands was available for 12 of the subjects with bone lesions. Kaplan-Meier analysis showed that the estimated mean age at clinical diagnosis in probands was 59.2 years [95% CI 52.7-65.7] as compared with 55.2 years [49.6 -60.9] for the children who had PDB lesions on bone scan, a difference that was not significant (p=0.25) (Figure 3).

Discussion
Genetic factors play a key role in the pathogenesis of PDB, and it has previously been suggested that genetic testing for SQSTM1 mutations and other susceptibility alleles might be used clinically to detect people with the disease (11) . The ZiPP study is unique in that it comprised a large-scale programme of genetic testing for SQSTM1 mutations in both probands and relatives coupled with the offer of targeted intervention to prevent the development of PDB in relatives that tested positive for SQSTM1 mutations. To our knowledge the acceptability of genetic testing for susceptibility to PDB has not previously been explored in this context. Our study showed that uptake of the offer of genetic testing was reasonably good in both probands (59%) and relatives (57%) indicating that this approach is acceptable to the target population and could be implemented in clinical practice in the future, should the ZiPP study show evidence of benefit of zoledronic acid in slowing or preventing the emergence of PDB in this high risk group.
Analysis of the baseline data confirmed the relevance of SQSTM1 mutations as a risk factor for PDB reflected by the fact that 20/222 (9.0%) of participants in the study already had evidence of early PDB on bone scan at the baseline visit. In most cases, the imaging findings were consistent with mild disease but one individual had already developed a striking deformity of two cervical vertebrae at the time of enrolment into the study ( Figure 2B). Interestingly, this participant, in common with other participants, was asymptomatic and had not experienced symptoms related to PDB.
Measurements of ALP were performed in all subjects at the baseline visit along with other markers of bone turnover. All biochemical markers of bone turnover that we measured were significantly higher in people with lesions as compared with those without lesions but the sensitivity of these markers in detecting people with lesions was poor. While previous studies have shown that markers such as ALP, BAP and PINP are elevated in a high proportion of individuals with established PDB, most of the participants in this study had very early disease which could explain why only a modest proportion of people with lesions had elevated marker levels. Whatever the reason it's clear that screening with biomarkers is not an effective method of detecting lesions in SQSTM1 mutation carriers.
Our findings are of interest in relation to the previous study by Bolland and colleagues who screened for evidence of PDB by scintigraphy in the children of individuals from 10 kindreds with familial PDB caused by SQSTM1 mutations (12,13) . In their initial study 4/23 (17%) offspring with SQSTM1 mutations had evidence of PDB on bone scintigraphy at an average age of 45 years (12) .
Subsequent follow up of this cohort 5 years later (13) revealed the emergence of PDB in two carriers who did not have evidence of the disease at baseline. Another study by Peeters evaluated 14 SQSTM1 mutation carriers from the Netherlands who were followed up for periods of up to 15 years (14) . The design of this study differed from that reported here since scintigrams were only performed in participants who had raised levels of ALP or PINP or both. Of the 8 subjects that underwent scintigraphy, new onset PDB was identified in one case (12.5%). The data presented here show that PDB lesions can be present on scintigraphy even though ALP and PINP values are normal and so it is possible some of the subjects studied by Peeters who did not undergo scintigraphy could have had occult lesions in the absence of biochemical abnormalities.
In contrast to the findings reported by Bolland (12) and Cundy (13) who used similar methodology to that reported here, we did not observe a delayed onset of PDB in the children of patients who had SQSTM1 mutations. It is important to point out however that in this study and previous studies, individuals were actively screened for PDB by radionuclide bone scanning and it is quite likely that several years may have elapsed before the diagnosis may have been picked up in routine clinical care. We think that the lower proportion of individuals with lesions observed in this study as compared with Bolland's study was we enrolled individuals who tested positive for SQSTM1 mutations irrespective of whether other family members were affected,whereas Bolland focused exclusively on nuclear families with SQSTM1 mutations.
The present study attests to the fact that PDB is a clinically silent condition in the early stages and may remain so for many years. This is in keeping with the findings of another study which revealed that in about 40% of individuals, PDB first presents with a complication of the disease such as deformity, fracture or deafness (15) . While nitrogen-containing bisphosphonates are highly effective at supressing elevated bone turnover in PDB for prolonged periods (16,17) and are effective at treating bone pain associated with PDB (18,19) the PRISM and PRISM-EZ trials showed that longterm supression of bone turnover with bisphosphonates carried no advantage over symptomdirected treatment in preventing complications of PDB such as progression of deafness, osteoarthritis. quality of life, or pathological fractures (20,21) . The results of these studies provided the impetus for undertaking the ZiPP trial which seeks to determine if intervening at a very early stage in the disease process may have a more favourable outcome.
The ZiPP study has confirmed the importance of SQSTM1 mutations as a genetic risk factor for PDB and has shown that a programme of genetic testing coupled to therapeutic intervention is both acceptable and feasible in the target population. Given the fact that most participants with lesions had early stage disease it is likely that longer term follow up of this cohort will be required to further investigate the effects of the intervention on symptoms and complications. This is being achieved through the ZiPP-long term extension (ZiPP-LTE) study (NCT03859895) in which participants who completed ZiPP will offered follow up for a further five years.  .

Figure 1. Recruitment into the ZiPP study
Disposition of probands and relatives of probands undergoing genetic testing for SQSTM1 mutations with the ZiPP study population. The numbers and proportion of subjects at each stage is indicated.