Gene Variants Related to Primary Familial Brain Calcification: Perspectives from Bibliometrics and Meta-Analysis

Discussion

This integrative analysis identifies SLC20A2 and MYORG as predominant genetic contributors to PFBC, with distinct neuropsychiatric manifestations (cognitive deficits, psychiatric symptoms) representing hallmark clinical features. The mid-adult disease onset pattern reinforces PFBC's characterization as a neurodegenerative disorder with presymptomatic calcification progression. These findings position SLC20A2 and MYORG variants as critical molecular markers while underscoring the need to disentangle mechanisms underlying its striking phenotypic heterogeneity.

In the keyword network analysis, SLC20A2 exhibited the largest node area among all related genes. Moreover, in the meta-analysis, variants associated with SLC20A2 showed the highest detection rates, and patients carrying these variants had significantly higher TCS scores. The pathogenic effect of SLC20A2 variants remains unclear, but one possible explanation is that variants in SLC20A2 disrupt the normal function of PiT-2, leading to phosphate homeostasis imbalance. This causes excessive phosphate absorption by brain cells, particularly astrocytes, resulting in calcium phosphate crystal deposition in brain tissue. These calcifications can accumulate in various regions of the brain, including the basal ganglia (Wallingford et al., 2017; Zhang et al., 2023).

The XPR1 gene also contributes to pathogenesis by directly impairing phosphate transport in brain cells and inhibiting the storage of phosphate by suppressing the intracellular accumulation of polyphosphates (Mailer et al., 2021; Xu et al., 2023). In addition, in our meta-analysis we observed that XPR1 was associated with lower TCS scores. The differential impact of XPR1 and SLC20A2 variants on calcification severity may stem from their distinct roles in phosphate homeostasis. Although both genes encode phosphate transporters, SLC20A2 (PiT-2) mediates constitutive cellular phosphate uptake across the blood–brain barrier and the choroid plexus (Jensen et al., 2016), thereby directly affecting the clearance of phosphate from the cerebrospinal fluid. Consequently, complete loss-of-function mutations in SLC20A2 may result in systemic phosphate dysregulation, predisposing patients to widespread calcification. In contrast, XPR1 primarily regulates phosphate efflux via its plasma membrane localization, functioning as a “phosphate sensor” that modulates cellular output through a feedback mechanism (Legati et al., 2015; Anheim et al., 2016). This functional specialization suggests that XPR1 variants may retain partial phosphate efflux capacity through residual transporter activity or compensatory pathways, thereby potentially limiting the calcification burden.

In addition to the phosphate metabolism pathway, some studies suggest that the pathogenic effects of variants in other genes may be related to damage to the NVU. The NVU, as the smallest functional unit of the brain, plays a crucial role in regulating cerebral blood flow and maintaining the integrity of the BBB. Dysfunction of the NVU can disrupt normal exchanges of substances between the blood and brain parenchyma. If the BBB is compromised, increased permeability may allow high concentrations of phosphate in the blood to infiltrate the brain, with PDGFRB, PDGFB, MYORG, and JAM2 genes potentially affecting BBB integrity (Xu et al., 2023). The dominant genes PDGFB/PDGFRB, which function as a ligand–receptor pair, help maintain neurovascular unit stability by regulating pericyte recruitment; the associated calcification phenotype may be related to local BBB disruption in the basal ganglia (Keller et al., 2013; Nicolas et al., 2013b). In contrast, the recessive genes MYORG and JAM2 impair the BBB through different mechanisms: MYORG regulates the astrocyte-mediated transport of endothelial tight junction proteins (Yao et al., 2018), while JAM2, being closely associated with tight junctions, directly maintains the integrity of the paracellular barrier (Yang et al., 2025). It is noteworthy that both “NVU” and “PiT-2” also appear as important nodes in the keyword network, marking significant directions for future research.

According to the results of the meta-analysis, the prevalence of cognitive disorders among PFBC patients was found to be the highest (45.3%) among all clinical phenotypes, followed by psychiatric disorders (30.8%). In the keyword network, “dementia” was also identified as an important node. One possible explanation for this is that excessive extracellular phosphate can lead to tissue toxicity (Hong et al., 2015) and neuroinflammation (Brown, 2020), which in turn damages surrounding brain tissue and impairs normal brain functions, causing common neurological and psychiatric symptoms in PFBC patients, such as memory loss, mood swings, and behavioral abnormalities (Suescun et al., 2019; Dunn et al., 2020). Additionally, “Parkinsonism” was also an important node in the keyword network. Although our results show that the prevalence of movement disorders was relatively low, the majority of studies in the meta-analysis reported on the occurrence of movement disorders, making this an important focus in PFBC research. Movement disorders may arise from motor neuron damage, abnormal neural conduction, or lesions in specific brain regions caused by brain calcification, especially when neural pathways between the basal ganglia and the cerebral cortex are affected. While this symptom is relatively rare among PFBC patients, when present, it significantly impacts the quality of life. Therefore, movement disorders remain an important area of concern in clinical research on PFBC and should be prioritized in early screening and clinical evaluation of the disease.

It is also worth noting that the meta-analysis observed that approximately one-fifth of PFBC patients were asymptomatic, while another one-fifth reported headaches, which is consistent with the rate of asymptomatic cases reported in a 2015 systematic review (Tadic et al., 2015). The genetic features of PFBC exhibit incomplete penetrance, meaning not all individuals carrying pathogenic variants will present with clinical symptoms. This incomplete penetrance may be related to variant type, the functional impact of the variants, and the timing of gene expression. Additionally, with advancing age, PFBC symptoms may gradually manifest, particularly in the context of neurodegeneration. Early carriers may not exhibit noticeable clinical signs due to their younger age or early physiological stage. As they age, the cumulative effects of brain calcification and neurodegenerative changes may promote the onset of clinical symptoms. Lastly, gene interactions, compensatory mechanisms, and gene–environment interactions may also influence the phenotypes of variant carriers. Certain carriers may have genetic interactions that are not yet fully understood, which could partially or fully compensate for the negative effects of the pathogenic variant, thereby preventing the appearance of symptoms. The heterogeneity of PFBC clinical phenotypes presents a challenge for clinical diagnosis and disease screening. Some asymptomatic variant carriers may only develop symptoms later due to specific environmental or physiological triggers, emphasizing the need for more research to support early screening and monitoring for PFBC.

In most of the outcomes from the meta-analyses, significant statistical heterogeneity was observed between studies. This variability may be attributed to differences in study populations, as the ratio of sporadic to familial cases varies across studies. Additionally, methodological heterogeneity in gene sequencing protocols contributes to the variability. Whole genome sequencing (WGS) method is more robust and less prone to missing variants compared with whole exome sequencing (WES; Belkadi et al., 2015). This suggests that future research should prioritize WGS when possible. The Baujat plot identified the study by Chen et al. in 2019 (Chen et al., 2019) as the primary source of heterogeneity, as it reported lower variant detection rates for PDGFRB compared with other studies. This study employed a retrospective cohort design, categorizing patients into a “systemic group” and a “nonsystemic group.” The “systemic group” consisted of 37 probands diagnosed at the Department of Neurology, Second Affiliated Hospital of Zhejiang University School of Medicine, while the “nonsystemic group” comprised 236 probands with cerebral calcification identified by imaging at other study centers. Patients with PFBC in the systemic group were ultimately diagnosed through outpatient clinics due to the presence of neurological symptoms, whereas those in the nonsystemic group were identified with calcification through imaging examinations and did not necessarily have neurological symptoms. It is possible that by including more asymptomatic individuals, this has expanded the study population to a certain extent, thereby introducing heterogeneity. It should be noted that, in addition to study design, the ratio of sporadic to familial cases included may also contribute to the high heterogeneity. For example, in the included studies, the research by Gebus et al. (Gebus et al., 2017) consisted solely of sporadic cases, which significantly differed from other studies (Table 2). Sporadic cases may be subject to substantial selection bias at the time of inclusion, with some sporadic patients potentially being overlooked due to atypical symptoms. For this reason, that study became the primary source of heterogeneity in the meta-analysis of SLC20A2 detection rates.

Our study offers several advantages. First, it is the first meta-analysis to focus on the variant detection rates of various genes and the prevalence of different phenotypes in brain calcification patients. Additionally, we incorporated bibliometric perspectives by quantitatively exploring the characteristics and trends of the studies, and we used a random-effects model to combine the relevant outcomes. Second, we conducted a systematic search of databases and comprehensively included all 18 studies that met the inclusion criteria, all of which exhibited high methodological quality. Despite the high heterogeneity in some outcomes, no significant publication bias was detected between studies, and sensitivity analysis suggested robust results. Lastly, our bibliometric analysis was based on the WOSCC database, which includes complete citation networks for high-quality literature, ensuring the data's quality and authority (Wang and Waltman, 2016; Yeung, 2019).

However, there are several limitations to our study. First, although we used various methods, such as the Baujat plot and “leave-one-out” analysis, to explore the sources of heterogeneity, the exact sources of heterogeneity between studies have not been fully elucidated. Due to the limited number of studies included for each outcome, we were unable to perform meta-regression for further investigation. Additionally, some early studies did not differentiate between primary and secondary brain calcification, and there is a lack of studies from the Arabian Peninsula and Africa. A higher representation of affected individuals in the studies may introduce selection bias, making the combined results less generalizable to other populations. Secondly, certain outcomes were based on a limited number of studies, such as those reporting MYORG variant detection rates and the prevalence of dysarthria phenotypes, which were derived from only three studies, potentially compromising the accuracy of the results. Furthermore, as JAM2 and CMPK2 were discovered later, no studies meeting the inclusion criteria were available for these genes. Earlier identified genes, such as SLC20A2, may introduce estimation bias due to the relatively higher number of reports, leading to higher variant detection rates and TCS scores. Some potential PFBC pathogenic genes have not yet been discovered, and the detection rates of these undiscovered genes in the calcification population cannot be determined. Finally, due to design limitations, we were unable to compare calcification scores and the incidence rates of asymptomatic patients with age, despite evidence suggesting that the clinical expressivity of calcification correlates with age.

Our study has significant clinical implications. First, it provides a deeper understanding of the pathogenic mechanisms and research trends associated with PFBC, offering new insights into the molecular mechanisms of human brain function and neurodegenerative diseases. Clinically, our findings have implications for diagnosis and genetic counseling and provide valuable reference points for assessing prognosis and treatment outcomes. Genetic screening can be used to aid the diagnosis of brain calcification patients, in conjunction with common clinical phenotypic features of PFBC, improving diagnostic accuracy and enhancing patients’ quality of life. Future research should focus on variants related to JAM2, CMPK2, and NAA60, using more clinical data for validation, especially as the pathogenic effect of CMPK2 on PFBC has yet to be independently replicated by other research groups. Further studies should also aim to expand the sample size and geographic scope, explore other potential pathogenic genes and variants, and analyze the associations and differences between different genotypes and phenotypes. Additionally, although current PFBC-related research is concentrated in China and France, it is essential to investigate pathogenic gene variants in brain calcification patients from other regions worldwide. Collaborative efforts between regional research institutions are needed to promote global research in this field. Finally, considering that current research focuses primarily on basic medical disciplines, and the keyword network shows a higher focus on “case report” type studies, future research could explore individual cases to identify potential therapeutic targets and emphasize clinical translational significance. The highly cited literature summarized in our study can provide valuable guidance.