Skip to main content
NCBI home page
Medicine logoLink to Medicine
. 2025 Aug 29;104(35):e44105. doi: 10.1097/MD.0000000000044105

Risk relationship between osteoporosis and plasma proteins

Cai Chen a, Qin Zeng b,*, Qianling Ye c, Futai Jin d
PMCID: PMC12401289  PMID: 40898453

Abstract

Osteoporosis (OP) has gradually become a major public health problem. It is clinically important to elucidate further the pathophysiologic mechanisms that induce OP and to identify more effective therapeutic targets. In the present study, we used Mendelian randomization analysis to assess the causal effects of 4907 plasma proteins on bone mineral density (BMD) outcomes to identify their potential therapeutic targets. The data files of 4907 plasma proteins were downloaded from the deCODE Genetics database, the bone density data files were obtained from the publicly available genome-wide association study database, the single nucleotide polymorphisms of weak instrumental variables in plasma proteins were removed based on the F test value of >10, the inverse variance weighted method was the main statistical method, and MR-Egger analysis was used for the test of diversity and difference. The obtained plasma proteins strongly associated with the disease were enriched by gene ontology (GO), pathway enrichment, and protein interaction networks were constructed in the GeneMANIA database. Based on P < .05 and false discovery rate value < 0.2, plasma proteins with consistent odds ratio values of 5 statistical methods were extracted, and a total of 22 plasma proteins strongly associated with diseases were obtained. The results of GO enrichment analysis showed that the disease-associated plasma proteins were functionally related to the Notch signaling pathway, trabecular morphogenesis, and so on, and their main enrichment was in the Notch signaling pathway and Toll-like receptor signaling pathway. A total of 20 interacting genes were predicted by the GeneMANIA database. Through least absolute shrinkage and selection operator regression, key variables were selected from 22 plasma proteins. There is a causal relationship between plasma proteins and BMD, with proteins such as recombinant pleckstrin homology domain containing family A, member 1, RAB6B, member RAS oncogene family (RAB6B), and UDP-glucose dehydrogenase exacerbating the disease process. In contrast, proteins such as lipopolysaccharide-binding protein, manic fringe homolog, and cartilage adhesion protein can exert a protective effect, and both act as prognostic markers for OP and as potential therapeutic targets.

Keywords: Mendelian randomization, OP, plasma proteins, therapeutic targets

1. Introduction

Osteoporosis (OP) is an age-related bone disease that primarily affects postmenopausal women and is characterized by a systemic condition of decreased bone mineral density (BMD), altered bone microarchitecture, and increased bone fragility, which can lead to fractures.[1,2] The prevalence of OP increases with age, and epidemiologic surveys show that more than 40 million women in the United States have been diagnosed with lowered bone density. As many as 400,000 hip fractures are caused by OP each year.[3,4] With the aging of the population, there will be a sharp increase in the number of patients with OP in the future, which will bring a great economic burden to individuals, families, and society.[5] Therefore, it is of great value to look for risk factors for its occurrence and to target prevention. As a systemic bone metabolic disease, OP often has no obvious clinical symptoms in the early stage and is usually detected only after a fracture occurs. It is imperative to predict OP promptly.[6] For the treatment of OP, medication is currently used.[7] The main drugs used in clinical practice to treat OP are bone formation enhancers, bone resorption inhibitors, and bone mineralizers. In contrast, bisphosphonates, estrogen receptor (selective) modulators, and calcitonin commonly alleviate the symptoms. Although effective, they all have some problems.[8] For example, bisphosphonates are associated with adverse effects such as gastrointestinal reactions, arthralgia, myalgia, and osteonecrosis of the jaw; in addition, some bisphosphonates have been reported to cause a range of nonnegligible side effects, such as an increased risk of venous thromboembolism and esophageal cancer.[8,9] Estrogen receptor (selective) modulators, despite their significant therapeutic effect in the early stages of OP, increase the risk of cardiovascular disease with long-term use.[10] Calcitonin is therapeutic in OP by regulating bone metabolism and inhibiting osteoclast activity. Still, it can also cause adverse effects such as limb pain, circulation disorders, and severe stroke.[11] Given the limitations of the side effects of many of the current drugs for OP, the search for new therapeutic targets for OP will not only help to address the limitations of the current treatments but also provide an essential direction for a deeper understanding of the disease mechanisms, precision medicine, and future drug development.

Blood is the most commonly used medium for clinical analysis, rich in a variety of proteins, and is currently an important biofluid for the study of biomarkers for the diagnosis and prognosis of human diseases.[12] Drugs approved by the US Food and Drug Administration in 2017 all targeted human proteins, and plasma proteins play a central role in a range of biological processes (BPs), as well as in a variety of diseases, according to the results of a survey.[13] Genome-wide association studies (GWAS) are a research strategy to study phenotype–genotype relationships at the population level. Currently, it has identified more than 100,000 single-nucleotide polymorphisms (SNPs) associated with human disease.[14] Plasma proteins are well-studied traits in GWAS because they represent an intermediate layer between genetic variation and disease progression. Several studies have identified protein quantitative trait loci (pQTL) in population-based cohorts and linked their findings to downstream disease risk.[15] Several recent studies have revealed a strong link between plasma proteins and OP. A study found that plasma fibroblast growth factor 23 levels before heart transplantation were positively correlated with BMD scores in the lumbar spine after transplantation.[16] Another study found that intermittent parathyroid hormone treatment enhanced the activity of the insulin-like growth factor system by upregulating pregnancy-associated plasma protein-A expression, thereby promoting bone formation.[17] It has also been proposed that the Annexin A2 protein may reduce BMD by promoting the migration of peripheral blood monocytes to bone-resorbing surfaces and increasing osteoclastogenesis and bone resorption.[18] More comprehensive and systematic evidence is needed to understand their relationship because of the complexity of plasma proteins and the increasing number of plasma proteins associated with OP.

Mendelian randomization (MR) is a novel analytical approach that uses genetic variation as an instrumental variable (IV) to assess the causal relationship between risk factors and outcomes.[19] Compared with traditional observational studies, MR studies have the advantages of large sample sizes, less influence of unknown confounders, and reverse causality. Due to these advantages, MR has been widely used in research in recent years.[20] In the present study, we analyzed the causal association between 4907 plasma proteins and BMD by MR study, and this study may help to reveal the genetic causality between 4907 plasma proteins and OP, which may help in the diagnosis and treatment of OP. The article structure of this paper is shown in Fig. 1.

Figure 1.

Figure 1.

Open in a new tab

Article structure. FDR = false discovery rate, GO = gene ontology, IVW = inverse variance weighted, KEGG = pathway enrichment, LASSO = least absolute shrinkage and selection operator, MR = Mendelian randomization, OR = odds ratio.

2. Materials and methods

2.1. Exposure data sources

The 4907 plasma proteins used as exposure factors in this study stem from those identified by the deCODE Genetics research team in 2021, which utilized plasma proteomics to massively integrate genomic data between gene sequence variants and plasma protein expression levels in 35,559 Icelandic populations and found sequence variants in plasma pQTL and protein levels 18,084 associations between sequence variants in plasma pQTLs, in addition, the team tested plasma protein levels for associations with 373 diseases and other traits, identifying 257,490 associations.[21] The human genetic data for the current study were obtained from the published GWAS in 2021, which had been applied for by the relevant ethical review board before the current research and, therefore, did not require additional ethical approval. This exposure file is from (https://www.decode.com/summarydata) database.

2.2. Sources of data on end results

Whole-body BMD was the outcome factor in this trial, which was derived from a study reported by Medina-Gomez et al[22] in 2018, which contained 56,284 European population samples containing a total of 16,162,733 SNPs.

2.3. Removing weak IVs

IV selection and MR analyses were conducted through 3 hypotheses: IV are strongly associated with exposure factors; IV are independent of confounders; and IV only affect the outcome through exposure factors and are not directly associated with the outcome.[23] SNPs strongly associated (P < 5e−8) were selected as IV, and chain imbalances were removed (set r2 = 0.001 and kb = 10000). The strength of IV was measured using the F test, calculated as F = R2 (N − k − 1)/k(1 − R2), where R2 represents the extent to which IV explains exposure, N represents the sample size of the GWAS study of exposure factors, and k represents the number of SNPs.[24] Screening SNPs with F > 10 to quantify IV strength and indicate a strong correlation.[25] We used the “TwoSampleMR” package and the “ieugwasr” package in the R software (Version 4.3.1) to filter the IVs that met the requirements.

2.4. MR analysis

All plasma protein exposure data files were read with the “VariantAnnotation” “gwasglue” “TwoSampleMR” package in R software, and information was extracted. The ending data ID was “ebi-a-GCST005348.” To assess the accuracy of the results, multiple sensitivity analyses were performed, including the MR-Egger method and the leave-one-out method. The MR-Egger method tests for horizontal multidirectionality and assesses whether IV affects outcomes through pathways other than exposure, and if P > .05 indicates no significant horizontal pleiotropy exists. The leave-one-out method evaluated the effect of individual SNPs on the overall MR estimate. Cochran Q test assessed heterogeneity, with P > .05 indicating no significant heterogeneity. The results were expressed as odds ratios (ORs) and 95% confidence intervals. Scatter plots, forest plots, and funnel plots were generated for key plasma proteins.

2.5. Screening for disease-associated plasma proteins

The results of the MR analysis and the “csv” file of the pleiotropy test were imported into the R language software, based on the inverse variance weighted (IVW) as the main statistical method, supplemented by the weighted median, MR-Egger, simple mode, weighted mode, and other 5 analytical techniques to screen out the same direction of OR value for the 5 analytical methods.

2.6. MR analysis of disease-related plasma proteins

To analyze the risk relationship between disease-related plasma proteins and BMD using MR, all the plasma protein exposure data files and BMD outcome data files were analyzed using the “TwoSampleMR” software package in R language software to obtain the risk relationship between disease-related plasma proteins and OP. A forest plot visualizing the relationship between plasma proteins and diseases was drawn using the “grid, readr, forestploter” software package.

2.7. Gene ontology and pathway enrichment analysis

We will use the “clusterProfiler,” “org. Hs. e.g.. db,” and “enrichplot” packages in the R language to perform gene ontology (GO) and pathway enrichment (KEGG) analysis on plasma proteins closely related to diseases under the screening conditions of P < .05 using the IVW method, and construct a visual bubble plot.

2.8. GeneMANIA database predicts plasma protein interaction genes

Using the database (http://genemania.org/), plasma proteins meeting the IVW test with a P value of <0.05 and the OR values of the 5 analytical methods in the same direction were predicted, and genes with interactions or regulatory roles were screened.

2.9. Least absolute shrinkage and selection operator regression screening feature variables

To further screen the related to disease proteins obtained from MR analysis, we used the least absolute shrinkage and selection operator (LASSO) regression to rank the importance of all feature variables and selected variables with higher importance. First, we downloaded the GSE56815 dataset from the NCBI database (https://www.ncbi.nlm.nih.gov/), which contains 40 control and 40 OP group samples. We took the average gene expression levels with the same symbol and obtained expression data of 17 corresponding genes out of the 22 plasma proteins from GSE56815. The data were homogenized within groups using the R software “base” package. Using the “glmnet” package in R software, we set the cross-validation coefficient to 10 and the regularization ratio to 1 to determine the optimal lambda value and screen key variables.

2.10. Statistical analysis

The results were expressed as OR and 95% confidence intervals, with P < .05 considered statistically significant. Differences between IVs existed if the Cochran Q statistic was P < .05, but not when P > .05. MR-Egger regression intercepts were considered horizontally multivariate if not 0 and P < .05. Conversely, they were considered not horizontally multivariate if P > .05.

3. Results

3.1. Removal of plasma protein files for weak IVs

Based on the original SNPs screening conditions (P < 5e−8, r2 = 0.001, kb = 10 000, F > 10), 31,315 SNPs closely related to the exposure factors were obtained using R language software.

3.2. Acquisition of plasma proteins strongly associated with disease

SNPs with F test value >10 and outcome data ID “ebi-a-GCST005348” were analyzed by MR analysis using R language software; the IVW method was the main statistical method, and the P value was set at <0.05. If the Cochran Q test suggested no heterogeneity, fixed effects IVW was chosen as the primary outcome; otherwise, random effects IVW was selected.[26] A total of 227 plasma proteins related to diseases were obtained. Based on P < .05 and false discovery rate (FDR) value < 0.2, plasma proteins with consistent OR values of the 5 statistical methods were extracted, a total of 22 disease-related key plasma proteins were obtained, and the visualized volcano diagram was constructed. The red in the volcano diagram represents proteins at high risk of disease, and the green color represents proteins at low risk. We screened 5 representative proteins among the high- and low-risk proteins based on significance, respectively, for further discussion. Twenty-two plasma proteins were tested for heterogeneity and pleiotropy with OP; none 19 were heterogeneous or pleiotropy (P > .05). In contrast, Cochran Q statistic for LRP4, FAM210A, and cartilage adhesion protein (CHAD) suggested heterogeneity for the IVs (P < .05). Further assessment of the pleiotropy of IVs using MR-Egger resulted in no evidence of significant pleiotropy (P > .05), and no SNP was found to significantly affect the overall results in the leave-one-out sensitivity analysis, indicating reliable results. The results of the MR analysis are detailed in Table 1, and the volcano map is shown in Fig. 2. Detailed results of the tests for pleiotropy and heterogeneity are shown in Table 2. Scatterplots, forest plots, and funnel plots are shown in Figs. 3 and 4 with Fig. 5.

Table 1.

Results of MR analysis of plasma proteins strongly associated with 22 diseases.

Plasma protein Methodology β Value OR value 95% CI FDR value
LRP4 IVW 0.114 1.121 1.076–1.168 0.000
PLEKHA1 IVW 0.105 1.111 1.064–1.160 0.003
NPTXR IVW −0.058 0.944 0.919–0.969 0.009
F9 IVW −0.291 0.747 0.643–0.868 0.048
TIMP4 IVW −0.061 0.941 0.912–0.971 0.048
FAM210A IVW −0.253 0.776 0.680–0.887 0.056
UGDH IVW 0.076 1.079 1.036–1.123 0.057
EPHB6 IVW 0.089 1.093 1.042–1.146 0.057
DLL4 IVW 0.083 1.087 1.039–1.136 0.057
CHAD IVW −0.448 0.639 0.501–0.814 0.061
ST3GAL6 IVW −0.036 0.964 0.945–0.984 0.066
ICAM2 IVW −0.153 0.858 0.787–0.935 0.078
CXCL10 IVW −0.138 0.871 0.805–0.942 0.078
POMC IVW 0.076 1.079 1.033–1.128 0.096
BPIFA2 IVW −0.073 0.930 0.891–0.971 0.108
ACP2 IVW −0.060 0.942 0.910–0.976 0.108
STAMBP IVW 0.279 1.322 1.121–1.558 0.108
MFNG IVW −0.180 0.835 0.748–0.932 0.148
RAB6B IVW 0.086 1.090 1.034–1.149 0.154
C1orf198 IVW −0.146 0.864 0.789–0.946 0.169
LBP IVW −0.044 0.957 0.931–0.983 0.171
ADAMTSL1 IVW 0.053 1.054 1.020–1.090 0.182
Open in a new tab

CI = confidence interval, FDR = false discovery rate, IVW = inverse variance weighted, MR = Mendelian randomization, OR = odds ratio.

Figure 2.

Figure 2.

Open in a new tab

Volcanic map of plasma proteins strongly associated with 22 diseases. Horizontal coordinates represent the beta values of 22 plasma proteins, vertical coordinates are −log10(P value), and risk factors for BMD are labeled in red and protective factors for BMD in green. BMD = bone mineral density.

Table 2.

Heterogeneity and pleiotropy tests for MR analysis of 22 plasma proteins and OP.

Exposure factor Outcome Heterogeneity test Pleiotropy test
IVW (Cochran Q test) P value MR-Egger (Cochran Q test) P value MR-Egger intercept MR-Egger intercept P value
LRP4 OP 30.629 .022 27.592 .035 −0.007 .203
PLEKHA1 OP 0.176 .916 0.000 .996 −0.004 .747
NPTXR OP 6.259 .618 6.257 .510 0.000 .974
F9 OP 4.297 .117 2.321 .128 0.018 .526
TIMP4 OP 7.676 .742 7.595 .668 0.001 .782
FAM210A OP 10.095 .039 9.969 .019 −0.004 .858
UGDH OP 1.934 .586 1.900 .387 −0.001 .871
EPHB6 OP 4.357 .628 3.508 .622 0.005 .399
DLL4 OP 3.102 .541 1.191 .755 −0.014 .261
CHAD OP 186.579 .000 183.157 .000 0.016 .675
ST3GAL6 OP 23.573 .704 22.070 .734 −0.003 .231
ICAM2 OP 5.175 .270 5.017 .171 0.005 .779
CXCL10 OP 3.858 .426 3.299 .348 0.010 .528
POMC OP 14.403 .420 14.207 .359 0.002 .679
BPIFA2 OP 13.795 .130 11.684 .166 0.006 .264
ACP2 OP 8.460 .294 7.973 .240 0.005 .567
STAMBP OP 3.851 .278 0.665 .717 0.017 .216
MFNG OP 1.597 .450 0.258 .612 −0.009 .454
RAB6B OP 3.094 .542 1.642 .650 0.007 .315
C1orf198 OP 1.128 .569 0.069 .792 0.007 .491
LBP OP 13.875 .309 9.691 .558 0.009 .065
ADAMTSL1 OP 8.881 .448 8.199 .414 −0.005 .438
Open in a new tab

IVW = inverse variance weighted, MR = Mendelian randomization, OP = osteoporosis.

Figure 3.

Figure 3.

Open in a new tab

Scatterplot of causal associations between 6 key plasma proteins and osteoporosis. (A–F) represents PLEKHA1, RAB6B, UGDH, LBP, MFNG, CHAD, and osteoporosis. CHAD = cartilage adhesion protein, LBP = lipopolysaccharide-binding protein, MFNG = manic fringe homolog, MR = Mendelian randomization, PLEKHA1 = recombinant pleckstrin homology domain containing family A, RAB6B = RAB6B, member RAS oncogene family, SNP = single-nucleotide polymorphism, UGDH = UDP-glucose dehydrogenase.

Figure 4.

Figure 4.

Open in a new tab

Forest plot of causal associations between 6 key plasma proteins and osteoporosis. (A–F) represents PLEKHA1, RAB6B, UGDH, LBP, MFNG, CHAD, and osteoporosis. CHAD = cartilage adhesion protein, LBP = lipopolysaccharide-binding protein, MFNG = manic fringe homolog, MR = Mendelian randomization, PLEKHA1 = recombinant pleckstrin homology domain containing family A, RAB6B = RAB6B, member RAS oncogene family, UGDH = UDP-glucose dehydrogenase.

Figure 5.

Figure 5.

Open in a new tab

Funnel plot of causal associations between 6 key plasma proteins and osteoporosis. (A–F) represent PLEKHA1, RAB6B, UGDH, LBP, MFNG, CHAD, and osteoporosis. CHAD = cartilage adhesion protein, LBP = lipopolysaccharide-binding protein, MFNG = manic fringe homolog, MR = Mendelian randomization, PLEKHA1 = recombinant pleckstrin homology domain containing family A, RAB6B = RAB6B, member RAS oncogene family, UGDH = UDP-glucose dehydrogenase.

3.3. MR analysis of 22 strongly correlated plasma proteins

MR analyzed the plasma proteins of 22 diseases based on the IVW method with P < 0.05, and all the “csv” files were obtained, and visual forest plots were constructed by using the installation packages of “grid, readr, forestploter” and so on in the R software. The forest plot is shown in Fig. 6.

Figure 6.

Figure 6.

Open in a new tab

Forest map of plasma proteins strongly associated with 22 diseases. CI = confidence interval, OR = odds ratio.

3.4. GO and KEGG enrichment analysis

Twenty-two disease-associated plasma protein genes were subjected to KEGG and GO enrichment analysis, and a total of 167 BP, 7 cellular components, and 21 molecular function results were obtained based on a P < .05 as the threshold. The results of GO enrichment analysis showed that the biological functions of the genes were mainly involved in the Notch signaling pathway, trabecular morphogenesis, and other functions, and the results of KEGG enrichment analysis showed that the genes were primarily enriched in the Notch signaling, Toll-like receptors (TLRs) signaling, and so forth. The GO bubble diagram is shown in Fig. 7, and the KEGG bubble diagram is shown in Fig. 8.

Figure 7.

Figure 7.

Open in a new tab

Bubble plots for GO enrichment analysis of plasma proteins strongly associated with 22 diseases. Darker colors represent more significant enrichment. GO = gene ontology.

Figure 8.

Figure 8.

Open in a new tab

Bubble map of KEGG enrichment analysis of plasma proteins strongly associated with 22 diseases. Darker colors represent more significant enrichment. KEGG = pathway enrichment.

3.5. Construction of an interaction network of 22 plasma proteins

To further investigate the biological functions mediated by the 22 plasma proteins, we obtained the interacting proteins as well as the functional enrichment by using tools such as coexpression, shared protein structural domains, colocalization, and gene interactions in the GeneMANIA database. The results showed that the 22 plasma proteins mainly interacted with 20 genes, including tissue inhibitor of metalloproteinase 1 (TIMP1), TIMP2, TIMP3, phospholipid transfer protein (PLTP), lunatic fringe (LFNG), and so forth; they were primarily involved in Notch signaling pathway, protein hydrolysis of extracellular domains of membrane proteins, and the regulation of metallopeptidase activity, and so forth. The GeneMANIA interaction network diagram is shown in Fig. 9.

Figure 9.

Figure 9.

Open in a new tab

Network map of 22 strongly related plasma protein interaction genes.

3.6. Selection of key variables for prediction

All risk factors were subjected to LASSO regression for dimensionality reduction, extracting the most important predictive factors and eliminating collinearity to avoid overfitting. The optimal parameter (lambda) in the final LASSO model was selected through cross-validation with lambda.1se chosen as the optimal value (λ = 0.080235124). The LASSO regression results showed that the 6 key predictive variables were UGDH, manic fringe homolog (MFNG), CHAD, RAB6B, member RAS oncogene family (RAB6B), lipopolysaccharide-binding protein (LBP), and recombinant pleckstrin homology domain containing family A (PLEKHA1). Figures 10 and 11, respectively, display the LASSO coefficient path plot and cross-validation curve of the predictors.

Figure 10.

Figure 10.

Open in a new tab

The LASSO coefficient path of predictive factors. LASSO = least absolute shrinkage and selection operator.

Figure 11.

Figure 11.

Open in a new tab

Cross-validation curve (λ.min = 0.050389996, λ.1se = 0.080235124).

4. Discussion

OP has become one of the most prevalent diseases in the world, with serious impacts on people's daily lives and huge socioeconomic implications.[27] Despite its high prevalence, it is often overlooked in its early stages, and the later stages of the disease usually lead to vertebral fractures, progressive spinal deformities, and neurological complications.[28] Therefore, the search for OP-specific diagnostic markers is crucial for exploring the pathogenesis of OP as well as for drug development and treatment. Plasma, one of the most abundant biofluids in the human body, contains 1000s of proteins that reflect the body's state of health, disease processes, and therapeutic responses. Therefore, plasma proteomics has great potential for the diagnosis, treatment, and prognostic assessment of diseases. Molecular research and bioinformatics technologies have developed rapidly in recent decades, and through molecular enrichment to analyze functions, BPs, and cellular components, molecular biology can provide clues for comprehensive and further studies of how genetic variation and coexpression affect protein function and disease progression.[29]

In the present study, we screened 4907 plasma proteins for genetic causality between plasma proteins and BMD using MR and combined with bioinformatics tools and screened plasma proteins that were strongly correlated with BMD by the F test value, with the thresholds set at P < .05, an FDR value of <0.2, and a consistent OR value for all 5 statistical methods, and a total of 22 plasma protein genes were obtained. Among them, STAM-binding protein, LRP4, PLEKHA1, EPHB6, RAB6B, DLL4, UGDH, proopiomelanocortin, and ADAMTS-like 1 were risk factors for OP, whereas ST3 β-galactoside α2,3-sialyltransferase 6, LBP, NPTXR, acid phosphatase 2, TIMP4, BPI fold containing family A member 2, CXC motif chemokine ligand 10, C1orf198, intercellular adhesion molecule 2, MFNG, FAM210A, F9, and CHAD were protective factors for OP.

Twenty-two plasma proteins were analyzed by GO enrichment, and their BPs were mainly enriched in the Notch signaling pathway, trabecular morphogenesis, and so forth. KEGG was primarily enriched on the Notch signaling and TLR signaling pathways. Several reports have demonstrated that the Notch signaling pathway,[30] bone trabecular morphogenesis,[31] and TLR signaling pathway are all closely related to the development of OP.[32] It is well known that the structure and function of bone trabeculae depend on the dynamic balance between osteoblast formation and osteoclast resorption. Osteoblasts dominate the synthesis and mineralization of bone matrix, osteoclasts are responsible for the resorption and renewal of bone tissue, and osteocytes regulate this process by sensing mechanical signals. All 3 work in synergy to maintain bone health, and their imbalance will lead to OP, fractures, and other diseases. The Notch signaling pathway is assembled and triggered by a complex mechanism. First, the Notch receptor protein is synthesized in the endoplasmic reticulum as a type I transmembrane protein and then translocated to the cell surface to form a heterodimer; subsequently, the binding of ligand from the signal-sending cell to the extracellular structural domain of the Notch receptor in the signal-receiving cell, or the activation of the nonligand-dependent Notch receptor, results in the dissociation of the receptor extracellular subunit and its transmembrane subunit, which leads to release of the activated intracellular structural domain of Notch receptors; finally, the intracellular structural domain of Notch receptors enters the nucleus and complexes with other proteins to form a transcription complex, thereby regulating gene transcription. Finally, the Notch receptor intracellular domain enters the nucleus and complexes with other proteins to form a transcription complex that regulates gene transcription.[33] Multiple studies have shown that the Notch pathway has multiple regulatory functions during skeletal development and bone remodeling.[34] Skeletal stem and progenitor cells (SSPC) perform skeletal maintenance and repair. With age, they produce fewer osteoblasts and more adipocytes, leading to loss of skeletal integrity. In contrast, Notch regulates SSPC cell fate decisions, and modulation of Notch signaling improves the skeletal aging phenotype and increases bone mass even beyond that of young mice.[35] TLRs can specifically recognize ligands, which subsequently induce the expression of a variety of genes and activate intracellular innate immunity.[36] TLRs, as typical representatives of intrinsic immune pattern recognition receptors, are widely involved in the recognition of pathogen-related molecular patterns and the perception of endogenous danger signal-related molecular patterns, which in turn initiate intrinsic immune responses.[37] Bone homeostasis is maintained due to the interplay of osteoblast-involved osteogenesis and osteoclast-involved bone resorption, and previous reports have confirmed that TLRs are abundantly expressed on the surface of osteoblasts, osteoclasts and osteoclasts and are involved in mediating bone loss.[38,39] It has been found that a reduction in TLR9, a member of the TLR family, significantly increases BMD, bone volume/tissue volume, and trabecular number, thereby promoting fracture recovery.[40]

GeneMANIA showed that TIMP1, TIMP2, TIMP3, PLTP, and LFNG interacted strongly with the plasma proteins obtained in this study. TIMP1, TIMP2, and TIMP3 have been reported to be involved in bone metabolism, and the inhibition of TIMP1 expression accelerated the apoptosis of osteoblasts and promoted OP.[41–44] It is well known that OP-induced bone loss must be accompanied by an increase in bone marrow adipose tissue, and inhibition of the bone marrow adipose tissue replacement process is the most promising therapeutic modality for OP, and elevated PTLP expression promotes fatty tissue production,[45] Despite the lack of studies directly correlating PTLP with OP, I cannot ignore its great potential. As one of the molecules in the Notch signaling pathway, it has been demonstrated that messenger RNA levels of Notch pathway molecules, such as LFNG, Notch1, and Nfatc1 were elevated in ovariectomized mice and that Icariin significantly reduced the expression of these genes and inhibited the Notch signaling pathway, and promoted osteogenic differentiation in vitro as well as attenuated the onset of OP in mice.[46] Based on the predicted interactions, we hypothesized that the 22 plasma proteins TIMP4, DLL4, intercellular adhesion molecule 2, UGDH, MFNG, acid phosphatase 2, and BPI fold containing family A member 2 may mediate OP by regulating the genes TIMP1, TIMP2, TIMP3, PLTP, and LFNG.

Six key plasma proteins were further selected using LASSO regression.PLEKHA1 was found to be closely related to the phosphatidylinositol-3-kinase and protein kinase B signaling pathway, which plays an important role in OP.[47,48] Yuan et al[49]confirmed that PLEKHA1 is one of the most important network hub genes for OP identified in their experimental study and that it highly interacts with JAG1, an OP-related gene that has been reported to be involved in normal trabecular bone formation. However, no experimental studies have confirmed the direct targeting of PLEKHA1 for OP. However, Hu et al[49]integrated multi-omics data and similarly found that PLEKHA1 is a key pleiotropic gene for OP development.

CHAD is a cartilage matrix protein that mediates the adhesion of detached chondrocytes. Its protein core consists of 11 leucine-rich repeats flanked by cysteine-rich structural domains.CHAD interacts importantly with collagen as well as with cell surface heparan sulfate proteoglycans and α2β1 integrins.[50] Capulli et al[51] observed lower CHAD messenger RNA and protein in bone sample biopsies from osteoporotic women and ovariectomized mice between the ages of 50 and 65 years, which has potential translational implications for the treatment of OP by decreasing osteoclasts and bone resorption through cyclic CHAD without affecting osteoblast parameters and bone formation. In this study, GO enrichment analysis suggests that CHAD is associated with trabecular morphogenesis, potentially influencing the progression of OP by regulating trabecular morphogenesis.

LBP is a key inflammatory marker whose main function is to bind lipopolysaccharide and promote its clearance.[52,53] Research findings indicate that LBP can enhance the binding of lipopolysaccharide to CD14 and TLR4 receptors, regulate cytokines, including tumor necrosis factor alpha, and induce osteoclast formation and function.[54,55] In this study, LBP was considered a protective factor for OP, showing a negative correlation with the development of OP. In a recent study by Cleminson et al[56], their team conducted a follow-up study on BMD data of 1149 men. Although they found a negative correlation between LBP and BMD in the mid-forearm, no significant associations were observed in the spine, whole body, or other sites. Further research is needed to explore the role of LBP in skeletal health.

The human MFNG gene is located on chromosome 22, band q13.1, encoding a protein with a relative molecular mass of 36,000. Currently, there are limited studies on its relationship with bone metabolism. The protein encoded by MFNG modifies EGF, which can bind to DLL1 to activate Notch1 while inhibiting the activation of Notch1 by Jag1.[57–59] Previous studies have found that the Notch1 signaling pathway plays a crucial role in the pathophysiological process of OP by regulating the dynamic balance between osteoblasts and osteoclasts. RAB6B is an isomer protein of the RAB6 subfamily, primarily located in the Golgi complex, and has been found to be associated with the development of various cancers.[60] Studies have shown that RAB6B may be involved in the regulation of the inflammatory factor tumor necrosis factor alpha, which participates in osteoclast differentiation and affects the balance between bone formation and resorption.[47,61] UGDH is a cytoplasmic hexameric enzyme that converts UDP-glucose to UDP-glucuronic acid, a key reaction in hormone and exogenous metabolism as well as in the production of extracellular matrix precursors, which plays an important role in a wide range of human cancers including lung, breast, esophageal, hepatocellular, prostate, ovarian, colorectal, and melanoma.[62] RAB6B and UGDH have been extensively studied in cancer research, while MFNG is mainly found to be associated with Notch1. No studies have been reported on these 3 genes in OP. In contrast, this study identified EPHB6 and UGDH as risk factors for OP development. Conversely, MFNG is considered a protective factor requiring further research and clinical validation.

Although this study's plasma protein exposure data were derived from European populations, the genetic mechanisms and causal relationships provide important references for OP research in Asian populations. Due to the racial specificity of genetic variants, the effectiveness of IVs, and limitations of cross-population causal inference, the results of this study are partially applicable to Asian populations. However, the findings of this study also have multifaceted promoting effects on OP research in Asian populations. This study can serve as one of the starting points for Asian research, guiding research directions, providing methodological references for the relationship between plasma proteins and OP in Asian populations, revealing racial differences, promoting research on Asian-specific mechanisms, facilitating multi-omics and interdisciplinary integration, advancing clinical translation and precision medicine, and promoting drug development.

5. Conclusion

In this study, we used an integrative approach to reveal a genetic susceptibility association between plasma proteins and OP risk. We obtained 22 plasma proteins that can be used as diagnostic and therapeutic targets for OP, which have great potential for OP diagnosis and treatment but need to be subsequently and continuously validated in the clinic. Our findings provide new insights into the diagnosis and treatment of OP and highlight potential targets that should be prioritized in the future in terms of early diagnosis and therapeutic strategies for OP.

This study has some limitations, as all GWAS data were derived from European populations. It is worth exploring whether similar genetic variants exist in other populations, and large databases should be utilized in the future to obtain cross-ethnic GWAS data to improve the generalizability of the findings. Second, the MR analysis in this study set slightly relaxed FDR values in the hope that more disease-related plasma proteins could be screened out, which is more common in other studies, and the F-statistic values of all selected SNPs were >10, suggesting that the IV was sufficiently robust. Finally, the level of evidence for MR was higher than that of traditional observational studies but still lower than that of randomized controlled trials, and its results need to be experimentally validated to exclude methodological limitations, confirm biological mechanisms, and provide a basis for clinical translation.

Author contributions

Conceptualization: Cai Chen, Qin Zeng.

Data curation: Cai Chen, Qianling Ye, Futai Jin.

Formal analysis: Cai Chen, Qin Zeng, Qianling Ye.

Funding acquisition: Cai Chen.

Methodology: Cai Chen, Qianling Ye.

Software: Cai Chen.

Writing – original draft: Cai Chen.

Writing – review & editing: Cai Chen, Qin Zeng, Qianling Ye.

Abbreviations:

BMD
bone mineral density
BP
biological process
CHAD
cartilage adhesion protein
FDR
false discovery rate
GO
gene ontology
GWAS
genome-wide association study
IVW
inverse variance weighted
KEGG
pathway enrichment
LASSO
least absolute shrinkage and selection operator
LBP
lipopolysaccharide-binding protein
LFNG
lunatic fringe
MFNG
manic fringe homolog
MR
Mendelian randomization
OP
osteoporosis
OR
odds ratio
PLEKHA1
recombinant pleckstrin homology domain containing family A
PLTP
phospholipid transfer protein
RAB6B
RAB6B, member RAS oncogene family
SNPs
single nucleotide polymorphisms
TIMP1
tissue inhibitor of metalloproteinase 1
TLR
Toll-like receptor
UGDH
UDP-glucose dehydrogenase.

The authors have no funding and conflicts of interest to disclose.

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

How to cite this article: Chen C, Zeng Q, Ye Q, Jin F. Risk relationship between osteoporosis and plasma proteins. Medicine 2025;104:35(e44105).

Contributor Information

Cai Chen, Email: chenc19941016@163.com.

Qianling Ye, Email: ye19941001@163.com.

Futai Jin, Email: 18702637782@163.com.

References

  • [1].De Martinis M, Sirufo MM, Nocelli C, Fontanella L, Ginaldi L. Hyperhomocysteinemia is associated with inflammation, bone resorption, vitamin B12 and folate deficiency and MTHFR C677T polymorphism in postmenopausal women with decreased bone mineral density. Int J Environ Res Public Health. 2020;17:4260. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [2].Zhang W, Gao R, Rong X, et al. Immunoporosis: role of immune system in the pathophysiology of different types of osteoporosis. Front Endocrinol. 2022;13:965258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [3].Bao C, Wu T, Zhu S, et al. Regulation of cholesterol homeostasis in osteoporosis mechanisms and therapeutics. Clin Sci (Colch). 2023;137:1131–43. [DOI] [PubMed] [Google Scholar]
  • [4].Wang S, Yuan Y, Lin Q, et al. Antiosteoporosis effect of tanshinol in osteoporosis animal models: a systematic review and meta-analysis. Front Pharmacol. 2022;13:937538. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [5].Adami G, Fassio A, Gatti D, et al. Osteoporosis in 10 years time: a glimpse into the future of osteoporosis. Ther Adv Musculoskelet Dis. 2022;14:1759720X221083541. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [6].Hernlund E, Svedbom A, Ivergård M, et al. Osteoporosis in the European Union: medical management, epidemiology and economic burden. A report prepared in collaboration with the International Osteoporosis Foundation (IOF) and the European Federation of Pharmaceutical Industry Associations (EFPIA). Arch Osteoporos. 2013;8:136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [7].Long SW, Li SH, Li J, et al. Identification of osteoporosis ferroptosis-related markers and potential therapeutic compounds based on bioinformatics methods and molecular docking technology. BMC Med Genomics. 2024;17:99. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [8].Yan K, Zhang RK, Wang JX, et al. Using network pharmacology and molecular docking technology, proteomics and experiments were used to verify the effect of Yigu decoction (YGD) on the expression of key genes in osteoporotic mice. Ann Med. 2025;57:2449225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [9].Langdahl BL. Overview of treatment approaches to osteoporosis. Br J Pharmacol. 2021;178:1891–906. [DOI] [PubMed] [Google Scholar]
  • [10].Hashemzadeh M, Romo R, Arreguin JM, Movahed MR. The effects of estrogen and hormone replacement therapy on cardiovascular systems. Future Cardiol. 2021;17:347–53. [DOI] [PubMed] [Google Scholar]
  • [11].Eastman K, Gerlach M, Piec I, Greeves J, Fraser W. Effectiveness of parathyroid hormone (PTH) analogues on fracture healing: a meta-analysis. Osteoporos Int. 2021;32:1531–46. [DOI] [PubMed] [Google Scholar]
  • [12].Cavalcante JDS, de Almeida DEG, Moraes MS, et al. Challenges and opportunities in clinical diagnostic routine of envenomation using blood plasma proteomics. Toxins. 2023;15:180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [13].Chen L, Peters JE, Prins B, et al. Systematic Mendelian randomization using the human plasma proteome to discover potential therapeutic targets for stroke. Nat Commun. 2022;13:6143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [14].Cui Y, Peng F, Wang D, et al. 3’aQTL-atlas: an atlas of 3’UTR alternative polyadenylation quantitative trait loci across human normal tissues. Nucleic Acids Res. 2022;50:D39–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [15].Schmitz D, Li Z, Lo Faro V, et al. Copy number variations and their effect on the plasma proteome. Genetics. 2023;225:iyad179. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [16].Löfdahl E, Ahmed S, Ahmed A, Rådegran G. Plasma biomarkers for clinical assessment of bone mineral density in heart transplanted patients-a single-center study at Skåne University Hospital in Lund. Transpl Int. 2022;35:10161. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [17].Clifton KB, Conover CA. Pregnancy-associated plasma protein-A modulates the anabolic effects of parathyroid hormone in mouse bone. Bone. 2015;81:413–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [18].Deng FY, Lei SF, Zhang Y, et al. Peripheral blood monocyte-expressed ANXA2 gene is involved in pathogenesis of osteoporosis in humans. Mol Cell Proteomics. 2011;10:M111.011700. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [19].Shen J, Zhou H, Liu J, et al. A modifiable risk factors atlas of lung cancer: a Mendelian randomization study. Cancer Med. 2021;10:4587–603. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [20].Wang J, Yang M, Xu K, et al. The causal associations between growth factors and constipation: a two-sample Mendelian randomization study. Front Physiol. 2023;14:1204146. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [21].Ferkingstad E, Sulem P, Atlason BA, et al. Large-scale integration of the plasma proteome with genetics and disease. Nat Genet. 2021;53:1712–21. [DOI] [PubMed] [Google Scholar]
  • [22].Medina-Gomez C, Kemp JP, Trajanoska K, et al. Life-course genome-wide association study meta-analysis of total body BMD and assessment of age-specific effects. Am J Hum Genet. 2018;102:88–102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [23].Meng H, Jiang L, Song Z, Wang F. Causal associations of circulating lipids with osteoarthritis: a bidirectional mendelian randomization study. Nutrients. 2022;14:1327. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [24].Palmer TM, Lawlor DA, Harbord RM, et al. Using multiple genetic variants as instrumental variables for modifiable risk factors. Stat Methods Med Res. 2012;21:223–42. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [25].Pierce BL, Ahsan H, Vanderweele TJ. Power and instrument strength requirements for Mendelian randomization studies using multiple genetic variants. Int J Epidemiol. 2011;40:740–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [26].Sun M, Yang H, Hu Y, et al. Differential white blood cell count and epigenetic clocks: a bidirectional Mendelian randomization study. Clin Epigenetics. 2024;16:118. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [27].Wang J, Huang Y, Guo L, Li J, Zhou S. Molecular mechanism of Achyranthis bidentatae radix and Morindae officinalis radix in osteoporosis therapy: an investigation based on network pharmacology, molecular docking, and molecular dynamics simulations. Biochem Biophys Rep. 2023;36:101586. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [28].Chaddha R, Agrawal G, Koirala S, Ruparel S. Osteoporosis and vertebral column. Indian J Orthop. 2023;57(Suppl 1):163–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [29].Lv JH, Hou AJ, Zhang SH, et al. WGCNA combined with machine learning to find potential biomarkers of liver cancer. Medicine (Baltimore). 2023;102:e36536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [30].Amroodi MN, Maghsoudloo M, Amiri S, et al. Unraveling the molecular and immunological landscape: exploring signaling pathways in osteoporosis. Biomed Pharmacother. 2024;177:116954. [DOI] [PubMed] [Google Scholar]
  • [31].Li X, Wang Z, Chen Y, et al. Regulation of monocyte polarization through nuclear factor Kappa B/inhibitor of Kappa B alpha pathway by Cuscuta chinensis Lam. In postmenopausal osteoporosis. J Ethnopharmacol. 2025;346:119710. [DOI] [PubMed] [Google Scholar]
  • [32].Zhu X, Du L, Zhang L, Ding L, Xu W, Lin X. The critical role of toll-like receptor 4 in bone remodeling of osteoporosis: from inflammation recognition to immunity. Front Immunol. 2024;15:1333086. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [33].Li X, Yan X, Wang Y, Kaur B, Han H, Yu J. The Notch signaling pathway: a potential target for cancer immunotherapy. J Hematol Oncol. 2023;16:45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [34].Ballhause TM, Jiang S, Baranowsky A, et al. Relevance of notch signaling for bone metabolism and regeneration. Int J Mol Sci. 2021;22:1325. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [35].Remark LH, Leclerc K, Ramsukh M, et al. Loss of Notch signaling in skeletal stem cells enhances bone formation with aging. Bone Res. 2023;11:50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [36].Seçme M, Dodurga Y, Demirkan NC, Kaçar N, Günel NS, Açikbaş I. Determination of T-cell clonality and expression profiles of Toll-like receptors signaling pathway genes and related miRNAs in patients with mycosis fungoides. Gene. 2024;891:147825. [DOI] [PubMed] [Google Scholar]
  • [37].Li D, Wu M. Pattern recognition receptors in health and diseases. Signal Transduct Target Ther. 2021;6:291. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [38].Han J, Zheng Q, Cheng Y, et al. Toll-like receptor 9 (TLR9) gene deletion-mediated fracture healing in type II diabetic osteoporosis associates with inhibition of the nuclear factor-kappa B (NF-κB) signaling pathway. Bioengineered. 2022;13:13689–702. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [39].Staab E, Thiele GM, Clarey D, et al. Toll-like receptor 4 signaling pathway mediates inhalant organic dust-induced bone loss. PLoS One. 2016;11:e0158735. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [40].Barreto G, Manninen M, K Eklund K. Osteoarthritis and toll-like receptors: when innate immunity meets chondrocyte apoptosis. Biology (Basel). 2020;9:65. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [41].Bektan Kanat B, Ulugerger Avci G, Bayramlar OF, et al. Matrix metalloproteinases/tissue inhibitors of metalloproteinases ratio: a biomarker of bone resorption in geriatric osteoporosis? Ann Geriatr Med Res. 2023;27:151–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [42].Xu SY, Shi P, Zhou RM. Post-menopausal oestrogen deficiency induces osteoblast apoptosis via regulating HOTAIR/miRNA-138 signalling and suppressing TIMP1 expression. J Cell Mol Med. 2021;25:4572–82. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [43].Yan TB, Li C, Jiao GJ, Wu WL, Liu HC. TIMP-1 suppressed by miR-138 participates in endoplasmic reticulum stress-induced osteoblast apoptosis in osteoporosis. Free Radic Res. 2018;52:223–31. [DOI] [PubMed] [Google Scholar]
  • [44].Wojdasiewicz P, Turczyn P, Lach-Gruba A, et al. The role of rosavin in the pathophysiology of bone metabolism. Int J Mol Sci. 2024;25:2117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [45].Scheideler M, Elabd C, Zaragosi LE, et al. Comparative transcriptomics of human multipotent stem cells during adipogenesis and osteoblastogenesis. BMC Genomics. 2008;9:340. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [46].Xu Y, Li L, Tang Y, Yang J, Jin Y, Ma C. Icariin promotes osteogenic differentiation by suppressing Notch signaling. Eur J Pharmacol. 2019;865:172794. [DOI] [PubMed] [Google Scholar]
  • [47].Zha L, He L, Liang Y, et al. TNF-α contributes to postmenopausal osteoporosis by synergistically promoting RANKL-induced osteoclast formation. Biomed Pharmacother. 2018;102:369–74. [DOI] [PubMed] [Google Scholar]
  • [48].Zheng D, Cui C, Yu M, et al. Coenzyme Q10 promotes osteoblast proliferation and differentiation and protects against ovariectomy-induced osteoporosis. Mol Med Rep. 2018;17:400–7. [DOI] [PubMed] [Google Scholar]
  • [49].Hu Y, Tan LJ, Chen XD, Greenbaum J, Deng HW. Identification of novel variants associated with osteoporosis, type 2 diabetes and potentially pleiotropic loci using pleiotropic cFDR method. Bone. 2018;117:6–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [50].Rämisch S, Pramhed A, Tillgren V, Aspberg A, Logan DT. Crystal structure of human chondroadherin: solving a difficult molecular-replacement problem using de novo models. Acta Crystallogr D Struct Biol. 2017;73(Pt 1):53–63. [DOI] [PubMed] [Google Scholar]
  • [51].Capulli M, Olstad OK, Onnerfjord P, et al. The C-terminal domain of chondroadherin: a new regulator of osteoclast motility counteracting bone loss. J Bone Miner Res. 2014;29:1833–46. [DOI] [PubMed] [Google Scholar]
  • [52].Qiao B, Chen Z, Huang J, et al. Lipopolysaccharide-binding protein as a biomarker in oral and maxillofacial tumors. Oral Dis. 2023;29:892–901. [DOI] [PubMed] [Google Scholar]
  • [53].Milbank E, Díaz-Trelles R, Dragano N, et al. Liver lipopolysaccharide binding protein prevents hepatic inflammation in physiological and pathological non-obesogenic conditions. Pharmacol Res. 2023;187:106562. [DOI] [PubMed] [Google Scholar]
  • [54].Triantafilou M, Triantafilou K. Lipopolysaccharide recognition: CD14, TLRs and the LPS-activation cluster. Trends Immunol. 2002;23:301–4. [DOI] [PubMed] [Google Scholar]
  • [55].Shuto T, Jimi E, Kukita T, Hirata M, Koga T. Granulocyte-macrophage colony stimulating factor suppresses lipopolysaccharide-induced osteoclast-like cell formation in mouse bone marrow cultures. Endocrinology. 1994;134:831–7. [DOI] [PubMed] [Google Scholar]
  • [56].Cleminson JR, Pasco JA, Bortolasci CC, et al. Lipopolysaccharide-binding protein and bone health: data from a population-based sample of men. Osteoporos Int. 2023;34:309–17. [DOI] [PubMed] [Google Scholar]
  • [57].LeBon L, Lee TV, Sprinzak D, Jafar-Nejad H, Elowitz MB. Fringe proteins modulate Notch-ligand cis and trans interactions to specify signaling states. Elife. 2014;3:e02950. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [58].Kakuda S, LoPilato RK, Ito A, Haltiwanger RS. Canonical Notch ligands and fringes have distinct effects on NOTCH1 and NOTCH2. J Biol Chem. 2020;295:14710–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [59].Kakuda S, Haltiwanger RS. Deciphering the fringe-mediated notch code: identification of activating and inhibiting sites allowing discrimination between ligands. Dev Cell. 2017;40:193–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [60].Peng H, Zhu E, Wang J, et al. RAB6B is a potential prognostic marker and correlated with the remolding of tumor immune microenvironment in hepatocellular carcinoma. Front Pharmacol. 2022;13:989655. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • [61].Domínguez Cadena LC, Schultz TE, Zamoshnikova A, et al. Rab6b localizes to the Golgi complex in murine macrophages and promotes tumor necrosis factor release in response to mycobacterial infection. Immunol Cell Biol. 2021;99:1067–76. [DOI] [PubMed] [Google Scholar]
  • [62].Harrington BS, Kamdar R, Ning F, et al. ; Australian Ovarian Cancer Study. UGDH promotes tumor-initiating cells and a fibroinflammatory tumor microenvironment in ovarian cancer. J Exp Clin Cancer Res. 2023;42:270. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Medicine are provided here courtesy of Wolters Kluwer Health

RESOURCES