Sage Journals: Discover world-class research

Abstract

Background

Recent evidence suggests a link between glycoprotein non-metastatic melanoma protein B (GPNMB) and Parkinson's disease (PD) pathogenesis. Although elevated plasma GPNMB levels associated with disease severity have been reported in PD, cerebrospinal fluid (CSF) alterations remain elusive.

Objective

To explore CSF GPNMB alterations and its clinical significance in PD.

Methods

This study enrolled 118 sporadic PD patients and 40 controls. We examined the potential associations between CSF GPNMB levels and the clinical characteristics or biomarkers of neurodegenerative pathogenesis.

Results

PD patients had higher CSF GPNMB levels than controls (p = 0.0159). In the PD group, CSF GPNMB levels correlated with age (age at examination: r_s= 0.2511, p = 0.0061; age at onset: r_s= 0.2800, p = 0.0021) and the severity of motor and cognitive dysfunction (MDS-UPDRS III score: r_s= 0.1998, p = 0.0347; Mini-Mental State Examination score: r_s= −0.1922, p = 0.0370). After correcting for multiple comparisons, the correlation with age at onset remained significant. CSF GPNMB levels were also positively correlated with CSF soluble triggering receptor expressed on myeloid cells 2 (sTREM2) levels in both the PD (r_s= 0.3582, p < 0.0001) and control (r_s= 0.4743, p = 0.0023) groups. Furthermore, multiple regression analysis revealed CSF sTREM2 level as the strongest determinant of CSF GPNMB levels in the PD group (t-value = 3.49, p = 0.0007).

Conclusions

Elevated CSF GPNMB levels, linked with age and microglial activation, may be a valuable marker for understanding the interplay between aging, neuroinflammation, and PD pathology.

Keywords

glycoprotein non-metastatic melanoma protein B parkinson's disease cerebrospinal fluid age soluble triggering receptor expressed on myeloid cells 2

Introduction

Parkinson's disease (PD), the second most prevalent neurodegenerative disease, affects approximately 6.1 million people worldwide.¹ Notably, both the incidence and prevalence of PD have increased in recent years.² In PD patients, a genome-wide association study uncovered several new genetic variants associated with autophagy–lysosomal pathway function, including polymorphisms in the gene encoding glycoprotein non-metastatic melanoma protein B (GPNMB).³ GPNMB is an endogenous type 1 transmembrane glycoprotein that was identified in the context of tissue damage and inflammation, and reportedly affects mitochondrial function⁴ and endo-lysosomal biology.⁵ Several studies have indicated a relationship between single nucleotide polymorphisms in GPNMB and PD development.^6–11 In addition, increased expression of GPNMB was proved in the substantia nigra of PD patients.¹² These above studies indicated a close association between elevated expression of GPNMB and PD pathology.

In addition to its role as a genetic risk factor, evidence proposes that GPNMB participates in the PD pathological process. Previous studies have shown elevated GPNMB protein levels and their association with disease severity and progression, primarily in lysosomal storage disorders.^13–15 Recently, consistent with the autophagy–lysosomal pathway abnormalities demonstrated in PD, an autopsy study found increased GPNMB levels in the substantia nigra of PD patients, and the experimental modeling of lysosomal stress caused elevated GPNMB levels in several brain regions in mice.⁵ Furthermore, GPNMB facilitates the neuronal uptake of alpha-synuclein, a protein that is associated with PD pathology, and elevated plasma GPNMB levels are correlated with more severe PD symptoms.⁶ However, alterations in GPNMB in cerebrospinal fluid (CSF) and their clinical significance in PD have not yet been thoroughly investigated.

In neurodegenerative diseases, including PD, microglial activation plays a crucial role in the inflammatory response, which contributes to disease pathogenesis.¹⁶ This microglial activation is characterized by the upregulation of a subset of genes, including GPNMB.¹⁷ Similarly, the gene encoding the triggering receptor expressed on myeloid cells 2 (TREM2), an innate immune receptor found on microglia,¹⁸ is also highly expressed during microglial activation. Soluble TREM2 (sTREM2) reflects this activation, and its levels in bodily fluids have been shown to correlate with disease severity in neurodegenerative conditions such as Alzheimer's disease (AD)^19,20 and PD.²¹ This suggests its potential as a marker of disease progression. Therefore, measuring sTREM2 levels alongside GPNMB may provide additional insights into the microglial contribution to PD pathology and further elucidate the relationship between neuroinflammation and disease progression, potentially offering a more comprehensive understanding of the mechanisms at play.

The objective of this study was to elucidate the link between GPNMB in the central nervous system and clinical presentations in PD. To do this, we compared CSF GPNMB levels between patients with PD and controls, and examined the relationships between CSF GPNMB levels and comprehensive clinical characteristics (such as age, sex, disease duration, and motor and non-motor manifestations). In addition, we assessed its associations with CSF or plasma biomarkers of neurodegenerative pathogenesis, including CSF sTREM2.^21–24

Materials and methods

Participants

We enrolled 118 consecutive PD patients who were admitted to Fujita Health University Hospital during the period from May 2020 to June 2023. All PD cases met the Movement Disorder Society (MDS) clinical diagnostic criteria for PD.²⁵ Patients with PD were further subdivided into PD with normal cognition (PD-NC), PD with mild cognitive impairment (PD-MCI), and PD with dementia (PDD) groups.²⁶ PD-MCI and PDD were diagnosed according to the MDS Task Force Level I.^27,28 Patients who did not meet the criteria for PD-MCI or PDD were defined as PD-NC patients. Additionally, we enrolled 40 age- and sex-matched participants. We are conducting an ongoing aging registry study that collects CSF from patients without neurodegenerative disease and undergoing epidural anesthesia for surgical treatment of urinary tract stones or benign prostatic hyperplasia in the urology department at Fujita Health University, Japan.

This research received approval from the ethics committee at Fujita Health University Hospital (approval numbers: HM23-205 and HG21-015), and all participants granted written informed consent for their participation, with the option to withdraw from the study.

Clinical evaluation

Motor and non-motor symptoms associated with PD were evaluated using the and the Japanese version of the Movement Disorder Society's Unified PD Rating Scale (MDS-UPDRS). Motor severity was calculated using Hoehn and Yahr and MDS-UPDRS III scores. Cognitive assessments were conducted using the Frontal Assessment Battery, Addenbrooke's Cognitive Examination-Revised (ACE-R), Mini-Mental State Examination (MMSE), and Japanese version of the Montreal Cognitive Assessment. We also assessed subjects using the Parkinson's Disease Questionnaire-39 Summary Index, Geriatric Depression Scale-15, Odor Stick Identification Test for Japanese, Japanese version of the REM Sleep Behavior Disorder Screening Questionnaire, Scales for Outcomes in Parkinson's Disease-Autonomic, Epworth Sleepiness Scale, and Japanese version of the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease. The levodopa equivalent daily dose (LEDD) was calculated using the latest conversion formula.²⁹ All neurological evaluations of PD patients were conducted during the ‘on’ condition. The interval period between clinical evaluation and CSF collection in the PD group was 7.6 ± 16.1 days (mean ± SD, range 0-137). In the control group, clinical evaluation and CSF collection were performed on the same day.

CSF sample preparation

CSF was obtained via standard lumbar puncture during fasting. Within 2 h of collection, CSF specimens were centrifuged at 1500 × g for 10 min at 4°C. They were then aliquoted into 500 µL portions in polypropylene tubes and stored at −80°C until their use in assays.

Measurements of CSF GPNMB and sTREM2 levels

The levels of CSF GPNMB and sTREM2 were measured using commercial enzyme-linked immunosorbent assay kits (DY2550; R&D Systems, Minneapolis, MN, USA, and ab224881; Abcam, Cambridge, UK). For both GPNMB and sTREM2 measurements, CSF samples were diluted by factors of 1 in 8 to obtain optical density measurements within the standard range. Absorbance at 450 nm was determined with the Bio-Rad Benchmark Microplate Reader and Microplate Manager version 5.2.1 (Bio-Rad Laboratories, Hercules, CA, USA). Samples and standards underwent duplicate measurements, and the means of these duplicates were used for statistical analysis. The average values of coefficient of variation across all samples analyzed for GPNMB and sTREM2 were 11.6% and 5.3%, respectively. Assays for CSF GPNMB were performed using the same reagents on the same day. Besides, replicate samples for sTREM2 on different days exhibited high reproducibility (Pearson r = 0.92).

Plasma sample collection and assays of ad-related biomarkers

Of the 118 patients with PD, we conducted measurements of AD-related biomarkers in the plasma of 65 consecutive PD patients admitted to Fujita Health University Hospital during the period from May 2020 to September 2021. To obtain plasma, blood samples were collected by venipuncture using EDTA blood collection tubes, following a fasting period of over 6 h. After collection, these samples were subjected to centrifugation at 1500 × g for 10 min, and 500 µL aliquots of plasma were transferred into polypropylene tubes, and promptly frozen and stored at −80°C until further analysis. Plasma levels of glial fibrillary acidic protein (GFAP), neurofilament light chain (NfL), amyloid-beta 40 (Aβ₄₀), Aβ₄₂, and tau phosphorylated at threonine 181 (pTau-181) were determined by single-molecule array using the Simoa Human Neurology 4-Plex E kit and Simoa pTau-181 V2 Advantage kit (Quanterix, Billerica, MA, USA). Each plasma sample was tested in duplicate.

We also compared the clinical characteristics between the 65 patients whose AD-related biomarkers were measured and the other 53 patients. Only the MDS-UPDRS IV score was notably higher in the measured group; the other clinical characteristics were not significantly different (Supplemental Table 1).

Statistical analysis

Statistical analyses were carried out using JMP software version 16 (SAS Institute, Cary, NC, USA). Statistical significance was defined as p < 0.05. To compare the distribution of sexes, Fisher's exact test was employed. The normality of variables and homoscedasticity were assessed with the Shapiro–Wilk and Levene's tests, respectively. Depending on the adherence to assumptions of normality or homogeneity of variance, comparisons of continuous variables between two groups were performed using Student's t-test or the Wilcoxon rank-sum test. As for the evaluation of CSF GPNMB, we added a generalized linear model for CSF GPNMB levels based on the presence of PD, age at examination, and sex. Spearman's rank correlation test was used to assess the correlations between continuous variables. To correct for multiple comparisons, we used a False Discovery Rate (FDR) correction using the Benjamini-Hochberg Procedure,³⁰ and the cutoff value for FDR was set at 0.05. In addition, we used a generalized linear model to evaluate the effects of the presence of PD, age at examination, and sex on the CSF GPNMB levels. Principal component analysis (PCA) was conducted on CSF GPNMB, CSF sTREM2, and AD-related plasma biomarkers. Principal component (PC) loading plots were used to identify the primary sources of variability among these biomarkers in PD patients. PC score plots were used to visualize individual people clustering based on biomarker profiles. To clarify the main factors related to CSF GPNMB levels, we conducted a multiple regression analysis with several explanatory variables based on simple regression analyses using Spearman's rank correlation test. Continuous variables are presented as the mean ± standard deviation.

Results

Participant characteristics

Table 1 presents the clinical features of participants in the control and PD groups. No notable differences were observed between the two groups in age at examination (p = 0.0845) or sex (p = 0.1661). In the PD group, 77 patients (65.3%) were identified as having PD-MCI, 26 (22.0%) as having PDD, and the remaining 15 (12.7%) as having PD-NC. The proportions of these three subgroups based on cognitive function were consistent with previous reports when disease duration was taken into account.^31,32 No significant differences were found between the PD and control groups in CSF protein (PD 50.23 ± 23.43 mg/dL, control 57.83 ± 37.38 mg/dL; p = 0.1797) or albumin (PD 30.48 ± 16.38 mg/dL, control 33.74 ± 19.92 mg/dL; p = 0.1906) levels.

Table 1.

Clinical characteristics of participants in the PD and control groups.

Characteristics	Control (N = 40) mean ± SD	PD (N = 118) mean ± SD	p
Age at examination (y)	68.25 ± 8.54	70.70 ± 8.44	0.0845
Sex (Male / female)	28/12	68/50	0.1661
Age at onset (y)		62.68 ± 9.82
Disease duration (mo)		91.05 ± 61.97
Duration of education (y)		12.38 ± 2.87
Total LEDD (mg)		639.05 ± 338.63
MDS-UPDRS I		110.38 ± 5.97
MDS-UPDRS II		14.43 ± 8.80
MDS-UPDRS III		35.07 ± 17.06
MDS-UPDRS IV		4.20 ± 4.31
HY		3.10 ± 1.02
PDQ-39 SI		51.08 ± 26.57
SCOPA-AUTO		14.92 ± 8.38
GDS-15		7.09 ± 3.77
J-QUIP		0.63 ± 1.07
RBDSQ-J		5.26 ± 3.18
ESS		8.77 ± 5.65
OSIT-J		3.53 ± 2.60
MMSE		25.60 ± 4.92
ACE-R total score		81.83 ± 16.46
MoCA-J total score		20.84 ± 5.20
FAB		12.56 ± 3.11
CSF protein level (mg/dL)	57.83 ± 37.38	50.23 ± 23.43	0.1797
CSF albumin level (mg/dL)	33.74 ± 19.92	30.48 ± 16.38	0.1906

Significance of continuous variables was tested using the Wilcoxon rank-sum test. Fisher's exact test was used to compare sex distributions. PD: Parkinson's disease; LEDD: Levodopa equivalent daily dose; MDS-UPDRS: Movement Disorder Society's Unified Parkinson's Disease Rating Scale; HY: Hoehn-Yahr scale; PDQ-39 SI: Parkinson's Disease Questionnaire-39 Summary Index; SCOPA-AUT: Scales for Outcomes in Parkinson's Disease-Autonomic; GDS-15: Geriatric Depression Scale-15; J-QUIP: Japanese version of the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease; RBDSQ-J: Japanese version of the REM Sleep Behavior Disorder Screening Questionnaire; ESS: Epworth Sleepiness Scale; OSIT-J: Odor Stick Identification Test for Japanese; MMSE: Mini-Mental State Examination; ACE-R: Addenbrooke's Cognitive Examination-Revised; MoCA-J: Japanese version of the Montreal Cognitive Assessment; FAB: Frontal Assessment Battery; CSF: cerebrospinal fluid.

Comparison of CSF GPNMB concentrations between the PD and control groups

Student's t-test confirmed significantly elevated log-transformed CSF GPNMB levels in the PD group compared with the control group (p = 0.0159) (Figure 1). Additionally, a generalized linear model analysis, including age at examination and sex as covariates, also revealed a significant effect of the presence of PD on CSF GPNMB (p < 0.0001) irrespective of the covariates (Supplemental Table 2).

Figure 1.

Comparison of CSF GPNMB levels. Comparison of log₁₀ CSF GPNMB between the control and PD groups. Significance was tested using Student's t-test. CSF: cerebrospinal fluid; GPNMB: glycoprotein non-metastatic melanoma protein B; PD: Parkinson's disease.

Associations between CSF GPNMB and clinical metrics and CSF biomarkers in each group

Clinical metrics. In the PD group, CSF GPNMB levels exhibited a positive correlation with age at examination (r_s= 0.2511, p = 0.0061), age at onset (r_s= 0.2800, p = 0.0021), and MDS-UPDRS III score (r_s= 0.1998, p = 0.0347) (Table 2). Moreover, in this group, CSF GPNMB levels demonstrated a negative correlation with the duration of education (r_s= −0.2455, p = 0.0074), and MMSE score (r_s= −0.1922, p = 0.0370). Furthermore, to investigate the effect of CSF GPNMB on motor and cognitive function in the PD group, we conducted multiple regression analysis on MDS-UPDRS III score and MMSE score, with CSF GPNMB included as a potential confounder. As shown in Supplemental Table 3, there was a trend towards a correlation between CSF GPNMB and the MDS-UPDRS III score, regardless of whether LEDD was included as a covariate, although this trend was not statistically significant (p = 0.0585, without LEDD; p = 0.0636, with LEDD). Meanwhile, no clear trend was found with the MMSE score.

Table 2.

Correlations between CSF GPNMB levels and clinical parameters or CSF biomarkers in the PD and control groups.

	Control group (N = 40)		PD group (N = 118)
	rs	p	rs	p
Age at examination (y)	0.2071	0.1998	0.2511	0.0061
Sex (Male / female)		0.6687		0.6493
Age at onset (y)			0.2800	0.0021
Disease duration (mo)			−0.0419	0.6520
Duration of education (y)			−0.2455	0.0074
Total LEDD (mg)			−0.0975	0.2938
MDS-UPDRS I			−0.0011	0.9907
MDS-UPDRS II			0.0557	0.5527
MDS-UPDRS III			0.1998	0.0347
MDS-UPDRS IV			0.0479	0.6207
HY			0.1058	0.2669
PDQ-39 SI			−0.0659	0.4822
SCOPA-AUTO			0.0579	0.5388
GDS-15			0.0501	0.5914
J-QUIP			−0.0666	0.4796
RBDSQ-J			0.0354	0.7057
ESS			−0.0461	0.6216
OSIT-J			−0.1132	0.2263
MMSE			−0.1922	0.0370
ACE-R total score			−0.1774	0.0546
MoCA-J total score			−0.1361	0.1433
FAB			−0.1462	0.1157
CSF protein level (mg/dL)	0.0626	0.7013	0.1418	0.1273
CSF albumin level (mg/dL)	0.0184	0.9103	0.1482	0.1123
CSF sTREM2 (pg/mL)	0.4743	0.0023	0.3582	<0.0001

Spearman's rank correlation test was used to determine significant correlations, which are indicated in bold. CSF: cerebrospinal fluid; GPNMB: glycoprotein non-metastatic melanoma protein B; PD: Parkinson's disease; LEDD: Levodopa equivalent daily dose; MDS-UPDRS: Movement Disorder Society's Unified Parkinson's Disease Rating Scale; HY: Hoehn-Yahr scale; PDQ-39 SI: Parkinson's Disease Questionnaire-39 Summary Index; SCOPA-AUT: Scales for Outcomes in Parkinson's Disease-Autonomic; GDS-15: Geriatric Depression Scale-15; J-QUIP: Japanese version of the Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease; RBDSQ-J: Japanese version of the REM Sleep Behavior Disorder Screening Questionnaire; ESS: Epworth Sleepiness Scale; OSIT-J: Odor Stick Identification Test for Japanese; MMSE: Mini-Mental State Examination; ACE-R: Addenbrooke's Cognitive Examination-Revised; MoCA-J: Japanese version of the Montreal Cognitive Assessment; FAB: Frontal Assessment Battery; sTREM2: soluble triggering receptor expressed on myeloid cells 2.

After correcting for multiple comparisons with FDR, the statistical significance between CSF GPNMB levels and age at onset remained significant (q-value = 0.0469) (Supplemental Table 4). In the control group, there were no significant associations between CSF GPNMB levels and age at examination (r_s= 0.2071, p = 0.1998) or sex (p = 0.6687).

CSF biomarkers. Table 2 also shows the associations between CSF GPNMB and CSF biomarkers in the PD and control groups. There was a significant positive correlation between CSF GPNMB and CSF sTREM2 levels in both the PD (r_s= 0.3582, p < 0.0001) and control (r_s= 0.4743, p = 0.0023) groups (Figure 2A, B).

Figure 2.

Correlation between CSF GPNMB and sTREM2 levels in each group. (A) Correlation between CSF GPNMB and sTREM2 levels in the PD group. (B) Correlation between CSF GPNMB and sTREM2 levels in the control group. Spearman's rank correlation test was used to determine significant correlations. CSF: cerebrospinal fluid; GPNMB: glycoprotein non-metastatic melanoma protein B; sTREM2: soluble triggering receptor expressed on myeloid cells 2; PD: Parkinson's disease.

Regression analysis for potential confounders of CSF GPNMB levels in the PD group

Considering the potential confounding variables from the aforementioned Spearman's rank correlation, we conducted multiple regression analysis to identify the determinants of CSF GPNMB levels in the PD group. As depicted in Table 3, the CSF sTREM2 level was the predominant influencer (t-value = 3.49, p = 0.0007), followed by the duration of education (t-value = −2.04, p = 0.0444).

Table 3.

Multiple regression analysis for CSF GPNMB and potential confounding factors in the PD group.

Explanatory variable	Unstandardized Coefficient			Standardized Coefficient	t	p
	Estimate	95% CI
	Estimate	Lower	Upper
Age at examination (y)	−41.19	−209.82	127.4	−0.0862	−0.48	0.6291
Age at onset (y)	78.99	−64.34	222.3	0.1922	1.09	0.2770
Duration of education (y)	−257.6	−508.59	−6.601	−0.1845	−2.04	0.0444
MDS-UPDRS III	33.47	−10.53	77.47	0.1413	1.51	0.1345
MMSE	−96.06	−264.7	72.54	−0.1039	−1.13	0.2611
CSF sTREM2 (pg/mL)	0.1297	0.0559	0.2034	0.3105	3.49	0.0007

Bold type indicates significant correlations. CSF: cerebrospinal fluid; GPNMB: glycoprotein non-metastatic melanoma protein B; PD: Parkinson's disease; CI: confidence interval; MDS-UPDRS: Movement Disorder Society's Unified Parkinson's Disease Rating Scale; MMSE: Mini-Mental State Examination; sTREM2: soluble triggering receptor expressed on myeloid cells 2.

Associations between CSF GPNMB levels and AD-related plasma biomarkers in the PD group

Regarding the observed negative correlation between CSF GPNMB and MMSE, we investigated whether elevated CSF GPNMB level in PD reflects AD pathology, considering the previous reports suggesting the presence of AD pathology in cognitive decline in PD.^33,34 We measured six AD-related plasma biomarkers (Aβ₄₀, Aβ₄₂, Aβ₄₂/Aβ₄₀, GFAP, NfL, and pTau-181), and examined their relationships with CSF GPNMB levels. Our results showed that those plasma biomarkers were not significantly associated with CSF GPNMB levels, except for plasma NfL (rs = 0.2681, p = 0.0308) (Supplemental Table 5). Figure 3 elucidates the interconnections between CSF and plasma biomarkers in the PD group. Spearman's rank correlation coefficients are systematically summarized in a correlation matrix, accompanied by their corresponding p-values (Figure 3A). A notable array of correlations was observed among the six AD-related plasma biomarkers. However, CSF GPNMB and sTREM2 demonstrated independence from those plasma biomarkers, with the exception of plasma NfL.

Figure 3.

Interconnections between CSF and plasma biomarkers in the PD group. (A) Spearman's rank correlation between plasma biomarkers in the PD group. Spearman's rank correlation test was used to determine significant correlations. *p < 0.05; **p < 0.01; ***p < 0.001. (B-D) Results of principal component analysis on CSF GPNMB, CSF sTREM2, and AD-related plasma biomarkers. (B) Proportion of variance for each principal component. Individual proportion was depicted with dotted lines, and cumulative proportion was depicted with bars. (C) PC loadings plot, showing an involvement of each biomarker in PC1 variance and PC2 variance. (D) Superimposed biplot with the PC loadings plot and the PC scores plot, which shows the distribution of individual patients in relation to PC1 and PC2. CSF: cerebrospinal fluid; PD: Parkinson's disease; GPNMB: Glycoprotein non-metastatic melanoma protein B; sTREM2: soluble triggering receptor expressed on myeloid cells 2; PC: principal component; AD: Alzheimer's disease; Aβ: amyloid-beta; GFAP: glial fibrillary acidic protein; NfL: neurofilament light chain; p-Tau181: tau phosphorylated at threonine 181.

PCA revealed two principal components. The PC loading plot demonstrated that the primary contributors to PC1 variance were plasma NfL, p-tau 181, GFAP, and Aβ₄₀. Conversely, CSF GPNMB and sTREM2 emerged as the primary determinants of PC2 variance, suggesting the independence of CSF GPNMB and sTREM2 from AD-related biomarkers in our PD patient group (Figure 3B-D).

Discussion

In the present study, CSF GPNMB levels were markedly elevated in the PD group compared with those in the control group. Further analysis in the PD group revealed that this elevation was significantly positively correlated with patient age and severity of motor and cognitive function. The age at onset remained significant after correction for multiple comparisons with FDR. A significant association was also observed between CSF GPNMB and sTREM2 levels in the PD group. This association was independent of AD-related plasma biomarkers except for NfL. Taken together, our study reveals that elevated CSF GPNMB levels in PD patients are associated with age at onset and show a unique correlation with sTREM2, while remaining independent of AD-related biomarkers, underscoring their potential relevance in the distinct pathology of PD.

Human GPNMB is expressed in the microglia of the central nervous system. The role of GPNMB in inflammation and its involvement in the pathology of neurodegenerative diseases have been underscored in several publications.^17,35,36 Experiments involving GPNMB overexpression in transgenic mouse models have provided evidence of its neuroprotective and anti-inflammatory effects.^35,37,38 Autopsy studies have shown that microglia in the midbrains of patients with PD have a pro-inflammatory trajectory, including elevated GPNMB levels.^5,39 Consequently, the specific role of GPNMB in inflammation in PD remains unclear. Thus, further research is necessary to investigate the potential of GPNMB, especially in the CSF, as an indicator of inflammation in the central nervous system of PD patients. A recent study reported increased GPNMB levels in the plasma but not CSF of PD patients.⁶ This discrepancy with our study's findings might be attributed to variations in sample size and patient backgrounds between the studies. To our knowledge, the present study is the most extensive examination conducted to date on patients with PD. Considering these factors, our extensive cohort not only amplifies the robustness of our findings but also provides a crucial contribution to the existing body of knowledge regarding the role of GPNMB in PD.

Diaz-Ortiz et al. observed a correlation between higher plasma GPNMB levels and more severe PD motor symptoms.⁶ We also investigated those clinical associations of CSF GPNMB, and revealed that CSF GPNMB levels correlated with age at examination and onset, suggesting a possible relationship between GPNMB and aging. In particular, after adjustment for multiple comparisons by FDR, the positive correlation between GPNMB and the age of onset remained significant. Previous evidence has revealed a close relationship between onset at an older age and increased severity of the clinical phenotype of PD.⁴⁰ Moreover, the temporal progression of motor and non-motor milestones in PD is influenced by both disease duration and age at diagnosis.^41–43 Given the well-documented relationships between aging, age at onset, and disease severity in PD,⁴⁴ our study aligns with previous findings that emphasize the role of GPNMB in aging and in the progression of patients with PD.^4,45 Furthermore, the relationship between GPNMB and MDS-UPDRS III or MMSE scores observed in this study could be affected by age-related factors. Alternatively, higher GPNMB levels in older people with PD might be a potential risk factor for the decline in motor and cognitive functions, as reflected in the deterioration of MDS-UPDRS III or MMSE scores. This could be because GPNMB facilitates the faster spread of alpha-synuclein, a protein associated with PD pathology.⁶ In line with this hypothesis, our multiple regression analyses on motor and cognitive functions revealed a trend indicating that CSF GPNMB levels might influence MDS-UPDRS III scores, although this trend was not statistically significant. Future research, particularly studies that evaluate alpha-synuclein, is needed to clarify the precise mechanisms underlying these associations.

Another finding from our study was the negative correlation between CSF GPNMB levels and MMSE score, indicating the potential influence of GPNMB on cognitive function in PD. Considering the complex pathology of cognitive decline in PD, we also investigated the relationship between CSF GPNMB levels and other biological markers that may reflect cognitive function in PD.²⁴ Notably, representative AD-related plasma biomarkers, such as Aβ 40, Aβ42, Aβ42/Aβ40, GFAP, and pTau-181, were not significantly associated with CSF GPNMB levels, except for NfL. As for the association with NfL, a marker of axonal damage which increases in various neurodegenerative diseases,⁴⁶ it might reflect the aspects of both CSF GPNMB and plasma NfL as disease progression biomarkers, involving both motor severity and cognitive decline in PD.²⁴

A noteworthy positive correlation was identified between CSF GPNMB and sTREM2 levels in PD patients. TREM2 serves as an innate immune receptor and is expressed by macrophages and microglia in the central nervous system.^47,48 Additionally, TREM2 promotes microglial activation and is closely correlated with the pathogenesis of neurodegenerative diseases including AD.⁴⁹ Therefore, given its role in promoting microglial survival and inducing inflammatory responses, sTREM2 considered as a biomarker of microglial activation.^47,50,51 Previous studies have shown that CSF sTREM2 levels are elevated in patients with PD and highlighting its potential as a predictor of cognitive decline in this disease.^21,52,53 We found that CSF GPNMB and sTREM2 levels were independent of AD-related plasma biomarkers. This suggests that these microglial changes, as reflected by CSF GPNMB and sTREM2 levels, exhibit different dynamics from AD-related pathology associated with cognitive decline in PD.²⁴ At present, the mechanisms by which GPNMB and TREM2 function cooperatively in the context of PD are unclear. In a previous study on intracellular protein activity in a sarcoma model analyzed by single-cell RNA sequencing, the authors reported that a specific group of TREM2-expressing macrophages also expressed Gpnmb.⁵⁴ Other studies have shown that in PD, microglia are more prone to causing inflammation and have increased GPNMB levels.³⁹ These findings suggested that TREM2 and GPNMB may act together through microglial inflammatory signaling. However, further research is needed to explore the precise mechanisms and roles of GPNMB and TREM2 in PD pathogenesis.

Our study had several limitations. First, the control group lacked information regarding clinical features (except for age and sex) and AD-related plasma biomarkers. Second, we did not conduct a combined analysis of CSF and plasma GPNMB levels to clarify their association with central PD pathology. This decision was influenced by the fact that human GPNMB is expressed in diverse cell types across multiple organs, such as bone and skin. Given the distinct implications of GPNMB in plasma and CSF, our analysis focused solely on CSF GPNMB in this instance. Third, in this study, we did not investigate CSF GPNMB levels in atypical parkinsonian syndromes, such as multiple system atrophy (MSA) and progressive supranuclear palsy (PSP). Recent studies using plasma samples have included MSA and PSP along with PD, revealing that GPNMB levels are significantly lower in MSA compared to PD.^55,56 Future investigations incorporating these conditions are necessary to evaluate the potential specificity of CSF GPNMB as a biomarker for PD. Fourth, we lack longitudinal clinical data and predictive models regarding CSF GPNMB. Fifth, this study did not include genetic testing for GPNMB, which could have contributed to a more comprehensive understanding of its involvement in the disease process. Lastly, the Simoa Advantage plasma Aβ42/Aβ40 assays are not sufficiently reliable to detect Aβ pathology, with low AUCs of ∼0.69 in detecting CSF and PET amyloid positivity.⁵⁷

Conclusions

In the present study, CSF GPNMB was significantly increased in PD, and its levels associated with patient age. Additionally, among known biomarkers related to neurodegenerative pathogenesis, CSF GPNMB levels were closely associated with CSF sTREM2 levels. Our study indicates that elevated CSF GPNMB levels in PD may reflect patient age and its related changes and might be associated with microglial activation.

Supplemental Material

sj-docx-1-pkn-10.1177_1877718X241288712 - Supplemental material for CSF GPNMB in Parkinson's disease: A potential association with age and microglial activation

Supplemental material, sj-docx-1-pkn-10.1177_1877718X241288712 for CSF GPNMB in Parkinson's disease: A potential association with age and microglial activation by Xi-Chen Zhu, Yasuaki Mizutani, Reiko Ohdake, Harutsugu Tatebe, Toshiki Maeda, Sayuri Shima, Akihiro Ueda, Mizuki Ito, Shinji Ito, Takahiko Tokuda and Hirohisa Watanabe in Journal of Parkinson's Disease

Supplemental Material

sj-xlsx-2-pkn-10.1177_1877718X241288712 - Supplemental material for CSF GPNMB in Parkinson's disease: A potential association with age and microglial activation

Supplemental material, sj-xlsx-2-pkn-10.1177_1877718X241288712 for CSF GPNMB in Parkinson's disease: A potential association with age and microglial activation by Xi-Chen Zhu, Yasuaki Mizutani, Reiko Ohdake, Harutsugu Tatebe, Toshiki Maeda, Sayuri Shima, Akihiro Ueda, Mizuki Ito, Shinji Ito, Takahiko Tokuda and Hirohisa Watanabe in Journal of Parkinson's Disease

Footnotes

Acknowledgements

We are grateful to Dr Ryoichi Shiroki and his colleagues in our institute for their help with the collection of CSF samples from healthy controls.

ORCID iDs

Xi-Chen Zhu

Yasuaki Mizutani

Sayuri Shima

Akihiro Ueda

Funding

The present work was supported by JSPS KAKENHI grant numbers JP21K07428 and JP22K07508 and AMED grant number 22dk0207055h0002.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

All data supporting this study's findings are available within the paper and its Supplemental Material. All participants ‘ clinical parameters and biological sample biomarkers are provided in .

Supplemental material

Supplemental material for this article is available online.

References

Bloem

Okun

Klein

. Parkinson's disease. Lancet 2021; 397: 2284–2303.

Dorsey

Sherer

Okun

, et al. The emerging evidence of the Parkinson pandemic. J Parkinsons Dis 2018; 8: S3–S8.

Kumaran

Cookson

. Pathways to parkinsonism redux: convergent pathobiological mechanisms in genetics of Parkinson's disease. Hum Mol Genet 2015; 24: R32–R44.

Cardoso

Fernandes

Aguilar-Pimentel

, et al. Towards frailty biomarkers: candidates from genes and pathways regulated in aging and age-related diseases. Ageing Res Rev 2018; 47: 214–277.

Moloney

Moskites

Ferrari

, et al. The glycoprotein GPNMB is selectively elevated in the substantia nigra of Parkinson's disease patients and increases after lysosomal stress. Neurobiol Dis 2018; 120: 1–11.

Diaz-Ortiz

Seo

Posavi

, et al. GPNMB Confers risk for Parkinson's disease through interaction with alpha-synuclein. Science 2022; 377: eabk0637.

Kia

Zhang

Guelfi

, et al. Identification of candidate Parkinson disease genes by integrating genome-wide association study, expression, and epigenetic data sets. JAMA Neurol 2021; 78: 464–472.

Kaiser

Zhang

Mollenhauer

, et al. A proteogenomic view of Parkinson's disease causality and heterogeneity. NPJ Parkinsons Dis 2023; 9: 24.

Chen

Liaw

Nfor

, et al. Epigenetic regulation of Parkinson's disease risk variant GPNMB cg17274742 methylation by sex and exercise from Taiwan Biobank. Front Aging Neurosci 2023; 15: 1235840.

10.

Liu

, et al. The association analysis of GPNMB rs156429 with clinical manifestations in Chinese population with Parkinson's disease. Front Genet 2020; 11: 952.

11.

Murthy

Blauwendraat

Guelfi

, et al. Increased brain expression of GPNMB is associated with genome wide significant risk for Parkinson's disease on chromosome 7p15.3. Neurogenetics 2017; 18: 121–133.

12.

Neal

Boyle

Budge

, et al. The glycoprotein GPNMB attenuates astrocyte inflammatory responses through the CD44 receptor. J Neuroinflammation 2018; 15: 73.

13.

Kramer

Wegdam

Donker-Koopman

, et al. Elevation of glycoprotein nonmetastatic melanoma protein B in type 1 Gaucher disease patients and mouse models. FEBS Open Bio 2016; 6: 902–913.

14.

Murugesan

Liu

Yang

, et al. Validating glycoprotein non-metastatic melanoma B (gpNMB, osteoactivin), a new biomarker of Gaucher disease. Blood Cells Mol Dis 2018; 68: 47–53.

15.

Zigdon

Savidor

Levin

, et al. Identification of a biomarker in cerebrospinal fluid for neuronopathic forms of Gaucher disease. PLoS One 2015; 10: e0120194.

16.

Kwon

Koh

. Neuroinflammation in neurodegenerative disorders: the roles of microglia and astrocytes. Transl Neurodegener 2020; 9: 42.

17.

Huttenrauch

Ogorek

Klafki

, et al. Glycoprotein NMB: a novel Alzheimer's disease associated marker expressed in a subset of activated microglia. Acta Neuropathol Commun 2018; 6: 108.

18.

Pocock

Vasilopoulou

Svensson

, et al. Microglia and TREM2. Neuropharmacology 2024; 257: 110020.

19.

Zhang

Tang

Yang

, et al. Soluble TREM2 ameliorates tau phosphorylation and cognitive deficits through activating transgelin-2 in Alzheimer's disease. Nat Commun 2023; 14: 6670.

20.

Tan

, et al. Dynamic changes of CSF sTREM2 in preclinical Alzheimer's disease: the CABLE study. Mol Neurodegener 2020; 15: 25.

21.

Wilson

Swarovski

Linortner

, et al. Soluble TREM2 is elevated in Parkinson's disease subgroups with increased CSF tau. Brain 2020; 143: 932–943.

22.

Rauchmann

Sadlon

Perneczky

, et al. Soluble TREM2 and inflammatory proteins in Alzheimer's disease cerebrospinal fluid. J Alzheimers Dis 2020; 73: 1615–1626.

23.

van der Ende

Morenas-Rodriguez

McMillan

, et al. CSF sTREM2 is elevated in a subset in GRN-related frontotemporal dementia. Neurobiol Aging 2021; 103: 158 e151–158 e155.

24.

Mizutani

Ohdake

Tatebe

, et al. Associations of Alzheimer's-related plasma biomarkers with cognitive decline in Parkinson's disease. J Neurol 2023; 270: 5461–5474.

25.

Postuma

Berg

Stern

, et al. MDS Clinical diagnostic criteria for Parkinson's disease. Mov Disord 2015; 30: 1591–1601.

26.

Tang

Han

, et al. Plasma GFAP in Parkinson's disease with cognitive impairment and its potential to predict conversion to dementia. NPJ Parkinsons Dis 2023; 9: 23.

27.

Dubois

Burn

Goetz

, et al. Diagnostic procedures for Parkinson's disease dementia: recommendations from the movement disorder society task force. Mov Disord 2007; 22: 2314–2324.

28.

Litvan

Goldman

Troster

, et al. Diagnostic criteria for mild cognitive impairment in Parkinson's disease: movement disorder society task force guidelines. Mov Disord 2012; 27: 349–356.

29.

Jost

Kaldenbach

Antonini

, et al. Levodopa dose equivalency in Parkinson's disease: updated systematic review and proposals. Mov Disord 2023; 38: 1236–1252.

30.

Benjamini

Hochberg

. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B (Methodol) 1995; 57: 289–300.

31.

Domellof

Ekman

Forsgren

, et al. Cognitive function in the early phase of Parkinson's disease, a five-year follow-up. Acta Neurol Scand 2015; 132: 79–88.

32.

Weintraub

Caspell-Garcia

Simuni

, et al. Neuropsychiatric symptoms and cognitive abilities over the initial quinquennium of Parkinson disease. Ann Clin Transl Neurol 2020; 7: 449–461.

33.

Myers

O'Donnell

Jackson

, et al. Proteinopathy and longitudinal cognitive decline in Parkinson disease. Neurology 2022; 99: e66–e76.

34.

Yang

Gong

. Changes of cerebrospinal fluid Abeta(42), t-tau, and p-tau in Parkinson's disease patients with cognitive impairment relative to those with normal cognition: a meta-analysis. Neurol Sci 2017; 38: 1953–1961.

35.

Saade

Araujo de Souza

Scavone

, et al. The role of GPNMB in inflammation. Front Immunol 2021; 12: 674739.

36.

Zhu

Liu

, et al. GPNMB Mitigates Alzheimer's disease and enhances autophagy via suppressing the mTOR signal. Neurosci Lett 2022; 767: 136300.

37.

Budge

Neal

Richardson

, et al. Transgenic overexpression of GPNMB protects against MPTP-induced neurodegeneration. Mol Neurobiol 2020; 57: 2920–2933.

38.

Budge

Neal

Richardson

, et al. Glycoprotein NMB: an emerging role in neurodegenerative disease. Mol Neurobiol 2018; 55: 5167–5176.

39.

Smajic

Prada-Medina

Landoulsi

, et al. Single-cell sequencing of human midbrain reveals glial activation and a Parkinson-specific neuronal state. Brain 2022; 145: 964–978.

40.

Pagano

Ferrara

Brooks

, et al. Age at onset and Parkinson disease phenotype. Neurology 2016; 86: 1400–1407.

41.

Prange

Danaila

Laurencin

, et al. Age and time course of long-term motor and nonmotor complications in Parkinson disease. Neurology 2019; 92: e148–e160.

42.

Raket

Oudin Astrom

Norlin

, et al. Impact of age at onset on symptom profiles, treatment characteristics and health-related quality of life in Parkinson's disease. Sci Rep 2022; 12: 526.

43.

Wickremaratchi

Knipe

Sastry

, et al. The motor phenotype of Parkinson's disease in relation to age at onset. Mov Disord 2011; 26: 457–463.

44.

Alves

Wentzel-Larsen

Aarsland

, et al. Progression of motor impairment and disability in Parkinson disease: a population-based study. Neurology 2005; 65: 1436–1441.

45.

Suda

Shimizu

Katsuumi

, et al. Glycoprotein nonmetastatic melanoma protein B regulates lysosomal integrity and lifespan of senescent cells. Sci Rep 2022; 12: 6522.

46.

Ashton

Janelidze

Al Khleifat

, et al. A multicentre validation study of the diagnostic value of plasma neurofilament light. Nat Commun 2021; 12: 3400.

47.

Filipello

Goldsbury

You

, et al.

Soluble TREM2: innocent bystander or active player in neurological diseases?

Neurobiol Dis 2022; 165: 105630.

48.

Byers

Jin

, et al. TREM-2 promotes macrophage survival and lung disease after respiratory viral infection. J Exp Med 2015; 212: 681–697.

49.

Gao

Jiang

Tan

, et al. Microglia in neurodegenerative diseases: mechanism and potential therapeutic targets. Signal Transduct Target Ther 2023; 8: 359.

50.

Teipel

Bruno

Plaska

, et al. Association of CSF sTREM2, a marker of microglia activation, with cholinergic basal forebrain volume in major depressive disorder. J Affect Disord 2021; 293: 429–434.

51.

Zhong

Chen

Wang

, et al. Soluble TREM2 induces inflammatory responses and enhances microglial survival. J Exp Med 2017; 214: 597–607.

52.

Peng

Qiu

Liu

, et al. Analysis of cerebrospinal fluid soluble TREM2 and polymorphisms in sporadic Parkinson's disease in a Chinese population. J Mol Neurosci 2020; 70: 294–301.

53.

Qin

Wan

Wang

, et al. The association of CSF sTREM2 with cognitive decline and its dynamic change in Parkinson's disease: analysis of the PPMI cohort. Front Aging Neurosci 2022; 14: 892493.

54.

Katzenelenbogen

Sheban

Yalin

, et al. Coupled scRNA-Seq and intracellular protein activity reveal an immunosuppressive role of TREM2 in cancer. Cell 2020; 182: 872–885 e819.

55.

Wang

, et al. The association between plasma GPNMB and Parkinson's disease and multiple system atrophy. Parkinsonism Relat Disord 2024; 120: 106001.

56.

Wan

Chen

, et al. No correlation between plasma GPNMB levels and multiple system atrophy in Chinese cohorts. Mov Disord 2023; 38: 1956–1961.

57.

Janelidze

Teunissen

Zetterberg

, et al. Head-to-head comparison of 8 plasma amyloid-beta 42/40 assays in Alzheimer disease. JAMA Neurol 2021; 78: 1375–1382.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.04 MB