Validating Glycoprotein Non-Metastatic Melanoma B (gpNMB, osteoactivin), a new biomarker of Gaucher Disease

Abstract

In the spleens of Gaucher disease mice and patients, there is a striking elevation of expression of glycoprotein non-Metastatic Melanoma B (gpNMB). We conducted a study in a large cohort of patients with Gaucher disease to assess the utility of serum levels of soluble fragment of gpNMB as a biomarker of disease activity. There was >15-fold elevation of gpNMB in sera of untreated patients with Gaucher disease. gpNMB levels correlated with overall disease severity as well as the severity of individual organ compartments: liver, spleen, bone and hematological disease. Imiglucerase enzyme replacement therapy resulted in significant reduction of gpNMB. Serum levels of gpNMB were highly correlated with accumulation of bioactive lipid substrate of Gaucher disease, glucosylsphingosine as well as established biomarkers, chitotriosidase and chemokine, CCL18. Our results suggest utility of gpNMB as a biomarker of Gaucher disease to monitor individual patients and cohorts of patients for disease progression or response to therapy. Investigation of gpNMB in Gaucher disease pathophysiology is likely to illuminate our understanding disease mechanisms.

Introduction

Gaucher Disease (GD) is one of the most common of lysosomal storage disease and it has served as a prototype for therapeutic innovation and biomarker discovery [1]. It occurs due to biallelic mutations in GBA which results in defective acid β-glucosidase and lysosomal accumulation of the primary substrate, glucosylceramide, most conspicuously evident as macrophages engorged with lipid-laden lysosomes, the eponymous Gaucher cells. Disease pathophysiology involves alternative metabolism of accumulating glucosylceramide via acid ceramidase that generates bioactive lipids including massive elevation of glucosylsphingosine (GlcSpn) which is responsible for chronic metabolic inflammation, a central feature of the disease. Clinically, Gaucher disease presents as a continuum of phenotypes involving variable combinations of marrow infiltration, cytopenia, hepatosplenomegaly, complex skeletal involvement including osteoporosis and avascular necrosis of the bone and occasionally pulmonary involvement [2]. Based on presence or absence of early onset neurological manifestations, it is further divided into 3 broad phenotypes: Type 1 presenting without primary neurodegenerative disease (although these patients are at increased risk of Parkinson’s disease later in life) and Type 2 and 3, characterized by infantile neuronopathic involvement or an indolent neuronopathic form of disease presenting in adolescence or early adulthood, respectively [3].

While multiple therapeutic options are available for Gaucher disease including several forms of enzyme replacement therapies and substrate reduction therapy, it is challenging to accurately assess relative efficacy due to extreme heterogeneity of the disease. Qualified biomarkers could serve as reliable surrogates for clinical trials for novel therapies in development as well as facilitate comparison of relative efficacy of existing therapies. Assessment of burden of active disease is currently based on indirect measures of severity of individual organ involvement such as liver volume, spleen volume and semi-quantitative measures of pathological bone marrow infiltration [4]. These measures of individual organ involvement however do not reliably reflect overall disease burden or activity due to extreme heterogeneity of Gaucher disease. For example, it is well-recognized that patients may harbor advanced skeletal disease even if the extent of hepatosplenomegaly is minimal and vice versa. Hence, there is a large unmet need for circulating biomarkers that reflect proximate metabolic disorder and overall disease burden in Gaucher disease. These issues were highlighted at meeting convened by the FDA’s Center for Drug Evaluation and Research on September 20, 2010. Currently used biomarkers of Gaucher disease in the clinic include serum chitotriosidase [5] and the chemokine, PARC/CCL18 [6], both massively secreted from glucosylceramide-laden macrophages. However, both biomarkers are elevated in other disease states and in case of chitotriosidase, common genetic polymorphism in its cognate gene, CHIT1 impacts on circulating levels which hinders inter-individual comparison of disease activity based on this biomarker. Moreover, the emerging appreciation of system-wide involvement beyond macrophages in the pathophysiology of Gaucher disease has been highlighted [7], underscoring the need for more representative biomarkers of disease activity.

We generated a mouse model of Gaucher disease type 1 by conditional deletion of GBA1 gene in hematopoietic and mesenchymal cell lineages using an Mx1 promoter. This model fully replicated human Gaucher disease type 1. We employed Ingenuity Pathway analysis to aid discovery of novel candidate biomarkers for subsequent validation in patients with Gaucher disease. We found the most highly expressed gene in spleens of GD mice and the top biomarker candidate encoded gpNMB (glycoprotein non-metastatic melanoma B). The expression of gpNMB was increased up to 18-fold in the spleens and as much as 216-fold in the livers of most severely affected GD mice. Other studies have also reported increased expression of gpNMB in spleen, brains, sera and CSF of GD patients and GD mice[8–10].

GpNMB is a type I membrane protein which is cleaved into a soluble form and secreted by various cell types through a disintegrin and metalloproteinase (ADAM10) sheddase activity [11–13]. Originally discovered as osteoactivin in animal studies [14], its expression serves as an anabolic regulator of osteoblastogenesis and decreases osteoclast mediated bone resorption [15–18]. Subsequent studies showed gpNMB upregulation and expression in melanocytes, neurons, astrocytes and in several malignant cell lines. It has also been implicated in modulating immune response [19]. Therefore, we conducted a study to assess biomarker properties in large cohort of patients with Gaucher disease with diverse spectrum of disease severity and its response before and after enzyme replacement therapy.

Our study demonstrates massive elevation of soluble fragment of gpNMB in sera of patients with Type 1 Gaucher disease, closely correlated with several indicators of disease severity including established biomarkers (Chitotriosidase, PARC/CCL18 and glucosylsphingosine), and its response to imiglucerase enzyme replacement therapy. Receiver operating characteristics (ROC) curve analysis in this large cohort of patients demonstrated characteristics of circulating soluble fragment of gpNMB that is remarkably comparable to chitotriosidase and PARC/CCL18.

Materials and Methods

Patients

The study comprised a total of 155 patients enrolled in the study. Diagnosis of GD was established by low levels of acid β-glucosidase activity in peripheral blood leukocytes (≤10% of controls) and the presence of GBA mutations. The study was approved by the institutional review boards of Yale University School of Medicine and New York University School of Medicine. Blood samples were collected to assess hemoglobin and platelet counts and established biomarkers. Imaging studies were performed to assess spleen volume, liver volume (extent of organomegaly expressed as multiples of normal) and severity of marrow and skeletal disease. Normal liver volume was 2.5% body weight and normal spleen volume 0.2% body weight as described previously [20]. Data to depict the burden of Gaucher disease were compiled for each patient, including Hermann Score, Bone Marrow Burden Score (BMB), Severity Score Index (SSI) and Disease Severity Scoring System (DS3) as described previously [20, 21]. Patients were divided into two cohorts. The discovery cohort consisted of 41 patients in whom we had baseline serum samples prior to initiation of ERT and subsequently on imiglucerase ERT. The validation cohort comprised 114 GD patients who were receiving enzyme replacement therapy imiglucerase or Velaglucerase) or substrate reduction therapy (eliglustat tartrate) at the time of obtaining serum samples. Patients in the validation cohort were serially followed up between 2012 and 2013 and serum gpNMB was analyzed at each visit.

Gaucher disease mice

For studies of gpNMB in GD mice, we used our mouse model developed via conditional deletion of GBA1 gene in hematopoietic and mesenchymal cell lineages using an Mx1 promoter. This model fully replicated human Gaucher disease type 1. The mice have been previously characterized in detail by us [7, 22]. In addition, we measured gpNMB in sera of mice with germ line deletion of Gba2 which encodes non-lysosomal glucocerebrosidase (which is not deficient in GD) [23] and double KO Gba1^−/−/Gba2^−/− previously characterized by us [24].

GpNMB by ELISA

Serum samples of GD patients and GD mice were diluted at 1:50 in phosphate buffered saline and GPNMB levels were measured using the enzyme-linked immunosorbent assay (ELISA) kits (#DY2550 for human and DY2330 for mouse) from R&D System (Minneapolis, MN) per the manufacturer’s instructions. Briefly, Capture Antibody were plated in 96-well plates for overnight at room temperature. After washing and blocking, 100ul of diluted samples and standards were added in duplication to each well and incubated at room temperature for 2 hours, followed by repeated washings and addition of 100ul of appropriately diluted Detection Ab. After 2 hours of incubation and wash, substrate was added to the plate to develop a color that was measured at 450 nm using the SpectraMax Microplate Reader (Molecular Device, Sunnyvale, CA).

LC-MS/MS Assay for glucosylsphingosine (GlcSpn, L-GL1)

Plasma GlcSpn (L-GL1) was measured as described previously [21]. 25 pg. of N, N-Dimethylsphingosine (Matreya, Pleasant Gap, and PA) was dried under nitrogen gas in an Eppendorf tube. Next 20 μL plasma and 1000 μL chloroform: methanol (2:3) was added to the tube. This was vortexed for 5 minutes and then centrifuged for 3 more minutes. 200 μL chloroform and 520 μL water were added to the tube for liquid-liquid extraction. This mixture was vortexed for 3 more minutes, followed by centrifugation at 16,000×g for 3 minutes. The lower phase of the solution was transferred to a separate auto-sampler vial. The upper phase was separated and the procedure was repeated again in this phase for re-extraction. The combined extracts were subsequently dried again with nitrogen gas followed by reconstitution in 100 uL methanol: water (9:1) for liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS). A calibration curve was prepared using the same procedure as above using GlcSpn standard (Matreya, Pleasant Gap, and PA) and N, N – Dimethylsphingosine.

Samples were placed into the LC–MS/MS system [Waters Acquity UPLC and an API-5000 triple-quadrupole mass spectrometer]. Separation of GlcSpn and other matrix components was achieved through Acquity BEH C18 column (2.1×50 mm, 1.7 μm column) under specific gradients with mobile phase A (0.1 % formic acid in water) and mobile phase B (0.1 % formic acid in acetonitrile). Mass spectrometry (MS) was subsequently performed in select ion monitoring mode with the following transitions: m/z 462.5> 282.4 for GlcSpn and m/z 490.3 > 292.4 for N, N-Dimethylsphingosine.

Chitotriosidase activity assay

Serum chitotriosidase activity assay was performed as previously described in our other paper [21]. Serum samples from patients were diluted at various concentrations of 1:10, 1:20 and 1:50. Healthy control samples remained undiluted. 5 μL of each diluted sample was then added to 50 μL of 4MU-Chitotrioside (Sigma Aldrich, St Louis, MO). This solution was already pre-diluted to 22 nM in Mcilvaine Buffer (0.1M Citric acid, 0.2M Na₂HPO₄, pH 5.2) previously. 0.5 mL stop buffer (0.3M glycine-NaOH, pH 10.6) was added to samples after incubation at 37C for 30 minutes. 100 μL of each this sample was then transferred to a plate and the enzyme activity was measured at the following frequencies: 366nm for excitation and 445nm for emission. A standard curve of β-Methylumbelliferone (4MU, Sigma Aldrich. St Louis, MO) was performed.

CHIT 1 genotyping

This was previously described in our other paper [21]. Genomic DNA from Gaucher disease patients and healthy subjects were extracted. The kit that was used for this purpose was Qiagen’s DNA isolation kit (Valencia, CA). CHIT 1 fragment containing the 24-bp duplication was amplified using the following primers: CHIT1F CAG CTA TCT GAA GCA GAA GG, and CHITR1 GAG AAG CCG GCA AAG TC. AmpliTaq Gold DNA Polymerase system (Thermo-Fisher Scientific, Waltham, MA) was used for polymerase chain reaction at temperatures of 95°C - 60°C - 72°C for 35 cycles. The amplified DNAs were subsequently separated on an agarose gel containing 2% Agarose GPG/LE (Amerianbio, Natick, MA) and 2% LMP Agarose (Thermo Fisher Scientific). In the presence or absence of the extra 24 bps, the band size was divided as either 99bp or 75bp.

CCL18 Enzyme-Linked Immuno Sorbent Assay (ELISA)

The CCL18 protein levels in patients and healthy controls were assayed according to the protocol provided in the DuoSet CCL18/PARC Kit (R&D System, MN) and as previously described [21].

Statistical Methods

Normally distributed data were analyzed using student’s t-test. Non-normally distributed data was analyzed using Wilcoxon rank sum test and Kruskall-Wallis test. Continuous variables were correlated using spearman’s rank correlation coefficient and categorical variables through rank sum test. ROC analysis was performed using STATA version 14.

Results

Serum gpNMB levels are increased in conditional KO mouse model and effect of germ-line deletion of Gba2

Previous expression microarray studies revealed increased overexpression of gpNMB in murine models of GD and in the spleen of a GD patient [7, 10]. In our present study, we used several mouse models to assess the determinants of serum gpNMB levels. The average serum gpNMB levels in WT mice was 35.9 ± 4.7 ng/ml whereas GD mice, it was increased to mean 144.1 ± 38.7 ng/ml (p = 0.01, Figure 1). Compared to wild type mice, there was no change in serum level of gpNMB in Gba2^−/− mice (Gba2 encodes neutral, non-lysosomal glucocerebrosidase, which is not deficient in GD). Deletion if Gba2 in GD mice, attenuated the elevation of gpNMB compared to GD mice with intact Gba2. These findings are consistent with previous findings suggesting a role for neutral glucocerebrosidase in inflammatory response in GD pathophysiology. [24].

Figure 1.

Serum GpNMB levels are increased in conditional KO mouse model and restored to normal in GBA KO/GBA2 KO double KO model

Demographic Description of patients

In the discovery cohort [Table 1] the average age of patients before treatment was 36.8 years (4 – 69 years). The average age of patients post treatment was 44.9 years (range 5 – 77 years). Seventeen patients had undergone prior splenectomy. The predominant genotype of the population was N370S heterozygous (n = 21). Fourteen patients were homozygotes for the N370S mutation and 5 patients were compound heterozygotes. Mean liver volume was 1.8 MN (range: 0.9 – 4) and spleen volume, 17.2 MN (range: 0.9 – 45.6) before initiation of ERT. After a mean of 6 years of imiglucerase ERT (range 0.5 – 19 years) the average liver volume decreased to 1.2 (range 0.5 – 3.9, p = 0.09) and spleen volume to 6.1 (range 0.9 – 14 p= 0.01). Hemoglobin at baseline was 122.6 g/L and increased to mean of 135.9 g/L after (p = 0.02). Platelet count was 158.7 × 10⁹/L at baseline and after ERT increased to 201.5 × 10⁹/L (p = 0.04). SSI and Herman scores were unchanged as expected and in our study DS3 scores were not impacted by ERT.

Table 1.

Baseline Characteristics of Discovery Cohort

Gender	Females: 24 (58.3%)
Genotype	N370S/N370S: 14 (34.1%)
	N370S/Other: 21 (51.2%)
	Other/Other: 6 (14.6%)
Age at First Symptom (years)	19.1 (0.5 – 54)
Age at Diagnosis (years)	21.5 (0.5 – 54)
Splenectomy	17 (41.4%)
Treatment (years)	3.6 (0.5 – 19)
	Before ERT	After ERT	P – value
Age (years)	36.8 (4–69)	44.9 (5–77)	-----
Liver Volume (X Normal)	1.8 (0.9 – 4)	1.3 (0.6 – 3.9)	0.09
Spleen Volume (X Normal)	17.2 (0.9 – 45.6)	6.1 (0.9 – 14)	0.01
SSI	7.2 (2 – 17)	7.5 (2 – 17)	0.97
DS3	2.6 (0 – 8)	2.8 (0 – 8)	0.89
Herman Score	3.6 (0 – 5)	3.6 (0 – 5)	0.94
Hemoglobin (g/L)	122.6 (95 – 164)	135.9 (108 – 172)	0.02
Platelets (x 109/L)	158.7 (39 – 364)	201.4 (44 – 366)	0.05
Serum Chitotriosidase (nm/ml/hour)	12451.9 (337 – 58707)	3880.5 (330 – 19433)	<0.001
Serum CCL 18 (ng/ml)	603 (40 – 1240)	276.7 (49 – 1039)	<0.001
Serum Lyso-GL1 (ng/ml)	180.9 (14 – 464)	99.3 (3 – 503)	<0.001

The average age of patient in the validation cohort [Table 2] was 48 years (range 4 – 83 years). These patients had already received ERT for 13 years at the time of the study (range 0.5 – 29 years). As expected the liver volume (average 0.9, range 0.5–2) and spleen volume (average 3.3, range 0.7 – 14.6) remained within near normal limits in a majority of the patients. Here the average hemoglobin and platelet counts were 13.97 g/dl and 168.8 × 10⁹/L respectively while average SSI and DS3 score were 6.9 and 2.5.

Table 2.

Baseline Characteristics of Validation Cohort

Cohort 2: Validation Cohort (n = 114)
Gender	Males = 38.6%
Genotype	N370S/N370S = 57.8%
Age	48.3 (4 – 83)
SSI	6.9 (2 – 17)
DS3	2.5 (0 – 8)
Hemoglobin (g/L)	139.7 (87 – 171)
Platelet (x 109/L)	168.8 (52 – 796)
Liver Volume X Normal	0.9(0.5 – 2)
Spleen Volume X Normal	3.3 (0.7 – 14.6)
Splenectomy	39 (20.7%)
Bone Marrow Burden Score	5.4 (0 – 16)
Herman Score	2.6 (1 – 5)
Years in Treatment	13.7 (.5 – 29)
Serum Chitotriosidase (nm/ml/hour)	973.3 (1 – 7262)
Serum CCL18 (ng/ml)	117.2 (1 – 936)
Serum GlcSpn (ng/ml)	155.9 (4 – 1160)

GPNMB levels are markedly elevated in untreated GD Type 1

We measured serum gpNMB levels in the discovery cohort before treatment with ERT and after mean 3.6 years of ERT. This data was further compared to healthy control subjects.

GPNMB levels in the pre- ERT patients were elevated with a mean level of 387 ± 31 ng/ml (range 4.5 – 854 ng/ml) compared to healthy controls (mean 25 ng/ml ± 4 (range 11 – 44 ng/ml, p<0.001). On ERT, serum GpNMB levels fell from 387 ng/ml (range 4.5 – 854 ng/ml) to 260 ± 37 ng/ml (range 41 – 1343 ng/ml, p<0.001). This represented a 15-fold elevation compared to healthy controls in untreated patients and 10-fold elevation after ERT [Figure 2].

Figure 2.

GpNMB levels (ng/ml) before ERT and after ERT in GD patients when compared to healthy controls

GpNMB levels correlate with indicators of disease severity

Next, we evaluated the correlation of plasma gpNMB levels with established biomarkers of disease activity namely chitotriosidase, CCL18 and GlcSpn (glucosylsphingosine, L-GL1) as well as standard indicators of disease severity, i.e., hematological, visceral and skeletal measurements of GD. We adjusted chitotriosidase by CHIT1 genotype as described previously [25] and removed patients with CHIT1 null homozygous polymorphism in the analysis. Serum chitotriosidase activity was positively correlated with serum level of soluble fragment of gpNMB levels (rho = 0.79 p < 0.001) [Figure 3.1]. Similarly, CCL18 levels correlated with gpNMB (rho = 0.79 p < 0.001) [Fig 3.2] and plasma GlcSpn was also significantly correlated with gpNMB (rho = 0.57 p < 0.001) [Fig 3.3]. Serum levels of soluble fragment of gpNMB correlated with liver volume (rho 0.47 p<0.001) [Fig 3.4] and with spleen volume (rho 0.54 p<0.001) [Fig 3.5]. Overall patients who had undergone prior splenectomy had higher levels of soluble fragment of gpNMB in GD patients when compared to GD patients with intact spleens (p = 0.003) [Fig 3.6]. Platelet counts were inversely correlated with serum levels of gpNMB (rho = −0.24 p < 0.001) [Fig 3.7]. Levels of soluble fragment of gpNMB in serum were weakly but significantly correlated with Zimran’s severity score index (rho = 0.22 p<0.001) [Fig 3.8]. Previous studies have demonstrated an osteoinductive effect of gpNMB on bone mass and this led us to examine the relationship of gpNMB with skeletal disease. A useful depiction of all facets of Gaucher skeletal disease is the Herman score and it was weakly but significantly correlated with serum gpNMB levels (rho 0.16 p=0.01) Fig 3.9.

Figure 3.

Correlation of serum GpNMB with other markers of GD severity:

1. Correlation of serum GpNMB with serum Chitotriosidase
2. Correlation of serum GpNMB with serum CCL18
3. Correlation of serum GpNMB with serum Lyso-GL1 (GlcSpn, glucosylsphingosine)
4. Correlation of serum GpNMB with liver volume
5. Correlation of serum GpNMB with spleen volume
6. Correlation of serum GpNMB with platelet count
7. Correlation of serum GpNMB with spleen status
8. Correlation of serum GpNMB with SSI
9. Correlation of serum GpNMB with Herman Scores

ROC analysis of gpNMB and comparison with Chitotriosidase and CCL18

To determine biomarker properties of gpNMB in relation to established biomarkers (chitotriosidase and CCL18), we performed ROC analysis. In a total of 284 serum samples, we compared the levels of soluble fragment of gpNMB with the levels of chitotriosidase and CCL18. The AUROCC (area under the ROC curve) for soluble fragment of gpNMB was almost identical to that for chitotriosidase and CCL18 (0.87 95% CI 0.81 – 0.91) with a p value of 0.5 indicating lack of statistical significance between the three biomarkers [Figure 4]. A cut-off of 40 ng/ml of soluble fragment of serum GPNMB levels, gave sensitivity of 71% and specificity of 90% in distinguishing between GD patient and healthy controls.

Figure 4.

ROC curve comparison between gpNMB, serum chitotriosidase and serum CCL 18

Discussion

gpNMB expression is increased in livers and spleens of GD mice as well as sera of GD patients[7, 8]. In a large cohort of GD patients, we confirmed markedly increased levels of the soluble fragment of gpNMB in proportion to severity of Gaucher disease indicated by overall disease severity as well as severity of hepatomegaly, splenomegaly, thrombocytopenia and bone disease. Serum levels of the soluble fragment of gpNMB were highly correlated with established biomarkers (chitotriosidase, CCL18 and glucosylsphingosine) and imiglucerase ERT in previously untreated patients was associated with significant reduction of gpNMB. Together, our studies support the utility of soluble fragment of GpNMB as a biomarker of Gaucher disease to longitudinally track disease severity and response to therapy. While our study was in progress, Kramer et al reported elevated serum levels of soluble fragment of gpNMB in mouse models of GD and its response to gene therapy and substrate reduction therapy [8]. This study also reported elevated levels of soluble fragment of gpNMB in smaller cohort of GD patients with results broadly in line with our study. In addition, Zigdon et al recently reported elevated gpNMB in mouse models of neuronopathic GD as well as in CSF and brain tissue of three patients with neuronopathic Gaucher disease [9]. Our study bolsters the evidence for utility of soluble fragment of gpNMB to monitor patients with Gaucher disease. The ROC characteristics of serum gpNMB in Gaucher disease is strikingly like established biomarkers, chitotriosidase and CCL18. It has the advantage over chitotriosidase as there is no genetic polymorphism that hinders its application to all patient populations.

Compared to recent report by Kramer et al, our study comprised a much larger patient population and we found stronger correlation of gpNMB with established biomarkers, GlcSpn, chitotriosidase and CCL18. Our study is the first to comprehensively evaluate correlation of levels of serum soluble fragment of gpNMB to clinical indicators of GD such as extent of hepatosplenomegaly, severity of thrombocytopenia, overall severity score index and severity of skeletal disease indicated by Hermann score. Moreover, for the first time, we also evaluated the effect of prior splenectomy on gpNMB. Our study show levels of serum levels of soluble fragment of gpNMB is significantly correlated not only with overall disease severity indicated by Zimran’s severity score index, it is correlated with severity of involvement of individual organ compartments, liver volume, spleen volume, bone disease, thrombocytopenia and splenectomy status. There was no significant difference for AUROCC between gpNMB, chitotriosidase and CLC18(p = 0.5), supporting its use as sole marker to monitor patients with GD. A cut off of 40 mg/ml, gpNMB gave a sensitivity of 87% and specificity of 82% for distinguishing between patients with GD and the healthy controls. Similar to chitotriosidase and CCL18, elevations serum levels of gpNMB have also been reported in other lysosomal diseases i.e., Niemann-Pick [26], Mucopolysaccharidosis Type VII [27], Tay Sachs and in Sandhoff disease [28, 29]. Taken together, while gpNMB exhibit similar biomarkers properties to chitotriosidase and CCL18, it is unlikely to be superior to glucosylsphingosine which is now emerging as key biomarker that reflects system-wide pathophysiology of Gaucher disease [21, 30, 31].

Validation of a biomarker such as gpNMB in Gaucher disease is not only clinically useful to monitor patients in the clinic to personalize management, it is also useful to probe the underlying biology of this highly complex lysosomal disease at the intersection of bioactive lipid accumulation and immune activation. Several studies suggest elevated gpNMB is a marker of lysosomal stress. The link between gpNMB and Gaucher disease may involve its role in the degradation of cellular debris and macroautophagy. Li et Al [32] studied its essential function in the fusion of phagosome with lysosome through the recruitment of an accessory protein, LC3. They proposed gpNMB involvement in mediating phagocytosis and the bulk degradation pathway [32]. Studies by Gabriel et al. further confirmed that lysosomal stress in a lipotoxic environment led to substantial increase in gpNMB with induction seen at both the mRNA and protein levels [31].

Microphthalmia transcription factor (MITF), recognized to signal lysosomal stress, appears to be the transcription factor that regulates gpNMB expression[33, 34]. Other studies have reported that inhibition of mTORC results in increased expression of gpNMB [33]. The bone is a major target of Gaucher disease and skeletal manifestations are responsible for most disabling morbidity. Early investigations suggest that the role of gpNMB in the bone is complex involving positive regulation of anabolic osteoblastic function as well as in resorptive osteoclast function. [15–18]. In the osteoclast, gpNMB is part of the regulon involving several proteins well-known to be aberrantly expressed in Gaucher disease, namely, acid phosphatase (ACPS) and cathepsin K (CTSK) [34]. Gaucher disease pathophysiology prominently involves NKT cell dysregulation mediated by bioactive lipids presented by antigen presentation cells. Other evidence suggests a role for gpNMB in regulation of modulation of T-cells via binding to Syndecan 4 on T cells through antigen presenting cells [35]. Finally, the proposed role of gpNMB in reflecting lysosomal stress is very interesting in view of the link between GBA mutations and Parkinson disease. Biallelic as well as heterozygotes carrier mutations in GBA are now established major genetic risk factors for Parkinson disease. Indeed, of some two dozen genes linked to Parkinson disease, at least 14 of the PD-linked genes encode proteins involved in endolysosomal trafficking, defining this as the key pathway in in pathophysiology of PD. Interestingly, gpNMB is one of these genes. The intersection of gpNMB and GBA mutations in disease pathophysiology suggest that deciphering underlying molecular basis would be a fruitful area of investigation.