Glucosylceramide and Glucosylsphingosine Quantitation by Liquid

Jul 7, 2017 - Gaucher disease (GD) is caused by mutations in the GBA1 gene that encodes the lysosomal enzyme acid β-glucosidase (GCase). Reduced ...
0 downloads 0 Views 1MB Size
Subscriber access provided by Uppsala universitetsbibliotek

Article

Glucosylceramide and Glucosylsphingosine Quantitation by Liquid Chromatography-Tandem Mass Spectrometry to Enable In Vivo Preclinical Studies of Neuronopathic Gaucher Disease Rick L Hamler, Nastry Brignol, Sean W. Clark, Sean Morrison, Leo B. Dungan, Hui H. Chang, Richie Khanna, Michelle Frascella, Kenneth J. Valenzano, Elfrida R. Benjamin, and Robert E. Boyd Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.7b01442 • Publication Date (Web): 07 Jul 2017 Downloaded from http://pubs.acs.org on July 10, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Analytical Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Glucosylceramide and Glucosylsphingosine Quantitation by Liquid Chromatography-Tandem Mass Spectrometry to Enable In Vivo Preclinical Studies of Neuronopathic Gaucher Disease Rick Hamler, Nastry Brignol, Sean W. Clark, Sean Morrison, Leo B. Dungan, Hui H. Chang, Richie Khanna, Michelle Frascella, Kenneth J. Valenzano, Elfrida R. Benjamin and Robert E. Boyd Amicus Therapeutics Inc, 1 Cedar Brook Drive, Cranbury, NJ 08512

The first two authors contributed equally to this work. *Correspondence: Rick Hamler, Amicus Therapeutics Inc., 1 Cedar Brook Drive, Cranbury, New Jersey 08512, USA. E-mail: [email protected]

1 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Abstract Gaucher disease (GD) is caused by mutations in the GBA1 gene that encodes the lysosomal enzyme acid β-glucosidase (GCase). Reduced GCase activity primarily leads to the accumulation of two substrates, glucosylceramide (GlcCer) and glucosylsphingosine (GlcSph). Current treatment options have not been shown to ameliorate the neurological pathology observed in the most severe forms of GD, clearly representing an unmet medical need. To better understand the relationship between GlcCer and GlcSph accumulation, and ultimately their connection with the progression of neurological pathology, we developed LC-MS/MS methods to quantify GlcCer and GlcSph in mouse brain tissue. A significant challenge in developing these methods was the chromatographic separation of GlcCer and GlcSph from the far more abundant isobaric galactosyl epimers naturally occurring in white matter. After validation of both methods, we evaluated the levels of both substrates in five different GD mouse models, and found significant elevation of brain GlcSph in all five, while GlcCer was elevated in only one of the five models. In addition, we measured GlcCer and GlcSph levels in the brains of wild-type mice after administration of the GCase inhibitor conduritol β-epoxide (CBE), as well as the nonlysosomal β-glucosidase (GBA2) inhibitor N-butyldeoxygalactonojirimycin (NB-DGJ). Inhibition of GCase by CBE resulted in elevation of both sphingolipids; however, inhibition of GBA2 by NB-DGJ resulted in elevation of GlcCer only. Taken together, these data support the idea that GlcSph is a more selective and sensitive biomarker than GlcCer for neuronopathic GD in preclinical models.

2 ACS Paragon Plus Environment

Page 2 of 20

Page 3 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Introduction Gaucher disease (GD) is an autosomal recessive lysosomal storage disorder (LSD) caused by mutations in the gene (GBA1) that encodes the lysosomal enzyme acid β-glucosidase (GCase), which is responsible for the hydrolysis of the substrates glucosylceramide (GlcCer) and glucosylsphingosine (GlcSph). Reduced GCase activity leads to accumulation of these sphingolipids in cells throughout the body1. GD is broadly classified into three clinical subtypes, which are differentiated by the severity and progression of clinical pathology, and whether the central nervous system (CNS) is affected (non-neuronopathic vs neuronopathic). Type 1 GD is the most common form, typically characterized by peripheral manifestations such as enlarged liver and spleen, reduced platelets, and bone disease2. Types 2 and 3 GD are neuronopathic and include CNS pathology in addition to the visceral manifestations seen in Type 1. The most common neurological symptoms of neuronopathic GD include encephalopathy, progressive myoclonic epilepsy, and ophthalmoplegia3. Type 2 GD is also referred to as the acute neuronopathic form and is characterized by significant neurological impairment in the first years of life, with life expectancies less than 2 years4. Type 3 GD, also referred to as the chronic neuronopathic form, is characterized by a slower progression of neurologic decline and survival often extending into adulthood. Today, enzyme replacement therapy (ERT) is the standard of care to treat the visceral symptoms, and is widely prescribed for the treatment of Types 1 and 3 GD. However, ERT has not been proven effective in treating the CNS pathology experienced by neuronopathic GD patients5, representing a clear unmet medical need. Pre-clinical research directed to this unmet need requires both appropriate animal models and identification of good CNS biomarkers to assess progression of the disease as well as the efficacy of potential therapeutic treatments. This goal is significantly complicated by the fact that brain tissue contains dramatically larger amounts of two related sphingolipids, namely galactosylceramide (GalCer) and galactosylsphingosine (GalSph). GalCer and GlcSph are epimers of GlcCer and GlcSph, respectively, and are indistinguishable with any type of mass spectrometric detection, consequently requiring chromatographic separation. We have developed and validated methods that yield baseline separation of the galactosyl and glucosyl versions of both sphingolipids, and enable accurate quantification of GlcCer and GlcSph to support preclinical studies in neuronopathic GD. It has been hypothesized recently that plasma 3 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

GlcSph levels may serve as a sensitive and selective biomarker for nonneuronopathic GD6-8; however, it is currently unclear if this hypothesis can be extended to CNS GlcSph levels and neuronopathic GD6-8. To evaluate these sphingolipids as potential biomarkers of neuronopathic GD, we used our methods to measure GlcCer and GlcSph in brain tissue from five different GD mouse models, and wild-type mice administered inhibitors of either GCase or non-lysosomal βglucosidase (GBA2). Our findings suggest that brain GlcSph is indeed a more sensitive and selective biomarker for neuronopathic GD in preclinical models.

Experimental Section Materials Glucosylceramide (GlcCer), glucosylsphingosine (GlcSph), and d3-C16:0 glucosylceramide were purchased from Matreya LLC (Pleasant Gap, PA). 13C6-GlcSph was synthesized at Amicus Therapeutics (Cranbury, NJ) based on chemistry previously reported9. Analytical chemicals and solvents were purchased from Thermo Fisher Scientific (Waltham, MA) unless otherwise stated. Wild-Type and GD Mouse Models Wild-type C57BL/6 mice were purchased from Taconic Farms (Germantown, NY). Wild-type C57BL/6-DBA/2 mice were provided by Dr. Marie-Francoise Chesselet (UCLA, Los Angeles, CA). GBAL444P/L444P mice containing a knock-in human L444P mutation were a kind gift from Dr. Richard Proia10-11 (National Institutes of Health, Bethesda, MD). The Proia mouse model carries a disruption of the endogenous GBA1 allele and an expression cassette for the human GBA1 L444P allele. The L444P allele leads to a neuronopathic form of Gaucher disease in humans, but is non-neuronopathic and thus less severe in the mouse. Brain tissues from hL444P-tg8 and hN370S-tg4 mice, which contain eight copies of the human L444P- or four copies of the human N370S-GCase mutation, respectively, were provided by Dr. Lorne Clarke12 (University of British Columbia, Vancouver, British Columbia, Canada). The Clarke mouse models carry multiple copies of either the human GBA1 N370S or L444P allele, and the endogenous GBA1 locus has been disrupted. As noted for the Proia L444P model above, the L444P mutation is severe in humans, but not in the Proia mouse. Conversely, the N370S 4 ACS Paragon Plus Environment

Page 4 of 20

Page 5 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

mutation leads to a milder, non-neuronopathic form of Gaucher disease in humans, but has been found to be lethal in the mouse. The presence of multiple copies of the N370S allele in the Clarke mouse provides for a viable model. Together, these two Clarke models represent an intermediate model of Gaucher disease. Transgenic mice containing a V394L point mutation backcrossed with mice expressing low levels (4L/PS-/-NA) or normal levels (4L/PS +/+NA) of prosaposin were obtained from Dr. Gregory Grabowski13 (Cincinnati Children’s Hospital, Cincinnati, OH). The Grabowski mice harbor three independent mutations. First, the prosaposin locus is disrupted (PS-/-) or left intact (PS+/+). Prosaposin is a polyprotein that is processed into four mature proteins. One of these, Saposin C, enhances the activity of lysosomal GCase. Second, these mice carry a transgene expressing prosaposin under control of the PGK promoter (NA). Together with the disruption of the endogenous prosaposin gene (PS-/-NA), this leads to lower than normal expression of prosaposin, and thus Saposin C. Third, these mice have a replacement of the endogenous GBA1 locus with a mouse GBA1 allele carrying a change that modifies the protein sequence (V394L). In isolation, the V394L form of GCase has no obvious impact on the phenotype of mice; however, in combination with the prosaposin hypomorph (PS/-NA), this leads to severe neuronopathic disease and visceral pathology in the mouse. The Grabowski mouse model thus represents a severe form of Gaucher disease. Brain Tissue Homogenate Preparation To prepare brain homogenates, either the whole brain or one hemisphere was processed. After thawing brain tissue on wet ice, it was diced and mixed thoroughly prior to aliquoting and weighing for homogenate preparation. Homogenates were prepared by adding 16 µL of deionized water per mg of tissue (typically 25 to 50 mg). The mixture was homogenized with lysing matrix A/D on a FastPrep-24 homogenizer (MP Biomedicals, Solon, OH). Care was taken to ensure complete homogenization of the tissue sample, and both tissue and tissue homogenates were kept on wet ice during preparation, processing, and extraction. Since both GlcCer and GlcSph are present at significant endogenous levels in mouse brain tissue, calibration standards and quality control (QC) samples were prepared using DMSO:MeOH (1:1) as the diluent. For both GlcCer and GlcSph, independent stock solutions for standards and QCs were prepared by dissolving each sphingolipid in CHCl3:MeOH (2:1) to a final concentration of 1.0 mg/mL. 5 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

GlcCer GlcCer calibration standard concentrations ranged from 20 ng/mL to 5000 ng/mL, and QC samples were prepared at nominal GlcCer concentrations of 20, 80, 400, 1000, 2000, and 4000 ng/mL. A stock solution of the GlcCer internal standard (IS) d3-C16:0 GlcCer was prepared by dissolving the sphingolipid in CHCl3:MeOH (2:1) to a final concentration of 1.0 mg/mL. This solution was then used to prepare the working IS solution in DMSO:MeOH (1:1) at a final concentration of 2000 ng/mL. A 50 µL aliquot of tissue homogenate and 25 µL of the working IS solution was added to a 13 x 100 mm silanized glass tube. Then, 0.4 mL of MeOH and 1.25 mL of Me2CO:MeOH (1:1) were added to the tube prior to shaking on a multi-tube vortexer (VWR, Radnor, PA) for 60 minutes. After shaking, 0.3 mL of water and 0.6 mL of H2O:MeOH (13:87) were added, and the tubes vortexed briefly and then centrifuged at 3220 x g for 10 minutes. The supernatant was loaded onto an Agilent Bond Elut-C18, 100 mg, 3 mL solid-phase extraction cartridge (Agilent Technologies, Santa Clara, CA) pre-conditioned with 2 mL of MeOH and then 2 mL of MeOH:Me2CO:H2O (67:23:10). The cartridge was subsequently washed with 2 mL of MeOH:Me2CO:H2O (67:23:10) and GlcCer was eluted with 2 mL of Me2CO:MeOH (90:10). The eluant was evaporated to dryness under nitrogen at 40 °C and reconstituted by addition of 50 µL of DMSO and 200 µL of MeCN:MeOH:H2O (95:2.5:2.5) containing 5 mM ammonium formate and 0.5% formic acid. The reconstituted samples were transferred to glass microvial inserts for subsequent liquid chromatography tandem mass spectrometry (LC-MS/MS) analysis. GlcSph GlcSph calibration standard concentrations ranged from 0.25 ng/mL to 200 ng/mL, and QC samples were prepared at nominal GlcSph concentrations of 0.25, 0.75, 4.0, 20, 80, and 160 ng/mL. A stock solution of GlcSph IS (13C6-GlcSph) was prepared by dissolving the sphingolipid in CHCl3:MeOH (2:1) to a final concentration of 0.3 mg/mL. This solution was used to prepare the working IS solution in DMSO:MeOH (1:1) at a final concentration of 125 ng/mL. A 50 µL aliquot of tissue homogenate and 25 µL of the working IS solution was added to a 13 x 100 mm silanized glass tube. Then, 1 mL of MeOH and 0.5 mL of 1 N HCl were added to the tube prior to sonicating at room temperature for approximately 5 minutes. The mixture was shaken on a multi-tube vortexer for 30 minutes and then centrifuged at 3220 x g for 10 minutes. 6 ACS Paragon Plus Environment

Page 6 of 20

Page 7 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

The supernatant was loaded onto an Oasis MCX 3cc, 60 mg sorbent solid-phase extraction cartridge (Waters, Milford, MA) pre-conditioned with 1 mL of MeOH and then 1 mL of water. The cartridge was washed with 2 mL of 0.1 N HCl and then 2 mL of MeOH. GlcSph was eluted with 2 mL of freshly prepared 5% ammonium hydroxide in MeOH. The eluant was evaporated to dryness under nitrogen at 40 °C and reconstituted by addition of 50 µL of DMSO and 200 µL of MeCN:MeOH:H2O (95:2.5:2.5) containing 5 mM ammonium formate and 0.5% formic acid. The reconstituted samples were transferred to glass microvial inserts for subsequent analysis by LC-MS/MS. HPLC-MS/MS Analysis The chromatographic methods for quantitation of GlcCer and GlcSph in brain tissue were both developed in the Amicus laboratory14-15. The liquid chromatography system consisted of an HTc autosampler coupled with two LC-20AD pumps from Shimadzu (Columbia, MD) and a 4000 QTrap mass spectrometer (Applied Biosystems, Foster City, CA) in positive ion electrospray ionization (+ESI) mode. Details of the chromatographic conditions and MS detection are described in Table 1. Nitrogen was used as the collision gas for both sphingolipids. For GlcCer, the collision energy was set at 53-57 V depending on the GlcCer isoform, and 49V for the IS. Quantitative MS/MS data were collected using multiple reaction monitoring (MRM) scan mode.

Table 1. HPLC and MS/MS parameters for the analysis of GlcCer and GlcSph. The GlcCer isoforms of varying acyl chain carbon lengths, ranging from C16 to C24, were monitored with the following precursor ion to product ion transitions: GlcCer C16:0 m/z 700 > 264, GlcCer C18:0 m/z 728 > 264, GlcCer C20:0 m/z 756 > 264, GlcCer C22:0 m/z 784 > 264,

7 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

GlcCer C23:0 m/z 798 > 264, GlcCer C24:1 m/z 810 > 264, GlcCer C24:0 m/z 812 > 264, and IS GlcCer d3-C16:0 m/z 703 > 264. These transitions represent a major product ion formed by the neutral loss of the acyl chain, glucose, and one water (Figure 1A). A linear calibration curve with weighting factor 1/x2 was generated by plotting the ratio of the summed peak areas for the seven isoforms of GlcCer to that of the IS versus increasing GlcCer calibration standard concentrations. This standard curve was used to quantify GlcCer levels in tissue homogenate samples. For GlcSph, the collision energy was set at 30 V for both GlcSph and the IS. Quantitative MS/MS data were collected using MRM scan mode. GlcSph data were collected as a single analyte and the following precursor to product ion transitions were monitored: GlcSph m/z 462 > 282 and IS 13C6-GlcSph m/z 468 > 282. These transitions represent the neutral loss of the glucose ring from GlcSph (Figure 1B). A linear calibration curve with weighting factor 1/x2 was generated by plotting the ratio of the peak area of GlcSph to that of the IS versus increasing GlcSph calibration standard concentrations. This standard curve was used to quantify GlcSph levels in tissue homogenate samples.

8 ACS Paragon Plus Environment

Page 8 of 20

Page 9 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 1. GlcCer and GlcSph Molecular Structures. (A) Structure of GlcCer, where R equals the acyl chain of varying carbon lengths. The most intense product ion is monitored at m/z 264 and results from the neutral loss of the acyl chain, the glucose, and one water. (B) Structure of GlcSph. The most intense product ion is monitored at m/z 282 and results from the neutral loss of glucose.

Method Validation For both the GlcCer and GlcSph methods, QC samples were used to assess accuracy and precision. Each QC concentration was analyzed as six independent replicates and all QCs were bracketed by two calibration standard curves. Accuracy, or percent bias (% Bias), was defined as 100 times the difference between the mean found concentration and the nominal concentration, and divided by the nominal concentration. Precision, or percent coefficient of variation (% CV), was defined as the standard deviation divided by the mean concentration. For GlcCer, the intraday accuracy for QCs ranged from -5.1% to 8.8% of nominal values, while the intra-day precision ranged from 7.1% to 13.8%. Mouse brain homogenates from both wild-type and those found to have elevated GlcCer levels due to oral administration of NB-DGJ were also analyzed as endogenous QCs (see Results and Discussion, and Figure 5), and the intra-assay precision ranged from 3.6% to 12.9%. For GlcSph, three accuracy and precision runs were conducted and the intra-day accuracy for QCs ranged from -7.8% to 4.1% of nominal values, while the inter-day accuracy ranged from -6.1% to 1.9%. The intra-day precision for QCs ranged from 1.4 to 14.4%, while the inter-day precision ranged from 2.4 to 13.4%. The intra-day accuracy for mouse brain homogenate QCs spiked with reference standard ranged from -7.9% to 11.5% of nominal values, while the inter-day accuracy ranged from -6.2% to 9.4% of nominal values. The intra-day precision for the homogenate QCs ranged from 1.3 to 4.2%, while the inter-day precision ranged from 2.8 to 4.3%. The brain tissue homogenate QC concentrations (20, 80, and 160 ng/mL) were chosen such that the endogenous GlcSph levels would not contribute more than 20% to the QC spike-in concentration. All runs used to assess accuracy and precision met the acceptance criteria that ≥2/3 of the total number of QC samples, and ≥1/2 of the QC samples at each concentration,

9 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

were within ± 20% (±25% for the LLOQ) of the nominal concentration, and the precision (measured as %CV) was ≤20% for each QC concentration (≤25% for the LLOQ). Parallelism experiments were conducted to confirm that the concentration-response of endogenous GlcCer and GlcSph in brain tissue homogenates match those of the GlcCer and GlcSph reference standards spiked into diluent DMSO:MeOH (1:1). These experiments were done in quadruplicate by diluting mouse brain tissue homogenate with water by an additional two- and five-fold. All parallelism results met the predetermined accuracy and precision acceptance criteria (diluted homogenate mean concentration within ± 20% of the undiluted homogenate mean concentration, and the precision ≤ 20% for each homogenate analyzed in quadruplicate). The dilution linearity of GlcSph was assessed in quadruplicate using brain tissue homogenate QCs spiked with GlcSph reference standard and dilution factors ranging from twoto fifty-fold. The solvent recovery, brain tissue homogenate recovery, and brain tissue homogenate matrix effect for GlcSph and 13C6-GlcSph were assessed. The average solvent recoveries for GlcSph and 13

C6-GlcSph were 96% and 102%, respectively, while the average brain tissue homogenate

recoveries were 71% and 76%, respectively. Matrix effect values of 1.00 and 0.99 for GlcSph and the IS, respectively, were indicative of no matrix-based ion suppression. Finally, the stability of GlcCer and GlcSph in mouse brain tissue homogenates was assessed in quadruplicate. GlcCer was found to be stable in mouse brain tissue homogenate for at least four hours at room temperature, at least four hours on wet ice, and after two freeze-thaw cycles. GlcSph was found to be stable in mouse brain tissue homogenate for six hours at room temperature, and after four freeze-thaw cycles. Statistical Methods GraphPad Prism 6 (GraphPad Software, San Diego, CA) was used for all statistical calculations. Data were log-transformed to improve normality. ANOVA with Tukey’s or Dunnett’s post-test was utilized where appropriate. In comparisons with unequal variance, a 2-sided t-test with Welch’s correction was used followed by a Bonferroni correction for multiple comparisons. Significance was set at p < 0.05. In Vivo Studies with CBE or NB-DGJ Administration to Wild-Type Mice

10 ACS Paragon Plus Environment

Page 10 of 20

Page 11 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

CBE (100 mg/kg intraperitoneal) was administered daily for four weeks to wild-type FVB/N mice. After the administration period, mice were sacrificed and brain tissue analyzed for GlcCer and GlcSph. C57BL/6 mice were dosed orally with either 3, 10, 30, 100, or 300 mg/kg NB-DGJ for 7 days, sacrificed on day 8 (1-day washout), and brain tissue analyzed for GlcCer and GlcSph.

Results and Discussion Nilsson and Svennerholm first reported the accumulation of GlcCer and GlcSph in the brains of infants and juveniles with neuronopathic GD in 19824. They found the highest levels of both sphingolipids in the most fulminant cases, suggesting that GlcCer and GlcSph contribute to the pathology of the disease. While there does not appear to be a correlation between GlcCer levels and GD pathology16, GlcSph has been correlated with a number of indicators of pathology17-20, including cytotoxicity21-23. More recent studies have not only identified a correlation between GlcSph levels in GD patients and the clinical manifestations of the disease6-7, but have also proposed that GlcSph is a sensitive and selective biomarker6-8. These studies, however, have focused only on GD patients’ plasma GlcSph levels. There were two goals for the work presented herein. The first was to develop robust LC-MS/MS methods to accurately quantify GlcCer and GlcSph in brain tissue. It should be noted that the presence of the far more abundant galactosyl sphingolipids (GalSph and GalCer) in brain tissue complicate the analysis of GlcCer and GlcSph due to the need for chromatographic separation of the glucose and galactose epimers. Based on previous work24, an un-derivatized silica column in combination with improved carbohydrate selectivity afforded by the addition of methanol and water to the mobile phase resulted in baseline separation (Figure 2). A validation protocol was then designed and implemented for both methods to ensure accurate and reliable quantitation of GlcCer and GlcSph brain levels.

11 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

12 ACS Paragon Plus Environment

Page 12 of 20

Page 13 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 2. MRM Chromatograms of GlcCer and GlcSph. (A) MRM chromatograms of each GlcCer isoform (shaded peaks) determined in wild-type (WT) mouse brain tissue, showing the respective baseline resolution of the later eluting and more abundant galactosyl epimers, GalCer (open peaks). (B) MRM chromatograms of GlcSph (shaded peaks) in wild-type (WT) and GBA L444P/L444P

mouse brain tissue, showing the respective baseline resolution of the later eluting and

more abundant galactosyl epimer, GalSph (open peaks). The second goal was to assess the value of these sphingolipids in brain tissue as potential biomarkers of neuronopathic GD. To accomplish this, we began by using our validated methods to measure GlcCer and GlcSph levels in brain tissue from five preclinical GD mouse models. Since “wild-type” litter mates were not available for all of the GD mouse strains tested, we investigated GlcCer and GlcSph brain levels across multiple strains and ages of wild-type mice (Figure 3). GlcCer levels ranged from 4 to 17 µg/g of brain tissue, while GlcSph levels ranged from 9 to 84 ng/g of tissue, an approximate 1000-fold difference in the absolute values of the two sphingolipids. Moreover, between the 4-, 20-, and 60-week-old C57/BL6-DBA/2 mice (WT6, WT7, and WT8, respectively), there was an age-dependent accumulation of GlcSph (p < 0.05), indicating that the method is capable of detecting significant differences between ages in a single strain. We next measured GlcCer and GlcSph levels in brain tissue from five different GD mouse models (Figure 3). It has been reported that GlcCer is not elevated in the brain of the Proia GBA L444P/L444P

mice25, or in the cerebellum and cerebrum of the hL444P-tg8 and hN370S-tg4 mice12;

our results are commensurate with these earlier reports. In contrast, Sun et al. previously reported a large accumulation of GlcCer in the cerebellum of PS-/-NA 4L mice, which we also observed as a greater than 25-fold increase in total brain GlcCer levels in this GD model relative to our wild-type measurements. Surprisingly, we found that GlcSph levels in the GBA L444P/L444P mice were generally increased by at least 4-fold compared to the wild-type mice. To our knowledge, this is the first report of elevated GlcSph in these animals. Sanders et al. observed elevated and progressive increases in GlcSph levels in the hL444P-tg8 and hN370S-tg4 mice12; again, our results are similar, with at least a 4-fold elevation in brain GlcSph levels being seen relative to wild-type mice. Even more dramatic was the elevation of brain GlcSph levels in the PS-/- NA 4L mice, with a greater than 150-fold increase relative to wild-type mice. Lastly, to our knowledge, GlcCer or GlcSph brain levels in PS+/+ NA 4L mice have not been previously 13 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

reported. Although we did not observe any brain GlcCer elevation in these mice, we did see a minimal 15-fold elevation in brain GlcSph levels relative to wild-type mice. Historically, both sphingolipids have been shown to be elevated in GD patients, though more recently GlcSph is becoming more widely accepted as a better biomarker for GD6-8. Therefore, it was not completely surprising to find that GlcCer brain levels were elevated in only one of the five GD models (PS-/- NA 4L). In sharp contrast to the similar GlcCer levels between strains, GlcSph levels were elevated in all five GD models, providing further support for the hypothesis that GlcSph is a more sensitive biomarker.

Figure 3. GlcCer and GlcSph Levels in Brain Tissue from Different Wild-Type and GD Mouse Strains. (A) GlcCer brain tissue levels in eight groups of wild-type mice (of varying ages and strains) and five GD mouse strains. (B) GlcSph brain levels in the same mice from Figure 3A. WT1=6-week-old FVB/N (n=6), WT2=8-week-old FVB/N (n=4), WT3=8-week-old 14 ACS Paragon Plus Environment

Page 14 of 20

Page 15 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

C57BL/6 (n=5), WT4=9-week-old C57BL/6 (n=7), WT5=10-week-old C57BL/6 (n=7), WT6=4week-old C57BL/6-DBA/2 (n=6), WT7=20-week-old C57BL/6-DBA/2 (n=17), WT8=60-weekold C57BL/6-DBA/2 (n=20), GBA L444P=18-week-old GBAL444P/L444P (n=10), hL444P-tg8=14week-old hL444P-tg8 (n=7), hN370S-tg4=14-week-old hN370S-tg4 (n=7), PS+/+ NA 4L= 20week-old PS+/+ NA 4L (n=8), and PS-/- NA 4L= 20-week-old PS-/- NA 4L (n=7). In recent years, there has been considerable interest in better understanding the origins and dispositions of both GlcCer and GlcSph. It is well understood that GlcCer exists in both the lysosome as well as other parts of the cell, and that two different β-glucosidase enzymes are responsible for its metabolism26. GCase is the lysosomal form and has been shown to metabolize both GlcCer and GlcSph, although the maximum velocity (Vmax) for GlcSph is considerably lower than for GlcCer27. GBA2 is known to metabolize non-lysosomal GlcCer; however, it is unclear if GlcSph is a substrate for GBA2. Further, it has been shown that GlcCer is deacylated in lysosomes by acid ceramidase28-30. We used our recently developed methodologies to better understand the connection between these two sphingolipids in brain tissue taken from mice after administration of the GCase inhibitor conduritol β-epoxide (CBE), or with an inhibitor of nonlysosomal GBA2, N-butyl-deoxy-galactonojirimycin (NB-DGJ). In vitro studies have shown that CBE selectively inhibits GCase31-34, while NB-DGJ selectively inhibits GBA235-36. For wild-type mice receiving CBE, approximate 40- and 400-fold increases were seen in GlcCer and GlcSph brain levels, respectively (Figure 4), relative to untreated mice (p < 0.05). These findings are consistent with a previous report37, as well as with our earlier work showing that increased levels of GlcSph result from both increased production of GlcSph by acid ceramidasemediated deacylation of GlcCer, and decreased metabolism of GlcCer due to GCase inhibition29. However, results were quite different when wild-type mice were administered NB-DGJ (Figure 5). A significant, dose-dependent increase in brain GlcCer levels was observed up to 100 mg/kg, with no additional increase seen at 300 mg/kg, likely due to inhibition of GlcCer synthase38 at higher NB-DGJ concentrations. Analysis showed that all NB-DGJ treatment groups have elevated brain GlcCer levels relative to the untreated group (p < 0.05). For GlcSph however, negligible increases were observed with NB-DGJ administration, none of which were found to be statistically significant, suggesting that non-lysosomal enzymatic deacylation of GlcCer is much less efficient than lysosomal deacylation. These results provide further evidence for

15 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

changes in GlcSph levels being more closely tied to lysosomal dysfunction than changes in GlcCer levels. Taken together with results from the GD mouse model measurements, these data extend the hypothesis that GlcSph is a sensitive and selective biomarker for neuronopathic GD, and may warrant further investigation in a clinical setting.

Figure 4. GlcCer and GlcSph in Brain Tissue from Wild-Type Mice Administered CBE. (A) GlcCer brain levels in four-week-old FVB/N mice that received 100 mg/kg CBE daily for four weeks by intraperitoneal injection. (B) GlcSph brain levels in the same mice from Figure 4A.

16 ACS Paragon Plus Environment

Page 16 of 20

Page 17 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

Figure 5. GlcCer and GlcSph in Brain Tissue from Wild-Type Mice Administered NBDGJ. (A) GlcCer brain levels in five-week-old male C57BL/6 mice that were orally administered in drinking water for 7 days 3, 10, 30, 100, or 300 mg/kg of NB-DGJ and sacrificed on day 8 (B) GlcSph brain levels in the same mice from Figure 5A.

1. Ballabio, A.; Gieselmann, V. BBA-Mol Cell Res. 2009, 1793, 684-696. 2. Cox, T. M.; Schofield, J. P., Baillieres Clin. Haematol. 1997, 10, 657-689. 3. Erikson, A.; Bembi, B.; Schiffmann, R., Baillieres Clin. Haematol. 1997, 711-723. 4. Nilsson, O.; Svennerholm, L., J. Neurochem. 1982, 709-718. 5. Goker-Alpan, O., Ther. Clin. Risk Manag. 2010, 315-323. 6. Dekker, N.; van Dussen, L.; Hollak, C. E. M.; Overkleeft, H.; Scheij, S.; Ghauharali, K.; van Breemen, M. J.; Ferraz, M. J.; Groener, J. E. M.; Maas, M.; Wijburg, F. A.; Speijer, D.; TylkiSzymanska, A.; Mistry, P. K.; Boot, R. G.; Aerts, J. M., Blood 2011, e118-e127. 7. Rolfs, A.; Giese, A.-K.; Grittner, U.; Mascher, D.; Elstein, D.; Zimran, A.; Böttcher, T.; Lukas, J.; Hübner, R.; Gölnitz, U.; Röhle, A.; Dudesek, A.; Meyer, W.; Wittstock, M.; Mascher, H., PLoS ONE 2013, e79732. 17 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

8. Fuller, M.; Szer, J.; Stark, S.; Fletcher, J. M., Clin. Chim. Acta 2015, 6-10. 9. Shibuya, H.; Kawashima, K.; Narita, N.; Ikeda, M.; Kitagawa, I., Chem. Pharm. Bul.l 1992, 1154-1165. 10. Liu, Y.; Suzuki, K.; Reed, J. D.; Grinberg, A.; Westphal, H.; Hoffmann, A.; Doring, T.; Sandhoff, K.; Proia, R. L., Proc. Natl. Acad. Sci. U. S. A. 1998, 2503-2508. 11. Mizukami, H.; Mi, Y.; Wada, R.; Kono, M.; Yamashita, T.; Liu, Y.; Werth, N.; Sandhoff, R.; Sandhoff, K.; Proia, R. L., J. Clin. Invest. 2002, 1215-1221. 12. Sanders, A.; Hemmelgarn, H.; Melrose, H. L.; Hein, L.; Fuller, M.; Clarke, L. A., Blood Cells. Mol. Dis. 2013, 109–115. 13. Sun, Y.; Quinn, B.; Witte, D. P.; Grabowski, G. A., J. Lipid Res. 2005, 2102-2113 14. Brignol, N.; Chang, K.; Hamler, R.; Schilling, A. E.; Khanna, R.; Lockhart, D. J.; Clark, S. W.; Benjamin, E. R., Mol. Genet. Metab. 2012, S22. 15. Hamler, R.; Brignol, N.; Morrison, S.; Hui, C. H.; Dungan, L.; Boyd, R. E.; Clark, S. W.; Khanna, R.; Wustman, B. W.; Flanagan, J. F.; Valenzano, K. J.; Lockhart, D. L.; Benjamin, E. R., Mol. Genet. Metab. 2014, S52. 16. Orvisky, E.; Sidransky, E.; McKinney, C. E.; Lamarca, M. E.; Samimi, R.; Krasnewich, D.; Martin, B. M.; Ginns, E. I., Pediatr. Res. 2000, 233-237. 17. Hannun, Y.; Bell, R., Science 1987, 670-674. 18. Igisu, H.; Hamasaki, N.; Ito, A.; Ou, W., Lipids. 1988, 345-348. 19. Spiegel, S.; Merrill, A. H., The FASEB Journal 1996, 1388-1397. 20. Mistry, P. K.; Liu, J.; Yang, M.; Nottoli, T.; McGrath, J.; Jain, D.; Zhang, K.; Keutzer, J.; Chuang, W. L.; Mehal, W. Z.; Zhao, H.; Lin, A.; Mane, S.; Liu, X.; Peng, Y. Z.; Li, J. H.; Agrawal, M.; Zhu, L. L.; Blair, H. C.; Robinson, L. J.; Iqbal, J.; Sun, L.; Zaidi, M., Proc. Natl. Acad. Sci. U. S. A. 2010, 19473-19478. 21. Atsumi, S.; Nosaka, C.; Iinuma, H.; Umezawa, K., Arch. Biochem. Biophys. 1993, 302-304. 22. Schueler, U. H.; Kolter, T.; Kaneski, C. R.; Blusztajn, J. K.; Herkenham, M.; Sandhoff, K.; Brady, R. O., Neurobiol. Dis. 2003, 595-601. 23. Sun, Y.; Liou, B.; Ran, H.; Skelton, M. R.; Williams, M. T.; Vorhees, C. V.; Kitatani, K.; Hannun, Y. A.; Witte, D. P.; Xu, Y.-H.; Grabowski, G. A., Neuronopathic Hum. Mol. Genet. 2010, 1088-1097. 24. Shaner, R. L.; Allegood, J. C.; Park, H.; Wang, E.; Kelly, S.; Haynes, C. A.; Cameron Sullards, M.; Merrill, A. H., Jr., J. Lipid Res. 2009, 1692–1707. 25. Sango, K.; Yamanaka, S.; Hoffmann, A.; Okuda, Y.; Grinberg, A.; Westphal, H.; McDonald, M. P.; Crawley, J. N.; Sandhoff, K.; Suzuki, K.; Proia, R. L., Nat. Genet. 1995, 170-176. 26. Yildiz, Y.; Matern, H.; Thompson, B.; Allegood, J. C.; Warren, R. L.; Ramirez, D. M.; Hammer, R. E.; Hamra, F. K.; Matern, S.; Russell, D. W., J. Clin. Invest. 2006, 2985-2994. 27. Vaccaro, A. M.; Muscillo, M.; Suzuki, K., Eur. J. Biochem. 1985, 315-321. 28. Yamaguchi, Y.; Sasagasako, N.; Goto, I.; Kobayashi, T., J. Biochem. 1994, 704-710. 29. Flanagan, J.; Ranes, B.; Brignol, N.; Hamler, R.; Clark, S., Mol. Genet. Metab. 2013, S40S41. 30. Ferraz, M. J.; Kallemeijn, W. W.; Mirzaian, M.; Herrera Moro, D.; Marques, A.; Wisse, P.; Boot, R. G.; Willems, L. I.; Overkleeft, H. S.; Aerts, J. M., BBA-Mol. Cell Biol. L. 2014, 811– 825. 31. Daniels, L. B.; Coyle, P. J.; Glew, R. H.; Radin, N. S.; Labow, R. S., Arch. Neurol. 1982, 550-556.

18 ACS Paragon Plus Environment

Page 18 of 20

Page 19 of 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

32. van Weely, S.; Brandsma, M.; Strijland, A.; Tager, J. M.; Aerts, J. M., Biochim. Biophys. Acta. 1993, 55-62. 33. Overkleeft, H. S.; Renkema, G. H.; Neele, J.; Vianello, P.; Hung, I. O.; Strijland, A.; van der Burg, A. M.; Koomen, G. J.; Pandit, U. K.; Aerts, J. M., J. Biol. Chem. 1998, 26522-26527. 34. Boot, R. G.; Verhoek, M.; Donker-Koopman, W.; Strijland, A.; van Marle, J.; Overkleeft, H. S.; Wennekes, T.; Aerts, J. M. F. G., J. Biol. Chem. 2007, 1305-1312. 35. Platt, F. M.; Neises, G. R.; Dwek, R. A.; Butters, T. D., J. Biol. Chem. 1994, 8362-8365. 36. Wennekes, T.; Meijer, A. J.; Groen, A. K.; Boot, R. G.; Groener, J. E.; van Eijk, M.; Ottenhoff, R.; Bijl, N.; Ghauharali, K.; Song, H.; O’Shea, T. J.; Liu, H.; Yew, N.; Copeland, D.; van den Berg, R. J.; van der Marel, G. A.; Overkleeft, H. S.; Aerts, J. M., J. Med. Chem. 2009, 689-698. 37. Marshall, J.; Sun, Y.; Bangari, D. S.; Budman, E.; Park, H.; Nietupski, J. B.; Allaire, A.; Cromwell, M. A.; Wang, B.; Grabowski, G. A.; Leonard, J. P.; Cheng, S. H., Mol. Ther. 2016, 1019-1029. 38. Walden, C. M.; Sandhoff, R.; Chuang, C.-C.; Yildiz, Y.; Butters, T. D.; Dwek, R. A.; Platt, F. M.; van der Spoel, A. C., J. Biol. Chem. 2007, 32655-32664.

All work presented herein was funded by Amicus Therapeutics.

19 ACS Paragon Plus Environment

Analytical Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

For Table of Contents Only Brain Glycosphingolipids

Gaucher Mouse

20 ACS Paragon Plus Environment

Page 20 of 20