Biochemical signatures of disease severity in Multiple Sulfatase Deficiency

Summary

Sulfatases catalyze essential cellular reactions, including degradation of glycosaminoglycans (GAGs). All sulfatases are post-translationally activated by the formylglycine generating enzyme (FGE) which is deficient in Multiple Sulfatase Deficiency (MSD), a neurodegenerative lysosomal storage disease. Historically, patients were presumed to be deficient of all sulfatase activities; however, a more nuanced relationship is emerging. Each sulfatase may differ in their degree of post-translational modification by FGE, which may influence the phenotypic spectrum of MSD.

Here, we evaluate if residual sulfatase activity and accumulating GAG patterns distinguish cases from controls and stratify clinical severity groups in MSD. We quantify sulfatase activities and GAG accumulation using three complementary methods in MSD participants. Sulfatases differed greatly in their tolerance of reduction in FGE-mediated activation. Enzymes that degrade heparan sulfate (HS) demonstrated lower residual activities than those that act on other GAGs. Similarly, HS-derived urinary GAG subspecies preferentially accumulated, distinguished cases from controls, and correlated with disease severity.

Accumulation patterns of specific sulfatase substrates in MSD provide fundamental insights into sulfatase regulation and will serve as much-needed biomakers for upcoming clinical trials. This work highlights that biomarker investigation of an ultra-rare disease can simultaneously inform our understanding of fundamental biology and advance clinical trial readiness efforts.

INTRODUCTION

Multiple sulfatase deficiency (MSD) is a severe, progressive disease of early childhood that results in substantial multisystemic burden of disease and neurologic regression¹. MSD is ultra-rare, with fewer than 150 cases in the literature ² and an estimated prevalence of approximately 1/500,000 individuals³. The natural history of MSD suggests two distinct severity cohorts: a severe form characterized by profound global impairment and an attenuated form characterized by acquisition of ambulation followed by neurologic regression^1,2,4 The broad clinical spectrum of MSD is only partially explained by genetics and residual enzyme activity. As trials of disease-modifying treatments are in development for MSD, ^4-6 there is an urgent, unmet need for a disease specific biomarker pattern that can differentiate the severe and attenuated phenotypes and serve as a marker of effective intervention.

MSD arises from pathogenic variants in SUMF1, which encodes the formylglycine generating enzyme (FGE). Previously characterized pathogenic variants primarily affect the stability of FGE enzyme^1,2,7. FGE is located in the endoplasmic reticulum and activates all cellular sulfatases by converting a conserved active site-cysteine residue to a formylglycine residue. This includes enzymes associated with mucopolysaccharidosis (MPS) disorders including N-acetylglucosamine 6-sulfatase (GNS), iduronate 2-sulfatase (IDS), sulfamidase (SGSH), N-acetylgalactosamine-6-sulfatase (GALNS) and arylsulfatase B (ARSB). MPS-related enzymes are responsible for the degradation of glycosaminoglycans (GAGs), long, unbranched polysaccharides components of proteoglycans ^8-11. These enzymes can be further classified by GAG substrates including: heparan sulfate (HS), chondroitin sulfate (CS), or dermatan sulfate (DS). When FGE is deficient, individual sulfatases may be activated to different extents; therefore, in MSD not all GAG subspecies may accumulate to the same extent. It remains unknown which GAG species are the predominant drivers of MSD pathology.

GAG analysis methods differ in sensitivity and specificity. Indirect methods of GAG quantification do not differentiate accumulation of specific GAG species ¹². Furthermore, healthy individuals can have substantial total GAG levels ¹³, which limits the utility of total GAGs as a biomarker for disease severity and diagnosis. Total GAG levels can be elevated in individuals with MSD, but within our previously-reported MSD cohort, some mild cases were noted to have normal total GAG levels^1,2. However, patterns of GAG accumulation in MSD patients have not been well studied.

While total GAGs may not be a useful MSD biomarker, measurement of specific GAG degradation products could prove more informative. GAG degradation occurs in a step-wise fashion, with distinct enzymes modifying or removing monosaccharides components from the non-reducing end of the GAG chain. Specific GAG nonreducing end species (GAG-NREs) have been validated as biomarkers for diagnosis, severity classification, and response to therapy in several MPS disorders ^{8,9,11,12,14-17}.

In single sulfatase MPS disorders, a single enzyme is deficient leading to accumulation of that enzyme’s specific GAG-NRE substrate. In MSD, multiple enzymes are deficient, so we hypothesize that several GAG-NREs will accumulate. The level of residual enzyme activity may vary for each sulfatase in MSD ^1,2, suggesting that sulfatases differ in their degree of FGE-mediated activation. The relative abundance of each GAG-NREs could indicate which sulfatases are most vulnerable to loss of FGE activity.

Multiple methods have been developed to evaluate GAG degradation products. Here we utilize three previously reported techniques: 1) the “internal disaccharide” method ^18,19, where GAG polymers are degraded in vitro with bacterial enzymes into resultant disaccharides, which are quantified via LC-MS/MS; 2) the “Sensi-Pro^®” method ^8,11,20, which quantifies the GAG non-reducing end (GAG-NRE) fragment via LC-MS/MS after its in vitro liberation from the GAG polymer via bacterial enzymes; and 3) the “endogenous NRE” method ^15-17, which quantifies small endogenously-produced GAG-NRE fragments from patient samples without in vitro degradation. These techniques how variable sensitivity and specificity when applied to single sulfatase disorders. Previous studies have suggested that analysis of endogenous GAG-NREs may provide greater differentiation between cases and controls as compared to the internal disaccharide and Sensi-Pro^® method ^15,17.

Here we combine analysis of sulfatase activities and GAG subspecies, using all three methods described above, in samples obtained from severe and attenuated MSD patients. We hypothesize that biomarker patterns in MSD patients will provide key insights into basic sulfatase biology and allow for the stratification of trial participants and establish inclusion/exclusion criteria for upcoming clinical trials. Our aim is to validate the use of GAG subspecies as a diagnostic and prognostic biomarker in MSD.

MATERIALS AND METHODS

Clinical severity

Clinical severity was classified as previously reported by age at symptom onset and genotype severity ¹. For each participant, we assigned a SUMF1 “genotype score” (Table 1) as previously described¹. Missense variants with biochemical or in silico modeling data supporting mild functional disruption were assigned a score of 1. Missense severe and nonsense alleles were each assigned two points each. As such, each participant could be assigned a total genotype score of 2 to 4. Individuals with a score of 2 were assigned an “attenuated” phenotype. Participants with a score of 3 or 4 were considered “severe”. Clinical severity was also reviewed from parental report (for MSD Biobank participants) or natural history records (for MDBP participants) to ensure that the reported severity matched genotype severity. For the subset of participants with GAG-NRE measurements, the correlation of GAG-NRE levels with motor symptom severity was captured using the Gross Motor Function Classification (GMFC-MLD)²¹. This scale was applied retrospectively using information available in the natural history dataset as previously described¹.

Table 1: Clinical classifications and MSD genotypes.

Subject	Severity Classification	Genotype score	Variant 1	Variant 1 score	Variant 2	Variant 2 score	Enzyme Analysis	Endogenous Biomarker GAG- NRE Analysis	Sensi-Pro GAG-NRE Analysis
1	attenuated	2	c.726-1_726delGA	1	c.640G>A (p.A214T)	1	yes	no	yes
2	severe	3	c.836C>T (p.A279V)	1	c.890A>C (p.N297T)	2	yes	yes	yes
3	severe	4	c.463C>T (p.S155P)	2	c.539G>T (p.W180L)	2	yes	yes	yes
4	severe	3	chr3:4404903-4429402 x 1 (GRCH37/hg19)	2	c.836C>T (p.A279V)	1	yes	yes	yes
5	severe	4	c.1113C>A (p.S359*)	2	c.463C>T (p.S155P)	2	yes	yes	no
6	attenuated	2	c.519+5_519+*delfGTAA	1	c.817G>A (p.D273N)	1	yes	no	yes
7	attenuated	2	c.519+5_519+*delfGTAA	1	c.817G>A (p.D273N)	1	yes	no	yes
8	severe	3	c.836C>T (p.A279V)	1	c.1091G>A (p.R364H)	2	yes	yes	yes
9	severe	4	c.374delA(p.Y125Sfs*16)	2	c.655A>C (p.T219P)	2	yes	yes	yes
10	attenuated	2	c.836C>T (p.A279V)	1	c.836C>T (p.A279V)	1	yes	yes	no
11	severe	3	c.1034G>A (p.R345H)	1	c.463C>T (p.S155P)	2	yes	yes	no
12	severe	3	c.836C>T (p.A279V) and c.797C>T(p.P266L)	1	c.776A>T (p.N259I)	2	no	no	yes
13	severe	3	chr3:4404949-4022886 x 1(GRCH37/hg19)	2	c.836C>T (p.A279V)	1	no	no	yes

Biochemical and in silico analysis of novel variants

cDNA and Expression constructs

All expression constructs used for biochemical and functional characterization were generated in the pBIX plasmid, a modified expression plasmid based on the pBI vector backbone that allows expression of two proteins under a doxycycline-inducible bi-directional promotor. The generation of pBIX plasmid and pBIX-FGE-HA expression plasmid was described earlier¹. The expression construct encoding MSD-causing FGE variant FGE-T219P, was generated by site-directed mutagenesis PCR using pBIX-FGE-HA as template and primers (FGE-T219P_Fwd, 5’-GCGGTTGCCTACTGCCCTTGGGCAGGGAAGCGG-3’, FGE- T219P _Rev, 5’-CCGCTTCCCTGCCCAAGGGCAGTAGGCAACCGC-3’). Either human steroid sulfatase (STS) or C-terminally His-tagged human Arylsulfatase A (ARSA-His) cDNA were cloned into MCS-I of pBIX vector or pBIX-FGE-HA to generate pBIX-STS/-ARSA or pBIX-FGE-HA+(STS/ARSA) plasmids, respectively. All expression plasmids were verified by sequencing of the entire coding region to exclude any undesired PCR-derived errors.

Cell culture and transfection

Previously described stable TetOn cell lines of HT1080 fibrosarcoma cells and immortalized MSD (MSDi) patient skin fibroblasts were used¹. The cells were maintained under 5% CO₂ at 37 °C in Dulbecco’s modified Eagle’s medium supplemented with 10% fetal calf serum, 1% penicillin/streptomycin (Invitrogen) and 800 μg/ml G418 (HT1080-TetOn cells) and 5 μg/ml blasticidin (MSDi- TetOn cells). Transfection of plasmids was performed with PEI and 4-6 h post-transfection, protein expression was induced with medium containing 1 μg/ml doxycycline for 22-24 h after which the cells were harvested.

In vitro sulfatase activity assays, intracellular stability analysis and western blotting

Biochemical and functional assays were performed as previously described ¹. Sulfatase activity assays were performed in MSDi-TetOn cell lysates that transiently expressed either the sulfatase alone or the sulfatase co-expressed together with FGE from pBIX expression plasmids. For western blot analyses, 75 μg of total protein were resolved by SDS-PAGE and proteins were transferred to PVDF membrane followed by incubation with respective antibodies diluted in blocking buffer (5% milk powder in 1X PBS). Rabbit polyclonal antisera against FGE (1:3000) and STS (1:5000) or mouse monoclonal anti-RGS-His antibody (1:2000; Qiagen #34650) for ARSA-His were used as primary antibodies and HRP-conjugated goat anti-rabbit antibodies or goat anti-mouse antibodies were used as secondary antibodies (diluted 1:10,000). The signals corresponding to sulfatases were normalized to Hsc70 signals (loading control) and used for specific activity calculation. Western blot signals were quantified using the AIDA 2.1 software (Raytest).

Cycloheximide-chase based intracellular protein stability analyses were performed in HT1080-TetOn cells that were transiently transfected with either pBIX-FGE-WT-HA or the variant pBIX-FGE-T219P-HA. 4-6 h post-transfection, the protein expression was induced by addition of 0.5 μg/ml doxycycline and allowed protein expression for 2.5 h. The doxycycline containing medium was removed, cells washed twice with 1X PBS to remove excess doxycycline. The cells were harvested immediately (0h) or incubated further with 1 ml media containing 250 μg/ml cycloheximide. At timepoints 2h, 4h and 6h the cells and media were harvested. To analyze the amount of FGE in the cells and medium, 75 μg of the cell lysate and immunoprecipitated FGE from the medium were resolved by SDS-PAGE and proteins transferred to PVDF membrane by western blotting. The membrane was probed with rat monoclonal anti-Hsc70 (1:5000; Thermofischer Scientific #MA1-26078) as loading control and mouse monoclonal anti-HA antibodies (1:2000; Roche; clone 3F10). The western blot signals were quantified using AIDA 2.1 software (Raytest). After normalizing the signals corresponding to FGE-HA in the cell lysates to Hsc70, the amount of FGE in both the cells and medium per mg of total protein were compared.

Sulfatase activity measurement

Sulfatases were quantified using a multiplex-liquid chromatography-tandem mass spectrometry (LC-MS/MS) method as previously published ²². Assays using dried blood spots were carried out with a 3 mm punch per assay. For leukocyte assays, mixed leukocytes from whole blood were isolated as described²³. Frozen mixed leukocyte pellets were thawed in assay buffer and vortexed for ~30 sec at room temperature. Samples were centrifuged at room temperature at ~12,000xg for 5 min to pellet insolubles, and protein in the supernatant was quantified with a Pierce micro-BCA protein assay kit (Thermo-Fisher) using bovine serum albumin as a standard. Typically, 2-5 ug of protein were used per assay.

Internal disaccharide and endogenous biomarker measurement

Quantification of urinary GAGs using both the internal disaccharide^18,19 and endogenous GAG-NRE biomarker¹⁶ methods was completed as previously described^15,17. Quantification of GAG endogenous biomarkers in MSD patient DBS was also completed. Full standard operating procedures are provided in the Supplementary Methods. For a full list of GAG biomarkers see Sup Table 1.

Sensi-Pro^® GAG-NRE measurement

Testing for total GAGs, heparan sulfate (HS), chondroitin/dermatan sulfate (CS), MPS II GAG-NRE (I2S0), MPS IVA GAG-NRE (a6) and MPS VI GAG-NRE (a4) in urine was performed at ARUP as a CLIA-certified clinical test as previously published ^8,11,20. Heparan sulfate, chondroitin/dermatan sulfate, and disease specific Non-Reducing Ends (NREs) were tested using the LC-MS/MS based Sensi-Pro^® assay⁸. Briefly, glycosaminoglycans were extracted from urine, enzymatically digested, identified, and quantified by LC-MS/MS. Total heparan sulfate concentration was calculated as the sum of the two internal disaccharides: D0A0 and D0S0; total chondroitin/dermatan sulfate was calculated as the sum of the three internal disaccharides: D0a0, D0a4, and D0a6. In addition, total GAGs were measured with Dimethylmethylene Blue (DMB) based assay²⁴.

Statistics

Given the small sample size, descriptive statistics, including means and standard errors of the mean are presented. To evaluate if biomarkers distinguish clinical groups, the minimal differentiation factor was calculated. This value represents the minimum fold difference between controls and disease states for biomarker (sulfatases: lowest control value/highest MSD case value; GAG levels: lowest case value/highest control value).

Study approval

All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration. Informed consent was obtained from all patients included in the study. Individuals with clinical and molecular confirmation of MSD were enrolled in one or both of the following IRB-approved protocols: 1) Myelin Disorders Biorepository Project (ClinicalTrials.gov Identifier NCT03047369, CHOP IRB #14-011236), 2) MSD Biobank (Genetic Alliance IRB #MSD-001).

RESULTS

Clinical characterization

Our cohort included 13 participants, including 9 severe and 4 attenuated cases (Table 1). The clinical and genetic information for each participant was collected. The genetic severity was assigned based on previously published work^1,2.

One participant (Table 1, Participant #9) harbored a novel missense variant, SUMF1 c.655C>A (p.T219P). This variant is absent from exome and genome sequencing data from control populations in gnomAD²⁵ and is located in a region of the protein with several other known pathogenic variants¹. To determine the functional effect of this novel variant, sulfatase activity was analyzed in MSDi-TetOn cell lysates that transiently express either the sulfatases STS or ARSA alone or together with wildtype FGE (FGE-WT) or the variant FGE-T219P (Fig 1A-B). Comparison of the specific activities of sulfatases revealed that the T219P variant showed a very low residual activity with less than 10% as active as FGE-WT towards STS and almost inactive towards ARSA (Fig 1C). These data indicate that the T219P mutation results in a severe reduction of FGE function in activating sulfatases.

Figure 1: The SUMF1 missense variant c.655A>C (p.T219P) severely impacts FGE stability and function.

MSDi-TetOn cells were transiently transfected with plasmids encoding the sulfatases (A) STS or (B) ARSA either alone or together with FGE-WT or FGE-T219P in the indicated fashion. After 24 h of doxycycline-induced expression, the activity of sulfatases was measured in respective cell lysates and specific activity calculated with amounts of sulfatases quantified from western blot after normalization of signals with Hsc70 (loading control). The data shown are representative of three independent experiments (n=3). C) Bar diagram depicting the comparison of the specific activity of analysed sulfatases in cells expressed in percentage, relative to activity of FGE-WT coexpressing cells as 100%. The values represent mean ± SEM of three independent experiments (n=3). D) HT1080-TetOn cells were transiently transfected with plasmids encoding either FGE-WT-HA (upper panel) or FGE-T219P-HA (lower panel). After 2.5 h of protein expression (induced with 0.5 μg/ml doxycycline) the cells were subjected to cycloheximide (CHX) chase analysis for 6 h with cells (C) and media (M) collected at the indicated time points. Cell lysates and FGE-antiserum-immunoprecipitates from media were resolved in SDS-PAGE and western blot decorated with anti-HA (to detect FGE) and anti-Hsc70 (as loading control). E) Plot depicting the percentage of FGE remaining after quantification of data from CHX chase analysis. After normalization of the anti-HA antibody signals (corresponding to FGE) to anti-Hsc70 signals, the total amount of FGE expressed as the percentage of that at the start of the chase (0 h).

Next, we analyzed the intracellular stability of FGE-T219P in comparison to FGE-WT by cycloheximide chase analysis. HT1080-TetOn cells transiently expressing FGE-WT or the FGE-T219P variant were treated with 250 μg/ml cycloheximide for up to 6 h and cells and medium collected at the indicated timepoints were analyzed by SDS-PAGE and western blot (Fig 1D). Comparison of the total FGE after quantification of FGE specific signals in cells and media showed that, while the amount of FGE-WT did not show a significant change over time, the amount of the T219P variant drastically decreased to about 50% in 2 h and gradually decreasing over time, exhibiting a half-life of around 2 h (Fig 1E). These data clearly indicate that the T219P mutation is detrimental to the intracellular stability of FGE rendering the protein highly unstable and thus destined for degradation.

Genotype severity was aligned with clinical severity. Participants with a genotype score of 3-4 presented with a severe phenotype, while those with scores of 2 presented with an attenuated phenotype.

Residual sulfatase activities in MSD patients

Next, we sought to evaluate biomarker levels in MSD cases and controls. As sample sizes were limited due the ultra-rare nature of MSD, we present descriptive statistics and describe observed between group differences. Multiplex tandem mass spectrometry-based analysis of enzyme activity levels in leukocytes from the severe MSD cohort demonstrated reduced activity of all quantified sulfatases (Fig 2). When comparing the attenuated and severe cohorts, the activities of N-acetylglucosamine 6-sulfatase (GNS, Fig 2A), arylsulfatase A (ARSA, Fig 2B), iduronate 2-sulfatase (IDS, Fig 2C), and sulfamidase (SGSH, Fig 2D) were reduced in both subgroups. Within the attenuated cohort, the activity of N-acetylgalactosamine-6-sulfatase (GALNS, Fig 2E) and arylsulfatase B (ARSB, Fig 2F) overlapped with controls. The activity of non-sulfatase enzymes was not affected across all MSD cases (Sup Fig 1).

Figure 2: Sulfatase activity levels in leukocytes differentiate MSD cases from controls, but individual enzymes differ in their residual activity after loss of FGE.

(A-D) Activity levels of n-acetylglucosamine 6-sulfatase (GNS), arylsulfatase A (ARSA), iduronate 2-sulfatase (IDS) and sulfamidase (SGSH) are severely reduced in both attenuated and severe cases. (E-F) Residual N-acetylgalactosamine-6-sulfatase (GALNS) and arylsulfatase B (ARSB) activities were higher as compared to other sulfatases, with attenuated case levels overlapping with control levels. (G) Varying degrees of residual sulfatase activity indicate that each sulfatase has a differential vulnerability to loss of FGE activity. Residual enzyme activities (individual case values normalized to the mean of control activity) can also help differentiate severe from attenuated cases. For all enzymes, the associated single sulfatase disorder is listed in parenthesis.

Comparing residual activity of individual sulfatases (Fig 2G) revealed that enzymes responsible for HS degradation as compared to those responsible for CS or DS degradation are more susceptible to loss of FGE activity. In addition, residual ARSA level was very low. The patterns of residual enzyme activity differentiated severe from attenuated MSD cases, suggesting a differential sensitivity to loss of FGE and the correlation with phenotype.

Internal disaccharide GAG subspecies as diagnostic biomarkers

Next, we sought to quantify GAG accumulation in urine samples from MSD patients. We first analyzed GAGs using the internal disaccharide method as previously described^15,17-19. Quantification of internal disaccharide GAG species derived from DS (D0a4, Fig 3A) and CS (D0a4 and D0a6, Fig 3B) did not differentiate cases from controls (see Sup Table 1 for full list of GAG markers). This is consistent with the relatively higher residual GALNS and ARSB activity levels observed in leukocytes from MSD patients (Fig 2G). However, it is important to note that CS and DS-derived internal disaccharide GAG species in both cases and controls had a high degree of variability.

Figure 3: Urinary heparan sulfate levels measured using the internal disaccharide method differentiate severe MSD cases from controls, while chondroitin sulfate and dermatan sulfate levels do not.

Using the internal disaccharide method, GAGs were quantified in urine from MSD patients. (A-B) The D0a4 and D0a6 markers of chondroitin sulfate and dermatan sulfate did not differ between cases and controls. (C-E) Conversely, heparan sulfate-derived species including D0A0, D0S0, and D0A6 differentiated severe cases from controls. The single attenuated participant evaluated had levels in between control and severe cases. For all markers, the associated single sulfatase disorders known to have marker elevations are listed in parenthesis.

Unlike DS and CS-derived GAG species, levels of HS-derived GAGs differed between MSD cases and controls, including D0A0 (Fig 3C), D0S0 (Fig 3D), and D0A6 subspecies (Fig 3E). There was only one sample available from an attenuated case for this analysis, but it demonstrated intermediate levels of HS-derived GAGs.

GAG endogenous NRE biomarkers as diagnostic biomarkers

GAG-NREs were measured with the endogenous NRE biomarker method in urine specimens from MSD patients and controls. We found that, unlike in the case of internal disaccharide biomarkers, all sulfatase-related endogenous biomarker species differed between severe MSD cases and controls. Specifically, both HS-derived endogenous biomarkers (Fig 4A-C) and CS-derived biomarkers (Fig 4D-E) were elevated in severe MSD cases as compared to controls, with the exception of a single MSD patient that had a normal level of the MPSIIID biomarker. The one available sample from an attenuated case demonstrated levels that fell between severe cases and controls for all endogenous biomarkers measured.

Figure 4: Endogenous biomarker GAG-NRE subspecies in urine are diagnostic biomarkers for MSD and highlight that sulfatases differ in their susceptibility to loss of FGE activity.

Using the endogenous biomarker method, GAG-NRE species were quantified in urine from MSD patients. The GAG-NRE species derived from HS including the biomarkers of (A) IDS deficiency (MPS II), (B) SGSH deficiency (MPS IIIA), and (C) GNS deficiency were all elevated in severe cases as compared to controls. In addition, chondroitin sulfate-derived biomarkers were also elevated in severe cases including the biomarkers of (D) GALNS (MPS IVA) and (E) ARSB (MPS VI) deficiencies. The single attenuated participant evaluated had levels in between control and severe cases. (F) Varying degrees of GAG-NRE accumulation indicates that each sulfatase has a differential vulnerability to loss of FGE activity. The fold increase in severe cases over the mean of the control specimens are shown as mean +/− SEM.

Evaluation of the fold increase in GAG-NRE accumulation in severe MSD patients (Fig 4F) again revealed that enzymes responsible for HS degradation (as opposed to CS or DS degradation) are more susceptible to loss of FGE activity. Similar patterns were seen in dried blood spot samples (Sup Fig 2).

Sensi-Pro^® GAG-NRE species as severity biomarkers

To evaluate if disease-associated urinary GAG-NRE differences could be reproduced using an alternative method, we next completed Sensi-Pro^® GAG-NRE biomarker analysis in samples from severe and attenuated patients. Analysis was completed using a CLIA-certified clinical assay that included measurement of: total GAG levels as measured by the dimethylmethylene blue method; CS/DS, and HS levels measured by the internal disaccharide method; and GAG-NRE subspecies measured by the Sensi-Pro^® assay. Participant samples were compared to normative values obtained from healthy controls in the clinical lab (dotted line, Fig 5). Total GAG levels were normal in most attenuated patients and only mildly elevated in severe cases (Fig 5A). Consistent with the internal disaccharide analysis, CS/DS levels were essentially normal (Fig 5B), as were the Sensi-Pro^® MPS IVA GAG-NRE (a6) and MPS VI GAG-NRE (a4) markers derived from CS (Fig 5C-D).

Figure 5: Heparan sulfate derived Sensi-Pro^® GAG-NRE subspecies in urine are disease severity biomarkers and correlate with motor outcomes in MSD.

(A) Total urinary GAGs were measured by the dimethylmethylene blue method. Total GAG levels in attenuated MSD patients were within the normal range (as shown by dotted line in all panels), while severe cases demonstrated elevated total GAG levels. (B) Combined CS and DS levels as measured by the sum of internal disaccharide markers were normal in most MSD cases. Similarly, the CS-derived (C) MPS IVA GAG-NRE and (D) MPS IV GAG-NRE were normal in most MSD cases. Conversely, (E) HS levels and (F) the HS-derived MPS II GAG-NRE were elevated in urine from severe MS patients and differed between attenuated and severe cases. (G) Progression of motor dysfunction differed between severe cases with elevated MPS II GAG-NRE levels (black) and attenuated cases with MPS II GAG-NRE levels in the normal range (red). Motor function was quantified using the GMFC-MLD score. Score corresponds to function as follows: 0 = walking without support; 1 = walking without support reduced quality of performance; 2 = walking with support; 3 = sitting without support AND crawling/rolling; 4 = either sitting without support OR crawling/rolling; 5 = no sitting without support AND no crawling/rolling; 6 = no head or trunk control.

In agreement with both the internal disaccharide and endogenous biomarker analysis, markers of HS degradation were more affected by loss of FGE activity than CS/DS-derived markers. Both HS (Fig 5E) and the HS-derived MPSII GAG-NRE (I2S0) (Fig 5F) were substantially elevated in severe patients as compared to both attenuated patients and healthy controls. MSD patients with MPSII GAG-NRE level above 0.1 mg/mmol creatinine had early-onset rapid neurologic regression as measured by the GMFC-MLD scale²¹ (Fig 5G). A score of 0 on the GMFC-MLD represents normal neurologic function, while a score of 6 represents loss of all motor function. All severe MSD patients had elevated MPSII GAG-NRE levels and demonstrated motor regression prior to age 4 (Fig 5G, black circles). Conversely, all attenuated patients had normal MPS II GAG-NRE levels on Sensi-Pro^® analysis and exhibited motor regression in later childhood (Fig 5G, red circles).

Integration of all biomarker data highlights that HS-related sulfatases are especially sensitive to loss of FGE activity

To evaluate if biomarkers distinguish clinical groups, the minimal differentiation factor for each enzyme activity level and GAG species was calculated. This value quantifies the fold difference between controls and disease states for biomarker (sulfatases: lowest control value/highest MSD case value; GAG levels: lowest case value/highest control value). In terms of diagnostic biomarkers, when comparing severe MSD cases and controls, markers associated with MLD, MPS IIIA and MPS II provided the best between-group discrimination (Table 2A). Interestingly, biomarkers associated MPS IVA and MPS VI were the least affected in most cases. In terms of severity biomarkers, when comparing severe and attenuated MSD cases, enzyme activity levels partially discriminated groups (Fig 2). However, HS-derived GAG subspecies, especially those that accumulate in MPS II, more clearly differentiated severity cohorts (Table 2B). Collectively, combined analysis of sulfatase activities and GAG-NREs can predict an individual’s disease severity.

Table 2: Biomarkers of MLD and the HS-related disorders MPS IIIA and MPS II are the most informative MSD biomarkers.

The minimal differentiation factor for each enzyme activity level and GAG species were calculated. For enzyme activity levels, the minimal differentiation factor is the lowest control value divided by the highest MSD case value. For GAG levels, the minimal differentiation factor is the lowest case value was divided by the highest control value. (A) When comparing severe MSD cases and controls, ARSA, SGSH, and IDS biomarkers are the most informative. (B) When comparing attenuated and severe MSD cases, IDS-derived GAGs demonstrated the greatest between-group differences. Markers with the greatest between-group differences are shown in red and represent a minimal fold difference of 6 or greater. Markers in dark green do not differ between groups.

A	Diagnostic biomarkers: Minimal differentiation factor between severe MSD cases and controls
		Enzyme Activity	Internal Disaccharide GAG*	Endogenous Biomarker GAG- NRE	Sensi-Pro GAG
	ARSA (MLD)	82.91	N/A	N/A	N/A
	SGSH (MPS IIIA)	29.74	5.55	10.24	N/A
	IDS (MPS II)	27.39	7.17	7.89	3.00
	GNS (MPS IIID)	4.47	4.37	0.66	N/A
	ARSB (MPS VI)	3.25	0.20	3.37	0.58
	GALNS (MPS IVA)	2.36	0.12	1.41	0.38
B	Severity biomarkers: Minimal differentiation factor between severe and attenuated MSD cases
		Enzyme Activity	Internal Disaccharide GAG*^	Endogenous Biomarker GAG- NRE^	Sensi-Pro GAG
	ARSA (MLD)	0.79	N/A	N/A	N/A
	SGSH (MPS IIIA)	0.09	2.21	1.90	N/A
	IDS (MPS II)	0.52	2.91	6.45	3.00
	GNS (MPS IIID)	0.17	3.58	0.39	N/A
	ARSB (MPS VI)	0.89	2.58	1.17	1.32
	GALNS (MPS IVA)	0.13	2.16	1.23	1.15

^*Internal disaccharides (D0S0, D0A6, D0A0) for MPS II, IIIA, IIID are markers shared by all three deficiencies

^{^}Only one attenuated MSD sample was availble for internal disaccharide and endogenous biomarker assays

N/A= not available or not applicable

DISCUSSION

Multiple sulfatase deficiency (MSD) is an ultra-rare fatal degenerative disorder with emerging therapeutic options ^5,26,27. Until now, clinically meaningfully biomarkers for MSD have been lacking. In MSD, the activity of individual sulfatases only partially correlates with clinical severity^1,2. This is hypothesized to be due to fluctuation of enzyme activity and challenges with ex vivo sulfatase enzymatic assays and difficulty directly measuring FGE activity in vitro. Pathogenic mutations in MSD are associated with decreased FGE stability and decreased half-life^1,2,7.

With the goal of identifying potential diagnostic and prognostic biomarkers, in this study we characterized the patterns of the byproducts of GAG degrading sulfatase activities in MSD, the glycosaminoglycan non-reducing end (GAG-NRE) subspecies. As anticipated, we found low sulfatase activity across the panel of tested enzymes and an elevation of GAG-NREs. Of note, there was a differential impact of the MSD-related FGE mutations on sulfatase activities. Heparan-sulfate (HS) related enzymes and degradation products were the most affected by FGE mutations and best differentiated cases from controls and severe and attenuated cohorts. These patterns were confirmed across testing platforms. The detectable CS- and DS-derived GAG-NRE species in MSD patients were similar to control samples, underscoring their limitations as diagnostic biomarkers and the importance of HS-related GAG-NREs. While total keratan sulfate levels were not measured in this study, we only saw a modest increase in the MPSIVA-associated KS-derived endogenous biomarkers in MSD samples. These biochemical findings mirror the clinical trends observed in MSD patients where neurologic symptoms (such as those seen in HS-related disorders) are much more prevalent than orthopedic and other somatic symptoms (such as those seen in MPSIVA) ^1,2.

In addition to supporting the potential role of using patterns of GAG-NRE accumulation in disease stratification, these findings suggest broader pathophysiologic differences in how sulfatases are activated in the cell. Further studies are needed to understand the unique sensitivity of HS-related enzymes to FGE activation.

In this work, we also expand upon our understanding of genotype-phenotype correlations in MSD through functional analysis of a novel SUMF1 variant c.655C>A (p.T219P). This variant demonstrated a very severe biochemical phenotype in vitro which correlated with a severe biomarker phenotype as measured by GAG-NRE and enzyme analysis of patient samples. Although this is a singular example, this suggests that biomarker data can help guide our interpretation of novel SUMF1 variants.

One limitation of this study is sample size. We were especially limited in our samples from attenuated cases. This is a challenge with ultrarare diseases, like MSD. To facilitate participation, individuals were able to remotely participate in sample collection, although more samples are needed to definitively evaluate if endogenous GAG-NRE biomarkers can be used to stratify patients by disease severity. Additionally, we used multiple complementary approaches to assess the impact of MSD mutations on GAG-NRE species. Despite the limited sample numbers, we found consistent results across platforms. It is important to note that GAG-NRE levels may fluctuate across the lifespan, so larger, longitudinal studies are needed. Finally, it is likely that sulfatide accumulation, which occurs because of loss of ARSA activity is also contributing to disease pathology. Future studies evaluating sulfatide levels in MSD patients are needed.

Our preliminary work suggests that HS-derived GAG-NRE species can serve as biomarkers of disease severity in MSD. Future studies will explore the longitudinal performance of these novel biomarkers and evaluate if their utility in measuring efficacy of novel therapeutics. As upcoming MSD gene therapy and small molecule clinical trials are initiated ^6,28, these biomarkers will be essential for defining appropriate study cohorts.

Synopsis:

Sulfatase activities and GAG nonreducing end accumulation patterns are disease severity biomarkers of Multiple Sulfatase Deficiency.

Data Availability:

The data that support the findings of this study are available from the corresponding author upon request.