Glycosaminoglycans in the blood of hereditary multiple exostoses patients: Half reduction of heparan sulfate to chondroitin sulfate ratio and the possible diagnostic application

Abstract

Hereditary multiple exostoses (HME) is an autosomal dominant skeletal disorder with wide variation in clinical phenotype and is caused by heterogeneous germline mutations in two of the Ext genes, EXT-1 and EXT-2, which encode ubiquitously expressed glycosyltransferases involved in the polymerization of heparan sulfate (HS) chains. To examine whether the Ext mutation could affect HS structures and amounts in HME patients being heterozygous for the Ext genes, we collected blood from patients and healthy individuals, separated it into plasma and cellular fractions and then isolated glycosaminoglycans (GAGs) from those fractions. A newly established method consisting of a combination of selective ethanol precipitation of GAGs, digestion of GAGs recovered on the filter-cup by direct addition of heparitinase or chondroitinase reaction solution and subsequent high-performance liquid chromatography of the unsaturated disaccharide products enabled the analysis using the least amount of blood (200 µL). We found that HS structures of HME patients were almost similar to those of controls in both plasma and cellular fractions. However, interestingly, although both the amounts of HS and chondroitin sulfate (CS) varied depending on the different individuals, the amounts of HS in both the plasma and cellular fractions of HME patient samples were decreased and the ratios of HS to CS (HS/CS) of HME patient samples were almost half those of healthy individuals. The results suggest that HME patients' blood exhibited reduced HS amounts and HS/CS ratios, which could be used as a diagnostic biomarker for HME.

chondroitinase, chondroitin sulfate, heparan sulfate, heparinase, hereditary multiple exostoses

Introduction

Hereditary multiple exostoses (HME) [OMIM 133700] is a skeletal bone disorder that results in the formation of benign cartilage-capped tumors or exostoses, primarily on the long bones of affected individuals (Solomon 1964). These exostoses develop shortly after birth and increase in number and size throughout childhood until closure of the growth plate at the end of puberty. Often, the presence of these benign bone tumors is accompanied by skeletal deformities such as shortening and bowing of the forearm and short stature (Shapiro et al. 1979). The main complications in HME are pain, caused by the pressure on neighboring tissues, disturbance of the blood circulation by compression of blood vessels and, in rare cases, spinal and/or cervical cord compression and myelopathy (Chiurco 1970). The disease is estimated to occur with an incidence of 1/50,000 (Schmale et al. 1994). HME appears to affect both sexes similarly (Schmale et al. 1994; Wicklund et al. 1995), but the penetrance in women is incomplete and the phenotype is often milder (Legeai-Mallet et al. 1997).

HME is a genetically heterogeneous disease linked with two genes, Ext1 and Ext2, and most patients have Ext1 or Ext2 haploinsufficiency (Zak et al. 2002; Jennes et al. 2009). The protein products of Ext-1 and Ext-2, exostosin-1 and exostosin-2, are type II transmembrane glycoproteins with glycosyltransferase activities, form a hetero-oligomeric complex, are located in the Golgi apparatus and are involved in the synthesis of heparan sulfate (HS) polysaccharides (Lind et al. 1998; McCormick et al. 1998, 2000; Senay et al. 2000; Wei et al. 2000). The majority of cases of HME are caused by frameshift or missense mutations in Ext-1 or Ext-2 that create truncated forms of the proteins these genes encode (Ahn et al. 1995; Stickens et al. 1996; Wuyts and Hul 2000; Zak et al. 2002). Data from chondrocyte-specific conditional homozygous loss of Ext-1 in mutant mice showed a complete local loss of HS, leading to the formation of multiple exostoses on long bones (Matsumoto et al. 2010). Recently, Zak et al. (2011) showed that exostoses also developed in mutant mice on long bones depending on gene dosage effects. However, Ext-1^+/− or Ext-2^+/− mutant mice, which are thought to be similar in genetic situations of human HME patients showed a decreased HS content in cells and/or tissues (Stickens et al. 2005; Zak et al. 2011). An immunostaining study from exostosis growth plates and electrophoretic/immunoblot analysis of proteoglycans (PGs) from HME patients gave some idea about the decreased distribution of HS in HME patients (Hecht et al. 2002). However, details of the contents and fine structure of HS in other tissues of HME patients such as blood remain to be studied.

PGs are a family of glycoconjugates characterized by the presence of one or two different types of glycosaminoglycan (GAG) chains covalently bound to a protein core. Human tissues, including blood, contain a variety of PGs (Bernfield et al. 1999). Two major types of GAGs, HS/heparin and chondroitin sulfates (CSs), are produced by most animal cells (Esko and Zhang 1996). Because of their differential sulfation, these linear polysaccharides are believed to be the most information-dense biopolymers found in nature. GAG interacts with hundreds of plasma proteins (Saito and Munakata 2007), including growth factors, cytokines, chemokines, proteases, protease inhibitors, coagulant and anticoagulant proteins, complement proteins, lipoproteins and lipolytic enzymes (Varki et al. 2009). More than 95% of plasma GAGs form complexes with plasma proteins (Calatroni et al. 1992). The ability to identify changes in blood GAG structures will advance the understanding and diagnosis of human diseases.

Human serum or plasma and cellular fractions of blood serve as a typical clinical specimen, since these are more accessible and convenient for long-term monitoring. It has been demonstrated that the low–molecular-weight fraction of human serum or plasma generated through enzymatic cleavage provides a rich source for potential biomarkers of diseases (Plaas et al. 1998; Chen et al. 2007). However, the minute concentration of GAGs in plasma and cellular fractions of human blood together with the complexity of the plasma and cellular fractions limit the information on the HS and CS component in human plasma and cellular fractions. The characterization of HS and CS structures seems to be important for elucidating the functions corresponding to those GAGs. Many efforts have been made to meet these requirements, involving a number of techniques such as multiple proteolytic enzyme digestions, trichloroacetic acid or heating precipitation, dialysis, lipid extraction and gel electrophoretic separation. (Duffy and Fried 2003; Higgins, 2008; Gill et al. 2009; Mulloy et al. 2009; Wang et al. 2010). However, to date, none of the above methods has yielded efficient extraction of HS and CS from blood. For this purpose, we have developed a purification method involving ethanol precipitation, desalting GAGs in the elution solution from the DEAE–Sephacel column using the filter-cup, subsequent digestion of GAGs on the filter membranes by direct addition of heparitinase or chondroitinase reaction solution and fluorometric postcolumn reverse phase high-performance liquid chromatography (HPLC). The method would be available only for a small amount of blood samples (the least 200 µL).

In this report, we first describe the establishment of the methods for the isolation of GAGs from plasma and cellular fractions of blood and then for the quantitative analysis of disaccharide compositions of HS and CS. By this method, we next examined the structures of HS and CS, as well as their amounts in both plasma and cellular fractions, of control and HME patients. We found that in blood samples of HME patients, the HS amount was reduced and the ratios of HS to CS (HS/CS) were always half those of controls and, thus, propose that the ratio of HS/CS could be used as a potential biomarker for the diagnosis of HME.

Results

Isolation and purification of GAGs from plasma and cellular fractions

The goal of this study was to examine how the GAG structures and amounts in blood are different between controls and HME patients and whether or not the possible differences could be used as a diagnostic biomarker. We first separated blood samples into plasma and cellular fractions. In preliminary experiments, we analyzed commercially available heparin, HS and CS by reversed-phase ion-pair chromatography. All disaccharides were detected at the expected elution positions (Figure 1).

Fig. 1.

Separation of unsaturated disaccharides by reverse-phase high-pressure liquid chromatography. Heparin (A) or HS (B) was digested with a heparitinase-I, -II and heparinase mixture, and CS (C) was digested with chondroitinase ABC. These digests were then subjected to reversed-phase ion-pair chromatography with postcolumn fluorescence labeling as described under the “Materials and methods” section. The chromatograms represent the amount of unsaturated disaccharides from each GAG. (A and B) The numbered disaccharide peaks correspond to 1, -HexA-GlcNAc-(ΔDi-0S); 2, -HexA-GlcNS-(ΔDi-NS); 3, -HexAGlcNAc(6S)-(ΔDi-6S); 4, -HexA-GlcNS(6S)-(ΔDi-(N,6)diS); 5, -HexA(2S)-GlcNS-(ΔDi-(N,2)diS); and 6, -HexA(2S)-GlcNS(6S)-(ΔDi-(N,6,2)triS). (C) The disaccharide peaks are 1′, -HexA-GalNAc-(ΔDi-0S); 2′, -HexA-GalNAc(4S)-(ΔDi-4S); 3′, -HexAGalNAc(6S)-(ΔDi-6S); 4′, -HexA-GalNAc(4,6)diS-(ΔDi-diS_E) and 5′, -HexA(2S)-GalNAc(6S)-(ΔDi-diS_D).

We developed a systematic approach for the isolation of GAGs that allowed us to identify and quantify GAG molecules in minute volumes of human blood (Figure 2). For the quantitative measurement of small amounts of GAG molecules in those fractions, we found that the following isolation steps were essential: Several rounds of ethanol precipitation are necessary to reduce the amount of contaminants such as lipids, lipid-derivatives and low-molecular-weight materials. DEAE–Sephacel column chromatography is also necessary to preferentially condense positively charged molecules. Then, GAGs were subjected to desalting the elution solution containing high concentrations of salt by filtering through the membrane in the filter-cup (5-kDa molecular weight cut-off filtering unit; Millipore Corp., Bedford, MA), and the retained GAGs on the filter membrane were digested by direct addition of the heparitinase or chondroitinase reaction solution and subsequent incubation at 37°C by which the recovery of the disaccharide product was improved. The disaccharide products thus obtained were analyzed by fluorometric postcolumn HPLC. We validated the processes by adding known amounts of CS or HS as internal standards to the samples (Tables Ia and Ib). The recovery of the purification and subsequent digestion processes gave an 80% yield. As far as we know, this method is feasible with the least amount of blood among the reported methods (up to 10 mL (Komosinska-Vassev et al. 2005; Lu et al. 2010)).

Table Ia.

Validation of the analysis of HS in blood

ΔDi-HS	Known percent compositiona	Analyzed percent composition
ΔDi-0S	46.1	46.4
ΔDi-NS	21.1	19.8
ΔDi-6S	8.1	8.3
ΔDi-(N,6)diS	6.6	4.7
ΔDi-(N,2)diS	11.5	12.4
ΔDi-(N,6,2)triS	6.5	8.4

The samples were prepared by adding known composition and amounts of HS to plasma fractions. GAGs (exogenous and endogenous) were isolated as described under the “Materials and methods” section. Disaccharide amounts derived from the exogenously added HSs were calculated by subtracting the ones from the endogenous HSs from the ones obtained in this experiment. The values are given as percentages of the total disaccharides and the means of two independent experiments. ΔDi-0S, ΔDi-NS, ΔDi-6S, ΔDi-(N,6)diS, ΔDi-(N,2)diS and ΔDi-(N,6,2)triS are the disaccharide units as described under “Abbreviations”.

^aFrom the company's report.

Table Ib.

Validation of the analysis of CS in blood

ΔDi-CS	Known percent compositiona	Analyzed percent composition
ΔDi-0S	0.9	1.0
ΔDi-4S	55.6	49.2
ΔDi-6S	32.9	40.1
ΔDi-diSE	1.3	1.5
ΔDi-diSD	8.2	8.3

The samples were prepared by adding known composition and amounts of CS to cellular fractions. GAGs (exogenous and endogenous) were isolated as described under the “Materials and methods” section. Disaccharide amounts derived from the exogenously added CSs were calculated by subtracting the ones from the endogenous CSs from the ones obtained in this experiment. The values are given as percentages of the total disaccharides and the means of two independent experiments. ΔDi-0S, ΔDi-4S, ΔDi-6S, ΔDi-diSE and ΔDi-diSD are the disaccharide units as described under “Abbreviations”.

^aFrom the company's report.

Fig. 2.

A schematic diagram representing GAG isolation from human blood. Blood samples were separated into plasma and cellular fractions and both were individually digested with DNase I and Proteinase-K after alkali treatment. The PGs in the extracts were purified from contaminating macromolecules by extensive ethanol precipitation, repeated washing through ion-exchange chromatography and GAGlyase digestion over filter membrane by direct addition of heparitinase and chondroitinase reaction solution. The unsaturated disaccharide products were analyzed by fluorometric postcolumn HPLC.

HS structure does not differ in blood of HME patients

Core protein-free GAGs were isolated from both plasma and cellular fractions of blood from controls and patients as described above. Portions of the GAGs were digested with a mixture of heparitinase I, heparitinase II and heparinase, and the disaccharide products were analyzed by reversed-phase ion-pair chromatography and subsequently by using a postcolumn fluorometric detection system as described in the “Materials and methods” section. Typical examples of the resulting chromatograms are shown in Figure 3. HS disaccharide peaks were identified by comparison of the elution profiles of the standard HS-derived unsaturated disaccharides.

Fig. 3.

Representative chromatograms show HS structures from cellular fractions of control (A) and HME patient (B) samples. The numbered disaccharide peaks correspond to 1, -HexA-GlcNAc-(ΔDi-0S); 2, -HexA-GlcNS-(ΔDi-NS); 3,-HexAGlcNAc(6S)-(ΔDi-6S); 4,-HexA-GlcNS(6S)-(ΔDi-(N,6)diS); 5,-HexA(2S)-GlcNS-(ΔDi-(N,2)diS) and 6, -HexA(2S)-GlcNS(6S)-(ΔDi-(N,6,2)triS).

The percent compositions of disaccharides based on their relative molar amounts obtained by lyase digestion from plasma fractions of control and patient blood were almost the same (Table II). A considerable amount of Δhexuronic acid-N-acetylgalactosamine (ΔDi-0S) (mean: 70 and 71 nmol ΔDi-0S units per 100 nmol HS disaccharide units from control and patient samples, respectively) was detected in the plasma fractions of both control and patient samples. ΔHexuronic acid-N-sulfated glucosamine (ΔDi-NS) was another major disaccharide unit in both plasma fractions, and their average amount was ∼15 and 14 nmol per 100 nmol HS disaccharide units in control and patient samples, respectively. Small amounts of unsaturated di-sulfated disaccharide units (Δhexuronic acid-N and 6-O-sulfated glucosamine (ΔDi-(N,6)diS) and ΔDi-(N,2)diS) were detected in both samples, and their relative molar amounts were not different between controls and HME patients; ΔDi-(N,6)diS, 3 and 2 nmol per 100 nmol HS disaccharide units in controls and HME patients, respectively, and ΔDi-(N,2)diS, 3 nmol per 100 nmol HS disaccharide units in both controls and HME patients. The presence of 2-O-sulfated Δhexuronic acid-N and 6-O-sulfated glucosamine (ΔDi-(N,6,2)triS), 4 and 6 nmol per 100 nmol HS disaccharide units in samples of controls and HME patients, respectively, may have derived from heparin in mast cells in both samples.

Table II.

HS disaccharide analysis of the plasma fractions of samples from controls and HME patients

Plasma fractions	Percentage of unsaturated disaccharides						Degree of sulfationa
	ΔDi-0S	ΔDi-NS	ΔDi-6S	ΔDi-(N,6)diS	ΔDi-(N,2)diS	ΔDi-(N,6,2)triS
Control
C-1	72	15	4.2	2.5	3.4	2.7	0.40
C-2	72	15	3.7	1.0	2.0	6.5	0.44
C-3	69	15	4.6	2.5	3.6	5.9	0.49
C-4	73	12	5.0	3.0	2.5	3.8	0.40
C-5	67	16	5.1	5.2	2.1	4.3	0.49
C-6	69	16	3.4	2.7	4.4	4.6	0.48
C-7	69	16	5.4	3.1	3.3	3.5	0.45
Mean	70 ± 2	15 ± 1	4.5 ± 0.7	2.8 ± 1.2	3.1 ± 1	4.5 ± 1.3	0.45 ± 0.04
Patient
P-1	71	15	3.8	2.3	3.5	4.6	0.44
P-2	71	14	4.1	1.6	3.0	6.8	0.47
P-3	69	15	5.5	2.2	3.4	5.1	0.47
P-4	69	15	4.9	3.5	2.6	4.7	0.47
P-5	71	13	4.0	1.3	1.7	9.4	0.51
P-6	72	12	4.4	1.2	2.6	7.6	0.46
P-7	71	14	5.0	1.3	2.4	6.6	0.46
P-8	72	12	4.3	2.1	2.1	7.6	0.47
P-9	74	12	4.1	1.7	3.5	4.1	0.39
P-10	68	18	4.4	2.6	4.9	3.1	0.46
P-11	72	14	3.5	2.0	3.7	4.8	0.43
P-12	68	17	4.7	3.7	1.7	5.0	0.44
P-13	69	17	4.5	1.8	1.7	5.1	0.44
P-14	74	11	3.7	2.7	1.7	7.3	0.45
P-15	72	17	4.2	0.8	1.4	4.9	0.40
P-16	74	12	3.5	2.8	1.7	6.0	0.43
Mean	71 ± 2	14 ± 2	4.3 ± 0.5	2.1 ± 0.8	2.6 ± 1	5.8 ± 1.6	0.45 ± 0.03

^aDegree of sulfation is the average number of sulfate groups/disaccharide calculated from the disaccharide analysis. The values are given as percentages of total disaccharides.

The disaccharide analysis of HS in the cellular fractions of control and HME patient samples showed almost identical disaccharide compositions (Table III). ΔDi-0S was the major HS disaccharide unit (63 and 62 nmol per 100 nmol disaccharide units from control and HME patient samples, respectively), while ΔDi-NS was another major disaccharide unit (average 13 and 14 nmol per 100 nmol of the HS disaccharide unit in control and HME patient samples, respectively). The amounts of these two major disaccharides, however, varied among individuals as well as between control and HME patient samples, with a mean deviation of 2 and 8% of ΔDi-0S and ΔDi-NS units, respectively. The cellular fractions have more highly sulfated disaccharide units than the plasma fractions (almost 25% more sulfation in both the control and HME patient samples), which may suggest that the HS in plasma was not only derived from the blood cells.

Table III.

HS disaccharide analysis of the cellular fractions of samples from controls and HME patients

Cellular fractions	Percentage of unsaturated disaccharides						Degree of sulfationa
	ΔDi-0S	ΔDi-NS	ΔDi-6S	ΔDi-(N,6)diS	ΔDi-(N,2)diS	ΔDi-(N,6,2)triS
Control
C-1	63	12	4.7	5.6	7.1	7.8	0.66
C-2	63	13	10.3	3.5	4.2	6.2	0.57
C-3	66	11	5.4	3.1	5.9	9.0	0.61
C-4	63	15	6.5	5.1	4.6	6.1	0.59
C-5	63	12	4.6	6.1	7.7	6.4	0.63
C-6	65	13	5.2	5.7	4.9	6.6	0.59
C-7	63	15	5.3	5.2	5.9	6.0	0.60
Mean	63 ± 1.2	13 ± 2	6.0 ± 2	4.9 ± 1.1	5.8 ± 1.3	6.9 ± 1.1	0.61 ± 0.03
Patient
P-1	56	16	7.5	5.6	8.2	7.0	0.72
P-2	62	14	7.4	5.5	6.3	4.9	0.60
P-3	57	15	8.6	5.9	8.2	5.5	0.68
P-4	62	14	6.0	5.5	6.5	5.7	0.61
P-5	59	15	8.4	4.2	5.1	8.1	0.66
P-6	57	14	9.6	5.4	4.8	9.1	0.71
P-7	64	16	5.1	4.3	4.5	6.2	0.57
P-8	64	14	7.6	3.7	4.4	6.7	0.58
P-9	64	14	7.5	5.2	6.2	3.3	0.54
P-10	62	18	2.9	6.2	7.1	4.2	0.60
P-11	58	15	15.0	3.8	1.6	6.4	0.60
P-12	68	14	7.0	6.0	4.3	1.4	0.45
P-13	62	14	8.5	6.7	6.5	2.0	0.55
P-14	72	9	3.7	7.2	4.8	3.7	0.47
P-15	64	13	8.8	5.2	5.6	2.9	0.52
P-16	59	13	7.3	5.3	9.3	5.5	0.66
Mean	62 ± 4.2	14 ± 2	7.6 ± 2	5.4 ± 0.9	5.8 ± 1.8	5.2 ± 2.1	0.60 ± 0.07

^aDegree of sulfation is the average number of sulfate groups/disaccharide calculated from the disaccharide analysis. The values are given as percentages of total disaccharides.

Thus, the HS structures remained unaffected in both the plasma and cellular fractions of HME patient blood, although HME patients had mutations in the HS chain synthetic enzymes.

CS structure does not differ in HME patient blood

Portions of both GAGs isolated from human plasma and cellular fractions were digested with chondroitinase ABC and the disaccharide products were analyzed by the method described in the “Materials and methods” section. Typical examples of the elution profiles for HPLC are shown in Figure 4.

Fig. 4.

Representative chromatograms show CS structures from the plasma fractions of control (A), HME patient (B) and disaccharides from cellular fractions (C) of control sample. The numbered disaccharide peaks correspond to 1′, -HexA-GalNAc-(ΔDi-0S); 2′, -HexA-GalNAc(4S)- (ΔDi-4S); 3′, -HexAGalNAc(6S)-(ΔDi-6S); 4′, -HexA-GalNAc(4,6)diS- (ΔDi-diS_E) and 5′, -HexA(2S)-GalNAc(6S)-(ΔDi-diS_D).

The relative molar composition analyses of CS in the plasma fractions showed that ΔDi-0S and Δhexuronic acid-4-sulfated N-acetylgalactosamine (ΔDi-4S) were the major types in both control and patient samples (Table IV). Similar results were reported elsewhere on normal plasma fractions (Huang et al. 1995). Interestingly, our present isolation and analytical method identified minor fractions of Δhexuronic acid-6-sulfated N-acetylgalactosamine (ΔDi-6S), ΔDi-diS_E and ΔDi-diS_D, which were undetected in previous reports (Huang et al. 1995). In particular, we detected 1.0 nmol of ΔDi-6S, 0.5 and 0.4 nmol of ΔDi-diS_E and 0.5 and 0.3 nmol of ΔDi-diS_D per 100 nmol CS disaccharide units in both control and patient plasma samples, respectively, indicating the high sensitivity of our detection method.

Table IV.

CS disaccharide analysis of the plasma fractions of samples from controls and HME patients

Plasma fractions	Percentage of unsaturated disaccharides					Degree of sulfationa
	ΔDi-0S	ΔDi-4S	ΔDi-6S	ΔDi-diSE	ΔDi-diSD
Control
C-1	67	31	0.8	0.6	0.2	0.34
C-2	67	32	0.5	0.1	0.1	0.33
C-3	63	35	1.1	0.6	0.3	0.38
C-4	61	36	1.4	0.7	0.7	0.40
C-5	62	35	1.2	0.6	0.9	0.39
C-6	67	31	1.0	0.8	0.9	0.35
C-7	66	32	1.2	0.6	0.8	0.36
Mean	65 ± 2.5	33 ± 2.2	1 ± 0.3	0.5 ± 0.2	0.5 ± 0.4	0.36 ± 0.03
Patient
P-1	50	49	0.8	0.2	0.1	0.50
P-2	58	41	0.7	0.2	0.1	0.43
P-3	59	40	0.9	0.2	0.0	0.41
P-4	56	43	1.0	0.2	0.1	0.45
P-5	61	38	0.8	0.1	0.0	0.39
P-6	49	50	1.4	0.1	0.0	0.52
P-7	55	44	1.4	0.2	0.0	0.46
P-8	68	31	0.8	0.2	0.1	0.33
P-9	65	33	1.2	0.6	0.6	0.37
P-10	66	32	1.0	0.7	0.9	0.36
P-11	66	31	0.9	0.7	1.0	0.37
P-12	58	40	0.8	0.7	0.4	0.45
P-13	44	54	0.8	0.3	0.8	0.57
P-14	64	34	0.8	0.6	0.4	0.37
P-15	69	28	1.2	0.6	0.4	0.32
P-16	68	30	0.9	0.7	0.4	0.33
Mean	60 ± 7.6	39 ± 7.8	1 ± 0.2	0.4 ± 0.2	0.3 ± 0.3	0.41 ± 0.07

^aDegree of sulfation is the average number of sulfate groups/disaccharide calculated from the disaccharide analysis. The values are given as percentages of total disaccharides.

The relative molar unit composition analyses of CS in the cellular fractions (Table V) showed that ΔDi-4S was the major disaccharide unit of the cellular fractions of both control and HME patient samples (92 and 90 nmol per 100 nmol CS disaccharide units in the control and HME patient samples, respectively). We also found that ΔDi-0S was one of the minor disaccharide units (5 and 8 nmol per 100 nmol CS units in the control and patient samples, respectively). The present results are well consistent with the previously reported ones (Hata et al. 1978; Vannucchi et al. 1980), which showed that human leukocytes contained CS consisting of both 4-sulfated and nonsulfated disaccharide units. However, one report (Murata 1974) described that human leukocytes contained CS consisting of only 4-sulfated disaccharide units. Furthermore, small and almost the same amounts of ΔDi-6S, ΔDi-diS_E and ΔDi-diS_D were detected in the digested products from all cellular fractions of the control and patient samples. Apparently, the degree of sulfation of CS in the cellular fractions was higher than that in the plasma fractions of both the control and HME patient samples (the cellular fractions of the control and HME patient samples had CS with a 2.7- and 2.3-fold higher degree of sulfation, respectively, than plasma fractions).

Table V.

CS disaccharide analysis of the cellular fractions of samples from controls and HME patients

Cellular fractions	Percentage of unsaturated disaccharides					Degree of sulfationa
	ΔDi-0S	ΔDi-4S	ΔDi-6S	ΔDi-diSE	ΔDi-diSD
Control
C-1	10	88	1.3	0.7	0.1	0.91
C-2	7	92	0.7	0.6	0.1	0.94
C-3	6	92	0.8	1.0	0.1	0.95
C-4	4	91	1.3	1.6	1.7	0.99
C-5	2	94	1.0	1.2	1.3	1.00
C-6	4	93	1.3	0.9	0.7	0.98
C-7	3	94	0.9	1.2	0.7	0.98
Mean	5 ± 2.5	92 ± 2	1 ± 0.3	1 ± 0.3	0.6 ± 0.6	0.97 ± 0.03
Patient
P-1	8	91	0.8	0.9	0.0	0.93
P-2	7	92	0.8	0.7	0.1	0.94
P-3	6	91	0.8	1.4	0.0	0.95
P-4	5	94	0.8	0.4	0.0	0.95
P-5	15	84	0.8	0.6	0.1	0.86
P-6	9	89	0.8	1.0	0.1	0.92
P-7	14	85	0.9	0.7	0.1	0.87
P-8	12	87	0.8	0.6	0.1	0.89
P-9	3	94	0.9	1.1	0.5	0.98
P-10	3	95	0.8	1.1	0.1	0.98
P-11	6	93	0.7	0.2	0.1	0.95
P-12	9	89	0.6	1.4	0.2	0.99
P-13	10	87	1.5	1.5	0.4	0.92
P-14	6	91	0.9	1.3	0.1	0.95
P-15	9	87	1.7	1.8	0.1	0.93
P-16	7	89	1.8	1.7	0.3	0.98
Mean	8 ± 3.3	90 ± 3.3	1 ± 0.3	1 ± 0.4	0.1 ± 0.1	0.94 ± 0.04

^aDegree of sulfation is the average number of sulfate groups/disaccharide calculated from the disaccharide analysis. The values are given as percentages of total disaccharides.

Plasma usually contains hyaluronan at the level of several 10 ng/mL and hyaluronan gave nonsulfated disaccharide products (ΔDi-HA) by chondroitinase ABC digestion, of which the elution position overlap with that of ΔDi-0S in HPLC chromatography. However, the content of bikunin-derived CS in plasma is usually ∼100-fold higher than that of hyaluronan, which has been shown to be mainly composed of nonsulfated disaccharide units (ΔDi-0S) (Zhuo et al. 2003, 2004; Ly et al. 2011) (Figure 4A and C). To confirm this, we actually measured hyaluronan content from the plasma fractions of control and HME patient samples and found that the contents of hyaluronan were very low compared with the CS contents. Furthermore, hyaluronan content was not significantly different between control and HME patient samples (See Supplementary Data). Therefore, the contribution of nonsulfated disaccharide products in the plasma fractions of both control and patient samples was due to the high amount of CS.

HS to CS ratio was half in blood of HME patients

We then quantified the HS and CS content in the blood of controls and HME patients. Although the HS and CS contents in blood varied among individuals and also between controls and HME patients, the ratios of the HS contents to the CS contents (HS/CS) were shown to be not so different among the individuals. Notably, the values of the HS/CS ratios from plasma and cellular fractions of HME patients were almost half those of the control samples (Table VI and Figure 5). That is, as for the plasma fractions, the HS/CS ratio in the samples from HME patients was 0.025 ± 0.004, while it was 0.055 ± 0.007 in the control samples. As for the cellular fractions, the HS/CS ratio in the samples from HME patients was 0.057 ± 0.01, while it was 0.154 ± 0.03 in the control samples.

Table VI.

HS to CS ratio from the blood of control and HME patients

	HME class	Plasma fractions		Cellular fractions		HS/CS
		HS (µg/mL)	CS (µg/mL)	HS (µg/mL)	CS (µg/mL)	Plasma fractions	Cellular fractions
Control
C-1	—	0.664	13.79	1.298	7.01	0.048	0.185
C-2	—	0.516	8.23	1.386	13.15	0.063	0.105
C-3	—	0.754	14.34	1.643	11.66	0.053	0.141
C-4	—	0.522	8.23	1.490	8.70	0.063	0.171
C-5	—	0.713	14.88	1.787	11.88	0.048	0.150
C-6	—	0.735	12.39	1.326	8.71	0.059	0.152
C-7	—	0.783	15.90	1.861	10.65	0.049	0.175
Mean	—	0.670 ± 0.11	12.54 ± 3.1	1.542 ± 0.2	10.25 ± 2.2	0.055 ± 0.007	0.154 ± 0.03
Patient
P-1	III	0.340	15.14	0.745	16.64	0.022	0.045
P-2	II	0.334	14.36	0.714	13.08	0.023	0.055
P-3	II	0.374	14.58	0.731	13.34	0.026	0.055
P-4	I	0.329	10.03	0.479	8.00	0.033	0.060
P-5	III	0.356	14.28	0.759	11.58	0.025	0.066
P-6	III	0.366	13.48	0.740	11.76	0.027	0.063
P-7	III	0.388	20.32	0.596	10.08	0.019	0.059
P-8	I	0.351	14.62	0.593	8.71	0.024	0.068
P-9	III	0.201	6.46	0.440	10.71	0.031	0.041
P-10	I	0.362	13.03	0.818	11.48	0.028	0.071
P-11	III	0.327	17.99	0.492	10.25	0.018	0.048
P-12	II	0.384	20.57	0.524	10.63	0.019	0.049
P-13	III	0.663	25.01	0.547	7.95	0.027	0.069
P-14	I	0.292	13.04	0.708	14.64	0.022	0.048
P-15	III	0.338	12.57	0.399	7.24	0.027	0.055
P-16	III	0.283	10.52	0.455	7.56	0.027	0.060
Mean	—	0.355 ± 0.09	14.75 ± 4.5	0.609 ± 0.1	10.85 ± 2.7	0.025 ± 0.004	0.057 ± 0.01

Fig. 5.

Analysis of the HS/CS ratio from healthy control individuals and HME patients. GAGs were isolated from blood as described under the “Materials and methods” section. The HS to the CS ratio in samples of HME patients is almost half that of the controls. Bars represent mean values. Statistical tests were performed using Student's t-test.

Discussion

The pathogenesis of HME was investigated by using a HME mouse model, which showed that multiple exostoses were observed along long bones accompanied by the conditional homozygous loss of Ext-1 leading to a nearly complete local loss of HS (Matsumoto et al. 2010). Furthermore, clonal homozygous inactivation of Ext-1 in chondrocytes developed frequent osteochondromas on the apendicular skeleton, mimicking the human HME phenotype (Jones et al. 2010). Most HME patients have been reported to be heterozygous in the Ext genes (Bernard et al. 2000; Legeai-Mallet et al. 2000; Hall et al. 2002). The genetic code provides for expression of genetic information in the form of proteins which are translated from mRNA that originates from one of the strands of the duplex DNA (Lewin 1994). So, we were wondering whether the behaviors of GAGs (HS and CS) in the blood of Ext heterozygous humans (HME patients) showed some abnormality or not. We showed for the first time that HS amount and HS to CS ratio (HS/CS) were significantly reduced in the blood of HME patients. These findings may correlate with previous reports on abnormal GAG biosynthesis with reduced amounts of HS levels in HME-derived chondrocytes (Hecht et al. 2002; Hecht et al. 2005). The reduced relative HS content in the blood of HME patients may be due to reduced EXT enzyme activity in HME patients. In support of this notion, an about more than 50% decline in HS formation and ∼75% decrease of GlcNAc and GlcA transferase activities were observed in EXT1 heterozygous ES cells (Lin et al. 2000). We also investigated whether there were any HS structural abnormalities between blood samples of controls and HME patients and whether there were any CS structural changes accompanied by the reduction of HS. We observed that with regard to HS in the blood, HME patients had the same structure as the controls had. This finding is correlated with the HS structure analysis from Ext1 gene trap mutant mice, which showed identical HS disaccharide structure with wild-type mice (Yamada et al. 2004). Both findings suggested that functional abnormalities in EXT do not interfere with the sulfation pattern on HS chains. The findings also prove the known HS biosynthetic pathway, where EXT is involved in chain elongation but refrained from diversified sulfation to decorate HS chains (Esko et al. 2009). We neither observed any significant differences in CS structures in plasma and cellular fractions between blood samples from controls and HME patients.

Recent studies have reported that EXT-2 can form complexes with EXT-1 as well as with N-acetylglucosamine N-deacetylase/N-sulfotransferase-1 (NDST1), an enzyme involved in sulfation of HS (Presto et al. 2008), which suggests that there may be pleiotropic effects of altering the level of Ext gene expression on both HS polymerization and sulfation patterns. In this study, we found almost little/no changes in the degrees of N- and O-sulfation of HS in plasma and cellular fractions of HME patients compared with those of controls (Tables II and III). Especially, N-sulfation appeared to be almost equal in all samples. It is, thus, likely that other isoforms of NDSTs may form complexes with EXTs to compensate for the N-sulfation, if N-sulfation by NDST1 was abrogated by Ext mutations. Alternatively, some reduction of the EXT protein activity may yet be sufficient to form complexes with NDST1.

A number of methods have been reported for the sensitive detection of GAGs such as glycan isotope labeling (Lawrence et al. 2008) or LC-mass spectrometry (LC-MS/MS) analysis (Wei et al. 2011). The prerequisite for this study was the development of a precise and high-yielding GAG isolation method from both human plasma and cellular fractions. One of the challenges was to isolate minute concentrations of GAGs from complex biological mixtures, which contain abundant proteins, lipids, glycolipids and lipoproteins, which might interfere with the analysis. The majority of proteins identified in human blood are comprised of albumin, immunoglobulins and transferrins, which account for ∼90% of the total protein content (Adkins et al. 2002). Other interfering molecules were lipids that are dispersed through their association with specific groups of proteins and lipoproteins that are associated with each other and with other serum molecules. Thus, to address issues of abundant interfering molecules, the plasma and cellular fraction samples were subjected to repeated ethanol precipitation. Another critical step was the GAGlyase digestion on the filter-cup by direct addition of the reaction solution containing heparitinase or chondroitinase. The present isolation method incorporated with those effective steps allowed us to measure differences from a minute GAG content from only 200 µL of sample solution.

Ext-1 and/or Ext-2 homozygous mutant mice died during development due to arrested growth and failed gastrulation. However, mice heterozygous for the Ext-1 or Ext-2 mutant allele appeared phenotypically normal. Some of these heterozygous mutant mice developed one or more exostoses in their ribs. However, exostoses were never identified on the long bones (tibia, femur, ulna and radius), where those often occur in HME patients (Stickens et al. 2005; Zak et al. 2011). Humans having HME are also genetically heterogeneous to Ext, which may be associated with the HS abnormalities. From human genetic mutation analyses, some of heterozygous mutants were also found to have reduced activities and/or contents of proteins as gene products. For example, patients with Albright's hereditary osteodystrophy diseases are heterozygous for the adenylyl cyclase stimulating protein encoding gene (GNAS1). The resultant effect of the mutation is the reduced activity of the protein (Pohlenz et al. 2003). Heterozygous mutants to the protein C gene (PROC) experience a mild form of protein C deficiency (Goldenberg and Manco-Johnson 2008) and protein C activity ranged from 18 to 60% in heterozygous mutants (Cafolla et al. 2012). Those examples provide a likely explanation why Ext heterozygous mutants have reduced HS contents, due to the reduction of the EXT enzyme protein activity and/or content.

We have analyzed the ratio of HS/CS from human blood and we successfully showed that the HS content is abnormal from the blood of HME patients (Table VI and Figure 5). Since we found that CS is the major GAG in human blood and its content had no significant changes between the samples of controls and HME patients, we expected that the ratio of HS content to the CS content could be affected in the blood of HME patients. We observed that the ratio was almost half in HME patients compared with normals. Therefore, the HS/CS ratio of blood GAGs has the potential to be developed as a diagnostic biomarker for the diagnosis of HME.

Materials and methods

Materials

Heparitinase-I (Flavobacterium heparinum, EC 4.2.2.8.), heparitinase-II (F. heparinum, no number assigned), heparinase (F. heparinum, EC 4.2.2.7), chondroitinase ABC (P. vulgaris, EC 4.2.2.4) and an unsaturated GAG disaccharide kit were purchased from Seikagaku Corp. (Tokyo, Japan). Sensyu Pak Docosil was purchased from Sensyu Scientific (Tokyo, Japan). DEAE–Sephacel was from Pharmacia (Uppsala, Sweden). Deoxyribonuclease I (bovine pancreas) was from Sigma (St. Louis, MO). Proteinase-K (Tritirachium album), potassium acetate, acetic acid, Triton X-100, ethylenediaminetetraacetic acid (EDTA) and other reagents were obtained from Nacalai Tesque (Kyoto, Japan).

Ethics statement

The Review Board Committees of the authors' institutions approved this study, and informed consent was obtained from all patients.

Patients and clinical study

In a prospective clinical study, the morphological and functional outcomes of 16 HME patients (13 families) were investigated. These patients were selected by clinical examination toward a clinical manifestation of exostoses. The mean age at the time of the study was 41.8 years (range, 29–71). Ten women and six men were included in this study. The clinical stages were evaluated by using Pedrini's classification (Pedrini et al. 2011). The following parameters were investigated for the evaluation of the disease: Patient's age, age at first clinical manifestation and details of any surgeries (indication, number and location of previous surgeries, and rate of recurrence). Clinical examination, including measurement of the range of motion in different joints (shoulder, elbow, wrist, hip, knee and upper ankle joint), was carried out and body mass index (BMI) and the lengths of the upper and lower arms and legs were measured. Previously taken radiographs were analyzed to detect further nonpalpable exostoses. Current radiographic images were taken if there were any complaints (pain, disturbances in sensitivity), clinical signs of new exostoses or growth of exostoses. Some of the parameters are given in Supplementary Data. All the patients were subjected to genetic mutation analysis and were found to have some exon mutations in the Ext-1 gene (unpublished observations).

Extraction of GAGs from blood samples

For the extraction of total GAGs, blood was collected in tubes containing EDTA, ∼200 µL samples were immediately centrifuged at 1500 × g for 15 min and were then separated into plasma and cellular fractions. Each fraction was mixed with three volumes of cold 95% (v/v) ethanol containing 1.3% (w/v) potassium acetate. The samples were kept at 0°C for 30 min and the resultant precipitates were collected by centrifugation at 15,000 × g and 4°C for 20 min and suspended in 500 µL H₂O. The ethanol precipitation described above was repeated three times. The precipitates were suspended in alkali with a final concentration of 0.4 M KOH and allowed to stand overnight at room temperature. The solutions were neutralized with acetic acid, subsequently treated with 10 µL of DNase I (25 µg DNA/mL of human blood (Adell and Ogbonna 1990)) at 37°C for 20 h and then with 120 µL of Proteinase-K (1 mg/mL) for 20 h at 45°C. Proteinase-K (0.5 mg/mL, 60 µL) was further added and the incubation was continued for another 20 h at 50°C. The reaction was stopped by boiling for 3 min and debris was removed by centrifugation at 12,000 × g and 4°C for 20 min. The supernatant was mixed with three volumes of 95% ethanol containing 1.3% (w/v) potassium acetate and the final precipitate was dissolved in 100 µL of 20 mM Tris–HCl (pH 7.5) containing 0.1% (v/v) Triton X-100. The solution was then loaded on a DEAE–Sephacel column (0.4 mL column volume), which had been equilibrated with the same buffer. The columns were washed with 20 column volumes of 20 mM Tris–HCl (pH 7.5), containing 0.2 M NaCl and 0.1% Triton X-100, and then eluted with four column volumes of 2 M NaCl in 20 mM Tris–HCl buffer (pH 7.5). A portion of the eluate was subjected to GAGlyase digestion and subsequent disaccharide composition analysis by HPLC as described below (see Figure 2).

GAGlyase digestion

Aliquots (40%) of the GAGs from the plasma and cellular fractions were applied to a filter-cup of Ultrafree-MC (5-kDa molecular weight cut-off filtering unit; Millipore Corp.) and the GAGs inside the filter membrane were centrifuged at 14,000 × g for 10 min. Before pouring GAGs into the filter-cup, the filter membranes were washed once with 70% ethanol and twice with deionized water to remove any impurities that might interfere with fluorescent labeling. With those washed filter cups, samples were desalted and the GAGs on the filter membranes were then digested with a mixture of 0.2 mU of heparitinase I, 0.1 mU of heparitinase II and 0.2 mU of heparinase in 50 μL of 50 mM Tris–HCl buffer (pH 7.2), 1 mM CaCl₂ and 5 μg bovine serum albumin at 37°C for 2 h. The other aliquots were treated in the same way. The GAGs on the membrane were digested with 5 mU of chondroitinase ABC in 50 μL of 50 mM Tris–HCl buffer (pH 7.2) and 1 mM CaCl₂ at 37°C for 2 h. After the digestion, the reaction mixtures containing unsaturated disaccharide products, enzymes, bovine serum albumin and (if any) undigested products on the membranes were centrifuged and the filtrates containing unsaturated disaccharide products were collected.

Analysis of unsaturated disaccharide products

The unsaturated disaccharide products in the filtrates were analyzed by reversed-phase ion-pair chromatography using the Sensyu Pak column Docosil with a fluorescence detector (Model RF-10AxL, Shimadzu Co. Kyoto, Japan) according to Toyoda's method (Toyoda et al. 2000) with slightly modified elution conditions. Briefly, after chromatographic separation, unsaturated disaccharide products were reacted with 2-cyanoacetoamide (Wako, Osaka, Japan) as a postcolumn reagent. Signals were identified and quantified by comparison of the elution positions and the peak heights with those of known amounts of standard disaccharides obtained in parallel runs.

Measurement of hyaluronan content

Hyaluronan content was measured as described previously (Wannarat et al. 2003). Briefly, immunoplates (Nunc, Denmark) were coated with hyaluronan-binding protein (HABP) (Seikagaku, Tokyo, Japan). Plasma samples were added on the coated plates and were incubated at 37°C for 1 h. Wells were washed and incubated with biotinylated HABP. After washing with phosphate buffer saline-0.05% Tween 20, horseradish peroxidase-streptavidin was added to each well, and color development was achieved by incubating with 3,3′,5,5′-tetramethylbenzidine solution (KPL, Gaithersburg, MD). The reaction was stopped using 1 M HCl and the absorbance was measured at 450/630 nm.

Statistical analysis

To evaluate the statistical significance of the experiments, we used Student's t-test. P < 0.05 were considered statistically significant.

Funding

This work was supported by Grants-in-aid for Scientific Research C (grant no. 20570113 to H.H.; 22590296 and 23570148 to K.K.) from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and Grant-in-Aid for Scientific Research on Priority Areas (14082206) (to K.K.) from the Ministry of Education, Culture, Sports, Science and Technology of Japan. This study was also supported in part by a grant from the Strategic Research Foundation Grant-aided Project for Private Universities from the Ministry of Education, Culture, Sports, Science and Technology, Japan (MEXT), 2011–2015 (grant no. S1101027).

Conflict of interest

None declared.

Abbreviations

CS, chondroitin sulfate; EDTA, ethylenediaminetetraacetic acid; GAG, glycosaminoglycan; HABP, hyaluronan binding protein; HME, hereditary multiple exostoses; HPLC, high-performance liquid chromatography; HS, heparan sulfate; PG, proteoglycan; ΔHexuronic acid, unsaturated β-d-gluco-4-enepyranosyluronic acid; ΔDi-0S, Δhexuronic acid-N-acetylgalactosamine; ΔDi-4S, Δhexuronic acid-4-sulfated N-acetylgalactosamine; ΔDi-4S,6S (ΔDi-diS_E), Δhexuronic acid-4,6-di-sulfated N-acetylgalactosamine; ΔDi-6S, Δhexuronic acid-6-sulfated N-acetylgalactosamine; ΔDi-UA2S,6S (ΔDi-diS_D), 2-O-sulfated Δhexuronic acid-6-sulfated N-acetylgalactosamine.

Acknowledgements

We thank Dr. Ishimaru and Dr. Takigami in the department of orthopedic surgery, Gifu University, for their great help in collecting patient blood samples and HME patient association “Bone Yuimarl” for their cooperation.