Journal of Inherited Metabolic Disease

, Volume 34, Issue 3, pp 755–761

The craniocervical junction following successful haematopoietic stem cell transplantation for mucopolysaccharidosis type I H (Hurler syndrome)

Authors

    • Department of Paediatrics, Children`s Hospital, Johannes Gutenberg-Universität Mainz
  • Heather Church
    • St Mary’s Hospital
  • Alan Cooper
    • St Mary’s Hospital
  • Jean Mercer
    • St Mary’s Hospital
  • Karen Tylee
    • St Mary’s Hospital
  • Robert F. Wynn
    • St Mary’s Hospital
  • J. Edmond Wraith
    • St Mary’s Hospital
Original Article

DOI: 10.1007/s10545-011-9309-5

Cite this article as:
Miebach, E., Church, H., Cooper, A. et al. J Inherit Metab Dis (2011) 34: 755. doi:10.1007/s10545-011-9309-5

Abstract

Mucopolysaccharidosis I Hurler (MPS IH) is a progressive multisystemic disorder caused by alpha-L-iduronidase deficiency. First choice of treatment in MPS IH children is haematopoietic stem cell transplantation (HSCT). The effect of HSCT has been shown to have limited influence on skeletal manifestations by poor penetration of musculoskeletal tissues by the enzyme derived from donor leucocytes. Aim of this study was to investigate the effect of HSCT on the craniocervical junction (CCJ) in Hurler patients. We analysed retrospectively sequential magnetic resonance imaging (MRI) scans of 30 patients with Hurler disease treated by HSCT since 1982 at the Royal Manchester Children’s Hospital, UK, in order to determine whether the patients suffer from dens hypoplasia. Results were compared with biochemical and clinical characteristics: Enzyme activity (EA), chimerism, urinary glycosaminoglycan (GAG) excretion and neurological status. Investigations were part of standard clinical procedures. Results are descriptive in presentation. In 26/30 patients a determination of odontoid hypoplasia was feasible. The majority showed a normal dens length and an increase with age. Only 3/26 revealed a dens hypoplasia. One of them had only partial donor engraftment (DE) with reduced EA, one of them suffered from chronic graft versus host disease (GVHD). One patient with only partial DE and reduced EA presented with initial dens hypoplasia until preadolescence but normalized later on. There may be a trend towards lower EA and the occurrence of DH in transplanted MPS patients - perhaps the dosage of enzyme plays a role in the correction of skeletal complications in this patient group. HSCT patients with incomplete DE and therefore lower EAs may require special attention and care.

Abbreviations

MPS

Mucopolysaccharidosis

H

Hurler

HSCT

Haematopoietic stem cell transplantation

GAG

Glycosaminoglycan

CCJ

Craniocervical junction

MRI

Magnetic resonance imaging

GVHD

Graft versus host disease

ERT

Enzyme replacement therapy

DE

Donor engraftment

EA

Enzyme activity

DS

Dermatan sulphate

CS

Chondroitin sulphate

MMC

Myelomalacia

Introduction

Mucopolysaccharidosis (MPS) I is a progressive and life-threatening lysosomal storage disorder caused by a deficiency of the enzyme α–L-Iduronidase that results in excessive accumulation of glycosaminoglycans (GAGs) in all tissues of the body. Hurler (H) syndrome is the classification used to describe the most severe clinical phenotype of this condition. Common clinical features include coarse facies, chronic rhinitis and otitis, corneal clouding, hepatosplenomegaly, valvular heart disease, respiratory insufficiency, joint stiffness and contractures, skeletal abnormalities and a progressive cognitive decline. The continued progression of disease leads to diminished physical activity and functional status, resulting in immobility. Death typically occurs by the age 10 due to fatal cardio-respiratory complications (Muenzer et al. 2009; Neufeld and Muenzer 2001; Pastores et al. 2007). Haematopoietic stem cell transplantation (HSCT) has been used in patients with the Hurler phenotype and has demonstrated some efficacy. There are improvements in liver and spleen storage, urinary GAG excretion, and in some functional areas (Aldenhoven et al. 2008; Cox-Brinkman et al. 2006; Peters et al. 2003). HSCT in patients younger than 2 years old, prior to developmental decline, has been shown to slow or prevent the inevitable mental degeneration of Hurler patients (Aldenhoven et al. 2008; Cox-Brinkman et al. 2006; Malm et al. 2008). However bone disease remains progressive after HSCT (Aldenhoven et al. 2008; Khanna et al. 2007; Staba et al. 2004; Tandon et al. 1996; Guffon et al. 1998; Weisstein et al. 2004; Field et al. 1994). In Hurler patients the odontoid process of the second cervical vertebra is often poorly developed and this can be associated with anterior atlantoaxial instability at the craniocervical junction (CCJ) putting the patient at risk of acute or chronic damage (myelopathy) to the upper spinal cord (Skeletal Dysplasia Group 1989). The CCJ consists of the occipital bone, foramen magnum and first two cervical vertebrae (Smoker and Khanna 2008). Dens abnormalities in MPS patients can vary from a total aplasia to different degrees of hypoplasia. In this study magnetic resonance imaging (MRI) films of patients with MPS I H who have undergone successful HSCT were reviewed. The presence of dens hypoplasia was noted. Sequential films gave an overview about dens development over time and correlation was sought between the nature of the abnormalities and the age at HSCT, degree of chimerism and residual level of GAG excretion.

Patients and methods

The patients chosen for this study were patients with the severe Hurler-variant treated by HSCT since 1982 at the Royal Manchester Children’s Hospital. The method of HSCT has not been specified, in some patients repeated procedures had been necessary. Following HSCT the biochemically and genetically confirmed Hurler patients have been seen regularly in the Willink Biochemical Genetics Unit based at the Royal Manchester Children’s Hospital, at least once a year. They have been examined in a systematic way: Measurement of enzyme activity (EA), chimerism, urinary GAG excretion and a clinical evaluation following standard procedures. The majority of patients have had a regular radiological scan of their spine. For this study we included all patients of the HSCT cohort after successful engraftment who had evaluable MRI scans of the CCJ, obtained by using different instruments offered by the neuroradiological department of the Royal Manchester Children’s Hospital. Different instruments have been used to perform MRI scans over years – images were of different qualities, therefore the comparability is somewhat limited. The midsagittal MRI films of the atlantoaxial region were analysed retrospectively. In the case of sequential films a descriptive longitudinal analysis was performed. The goal of the MRI evaluation was the determination of dens hypoplasia. We used T1-weighted MRI scans to evaluate the CCJ. Images of the upper cervical region were magnified into original sizes based on the MRI scale imprinted on films. To determine the overall length of the odontoid process on midsagittal T1-weighted MRI images we used the technique described by Cokluk et al. 2006 (Cokluk et al. 2006). After the age of 5 years, the odontoid process fuses with the body. In adulthood, the remnant of the dentocentral synchondrosis can be imaged by MRI as a hypointense structure between the inferior end of the odontoid and the superior margin of the body of C2. The length of the odontoid process was measured from its tip to the (remnant) synchondrosis (Cokluk et al. 2006). This technique is comparable to the measurement done by Mc Manners 1983 to evaluate height of the odontoid on plain radiographs of the lateral neck (McManners 1983). Several authors have stated that the definition of odontoid hypoplasia is that the tip of the odontoid process does not reach the upper margin of the anterior arch of the atlas (Smoker and Khanna 2008; Perovic et al. 1973; Kirks 1984). This method cannot be transferred to children. In normal children, the tip of the odontoid may fall well short of the upper margin of the anterior arch of the atlas, subsequently the overall height of the odontoid increases with age. This has important implications for the assessment of dens hypoplasia in children. Elliott analysed in 1988 the height of the odontoid in 508 normal children aged between 3 and 18.9 years. To assess the overall height of the odontoid peg he used the method advocated by McManners (Elliott 1988). To determine the presence of odontoid hypoplasia in our patients we compared our own values to Elliott’s: A dens hypoplasia was defined if our own value fell below the range defined by mean of age- and gender-matched Elliot-data minus to standard deviations. We were not able to find childhood odontoid norm values on a basis of MRI scans. It has to be considered that Elliott’s norm values were evaluated by the measurement of plain X-rays. In patients younger than 3 years we measured the length of the odontoid process but did not determine a dens hypoplasia because of missing norm values and insufficient ossification. An increased signal intensity on the T2-weighted midsagittal image of the cervical spine was interpreted as myelomalacia (MMC) – a radiological sign of cervical myelopathy. Results are descriptive in presentation. To analyse and compare our data we divided our patients into seven age groups: 1 = 0 to 2.9 years, 2 = 3 to 5.9 years, 3 = 6 to 8.9 years, 4 = 9 to 11.9 years, 5 = 12 to 14.9 years, 6 = 15 to 18.9 years, 7 = older than 18.9 years. The results of the last investigations were used in those patients who were studied longitudinally.

Results

There was a total cohort of 42 Hurler patients evaluated following HSCT at the Royal Manchester Children`s Hospital during the study period. 4/42 deceased during early childhood from GVHD (graft versus host disease), in 8/42 analysable reports and images were missing. We analysed 30 patients at the age of 2.3 to 20.3 years (mean 8 years), 18 male and 12 female patients. HSCT was performed at the age of 7 to 23 months, the mean age was 13.5 months. In 10 patients the first HSCT procedure failed. The second transplantation took place at a mean age of 22.5 months. The follow-up time varied from 1 to 19.1 years with a mean follow-up interval of 5.10 years, relating to first transplantation. Table 1 summarizes the clinical characteristics of our patients. Three patients had only partial donor engraftment (DE) following HSCT: patient 7: 80%, patient 9: 80% and patient 29: 50%. Among them, patients 9 and 29 had a “low“ enzyme activity between 5 and 10 μmol/g/hr (normal range 10–50 μmol/g/hr) - corresponding to heterozygous levels. Patient 15 received bone marrow from his heterozygous brother – he has a complete DE but also enzyme activities in the heterozygous range. The remaining patients had normal α-L-Iduronidase activity following HSCT and all patients showed a significant reduction of urinary GAGs. Only 1 patient had objective neurological signs indicating a cervical myelopathy. This patient (9) developed hyperreflexia in upper and lower extremities at the age of 12 years and 9 months. A hyperintensity on the cervical cord as an indicating factor for cervical myelopathy was seen in only two patients (9; 18). Patient 9 developed signal changes at the age of 7 years. He was also suffering from severe dens hypoplasia, atlantoaxial instability and cord compression degree 2. His neurological evaluation also revealed signs of cervical myelopathy as described above. In patient 18 cervical myelomalacia on MRI imaging was first documented at the age of 4 years. He did not develop any clinical symptoms but went on to have a craniocervical decompression at the age of 6 years.
Table 1

Summary of present results (Partial engraftment defined as engraftment < 90%)

ID

sex

Age (years)

HSCT (months)

DH

DE

EA (μmol/g/hr )

GAG (mg/mmol creatinine)

Clinic evaluation abnormal?

MMC

Total amount of scans

1

m

4.8

15

complete

18

23

3

2

m

3.0

17

+

complete

12

20

1

3

m

8.5

17

complete

20

10

3

4

m

12.1

15

complete

22

7

2

5

f

4.2

10/15

complete

25

19

3

6

m

17.1

10

complete

24

7

NA

2

7

m

4.9

12/20

partial

16

16

3

8

f

2.3

16

 

complete

44

28

1

9

m

13.3

9

+

partial

9

18

+

+

6

10

m

5.5

10

complete

26

14

5

11

m

3.8

15

complete

25

37

3

12

f

20.3

15

complete

18

NA

4

13

m

10.1

19

+

complete

14

12

8

14

f

7.0

13/21

complete

16

23

6

15

m

3.7

16/32

complete

5

33

3

16

f

3.0

11

complete

40

17

2

17

f

15.1

15/18

complete

32

8

1

18

m

6.3

15

complete

22

12

+

2

19

f

7.5

23

complete

29

NA

5

20

m

5.8

12/27

complete

27

16

3

21

m

10.7

12

complete

19

10

3

22

f

2.3

15

 

complete

27

42

1

23

f

2.8

9

 

complete

25

24

1

24

f

5.4

11/16

complete

13

30

2

25

m

13.2

7/26

complete

20

8

6

26

f

13.5

14/30

complete

41

8

5

27

f

4

20

complete

17

26

2

28

m

2.8

21

 

complete

13

28

1

29

m

9.3

17/20

partial

8

40

4

30

m

17.8

11

complete

21

4

2

MRI scans were available for review in 30 patients with MPS I H with a total amount of 93 MRI scans. In a fourth of the images (23/93) the patients were younger than 3 years. In general, the first post-HSCT-scan was performed 12 months after the procedure at the earliest. In our patient cohort the median age of the first post-transplant review was 5.1 years with a range from 1.7 years to 16.10 years. 22/30 had at least one MRI follow up with a minimal time span of 12 months. In 26/30 patients a determination of the presence of odontoid hypoplasia by measurement of MRI scans was feasible. As mentioned above we had to take out the four patients younger than 3 years.

Of the remaining 26 individuals 23 (88.5%) currently showed a normal dens length, in 19 of them there was a normal odontoid height in each available scan. Most children presented with a normal dens length at the age of 3 years (at least 12 months after HSCT). Dens hypoplasia with subsequent normalization was detected in one patient (29): In this case dens hypoplasia was diagnosed in early childhood and confirmed in yearly follow-up reviews. By the age of 9.3 years he subsequently developed a normal sized dens at the lower range with a height of 12.3 mm. Patients 1, 18 and 25 revealed the diagnosis of dens hypoplasia in only one of all available measurements at the age of 3.8, 5.7 and 7.2, respectively. In their last evaluation at the age of 4.8, 6.3 and 13.2 years they presented with normal odontoid length.

3/26 revealed a dens hypoplasia over time (2; 9; 13). Case 2 and 9 showed an odontoid dysplasia in each available MRI scan. Patient 13 had several MRI scans between the age of 4.6 and 7.6 years, where the peg length varied between 7.6 and 8.4 mm without any significant dens growth. Depending on the age-related norm values the odontoid was classified as a hypoplastic dens or as normal but in the lower range. At the age of 9 years a growth spurt was noted but growth remained below the lower normal range.

Figure 1 shows the odontoid length divided in different age groups (not gender-matched). Each available scan is included, so some patients are included repeatedly. The overall odontoid height increases with age. Looking specifically at the dens length of our patients with DH the overall odontoid heights range between 7 and 10 mm, regardless of age and sex of our patients (pt 2 = 7 mm; pt 9 = 7 mm; pt 13 = 10 mm; pt 18 = 9 mm). The mean age at HSCT in the patient group with normal developed dens was 14 months, the mean age at HSCT in the patient group with dens hypoplasia was 14.3 months. Mean urinary GAGs in the patient group with normal developed dens were 20 mg/mmol creatinine, GAGs in the patient group with dens hypoplasia were 17 mg/mmol creatinine. We found a mean ratio of 0.57 in the patients with normal developed dens and a dermatan sulphate/chondroitin sulphate (DS/CS) ratio of 0.7 in the patients with abnormal dens length. Figure 2 shows MRI examples of the CCJ of three of our patients to illustrate the difference between normal age dependent dens development and the presence of dens hypoplasia in childhood.
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9309-5/MediaObjects/10545_2011_9309_Fig1_HTML.gif
Fig. 1

Odontoid length of MPS I H patients following successful HSCT divided in different age groups

https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9309-5/MediaObjects/10545_2011_9309_Fig2_HTML.gif
Fig. 2

(a) patient 25 at the age of 13.2 years – normal dens length (16.9 mm); (b) patient 10 at the age of 5.5 years – normal dens length (11.7 mm); (c) patient 9 at the age of 13.3 years – dens hypoplasia (6.9 mm)

Discussion

Influence of differences relating HSCT protocols through the years has not been considered in our study. Nevertheless, the incidence of DH in HSCT – treated MPS I patients is likely to be lower than in untreated patients. Unfortunately, MRI scans of untreated Hurler patients were not available at the Royal Manchester Children`s Hospital, therefore is no possibility to compare both patient groups directly. Baseline incidence of odontoid hypoplasia in untreated Hurler patient population is not available. Wynne-Davies et al. reported findings of spine radiographs of six untreated MPS I H patients at the age of 1 to 13 years; 4/6 presented with odontoid hypoplasia (Skeletal Dysplasia Group 1989). In our MPS I H HSCT cohort an odontoid hypoplasia was only present in less than 1/3 of patients, so the majority showed a normal dens growth. There was no dens aplasia in our patient group. Referred to Wynne Davies` literature, where 2/3 of untreated MPS I patients suffer from DH, transplantation seems to have a favourable impact on prognosis of dens development (Skeletal Dysplasia Group 1989). To confirm this presumption further studies are required.

Tandon et al. investigated 12 MPS I H patients after HSCT. In 4/12 a dens hypoplasia was diagnosed on the initial radiological scans, two of them subsequently normalized (Tandon et al. 1996). Numbers fit to our result that overall incidence of odontoid hypoplasia in transplanted Hurler patients seems to be low.

In spite of missing statistical evidence because of the comparatively small amount of patients our results support the conclusion of the Canadian working group that reported correction of odontoid dysplasia following successful HSCT (Hite et al. 2000). However it is important to distinguish the following issue: Hite et al. analysed the improvement of morphologic changes (odontoid dysplasia) whereas this study focused on the dens hypoplasia without respect to morphology intending to develop a quantitative measure. In many of the early scans we found an abnormal odontoid morphology with different severity of dysplasia (e.g. deformed base of the tip). In most cases dens morphology improved over time.

Odontoid height in our patient group compared to the values measured by Elliott in normal children does not vary significantly (Fig. 3). Cokluk et al. investigated the size of the peg in 66 adult patients (Cokluk et al. 2006). Mean height in normal adults was 21.3 mm, these values are also comparable to ours. Dens length in successfully HSCT-treated patients does not lie in the lower range of norm values.
https://static-content.springer.com/image/art%3A10.1007%2Fs10545-011-9309-5/MediaObjects/10545_2011_9309_Fig3_HTML.gif
Fig. 3

Odontoid length in HSCT-treated MPS I H patients and odontoid length in normal children evaluated by Elliott 1988

There was no correlation between the presence of odontoid hypoplasia and age of HSCT and between dens hypoplasia and urinary GAG excretion. Regarding the DS/CS ratio the mean ratio in the patients with dens hypoplasia is slightly higher than in the rest of our patients.

One of the 3 patients with odontoid hypoplasia had only a partial DE (9). One of the remaining cases suffered from chronic graft versus host disease (GVHD) (13). Patient 2 with his 36 months (19 months after HSCT) is a very young patient, who has only a mild dens hypoplasia (dens length 6.9 mm, normal values 7 – 8.6 mm) and may improve over the course of time. Case 9 raises the question if patients with partial donor engraftment may be on a higher risk to evolve a dens hypoplasia. Looking specifically at the remaining 2 patients with only partial donor engraftment (29 and 7) they currently present with a normal dens size. Patient 29 showed an initial dens hypoplasia until preadolenscence. Patient 7 has only partial donor engraftment but reveals normal enzyme activity levels. There is an additional patient to the total Manchester cohort of 42 HSCT patients that is not included in this study because his MRI scans were not evaluable for measurements. Nevertheless, we got information from his MRI reports. This patient was transplanted at the age of 12 months. DE in this patient was partial with 65%, EA between 5 and 10 μmol/g/hr. He initially presented at the age of 3.5 years with a normal peg length but then fell out of norm by the age of 5.5 years. This additional case may support a thesis that there is a trend towards lower enzyme levels and poorer MPS clearance in patients with persistent dens hypoplasia after HSCT. Hence we propose the thesis that the level of enzyme in our transplanted patients may influence dens development. Further studies with more patients are necessary to confirm if patients with incomplete DE and therefore lower enzyme activities may really be at a higher risk to evolve odontoid hypoplasia.

Conclusion

Successful HSCT seems to have an early and positive influence on the CCJ. Our retrospective study suggests that patients with MPS I H show a progressive improvement of the underdeveloped axis after successful engraftment following HSCT. Despite the fact that the use of radiological techniques to determine DH and the definition of hypoplasia are missing in Wynne-Davies´ report, the difference of incidences in his untreated Hurler patient group and our HSCT-treated patient cohort is impressive and emphasizes the hypothesis that HSCT influences the development of the odontoid (4/6 for all MPS versus 3/26 in the present study).

The present study suggests that the level of enzyme in our transplanted patients may play an important role in dens development. Transplanted patients with lower EA need special attention and care. Tsukimura et al. recently published their results concerning the uptake of a recombinant human alpha-L-iduronidase by cultured fibroblasts and osteoblasts (Tsukimura et al. 2008). Authors suggest the administration of a high dose of enzyme for further improvement of enzyme replacement therapy (ERT) for skeletal disorders caused by MPS I. They obviously also think about a dose-depending efficacy of enzyme on the bone disease. First, the influence of ERT standard therapy on dens development should be analysed. Further studies with different doses of ERT may be reasonable, to check influence of higher doses on dens abnormalities as such and on other residual disease manifestations in HSCT patients. Administration of higher doses of enzyme may possibly improve the dens morphology even in MPS patients under ERT.

Study limitations

The absence of a concurrently followed untreated control group limits the strength of the conclusions regarding the effect of HSCT on the CCJ in Hurler patients.

Different instruments have been used to perform MRI scans over years – images were of different qualities, therefore the comparability is somewhat limited. Additionally, we were not able to find childhood odontoid norm values on a basis of MRI scans. Elliott’s norm values were evaluated by the measurement of X-rays not MRI images – which leads to a further methodical constriction. The way of HSCT has not been specified, in some patients repeated procedures had been necessary.

Details of funding

The study is sponsored by an independent educational grant by the Deutsche Forschungsgesellschaft (DFG).

Copyright information

© SSIEM and Springer 2011