Mendy Welsink Karssies

Classical Galactosemia is an inborn error of metabolism, a disease in which galactose (milk sugar) is not processed correctly in the cells of the body. Lactose from milk and dairy products is converted into galactose and glucose, mainly in the gut. Normally, galactose is transported through the bloodstream and used by the body for energy after it has been converted into glucose in the liver. The enzyme galactose-1-phosphate uridylyltransferase (GALT) is an enzyme essential for one of the steps in the conversion of galactose into glucose. Galactosemia literally means galactose in the blood. In patients with Classical Galactosemia (CG), the enzyme GALT is not working properly as a result of inherited changes in the DNA. The accumulation of galactose in the blood leads to toxic products upstream of the GALT enzyme and a shortage of products downstream of the GALT enzyme. Both mechanism are presumed to contribute to the complications observed in patients. Patients are diagnosed with CG if the activity of the GALT enzyme measured in red blood cells is below 15% of controls and/or if two disease-causing mutations are found in the DNA.

In newborns with CG, the ingestion of galactose from breast milk or infant formula results in a life-threatening disease in the first weeks of life affecting multiple organs, such as the liver, kidneys, the eyes and the brain. The only available treatment is to strictly limit the intake of galactose from the diet. After the start of a galactose-restricted diet, newborns with CG quickly recover. However, most patients develop long-term complications affecting the brain resulting in motor developmental delay and speech developmental delay, a below average IQ, cognitive impairment and movement disorders. In most female patients, the ovaries are affected as well, leading to an early menopause, also known as premature ovarian insufficiency (POI). Even the early detection of patients by family screening (the screening of siblings of CG patients) with an immediate start of galactose restriction after birth, and the inclusion of CG in the newborn screening program (in the Netherlands since 2007) leading to a diagnosis and start of treatment in the first week of life, does not prevent long-term complications. The exact disease mechanism and timing of damage is not yet clear. The clinical outcome of patients is highly variable and ranges from severely affected to completely normal, which is poorly understood. To be able to prevent long-term complications in CG, more insight into the long-term outcome of patients is needed. Markers that are able to predict the course of disease are yet to be discovered but urgently needed as they can provide clarity for patients and families. In addition, these markers may serve as target for the development of new treatment options and could be used to assess their effect.

The aim of this thesis is to elucidate the broad spectrum of clinical outcome of CG and to search for (prognostic) markers of clinical outcome. Part 1 of this thesis explores the clinical and biochemical outcome of patients and in Part 2 potential (bio)markers and their association with clinical outcome are investigated. As galactose is present in many products, a lifelong galactose-restricted diet can be a challenge for patients. Therefore, we developed patient education materials for children and adolescents with inborn errors of metabolism and tested if these materials improve knowledge on the disease and treatment in Part 3.

Part 1 The spectrum of clinical outcome
The clinical outcome of patients is highly variable and patients and even siblings with identical disease-causing mutations can be found on opposite ends of the clinical outcome spectrum. In Chapter 2, both the symptoms in the newborn period and the prevalence and extent of long-term complications were evaluated. In our cohort of 56 patients, differences in mutations (genotypes) and the severity of complications (phenotypes) were found. In our cohort, we were able to distinguish three different groups of patients.

The first group of patients includes 47 patients with classical phenotypes with low to undetectable GALT enzyme activity in red blood cells (<3%) and two disease-causing mutations (hereafter named classical patients). This group consists of 32 patients who were diagnosed before 2007 after they presented with CG related symptoms (on average diagnosed and treated within the second week with a maximum of six weeks) and of 15 patients detected by family screening or newborn screening since 2007 who were treated early (within the first week of life). A majority of the classical patients have long-term complications, but the clinical outcome varies considerably within this group. The second group of patients includes seven patients detected by newborn screening with previously unreported genotypes, residual GALT enzyme activity in red blood cells up to 10% and no symptoms at diagnosis (hereafter variant patients). The third group includes two patients with two specific ‘SNPSL’ mutations found in patients of African descent, causing GALT deficiency in red blood cells but residual GALT enzyme activity in other tissues (hereafter patients with two SNPSL mutations). Both patients with two SNPSL mutations were diagnosed before 2007 and were diagnosed late after they presented with CG related symptoms at age 7 months (feeding difficulties and cataract) and age 10 years (visual impairment due to cataract). To investigate if an early treatment resulted in a more favorable long-term outcome, the prevalence of long-term complications was investigated within the group of classical patients. Even though patients detected through family screening demonstrated no symptoms in the newborn period and the patients detected by newborn screening demonstrated fewer and less severe illness in the newborn period than patients diagnosed before 2007, an early dietary treatment did not result in a more favorable long-term outcome. The long-term outcome described in Chapter 2 includes the neurological outcome of patients and the endocrinological outcome (hormonal and reproductive status) of female patients. The intellectual outcome and cognitive abilities of patients (neuropsychological outcome) are described in Chapter 4. Neurological outcome Of the 56 patients in our cohort, 36 were examined for neurological signs and symptoms. In 17 out of 36 patients (47%) a movement disorder was found, which was more frequently in adults (8 out of 14, 57%) than in children (9 out of 22, 41%). All patients with a movement disorder were classical patients. All variant patients and both patients with two SNPSL mutations were examined for neurological signs and symptoms, but no movement disorder was found. The movement disorders varied from a mild tremor to a severe tremor (involuntary muscle contractions leading to shaking of one or more parts of the body) and dystonia (involuntary muscle contractions leading to abnormal posture and spasms) interfering with daily functioning. In patients with a movement disorder, the frequency of an abnormal motor development (42%) and speech development (75%) was higher when compared with patients without a movement disorder in which 16% had an abnormal motor development and 38% had an abnormal speech development. Endocrinological outcome Data were available for 21 female patients aged 12 years and older. In 8 out of 19 female patients, a delayed puberty was induced with medication. In two patients information was missing. Primary ovarian insufficiency (POI) was diagnosed in 12 out of 17 (71%) female patients and was uncertain in four. The female variant patients were too young for a endocrinological evaluation. Both patients with two SNPSL mutations were not diagnosed with POI. Neuropsychological outcome Besides the intellectual abilities, neuropsychological functioning of CG patients was investigated both in the literature (Chapter 3) and in our own cohort (Chapter 4). Most studies focus on the IQ of patients rather than investigating the underlying and more specific abilities involved in knowledge gain and comprehension (cognition). Cognitive functioning can be divided into multiple domains such as information processing speed, attention, memory, visuospatial functioning (the ability to process and interpret visual information) and executive functioning (an overall term used for skills that include working memory, flexible thinking and self-control). As these domains provide information on the specific abilities of patients, it is important to investigate them separately. In Chapter 3, we reviewed the available literature on cognitive functioning in CG patients and identified 11 studies. As a group, patients had below average scores on all cognitive domains. However, the results differed between studies, most studies used a limited number of tests or reported on small cohorts. Also, most studies were of limited quality. In Chapter 4, we investigated multiple cognitive domains and social functioning in our cohort of patients. The 48 included patients had an average IQ of 77 (range 45-103), which is below average since the average IQ of the general population is 100 and an IQ below 85 is below average. All variant patients had an IQ above 85, while both patients with two SNPSL mutations had an IQ well below 85. In the group of classical patients, a majority of the patients (68%) had an IQ below 85. Compared to the general population, CG patients demonstrated impaired cognitive functioning without a clear profile. The results indicated that cognitive outcomes of patients may exceed IQ based expectations and thus the IQ may not reflect the abilities of patients. As it has been suggested that CG patients demonstrate autistic traits, we investigated certain aspects of social functioning (i.e. social responsiveness). On average, the (parents of) patients reported normal levels of social responsiveness without features suggestive of autism. Most results were comparable to the general population, implying that social responsiveness is not impaired in CG. Since social problems, withdrawn behavior and anxiety were frequently reported by (parents of) patients, social functioning should be further investigated. Based on the large variability in cognitive, behavioral and social functioning without a clear profile, a neuropsychological evaluation and follow-up is advised in all patients to provide patients with support if needed. Brain abnormalities on MRI As the long-term complications mainly involve the brain, the brain MRIs of CG patients were evaluated in Chapter 2. In 86% of the patients brain abnormalities were found, most frequently a reduced brain volume (i.e. atrophy of the cerebrum and cerebellum). White matter lesions indicating damage to the white matter were found in patients, but the extent of visible white matter lesions was limited. The presence of white matter lesions was not associated with IQ nor with movement disorders, but atrophy of the cerebrum and cerebellum was more frequently present in patients with a lower IQ and/or a movement disorder. As changes in the brain may not be visible on MRI, computer programs were used to evaluated gray- and white matter in detail and the results are described in Chapter 7. Biochemical outcome The biochemical outcome of patients is described in Chapter 2. The biochemical outcome of the seven variant patients and two patients with two SNPSL mutations differed from the 47 classical patients. In CG patients, the accumulation of toxic products upstream of the GALT enzyme, such as Gal-1-P and galactitol can be measured in red blood cells and urine respectively. In classical patients, both Gal-1-P and galactitol remain elevated despite restriction of galactose in the diet. In all but one variant patient, Gal-1-P levels were below the detection limit during dietary treatment just as in a healthy population and galactitol levels were within the normal range. This indicates that the variant patients, who currently demonstrate no long-term complications, biochemically differ from classical patients. Both patients with two SNPSL mutations demonstrated lower Gal-1-P levels and galactitol levels when compared with classical patients, but still higher than the levels observed in the variant patients. Therefore, we investigated if differences in the clinical outcome of CG patients are caused by individual differences in the extent of galactose intoxication (i.e. higher Gal-1-P levels indicate more galactose intoxication) and could be associated with a poor clinical outcome. - Gal-1-P The most recent Gal-1-P level of CG patients was not significantly different between patients with a poor clinical outcome (IQ<85, the presence of movement disorders and/or POI in females) and with a normal clinical outcome (IQ≥85, the absence of movement disorders and/or POI). In the group of classical patients, the differences in Gal-1-P levels between patients were small and were not associated with clinical outcome. - Lifetime Gal-1-P As the most recent Gal-1-P may not represent overall galactose intoxication, lifetime Gal-1-P was evaluated as well. Gal-1-P has been reported to reach a steady state within a few months up to one year after the start of a galactose-restricted diet. Therefore, we defined lifetime Gal-1-P as the average of all available Gal-1-P levels reported after 12 months. As variant patients and patients with two SNPSL mutations demonstrated a different biochemical profile, lifetime Gal-1-P was only evaluated in classical patients. Just as the most recent Gal-1-P level, lifetime Gal-1-P was not significantly different between patients with and without long-term complications. To investigate if Gal-1-P had reached a steady state after 12 months, we studied its association with age. The finding of a significant negative correlation indicates a more gradual decline of Gal-1-P after 12 months. This is an important finding as Gal-1-P is still considered as a potential marker of clinical outcome. A shortage of products downstream of the GALT enzyme leads to a reduced production of ‘UDP sugars’, which are needed for the production of proteins and lipids. Glycosylation is a process in which sugars are added to proteins and lipids creating special glycoproteins and glycolipids that are involved in many processes of the body. In CG patients, glycosylation abnormalities have been demonstrated. Therefore, we investigated if glycosylation abnormalities (as indicated by N-glycan profiles) were associated with the clinical outcome of patients. - IgG N-glycans To be able to investigate glycosylation abnormalities, glycoproteins and more specifically N-linked glycans on Immunoglobulin G (IgG) were analyzed, as IgG has well defined glycan structures which have been associated with the development of diseases (e.g. autoimmune diseases, infectious diseases and cancer). Considering IgG is present in the blood and blood samples are easy to collect, this method is minimally invasive for patients. The N-glycans were analyzed with the use of an automated glycan preparation method linked to ultra-performance liquid chromatography, which is used to separate and distinguish N-glycans from other components. Even though significant differences in N-glycans were found between patients and controls, N-glycans were not significantly different between patients with a poor and a normal intellectual, neurological and endocrinological outcome. Part 2 Towards individual prognostication Galactose oxidation capacity In this thesis, we hypothesized that differences in clinical outcome are caused by differences in galactose oxidation capacity. A (slightly) higher enzyme activity might lead to a (clinically relevant) higher ability to convert galactose (i.e. galactose oxidation capacity), which will cause lower Gal-1-P levels and less abnormal glycosylation and possibly a more favorable clinical outcome. In Chapter 5, we studied galactose oxidation capacity in cultured skin cells (i.e. fibroblasts) of CG patients and controls. We chose fibroblasts because performing a skin biopsy is minimally invasive and skin cells can be multiplied easily by culturing. We used a previously developed method (galactose metabolite profiling, GMP) to determine galactose oxidation capacity. The Galactose Index (GI) (i.e. the ratio of [U-13C]Gal-1-P / [13C]UDP-galactose) distinguished classical patients from controls, from patients with two SNPSL mutations and from variant patients. As Gal-1-P decreases and UDP-galactose increases in patients with higher galactose oxidation capacity, a lower GI indicates more residual galactose oxidation capacity. The controls in our cohort demonstrated a barely detectable GI (0.84 – 1.12) and the variant patients demonstrated a GI in the low range (1.61 – 7.17). Interestingly, the patients with two SNPSL mutations demonstrated a lower GI than most variant patients (2.47 and 5.40), indicating more residual galactose oxidation capacity in fibroblasts. The classical patients demonstrated a large variability in GI (9.04 – 22.37) which was not associated with clinical outcome (i.e. a higher GI was also found in patients with a normal clinical outcome). In Chapter 6, we determined whole body galactose oxidation capacity in patients and controls by using a non-invasive 1-13C galactose breath test. Galactose is a sugar which consists of six carbon atoms. For this study one out of six atoms was labeled. Breath samples were collected for patients and controls both before and 60, 90 and 120 minutes after they ingested a harmlessly small amount of labeled galactose. The amount of labeled galactose in the breath samples was measured with the use of gas-isotope-ratio mass spectrometry, which can reliably distinguish small molecules. Hereafter, the percentage of oxidized galactose (CUMPCD) was calculated. The 1-13C galactose breath test was able to distinguish classical patients (median CUMPCD TN2O: 0.35) from variant patients (median CUMPCD TN2O: 13.79), from patients with two SNPSL mutations (median CUMPCD TN2O: 9.44) and from controls (median CUMPCD TN2O: 9.29). In the group of variant patients with residual GALT enzyme activity in red blood cells up to 10% and undetectable Gal-1-P levels, the whole body galactose oxidation capacity was within the control range. The higher galactose oxidation capacity and lower Gal-1-P levels compared with classical patients and the finding that the variant patients currently have a normal clinical outcome supports our hypothesis. As the differences in galactose oxidation capacity within the group of classical patients were small and galactose oxidation capacity was not able to differentiate between classical patients with a poor and normal clinical outcome, the potential of residual galactose oxidation capacity as (bio)marker in patients with CG seems limited. MRI of the brain studied in more detail In Chapter 7, various MRI techniques and analysis programs were used to investigate the brain of 21 CG patients and 24 controls. Of the 21 patients, 15 patients (71%) had an IQ below 85. In total 16 out of 21 patients were examined for neurological signs and symptoms and in 9 out of 16 patients (56%) a movement disorder was found. In the brain, white matter is responsible for the communication between nerves and gray matter processes the information. On MRI, the difference between gray- and white matter is visible due to the higher content of myelin in white matter. Myelin is a mixture of proteins and lipids that form a protecting sheath around nerves. White matter abnormalities are frequently reported in CG and therefore we investigated both total white matter and a specific white matter tract, the corticospinal tract, which is responsible for controlling movements of the limbs and torso. The gray- and white matter volume, white matter integrity and myelin content in both gray matter, white matter and the CST was lower for patients when compared to controls. Moreover, these values were associated with the neurological and intellectual outcome of patients (i.e. gray and white matter volumes, white matter integrity and myelin content were lower for patients with more severe movement disorders and/or a lower IQ when compared to patients without long-term complications). The extent of visible white matter lesion was limited in CG patients as the white matter lesion volume of patients was comparable to healthy subjects. The results of this study demonstrate that not only white matter is affected in CG patients, but gray matter as well. The finding that the structural changes in the brain of CG patients were associated with neurological and intellectual outcome indicate that MRI can be of use to further unravel the disease mechanism of CG. The clinical outcome of patients (i.e. cognitive impairment and movement disorders), the absence of widespread white matter lesions and the presence of both gray- and white matter changes on MRI could indicate that CG is primarily a gray matter disease with secondary damage to the white matter. As this was an explorative study in a small patient group, more research is needed in a larger cohort in which the course of gray- and white matter should be investigated by repeating the MRI over time. Part 3 Patient education The final part of this thesis addresses patient education. The galactose-restricted diet is currently the only available treatment and patients are advised to stick to this diet for the rest of their lives. Galactosemia and inborn errors of metabolism in general are complex, and children and adolescents suffering from these disorders may benefit from patient education materials that improve knowledge on disease and treatment. In the study described in Chapter 8, patient education materials were developed for children and adolescents with galactosemia as well as other inborn errors of metabolism and were tested. As results demonstrated a significant knowledge gain in patients, these materials could be of use to explain complex diseases and treatment to children and adolescents suffering from inborn errors of metabolism.