Original Research Article

Dement Geriatr Cogn Disord 2002;14:128136

Accepted: May 7, 2002

Alzheimer Disease Effect of Continuous Intracerebroventricular Treatment with GM1 Ganglioside and a Systematic Activation Programme
Lars Svennerholm Grel Brne Ingvar Karlsson Annika Lekman Ingalill Ramstrm Carsten Wikkels
Department of Clinical Neuroscience, Gteburg University, Gteborg, Sweden

Key Words Alzheimer disease W GM1 ganglioside W Intracerebroventricular W Training programme W Transmitter substances

Abstract Five patients with the early-onset form of Alzheimer disease (AD) received GM1 ganglioside by continuous injection into the frontal horns of the lateral ventricles for a period of 12 months. The optimal GM1 dose varied between 20 and 30 mg/24 h. The patients were trained twice a week for 45 h with an individually designed cognitive programme, which included the use of a word processor. Neurological, neuropsychological, psychiatric and neurochemical examinations were performed a week before surgery and on days 30, 90, 180, 270 and 365 after surgery. The cerebrospinal fluid levels of the monoamine metabolites homovanillic acid and 5-hydroxyindoleacetic acid and the neuropeptide somatostatin increased. The regional cerebral blood flow showed a tendency to increase. The progression of deterioration was stopped, and motor performance and neuropsychological assessments improved. The patients became more active and felt safer in relation to other peo-

ple and performing various activities. They had improved reading comprehension and a better feeling for language. They were able to write reports and short letters on a word processor. When interviewed at the end of the study, all 5 patients stated that they felt better, and their relatives reported that they had regained integrity and their joie de vivre.
Copyright 2002 S. Karger AG, Basel

Introduction

Alzheimer disease (AD) is characterized by progressive loss of nerve processes, which is the neurobiological basis for the progressive deterioration [1]. Gangliosides, normal lipid components of the mammalian plasma membrane, are particularly abundant in the outer leaflet of the neuronal membrane [2]. A significant reduction of gangliosides was found in some studies of brains from classical (early-onset) AD patients [3, 4]. Indeed, in our own studies of brains from early-onset AD cases, we found a marked diminution of gangliosides in all grey matter areas examined frontal and temporal cortices, hippocampus and caudate nucleus [4] suggesting a pronounced loss of nerve processes. Because gangliosides have been shown to

ABC
Fax + 41 61 306 12 34 E-Mail karger@karger.ch www.karger.com

2002 S. Karger AG, Basel 14208008/02/01430128$18.50/0 Accessible online at: www.karger.com/journals/dem

Dr. Annika Lekman Department of Clinical Neuroscience Sahlgrens University Hospital/Mlndal S431 80 Mlndal (Sweden) Tel. +46 31 343 2415, Fax +46 31 343 2426, E-Mail annika.lekman@vgregion.se

Table 1. Characteristics of 5 Alzheimer patients at the beginning of Gm1 treatment and the cognitive and social training programme

Patient 1 Age, years Sex Education years Previous occupation Heredity Duration of symptoms, years Neurological symptoms EEG MRI + CT rCBF flow reduction 69 female 8 instructor sister, presenile AD 5 rigidity subclonus normal parieto-occipital atrophy (+)

Patient 2 67 female 10 nurse 5

Patient 3 65 female 7 + English courses shipping agent 5

Patient 4 68 male 9 + lower business school accountant 67 gait disturbance, hypokinesia

Patient 5 54 female 7 housewife 34

normal frontotemporal atrophy

slow potentials, slow basic rhythm no abnormal finding left frontal lobe and Sylvian fissure region + 22/30

widening of lateral ventricles

strong episodic abnormality parieto-occipital atrophy (+)

right parietal lobe ++ frontal lobes + right temporal lobe (+)

right parietal and parietal and parietotemporal parietotemporal lobe ++, same regions lobes bilateral + on the left (+) 24/30 24/30

MMSE

23/30

24/30

have neuronotrophic and neuritogenic activity [5], a trial in which AD patients were treated for 3 months with subcutaneous or intramuscular injection of GM1 ganglioside was attempted [6]. No clinical or biochemical effects on the CNS were observed, suggesting to us that a negligible amount of GM1 had crossed the blood-brain barrier. We therefore developed a technique whereby GM1 ganglioside could be continuously administered to the CNS and treated 5 patients with classical early-onset AD with GM1 administered by intracerebroventricular infusion over a period of 1 year [7]. The results of this preliminary report suggested that the pharmacological treatment should be complemented by a carefully designed programme of cognitive, social and physical training. These principles have been followed in the new study, and this report describes the encouraging results of the study in improving and stabilizing some of the impairments associated with AD and documents changes in clinical, biochemical and neurophysiological variables. As a disorder of progressive impairment of mentation, AD has no animal model with which to study basic mechanisms and potential therapies. Thus, feasibility and safety and, if possible, efficacy must be examined in human studies.

Patients and Methods

Selection of Patients Patients had to have classical AD [8] based on the NINCDSADRDA criteria for probable AD [9]. Before entry, each patient underwent a physical, neurological and psychiatric evaluation, neuropsychological tests rating with the Gottfries-Brne-Steen scale [10], measurement of regional cerebral blood flow (rCBF), computed tomography (CT) and/or magnetic resonance imaging (MRI), electroencephalography (EEG) and an extensive clinical chemical and neurochemical examination. On the basis of these studies, entry criteria were: age, 5070 years; at least one relative willing to and capable of evaluating the effect of treatment; mild to moderate dementia [a score of 4 on items measuring intellectual impairment in the Gottfries-Brne-Steen scale and a score of 1525 (preferably closer to 25) on the Mini Mental State Examination (MMSE)]. Exclusion criteria were: presence of signs of major medical illness or mental illness in addition to AD, history of inflammatory brain disease, multiple episodes of head trauma, alcohol or drug abuse, or concomitant therapy with CNS-active drugs that might markedly influence brain function. The characteristics of the 5 patients and their clinical symptoms are shown in tables 1 and 2. All patients and their spouses signed consent forms after obtaining detailed oral and/or written information about the treatment. They were informed that they could interrupt the treatment at any time. The study was approved by the Ethics Committee, Faculty of Medicine, Gteborg University, and the Swedish Medical Products Agency.

GM1 Ganglioside Treatment and Activation Programme

Dement Geriatr Cogn Disord 2002;14:128136

129

Table 2. Function and/or symptom and neuropsychological tests of 5 Alzheimer patients

Patient 1 Day: Short-term memory impairment (recent memory, item I 4, G scale) Long-term memory impairment (distant memory, item I 5, G scale) Personality change Emotional function, item E 1, G scale Emotional lability, item E 2, G scale Emotional function, item E 3, G scale Stress tolerance (inability to increase tempo, item I 8, G scale) Disturbance of consciousness (confusion + clinical observation) Abstract thinking (similarities) Dyscalculia (arithmetic test) Sensoric aphasia (ADAS, item 10) Episodic memory (Fulds object memory test, total number recall, 5 trials) Attention (Trailmaking test part A) Capacity of simultaneous thinking (Trailmaking test part B) Motor aphasia (ADAS, item 9) Tactile agnosia (Fulds object memory test) Prosopagnosia (face identity) Visuospatial inability (Kohs block test) Executive functions (ADAS) MMSE Intellectual functions (G scale) 0 365 Patient 2 0 365 Patient 3 0 365 Patient 4 0 365 Patient 5 0 365

4 4 1 0 0 2 1 10 10 0 18 62 s 210 1 10 44 3 0 22 17

3 (+) 2 (+) 0 (+) 0 (u) 0 (u) 2 (u) 0 (+) 20 (+) 8 () 0 (u) 23 (+) 53 s (+) 235 s () 0 (+) 10 (u) 57 (+) 12 (+) 0 (u) 27 (+) 10 (+)

4 2 0 0 0 2 1 2 8 0 18 39 s n.t. 1 10 30 7 0 24 14

4 (u) 2 (u) 0 (u) 0 (u) 0 (u) 2 (u) 1 (u) 6 (+) 10 (+) 1 () 20 (+) 65 s () n.t. 1 (u) 10 (u) 33 (+) 8 (+) 0 (u) 24 (u) 13 (+)

4 1 0 0 0 3 1 8 9 0 13 65 s n.t. 2 10 34 0 2 23 19

3 (+) 2 () 0 (u) 0 (u) 0 (u) 1 (+) 1 (u) 8 (u) 5 () 0 (+) 13 (+) 95 s () n.t. 3 () 10 (u) 35 (+) 0 (u) 3 () 18 () 11 (+)

2 2 0 0 0 3 1 19 8 0 26 n.t. n.t. 1 10 31 0 2 23 15

2 (u) 2 (u) 0 (u) 0 (u) 0 (u) 2 (+) 1 (u) 20 (+) 7 () 0 (u) 35 (+) 305 s (+) n.t. 1 (u) 10 (u) 48 (+) 0 (u) 3 () 23 (u) 9 (+)

3 2 1 0 0 2 0 9 5 0 22 45 s n.t. 0 10 44 9 2 24 17

3 (u) 2 (u) 0 (+) 0 (u) 0 (u) 2 (u) 1 () 12 (+) 4 () 0 (u) 20 () 48 s () n.t. 1 () 10 (u) 54 (+) 6 () 0 (+) 24 (u) 9 (+)

Similarities, dyscalculia, Fulds tests and MMSE [7]: increasing values = improvement (+); (ADAS) [7] and G scale [10]: increasing values = impairment (); u = unchanged; n.t. = not testable; I = intellectual scale; E = emotional scale.

Surgical Technique All surgery was performed under general anaesthesia. Two small horseshoe-shaped skin flaps were prepared frontally and corresponding small burrholes were made. Co-ordinates for targets in the ventricles were calculated from plain X-ray films. Common shunt catheters were implanted, connected to Rickham reservoirs and then connected to one another and to a single extension catheter via a Yconnector. This catheter was tunnelled down to a subcutaneous pouch in the left fossa iliaca, where a programmable pump (Synchromed, Medtronic BV, Kerkrade, The Netherlands) was implanted and connected to it. The pump was filled with 17 ml of ganglioside GM1 (60 mg/ml in isotonic phosphate-sodium chloride solution, pH 7.4). The GM1 was prepared from cattle living in areas free from subacute spongiform encephalopathies and was processed as described by Di Martino et al. [11] to give non-infectious ganglioside

preparations from scrapie-infected brain tissue. The GM1 preparation was also shown to contain no foreign genetic material. The dose of GM1, 30 mg/24 h, was selected on the basis of previous experience [7]. When 2 ml of GM1 solution remained in the pump, infusion was interrupted and the remaining solution was sent for bacterial culture and determination of endotoxins. Neuropsychiatric Examination Patients underwent a semistructural neuropsychiatric examination to evaluate disturbances of mood, emotional function, psychotic or paranoid symptoms, symptoms of delirium, symptoms related to the frontal lobe and a neurological examination [12] (table 2).

130

Dement Geriatr Cogn Disord 2002;14:128136

Svennerholm/Brne/Karlsson/Lekman/ Ramstrm/Wikkels

Neuropsychological Assessment and Gottfries-Brne-Steen Scale Rating The patients baseline cognitive functions and treatment response were assessed by the same set of neuropsychological tests as used in the preliminary study [7]. Patients were examined by the same neuropsychologist before noon on 2 consecutive days and examinations were carried out before treatment (day 0) and on days 30, 90, 180, 270 and 365. Table 2 shows the results from days 0 and 365. The Gottfries-Brne-Steen scale [10] was rated by the research nurse before treatment and every second week for the first 3 months and then every month during treatment to examine changes in intellectual and emotional dysfunction and activities of daily living (impairment of motor performance). CT, MRI and Single-Photon Emission to Assess rCBF Patients were examined twice, before and immediately after the treatment for 1 year. Ten normal persons (mean age 61 B 14 years) served as controls for rCBF measurements, which were recorded with d,l-hexamethylpropylene amine oxime (HMPAO) and a threeheaded SPECT camera system (General Electric, Neurocam) equipped with high-resolution collimators. Tc-HMPAO (900 1,000 MBq) was administered intravenously with the patient resting with eyes closed. Reconstruction was done as described by Larsson et al. [13], and the relative quantification of the rCBF was determined by the maximum-minimum method as we had described previously [14]. Clinical Chemical and Neurochemical Laboratory Examinations All examinations were performed 7 days before surgery (day 7), on the day of surgery (day 0) and on days 30, 90, 180, 270 and 365. They included complete haematological status, electrolytes, kidney, liver and thyroid function tests, vitamin B12 and folate levels. Neurochemical tests included measurement of proteins in serum and CSF and of the monoamine metabolites homovanillic acid (HVA), 5hydroxyindoleacetic acid (5-HIAA) and hydroxymethoxyphenylglycol (HMPG), and of somatostatin in CSF as well as CSF cytology [7]. GM1 levels in CSF were measured by the resorcinol method [15] on the same days and also when a disturbance of the GM1 infusion was suspected. Antiganglioside IgG and IgM antibodies were determined with an ELISA [16] with the following glycosphingolipid antigens with GM1 epitopes: GM1, GA1, GD1a and GD1b. Stimulation-Activation-Training Programme This programme was developed by identifying areas of function and processes that are relatively unimpaired and maximizing their use, while placing less reliance on areas of greater impairment [17]. Because focused attention was found to be less impaired than divided attention in our patients, we performed all training in a 3room apartment, with a well-equipped kitchen and furnished with objects familiar to the patients in their own homes. To avoid competing demands on attention, the research assistant trained only 2 patients at the same time, and, when working on the computer, patients were alone in the room. Patients received training twice a week for 45 h. The programme gives systematic training with a multidisciplinary approach to the whole life situation and surroundings of the patient. An important part of the training was the preparation of daily meals: patients had to plan nutritionally well-balanced meals, decide which provisions had to be bought, remember routes to the self-service shop, find the articles on the shelves, pay for the articles, prepare the meal and lay the table. Intellectual activities includ-

ing encoding and retrieval were trained by reading short stories and newspapers followed by discussion of the topics, and individual conversation was held with each patient to increase his/her self-esteem. The computer programme included word comprehension, spelling, calculation, planning techniques and writing of daily protocols, memoranda, notes, letters and short stories. The stimulation-activation-training (SAT) programme supplied information to caregivers to help them understand the strategy of the treatment.

Results

Variation of the GM1 Dose All patients received an initial dose of 30 mg of GM1/ 24 h. If the patient appeared to be overactive or showed signs of mild confusion, the level of GM1 in the lumbar CSF was measured. Our preliminary study [7] had shown that patients with GM1 levels in CSF over 60,000 nmol/l reacted with overactivity and mild confusion but that when the dose was lowered to 20 mg/24 h all symptoms disappeared. Patient 2 showed overactivity after only 2 months of treatment and her CSF GM1 level was over 50,000 nmol/l (fig. 1). Since this patient had mild depression when she entered the study, which disappeared after treatment with GM1, the dose was not lowered until she became confused on day 240. Patient 4 had a level of over 57,000 by day 30 and showed signs of confusion by day 50, at which time the dose was lowered to 20 mg/24 h. His wife felt that his power of initiative gradually diminished on this dose; therefore we increased the dose to 25 mg/24 h, and his wife thought that her husbands activity was restored (fig. 1). Ganglioside Level in CSF Before treatment the level of GM1 in ventricular and lumbar CSF ranged from 16 to 32 nmol/l. Treatment with 30 mg of GM1/24 h increased the concentration about 2,000 times in lumbar CSF (fig. 1). The steady-state level of GM1 on this dose varied between 40,000 and 60,000 nmol/l. In patient 1, the pump was not functioning properly for about 50 days, and we estimate that the patient did not receive any GM1 from day 165 to day 210. During this period, GM1 dropped from 40,000 to 280 nmol/l (fig. 1). The concentrations of the other gangliosides of the gangliotetraose series, GD1a, GD1b and GT1b, did not change significantly during the treatment; for example, the proportion of GD1a increased by ! 2%.

GM1 Ganglioside Treatment and Activation Programme

Dement Geriatr Cogn Disord 2002;14:128136

131

Fig. 1. GM1 ganglioside dose (mg/24 h) and GM1 concentration (nmol/l) in lumbar CSF of 5 AD patients.

132

Dement Geriatr Cogn Disord 2002;14:128136

Svennerholm/Brne/Karlsson/Lekman/ Ramstrm/Wikkels

Laboratory Tests for Exclusion of Complications No abnormality was found in blood chemistry, haematology or urinalysis measured every 3 months during the 1-year study. Samples of lumbar CSF were also taken from the patients every 3rd month, except for patient 3, who refused further lumbar punctures due to headache after the previous one. Therefore no CSF was taken on days 180 and 270. CSF cytology, CSF albumin, IgG and IgM were normal in all samples. No patient had increased anti-GM1 or anti-GM1 epitope carrying IgG or IgM reactivity in serum. Bacterial culture and endotoxin determinations of the pump content at each refilling were negative. Transmitter Substances in CSF Prior to treatment, the level of HVA was low in patients 25 but normal in patient 1 (table 3). During treatment, values for HVA increased significantly in patients 1, 2, 3 and 5 by day 90, but showed little change in patient 4. The initial 5-HIAA values were low in patients 3, 4 and 5 and increased during treatment in all patients except patient 4, but less than for HVA. HMPG was normal and did not change with treatment. Somatostatin levels were low in all 5 patients before treatment but increased rapidly so that there were significant increases by day 30. Since, as already described, the micropump for patient 1 was not working for several weeks, we were able to examine the effect of SAT training without ganglioside treatment at this time. We found that a CSF sample taken about 25 days after the pump was not working had decreased HVA levels and even more marked decreases in somatostatin. The levels increased when GM1 was readministered. A further 2 patients who received SAT training twice a week for a year without GM1 treatment did not show changes in monoamine metabolites or somatostatin. Regional Cerebral Blood Flow The 5 patients had significantly lower rCBF in the basal frontal, frontal association and temporoparietal regions as well as white matter parietal and semi-ovale center regions (p ! 0.01). Less significant reductions were recorded in the hippocampus, mesencephalon, gyrus singuli, motor cortex, parietal association cortex and frontal white matter (p ! 0.05). Although the relative blood flow was higher in many regions in 4 of the 5 patients after treatment for 1 year, the changes were not significant.

Table 3. Transmitter substances: levels of CSF monoamine metabo-

lites and somatostatin in GM1-treated AD patients Day 0 Patient 1 HVA 5-HIAA HMPG Somatostatin Patient 2 HVA 5-HIAA HMPG Somatostatin Patient 3 HVA 5-HIAA HMPG Somatostatin Patient 4 HVA 5-HIAA HMPG Somatostatin Patient 5 HVA 5-HIAA HMPG Somatostatin 30 90 180 270 365

338 141 45 30 145 135 36 31 162 60 39 30 155 97 54 24 153 58 33 26

443 173 45 39 133 95 35 39 191 46 22 46 171 84 37 48 101 66 57 37

485 172 39 53 172 91 26 45 287 59 25 56 138 51 39 33 288 93 39 40

294 158 42 9 175 107 32 54

371 163 37 57 191 124 29 40

516 209 35 47 244 244 35 48 349 87 45 45

173 89 57 36 236 85 30 34

168 79 45 33 376 124 31 59

170 72 45 56 279 117 23 51

Values are given in nanomoles per litre for the monoamine metabolites and in picomoles per litre for somatostatin. Mean values (BSD) of monoamine metabolites in lumbar CSF of 114 healthy individuals, 1888 years of age: HVA 253 B 119 nmol/l; 5-HIAA 125 B 54 nmol/l; HMPG 47 B 10 nmol/l [25].

Neurological, Neuropsychiatric and Neuropsychological Assessments None of the patients showed symptoms of emotional disturbance, psychosis or paranoia. Patient 3 had been treated with citaloprame (Cipramil, Lundbeck, Denmark) before she entered the study; treatment was continued during the study and she showed no symptoms of depression. Patients 1 and 4 had mild extrapyramidal symptoms at the start of the study (table 1), which disappeared within the first month, and no other neurological symptoms occurred during the study period. The assessment of various measures in the neuropsychological examination is shown in table 2. Results are given for day 0 and day 365. Scoring of most items was unchanged, but more items had improved than had dete-

GM1 Ganglioside Treatment and Activation Programme

Dement Geriatr Cogn Disord 2002;14:128136

133

riorated on day 365. Intellectual function tested by the Gottfries-Brne-Steen scale, designed especially to rate patients with dementia, improved in all patients. All patients were sensitive to stress, and except for the initial testing, they had to stay in the ward, where it was difficult for them to relax in the presence of many agitated and screaming patients, which had a negative effect on their performance. Testing with the Gottfries-Brne-Steen scale took place in the patients home or in the training department, which may explain, in part, the better results in this test. All patients had increased their capacity for making contact with others and were more spontaneous in their actions, which helped them improve their social functioning. They were also able to integrate their mental functions better, which facilitated their interaction with relatives and accidental contacts with people. The general impression among the staff members was that all 5 patients were slightly improved after 1 year of treatment. Evaluation of the SAT Programme The patients received training twice a week for 45 h. Each day began with planning what the patient would do during the day, and a discussion about subjects in which the patient was particularly interested. This helped the patients to relax and enabled them to focus on the days required activities. One patient was responsible for the daily meal. At the end of the year all patients had learned to choose the menu, make a list of items to buy, walk to the shop and buy them, prepare the meal, lay the table, wash up and put appliances, cutlery, glassware and crockery where they belonged. Patients were trained to compensate for their reduced sense of locality by looking for characteristic markers on the road to the bus stop, supermarket, hospital, etc. They became so successful that patient 4, who had a severe visuospatial deficit, succeeded in going from home to the training department by bus, including a change of bus on the way. Memory was trained by having patients read newspapers and short stories and then select the essential topic. All patients performed better, at the end of the year; they were able, at the end of the day, to write down what they had done during the day. They also learned to write notes and short letters on the computer and increased their reading comprehension by using selected computer programs. The speech therapist that examined them at 3-month intervals found that their linguistic performance had improved. Physical activity training consisted mainly of walking or riding a stationary bicycle during inclement weather. The physical therapist found that both the patients motivation for and ability to initiate movements were improved.

Functional Assessments by Patients and Caregivers The patients appetites improved during the treatment. Three patients gained weight and the other 2 showed no weight change. The man (patient 4) reported increased hair growth, and one womans hair became curly. All patients reported increased sexual desire. All patients and their husbands/wife felt a significant improvement throughout the study. Patients were more secure in their relations with other persons and showed a new maturity in conversation and in performing various activities. They showed increased initiative and were more selfassured in various forms of activities and training.

Discussion

The primary goal of our treatment of AD patients with GM1 ganglioside was to stop the progression of the disease. According to a variety of measures, we appear to have succeeded. Clinical evaluation overall suggested a slight improvement in all 5 patients; no significant impairment of somatic, neurological or psychiatric symptoms was observed. Gait disturbance and extrapyramidal or extrapyramidal-like symptoms that were observed before treatment in patients 1 and 4 disappeared within the first 23 weeks of GM1 administration, and the physiotherapist noted improvements in certain items in the movement protocols (data not shown). Cerebral blood flow increased slightly in 4 of the 5 patients. The finding of a tendency to increased rCBF in the treated patients is important, because untreated AD patients show a continuous diminution of the rCBF with time. The GottfriesBrne-Steen scale, developed to assess patients with dementia, showed improved values in all patients. In his master review of AD, Sjgren [18] drew particular attention to inactivity as an early and general symptom in AD, which might contribute markedly to the patients deterioration. After only 23 weeks of treatment with GM1, our patients showed increased activity; moreover, their motivation for training was enhanced, in particular with respect to working with a word processor, which none of them had used before. It was remarkable how quickly they learned this skill. Planning and performing daily household tasks was also improved and was an important part of the training as patients gained selfrespect as their ability to perform these acts increased. With their enhanced self-confidence they were able to perform other activities and duties that they would not have dared to attempt before they started training.

134

Dement Geriatr Cogn Disord 2002;14:128136

Svennerholm/Brne/Karlsson/Lekman/ Ramstrm/Wikkels

The study also showed that ganglioside GM1 is a safe drug. In our first attempt to give GM1 to AD patients [6], 17 patients received 100 mg of GM1 subcutaneously or intramuscularly for 3 months, and no patient showed any adverse reaction. Antiglycolipid antibody titres were normal during the period of treatment and 6 months after treatment had been stopped. In our recent studies [7], and the present one, 10 patients have received GM1, 20 30 mg/24 h, for at least 12 months without any signs of inflammation or toxic reactions at any of the monthly assessments; lumbar CSF showed no increase in cell counts and protein concentrations, and electrophoretic profiles were normal. Consistent with our previous study [7], it was of interest that the therapeutic window for intraventricular administration of GM1 was extremely narrow, between 20 and 30 mg/24 h. Doses above 30 mg/ 24 h resulted in overactivity and some confusion among patients, whereas doses below 20 mg reduced activity. GM1 has been used previously in large-scale studies of stroke, Parkinsons disease and spinal cord injuries [19]; doses of 100300 mg of GM1/24 h were given intravenously, subcutaneously or intramuscularly. Results were not encouraging, and our present findings suggest that therapeutic levels were not reached in the brain or spinal cord. In many hundred animal studies worldwide, ganglioside doses of 2550 mg/kg body weight/24 h were required before efficacy was achieved, doses that are 10 40 times higher than those used in human studies. This translates to doses of 14 g daily if given to patients by traditional parenteral routes. When patients with Parkinsons disease were treated with intravenous doses of 2 g of GM1 3 days a week for 8 weeks, they all developed hyperlipidaemia [20]. Therefore we think that the intracerebroventricular administration of GM1 described here is the method of choice for treatment of brain disorders: our two studies [7] and the present one have shown that the method is safe, the surgery is simple, the catheters and micropump are well tolerated and, of greatest importance, a therapeutic effect is obtained. In a large number of studies, gangliosides have been shown to have neuritogenic and neurotrophic activity [5]. We have no evidence of neuritogenic effects in our present study, i.e. formation of new synapses that might have restored lost functions. Rather, the arrest of progressive deterioration in our AD patients suggests that GM1 is exerting a neurotrophic effect. Moreover, consistently with our previous study [7], GM1 also promotes the turnover of transmitter substances. Levels of HVA, 5-HIAA and somatostatin increased in the CSF of our patients after treatment with GM1. The particularly marked effect

on HVA is in agreement with a large number of studies in rodents in which GM1 restored or improved the function of dopaminergic neurons after mechanical or biochemical lesions [19]. Notably, the effect of GM1 on dopamine turnover is long lasting: table 3 shows that the level of HVA in CSF continued to increase throughout the study and reached its highest level at the end of the period. A marked increase in somatostatin was also observed, suggesting that it might be a suitable marker with which to follow the biological effects of drugs in AD. In speculating about the mechanism of action of GM1, it is of interest that when we developed the partition method for separation of gangliosides from other lipids [21] we had observed the strong binding of CA2+ to gangliosides. Several studies have confirmed this strong binding and also that the apparent association constant depends on the surrounding ionic strength [22]. Svennerholm [23] suggested that Ca2+-associated gangliosides on the outer surface of the presynaptic membranes might be replaced by Na+ as counter-ion by the action potential; this would facilitate transformation of the membrane lipid bilayer into a micellar state necessary for the fusion of the synaptic vesicle and the plasma membrane and release of transmitter. The Ca2+ replaced by Na+ as counterion to gangliosides would then participate in the influx of Ca2+ into the synapse. We suggest that treatment with GM1 would lead to insertion of GM1 molecules in the neuronal membrane. The new GM1 molecules would bind Ca2+ as counter-ion and increase the inflow of Ca2+ during synaptic transmission and the release of transmitter. The combination of GM1 treatment and systematic training has been important. Both patients and their spouses observed that the ganglioside treatment increased the patients activity and that the training programme channelled this increased activity toward rational functioning. The patients were taught to make a schedule for their daily activities and gained considerable satisfaction from being able to do their housework once more. Using the word processor permitted them to write protocols, notes and letters, which increased their self-esteem. Of particular note, the use of the word processor required the learning of a new skill. Reading and discussion of articles in newspapers and magazines increased their reading comprehension and feel for language and enhanced their self-assurance as well as improving their safety. They regained their joie de vivre. Patients and spouses all expressed their gratitude for being allowed to participate and wished to continue for a further year, as they felt that their safety and self-assurance could be further improved.

GM1 Ganglioside Treatment and Activation Programme

Dement Geriatr Cogn Disord 2002;14:128136

135

However, we were unable to raise the necessary grants to continue the training. Our treatment schedule has been valuable for patients with AD: the disappearance of Parkinson-like symptoms in a few weeks, the significant increase in dopamine turnover and the beneficial effect of parenteral administration of GM1 to patients with Parkinsons disease [24] suggest that our treatment method will be particularly useful for patients with Parkinsons disease. It is our sincere hope that other research groups will take note of these findings and extend our observations. In conclusion, treatment with ganglioside appears to have stopped the progression of the continuous deterioration of nerve processes in AD and increased the turnover of the transmitter substances examined. The drug increased the patients activity and motivation for training.

Acknowledgements
We thank the participating patients and their families, Dr. L.-E. Augustinsson for the neurosurgical implantation of the catheters, Dr. P. Fredman for the CSF protein and serum antibody determinations, Dr. R. Ekman for somatostatin determinations and Fidia Research Laboratories, Abano Terme, Italy, for supplying GM1 ganglioside. We also owe a special debt of gratitude to Dr. Pamela Talalay, Department of Neurology, Johns Hopkins School of Medicine, Baltimore, for her dedication and invaluable assistance in the preparation of the manuscript. The study was supported by project grants from Health Care and Medical Service, National Board of Welfare, the Royal Society of Arts and Sciences in Gteborg, Old Servants Fund, the County of Bohusln and the Town of Mlndal.

References
1 Terry RD, Masliah E, Hansen LA: Structural basis of the cognitive alterations in Alzheimer disease; in Terry RD, Katzman R, Bick KL (eds): Alzheimer Disease. New York, Raven Press, 1994, pp 179196. 2 Hansson H-A, Holmgren J, Svennerholm L: Ultrastructural localization of cell membrane GM1 ganglioside by cholera toxin. Proc Natl Acad Sci USA 1977;74:37823786. 3 Suzuki K, Katzman R, Korey SR: Chemical studies on Alzheimers disease. J Neuropathol Exp Neurol 1965;24:211214. 4 Svennerholm L, Gottfries C-G: Membrane lipids, selectively diminished in Alzheimers brain suggest synapse loss as a primary event in early onset form (type I) and demyelination in late onset form (type II). J Neurochem 1994; 62:10391047. 5 Leden RW: Biology of gangliosides: Neuritogenic and neuronotrophic properties. J Neurosci Res 1984;12:147159. 6 Svennerholm L, Gottfries C-G, Blennow K, Fredman P, Karlsson I, Mnson J-E, Toffano G, Wallin A: Parenteral administration of GM1 ganglioside to presenile Alzheimers patients. Acta Neurol Scand 1990;81:4553. 7 Augustinsson L-E, Blennow K, Blomstrand C, Brne G, Ekman R, Fredman P, Karlsson I, Kihlgren M, Lehman W, Lekman A, Mnsson J-E, Ramstrm I, Wallin A, Wikkels C, Gottfries C-G, Svennerholm L: Intracerebroventricular administration of GM1 ganglioside to presenile Alzheimer patients. Dementia Geriatr Cogn Disord 1997;8:2633. 8 Wallin A, Blennow K, Scheltens P: Research criteria for the clinical diagnosis of pure Alzheimers disease. Drugs Today 1994;4:265 273. 9 McKhann G, Drachman D, Folstein M, Katzman R, Price D, Stalan EM: Clinical diagnosis of Alzheimers disease: Report of the NINCDS ADRDA work group under the auspices of Department of Health, and Human Services Task Force on Alzheimers disease. Neurology 1984;34:939944. 10 Gottfries C-G, Brne G, Gullberg B, Steen G: A new rating scale for dementia syndromes. Arch Gerontol Geriatr 1982;1:311330. 11 Di Martino A, Safar J, Creoni M, Gibbs CJ Jr: Purification of non-infectious ganglioside preparations from scrapie-infected brain tissue. Arch Virol 1992;124:111121. 12 Wallin A, Edman , Blennow K, Gottfries CG, Karlsson I, Regland B, Sjgren M: Stepwise Comparative Status Analysis (STEP): A tool for identification of regional brain syndromes in dementia. J Geriatr Psychiatr Neurol 1996; 9:185199. 13 Larsson A, Bergh A-C, Bilting M, rlig , Jacobsson L, Stephensen H, Wikkels C: Regional cerebral blood flow in normal pressure hydrocephalus: Diagnostic and prognostic aspects. Eur J Nucl Med 1994;21:118123. 14 rlig , Larsson A, Bergh A-C, Jacobsson L, Wikkels C: A new method for the relative quantification of rCBF examined by 99TcmHMPAO-SPECT. Nucl Med Commun 1994; 15:814823. 15 Svennerholm L: Quantitative estimation of sialic acids. II. A colorimetric resorcinol-hydrochloric acid method. Biochim Biophys Acta 1957;24:604611. 16 Fredman P, Lycke J, Andersen O, Vrethem M, Ernerudh J, Svennerholm L: Peripheral neuropathy associated with monoclonal IgM antibody to glycolipids with a terminal glucoronyl3-sulfate epitope. J Neurol 1993;240:381387. 17 Richards M: Neurobiological treatment of Alzheimers disease; in Morris RG (ed): The Cognitive Neuropsychology of Alzheimer Type Dementia. Oxford, Oxford University Press, 1996, pp 327342. 18 Sjgren H: Clinical analysis of morbus Alzheimer and morbus Pick. Acta Psychiatr Neurol Scand Suppl 1952;82:67115. 19 Svennerholm L: Gangliosides: A new therapeutic agent against stroke and Alzheimers disease. Life Sci 1994;55:21252134. 20 Roberts JW, Hoeg JM, Mouradian MM, Linfante I, Chase TN: Iatrogenic hyperlipidaemia with GM1-ganglioside. Lancet 1993;342:115. 21 Svennerholm L: The quantitative estimation of cerebrosides in nervous tissue. J Neurochem 1956;1:4253. 22 Thomas PD, Brewer G: Gangliosides and synaptic transmission. Biochim Biophys Acta 1990;1031:277289. 23 Svennerholm L: Gangliosides and synaptic transmission. Adv Exp Med Biol 1980;125: 533544. 24 Schneider JS, Roeltgen DP, Mancall EL, Chapas-Crilly J, Rothblat DS, Tatarian GT: Parkinsons disease: Improved function with GM1 ganglioside treatment in a randomized placebo-controlled study. Neurology 1998;50:1630 1636. 25 Blennow K, Wallin A, Gottfries C-G, Karlsson I, Mnsson J-E, Skoog I, Wikkels C, Svennerholm L: Cerebrospinal fluid monoamine metabolites in 114 healthy individuals 1888 years of age. Eur Neuropharmacol 1993;3:55 61.

136

Dement Geriatr Cogn Disord 2002;14:128136

Svennerholm/Brne/Karlsson/Lekman/ Ramstrm/Wikkels