Problems of Research into the Biochemistry of Schizophrenia
by Harold Hillman

Note: This paper has been included to illustrate the problems of medical research; in a world of facile comments and supposedly scientific assertions, it's interesting to see how many very simple problems and pitfalls exist - Rae West.

H. Hillman*
*Present address: Unity Laboratory of Applied Neurobiology, 76 Epsom Road, Guildford GU1 2BX UK

Institute of Biological Psychiatry, Bangor, Gwynedd, LL57 2UW, UK

      Problems of research into the biochemistry of schizophrenia were identified during interviews with 200 schizophrenic patients in the presence of 300 relatives or friends. The former were examined clinically. There is difficulty in establishing the diagnosis, because of the multiplicity and variability of duration of the symptoms. It was difficult to find statistically useful populations, but the ones studied here were (i) Patients who were being treated with neuroleptics; (ii) patients who were not taking neuroleptics; (iii) first degree relatives of patients; (iv) healthy control subjects. Venous blood and skin biopsy samples were taken. The biochemical lesions produced in the brain or blood may be obscured by the effects of the drugs given for treatment, or the life styles of the patients consequent upon the disease. Experiments on healthy animals can indicate the probable effects of drugs, but there is no satisfactory animal model of the disease. The many theories arising out of reported biochemical changes in schizophrenia indicate the complexity of its effects, but none of them have been shown unequivocally to be related to the genesis of the disease, rather than its effects. This paper examines some of the difficulties of research on patients with this disease.

      Despite the enormous amount of research work on the biochemistry of schizophrenia since the Second World War (1-5), there is no general agreement either whether the disease is primarily biochemical, genetic, psychiatric or anatomical, nor -if it has a biochemical aetiology, - what the nature of the biochemical lesion is. In this paper, it was assumed after consideration of a great deal of literature, that the lesion was biochemical, and possible biochemical disorders were sought in the blood from a population of schizophrenic patients. These were patients from all over the United Kingdom, who came accompanied by their friends, to the Schizophrenia Association of Great Britain. It was difficult to do statistically valid studies on a population of volunteers, who were already a selected sample. Nevertheless, we had patients with close relatives and friends, and some unrelated spouses and healthy control subjects; some of the patients were taking a variety of neuroleptics, but some had not taken them for 3 months or more, or had never taken them.
      It was thought useful to compare the biochemistry of the blood of the four groups; to this end, the criteria had to be established for the diagnosis; these studies are being submitted for publication elsewhere. However, during the course of this research, several of the difficulties of research into the biochemistry of schizophrenia became evident, and these are the subject of this paper.

Patients and procedures
      Between 1990 and 1993, about 200 patients, 300 relatives and 50 healthy 'control' subjects came to the Schizophrenia Association of Great Britain in Bangor, Wales, to have venous blood samples taken. The patients were interviewed in the presence of their relatives. The patients only were examined clinically. Skin biopsies were taken from those patients, the relatives and the control subjects, who gave their informed and written consent beforehand.
      Advertisements were placed in national and local newspapers inviting schizophrenic patients, their blood relatives, their spouses and healthy subjects to come to Bangor. They fasted and only drank water from 20.00 the evening before, until 8.00 the next morning, when the blood samples were taken. The patients, relatives and healthy subjects gave written consent for the blood sampling, and signed that they did not have, or had not recently had, bacterial or viral diseases; they were also asked about any drugs, vitamins or mineral supplements they were taking. Samples of blood (20-25 ml) were drawn from the antecubital veins, and, if necessary, cooled immediately to 0øC. They were placed in tubes for whole blood, plasma or serum samples, to examine catecholamines, vitamins, antibodies to milk and grain, trace elements and enzymes. The skin biopsies were used to grow fibroblasts and keratinocytes in culture.
      All subjects gave urine specimens which were examined with 'Multistix' reagent sticks (Ames Ltd), to detect the presence of urobilinogen, blood, bilirubin, ketone, glucose and protein, and to assess the pH and specific gravity of the specimens. The patients were then examined clinically to see if they had any detectable concurrent illnesses, and they and their general practitioners (if the patients wished it) were informed immediately, if any signs of illness, in addition to schizophrenia, were found. No treatment was suggested or given to patients, who were always advised to seek advice from their psychiatrists and general practitioners.
      Some of the patients were taking neuroleptics, others had never taken them at all or had not during the three months prior to their visits to Bangor. The chemistry of the blood of these two groups was compared with that of the blood relatives; a fourth group acting as control subjects consisted of healthy spouses, partners and healthy students.

The population of patients under study
      Patients, relatives and healthy control subjects volunteered for the research, and so did not constitute random samples of each of the populations. This diminished the value of any comparisons between them from a statistical viewpoint. In the present study, the blood from the patients was compared with that from healthy first degree relatives, who were genetically close to them, in the hope that this would take account of differences found between the two populations due to their similarity in genetic contribution. Unfortunately, most of the relatives were parents, who were, of course, one generation older than the patients. Schizophrenia usually starts during the first two decades of life, so that one should really only compare chemical components of the blood, which are known not to change significantly with middle age. Alternatively, research workers should demonstrate that the parameters they were measuring did not themselves change with time.
      Patients who came to the Schizophrenia Association in Bangor from all parts of Britain, were well enough to co-operate with the research workers. This raised an important question. If schizophrenia was a biochemical disease, would one expect to find the lesion in patients who were well between relapses?

Taking a history from the patient
      When patients came by themselves, the information was often very sparse compared with that imparted when they were interviewed in the presence of members of their families or friends. Often previous doctors had asked the same questions 10-20 times before, and there was a certain amount of learning and wishing to please or irritate the interviewers. Sometimes, the patients could not or did not wish to remember information such as the number of times they had stayed in hospital, how many of their admissions were voluntary, the doses of therapeutic or addictive drugs that they were taking, the number of contacts with the police they had had, their weekly cigarette and alcohol consumption. Often, even alert healthy people not under the influence of neuroleptics, found it difficult to be accurate about their habits.
      Patients sometimes said that they had been admitted to hospital 'voluntarily', because they were told that they would be 'sectioned' if they did not. It was difficult to decide whether these admissions were voluntary or not.
      They gave misleading information, either because they were under the influence of disease, or they were taking drugs which impaired their memories.

Patient compliance
      The problems of patient compliance with drug prescriptions have been well documented over a number of years (5-8). With psychiatric patients, these were even more difficult. It was impossible to know for certain whether patients had been taking the drugs prescribed, unless they were in hospital, or were given injections by medical personnel. The patients sometimes did not take neuroleptics because of the side effects, such as inability to concentrate, unnecessary movements, increase in weight, dryness of the mouth, akathisia and impotence. They often felt that doctors did not have enough time to interview then in depth -especially as out-patients - and did not listen to them, or were not particularly concerned about side effects, which were distressing to the patients. They felt sometimes that the psychiatrists and general practitioners were too busy to listen to meaningful accounts of their current states. The doctors were therefore tempted to prescribe drugs rather than give them the supportive psychotherapy and attention they sought.

The diagnosis
      The criteria of the American Psychiatric Association's Diagnostic and Statistical Manual III were used (3) in preference to the 1987 DSM III R criteria(9), in view of the comments of Fenton et al.(10). DSM III consisted of 10 symptoms summarised in Table 1.

Table 1. Simplified summary of symptoms of schizophrenia according to DSM III, American Psychiatric Association (3). One or more of these symptoms should be present during the active phase of the disease to make a diagnosis. Symptoms are often accompanied by difficult personal relationships with family, friends and colleagues.
1. Thoughts of being controlled, or actions inserted into their minds from outside.
2. Feelings that patients' thoughts are being broadcast to the world.
3. Thoughts of other people (other than God) inserted into the patients mind.
4. Feeling that thoughts are being removed from the patients mind.
5. Absurd, fantastic, implausible or bizarre delusions.
6. Grandiose, religious, nihilistic delusions.
7. Delusions with hallucinations.
8. Auditory hallucinations with one or more voice.
9. Auditory hallucinations on several occasions not connected with depression or elation.
10. Incoherence, loosening of associations, marked illogicality, poverty of content of speech, accompanied by inappropriate affect, delusions, hallucinations or catatonia.

One of the difficulties of this, or any, classification were that some degree of subjective judgement was necessary to decide how much weight one gave to each symptom. An important limitation was the accuracy with which the psychiatrist could judge in a 10-30 minute interview how many of the symptoms the patient showed; we normally saw the patients for 2-3 hours. The patients were asked if they had previously been diagnosed as schizophrenic and one could not know whether the previous diagnosis was made on the basis of the same symptoms. Furthermore, one was influenced when one made a diagnosis by knowledge of whether the neuroleptics were effective for the particular patient; many of the neuroleptic drugs are also tranquillisers, sedatives, anxiolytics, anti-psychotics, etc. However hard one tries, it is difficult to make an entirely fresh diagnosis.

Choice of system to be studied
      It is generally believed that schizophrenia is an illness of the brain, although if the primary lesion were in the liver, the autonomic nervous system or the alimentary tract, this could produce secondary effects in the brain. This problem will be resolved if a biochemical lesion is detected in schizophrenia. If this were found, one would then investigate whether other organs or systems were affected.

(i) non invasive techniques
      Examination of the brain by non-invasive techniques of x-rays, air encephalography, magnetic resonance imaging, and positron emission tomography, have given some indication of abnormalities in the left temporal lobe of schizophrenics, possibly different between the two sexes (11-14). Such findings in a number of institutes need to be shown unequivocally to be specific for schizophrenia. Subsequently the abnormal parts of the brain should be examined in detail, in post mortem specimens, by anatomical, biochemical and pathological techniques, since one would expect that lesions due to schizophrenia would be concentrated there.

(ii) biopsy
      When a lesion is localised to a particular part of the brain, it would be possible theoretically to take a biopsy from that part. However, it would be unlikely to be confined to one part of the brain only. Even if the localisation were precise and certain, it is doubtful whether it would be ethically acceptable to take biopsies of the brain, unless they would be required for a treatment, which would result in a certain cure.

(iii) cerebrospinal fluid
      Since this bathes the whole central nervous system, it has been used to take samples to study catecholamines, endorphins, and aminoacids (15-18), but, until now no certain biochemical findings have been made, which would justify such an uncomfortable procedure, which also requires co-operation of the patient. It is also not certain whether it is ethically justifiable in respect of treatment of the patient, who may not be able to give informed consent.

(iv) post mortem material
      The next best tissues are brain and spinal cord taken post mortem from patients, who have died during an acute relapse of schizophrenia (19-21). Few do, so most post mortem studies must be done on brains from patients who died after many years illness. Most of them have been treated with neuroleptics for many years, and often they are undernourished; they may also have taken drugs of addiction and be alcoholics or heavy smokers, so findings are often difficult to interpret unequivocally. One normally obtains post mortem material from four hours to three days after the death of the patient. During this time, the biochemistry of the tissues changes due to hypoxia, agonal bacteraemia, autolysis and bacterial action. Research workers hope that these changes do not obscure any differences between samples of brain from schizophrenic patients and healthy subjects.
      It is difficult to decide whether any abnormalities found in schizophrenia - as in cancer, multiple sclerosis or manic depression - are the cause of the lesions or the results of them. Research workers usually want to know the former. It is also difficult to distinguish between the specific causes and effects of a particular disease, and the more general effects of the life style resulting from the disease. For example, a patient with 'flat' schizophrenia is often inactive, while someone in a hyperactive phase may indulge in energetic physical tasks; both of these states would affect the biochemistry of the autonomic nervous system, the dopaminergic system and the muscles, in opposite senses.

(v) blood, plasma or serum
      Most studies of schizophrenia have involved taking single venous samples from the antecubital vein. Obviously, the blood sample then represented the biochemistry at the instant at which the sample was taken, influenced by the disease, and any drugs whose action or excretion were slow. Therefore, we thought it necessary to create a simple classification of the severity of the disease at the time the patients were examined. Although the classification was arbitrary, it was intended to be one that could be judged simply, and agreed between different research workers. The grades were as follows: 1, very ill, usually in hospital; 2, ill, but at home, not representing a danger to themselves or others; 3, not severely ill, but having some symptoms; 4, capable of working, or working - if not retired; 5, working and leading a normal life.
      A fundamental problem associated with the use of blood samples was that the blood flow in the brain is less than 20% of that of the whole body. Consequently any blood which had perfused the whole brain would be diluted at least 5 times by the time it reached the antecubital vein. If the cerebral lesion of schizophrenia were confined to a small region of the brain, such as the temporal lobe or nigrostriatal system, the dilution would be much greater. Therefore, there would have to be a huge change in the local site of the schizophrenic lesion to permit its detection in the peripheral blood. This could be much diminished if one could take carotid artery and jugular vein samples of blood, but this would not be ethically acceptable.
      The platelets have been examined in schizophrenia particularly in relation to their monoamine metabolism (22-24), but it is not certain that they have not changed during isolation, nor how relevant they are to brain biochemistry.
      The care and accuracy with which blood samples are taken and measured can obviously affect the results. The following variables have to be taken into account; whether the subjects are apprehensive; whether the patients are sitting up or lying down; what drugs and food supplements they are taking; whether they have eaten recently; what time of day it is; how long the blood samples are stored and at what temperature; how they are separated; how they are pipetted; how they are stored after separation; the accuracy, specificity and appropriate concentration range of the procedures used; the method of calibration; the blanks used; the kind of vessels in which samples are stored; the conditions under which they may be transported long distances; the relevant Health and Safety Regulations for the procedures used.
      The statistics represent a further repertoire of problems (25-27). Often one does not have matched subjects. One may be unable to find enough subjects in one group. Groups which are being compared may consist of grossly different numbers. Often the number in the group is too small to know if the values measured are randomly distributed. When can one exclude particular values, which appear outside all the others? Many of these problems can be avoided by consulting a statistician before embarking on the research project.

(vi) tissue culture
      Explants of tissue are generally small single pieces, so that one has to passage the tissue through several generations to obtain enough material to measure any biochemical parameters in it. Cultures may be used to study growth characteristics, microscopy and histology of the tissue. Lymphocytes, skin fibroblasts and keratinocytes, can be grown in tissue cultures (28-30). Of course, the longer a tissue is grown and the more passages the cells go through, the more its biochemistry reflects that of the culture medium. The hope of detecting changes in tissue culture, which reflect those in the brain depend upon the following :
      (a) there are biochemical changes in the brain, liver, or intestine in schizophrenia;
      (b) such changes are also shared by the lymphocytes, fibroblasts, or keratinocytes;
      (c) such changes survive placing into tissue culture;
      (d) they survive passaging;
      (e) there is enough material to measure the biochemical changes in the tissue culture;
      (f) one assumes that other aspects of the life style of the schizophrenia patients, or the neuroleptics, drugs of addiction, alcohol or cigarettes that they are taking at the time of sampling, are unlikely to survive culture, and thus obscure biochemical differences between the schizophrenic patients and the healthy control subjects.
      It has been claimed that biochemical and morphological differences between cultures from schizophrenic and healthy control patients have been found, and it will be interesting to translate such findings into understanding of the genesis of the disease or its treatment.

(vii) and (viii) sputum and perspiration
      These have not yet been investigated under laboratory conditions to find out if they are useful sources of material to seek biochemical differences between ill and healthy subjects. This will depend upon the quantity of material, its variability when taken in standard conditions and how many cells there are in it. These aspects would be worthwhile exploring in view of the potential value of these tissues which can be obtained without discomfort or invasion of the subjects.

(ix) animals, particularly monkeys (31).
      It is usually unwise to use animals as models for human illness, unless one can produce precisely the same lesion in them, such as septicaemia or tuberculosis. Administering high doses of these drugs to rats, cats and monkeys has yielded useful findings in respect of the biochemical dependence of these drugs, probably because one is indeed looking at the same biochemical lesion in these mammals as occurs in human beings. However, there is no satisfactory animal model of schizophrenia, therefore this is unlikely to be of use for this purpose. Although animals exhibit wide spread species variation, it is reasonable to suppose that the mechanisms of drug action are similar in animals to what they are in patients. The more difficult question is whether the drugs also act similarly on healthy subjects and schizophrenic patients;

(x) and (xi) hair and nails
      A large number of trace elements and enzymes have been measured in these tissues, but it is unlikely that these measurements are useful, because the chemistry of the tissues depends upon the following: (a) how recently they have been washed, dyed, bleached or conditioned, and with which soaps, detergents and shampoos;
      (b) whether the hair is cut close to the roots or near the ends;
      (c) the age of the subject;
      (d) the relationship which the concentration of any substances in the hair or nails bears to that in the blood, which is not known;
      (e) the normal range of concentrations of the trace elements in tissue from healthy people.

(xii) urine
      In hospitals, one can collect complete 24 hour urine samples, and measure a range of catecholamine metabolites, ions, peptides, steroids, etc., since these substances are excreted by this route. One must obviously beware lest these substances are degraded during storage or by bacteria. Examination of urine has indicated a large number of anomalies in the urine of schizophrenics, but most of these have later yielded to other explanations (2,32-34). However, the reports of Reichelt and his collaborators of finding novel peptides in the urine of schizophrenic and autistic patients is of great potential interest (35).

      An indication of the difficulties of research into the genesis of schizophrenia can be appreciated by examining a list of most of the findings and theories, which have been proposed over the last half century on the basis of experiments carried out in good faith (Table 2). Many of the problems of research into schizophrenia are shared with research into all mental diseases and many other clinical conditions.

Table 2. Some of the systems believed to be involved in schizophrenia, the lesions found and the substances believed to change. This list is not exhaustive, and has been drawn up mainly from reviews (1-3, 5, 32, 36, 37).
A. It is a purely mental condition
B. Systems involved
      Temporal or frontal cortex
      Limbic system
      Immune system
      Autonomic nervous system
      Extrapyramidal system
C. Chemicals in addition to latter
      Essential free fatty acids
      3, 4 dimethyloxyphenylethylamine
      Folic acid
      Ascorbic acid
      Creatine fructo-kinase
      Antibodies to milk and grain
D. Physiological and psychological findings
      Electroencephalographic changes
      Electrocardiographic changes
      Evoked potentials
      Abnormal sleep patterns
      Abnormal eye movements
      Cognitive deficits
E. Other causes suggested
      Psychological or physical abuse
      Neonatal asphyxia
      Drugs of addiction
      Viral diseases


HTML, scanning Rae West. First uploaded 98-03-01 revd 98-06-22 © H Hillman. All rights reserved.