A Biochemical Approach to Depressive Illness

We also have a fourth possibility, namely that the changes in metabolism may be taking place only in the brain although the same metabolic process is to be found also in other tissues. If this situation exists it may be impossible to infer changes in cerebral tissue from changes in metabolites in blood and urine. Studies of cerebral metabolism in depressive illness are necessary in each of the above situations. In 1, 2 and 3 above they are necessary to justify the assumptions relating general and cerebral metabolism, whilst in case 4 the information is not obtainable by any other technique. Copyright Royal Medical Society. All rights reserved. The copyright is retained by the author and the Royal Medical Society, except where explicitly otherwise stated. Scans have been produced by the Digital Imaging Unit at Edinburgh University Library. Res Medica is supported by the University of Edinburgh’s Journal Hosting Service: http://journals.ed.ac.uk ISSN: 2051-7580 (Online) ISSN: 0482-3206 (Print) Res Medica is published by the Royal Medical Society, 5/5 Bristo Square, Edinburgh, EH8 9AL Res Medica, Winter 1967-68, 6(1): 21-25 doi: 10.2218/resmedica.v6i1.830 A BIOCHEMICAL APPROACH TO

W e also have a fourth possibility, namely that the changes in metabolism may be taking place only in the brain although the same metabolic process is to be found also in other tissues. If this situation exists it may be im possible to infer changes in cerebral tissue from changes in metabolites in blood and urine.
Studies of cerebral metabolism in depressive illness are necessary in each of the above situ ations. In 1, 2 and 3 above they are necessary to justify the assumptions relating general and cerebral metabolism, whilst in case 4 the information is not obtainable by any other technique.

TECH N IQ U ES FOR THE ST U D Y OF C E R E B R A L M ET A BO LISM
Advances in the direct biochemical study of diseases affecting the brain have been made using two main types of technique : 1. T h e analysis of brain tissue obtained at post mortem or as biopsy material during neurosurgical operation. This technique has been particularly valuable when the biochemical abnormality involves structural components of neural tissue, e.g. in the cerebral lipoidoses. 2. Th e establishment of animal models of the disease allowing careful control of variables and biochemical studies at varying stages of the disease. This approach has provided valuable evidence, e.g. in the vitamin deficiency states. (Medical Research Council) T h e suggestion that depressive illness may be associated with biochemical or metabolic dis turbance is not new but added impetus has been given to this approach in recent years by the success of physical methods of treatment, e.g. electroconvulsive therapy and the antide pressant drugs. Studies of electrolyte distribu tion, adrenal steroid hormones and the meta bolism of the biogenic amines have occupied most attention, and excellent reviews of changes reported in depression are available by Durell and Schildkraut, and Coppen (1967).
Biochcm ical studies in depression. T he majority of the published studies have been carried out in relation to " whole body" metabolism, using blood and urinary estimation of metabolites. W hilst we would agree that depressive illness is a disease of the whole person we would argue that the most important and significant disturbances involve the higher nervous functions. Thus studies of general body metabolism will be of relevance only if certain special conditions are fulfilled : 1. If it is known that the whole body changes in metabolism are paralleled by cerebral changes; or 2. If the metabolites measured in blood and urine have brain as their only source and and are not produced in any other tissue; or 3. If changes in blood chemistry are found the effects of which arc known from previous studies, e.g. hypoxia or hypoglycaemia.
W h ilst tests of the biochem ical functions of m any organs are in regular clinical use, e.g. liver function tests, renal function tests, the design of studies of biochemical functions of the brain in living man are lim ited by its inaccessibility. Logically it would seem that there are only two practical approaches to the study of the biochem ical functioning of the living brain, the estim ation of arteriovenous differences of metabolites and the estimation of m etabolites released from brain into the cerebrospinal fluid.
In depressive illness the biochem ical approach to the brain is further lim ited for our patients do not die from the disease process, though they may die because of it, e.g. by suicide. In addition we are unable to devise an acceptable animal model of the disease. W e must look, therefore, to studies in th e living patient to provide us with our inform ation. T h e A m in e H ypothesis -a specific hypothesis relating depressive illness to changes in the cerebral m etabolism of the biogenic amines.
A num ber of biogenic amines are found in brain tissue including histam ine, 5-hydroxytryptamine, noradrenaline and dopamine. W h ilst there is a rough parallel between the distribution of the amines in brain in that they tend to be concentrated in central structures and basal ganglia, there are also differences in the distribution of the individual amines which suggest specific relationships to certain physiological systems. Studies of the distribution of the amines have been facilitated by the application of the technique of fluorescence m icroscopy (Fuxe et al (1965) ) by which the aminecontaining cells and fibre tracts can be visualized in histological preparations. T h e suggestion that th e amines function as transmitter substances at central synapses has received support from the. fluorescence microscopy studies. Such studies have suggested a possible role of dopam ine as a transmitter in the extrapyramidal system and changes in dopamine levels in parts of this system arc reported in Parkinsonism.
5-Hydroxytryptam ine is concentrated in parts of the lim bic system, e.g. the hypothalamus, amygdala and hippocam pus, whilst noradrena line is concentrated in hypothalam us and mid brain, possibly in relationship to systems in volved in central autonom ic control.
T h e evidence for a disturbance of amine m etabolism is presented by D urell and Schildkraut (1964) and C o p p en (1967), and is based on the known actions of drugs which either precipitate' or are used to treat depressive illness. In general terms drugs which can precipitate depression in susceptible individuals, e.g. reserpine, reduce the concentrations of 5-H .T. and noradrenaline, whilst the antidepressant drugs either increase brain am ine levels (m onoam ine oxidase inhibitors) or potentiate the effects of the released amines (tricyclic group of drugs). T h u s a general hypothesis can be advanced to relate changes in am ine m eta bolism to depressive illness.

WORKING HYPOTHESIS
" T h at depression will occur whenever the levels of biogenic amines arc reduced at reactive sites in B rain" .
T h is hypothesis makes no attem pt to dis tinguish between the relative im portance of changes in noradrenaline, dopam ine and 5-H .T, metabolism . O ur first attem pts to investigate this problem were, however, limited to investigation of 5-H .T. Later studies and also studies by D encker et al (1966)  T h e results of such a study are shown in T ab le 1. (Ashcroft et al (1966)). T h e y appear to confirm the hypothesis of a disturbance of cerebral 5-H .T. m etabolism in depressive illness, the low levels of s-H IA A in the lum bar cerebrospinal fluid of the depressed patients being consistent with a defect in the release or synthesis of the am ine. Furth er studies revealed' a correlation between the severity of the de- T h e am ino acid tryptophan is actively transported from blood into brain tissue where it is hydroxylated to 5-hydroxytryptophan.
T h e 5-hydvoxytrytophan is then decarboxylated to the am ine 5-hydroxytryptamine which is metabolised to 5-hydroxyincloI -3ylacetie acid (5-H IA A ) after release. T h e 5-H IA A is then cleared into blood directly or via the cerebro spinal fluid. Since 5-H IA A does not pass in the reverse direction from blood to cerebrospinal fluid we argued that the levels of 5 -IIIA A in the fluid would reflect the levels o f the acid in the brain and that this in turn would reflect the turnover of parent am ine 5-hydroxytrypta m ine. T h e identification of the acid in human cerebrospinal fluid gave prelim inary results indicating a difference between the levels in patients w ith depressive illness and a group of nondepressed patients with neurological disease. A shcroft and Sharm an (1960) suggested that this approach m ight be a fruitful one. Su b sequent studies have been directed to further investigation of these problem s and have been of two types: 1. Studies of C S F levels in m an; 2. Anim al studies designed to exam ine the relationship between cerebrospinal fluid pression and 5-H IA A levels and also a return towards normal of the levels w ith remission from the illness irrespective of the type of treatm ent ( E C T or im ipramine). Confirm ation of these results was provided in a study by D encker et al (1966) who also extended the findings by m easuring homovanillic acid, the m etabolite of dopamine. W h ilst they found the levels-of 5-H IA A to be reduced the levels of H V A were w ithin normal lim its in depressed patients.
C loser inspection of the results in T a b le ] reveals an alternative explanation of the low levels of 5-HIAA in depression which must be considered before we can accept them as indicating an alteration of cerebral metabolism. In nondepressed subjects there is a gradient in levels from ventricular fluid to lumbar C SF. Fluid obtained at air encephalography which may be considered as a mixture of ventricular fluid displaced by injected air with C S F from other levels has a concentration of 5-IIIAA part way between that of lumbar and ventricular fluid. These results suggest that the 5-H IAA is added to the ventricular fluid and removed as the fluid passes down the cerebrospinal axis. Thus low levels in lumbar cerebrospinal fluid could represent either less addition at ventricular levels or a greater degree of removal as the C S F passes down to the lumbar region.

ANIMAL EXPERIMENTS
The relationship between the concentration of 5 H IA A in C .S.F . obtained from different levels of the cerebrospinal axis and the meta bolism of the parent amine in brain was thus revealed as a complex one and it seemed un likely that the problems could be resolved in clinical studies. Animal studies were thus utilised and have provided the vital links in the chain which may now allow us to move forward again with studies in man. 1. Relationship between Cerebrospinal fluid and brain levels. Studies were made using a technique allow ing the sampling of C S F from the lateral ventricles and cisterna of the dog and the fol lowing results were obtained.
(1) The levels of homovanillic acid in the lateral ventricular fluid reflects the levels of the acid in the underlying brain tissue, i.e., the caudate nucleus. Levels of 5-HIAA are higher than might be expected if caudate nucleus were the only source of the acid and it suggested that structures in the inferior horn of the lateral ventricle, viz. the amygdala and hippocampus, may contribute to the levels. (Guldberg et al (1966)).

Active transport of acid metabolites from
brain and C S F to blood.
Experimental studies by Pappenheimer et al (1961) and Davson et al (1962) have demonstrated the presence of a transport mechanism capable of actively transporting certain organic acids, e.g., diodrast and para amino hippuric acid from cerebrospinal fluid into blood. This mechanism appears similar to the renal tubular mechanism which transports organic acid, and the suggestion is made that it is a property of the choroid plexus. W e reasoned that the gradient in levels of 5-H IAA noted in our patients between ventricular and lumbar fluid could have resulted from the action of such a transport mechanism and animal experiments were designed to test this hypothesis.
(a) A comparison of ventricular and cisternal C S F levels of 5-H IAA and HV A in the dog revealed a similar gradient to that observed in man.
Pretreatment of the animals with probenieid, a drug which blocks the renal transport mechanism for organic acids, was shown to reduce significantly the gradient levels of 5-IIIA A and H V A between ventricles and cisterna with a marked relative rise in cisternal acid levels. Guldberg et al (1967). Such a result is consistent with the presence of an active transport mechanism for the organic acids situated in the C S F pathway.
(b) A technique was developed for the per fusion of the cerebral ventricles in the conscious dog similar to that used in the goat by Pappenheimer (1961). Such a method (Ashcroft et al (1967)) was used to study the clearance of 5-H IAA and H V A and inulin from the C SF , the results confirming the presence of an active transport mechanism for the acids and localising it in the fourth ventricle. Neff et al (1964) have provided evidence for a similar transport mechanism for 5-HIAA from brain tissue to blood. CONCLUSIONS T he studies described in brief above represent an attempt to develop a technique for the study of the cerebral metabolism of 5-hydroxytryptamine in patients with depressive illness. The progressive evaluation of the technique has required the use of animal experimental procedures to examine the relationship between C S F and brain in addition to studies in man.
Our simple model (Fig. 1) must be amended in light of these studies to show a variation in 5-HIAA levels with the site of C S F sampling and to include the concept of active transport rather than the passive diffusion o f acid m etabolites from brain and C S F into blood.
T h e results of the C S F studies in depressed patients showed lowered levels o f 5-H IA A but norm al H V A levels in lum bar fluid. Such a selective low ering of 5-H IA A levels argues against an increased rate of transport of the acid from C S F as disturbance o f such a m ech anism would be expected to low er concentra tions of both H V A and 5-H IA A . W e m ust conclude, therefore, that the findings appear to indicate a dim inished release or synthesis o f the am ine 5-hydroxytryptam ine in the brain of the depressed patient and would suggest that the m ost likely defect is in the hydroxylation of tryptophan.
T h e results, however, give us only a glim pse o f the problem and the investigations have left us with m ore questions to answer than when we started, e.g. are the sym ptom s of depressive illness causally related to the change in am ine m etabolism ? D o different depressive clinical syndromes have different associated bio chem ical findings? C an the biochem ical find ings be shown to have any predictive value in terms of treatm ent and prognosis?