lithium in drinking water and dementia incidence

Posted comment on ´Association of Lithium in Drinking Water with the Incidence of Dementia` by L.V. Kessing, T.A. Gerds, N.N. Knudsen, L.F. Jorgensen, S.M. Kristiansen, D. Voutchkova, V. Ernsten, J. Schulllehner, B. Hansen, P.K. Andersen and A.K. Ersboll and published in JAMA Psychiatry 2017 74(10) 1005-1010 doi 10.1001/jamapsychiatry.2017.2362


Kessing and colleagues` article addresses the question of whether the incidence of dementia in the general population of Denmark co-varies with long-term exposure to micro-levels of lithium in the drinking water. Their study suggested that higher levels of lithium could be associated with a decreased incidence of dementia.

A nationwide, population-based nested case control study with 733,653 control individuals and 73,731 dementia sufferers was performed in Denmark. Dementia sufferers were aged between 50 and 90 years of age and had received a diagnosis of dementia from 1st January 1970 to 31st December 2013 and had hospital contact either as an in- or  out-patient. Each dementia sufferer was matched by both age (median age 80.3 yrs old) and sex (approx. 60% female and 40% male) to 10 controls. All control subjects had to be alive and have no diagnosis of dementia when their matched subject had been diagnosed. The residence of each study subject was recorded from 1986 and each location was cross-referenced to the level of lithium in the drinking water at that time measured according to municipality between 2000 and 2010. The lithium level was assumed to be stable within the study period.  Data was analysed from 1st January 1995 to 31st December and Kessing and colleagues looked at 4 levels of lithium exposure (2.0uG/L to 5.0; 5.01 to 10; 10.1 to 15; and greater than 15) and recorded the incidence of dementia in their test groups.

Kessing and colleagues found in their study that the level of lithium exposure was lower for patients with a diagnosis of dementia than for the controls (dementia – median 11.5uG/L whereas controls -median 12.2 uG/L). Also, this association between exposure and dementia was found to be non-linear. A comparison of incidence rate ratio (IRR) for dementia for lithium exposure of 2.0 to 5.0uG/L to exposure greater than 15.0uG/L showed a decreased value for the higher exposure (IRR 0.83, 95% CI, 0.81-0.85 P>0.001). However, the ratio value was the same for a comparison of exposures between 10.1 to 15.0 uG/L (IRR 0.98, 95% CI, 0.96 – 1.01; P= 0.17), but the value increased when it was compared to 5.1 to 10.0uG/L (IRR 1.22, CI 95%, 1.19-1.25 P>.001. Similar patterns were obtained for both Alzheimer and vascular dementia sufferers.

Therefore, Kessing and colleagues summarised that their results suggested that increased lithium exposure in drinking water may be associated with a lower incidence of dementia in a non-linear manner. However, they described the association as not being definitive since the nature of their study set-up meant that patterns and links could be identified between factors, ie. lithium exposure and incidence of dementia, but no definite conclusion could be made because other factors may have influenced  their results.


What makes this article interesting is that it looks at the topic of dementia not from cause or treatment, but from protection. This article hints that naturally occurring high levels of lithium in drinking water can have a protective effect from dementia, ie. there is a slightly lower risk of dementia occurring with exposure to high levels of naturally occurring lithium in this case sourced in the drinking water. Others have jumped to suggesting that adding lithium to drinking water may provide some form of protection from developing this disease, but it should be said that Kessing and colleagues indicate that their results are not definitive and the incidence of dementia could be influenced by other unknown factors. Unfortunately, because of the nature of the disease and the fear that it induces, the field of dementia research and non-academic thinking is awash with different hypotheses about causes, treatment and protection and also unfortunately, although we may be gradually explaining its biochemical basis, we are still not far enough forward to finding a ´one-tablet` cure. The problem with dementia is it`s elusiveness – many causes and triggers, many possible culminations of causes and its wide-ranging physiological effects that may be individual in both degree and nature. The value of this article is that it provides another element to studying this disease in the hope that every fragment of information is another piece of the puzzle.

The article commented on here in this blog is about the effect of lithium exposure on dementia. The beneficial health effects of lithium comes about from studies on bipolar disorder which is where sufferers experience periods of mania and periods of depression.  Sufferers of this particular disorder exhibit associated negative cognitive effects with both of the two phases and it often develops later into dementia. Treatment with chronic administration (but not acute) lithium helps the sufferers by stabilising moods so that the periods of mania and depression are reduced and has also been found to delay the onset of dementia if occurring.

The first point about lithium exposure and dementia is that cognitive disorders, especially those not genetically predetermined and where the emotional status and system are involved, seem to appear over a period of time and require treatments that are also needed over a period of time before reduction of symptoms is observed. This delay between administration and effect means that the long-term ´fix` is not reliant on instantaneous changes of biological processes. There may be observable influences on processes after single administrations, but these may not be the same as those occurring after long-term administration, or even be related to symptoms.  The exploration of the effects of lithium exposure on dementia hence requires two approaches – the instantaneous effects observed perhaps in the ´test-tube` although more advisable on the whole body after single exposure and secondly, what is happening at the biological and physiological levels after a longer period with or without further administration of the ligand or sustained administration. This type of exploration mirrors the approaches to research on depression where we know that the beneficial effects of antidepressant administration require 3 weeks to occur whereas certain biochemical changes are instantaneous eg. the immediate effect on neuronal adenyl cyclase activity compared to long-term effects on neuronal connectivity and neurotransmitter receptor number . The therapeutic effect of any drug (or therapy) has to either reverse these changes in order that treatment is deemed successful, or provide alternative means by which correct neuronal functioning may occur. In the case of some treatments this reversal is also temporary and administration must continue. Without sustained administration the biological processes and systems can fall back into ´dysfunctioning status` resulting in the reappearance of symptoms and an example of this is lithium itself and its role in the treatment of bipolar disorder.

If we are to understand what is happening in dementia we have to look at the conditions that cause it and the mechanisms that are subsequently put into play. Kessing and colleagues` article prompts an investigation of the metal, lithium and its supposed effect on the incidence of dementia. We have to assume that the biological mechanisms involved in dementia and associated with cell death and destruction of neuronal pathways are independent of the cause of the disorder, ie. the biological mechanisms with abnormal functioning are the same whether the dementia observed is a result of long-term bipolar disorder or injury for example. In the same vein, as given above treatments can either normalise the dysfunctioning processes or provide alternatives that compensate for those dysfunctions caused by the disease or injury. If we apply this hypothesis to lithium and dementia, the biochemical effects of lithium which lead to the proposed neuroprotective effect can then be grouped according to whether they affect neuronal firing (eg. reducing levels of over-excitation in neuronal firing in certain brain areas as observed in dementia) or by reducing cell apoptosis (ie. reducing the mechanisms employed in cell degradation and cell death, also perceived as important in dementia). Therefore, lithium could be said to promote a reduction in susceptibility to dementia by affecting one or both of these groups of functions.

If we look at the first group of functions relating to lithium exposure that of effects on neuronal cell firing we can see that the action of short term administered lithium acts at many different neuronal sites and cellular functions. Biochemically, neuronal over-excitation results in excessive firing in individual cells or in groups of cells within brain areas at a level which is outside normal expectations for that cell or area. For lithium to have an effect on this mechanism then it must interact with normal firing mechanisms. (Here in this comment we concentrate just on the brain and neuronal systems since we are looking in general at possible mechanisms for cognitive failure, emotional system upset and dementia.) Lithium ions due to their size and electronic charge can act where two other common ions in the brain act. These are sodium ions (a cation like lithium with a single charge) and magnesium ions (also a cation, but with an electronic charge of two, however having the same ionic and hydrated radius as the lithium ion). Since a lithium compound is medically administered then the lithium is already in ionic form and so is already able to accept an electron. Therefore, electronically lithium ions can replace sodium ions in firing mechanisms.  One way in which it can do this is to transfer through the sodium channels of the neuronal cell membrane. The possible result of this action is cell depolarisation and this is observed in brain areas where hyperpolarisation occurs. Researchers have also found that lithium can regulate the expression of different isoforms of sodium channel and therefore, effect on firing through increased presence of sodium channels is possible. Cellular firing levels can also be affected by lithium causing the release of the neurotransmitters, serotonin (5HT) and noradrenaline (NA) which can result in cells that are activated by these molecules actually depolarising. This is important since in some brain areas, 5HT and NA have an inhibitory effect on cell activity, or have an inhibitory effect on cells further down the pathway.  In the absence of these neurotransmitters for whatever reason, the administration of lithium ions could substitute for their loss and firing whether excitatory or inhibitory results. It should be noted however, that under normal firing circumstances, lithium ions do not cause excessive firing and cellular depolarisation.

Whereas lithium ions can directly replace sodium ions in neuronal firing mechanisms, their action relating to magnesium ions is a little more complicated since they disrupt processes important in cellular firing that rely on the involvement of magnesium ions for correct molecular conformational structure.  For example, lithium ions can have an effect on cell firing via its action on sodium potassium ATPase ( Na+K+ATPase) which is magnesium ion sensitive and responsible for ionic gradients across the neuronal cell membrane during firing and another transport system, the mitochondrial sodium-calcium exchanger. The sodium-calcium exchanger is important in the removal of calcium ions from the mitochondrial matrix. The activity of this depends on previous action of the Na+K+ATPase which pumps sodium ions out of the cell and potassium ions in during firing. This action allows the entry of calcium ions into the cell in the presence of high concentrations of sodium ions (ie. at depolarisation). Hence, failure of Na+K+ATPase action (as observed as impaired in bipolar disorder) leads to potassium ion depletion inside the cell and sodium ion accumulation. Therefore, the sodium calcium exchanger begins to pump calcium ions in leading to an increase in the cells hyper-excitability (also observed in bipolar disorder). Lithium ions in this case activate Na+K+ATPase, hence normalising cellular ion concentrations and depolarising the cells where hyper-excitability was previously observed.  It was found that lithium ions at high doses actually replace the magnesium ions in the complex. The presence of the metal ion is important because of its role in the nucleotide phosphorylation stage of the process eg. in the breakdown of ATP (the transport of ions requires energy) in this case. Nucleoside triphosphates require magnesium ions or a manganese complex to be active since the magnesium ions neutralise some of the negative charges present on the physical polyphosphate chains of the molecules, hence reducing non-specific ionic interactions between the enzymes and the polyphosphate groups of the nucleotide.  This link with nucleotides, magnesium function and lithium action as a substitute ion can also be observed with cellular cyclic nucleotides, eg. in the case of adenyl cyclase and the production of cAMP. Here the enzyme interacts indirectly with the magnesium ion through hydrogen bonds to coordinated water molecules. The inhibitory effects or activation effects of lithium ions on adenyl cyclase (cAMP levels are reduced in depression and increased in mania) relates to G protein binding (likely to be inhibited) and hence, this is another example where lithium has an effect dependent on what the normal functioning of that area or system is.

Therefore, we have seen how lithium ions can work at the level of cell firing, but how does it reduce levels of neuronal firing over-excitation and ultimately, levels of excitotoxicity where cells begin to die? Both over-excitation and increased excitotoxicity have already been reported as involved in dementia and occur in areas such as the entorhinal cortex and hippocampus. The end effect can be local cell death and in dementia there appears to be a natural progression of cellular degradation as inactivity and death in one area leads to underactivity and cellular death in another and so on. It should be remembered that only the hippocampus appears to be capable of high levels of neurogenesis (new cell formation functionally linked to memory) and hence, any loss of cells in other areas will have serious effects on firing of the system as a whole.

So, how can lithium ions reduce over-excitation where needed? The different mechanisms influenced by lithium ions will lead to different effects on neuronal cell functioning restoring normality. The administration of lithium ions can lead to neuronal firing in areas which would normally have an inhibitory effect or could force depolarisation in areas which are normally hyperpolarised. Neuronal firing forced by lithium ion presence could be by substitution with sodium ions in membrane-bound sodium channels leading to cellular depolarisation directly or through its action on neurotransmitters 5HT and NA release. Normalisation of the other ion transfer systems required such as Na+K+ATPase and sodium-calcium exchangers can also be induced. The overall effect is that lithium ions substitute for dysfunctional firing systems and cause neuronal activation which leads to inhibition of cells further down the pathway. This knock-on effect on others could remove the higher levels of excitation seen in some conditions that lead to excitotoxicity and cell death. The overall inhibition of hyper-excitable cells may also explain the activation effects of lithium on adenyl cyclase (cAMP levels are reduced in depression and increased in mania) where G protein binding is inhibited and also reports where lithium administration causes an increase in calcium ions through likely activation of the PI3 / Akt pathway as observed with action against alpha-bungarotoxin and muscarinic receptor binding. Therefore, a reduction in firing occurs and since areas work together and firing is interconnected then functioning patterns of firing will also change and long term alterations ensue. Since in the case of lithium ions and bipolar disorder, mood is stabilised on long-term administration the physiological changes that occur after the sustained period of administration of the metal ion result in a likely ´normalising` action of individual cell firing and ultimately, connectivity patterns seen in the disorder.

The second area of functioning where lithium ions may have their effect is on reducing cell apoptosis or death.  Cell death methods can be intrinsic (eg. requires transcription factor effects and subsequent DNA transcription changes) or extrinsic (eg. requires extracellular receptor binding and caspase cascades), but independent of cause (eg. injury, cell membrane signalling detrimental changes, neurotrophic signal changes) they appear essentially to have the same mechanisms. A change in apoptosis because of lithium treatment leads to a change in mitochondrial/ER dysfunction, reduction of negative epigenetic effects, reduction of glial dysfunction, reduction of oxidative stress and inhibition of the enzyme, glycogen synthase kinase- 3 (GSK-3). For example it has been reported that there is a change in IP3 in bipolar disorder which results in a change in calcium ion signalling in offending cells. The rise in intracellular calcium (also reported in bipolar disorder) results in dysfunction of the mitochondria (observed by increased Bcl-2 levels and decreased Bax levels) and increased oxidative stress of the cell (an effect that can be decreased by glutathione administration also observed in bipolar disorder). The increased level of oxidative stress results in cellular apoptosis. Lithium is reported to inhibit GSK-3 which is one of the factors controlling cellular apoptosis and it can either increase activity or decrease it depending on circumstances. This again could be interpreted as lithium ions ´normalising` function in dysfunctioning cells, or having no effect if the cells are functioning normally.

Increasing apoptosis occurs by the disruption of correct mitochondrial functioning and affects the regulation of the expression of the mitochondrial transcription factor, Bcl-2. Decreased apoptosis occurs in the case of lithium administration when GSK-3 inhibits the early phase of the caspase cascade (caspase 8) or by having an active PI3-AKT pathway which leads to a rise in calcium ion release or an increased beta-catenin/wnt pathway. The inhibition of the GSK3 enzyme has wide ranging effects since the enzyme has multiple functions such as phosphorylation of glycogen synthase and ultimately regulation of glucose metabolism, effects on transcription factors such as cJun and cell cycle mediators such as cyclin D. There are also observations that GSK-3 affects proteins bound to microtubules and this could relate to an observed build-up of beta-amyloid protein in the cells – a process which is linked to the observation of dementia.

Therefore, the administration of lithium ions can result in decreased apoptosis or a protective effect  against apoptosis. The latter may be the case if the cell senses that the normal firing function is defective and another unnatural ion has taken the system over as suspected in the case of lithium ions actions on cellular functioning described above. The result is that there is a normalisation of function in the relevant areas which in bipolar disorder may exhibit extensive induced apoptosis. The increase in cell number and functioning cells is supported also by the observation that lithium administration leads to an increase in volume of the hippocampus, an area whose function is observed to be reduced in dementia. Renewing cells is essential particularly for the hippocampus as stated above since neurogenesis here is linked to the cognitive input of new experiences, binding of information and memory. An increase in growth of the hippocampus has been observed with lithium ion administration and this counteracts the area`s shrinkage which has been reported in depression. An increase in volume implies the production of new cells.

Therefore, lithium ions exhibit a wide range of cellular effects, but their action appears to be linked to only those brain areas whose neuronal functioning is abnormal, ie. areas which exhibit over-excitation or are subjected to over-inhibition. Exposure to the lithium ions can ´normalise` firing and also ´switch off` the apoptosis mechanism so that cells can return to normal functioning levels even if not by natural mechanisms. Cells that are functioning normally appear not to be affected and this is supported for example by the action of lithium ions on guanylyl cyclase which is part of the photo-reduction mechanism in humans and important in visual processing.  Guanylyl cyclase activity leads to an increase in cGMP keeping sodium ion channels open whereas light reduces the level of cGMP causing the sodium channels to close and the cell to hyperpolarise. Sufferers of bipolar disorder report reduced visual motion perception, but lithium has no effect on photo-reduction or on ocular functioning even though it is a potent inhibitor of guanylyl cyclase activity. Therefore, just like with neuronal firing, this brings about a discrepancy since the biochemical effects described for lithium ions are independent of cell status and therefore should occur whether the cell is in either an over-excited/over-inhibited state or not and the action should be on all cells and not just those exhibiting excitotoxicity. The action of lithium ions therefore, implies that those cells exhibiting excitotoxicity or lack of inhibition have already different functioning or different physiological characteristics that favour lithium ion action and binding in preference to its action on normal cells. It could mean that the propensity of binding of lithium ions is lower on these abnormally functioning cells than the cells own natural substrates and hence, it only works when these natural substrates are not available. Since the first port of call of the lithium ion is the neuronal cell membrane then it is possible that the over-excited or over-inhibited cell already has a different cell membrane structure or functioning and it is here where the natural substrates are normally favoured, but when absent, then the presence of lithium ions after administration will have an effect. This hypothesis is supported by the biochemical property of magnesium ions which is that they coordinate groups of ions or molecules so that the correct arrangement and correct conformation of the molecule is attained. This could be of importance when looking at lipid rafts and protein conformations as part of the cell membrane physiological structure and functioning. Lithium ions could replace magnesium ions if absent to re-stabilise the neuronal cell membrane lipid raft structure so that normal firing mechanisms can be induced.

Therefore, a look at the biochemical basis of lithium ion action in bipolar disorder and the mechanisms behind its mood stabilising effect gives an indication to the complexity of dementia and to the problem of defining biochemically what causes it and what can be done about it. Lithium ion administration can be described as a ´blanket` treatment affecting many different systems even if the overall affect is a reduction of emotional upsets and a stabilisation of mood. Its possible mode of action is likely to be through normalising cellular firing in areas experiencing over-excitation or lack of firing and this could be caused by imbalances of sodium ions and magnesium ions. It could also act by affecting the cellular apoptosis processes that may be called into play if cell death is ordered due to cell dysfunction or to prevent cell death ordered because of the abnormal lithium ion effects on the neuronal firing mechanisms. Lithium ion exposure also does provide some insight into a possible initial ´limiting` cause for dysfunctional cell firing and this appears to be the cell`s exterior membrane.

Since we`re talking about the topic….

…would it be possible to use radioactive lithium to map areas of neuronal cell action with time using mouse or rat models of disorders such as dementia? Would such studies give an idea of where the ´limiting` brain area for dementia is?

……lithium is said to be an important inhibitor of glycogen synthase kinase 3 which is known to phosphorylate the voltage gated potassium channel type, KCNQ2. Phosphorylation of the channel decreases its activity and ultimately affects the transport of potassium ions important in depolarisation. Therefore, can we assume that lithium ion action ultimately has an effect on KCNQ2  channels` number or function and this can be observed by measuring potassium ion transport and concentration in the neuronal cells using inhibitors or blockers of KCNQ2?

……prolonged lithium ion treatment is said to lead to probably indirect inhibition of PARP-1 (poly-(ADP-ribose) polymerase) by inhibiting 3´5´-phosphoadenosine phosphatase (pAp-phosphotase) function. This causes a progressive accumulation of pAp in the cell that binds to the PARP-1 inhibiting it. However, the hypothesis is under dispute. Can we assume that deletion of pAp phosphatase gene could clarify the observation since in its absence poly -ADP-ribosylation activity inside the cell would be normal and PARP-1 inhibition would not be observed?


This entry was posted in dementia, lithium, neuronal firing, Uncategorized and tagged , , . Bookmark the permalink.