Skip to main content

Towards gene identification and personalized prophylactic medication for lithiopathies in lithium-using patients

Final Report Summary - LITHIOPATHIES (Towards gene identification and personalized prophylactic medication for lithiopathies in lithium-using patients)

Summary of achieving the project's objectives

Introduction. Lithium is widely prescribed and the mainstay treatment for bipolar disorders. Unfortunately, lithium treatment often leads to adverse side effects (lithiopathies) like nephrogenic diabetes insipidus (NDI), metabolic acidosis, hypercalcaemia and chronic kidney disease (CKD). Lithium-induced NDI (Li-NDI) is a disorder in which the kidney cannot concentrate its pro-urine anymore, while a metabolic acidosis stands for the increased acidification of the blood as a result from increased acid production by the body, or a decreased ability of the kidney to excrete acid. Hypercalcaemia is increased blood calcium levels, while CKD stands for a progressive decline in kidney function. As these lithiopathies are largely irreversible, medication for the lithiopathies should be administered with the start of lithium treatment. However, prophylactic administration of these medications to lithium-using patients is undesirable, because not all patients develop lithiopathies.

Aim. To target for lithiopathy-specific personalized prophylactic medication, the aim of this study was to identify the susceptibility genes for development of the lithiopathies using genome wide association studies (GWAS) in mice.

1) Analyze development lithiopathies and perform GWAS studies.
2) Isolate the responsible gene within the cluster and rationally hypothesize the role of the gene in lithiopathy development.
3) Confirming role of susceptibility gene in lithiopathy models.
4) Translation to humans.

Methods. 29 different inbred mouse strains were treated with a control or lithium diet and the lithium-induced side effects were analyzed. For this purpose mice were housed in metabolic cages and phenotype data was collected after 10 days, 1 month and 1 year. Using the phenotype data, GWAS were performed.

Results. To evaluate the development of Li-NDI, urine output and osmolality were analyzed. After 10 days in many strains a significant increase in urine output was noted in the lithium-treated mice compared to the control groups. However there were also strains without a change in urine output after lithium treatment. The urine osmolality ratio between control and lithium-treated mice was strongly related to the urine output ratio (R2=0.93290 P<0.0001). The analysis from day 28 demonstrated similar urine output ratios as compared to day 10, although there were a few strains with a slightly altered response to lithium. Remarkably, after 1 year, there were a few strains in which Li-NDI strongly progressed, while in some other strains urine output was decreased. Strains without Li-NDI at day 10 and 28 did also not show an increased urine output after 1 year of lithium treatment.

To analyze the development of lithium-induced metabolic acidosis, blood pH, bicarbonate and CO2 were analyzed at day 10 and 28. While at both time points most strains showed a tendency for a reduced blood pH, there was only one strain at both time points that demonstrated a significant reduction in blood pH. While blood pCO2 levels were not significantly altered for any strain, blood bicarbonate concentration was decreased with a few strains, including one that demonstrated a reduced blood pH. These results demonstrated that there were two strains with a lithium-induced metabolic acidosis. To determine whether lithium affects renal proton secretion, we also measured urine pH and corrected this for urine output in order to determine free proton secretion. At both day 10 and 28 most strains demonstrated an increased proton excretion after lithium treatment, however only at day 27-28 this reached significance. The coinciding decrease in blood pH and an elevated free proton excretion suggests that the lithium-induced metabolic acidosis is not due to a decreased renal excretion of acid, but an increased body production of acid.

In addition to Li-NDI and metabolic acidosis, we also analyzed the development of hypercalcaemia, by measuring the blood concentration of ionized calcium (iCa2+). At day 10 there was a lithium-induced increase of (iCa2+) in four strains, while in most other strains lithium seemed to increase (iCa2+), however this was not significant. A similar tendency was observed at 28 days, however no significant differences were observed.

Finally, the development of lithium-induced CKD was analyzed after 1 year of lithium treatment, as in humans this disease only occurs after many years of lithium treatment. For this purpose, kidneys were isolated and the presence of interstitial fibrosis, a hallmark of Li-CKD was analyzed together with other signs of Li-CKD, including tubular atrophy and glomerulosclerosis. Although the majority of strains seemed to have an increased presence of both interstitial fibrosis, tubular atrophy and glomerulosclerosis, these differences were not significant. Remarkably, the non-significant differences of lithium on the presence of interstitial fibrosis, tubular atrophy and glomerulosclerosis were highly correlated among strains. In strains that demonstrated an increased tendency for interstitial fibrosis, a similar tendency for tubular atrophy and, to a lesser extent, glomerulosclerosis was found. Finally, in these strains we also analyzed other markers for renal damage, including blood urea nitrogen (BUN) and beta-2 microglobulin (B2M). While BUN levels were not significantly altered in these strains, B2M levels were significantly increased in strains with an increased tendency for interstitial fibrosis and tubular atrophy, while strains without this tendency did not show this increase.

Having analyzed all side effects of lithium treatment we found a large, significantly different variability among strains in the development in Li-NDI. As variability among the 29 strains was very limited for the other side effects, GWAS were only performed on the NDI data. Using data from all time points, GWAS identified various susceptibility genes for the development of lithium-induced NDI. Highly significant associations with NDI development were found in 23 regions, which contained 1-7 genes. Using literature and the NIH proteome databases of native inner medullary collecting duct cells and cortical collecting duct cells, the renal expression of these genes was assessed. In addition, we performed qRT-PCR to investigate the expression of several of our candidates along the renal tubule. Based on the strength of association, their expression in the collecting duct and their potential role in Li-NDI, we developed a first list of genes that are potential susceptibility genes in the development of Li-NDI. Finally, we have tried to confirm the role of the susceptibility genes in mpkCCD cells, as an in vitro model for lithium-induced NDI, by using a CRISPR and siRNA approach, however it was not possible to successfully downregulate the different candidate genes and maintain the integrity of the cell model. This confirmation is a necessary step before allowing translation to humans. As such, we did not commence on this part of the project, however we are currently investigating potential solutions and expect to find one soon.

Conclusion. This study identified that the development of lithium-induced side effects largely depends on the genetic background. Furthermore, we identified strains in which lithium-induced NDI, metabolic acidosis and hypercalcaemia can be studied, while we also identified strains with signs of chronic kidney damage after lithium treatment. Further studies are required to study whether prolonged exposure will lead to Li-CKD. Finally, using GWAS, we established a first list of susceptibility genes in the development of lithium-induced NDI, that remain to be established and extrapolated to humans.

Potential impact and use of results and any socio-economic impact of the project.

First, our identification that the response of mice to lithium is variable among strains is consistent with the fact that humans also react differently to lithium and confirms the involvement of a genetic background in Li-NDI development. Secondly, normally studies on the side effects of lithium are performed in the C57Bl/6J mouse strains, however we identified that other mouse strains are more suitable to study lithium-induced NDI, acidosis, hypercalcaemia and possibly CKD. This finding will allow other researchers to study specific side effects of lithium in the proper mouse strain. As the identification of a proper animal model is often key to study and cure a disease, our findings will strongly contribute to the acceleration of research in these areas. Third, the identification of susceptibility candidate genes will not only provide more information on the development of lithium-induced NDI, but also on the underlying physiological processes, being regulation of water homeostasis. As lithium-induced NDI is the most relevant side effect of lithium treatment and the most common form of NDI, this list of candidate genes is anticipated to be instrumental to the identification of susceptibility genes in humans with lithium-induced NDI. Prophylactic analysis of the susceptibility profile for lithiopathies of starting lithium-using patients will strongly increase the safety of the use of lithium as a cost-effective medication for the many patients suffering from bipolar and schizoaffective disorders, and possibly alcoholism, AIDS, dementia and cluster headaches, as lithium is also considered as a treatment for these disorders.

Contact information:
Prof. Peter Deen, 286 Dept. Physiology, Radboud Institute for Molecular Life Sciences, Radboud University Medical Center, Nijmegen, The Netherlands.
Email:; Phone: +31-243617347