What we need to know about the effect of lithium on the kidney

Abstract

Lithium has been a valuable treatment for bipolar affective disorders for decades. Clinical use of lithium, however, has been problematic due to its narrow therapeutic index and concerns for its toxicity in various organ systems. Renal side effects associated with lithium include polyuria, nephrogenic diabetes insipidus, proteinuria, distal renal tubular acidosis, and reduction in glomerular filtration rate. Histologically, chronic lithium nephrotoxicity is characterized by interstitial nephritis with microcyst formation and occasional focal segmental glomerulosclerosis. Nevertheless, this type of toxicity is uncommon, with the strongest risk factors being high serum levels of lithium and longer time on lithium therapy. In contrast, in experimental models of acute kidney injury and glomerular disease, lithium has antiproteinuric, kidney protective, and reparative effects. This paradox may be partially explained by lower lithium doses and short duration of therapy. While long-term exposure to higher psychiatric doses of lithium may be nephrotoxic, short-term low dose of lithium may be beneficial and ameliorate kidney and podocyte injury. Mechanistically, lithium targets glycogen synthase kinase-3β, a ubiquitously expressed serine/threonine protein kinase implicated in the processes of tissue injury, repair, and regeneration in multiple organ systems, including the kidney. Future studies are warranted to discover the exact “kidney-protective dose” of lithium and test the effects of low-dose lithium on acute and chronic kidney disease in humans.

lithium, a trace element essential for health (35), has been a valuable first-line treatment for bipolar affective disorders for the past 50 years (8, 24, 37). In addition to serving as a mood stabilizer, lithium has been successfully used to treat a variety of other disorders in humans, including migraine and cluster headache (29), arthritis (12), leukopenia (13, 30), neurodegenerative diseases (14, 19), and more (10, 12). Despite its affordability and effectiveness in diverse diseases, lithium has been less and less used in the past several decades, probably due to its adverse effects, such as nephrotoxicity, and the narrow therapeutic index (11, 25, 37). On the contrary, experimental evidence suggests that short-term low dose of lithium may have a kidney-protective effect (1). This perspective considers the seemingly paradoxical effects of lithium on the kidney.

Adverse effect of lithium on the kidney.

Within days to weeks following administration of lithium at psychiatric doses, almost 50% of the patients exhibit some degree of polyuria. Among those, 15∼20% patients develop nephrogenic diabetes insipidus (NDI), characterized by loss of ability of the kidney to concentrate urine due to resistance to vasopressin (6, 17, 37). Uncontrolled NDI may result in fluid and electrolyte disturbance, such as hypovolemia, hypernatremia, hyperchloremic metabolic acidosis, and distal renal tubular acidosis. Multiple mechanisms have been proposed to account for lithium-induced NDI. The epithelial sodium channel (ENaC)-mediated lithium entry in the collecting duct principal cells seems to contribute (7). Indeed, knockout of αENaC specifically in the collecting duct prevented the increase in water intake, polyuria, and reduced urine osmolality in mice chronically treated with lithium (7). Similarly, in rats, lithium treatment for 10 days downregulated aquaporin-2 expression, reduced the principal-to-intercalated cell ratio, and caused polyuria and renal concentrating deficit. Concomitant administration of amiloride, an ENaC inhibitor, attenuated all of these alternations (20), providing a rationale for its use in treating this disorder (4). In addition, lithium-induced NDI in mice is accompanied by increased cyclooxygenase-2 expression in renal interstitial cells and increased urine prostaglandin E₂ excretion. Selective cyclooxygenase-2 blockade or indomethacin counteracts lithium-induced NDI in mice and in humans, suggesting involvement of renal interstitial prostaglandin E₂ (32). Moreover, lithium blocks the activity of glycogen synthase kinase (GSK)-3 and recent data suggest that GSK3 might regulate vasopressin signaling and aquaporin-2 expression in collecting duct cells (32). The α- rather than the β-isoform of GSK3 seems to be involved because only GSK3α knockout mice have a urine-concentrating defect (15, 28). That effects of lithium other than GSK3 inhibition are mainly responsible for NDI is suggested by the fact that highly selective small-molecule inhibitors of GSK3 have only modest polyuric effects in rodents (8). Lithium may also compete with magnesium ions and thereby inhibit vasopressin-sensitive adenylyl cyclase (37). Moreover, lithium dysregulation of renal protein kinase-Cα, a kinase involved in phosphatidylinositol signaling, has also been implicated in NDI (36). Lithium-induced polyuria and NDI are usually benign conditions. Polyuria or NDI usually remit spontaneously with discontinuation of lithium, consistent with functional effects on the renal tubules (25).

For psychiatric patients, lithium may be prescribed long term for 10 to 20 years. Chronic lithium nephrotoxicity, including kidney dysfunction and urinary abnormalities such as proteinuria, has been described in such patients (17). In one of the largest published case series, Markowitz et al. (23) described 24 cases of biopsy-proven lithium nephrotoxicity, which represented 0.37% of 6,514 native kidney biopsies. All of the 24 patients presented with kidney dysfunction, 41.7% had proteinuria and 87% had NDI. Kidney biopsy demonstrated chronic tubulointerstitial nephritis in all patients, characterized by microcyst formation arising from distal tubules or collecting ducts. Glomerular lesions consistent with focal segmental glomerulosclerosis (FSGS) were found in 12 of the 24 patients in parallel with severe tubulointerstitial nephritis, possibly resulting from nephron loss and compensatory hypertrophy of surviving nephrons. Thus, FSGS in lithium-treated patients may be a postadaptive rather than a direct injurious effect of lithium on the glomerulus.

Despite case reports of lithium nephrotoxicity, the overall incidence of clinically significant chronic lithium nephrotoxicity appears to be very low. In a large-scale epidemiologic study in two regions of Sweden with 2.7 million inhabitants, the prevalence of chronic kidney disease in the lithium-treated population was about 1.2%, which is comparable to that in the general population (5). Time on therapy and thus a cumulative high dose of lithium were significant risk factors identified in this study (5).

Beneficial effect lithium on the kidney.

Lithium has been known to possess a proreparative activity following injury in multiple organ systems for many years (16). In the central nervous system, lithium has beneficial effects in acute neural injury and in chronic degenerative conditions, including Alzheimer's disease (14, 19). In patients with psychiatric disorders, lithium treatment induced an increase in brain gray matter (26). Lithium has been used to treat leukopenia (30) and may help mobilize hematopoietic stem cells in bone marrow transplant donors (13). Historically, lithium was thought to be beneficial for patients with hepatitis (18, 27, 34). The mechanism(s) of these beneficial effects of lithium has been unclear for years but has recently been attributed to the blocking of GSK3β, a well-conserved ubiquitously expressed multitasking serine/threonine protein kinase that plays a pivotal role in multiple pathophysiological processes. Originally described as a regulator of glycogen metabolism, GSK3β is now known to modulate inflammation, immunomodulation, embryo development, tissue injury, repair, and regeneration. Accumulating evidence suggests that low-dose lithium also has a kidney-protective effect in multiple experimental models of kidney disease (Fig. 1).

Fig. 1.

Paradoxical effects of lithium on the kidney. Lithium has been reported to confer both detrimental (black area and panels on left) and protective (white area and panels on right) effect on the kidney. At psychiatric doses, lithium quickly elicits polyuria and in some patients causes nephrogenic diabetes insipidus (NDI), which may uncommonly result in distal renal tubular acidosis (dRTA). Long-term lithium treatment in psychiatric patients occasionally causes chronic lithium nephrotoxicity, characterized by kidney dysfunction and interstitial nephritis with microcyst formation, and, occasionally, focal segmental glomerulosclerosis (FSGS). A number of molecular mechanisms have been implicated in the genesis of lithium nephrotoxicity, including the suppression of phosphatidylinositol (3,4,5)-trisphosphate (PIP₃) signaling, inhibition of glycogen synthase kinase (GSK)-3α, competition with magnesium ions (Mg²⁺), overactivity of cyclooxygenase (COX)-2, and more. In contrast, short-term low-dose lithium has beneficial effects in diverse experimental kidney diseases, including acute kidney injury (AKI) and glomerulonephropathies (GN). In experimental AKI, lithium protects against renal tubular death, suppresses kidney inflammation, and promotes renal repair. In animals with glomerular disease, lithium therapy reduces proteinuria, ameliorates podocytopathy, and glomerular damage. Mechanistically, the kidney-protective effect of lithium has been attributed to GSK3β inhibition.

Protective and proreparative effect of lithium on acute kidney injury.

The effect of lithium on experimental acute kidney injury (AKI) has been recently studied. In a murine model of lipopolysaccharide (LPS)-elicited AKI, a single intraperitoneal injection of a low dose of lithium chloride (40 mg/kg) at the time of LPS insult considerably reduced renal dysfunction, histological injury, and renal inflammation, associated with a marked improvement in survival (38). Lithium-dependent kidney protection was associated with inhibition of GSK3 and reduction in tubular cell apoptosis (38). This was confirmed in rats with gentamycin-elicited AKI (31). To determine whether lithium is effective in treating established AKI, a single dose of lithium was given to mice 3 days after cisplatin-induced AKI (3). Creatinine levels had peaked, and marked renal histological injury was evident at the time of lithium administration. Delayed treatment with low-dose lithium promoted renal tubular epithelia repopulation and accelerated recovery of renal function (3). The beneficial effect of lithium on AKI was further validated in a murine model of bilateral ischemia-reperfusion injury (IRI) (3). A single low dose of lithium 8 h after IRI promoted both functional recovery and histological repair of the acutely injured kidney, suggesting that lithium is effective in diverse forms of AKI (3). Mechanistically, the reparative effect of lithium was attributed to inhibition of GSK3β, leading to accelerated proliferation of surviving tubular epithelial cells. This was associated with increased expression of proproliferative molecules in renal tubular epithelial cells, including cyclin D1, c-Myc, and hypoxia-inducible factor-1α (3). Consistent with this hypothesis, GSK3β colocalized and physically interacted with these molecules in renal tubular epithelial cells and dictated their phosphorylation and degradation (3). Whether low-dose lithium can similarly prevent or ameliorate human AKI is currently unknown.

Beneficial effect of lithium in glomerular disease.

Proteinuria is an uncommon finding in psychiatric patients receiving lithium carbonate treatment, but the direct effect of lithium on proteinuria in glomerular diseases has been less examined. In female NZB/W mice with a spontaneous lupus-like autoimmune disease, treatment with lithium attenuated proteinuria and markedly improved survival (21). This antiproteinuric property was felt to be a direct glomerular-protective effect rather than the result of systemic immune modulation. Consistent with this view, serum levels of autoantibodies such as anti-retroviral glycoprotein 70 and anti-single-stranded DNA were minimally affected by lithium therapy (21).

Podocytes are a critical component of the glomerular filtration barrier contributing to glomerular permselectivity. Recent data suggest that low-dose lithium exerts a podocyte-protective effect. In cultured podocytes, lithium attenuated doxorubicin-induced actin cytoskeleton disorganization, reinstated actin cytoskeleton integrity, and prevented podocyte hypermotility, associated with downregulation of focal adhesion turnover (40). Mechanistically, lithium overrode the doxorubicin-elicited GSK3β activation, as well as hyperphosphorylation and activation of paxillin, a crucial element of cellular focal adhesion apparatus and substrate of GSK3β (40). In addition to paxillin, multiple other GSK3β substrates, such as NF-κB (2), nuclear factor (erythroid-derived)-like 2 (43), and a component of the mitochondrial transition pore, cyclophilin F (39), are phosphorylated and regulated by GSK3β in podocytes. It is tempting to speculate that lithium protects podocytes by intercepting a GSK3β-facilitated NF-κB-dependent inflammatory response, mitochondria permeability transition, and switching off of the nuclear factor (erythroid-derived)-like 2 antioxidant response (22). In vivo in mice with doxorubicin nephropathy, a single low dose of lithium lessened proteinuria and glomerulosclerosis (40). This was associated with reduction of GSK3β activity and preservation of actin cytoskeleton integrity in glomeruli, paralleled by early attenuation of podocyte foot process effacement (40). Recent studies demonstrate that inhibition of GSK3β in podocytes by other highly selective small-molecule inhibitors or by genetic knockout has podocyte-protective and antiproteinuric effect in multiple experimental glomerular diseases, including nephrotoxic serum nephritis (43), LPS podocytopathy (2), and lupus nephritis (36, 42). These data suggest that lithium might have beneficial effects in many other types of glomerular disease as well.

How to reconcile the seemingly paradoxical effect of lithium on the kidney?

“The dose makes the poison” (37) might be the best explanation for the seemingly paradoxical effects of lithium on the kidney (6). As mentioned above, epidemiological evidence suggests that high serum levels of lithium and the duration of lithium therapy are the strong predictors of its renal toxicity (5). In contrast, almost all studies demonstrating a kidney-protective effect of lithium employed short-term low-dose lithium (2, 3, 38–40, 43). Because it is ionized in body fluids, lithium cannot easily pass through the blood-brain barrier, and relatively high doses are needed to produce therapeutic effects in psychiatric disorders (33). The dose of lithium approved by the Food and Drug Administration for psychiatric disorders is comparatively high with a narrow therapeutic index (6, 17, 37). In mice, high-dose lithium provoked acute injury in renal tubular principal cells in days (11). Similarly, an ultrahigh dose of lithium (16 mmol/kg), almost two times the median lethal dose of lithium chloride in mice, rapidly induced transient albuminuria and histological signs of podocytopathy in five hours (9). While the exact “kidney-protective dose” of lithium is unknown, in our experience, a single dose that is one-third to one-half of the neurobiological dose is sufficient to block GSK3β in the diseased kidney and alter the phosphorylation status of GSK3β substrates, resulting in proliferation of renal tubular cells and cytoskeleton remodeling in podocytes (2, 3, 40, 41). Further studies are needed to determine the optimal dose of lithium for the treatment of kidney diseases in humans. Alternatively, given the counteracting effect of amiloride on lithium nephrotoxicity (4, 20), it is tempting to speculate that combined therapy with lithium and amiloride might tilt the balance to benefit. This certainly warrants further exploration. If confirmed in clinical trials, lithium might represent a novel, pragmatic, and affordable treatment for diverse types of kidney disease.

GRANTS

The research work of the authors has been supported in part by National Institute of Diabetes and Digestive and Kidney Diseases Grant R01-DK-092485 and the Foundation for Health (to R. Gong).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

R.G. and L.D.D. prepared figures; R.G. and L.D.D. drafted manuscript; R.G., P.W., and L.D.D. edited and revised manuscript; R.G., P.W., and L.D.D. approved final version of manuscript.