Significant outcomes
-
• The views on lithium mechanisms have evolved from the uric acid concept in the 19th century through various biochemical and pharmacological theories in the second part of the 20th century up to the 21st century findings obtained from in vitro and genetic studies, as well as related to new pathogenic hypotheses of bipolar disorder.
-
• The inhibition by lithium of GSK-3ß and the PI system may remain crucial for the therapeutic effect of this ion.
Limitations
-
• The variety and amount of data make it hardly possible to establish a cohesive concept of the lithium therapeutic mechanism.
-
• The novel information is constantly being discovered.
Introduction
Nowadays, lithium belongs to the most important psychotropic drugs. In all major guidelines, it is recommended as a drug of first choice for preventing recurrences in mood disorders, especially bipolar disorder (BD). Lithium is also useful for the treatment of manic states, less so for acute depressive ones; however, it can be used for the augmentation of antidepressant drugs in treatment-resistant depression (Rybakowski et al., Reference Rybakowski and Ferensztajn-Rochowiak2023). During long-term administration, lithium exerts a suicide-preventive effect (Tondo & Baldessarini, Reference Tondo and Baldessarini2025). Other properties of lithium include immunomodulatory and antiviral activity. In recent decades, the potential neuroprotective and anti-dementia effects of lithium have attracted increasing interest (Rybakowski, Reference Rybakowski2022).
In this paper, the evolution of views on lithium’s mechanisms of action will be presented, with particular emphasis on its therapeutic effect in mood disorders. Such an effect was observed in the second part of the 19th century and revived in 1949, the year lithium was introduced into modern psychiatry. The findings of the 21st century place the therapeutic properties of lithium beyond mood disorders. The mechanisms of its neuroprotective and anti-dementia properties will also be mentioned. This is a narrative historical review.
Lithium in the 19th century: The uric acid hypothesis
The observations of the high solubility of lithium urate and elevated urate levels in subjects with gout prompted British physician Alfred Baring Garrod (1819–1907) in 1859 to use lithium carbonate to treat gout and increase urate elimination in such patients (Garrod, 1959). A quarter of a century later, this concept was taken for psychiatric patients by the Danish scientist Carl Lange (1834–1900), who proposed the pathogenesis of depression as ‘uric acid diathesis’ where depression could be a gout of the brain. He successfully applied lithium carbonate for the treatment and prophylaxis of periodic depression (Lange, 1986). In the meantime, in 1871, an American neurologist, William Alexander Hammond (1828–1900), used lithium bromide to treat mania (Hammond, Reference Hammond1871). However, in the mechanism of antimanic action, a sedative effect of bromide cannot be excluded.
Following this, a hiatus for lithium administration in psychiatric patients lasted up to 1949, when Australian psychiatrist John Cade reported on the spectacular effect of lithium carbonate in the treatment of mania (Cade, Reference Cade1949). Interestingly, the association with uric acid was present here. Before giving lithium clinically, Cade observed that the urine of manic patients was particularly toxic to guinea pigs, attributing this to the excess of urates. He gave lithium urate to the animals, obtaining a sedative effect, and subsequently lithium carbonate with the same result.
Studies on the lithium mechanism in the second part of the 20th century
Electrolyte metabolism and membrane transport
As lithium belongs to the first group of the periodic table, along with sodium and potassium, its therapeutic mechanisms, involving electrolyte metabolism and cation transport, have become a reasonable topic of investigation. Initially, it was suggested that the therapeutic response to lithium may be associated with the magnitude of its retention in the organism. The lithium excretion test was elaborated, showing that manic patients who accumulated lithium better responded therapeutically than those who excreted more (Serry, Reference Serry1969).
A specific mechanism of transmembrane lithium transport was discovered in 1975 by Harvard investigators. This was identified by studying erythrocytes and was named lithium-sodium countertransport (LSC) (Haas et al., Reference Haas, Schooler and Tosteson1975). It has been shown that the LSC regulates lithium efflux from cells, and reduced LSC activity may be associated with increased lithium accumulation in certain cells (e.g. erythrocytes) (Rybakowski et al., Reference Rybakowski, Frazer, Mendels and Ramsey1977). Therefore, numerous studies have focused on the magnitude of the erythrocyte lithium ratio (i.e. the quotient of erythrocyte lithium concentration to serum lithium concentration). In several studies, a higher erythrocyte lithium ratio, that is, lower LSC activity, was found in patients with mood disorders (Pandey et al., Reference Pandey, Dorus, Casper, Janicek and Davis1984; Rybakowski et al., Reference Rybakowski, Frazer, Mendels and Ramsey1978). Additionally, a genetic component of this ratio was established (Dorus et al., Reference Dorus, Pandey, Shaughnessy and Davis1980). In some studies, lithium treatment has been found to inhibit LSC activity (Ehrlich et al., Reference Ehrlich, Diamond and Gosenfeld1981). However, a possible connection between LSC and the lithium mechanism in mood disorders has not been further pursued.
Another biological system involved in transmembrane cation transport, namely the sodium-potassium ATPase, was also investigated in relation to the lithium mode of action. Studies of erythrocytes conducted in the 1970s and 1980s found decreased activity of this enzyme in BD (Hokin-Neaverson & Jefferson, Reference Hokin-Neaverson and Jefferson1989; Naylor et al., Reference Naylor, Smith, Dick, Dick, McHarg and Chambers1980; Rybakowski et al., Reference Rybakowski, Potok and Strzyzewski1983a). In the 1990s, the sodium-potassium ATPase hypothesis of BD was formulated, where the authors proposed reduced Na+, K+-ATPase activity in this illness, which can lead to both mania and depression by increasing membrane excitability and decreasing neurotransmitter release, respectively (El-Mallakh & Wyatt, Reference El-Mallakh and Wyatt1995). The contribution of Na+, K+-ATPase activity to the pathogenesis of BD was confirmed in subsequent molecular-genetic studies (Goldstein et al., Reference Goldstein, Lerer, Laiba, Mallet, Mujaheed, Laurent, Rosen, Ebstein and Lichtstein2009; Mynett-Johnson et al., Reference Mynett-Johnson, Murphy, McCormack, Shields, Claffey, Manley and McKeon1998). Although Na+-K+-ATPase does not transport lithium, lithium treatment may enhance the enzyme’s activity, suggesting such a therapeutic mechanism in BD (Hokin-Neaverson & Jefferson, Reference Hokin-Neaverson and Jefferson1989; Rybakowski et al., Reference Rybakowski, Potok, Strzyzewski and Chłopocka-Woźniak1983b).
The adenyl cyclase (AC) and cyclic adenosine monophosphate (cAMP) system
The AC and cAMP belong to the second messenger systems, which are activated by monoaminergic neurotransmission. In the 1980s, Israeli researchers demonstrated that lithium treatment inhibits AC activity and reduces cAMP accumulation in the brain (Newman & Belmaker, Reference Newman and Belmaker1987). This was confirmed in 1990 by showing that, in rats’ cortical regions, lithium-induced decreases in cAMP levels are caused by increased activity of cAMP-phosphodiesterase and increased cGMP levels (Harvey et al., Reference Harvey, Carstens and Taljaard1990). Further evidence for the pathogenic role of adenylate cyclase in BD and the therapeutic mechanism of lithium was obtained in the 21st century. It was shown that lithium and carbamazepine inhibit AC5 isoforms, which can be associated with their antidepressant effect (Mann et al., Reference Mann, Heldman, Bersudsky, Vatner, Ishikawa, Almog, Belmaker and Agam2009). The genome-wide association study (GWAS) pointed to the adenylate cyclase-2 gene (ADCY2) as a risk gene for BD (Muhleisen et al., Reference Mühleisen, Leber, Schulze, Strohmaier, Degenhardt, Treutlein, Mattheisen, Forstner, Schumacher, Breuer, Meier, Herms, Hoffmann, Lacour, Witt, Reif, Müller-Myhsok, Lucae, Maier, Schwarz, Vedder, Kammerer-Ciernioch, Pfennig, Bauer, Hautzinger, Moebus, Priebe, Czerski, Hauser, Lissowska, Szeszenia-Dabrowska, Brennan, McKay, Wright, Mitchell, Fullerton, Schofield, Montgomery, Medland, Gordon, Martin, Krasnow, Chuchalin, Babadjanova, Pantelejeva, Abramova, Tiganov, Polonikov, Khusnutdinova, Alda, Grof, Rouleau, Turecki, Laprise, Rivas, Mayoral, Kogevinas, Grigoroiu-Serbanescu, Propping, Becker, Rietschel, Nöthen and Cichon2014). This was further confirmed by Iranian researchers, who suggested that the ADCY2 polymorphism may serve as a potential predictive marker of lithium response (Aghabozorg et al., Reference Aghabozorg Afjeh, Shams, Hamednia, Boshehri, Olfat and Omrani2020).
Related to cAMP is a transcription factor known as cAMP response element-binding protein (CREB). This is associated with the activity of brain-derived neurotrophic factor (BDNF), one of the most important components of neuronal plasticity. The role of BDNF in BD, as well as in the lithium mechanism, will be discussed later.
Phosphatidylinositol (PI) system
The PI system is involved in the release of inositol 1,4,5-trisphosphate (IP3), a second messenger in signal transduction. It was demonstrated that lithium inhibits inositol monophosphatase (IMPase), a crucial step in this mechanism (Hallcher and Sherman, Reference Hallcher and Sherman1980). Several years later, an inositol depletion hypothesis of BD was formulated, in which lithium influences the PI system, thereby weakening the second signalling pathway. The proposed lithium mode of action includes inhibition of IMPase, leading to increased inositol-1-monophosphate levels, followed by phosphoinositide accumulation, decreased PI concentrations, and altered levels of key membrane phospholipids (Berridge et al., Reference Berridge, Downes and Hanley1989). Furthermore, it was shown that, concurrent with IMPase inhibition, lithium decreases the amount of PI-specific phospholipase C, which is critical for the generation of IP3, diacylglycerol, and cAMP, resulting in downregulation of inositol phosphates (Soares et al., Reference Soares, Chen, Dippold, Wells, Frank, Kupfer, Manji and Mallinger2000). In the 21st century, experimental studies in mice by knocking out one of the two genes associated with the PI system: IMPA1 (inositol monophosphatase-1) or SMIT1 (sodium myo-inositol transporter-1) demonstrated that both procedures produced behaviour similar to that observed after lithium administration. Thus, both inositol depletion and phosphoinositide accumulation may contribute to lithium’s effects (Agam et al., Reference Agam, Bersudsky, Berry, Moechars, Lavi-Avnon and Belmaker2009). Molecular genetic studies have confirmed the role of the PI system in the pathogenesis of BD. A 2008 GWAS study of BD patients identified an association with the diacylglycerol kinase eta (DGKH) gene, which encodes an essential protein in the PI pathway (Baum et al., Reference Baum, Akula, Cabanero, Cardona, Corona, Klemens, Schulze, Cichon, Rietschel, Nöthen, Georgi, Schumacher, Schwarz, Abou Jamra, Höfels, Propping, Satagopan, Detera-Wadleigh, Hardy and McMahon2008). Also, an association between the IMPA2 gene polymorphism and BD has been reported (Bloch et al., Reference Bloch, Weller, Doyle, Ferraro, Berrettini, Hodge and Lohoff2010). The effect of lithium on the PI system is regarded as one of the most important mechanisms and has been observed in several lithium processes discovered in the 21st century.
Glycogen synthase kinase-3 beta (GSK-3ß)
GSK-3ß is involved in numerous signalling pathways and multiple physiological functions, including, inter alia, neurogenesis, neuronal growth, and biological rhythms. Thirty years ago, two papers were published that described lithium’s competitive mechanism of GSK-3β inhibition (Klein & Melton, Reference Klein and Melton1996; Stambolic et al., Reference Stambolic, Ruel and Woodgett1996). Since then, the inhibition of GSK-3β by lithium has been postulated to be the best-evidenced molecular mechanism underlying this ion’s therapeutic action in BD. Whereas the most recent GWAS study of bipolar patients found no common variants in the GSK-3β gene associated with an increased risk of the illness, and it is not clear which of the multiple target molecules of this enzyme may explain the drug’s therapeutic activity in BD (Mullins et al., Reference Mullins, Forstner, O’Connell, Coombes, Coleman, Qiao, Als, Bigdeli, Børte, Bryois, Charney, Drange, Gandal, Hagenaars, Ikeda, Kamitaki, Kim, Krebs, Panagiotaropoulou, Schilder, Sloofman, Steinberg, Trubetskoy, Winsvold, Won, Abramova, Adorjan, Agerbo, Al Eissa, Albani, Alliey-Rodriguez, Anjorin, Antilla, Antoniou, Awasthi, Baek, Baekvad-Hansen, Bass, Bauer, Beins, Bergen, Birner, Bocker Pedersen, Boen, Boks, Bosch, Brum, Brumpton, Brunkhorst-Kanaan, Budde, Bybjerg-Grauholm, Byerley, Cairns, Casas, Cervantes, Clarke, Cruceanu, Cuellar-Barboza, Cunningham, Curtis, Czerski, Dale, Dalkner, David, Degenhardt, Djurovic, Dobbyn, Douzenis, Torn, Escott-Price, Ferrier, Fiorentino, Foroud, Forty, Frank, Frei, Freimer, Frisen, Gade, Garnham, Gelernter, Giortz Pedersen, Gizer, Gordon, Gordon-Smith, Greenwood, Grove, Guzman-Parra, Ha, Haraldsson, Hautzinger, Heilbronner, Hellgren, Herms, Hoffmann, Holmans, Huckins, Jamain, Johnson, Kalman, Kamatani, Kennedy, Kittel-Schneider, Knowles, Kogevinas, Koromina, Kranz, Kranzler, Kubo, Kupka, Kushner, Lavebratt, Lawrence, Leber, Lee, Lee, Levy, Lewis, Liao, Lucae, Lundberg, MacIntyre, Magnusson, Maier, Maihofer, Malaspina, Maratou, Martinsson, Mattheisen, McCarroll, McGregor, McGuffin, McKay, Medeiros, Medland, Millischer, Montgomery, Moran, Morris, Muhleisen, O’Brien, O’Donovan, Olde Loohuis, Oruc, Papiol, Pardinas, Perry, Pfennig, Porichi, Potash, Quested, Raj, Rapaport, DePaulo, Regeer, Rice, Rivas, Rivera, Roth, Roussos, Ruderfer, Sanchez-Mora, Schulte, Senner, Sharp, Shilling, Sigurdsson, Sirignano, Slaney, Smeland, Smith, Sobell, Soholm Hansen, Soler Artigas, Spijker, Stein, Strauss, Swiatkowska, Terao, Thorgeirsson, Toma, Tooney, Tsermpini, Vawter, Vedder, Walters, Witt, Xi, Xu, Yang, Young, Young, Zandi, Zhou, Zillich, Agartz, Alda, Alfredsson, Babadjanova, Backlund, Baune, Bellivier, Bengesser, Berrettini, Blackwood, Boehnke, Borglum, Breen, Carr, Catts, Corvin, Craddock, Dannlowski, Dikeos, Esko, Etain, Ferentinos, Frye, Fullerton, Gawlik, Gershon, Goes, Green, Grigoroiu-Serbanescu, Hauser, Henskens, Hillert, Hong, Hougaard, Hultman, Hveem, Iwata, Jablensky, Jones, Jones, Kahn, Kelsoe, Kirov, Landen, Leboyer, Lewis, Li, Lissowska, Lochner, Loughland, Martin, Mathews, Mayoral, McElroy, McIntosh, McMahon, Melle, Michie, Milani, Mitchell, Morken, Mors, Mortensen, Mowry, Muller-Myhsok, Myers, Neale, Nievergelt, Nordentoft, Nothen, O’Donovan, Oedegaard, Olsson, Owen, Paciga, Pantelis, Pato, Pato, Patrinos, Perlis, Posthuma, Ramos-Quiroga, Biernacka, Nurnberger, Cichon, Edenberg, Stahl, McQuillin, Di Florio, Ophoff and Andreassen2021). Recently, GSK-3ß substrates have been proposed to contribute to lithium’s effects in BD, including the transcription factor CREB, the RNA-binding protein FXR1 (Fragile X-related protein 1), kinesin subunits, and the cytoskeletal regulator CRMP2 (Collapsin response mediator protein 2). They are all associated with the pathogenesis of BD, although in different ways (Chatterjee et al., Reference Chatterjee and Beaulieu2022). It was also found that lithium’s inhibition of GSK-3β affects the PI3K/Akt pathway, increasing Akt-1 activity by inhibiting phosphoinositide 3-kinase-mediated phosphorylation of Akt (Tian et al., Reference Tian, Kanno, Jin and Nishizaki2014).
Research in the 21st century has identified many effects of lithium associated with GSK3β inhibition. For example, lithium’s influence on GSK-3β might play a role in lithium-induced kidney impairment (Rybakowski et al., Reference Rybakowski, Abramowicz, Szczepankiewicz, Michalak, Hauser and Czekalski2013) and in its anti-suicidal effects (Malhi et al., Reference Malhi, Das, Outhred, Irwin, Morris, Hamilton, Lynch and Mannie2018). Recently, the inhibitory effect of lithium on GSK3ß has been regarded as the most important mechanism underlying its favourable effects in Alzheimer’s disease (AD) (Shen et al., Reference Shen, Zhao, Zhao, Meng, Zhang, Zhang, Taishi and Sun2024).
Neurotransmitters
The effect of lithium on dopaminergic neurotransmission has been postulated since the 1970s. It was found that lithium reduces dopaminergic activity, which may be related to its antimanic effect (Gallagher et al., Reference Gallager, Pert and Bunney1978). However, the most definitive papers and reviews on this subject appeared in the 21st century. Can et al. (Reference Can, Frost, Cachope, Cheer and Gould2016) showed that chronic lithium treatment normalises dopamine release in the nucleus accumbens. The most recent review by Mohamadian et al. (Reference Mohamadian, Fallah, Ghofrani-Jahromi, Rahimi-Danesh, Shokouhi Qare Saadlou and Vaseghi2023) highlights the regulation of dopamine transmission by lithium and the relationship of that effect to lithium’s effect on GSK-3β.
The serotonin hypothesis regarding the mechanism of long-term lithium action, involving stimulation of serotonergic neurotransmission, was proposed more than forty years ago (Müller-Oerlingausen, Reference Muller-Oerlinghausen1985). Several years later, it was demonstrated that bipolar patients treated with lithium have increased plasma serotonin concentration (Artigas et al., Reference Artigas, Sarrias, Martínez, Gelpí, Alvarez and Udina1989). In their review paper, Price et al. (Reference Price, Charney, Delgado and Heninger1990) pointed out that both presynaptic and postsynaptic effects of lithium on the serotonergic system enhance serotonergic neurotransmission. More recently, in the second decade of the 21st century, the serotonin link of the lithium mechanism was reiterated. A study using positron emission tomography conducted in BD patients during a depressive episode demonstrated that those with lower baseline serotonin transporter binding and serotonin 5-HT1A receptor binding showed greater clinical improvement after lithium treatment (Ananth et al., Reference Ananth, Bartlett, DeLorenzo, Lin, Kunkel, Vadhan, Perlman, Godstrey, Holzmacher, Ogden, Parsey and Huang2020). It was also hypothesised that the therapeutic effect of lithium is linked to activation of gastrointestinal enterochromaffin cells that secrete serotonin and modulate vagal and area postrema inputs to the brain within the framework of the gut-brain axis. Such activation is dependent on the Transient Receptor Potential Melastatin 2 (TRPM2) channel (Thorpe et al., Reference Thorpe, Ootsuka, Peterson, Jones, Humenick, Martin, Zivkovic, Sakaguchi, McArdle, Ittner, Brookes, Habib, Sirimaharaj, Tawodros, Roach, Edwards, Clevers, Beumer, Puschof, Wei, Singh, Ha, Ro, Zubcevic, Otmanowski, Mendez-Hernandez, de Lartigue, Blessing and Keating2026). The TRPM2 gene is a susceptibility gene for BD and regulates GSK-3ß activity in the brain (Jang et al., Reference Jang, Lee, Lee, Jung, Khalid, Uchida, Tominaga, Jeon and Oh2015).
Research on the effects of lithium on glutamatergic neurotransmission began in the late 1990s. It was found that chronic lithium administration protects neurones in the central nervous system against neurotoxicity by inhibiting NMDA receptor-mediated calcium influx (Hokin et al., Reference Hokin, Dixon and Los1996; Nonaka et al., Reference Nonaka, Hough and Chuang1998). This was supposed to be a mechanism of lithium’s mood-stabilising and antimanic effect (Ghasemi & Dehpour, Reference Ghasemi and Dehpour2011; Öngür et al., Reference Öngür, Jensen, Prescot, Stork, Lundy, Cohen and Renshaw2008). A possible role of glutamatergic mechanisms was further confirmed in GWAS studies (Amare et al., Reference Amare, Thalamuthu, Schubert, Fullerton, Ahmed, Hartmann, Papiol, Heilbronner, Degenhardt, Tekola-Ayele, Hou, Hsu, Shekhtman, Adli, Akula, Akiyama, Ardau, Arias, Aubry, Hasler, Richard-Lepouriel, Perroud, Backlund, Bhattacharjee, Bellivier, Benabarre, Bengesser, Biernacka, Birner, Marie-Claire, Cervantes, Chen, Chillotti, Cichon, Cruceanu, Czerski, Dalkner, Del Zompo, DePaulo, Étain, Jamain, Falkai, Forstner, Frisen, Frye, Gard, Garnham, Goes, Grigoroiu-Serbanescu, Fallgatter, Stegmaier, Ethofer, Biere, Petrova, Schuster, Adorjan, Budde, Heilbronner, Kalman, Kohshour, Reich-Erkelenz, Schaupp, Schulte, Senner, Vogl, Anghelescu, Arolt, Dannlowski, Dietrich, Figge, Jäger, Lang, Juckel, Konrad, Reimer, Schmauß, Schmitt, Spitzer, von Hagen, Wiltfang, Zimmermann, Andlauer, Fischer, Bermpohl, Ritter, Matura, Gryaznova, Falkenberg, Yildiz, Kircher, Schmidt, Koch, Gade, Trost, Haussleiter, Lambert, Rohenkohl, Kraft, Grof, Hashimoto, Hauser, Herms, Hoffmann, Jiménez, Kahn, Kassem, Kuo, Kato, Kelsoe, Kittel-Schneider, Ferensztajn-Rochowiak, König, Kusumi, Laje, Landén, Lavebratt, Leboyer, Leckband, Tortorella, Manchia, Martinsson, McCarthy, McElroy, Colom, Millischer, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, O’Donovan, Ozaki, Pfennig, Pisanu, Potash, Reif, Reininghaus, Rouleau, Rybakowski, Schalling, Schofield, Schweizer, Severino, Shilling, Shimoda, Simhandl, Slaney, Squassina, Stamm, Stopkova, Maj, Turecki, Vieta, Veeh, Witt, Wright, Zandi, Mitchell, Bauer, Alda, Rietschel, McMahon, Schulze, Clark and Baune2023; Chen et al., Reference Chen, Lee, Lee, Ouyang, Chen, Chong, Wu, Tan, Lee, Chuo, Chiu, Tsang, Chang, Lung, Chiu, Chang, Chen, Hou, Chen, Lai, Tung, Chen, Lane, Su, Feng, Lin, Chang, Teng, Liu, Chen, Liu, Chen, Lu, Fan, Wu, Li, Wang, Wu, Peng, Chang, Lu, Chen and Cheng2014; Perlis et al., Reference Perlis, Smoller, Ferreira, McQuillin, Bass, Lawrence, Sachs, Nimgaonkar, Scolnick, Gurling, Sklar and Purcell2009).
Antiviral activity
Lithium exhibits significant antiviral activity against herpesvirus infections. In the late 1970s, two patients with recurrent herpesvirus infection were described to have remission during lithium treatment (Lieb, Reference Lieb1979). About the same time, researchers at the University of Birmingham showed that lithium inhibited the replication of herpes simplex virus (HSV) in hamster kidney cells (Skinner et al., Reference Skinner, Hartley, Buchan, Harper and Gallimore1980). Eleven years later, a retrospective analysis of labial herpes recurrences in patients on prophylactic lithium was conducted in a common research project between the Department of Adult Psychiatry at Poznan University of Medical Sciences and the Department of Psychiatry at the University of Pennsylvania. During lithium administration, the disappearance of herpes recurrences occurred in 46% of Polish patients with labial herpes, and the overall reduction in recurrence rate was 64%. The bigger effect was revealed in subjects with serum lithium concentrations >0.65 mmol/l and erythrocyte lithium concentrations >0.35 mmol/l. Among American lithium-treated BD patients, the recurrence rate of labial herpes was reduced by 73% compared with the five-year pre-lithium period. Whereas in patients with recurrent depression receiving antidepressants, no difference in the recurrence compared to the five years before treatment was observed (Rybakowski & Amsterdam, Reference Rybakowski and Amsterdam1991). From the current perspective, the anti-herpes effect of lithium could play a role in its prophylactic and therapeutic effects in dementia, as the infection with HSV type 1 (HSV-1) is a significant risk factor for AD (Itzhaki, Reference Itzhaki2018). In patients with BD, infection with HSV-1 was connected with impairment of cognitive functioning (Dickerson et al., Reference Dickerson, Boronow, Stallings, Origoni, Cole, Krivogorsky and Yolken2004), and the effect of lithium may be related to a favourable effect on cognition in this illness (Rybakowski, Reference Rybakowski2016).
Despite its effect on herpesviruses, lithium may also influence RNA viruses responsible for respiratory infections. In a retrospective assessment, a remarkable reduction in the mean annual rate of flu-like infections was observed among patients taking lithium, but not among those receiving antidepressants (Amsterdam et al., Reference Amsterdam, García-España and Rybakowski1998). Recently, a 1/3 decrease in respiratory infections, mostly due to RNA virus infections, was noticed during long-term lithium treatment (Landén et al., Reference Landén, Larsson, Lichtenstein, Westin and Song2021). Also, the antiviral potential of lithium against coronaviruses has been demonstrated in experimental studies (Nowak & Walkowiak, Reference Nowak and Walkowiak2020), raising hope for its use in the treatment of COVID-19 (Murru et al., Reference Murru, Manchia, Hajek, Nielsen, Rybakowski, Sani, Schulze, Tondo and Bauer2020). A randomised clinical study found that lithium augmentation can reduce the severity of the COVID-19 course (Spuch et al., Reference Spuch, López-García, Rivera-Baltanás, Cabrera-Alvargonzález, Gadh, Rodrigues-Amorim, Álvarez-Estévez, Mora, Iglesias-Martínez-Almeida, Freiría-Martínez, Pérez-Rodríguez, Pérez-González, López-Domínguez, Longueira-Suarez, Sousa-Domínguez, Araújo-Ameijeiras, Mosquera-Rodríguez, Crespo, Vila-Fernández, Regueiro and Olivares2022). The review of more than 26,000 electronic records found a lower infection frequency in those with lithium levels 0.5–1.0 mmol/l compared with those with lithium concentrations <0.5 mmol/l (De Picker et al., Reference De Picker, Leboyer, Geddes, Morrens, Harrison and Taquet2022). Also, a recent analysis of electronic records in Israel showed that lithium use was associated with decreased frequency of severe COVID-19, even though the users were older and had more medical comorbidities. In contrast, valproate was linked with a higher risk (Avni et al., Reference Avni, Blasbalg and Toren2026).
The studies on lithium mechanisms in the 21st century: Focus on lithium responders
In vitro studies
New avenues of lithium research were explored using in vitro methods with peripheral blood mononuclear cells (PBMCs), lymphoblastoid cell lines (LCLs), and induced pluripotent stem cells (iPSCs). In many instances, different results were obtained in the cells derived from lithium responders (LRs) and lithium non-responders (LNRs).
The most important finding from the PBMC studies concerns the association between lithium response and GSK-3ß phosphorylation at Ser9 (Li et al., Reference Li, Friedman, Zhu, Wang, Boswell, May, Davis and Jope2007). In another study, cells from BD patients and controls showed distinct immunophenotypes, suggesting immunologic dysfunction in patients (Wu et al., Reference Wu, Lee, Wu, Sun, Chang, Chen, Chen, Wu and Cheng2019). The difference between LR and LNR was also observed in several studies using LCLs. After in vitro lithium treatment, 2,060 genes in LCLs showed differential expression between LR and LNR, with insulin-like growth factor 1 significantly associated with LR (Squassina et al., Reference Squassina, Costa, Congiu, Manchia, Angius, Deiana, Ardau, Chillotti, Severino, Calza and Del Zompo2013). In another study, increased expression of HDGFRP3 (hepatoma-derived growth factor-related protein-3) and ID2 (inhibitor of DNA binding 2), genes associated with neurogenesis and neurotrophic processes, was found in the LR (Milanesi et al., Reference Milanesi, Voinsky, Hadar, Srouji, Maj, Shekhtman, Gershovits, Gilad, Chillotti, Squassina, Potash, Schulze, Goes, Zandi, Kelsoe and Gurwitz2019). Lithium treatment of LCLs normalised cell viability by modulating GSK-3 expression in LR (Paul et al., Reference Paul, Iyer, Nadella, Nayak, Chellappa, Ambardar, Sud, Sukumaran, Purushottam, Jain and Viswanath2020), and immunoglobulin-related genes in LCLs differed between LR and LNR (Mizrahi et al., Reference Mizrahi, Choudhary, Ofer, Goldberg, Milanesi, Kelsoe, Gurwitz, Alda, Gage and Stern2023). The dissimilarity between LR and LNR in LCLs was also obtained by investigating lithium treatment’s effects on microRNA (miRNA) expression. The differences were particularly evident in miRNAs involved in immune processes (Cattane et al., Reference Cattane, Courtin, Mombelli, Maj, Mora, Etain, Bellivier, Marie-Claire and Cattaneo2022; Chen et al., Reference Chen, Wang, Burmeister and McInnis2009).
Studies on iPSCs also yielded several findings. The iPSC-derived neurones and cortical spheroids from BD patients exhibited hyperexcitability compared with those from control subjects. Treatment with lithium normalised excitability, a normalisation more pronounced in LR (Osete et al., Reference Osete, Akkouh, Ievglevskyi, Vandenberghe, de Assis, Ueland, Kondratskaya, Holen, Szabo, Hughes, Smeland, Steen, Andreassen and Djurovic2023; Mertens et al., Reference Mertens, Wang, Kim, Yu, Pham, Yang, Zheng, Diffenderfer, Zhang, Soltani, Eames, Schafer, Boyer, Marchetto, Nurnberger, Calabrese, Oedegaard, McCarthy, Zandi, Alda, Nievergelt, Mi, Brennand, Kelsoe, Gage and Yao2015; Stern et al., Reference Stern, Santos, Marchetto, Mendes, Rouleau, Biesmans, Wang, Yao, Charnay, Bang, Alda and Gage2018). Reducing excitability was associated with inhibition of inositol monophosphatase 1, a component of the PI system (Saha et al., Reference Saha, Krishnan and Raghu2023), and activation of the Akt signalling pathway, which is associated with cell growth and proliferation (Khayachi et al., Reference Khayachi, Abuzgaya, Liu, Jiao, Dejgaard, Schorova, Kamesh, He, Cousineau, Pietrantonio, Farhangdoost, Castonguay, Chaumette, Alda, Rouleau and Milnerwood2024). As to LNRs, a reduced expression of the LEF1 (lymphoid enhancer-binding factor 1) gene, modulating neuronal activity, was demonstrated (Santos et al., Reference Santos, Linker, Stern, Mendes, Shokhirev, Erikson, Randolph-Moore, Racha, Kim, Kelsoe, Bang, Alda, Marchetto and Gage2021).
The expression of 28 cellular pathway proteins in monocytes and CD4+ lymphocytes from LR and LNR was also investigated using an enhanced flow cytometry method (CellPrint). No differences were observed between LR and LNR before lithium treatment. Whereas after treatment, most proteins had increased expression in LR and decreased in LNR, especially PDEB4 (cAMP-specific 3’,5’-cyclic phosphodiesterase 4B) and NR3C1 (nuclear receptor subfamily 3, group C, member 1) proteins (Gao et al., Reference Gao, Kaye, Ayati, Koyuturk, Calabrese, Christian, Lazarus and Kaplan2023).
Lithium’s effects on neural cells were also investigated in experimental research on animal models. In mouse neuroblasts, lithium enhanced the mRNA expression of the mesencephalic astrocyte-derived neurotrophic factor (Abu-Hijleh et al., Reference Abu-Hijleh, Prashar, Joshi, Sharma, Frey and Mishra2021). In mouse cortical neurones, chronic lithium treatment inhibited calcium influx by activating the metabotropic glutamate receptor Glu-5 and inhibiting protein kinase C and GSK-3ß, resulting in diminished neuronal excitability (Khayachi et al., Reference Khayachi, Ase, Liao, Kamesh, Kuhlmann, Schorova, Chaumette, Dion, Alda, Séguéla, Rouleau and Milnerwood2021). In a knockout mouse model lacking Ank3 expression, lithium normalised dendritic anomalies caused by Ank3 deficiency by inhibiting GSK3-ß and stimulating AC (Piguel et al., Reference Piguel, Yoon, Gao, Horan, Garza, Petryshen, Smith and Penzes2023).
Genetic studies
The main topic of genetic studies on lithium was mostly seeking markers of prophylactic response in BD. The first wave of these studies was associated with so-called candidate genes and was reviewed by Rybakowski (Reference Rybakowski2013). In subsequent years, such genes were supplemented at the Poznan centre with those associated with stress response (Szczepankiewicz et al., Reference Szczepankiewicz, Narozna, Rybakowski, Kliwicki, Czerski, Dmitrzak-Węglarz, Skibińska, Twarowska-Hauser and Pawlak2018). In 2019, American researchers evaluated polymorphisms in 45 candidate genes linked to lithium’s prophylactic efficacy. They found the strongest association with the CACNG2 (calcium voltage-gated channel auxiliary subunit gamma 2) gene, known as stargazin, and the NRG1 (neuregulin 1) gene, previously linked to schizophrenia (Miranda et al., Reference Miranda, Shekhtman, McCarthy, DeModena, Leckband and Kelsoe2019). A more recent analysis listed six candidate genes associated with the prophylactic effect of lithium, evidenced in at least two studies, such as GSK3β, BDNF, serotonin transporter (5HTTLPR), calcium voltage-gated channel auxiliary subunit gamma 2 (CACNG2), nuclear receptor subfamily 1 group D member 1 (NR1D1), and dopamine receptor D1 (DRD1) gene (Senner et al., Reference Senner, Kohshour, Abdalla, Papiol and Schulze2021).
Shortly after candidate gene research, GWAS studies were conducted. In the study by Perlis et al. (Reference Perlis, Smoller, Ferreira, McQuillin, Bass, Lawrence, Sachs, Nimgaonkar, Scolnick, Gurling, Sklar and Purcell2009), response to lithium was associated with the GRIA2 gene, which codes for the ionotropic alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) glutamatergic receptor. Chinese researchers demonstrated an association between lithium response and the GADL1 gene, which encodes glutamate decarboxylase-like protein 1 (Chen et al., Reference Chen, Lee, Lee, Ouyang, Chen, Chong, Wu, Tan, Lee, Chuo, Chiu, Tsang, Chang, Lung, Chiu, Chang, Chen, Hou, Chen, Lai, Tung, Chen, Lane, Su, Feng, Lin, Chang, Teng, Liu, Chen, Liu, Chen, Lu, Fan, Wu, Li, Wang, Wu, Peng, Chang, Lu, Chen and Cheng2014). In another study, an association between lithium’s beneficial effects and the SEC14 and spectrin domain 1 (SESTD1) gene, which encodes a synapse protein that directly binds to phospholipids, was found (Song et al., Reference Song, Bergen, Di Florio, Karlsson, Charney, Ruderfer, Stahl, Chambert, Moran, Gordon-Smith, Forty, Green, Jones, Jones, Scolnick, Sklar, Smoller, Lichtenstein, Hultman, Craddock, Landén, Smoller, Perlis, Lee, Castro, Hoffnagle, Sklar, Stahl, Purcell, Ruderfer, Charney, Roussos, Michele Pato, Medeiros, Sobel, Craddock, Jones, Forty, Florio, Green, Jones, Gordon-Smith, Landen, Hultman, Jureus, Bergen, McCarroll, Moran, Smoller, Chambert and Belliveau2016). Recently, Wolter et al. (Reference Wolter, Le, Matoba, Lafferty, Aygün, Liang, Courtney, Song, Piven, Zylka and Stein2023) identified, in a GWAS, a region on chromosome 3 at p21.1 containing the gene GNL3 (G protein nucleolar 3), which is essential for neuronal progenitor cell proliferation in response to lithium. Whole exome sequencing research in a family in which monozygotic twins differed in their prophylactic response to lithium identified differences in six genes, including DISC-1 (Disrupted-in-Schizophrenia 1) (Jacobs et al., Reference Jacobs, Hagin, Shugol, Shomron, Pillar, Fañanás, Serretti, Vieta and Popovic2020). Previously, this gene, initially identified as being associated with a predisposition to schizophrenia, was also found to be linked to BD (Perlis et al., Reference Perlis, Purcell, Fagerness, Kirby, Petryshen, Fan and Sklar2008).
The Consortium of Lithium Genetics (ConLiGen), established in 2009, was intended to make a significant contribution to understanding the genetic factors associated with lithium prophylactic efficacy (Schulze et al., Reference Schulze, Alda, Adli, Akula, Ardau, Bui, Chillotti, Cichon, Czerski, Del Zompo, Detera-Wadleigh, Grof, Gruber, Hashimoto, Hauser, Hoban, Iwata, Kassem, Kato, Kittel-Schneider, Kliwicki, Kelsoe, Kusumi, Laje, Leckband, Manchia, Macqueen, Masui, Ozaki, Perlis, Pfennig, Piccardi, Richardson, Rouleau, Reif, Rybakowski, Sasse, Schumacher, Severino, Smoller, Squassina, Turecki, Young, Yoshikawa, Bauer and McMahon2010). In 2016, the first GWAS involving 2,563 patients from 22 centres showed that the chromosome 21 region, which includes two long non-coding RNAs (lncRNAs) that regulate gene expression in the central nervous system, was associated with lithium efficacy (Hou et al., Reference Hou, Heilbronner, Degenhardt, Adli, Akiyama, Akula, Ardau, Arias, Backlund, Banzato, Benabarre, Bengesser, Bhattacharjee, Biernacka, Birner, Brichant-Petitjean, Bui, Cervantes, Chen, Chen, Chillotti, Cichon, Clark, Colom, Cousins, Cruceanu, Czerski, Dantas, Dayer, Étain, Falkai, Forstner, Frisén, Fullerton, Gard, Garnham, Goes, Grof, Gruber, Hashimoto, Hauser, Herms, Hoffmann, Hofmann, Jamain, Jiménez, Kahn, Kassem, Kittel-Schneider, Kliwicki, König, Kusumi, Lackner, Laje, Landén, Lavebratt, Leboyer, Leckband, Jaramillo, MacQueen, Manchia, Martinsson, Mattheisen, McCarthy, McElroy, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Ösby, Ozaki, Perlis, Pfennig, Reich-Erkelenz, Rouleau, Schofield, Schubert, Schweizer, Seemüller, Severino, Shekhtman, Shilling, Shimoda, Simhandl, Slaney, Smoller, Squassina, Stamm, Stopkova, Tighe, Tortorella, Turecki, Volkert, Witt, Wright, Young, Zandi, Potash, DePaulo, Bauer, Reininghaus, Novák, Aubry, Maj, Baune, Mitchell, Vieta, Frye, Rybakowski, Kuo, Kato, Grigoroiu-Serbanescu, Reif, Del Zompo, Bellivier, Schalling, Wray, Kelsoe, Alda, Rietschel, McMahon and Schulze2016). Some of the further ConLiGen studies evaluated the connection with polygenic risk score (PRS) for various psychiatric disorders. The presence of a PRS for schizophrenia was associated with poorer lithium’s prophylactic efficacy (Amare et al., Reference Amare, Schubert, Hou, Clark, Papiol, Heilbronner, Degenhardt, Tekola-Ayele, Hsu, Shekhtman, Adli, Akula, Akiyama, Ardau, Arias, Aubry, Backlund, Bhattacharjee, Bellivier, Bengesser, Biernacka, Birner, Brichant-Petitjean, Cervantes, Chen, Chillotti, Cichon, Cruceanu, Czerski, Dalkner, Dayer, Del Zompo, DePaulo, Étain, Falkai, Forstner, Frisen, Frye, Fullerton, Garnham, Goes, Grigoroiu-Serbanescu, Grof, Hashimoto, Hauser, Herms, Hoffmann, Hofmann, Jamain, Jiménez, Kahn, Kassem, Kuo, Kato, Kelsoe, Kittel-Schneider, Kliwicki, König, Kusumi, Laje, Landén, Lavebratt, Leboyer, Leckband, Tortorella, Manchia, Martinsson, McCarthy, McElroy, Colom, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, O’Donovan, Ozaki, Ösby, Pfennig, Potash, Reif, Reininghaus, Rouleau, Rybakowski, Schalling, Schofield, Schweizer, Severino, Shilling, Shimoda, Simhandl, Slaney, Squassina, Stamm, Stopkova, Maj, Turecki, Vieta, Volkert, Witt, Wright, Zandi, Mitchell, Bauer, Alda, Rietschel, McMahon, Schulze and Baune2018), which corresponds with a clinical study made in Poznan showing a negative association between lithium’s prophylactic efficacy and predisposition to psychotic symptoms (Dembińska-Krajewska et al., Reference Dembińska-Krajewska, Kliwicki, Chłopocka-Woźniak and Rybakowski2012). It was also demonstrated that a PRS for recurrent depression negatively correlates with lithium’s prophylactic efficacy (Amare et al., Reference Amare, Schubert, Hou, Clark, Papiol, Cearns, Heilbronner, Degenhardt, Tekola-Ayele and Hsu2021) as well as a PRS for attention deficit hyperactivity disorder (Coombes et al., Reference Coombes, Millischer, Batzler, Larrabee, Hou, Papiol, Heilbronner, Adli, Akiyama, Akula, Amare, Ardau, Arias, Aubry, Backlund, Bauer, Baune, Bellivier, Benabarre, Bengesser, Bhattacharjee, Cervantes, Chen, Chillotti, Cichon, Clark, Colom, Cruceanu, Czerski, Dalkner, Degenhardt, Del Zompo, DePaulo, Étain, Falkai, Ferensztajn-Rochowiak, Forstner, Frisen, Gard, Garnham, Goes, Grigoroiu-Serbanescu, Grof, Hashimoto, Hauser, Herms, Hoffmann, Jamain, Jiménez, Kahn, Kassem, Kato, Kelsoe, Kittel-Schneider, König, Kuo, Kusumi, Laje, Landén, Lavebratt, Leboyer, Leckband, Maj, Manchia, Martinsson, McCarthy, McElroy, Mitchell, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, O’Donovan, Osby, Ozaki, Pfennig, Pisanu, Potash, Reif, Reininghaus, Rietschel, Rouleau, Rybakowski, Schalling, Schofield, Schubert, Schweizer, Severino, Shekhtman, Shilling, Shimoda, Simhandl, Slaney, Squassina, Stamm, Stopkova, Tortorella, Turecki, Vieta, Witt, Zandi, Fullerton, Alda, Frye, Schulze, McMahon and Biernacka2021).
The study on the association of lithium’s prophylactic efficacy with human leukocyte antigen (HLA) genes made by ConLiGen showed that good efficacy was associated with HLA genes linked to lower inflammation. In contrast, a poorer response was connected with HLA genes linked to higher inflammation (Le Clerc et al., Reference Le Clerc, Lombardi, Baune, Amare, Schubert, Hou, Clark, Papiol, Cearns, Heilbronner, Degenhardt, Tekola-Ayele, Hsu, Shekhtman, Adli, Akula, Akiyama, Ardau, Arias, Aubry, Backlund, Bhattacharjee, Bellivier, Benabarre, Bengesser, Biernacka, Birner, Brichant-Petitjean, Cervantes, Chen, Chillotti, Cichon, Cruceanu, Czerski, Dalkner, Dayer, Del Zompo, DePaulo, Étain, Jamain, Falkai, Forstner, Frisen, Frye, Fullerton, Gard, Garnham, Goes, Grigoroiu-Serbanescu, Grof, Hashimoto, Hauser, Herms, Hoffmann, Jiménez, Kahn, Kassem, Kuo, Kato, Kelsoe, Kittel-Schneider, Ferensztajn-Rochowiak, König, Kusumi, Laje, Landén, Lavebratt, Leckband, Tortorella, Manchia, Martinsson, McCarthy, McElroy, Colom, Millischer, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, O’Donovan, Ozaki, Ösby, Pfennig, Potash, Reif, Reininghaus, Rouleau, Rybakowski, Schalling, Schofield, Schweizer, Severino, Shilling, Shimoda, Simhandl, Slaney, Pisanu, Squassina, Stamm, Stopkova, Maj, Turecki, Vieta, Veeh, Witt, Wright, Zandi, Mitchell, Bauer, Alda, Rietschel, McMahon, Schulze, Spadoni, Boukouaci, Richard, Le Corvoisier, Barrau, Zagury, Leboyer and Tamouza2021). In another study, 36 candidate genes associated with a good prophylactic response to lithium were identified, primarily involving the cholinergic and glutamatergic systems (Amare et al., Reference Amare, Thalamuthu, Schubert, Fullerton, Ahmed, Hartmann, Papiol, Heilbronner, Degenhardt, Tekola-Ayele, Hou, Hsu, Shekhtman, Adli, Akula, Akiyama, Ardau, Arias, Aubry, Hasler, Richard-Lepouriel, Perroud, Backlund, Bhattacharjee, Bellivier, Benabarre, Bengesser, Biernacka, Birner, Marie-Claire, Cervantes, Chen, Chillotti, Cichon, Cruceanu, Czerski, Dalkner, Del Zompo, DePaulo, Étain, Jamain, Falkai, Forstner, Frisen, Frye, Gard, Garnham, Goes, Grigoroiu-Serbanescu, Fallgatter, Stegmaier, Ethofer, Biere, Petrova, Schuster, Adorjan, Budde, Heilbronner, Kalman, Kohshour, Reich-Erkelenz, Schaupp, Schulte, Senner, Vogl, Anghelescu, Arolt, Dannlowski, Dietrich, Figge, Jäger, Lang, Juckel, Konrad, Reimer, Schmauß, Schmitt, Spitzer, von Hagen, Wiltfang, Zimmermann, Andlauer, Fischer, Bermpohl, Ritter, Matura, Gryaznova, Falkenberg, Yildiz, Kircher, Schmidt, Koch, Gade, Trost, Haussleiter, Lambert, Rohenkohl, Kraft, Grof, Hashimoto, Hauser, Herms, Hoffmann, Jiménez, Kahn, Kassem, Kuo, Kato, Kelsoe, Kittel-Schneider, Ferensztajn-Rochowiak, König, Kusumi, Laje, Landén, Lavebratt, Leboyer, Leckband, Tortorella, Manchia, Martinsson, McCarthy, McElroy, Colom, Millischer, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, O’Donovan, Ozaki, Pfennig, Pisanu, Potash, Reif, Reininghaus, Rouleau, Rybakowski, Schalling, Schofield, Schweizer, Severino, Shilling, Shimoda, Simhandl, Slaney, Squassina, Stamm, Stopkova, Maj, Turecki, Vieta, Veeh, Witt, Wright, Zandi, Mitchell, Bauer, Alda, Rietschel, McMahon, Schulze, Clark and Baune2023). A joint study of ConLiGen and the Pharmacogenomics of Bipolar Disorder group found a link between lithium’s prophylactic efficacy and the function of focal adhesions, which are important for neuronal growth, and the PI3K-Akt signalling pathway, which is related to cell proliferation and growth (Ou et al., Reference Ou, Rosenthal, Adli, Akiyama, Akula, Alda, Amare, Ardau, Arias, Aubry, Backlund, Bauer, Baune, Bellivier, Benabarre, Bengesser, Bhattacharjee, Biernacka, Cervantes, Chen, Chen, Chillotti, Cichon, Clark, Colom, Cousins, Cruceanu, Czerski, Dantas, Dayer, Del Zompo, Degenhardt, DePaulo, Étain, Falkai, Fellendorf, Ferensztajn-Rochowiak, Forstner, Frisén, Frye, Fullerton, Gard, Garnham, Goes, Grigoroiu-Serbanescu, Grof, Gruber, Hashimoto, Hauser, Heilbronner, Herms, Hoffmann, Hofmann, Hou, Jamain, Jiménez, Kahn, Kassem, Kato, Kittel-Schneider, König, Kuo, Kusumi, Lackner, Laje, Landén, Lavebratt, Leboyer, Leckband, Jaramillo, MacQueen, Maj, Manchia, Marie-Claire, Martinsson, Mattheisen, McCarthy, McElroy, McMahon, Mitchell, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, Ösby, Ozaki, Papiol, Perlis, Pisanu, Potash, Pfennig, Reich-Erkelenz, Reif, Reininghaus, Rietschel, Rouleau, Rybakowski, Schalling, Schofield, Schubert, Schulze, Schweizer, Seemüller, Severino, Shekhtman, Shilling, Shimoda, Simhandl, Slaney, Squassina, Stamm, Stopkova, Tighe, Tortorella, Turecki, Vieta, Volkert, Witt, Wray, Wright, Young, Zandi and Kelsoe2024).
Nunes et al. (Reference Nunes, Stone, Ardau, Berghöfer, Bocchetta, Chillotti, Deiana, Degenhardt, Forstner, Garnham, Grof, Hajek, Manchia, Mattheisen, McMahon, Müller-Oerlinghausen, Nöthen, Pinna, Pisanu, O’Donovan, Rietschel, Rouleau, Schulze, Severino, Slaney, Squassina, Suwalska, Turecki, Uher, Zvolsky, Cervantes, Del Zompo, Grof, Rybakowski, Tondo, Trappenberg and Alda2021) investigated the genetics of a ‘clinical exemplar score’ which predicts LR versus LNR, using data from seven collaborating centres. Clinically, the LR exemplars had a later illness onset, episodic course, no rapid cycling or psychosis, and minimal psychiatric comorbidities. Their genetic predictors included variants in the Alzheimer’s amyloid-secretase pathway, G protein-coupled receptors, muscarinic acetylcholine receptor signalling, and histamine H1 receptor signalling pathways.
The biomarkers associated with lithium’s beneficial effects could also be identified through epigenetic studies. A genome-wide methylomic approach (SeqCapEpi) revealed significant methylation differences between LRs and LNRs (Marie-Claire et al., Reference Marie-Claire, Lejeune, Mundwiller, Ulveling, Moszer, Bellivier and Etain2020). Combining clinical and epigenetic markers enabled the identification of a lithium response in 86% of cases (Marie-Claire et al., Reference Marie-Claire, Courtin, Bellivier, Gard, Leboyer, Scott and Etain2023). Spanish investigators found 130 differentially methylated positions and 16 differentially methylated regions between LR and LNR. Among them, 122 genes were identified, and those related to the GSK3ß pathway were highly represented (Zafrilla-Lopez et al., Reference Zafrilla-López, Acosta-Díez, Mitjans, Giménez-Palomo, Saiz, Barrot-Feixat, Jiménez, Papiol, Ruiz, Gavín, García-Portilla, González-Blanco, Bobes, Schulze, Vieta, Benabarre and Arias2024).
Studies on lithium related to new pathogenic hypotheses of bipolar disorder
The approach to studying lithium’s mechanisms in the last two decades was also directed towards a relationship with new pathogenic hypotheses of BDs. Some of the findings were already reported in in vitro studies and genetic research.
An increasing number of studies suggest that lithium may act through immunomodulatory and anti-inflammatory mechanisms, which are related to the hypothesised role of low-grade inflammation in BD. This phenomenon is supported by previously mentioned in vitro research (Cattane et al., Reference Cattane, Courtin, Mombelli, Maj, Mora, Etain, Bellivier, Marie-Claire and Cattaneo2022; Mizrahi et al., Reference Mizrahi, Choudhary, Ofer, Goldberg, Milanesi, Kelsoe, Gurwitz, Alda, Gage and Stern2023; Wu et al., Reference Wu, Lee, Wu, Sun, Chang, Chen, Chen, Wu and Cheng2019) and genetic studies (Le Clerc et al., Reference Le Clerc, Lombardi, Baune, Amare, Schubert, Hou, Clark, Papiol, Cearns, Heilbronner, Degenhardt, Tekola-Ayele, Hsu, Shekhtman, Adli, Akula, Akiyama, Ardau, Arias, Aubry, Backlund, Bhattacharjee, Bellivier, Benabarre, Bengesser, Biernacka, Birner, Brichant-Petitjean, Cervantes, Chen, Chillotti, Cichon, Cruceanu, Czerski, Dalkner, Dayer, Del Zompo, DePaulo, Étain, Jamain, Falkai, Forstner, Frisen, Frye, Fullerton, Gard, Garnham, Goes, Grigoroiu-Serbanescu, Grof, Hashimoto, Hauser, Herms, Hoffmann, Jiménez, Kahn, Kassem, Kuo, Kato, Kelsoe, Kittel-Schneider, Ferensztajn-Rochowiak, König, Kusumi, Laje, Landén, Lavebratt, Leckband, Tortorella, Manchia, Martinsson, McCarthy, McElroy, Colom, Millischer, Mitjans, Mondimore, Monteleone, Nievergelt, Nöthen, Novák, O’Donovan, Ozaki, Ösby, Pfennig, Potash, Reif, Reininghaus, Rouleau, Rybakowski, Schalling, Schofield, Schweizer, Severino, Shilling, Shimoda, Simhandl, Slaney, Pisanu, Squassina, Stamm, Stopkova, Maj, Turecki, Vieta, Veeh, Witt, Wright, Zandi, Mitchell, Bauer, Alda, Rietschel, McMahon, Schulze, Spadoni, Boukouaci, Richard, Le Corvoisier, Barrau, Zagury, Leboyer and Tamouza2021). The immunomodulatory and anti-inflammatory mechanisms of lithium have been the subject of reviews in which the documentation was given for the reduction by lithium of proinflammatory cytokines such as interleukin (IL)-1β, interferon-γ, tumour necrosis factor, IL-6, and monocyte chemoattractant protein-1, as well as the enhancement of anti-inflammatory cytokine IL-10 (Sakrajda & Szczepankiewicz, Reference Sakrajda and Szczepankiewicz2021; Szałach et al., Reference Szałach, Lisowska, Cubała, Barbuti and Perugi2023).
The action of lithium on the inflammatory system might be related to its inhibition of GSK-3β, as this enzyme is an important regulator in NOD-like receptor protein 3 inflammasome stimulation, thereby affecting the production of inflammatory cytokines (Arumugam et al., Reference Arumugam, Qin, Liang, Han, Boodapati, Li, Lu, Flavell, Mehal and Ouyang2022). The most recent molecular and genetic study evaluating the association between a polygenic score for immune system genes and lithium’s prophylactic effect revealed an overrepresentation of genes involved in cell adhesion and intercellular communication, particularly those associated with GSK-3β inhibition. In contrast, the relationship with immune system genes was weak to moderate (Herrera-Rivero et al., Reference Herrera-Rivero, Gutiérrez-Fragoso, Kurtz and Baune2024).
A pathogenic theory of BD related to mitochondrial dysfunction and oxidative stress has also been postulated in recent years (Lam et al., Reference Lam, Xu, Yeo, Cheah and Ling2023; Madireddy & Madireddy, Reference Madireddy and Madireddy2022). In the frontal cortex of postmortem BD subjects, lithium was found to influence the electron transport chain (ETC) and oxidative phosphorylation (OXPHOS) pathways, thereby enhancing OXPHOS (Maurer et al., Reference Maurer, Schippel and Volz2009). In an in vitro study, chronic lithium treatment of human neurones (SH-SY5Y) resulted in increased resistance to oxidative stress, enhanced glucose consumption, and increased glycolytic activity (Nciri et al., Reference Nciri, Desmoulin, Allagui, Murat, Feki, Vincent and Croute2013). In a clinical research of patients with BD, lithium treatment increased OXPHOS, which correlated with plasma lithium levels (De Sousa et al., Reference De Sousa, Streck, Zanetti, Ferreira, Diniz, Brunoni, Busatto, Gattaz and Machado-Vieira2015). An analysis of RNA-seq data from lithium-treated BD patients showed an overrepresentation of genes involved in mitochondrial function (ETC and OXPHOS-related genes) and downregulation of their expression in LR compared to LNR (Stacey et al., Reference Stacey, Schubert, Clark, Amare, Milanesi, Maj, Leckband, Shekhtman, Kelsoe, Gurwitz and Baune2018). It is postulated that lithium’s influence on mitochondria involves inhibition of GSK-3ß, as that enzyme regulates various mitochondrial functions (Yang et al., Reference Yang, Chen, Gao, Shi, Li, Jiang, Hu, Liu, Xu and Wu2017).
The disruption of biological rhythms in BD has been well documented, with most studies indicating a predominance of an evening chronotype in these patients (Freund & Haussleiter, Reference Freund and Haussleiter2023). Research in this area, including studies using skin fibroblast models, indicates that LRs tend to have a higher level of morningness chronotype (McCarthy et al., Reference McCarthy, Wei, Nievergelt, Stautland, Maihofer, Welsh, Shilling, Alda, Alliey-Rodriguez, Anand, Andreasson, Balaraman, Berrettini, Bertram, Brennand, Calabrese, Calkin, Claasen, Conroy, Coryell, Craig, D’Arcangelo, Demodena, Djurovic, Feeder, Fisher, Frazier, Frye, Gage, Gao, Garnham, Gershon, Glazer, Goes, Goto, Harrington, Jakobsen, Kamali, Karberg, Kelly, Leckband, Lohoff, McInnis, Mondimore, Morken, Nurnberger, Obral, Oedegaard, Ortiz, Ritchey, Ryan, Schinagle, Schoeyen, Schwebel, Shaw, Shekhtman, Slaney, Stapp, Szelinger, Tarwater, Zandi and Kelsoe2019), and lithium treatment normalises circadian rhythm disturbances (Federoff et al., Reference Federoff, McCarthy, Anand, Berrettini, Bertram, Bhattacharjee, Calkin, Conroy, Coryell, D’Arcangelo, DeModena, Fisher, Feeder, Frazier, Frye, Gao, Garnham, Gershon, Alliey-Rodriguez, Glazer, Goes, Karberg, Harrington, Jakobsen, Kamali, Kelly, Leckband, Lohoff, Maihofer, McInnis, Mondimore, Morken, Nurnberger, Oedegaard, Ritchey, Ryan, Schinagle, Schoeyen, Schwebel, Shaw, Shilling, Slaney, Stautland, Tarwater, Calabrese, Alda, Nievergelt, Zandi and Kelsoe2022). This aligns with findings from Poznan and Kraków centres, where BD patients on lithium showed a tendency towards a morningness chronotype (Dopierała et al., Reference Dopierała, Chrobak, Tereszko and Rybakowski2017). Mishra et al. (Reference Mishra, Wei, Rohr, Ko, Nievergelt, Maihofer, Shilling, Alda, Berrettini, Brennand, Calabrese, Coryell, Frye, Gage, Gershon, McInnis, Nurnberger, Oedegaard, Zandi, Kelsoe and McCarthy2023) demonstrated differences in the expression of circadian clock genes between BD patients and healthy individuals, as well as between patients with different lithium responses (LR vs. LNR) using a fibroblast model. The most significant differences involved the Per1 (clock protein PERIOD 1) and Per3 genes. Recently, it was found that a higher genetic predisposition for morningness may increase the chance of a lithium response in BD patients (Zafrilla-Lopez et al., Reference Zafrilla-López, Herrera-Escartín, Guardiola-Ripoll, García-Fernández, Acosta-Díez, Giménez-Palomo, Saiz, García-Portilla, Jiménez, Papiol, Fañanás, Ruiz, Gavín, González-Blanco, Vieta, Schulze, Fatjó-Vilas, Benabarre, Mitjans and Arias2026).
BDNF is essential for neuronal function and survival, the regulation of many neurotransmitters, and the pathogenesis and treatment of mood disorders (Hashimoto et al., Reference Hashimoto2010). The association of the Val66Met polymorphism in the BDNF gene with BD has been demonstrated (Li et al., Reference Li, Chang and Xiao2016). Serum BDNF concentrations are reduced during episodes of both mania and depression (Fernandes et al., Reference Fernandes, Molendijk, Köhler, Soares, Leite, Machado-Vieira, Ribeiro, Silva, Sales, Quevedo, Oertel-Knöchel, Vieta, González-Pinto, Berk and Carvalho2015). Lithium increases BDNF expression in the rat brain by stimulation of CREB, which activates the BDNF gene expression (Fukumoto et al., Reference Fukumoto, Morinobu, Okamoto, Kagaya and Yamawaki2001). A low concentration of BDNF is also regarded as a marker of the late stage of BD (Kauer-Sant’Anna et al., Reference Kauer-Sant’Anna, Kapczinski, Andreazza, Bond, Lam, Young and Yatham2009); however, in excellent LRs, despite a long duration of BD (20 years or more), serum BDNF was similar to that of healthy subjects, supporting the stimulating effect of lithium on BDNF expression and maintaining a proper BDNF level (Rybakowski & Suwalska, Reference Rybakowski and Suwalska2010). We also showed that the prophylactic effect of lithium in BD is associated with the Val66Met polymorphism of the BDNF gene (Rybakowski et al., Reference Rybakowski, Suwalska, Skibinska, Szczepankiewicz, Leszczynska-Rodziewicz, Permoda, Czerski and Hauser2005).
Telomere length is considered a marker of ageing processes. GWAS studies have shown that this phenomenon is genetically determined, with the strongest association with the leucine-rich repeat gene (LRRC34) (Coutts, Reference Coutts, Palmos, Duarte, de Jong, Lewis, Dima and Powell2019). Several studies reported shorter telomeres in BD patients and/or their restoration to normal length with lithium treatment (Ferensztajn-Rochowiak et al., Reference Ferensztajn-Rochowiak, Kurczewska, Rubiś, Lulkiewicz, Hołysz, Rybakowski and Rybakowski2021; Pisanu et al., Reference Pisanu, Congiu, Manchia, Caria, Cocco, Dettori, Frau, Manca, Meloni, Nieddu, Noli, Pinna, Robledo, Sogos, Ferri, Carpiniello, Vanni, Bocchetta, Severino, Ardau, Chillotti, Zompo and Squassina2020). It was found that the duration of lithium use correlates with the expression of the telomerase reverse transcriptase enzyme (Lundberg et al., Reference Lundberg, Biernacka, Lavebratt, Druliner, Ryu, Geske, Colby, Boardman, Frye and Schalling2020). However, the latest study involving 591 patients on lithium did not show a correlation between the duration of lithium use and ageing markers, including telomere length (Mutz et al., Reference Mutz, Wong, Powell, Young, Dawe and Lewis2024).
The old purinergic hypothesis of mood disorders was linked to uric acid, while the new one is associated with the purinergic signalling system. This system contains two classes of receptors – P1 receptors, responsive to adenosine, and P2 receptors, responsive to ATP. In patients with BD, adenosine concentration is lower compared to the control group, and there is a negative correlation between adenosine concentration and depression severity (Gubert et al., Reference Gubert, Jacintho Moritz, Vasconcelos-Moreno, Quadros dos Santos, Sartori, Fijtman, Kauer-Sant’Anna, Kapczinski, Battastini and Magalhães PVda2016). In an experimental study, lithium was found to increase adenosine concentrations by inhibiting ectonucleotidase (Wilot et al., Reference Wilot, Da Silva, Ferreira, Bonan, Sarkis, Rocha and Battastini2004). Recently, the P2X7 receptor has emerged as the most promising candidate in the pathogenesis of BD (Gonçalves et al., Reference Gonçalves, Andrejew and Gubert2022). In the study by Wilot et al. (Reference Wilot, Bernardi, Frozza, Marques, Cimarosti, Salbego, Rocha and Battastini2007), the neuroprotective effect of lithium against ATP-induced cell death in rat hippocampal slices was mediated by the P2X7 receptor. Although there have been few findings on the mechanism of lithium in purinergic signalling related to mood disorders, they may be more closely linked to its neuroprotective potential, which will be discussed in a subsequent subchapter.
The mechanisms of lithium’s neuroprotective and anti-dementia activity
So far, the history of lithium mechanisms described here has been connected with the therapeutic effect of this ion in mood disorders. However, in recent decades, there has also been increasing interest in the mechanisms underlying lithium’s neuroprotective properties and possible therapeutic activity in dementia. As a neuroprotective effect, lithium has been observed to increase human grey matter in the prefrontal cortex, anterior cingulate, and hippocampus (Hajek & Weiner, Reference Hajek and Weiner2016; Moore et al., Reference Moore, Bebchuk, Wilds, Chen and Manji2000). Lithium has also been shown to reduce the risk of stroke and, in animal models, to produce significant benefits in such conditions (Almeida et al., Reference Almeida, Singulani, Ford, Hackett, Etherton-Beer, Flicker, Hankey, De Paula, Penteado and Forlenza2022). The population studies have shown that lithium treatment results in a reduction of the risk of dementia and dementia severity (Donix & Bauer, Reference Donix and Bauer2016). In patients with BD, long-term treatment with lithium was associated with a diminished rate of dementia, in contrast to such treatment with anticonvulsants, antidepressants, and antipsychotics (Kessing et al., Reference Kessing, Forman and Andersen2010). In patients with mild cognitive impairment (MCI) and AD, some favourable effects of lithium on cognition were found that correlated with treatment duration (Matsunaga et al., Reference Matsunaga, Kishi, Annas, Basun, Hampel and Iwata2015, Reference Matsunaga, Fujishiro and Takechi2019). Japanese authors analysed trials with 6,547 AD participants and concluded that low-dose lithium was more effective and better tolerated than aducanumab (Terao & Kodama, Reference Terao and Kodama2024). However, more recent evidence indicates that, despite a satisfactory safety profile, lithium does not consistently translate into measurable cognitive or functional improvement in AD dementia (Pereira da Silva et al., Reference Pereira da Silva, de Deus, Ribeiro, Tudella, Cabeça, Silva, Han, Franco, Costa, do Nascimento MDV, Andrade Fernandes, Franco and de Sousa Maia2026).
The neuroprotective potential of lithium may be linked to its effects on purinergic signalling. Chronic lithium treatment inhibits ATP-induced cellular death by decreasing ATP and increasing adenosine levels (Gubert et al., Reference Gubert, Andrejew, Figueiro, Bergamin, Kapczinski, Magalhães PVda and Battastini2021). For potential neuroprotective mechanisms of lithium in ischaemic stroke, its effects on GSK3-ß, the Wnt signalling pathway, mitogen-activated protein kinase, extracellular signal-regulated kinase, BDNF, mammalian target of rapamycin, and glutamatergic receptors have been proposed (Wang et al., Reference Wang, Lu, Shi and Wang2024).
There is some evidence linking dementia with decreased lithium concentration. Aron et al. (Reference Aron, Ngian, Qiu, Choi, Liang, Drake, Hamplova, Lacey, Roche, Yuan, Hazaveh, Lee, Bennett and Yankner2025) showed that the endogenous levels of lithium in the prefrontal cortex were significantly reduced in subjects with both MCI and AD. Also, a negative correlation between dementia occurrence and lithium concentration in drinking water was demonstrated (Fajardo et al., Reference Fajardo, Fajardo, LeBlanc and MacPherson2018; Fraiha-Pragado et al., Reference Fraiha-Pegado, de Paula, Alotaibi, Forlenza and Hajek2024; Kessing et al., Reference Kessing, Gerds, Knudsen, Jørgensen, Kristiansen, Voutchkova, Ernstsen, Schullehner, Hansen, Andersen and Ersbøll2017). Among biochemical mechanisms underlying lithium’s favourable effects in dementia, inhibition of GSK-3ß plays a dominant role, primarily by reducing tau phosphorylation and amyloid deposition in the brain (Algazzawi et al., Reference Algazzawi, Abujamai, Alshanberi, Satar and Ansari2026; Shen et al., Reference Shen, Zhao, Zhao, Meng, Zhang, Zhang, Taishi and Sun2024; Xiao et al., Reference Xiao, Sayed, Xing, Zhang, Ai, Kirichek, Le, Wong, Valentino, Teopiz, Vinberg, Rosenblat, Lo, Zhang and McIntyre2026). An experimental study showed that low doses of lithium inhibited GSK3-ß in the prefrontal cortex of C57BL/6J mice, which may be related to the clinical therapeutic effect of low lithium doses in AD (Fenech et al., Reference Fenech, Hamstra, Finch, Ryan, Marko, Roy, Fajardo and MacPherson2023). Other mechanisms may involve lithium’s effects on autophagy, inflammation, and oxidative stress. In addition, lithium’s inhibition of herpesviruses may play a role, as this infection has been identified as an important pathogenic factor in AD (Itzhaki, Reference Itzhaki2018; Rybakowski, Reference Rybakowski2019).
Concluding remarks
This historical overview highlights the multifaceted mechanisms of lithium action, especially those related to nervous system function. As to the mechanisms of lithium’s therapeutic (mostly prophylactic) efficacy in mood disorders, especially BD, the subjects of mechanisms have progressed from the uric acid concept in the 19th century through various biochemical and pharmacological theories in the second part of the 20th century up to the 21st century findings obtained from in vitro and genetic studies as well as related to new pathogenic hypotheses of BD.
For the mechanisms of lithium in mood disorders, two types of data can be distinguished. The first is connected with the pathogenesis of mood disorders, mainly BD. In this matter, the lithium mechanisms of action may correct the pathophysiological abnormality underlying the illness. Such an appearance may be associated with positively influencing (e.g. abnormal ATPase activity, low-grade inflammation, mitochondrial disturbance, or dysfunction of biological rhythms). This can also be connected with genes underlying BD and identified as playing a role in lithium therapeutic activity (e.g. CACNG2 and BDNF genes). The second topic is the possibility of applying the lithium mechanism to the quality of its prophylactic effect in BD. In this respect, differences were identified between patients with a favourable lithium prophylactic effect (LRs) and those with a poor effect (LNRs). A large proportion of these findings has been obtained by demonstrating molecular and genetic distinctiveness using various research methods. They will complement the clinical predictors of lithium response, which remain the main factors used in practice.
It should be noted that the 21st-century studies have shown that inhibition of GSK-3ß and the PI system plays a significant role in many of lithium’s mechanisms of action in mood disorders, as postulated primarily in the 1980s and 1990s. Also, in studies on the mechanisms of lithium’s neuroprotective and anti-dementia effects performed over the past two decades, inhibition of GSK-3ß stands out. Thus, it may be hypothesised that these processes have remained the most crucial for the therapeutic effects of this ion (Brown & Tracy, Reference Brown and Tracy2013).
Open access
