Long-term use of mood-stabilizing drugs has long been a cornerstone of clinical treatment for bipolar disorder (BD). Lithium, which has been in use longer than any other, can be highly effective. In particular it can prevent or reduce the intensity of episodes of mania. Symptoms of mania include hyperactivity, euphoria or highly elevated mood, rushed speech, poor judgment, reduced need for sleep, aggression, and anger.

But lithium doesn’t help every patient; in fact, only about 1 patient in 3 responds to it. Among those who do, an important added benefit is that suicide and overall mortality rates are significantly reduced. But toxicity has been associated with lithium administration over the long-term, and its use has been replaced in some patients with drugs not originally approved to treat bipolar disorder, including anticonvulsants such as valproate and lamotrigine.

Given lithium’s well-documented ability to be of great benefit to a sizeable subset of BD patients, researchers have taken great pains to discover why it works for them: how the drug affects the central nervous system, at the level of cells, networks, and circuits. The answer has remained elusive, but new research co-led by BBRF grantees and published in the journal Lancet not only sheds light on the drug’s mechanism of action, but also points to novel therapeutic approaches for patients, including those who do not respond to lithium.

2022 BBRF Young Investigator Anouar Khayachi, Ph.D., of McGill University, Canada, is first author of the new paper, and part of a Canadian team that includes co-leaders Guy A. Rouleau, M.D., Ph.D., a 2010 BBRF Distinguished Investigator, Austen J. Milnerwood, Ph.D., and Martin Alda, M.D., FRCPC, 2020 BBRF Colvin Prize winner and a 2003 and 1999 BBRF Independent Investigator.

Among the many insights about lithium’s mechanism of action is a recent finding by members of the Canadian team that in neurons of the mouse cerebral cortex, long-term exposure to lithium decreased the flow of sodium ions, reduced firing rates, and lowered the range of calcium ion levels. Ions are atoms that carry an electric charge; their flow into and out of neurons is one of the ways in which the activation of brain cells is regulated.

In the current research, the team used a stem cell-based technology called iPSC (induced pluripotent stem cell). Cells—in this case, blood cells—are harmlessly sampled from individuals both with the illness under study as well as healthy controls. In the lab, these cells are gown in culture and brought back to a stem cell-like state, then re-programmed to re-develop as neurons. The team grew cultures of this kind from 5 BD patients who were responsive to lithium, 4 who were not responders, and 5 age and sex-matched healthy controls (all participants in the study were male).

The point of such research is to discover processes in cells from patients that differ from those in cells from controls. In this case, the BD patient-derived neurons enabled further comparison—between cells derived from patients who were lithium responders and those who weren’t.

One important observation made in prior work by the team as well as other researchers was replicated in this work: all of the patient-derived neurons, regardless of the donors’ lithium response status, displayed hyperexcitability. When these cells were treated with lithium over 7 days, the hyperactivity was reversed in cells derived from lithium responders, but, as expected, it was not reversed in cells from lithium non-responders.

The next step was to conduct many different kinds of tests on these three sets of neurons grown in culture dishes. Many things were revealed. First: In cells grown from lithium responders—in which hyperactivity was reversed with lithium treatment—the team noted changes in the ability of positively charged sodium ions to flow into and out of the cells, relative to cells grown from lithium non-responders.

Additional experiments that included analyses of protein activity and gene expression revealed that the potentially therapeutic effect of lithium on neurons derived from lithium responders was associated with a specific intracellular signaling pathway, called the Akt signaling pathway. Neurons (and other cells) regulate their survival and growth, in part, via this important pathway.

Additional experiments demonstrated that a compound that activates the Akt pathway mimicked the effect that lithium has on neurons—it reverses their hyperexcitability—but only in neurons grown from patients who responded to lithium, not in those from non-responders. Among the implications of this discovery is that it may make sense to develop and test Akt pathway activators to treat mania in bipolar patients. If the activity of such an agent was as therapeutically beneficial as that of lithium, and the agent was found to be less toxic or have fewer long-term side effects, it might be considered as a replacement for lithium.

Another key finding from the team’s experiments also has therapeutic implications. In all BD patient-derived neurons grown in culture—neurons from both lithium responders and non-responders—activation of a protein complex called AMPK (AMP-activated protein kinase) reduced heightened neural network activity that seems to be characteristic in BD. AMPK is an energy sensor inside of cells, a major cellular regulator of lipid and glucose metabolism. Targeting AMPK in neurons might be a strategy to address neuropathology in lithium non-responders and responders alike.

One approved AMPK activator, metformin, is taken by millions of people. People with BD have a 2-fold increased risk of type II diabetes, and insulin resistance is present, the team notes, in about 50% of patients, “which might correlate with disease severity/progression.” There has been some suggestion that lithium exerts therapeutic effects in BD via its impact on insulin signaling. Akt and AMPK are both also involved in insulin signaling and the development of insulin resistance. One study has found that metformin improved clinical outcomes in BD patients, not only lowering insulin resistance but improving mood symptoms as well. This preliminary finding “and the results of our work here support use of AMPK activation for BD,” the team said, although, of course, this will remain a hypothesis until considerable additional research is performed.

The team noted the main limitation of their study. Studying cells grown in culture in the lab, even those derived from BD patients, are by definition only suggestive, as they were conducted outside the context of the full human (or any animal) system. Reversal of hyperactivation in neurons derived from patients needs to be tested in living organisms—but at this point, no animal model is available to test it.

Nevertheless, the reported experiments suggest specific alterative strategies for treating BD, and provide “a framework for a personalized drug screening platform to accelerate development of alternative therapeutic strategies for BD,” the team wrote. Personalized medicine for BD, if realized, could potentially address the frequent lag between diagnosis and therapy selection, as well as significantly reduce the risk of suicide, they said.