Email Icon Facebook icon Twitter Icon GooglePlus Icon Contact

User Top Menu

Patient-Derived Bipolar Neurons Reproduce Clinical Response to Lithium

2 Nov 2015

November 3, 2015. Lithium is effective at calming the mania associated with bipolar disorder—when it works. For reasons that are not understood, a significant number of patients fail to respond to lithium. Now, using patients' skin cells and induced pluripotent stem cell (iPSC) technology, researchers have created a cell-based model of bipolar disorder that displays a distinct neuronal hyperexcitability phenotype with a variable response to lithium. Cells derived from lithium-sensitive patients show a reversal of hyperexcitability upon lithium treatment, while cells from lithium-resistant patients show no change.

The work was published October 28 in Nature, from the labs of Jun Yao of Tsinghua University in Beijing, China; Fred Gage at the Salk Institute in La Jolla, California; and John Kelsoe of the University of California, San Diego.

"This work provides a model for the effect of lithium in bipolar disorder," said Melvin McInnis of the University of Michigan in Ann Arbor, who was not involved in the study. As such, it could reveal markers for lithium response, and could be useful for screening for new therapeutic agents.

The finding of hyperexcitability aligns with work McInnis and collaborator K. Sue O'Shea published recently, showing that patient-derived neurons display increased calcium mobilization, which can be reversed by lithium (Chen et al., 2014; O'Shea and McInnis, 2015). The results raise the possibility that neuronal firing characteristics could serve as a useful endophenotype for identifying novel therapeutics. "We want to have new medications or interventions that would model the same sort of behavior we see with lithium in these examples," said McInnis.

In the new study, first authors Jerome Mertens and Qiu-Wen Wang and colleagues started with fibroblasts from six clinically well-characterized patients with manic type 1 bipolar disorder (three lithium responders and three non-responders) and four unaffected subjects. The researchers first created iPSCs by cellular reprogramming. Then, they differentiated the cells to neurons using a protocol developed in the Gage lab that produces highly homogeneous cultures of hippocampal dentate gyrus-like neurons (Yu et al., 2014). Changes in hippocampal neurons have been implicated in bipolar disorder and other psychiatric diseases.

The neurons derived from bipolar cases were clearly different from those of unaffected controls. A comparison of gene expression using RNAseq in young neurons pointed to increases in genes for mitochondrial function and action potential firing. Patch-clamp recording revealed a cellular phenotype of hyperexcitablity, with the bipolar cells displaying higher rates of spontaneous firing and evoked action potentials, along with lower firing thresholds and higher action potential amplitude compared to unaffected cells. The results add up to "a physiological phenotype that can be detected in a cell-autonomous manner in vitro," Gage told SRF.

The clinical phenotype of lithium responsiveness also played out in the cell cultures. In cells from patients sensitive to lithium, one week of chronic application of the drug reduced the frequency of spontaneous firing and evoked action potentials, and also largely normalized gene expression. Cells from patients who did not respond to lithium showed neither effect. The resistant cells could be quieted, however, with lamotrigine, an anticonvulsant also effective in bipolar disorder.

A cellular model of drug responsiveness could offer a shortcut to the trial-and-error approach that sometimes takes years to settle on an effective regimen for bipolar patients. "One could imagine in the future having a simpler assay to predict if the patients might be responsive or nonresponsive, and even to screen in vitro to see what drugs they might respond to," said Gage. "The encouragement for us as cell and molecular biologists is that we have a detectable phenotype that allows us to begin to dissect this as an endophenotype at a cellular level."

Gage said he is continuing to work with clinical colleagues to extend the initial findings to more patients and different conditions. "We have a new cohort with double the number of patients from different locations, with different genetics and ethnicity, and we are starting with a different somatic cell type—we know how important it is to generalize these results."

In previous work, the Gage lab created the first reported neurons from schizophrenia patient cells, but did not detect electrophysiological changes compared to normal cells (see SRF related news report). However, a more recent study using the hippocampal neuron differentiation protocol did find a physiological phenotype in schizophrenia neurons, which appeared to be less excitable than neurons from unaffected subjects (Yu et al., 2014). That would make them different from the bipolar neurons, but Gage cautions against comparing the two types of cells in detail, because the experiments were done at different times, and the neuronal phenotypes are exquisitely sensitive to the conditions of differentiation and timing. He said his group is now doing a head-to-head experiment where one person will compare cells derived from patients with schizophrenia, bipolar disorder, depression, and autism at the same time to determine what similarities and differences might be detectable between diseases.—Pat McCaffrey.


Mertens J, Wang QW, Kim Y, Yu DX, Pham S, Yang B, Zheng Y, Diffenderfer KE, Zhang J, Soltani S, Eames T, Schafer ST, Boyer L, Marchetto MC, Nurnberger JI, Calabrese JR, Odegaard KJ, McCarthy MJ, Zandi PP, Alba M, Nievergelt CM, Pharmacogenomics of Bipolar Disorder Study, Mi S, Brennand KJ, Kelsoe JR, Gage FH, Yao J. Differential responses to lithium in hyperexcitable neurons from patients with bipolar disorder. Nature. 2015 Oct 28. Abstract