Scientists have zeroed in on a certain set of cortical neurons as instigators of a toxic circuit in a mouse model of TDP-43 proteinopathy. Somatostatin interneurons amp up their activity when the mice are just a few weeks old, resulting in hyperexcitability and excitotoxicity in downstream pyramidal neurons in layer 5 (L5PNs) of the cortex, report the authors of a study in the February 23 Nature Neuroscience online. The researchers fixed the problem by ablating the somatostatin interneurons. This suggests that if the same circuit were to go awry in people with TDP-43-based diseases, such as amyotrophic lateral sclerosis and frontotemporal dementia, then those neurons could be a possible target for therapeutics.

Out of control.

In mice expressing mutant TDP-43, hyperactive somatostatin neurons suppress the normal inhibition of layer 5 neurons by parvalbumin interneurons, causing hyperexcitability and excitotoxicity. [Courtesy of Yun Li.]

Excitotoxicity is a well-known culprit in the demise of motor neurons in ALS. In fact, the only FDA-approved treatment for the disease, riluzole, works by preventing the release of excess glutamate, which becomes excitotoxic. Glutamatergic layer 5 pyramidal neurons, which include upper motor neurons, transmit signals to the spinal cord and other subcortical regions. In the same region of the brain are two populations of interneurons, categorized by their expression of either somatostatin or parvalbumin. Joint senior authors Yun Li and Da-Ting Lin propose that somatostatin interneurons inhibit parvalbumin interneurons, which in turn temper the activity of pyramidal neurons. In toto, this system gives the brain a way to fine-tune motor activity. 

Li and Lin, a married couple, began their studies at the Jackson Laboratory in Bar Harbor, Maine, and continued them after moving to the National Institute on Drug Abuse in Baltimore. The scientists first got interested in the activity of cortical layer 5 pyramidal neurons (L5PNs) in TDP-43 mice when they used two-photon microscopy to image live neurons in the brain through a window in the skull (see Apr 2012 news series). They crossed mice expressing yellow fluorescent protein (YFP) in L5PNs with mice expressing the ALS-linked A315T variant of human TDP-43 in all neurons (see Oct 2009 news). By a few months of age, this well-known model of TDP-43 pathology starts to have trouble moving; death comes soon afterward. Compared with normal mice, these transgenic mice have fewer motor neurons at death, and the remaining ones contain ubiquitinated protein aggregates.

When study first author Lifeng Zhang peered into the brains of YFP/A315T double transgenics, he saw numerous blebs along the layer 5 dendrites in mice as young as six weeks—well before motor symptoms or ubiquitin inclusions arise. Li recognized this blebbing as a sign of excitotoxicity. Zhang and co-first author Wen Zhang (no relation) then looked to see what might cause this. First, they measured action potentials in brain slices from three-week-old mice. Already, pyramidal neurons were hyperactive, firing more often than the same neurons in brains from wild-type littermates. Notably, weaker inhibitory currents modulated the pyramidal neurons, while afferent excitatory currents were as strong as in slices from normal tissue. In essence, the neurons were getting the right amount of stimulation, but too little inhibition.

To pinpoint why, the authors investigated the somatostatin and parvalbumin interneurons as possible troublemakers. In brain slices from TDP-43 mice, somatostatin interneurons were hyperactive, while parvalbumin interneurons were hypoactive. If somatostatin interneurons inhibit parvalbumin interneurons, which in turn inhibit layer 5 cortical motor neurons, then hyperactive somatostatin interneurons would elicit more action from the L5PNs, the authors concluded (see image above). These motor neurons would then release excess glutamate, Lin said, which could come back to haunt them.

Zhang and Zhang tested this theory in two ways. First, they used optogenetics to stimulate or inhibit somatostatin interneurons in brain slices from wild-type mice, measuring the effects on pyramidal neurons. Sure enough, activating the somatostatin interneurons reduced the parvalbumin inhibitory currents coming into pyramidal cells, and this led to hyperexcitability and suppression of L5PNs. Shutting off the somatostatin interneurons had the opposite effect, tightening inhibition of the pyramidal neurons.

Second, the scientists tested whether ablating somatostatin interneurons in TDP-43 mice would revive pyramidal neurons. They expressed the receptor for diphtheria toxin in the mouse somatostatin interneurons, then injected the toxin into the cortex once the animals reached six weeks old. Two weeks later, they sacrificed the mice and tested brain slices. The inhibitory currents coming into the layer 5 pyramidal neurons were stronger than those in slices from untreated animals, if not completely back to wild-type levels. In another set of mice sacrificed six weeks after the diphtheria treatment, the authors found fewer ubiquitin aggregates, and more motor neurons, than in untreated TDP-43 mice.

The authors do not know if correcting the activity of their microcircuit could alleviate symptoms in people with ALS, or even in mouse models. The mice they studied were younger than when symptoms began in the TDP-43-A315T mice, and Li notice no major changes in their mobility. The authors were unable to assess effects on lifespan because these animals die of bowel blockage before the motor neuron disease becomes as severe as in human ALS (see Sep 2012 news).

For Li, the next question will be what makes the somatostatin interneurons at the top of the circuit hyperexcitable. She plans to sequence RNAs from those neurons to identify altered expression patterns caused by the mutant TDP-43.

Does the same microcircuit go awry in people with ALS or FTD? “It is too early to make conclusions about whether similar mechanisms are occurring in ALS,” commented Anna King of the University of Tasmania in Australia, who was not involved in the study (see full comment below). However, King noted that studies in people hint that dysfunction of interneuron populations could contribute to upper motor neuron hyperexcitability in ALS.

For example, transcranial magnetic stimulation (TMS) detects unusually high activity in the motor cortices of people with ALS, often well before the physical symptoms arise (see Sep 2015 news). Steve Vucic of the University of Sydney, who uses TMS to study and help diagnose ALS, said interneuron and layer 5 defects could neatly explain the early cortical hyperexcitability he observes in people. He added that the new findings highlight the importance of upper motor neurons, not just the spinal cord, in ALS (see Jan 2015 news; news). 

Vucic and King both think the new paper suggests that modulating the activity of the somatostatin-parvalbumin-L5PN microcircuit could eventually become a treatment for ALS. Lin has some ideas as to how one might do that. If scientists could identify a receptor or ion channel specific to somatostatin interneurons, he suggested, they could seek out small molecules that tune them down. Alternatively, he speculated that physicians could use TMS to focally tune the microcircuit, for example, by activating parvalbumin interneurons to turn down the L5PNs. While current TMS technology cannot activate neurons quite so specifically, Lin suspects that future improvements might make this feasible.—Amber Dance


  1. This article addresses an important area of ALS research that has been relatively neglected. There is substantial clinical evidence that dysfunction of interneuron populations could be a potential contributor to upper motor neuron hyperexcitability in ALS, which has been known for nearly 20 years. This has been demonstrated using paired pulse transcranial magnetic stimulation (TMS) techniques. For example, reduced short-interval intracortical inhibition (SICI) has been demonstrated in sporadic ALS as well familial disease, and in familial disease it occurs early, before symptom onset (Grieve et al., 2015; Vucic et al., 2008). There have been few follow-up studies in animal models to dissect out mechanisms or try to determine if a specific type of interneuron is affected, but very little work has been done to dissect out mechanisms of hyperexcitability in TDP-43 models. It is exciting to see progress in this area. In terms of FTD the contribution of hyperexcitability is far less clear and less work has been done in this area.

    This paper examines hyperexcitability in a mutant TDP-43 mouse model. It identifies dysfunction in a specific class of interneurons that are classified by their expression of somatostatin. Signs of excitotoxic injury, such as dendritic blebbing were present in layer 5 pyramidal neurons from early timepoints. This paper is elegant in that it not only shows reduced inhibitory input acting through parvalbumin interneurons, but also that ablating somatostatin immunoreactive interneurons prevents subsequent pathology and neuronal loss. Unlike many studies in ALS research, this paper focuses on the motor cortex rather than the spinal cord. It is currently unclear if pathology is driven from upper or lower motor neurons or if both occur independently, although cortical hyperexcitability is demonstrated in clinical studies.

    It is of course too early to make conclusions about whether similar mechanisms are occurring in human ALS. Considerable work needs to be done here. In human tissue there have been few histological studies looking at interneurons, particularly in recent times, and findings are in some cases contradictory. It will be important to follow up this study by looking at whether mutated TDP-43 expression directed toward specific cell types, such as interneuron populations, can alone drive pathology. And, while the authors identified a GABAergic mechanism, it still needs to be determined if somatostatin plays a role. Additionally, studies will need to determine whether this mechanism is specific to mutant TDP-43. In this regard, studies in other models (such as G93A mSOD1) and in iPSCs have suggested increased excitability in motor/pyramidal neurons.

    In summary, this paper supports a growing body of evidence that targeting excitability in ALS is an important avenue to pursue and offers hope that modulation of inhibitory networks may be a potential therapeutic option for treatment of this disease.

  2. This is a very exciting paper with remarkable implications. The key finding that alterations in somatostatin interneurons may be driving the degeneration of layer 5 cortical neurons is potentially paradigm-shifting. The authors were able to show that blocking the action of those inhibitory interneurons was sufficient to protect layer 5 cortical neurons from excitotoxic death. Further work will be needed to determine if a similar pathway occurs in other mouse models, and ultimately in humans where cortical hyperexcitability is also an early finding. But there is no doubt it brings to light that targeting specific interneuron populations may be therapeutically protective, a very exciting prospect.

Make a Comment

To make a comment you must login or register.


Therapeutics Citations

  1. Riluzole

News Citations

  1. Modern Microscopy Skims Surface of Living Minds and Spines
  2. Meet the First Published TDP-43 Mouse
  3. Are TDP-43 Mice Living Up to Expectations?
  4. Magnet Test Finds Cortex Overexcitable in All ALS
  5. Earliest ALS Defects Said to Start in Disparate Places
  6. Are Upper Motor Neuron Gaffes a Prelude to Disease?

Further Reading


  1. . Excitotoxicity in ALS: Overstimulation, or overreaction?. Exp Neurol. 2016 Jan;275 Pt 1:162-71. Epub 2015 Nov 13 PubMed.
  2. . Inhibitory dysfunction in amyotrophic lateral sclerosis: future therapeutic opportunities. Neurodegener Dis Manag. 2015 Nov 30; PubMed.
  3. . Early interneuron dysfunction in ALS: Insights from a mutant sod1 zebrafish model. Ann Neurol. 2013 Feb;73(2):246-58. PubMed.
  4. . Inhibitory synaptic regulation of motoneurons: a new target of disease mechanisms in amyotrophic lateral sclerosis. Mol Neurobiol. 2012 Feb;45(1):30-42. PubMed.
  5. . Cortical hyperexcitability may precede the onset of familial amyotrophic lateral sclerosis. Brain. 2008 Jun;131(Pt 6):1540-50. Epub 2008 May 9 PubMed.
  6. . Amyotrophic lateral sclerosis affects cortical and subcortical activity underlying motor inhibition and action monitoring. Hum Brain Mapp. 2015 Aug;36(8):2878-89. Epub 2015 Apr 24 PubMed.
  7. . Sensitivity and specificity of threshold tracking transcranial magnetic stimulation for diagnosis of amyotrophic lateral sclerosis: a prospective study. Lancet Neurol. 2015 May;14(5):478-84. Epub 2015 Apr 3 PubMed.

Primary Papers

  1. . Hyperactive somatostatin interneurons contribute to excitotoxicity in neurodegenerative disorders. Nat Neurosci. 2016 Apr;19(4):557-9. Epub 2016 Feb 22 PubMed.