Chronic Ca 2+ imaging of cortical neurons with long-term expression of GCaMP-X

Jinli Geng, Wenxiang Li, Yingjun Tang, Yunming Gao, Yitong Lu, Ping Li, Xiaomei Li, Ping Wang, Yubo Fan, Yaxiong Yang, Zengcai V. Guo, Xiaodong LiubioRxiv 2022 · DOI

TL;DR

Researchers have developed GCaMP-X, a modified version of the popular calcium indicator GCaMP that eliminates the probe's inherent neurotoxicity. This allows for stable, high-fidelity calcium imaging of neurons over months, revealing for the first time how spontaneous calcium oscillations directly drive the physical growth and development of neural circuits.

What Problem Does This Paper Address?

Calcium ($Ca^{2+}$) is the "universal currency" of neuronal signaling, reflecting both immediate electrical activity (like action potentials) and long-term developmental processes (like the growth of new connections, or neuritogenesis). To see these signals, scientists use Genetically Encoded Calcium Indicators (GECIs), with the GCaMP series being the gold standard.

However, GCaMP has a dark side: it is toxic to neurons over time. Because GCaMP is built using calmodulin (CaM), it interferes with the cell's natural signaling machinery. Specifically, when GCaMP isn't bound to calcium, it "steals" binding partners from the cell's endogenous calmodulin, particularly L-type calcium channels. This leads to "nuclear filling" (where the probe accumulates in the nucleus), stunted growth, and eventually cell death. This has historically limited "chronic" imaging to a narrow "safe window" of a few weeks, making it nearly impossible to study long-term neural development or disease progression accurately.

Methods

The authors utilize GCaMP-X, a variant they previously designed to be "neuron-compatible." The key innovation is the addition of an extra protective motif (derived from the IQ domain of neuromodulin) that prevents the probe from interfering with the cell's natural apo-calmodulin targets.

The study validates GCaMP-X through several rigorous approaches:

  • Electrophysiology: Using patch-clamp recordings in HEK293 cells to prove GCaMP-X doesn't disrupt calcium channel gating.
  • Long-term In Vitro Imaging: Monitoring cultured cortical neurons for over a month (up to DIV 42) to track morphology and spontaneous activity.
  • Chronic In Vivo Two-Photon Imaging: Observing the primary somatosensory cortex of living mice for over 8 weeks, comparing sensory-evoked responses (whisker stimulation) between GCaMP and GCaMP-X.
  • Subcellular Targeting: Deploying versions of the sensor specifically localized to either the cytosol ($GCaMP-X_C$) or the nucleus ($GCaMP-X_N$).

Datasets

While not a traditional "big data" study, the paper relies on several biological "benchmarks":

  • Primary Cortical Cultures: Derived from neonatal ICR mice and transgenic lines.
  • Transgenic Mouse Lines: Specifically the Ai148 line (a TIGRE2.0 line expressing GCaMP6f), which is widely used in the field but shown here to suffer from significant chronic side effects.
  • Viral Delivery: Using AAV (Adeno-associated virus) serotypes to deliver sensors to specific brain regions in vivo.

Key Insights and Ideas

  • The "Safe Window" is a Mirage: The authors show that even within the standard 3-8 week imaging window, conventional GCaMP is already subtly degrading neuronal health and response success rates. GCaMP-X extends this window indefinitely.
  • The Calcium-Growth Link: By being able to image the same neurons for a month, the authors discovered a direct linear correlation between the frequency of spontaneous calcium oscillations (specifically in the 10-100 mHz range) and the rate of neurite elongation.
  • Nuclear Calcium as an Integrator: Using $GCaMP-X_N$, they observed that the nucleus acts as a "low-pass filter" or integrator of cytosolic calcium spikes. This suggests the nucleus "counts" activity to trigger gene expression for growth.
  • Fixing Transgenic Models: The study demonstrates that the common "epileptiform" activity and stunted growth seen in popular transgenic GCaMP mice (like Ai93 and Ai148) are likely artifacts of GCaMP toxicity, which GCaMP-X avoids.

Key Results

  • Stability In Vivo: Beyond 8 weeks post-injection, GCaMP6m-expressing neurons showed a significant drop in success rates for whisker-evoked responses, whereas GCaMP-X remained stable and responsive.
  • Morphological Preservation: At 28 days in culture, neurons expressing GCaMP6m had a soma size ~33% smaller and significantly shorter, less complex neurites compared to GCaMP-X or GFP controls.
  • Signal-to-Noise Ratio (SNR): GCaMP-X provided a 2-fold improvement in SNR for nuclear imaging compared to conventional GCaMP, largely because it doesn't cause the "blinding" background haze associated with nuclear accumulation/toxicity.
  • Growth Correlation: A near-perfect inverse correlation ($R^2=0.99$) was found between oscillation frequency and neurite outgrowth, providing a quantitative "growth code" for developing neurons.

Limitations and Future Work

The authors acknowledge that while GCaMP-X solves the calmodulin-interference problem, any high-level overexpression of a foreign protein can eventually stress a cell.

Future Work includes: * Applying the "X-design" principles to other types of sensors, such as red-shifted calcium indicators or voltage sensors. * Investigating the specific molecular pathways that translate the 10-100 mHz calcium "clock" into physical structural changes in the neuron. * Using GCaMP-X to re-evaluate long-term disease models (like Alzheimer's or Parkinson's) where calcium dysregulation is a suspected factor but was previously difficult to image without probe-induced artifacts.

Summary updated

Abstract

Abstract Dynamic Ca 2+ signals reflect acute changes in membrane excitability (e.g. sensory response), and also mediate intracellular signaling cascades normally of longer time scales (e.g., Ca 2+ -dependent neuritogenesis). In both cases, chronic Ca 2+ imaging has been often desired, but largely hindered by unexpected cytotoxicity intrinsic to GCaMP, a popular series of genetically-encoded Ca 2+ indicators. Here, we demonstrate that the recently developed GCaMP-X outperforms GCaMP in long-term probe expression and/or chronic Ca 2+ imaging. GCaMP-X shows much improved compatibility with neurons and thus more reliable than GCaMP as demonstrated in vivo by acute Ca 2+ responses to whisker deflection or spontaneous Ca 2+ fluctuations over an extended time frame. Chronic Ca 2+ imaging data ( ≥ 1 month) are acquired from the same set of cultured cortical neurons, unveiling that spontaneous/local Ca 2+ activities would progressively develop into autonomous/global Ca 2+ oscillations. Besides the morphological indices of neurite length or soma size, the major metrics of oscillatory Ca 2+ , including rate, amplitude, synchrony among different neurons or organelles have also been examined along with the developmental stages. Both neuritogenesis and Ca 2+ signals are dysregulated by GCaMP in virus-infected or transgenic neurons, in direct contrast to GCaMP-X without any noticeable side-effect. Such in vitro data altogether consolidate the unique importance of oscillatory Ca 2+ to activity-dependent neuritogenesis, as one major factor responsible for the distinctions between GCaMP vs GCaMP-X in vivo . For the first time with GCaMP-X of long-term expression in neurons, spontaneous and sensory-evoked Ca 2+ activities are imaged and evaluated both in vitro and in vivo , providing new opportunities to monitor neural development or other chronic processes concurrently with Ca 2+ dynamics.
BibTeX
@misc{geng2022chronic,
  author = {Jinli Geng and Wenxiang Li and Yingjun Tang and Yunming Gao and Yitong Lu and Ping Li and Xiaomei Li and Ping Wang and Yubo Fan and Yaxiong Yang and Zengcai V. Guo and Xiaodong Liu},
  title = {{Chronic Ca <sup>2+</sup> imaging of cortical neurons with long-term expression of GCaMP-X}},
  year = {2022},
  doi = {10.1101/2022.01.09.475579},
  howpublished = {bioRxiv},
}
Figure