Respiratory neurons in the ventrolateral medulla are closely intertwined with astrocytes that play an important role in stabilizing the autonomous activity of this network (Hulsmann et al., 2000; Grass et al., 2004; Schnell et al., 2011; Sheikhbahaei et al., 2018; Tacke et al., 2023). Astrocytes provide essential support to the neuronal network by controlling neurotransmitter homeostasis, ensuring proper neurotransmission (Gomeza et al., 2003; Szoke et al., 2006; Eulenburg and Hülsmann, 2022). Furthermore, various neuromodulators evoke intracellular calcium elevations in astrocytes, which may impact respiratory activity in turn (Funk et al., 2008; Härtel et al., 2009; Huxtable et al., 2010; Schnell et al., 2016). However, whether spontaneous calcium activity influences or drives respiratory neuronal activity is still a subject of ongoing discussion (Schnell et al., 2011; Okada et al., 2012; Oku et al., 2016; Tacke et al., 2023). Here, we built and tested a fiber optical setup to record and analyze astrocytic calcium signaling in the intact respiratory network of the working heart brainstem preparation. One aim of the study was to test whether calcium signals from individual astrocytes can be recorded with an imaging fiber.

To induce reliable calcium signals in astrocytes, 10 µM norepinephrine (NE) was added to the perfusion. NE induced an increase of the apparent perfusion pressure (aPP) of 49.4 ± 26.3 mmHg (n = 11). Additionally, there was also a variable increase of the phrenic nerve activity (PNA) with larger amplitudes and an increase of the respiratory rate (Figures 2, 3). The peak of the PNA-response followed the initial increase of the aPP after 104.7 ± 5.2 s (n = 11). In four recordings, the increase of the aPP was accompanied by a short period of reduced PN activity (Figure 2).

Analysis of the emission of the full image core of the fiber revealed an increase of the calcium activity of the 470/405 nm ratio signal in 10 out of 11 recordings/preparations. Interestingly, the astrocytic calcium rise preceded the increase of PNA in 8 preparations (average delay 57.1 ± 21.0 s; p = 0.008; Mann-Whitney Rank Sum Test).

In a second step we analyzed the calcium responses in more detail. In 4 of the preparations individual cells could be identified by their individual, non-uniform calcium signaling (Figures 2, 3). Since the individual signals show different timing and duration, we assume that these signals represent the calcium response of individual cells. It is important to note that these individual signals are clearly distinguishable from the integrated emission of the full fiber (gray trace in Figures 2A, 3A). Furthermore, in 2 preparations the analysis of the different cells revealed spontaneous calcium signals in individual cells (Figure 3). Calcium signals that were phase locked to the inspiratory phrenic bursts were not observed (Figure 3).

In this paper, we were able to prove that calcium imaging of astrocytes in the working heart brainstem preparation is possible by using a commercially available imaging fiber and a custom-made fiber-coupled microscope. It was possible to discern NE-induced and spontaneous calcium signals from individual cells. Furthermore, we also showed that NE-induced calcium signals in astrocytes precede the NE-induced increase of the respiratory network activity, which has been described in rodents (Viemari, 2008).

Recording calcium signals from individual cells represents a clear advantage compared to the classical fiber photometry, which can only capture ‘bulk’ or population-level calcium (Ca²⁺) activity (Gunaydin et al., 2014). This is particularly important when analyzing spontaneous astrocytic signals, which might not be detected in the bulk signal (see Figure 3).

Interestingly, in four of the experiments there was a short period of a reduced PNA, which started simultaneously with the NE-induced increase of the perfusion pressure (aPP). Activation of arterial baroreceptors by the elevated aPP or directly by NE (Holmes and Ledsome, 1984), would negatively modulate the network in the VRC (Simms et al., 2007; Barnett et al., 2021). After adaptation of the baroreflex, the neuron-mediated stimulation of the network by NE will become dominant (Zanella et al., 2014; Tacke et al., 2023). Further experiments are required to test if the depression is eventually a response of the network to the astrocytic calcium rise and release of an inhibitory gliotransmitter (Gourine et al., 2010; Forsberg et al., 2017).

In response to NE, individual cells showed different delay and signal duration. These signals result most likely from the soma of the astrocyte, since it is known that NE induces strong somatic calcium-signals (Schnell et al., 2016). However, with the fiber imaging system we were not able to discriminate between different parts of the cells, thus making it impossible to distinguish calcium signals of processes from somata of individual cells. Thus, during the peak of the NE-response (Figure 2D), we observe a rather blurred image consisting of the overlapping calcium signals in somata and processes of neighboring astrocytes. Indeed, subcellular structures were not visible in the hGFAP-EGFP mice, in which high baseline fluorescence usually allows this discrimination using epifluorescence or 2P-imaging (Grass et al., 2004; Hirrlinger et al., 2004). Norepinephrine was applied to the perfusate and therefore reached the medulla via the blood stream. We cannot exclude that the perfusion with the artificial solution is altering the blood brain barrier (BBB) integrity, but even if this was intact, it is known that brain capillary endothelial cells and astrocytes express norepinephrine transporter that allow NE to enter the brain (Wakayama et al., 2002; Inazu et al., 2003; Schnell et al., 2016). The apparent delay between the vascular response and the astrocytic calcium signal is in favor of an intact BBB.

We used a ratiometric approach for the calcium imaging with two wavelengths for excitation. 470 nm allows the detection of calcium-induced fluorescence change and 405 nm, which is very close to the isosbestic point of GCaMP6s (Dana et al., 2019). Under baseline condition, individual cells, however, were not always distinguishable (Figure 2B) but appeared only after application of NE (Figure 2). One reason is probably that the lateral resolution is limited by the size of the fiber cores (2 µm). Since at its isosbestic point the emission of GCaMP6s (Ca²⁺-free or Ca²⁺-saturated) is below 10% of the maximal fluorescence of the Ca²⁺-saturated GCaMP6 at 470 nm (Dana et al., 2019), low baseline calcium levels might not be detectable. Thus, it can be concluded that the fiber imaging setup requires a higher signal to noise ratio for the detection of (baseline) calcium levels as e.g., two-photon imaging, where most astrocytes are visible at rest (Tacke et al., 2023). One possible source of noise is autofluorescence of the imaging fiber (Udovich et al., 2008). The transmissibility (t) of the fiber and the objective (t @410 nm 83.8% vs. t @470 nm 89.9%), however, should not be a limitation. Nevertheless, we could prove that NE-induced as well as spontaneous calcium signals can be recorded with the setup.

The imaging fiber has a relatively short working distance <100 µm (Feng et al., 2022), thus, the fiber has to be inserted into the tissue and could potentially damage critical structures. We cannot exclude that this manipulation induces calcium signals in astrocytes (see high baseline signal in cell 1, Figure 3A) in response to the mechanical stimulus (Lee et al., 2015). For this project, we limited the size of the imaging fiber to an outer diameter of about 200 µm to minimize the impact of the fiber when inserting into the tissue to reach the ventral respiratory column. Further trials with larger fibers and a higher number of cores need to be tested.

Utilizing an imaging fiber enables the measurement of calcium signals from individual astrocytes. This is particularly valuable when analyzing spontaneous calcium signals, which can be very heterogeneous in astrocytes (Schnell et al., 2011). Although the spatial resolution of the imaging fiber system is not as high as that for our two-photon microscope in acute brains slices (Schnell et al., 2016; Tacke et al., 2023), it allows deep tissue recordings in a complex preparation, such as the WHBP, where the complex intricate circuitry of the respiratory network remains intact.