Astrocytes, which are abundant glial cells in the CNS, have become increasingly recognized as critical elements regulating neuronal function [1] including somatosensory information processing in the spinal dorsal horn (SDH) [2, 3] and brain [4]. By using a method of in vivo Ca2+ imaging in the SDH [5], several studies have shown that SDH astrocytes have increased intracellular Ca2+ levels ([Ca2+]i) following strong mechanical pressure (pinch) to the hindpaw [6] and intraplantar injection of chemical irritants (capsaicin and formalin) [7, 8], suggesting that SDH astrocytes respond to noxious stimuli in the periphery. However, the mechanism underlying the increase in astrocytic [Ca2+]i is not fully understood. Our recent study demonstrated that [Ca2+]i increases in SDH astrocytes after intraplantar capsaicin are mediated by the activation of α1A-adrenaline receptors (α1A-ARs) through descending noradrenergic (NAergic) neurons from the locus coeruleus (LC) to the SDH [7]. However, whether the α1A-AR-mediated descending LC-NAergic signals commonly contribute to astrocytic Ca2+ responses evoked by noxious stimuli remains unclear. In this study, we investigated astrocytic Ca2+ responses to noxious irritant formalin using multiple approaches, including in vivo Ca2+ imaging, circuit-specific neuronal ablation, conditional gene knockout, and pharmacological intervention.

For in vivo Ca2+ imaging in SDH astrocytes, the Ca2+ indicator, GCaMP6m, was selectively expressed in SDH astrocytes following microinjection of an adeno-associated virus (AAV) vector expressing GCaMP6m under the control of the astrocytic promoter, gfaABC1D, into the left SDH (Additional file 1: Figure S1; Additional file 2), as reported previously [7, 8]. Using GCaMP6m-expressing mice under anesthesia, we confirmed that intraplantar injection of formalin, but not vehicle, induced robust increases in [Ca2+]i in SDH astrocytes (Fig. 1a). To examine the involvement of the descending LC-NAergic pathway, we employed a circuit-specific ablation method using diphtheria toxin (DTX) and its receptor (DTR). To ablate SDH-projecting LC-NAergic neurons, AAVretro-Cre was microinjected into the left SDH of wild-type mice, and AAV-FLEX-DTR-EGFP or AAV-FLEX-AcGFP (control) was injected into the bilateral LC (Fig. 1b). In these mice, GFP expression was observed in the LC, and GFP+ LC neurons were immunolabeled with an antibody for tyrosine hydroxylase (TH), a marker for catecholaminergic neurons (mostly NAergic neurons in the LC) (Fig. 1c). Systemic administration of DTX eliminated GFP+ LC neurons in mice with AAV-FLEX-DTR-EGFP, but not in those with AAV-FLEX-AcGFP (Fig. 1c). In GCaMP6m-expressing mice with an ablation of descending LC-NAergic neurons, we found that the percentage of SDH astrocytes with increased [Ca2+]i evoked by intraplantar formalin injection was significantly lower (Fig. 1d). The average trace of Ca2+ responses and the area under the curve (AUC) of Ca2+ traces from individual SDH astrocytes during the first 600 s after formalin injection were also suppressed. These results indicate that the descending LC-NAergic pathway contributes to formalin-induced astrocytic Ca2+ responses in SDH.

Fig. 1
figure 1

Intraplantar injection of formalin activates SDH astrocytes via α1A-ARs through descending LC-NAergic signals. a Averaged trace and AUC during the first 600 s (AUC0–600 s) of astrocytic Ca2+ signals in the SDH after intraplantar injection of vehicle or formalin (vehicle, n = 47 ROIs, 4 mice; formalin, n = 123 ROIs, 4 mice, ****P < 0.0001, Mann–Whitney U test). b Schematic illustration of retrograde transduction strategy in descending LC-NAergic neurons using the FLEX-switch system. c Representative images of LC-NAergic neurons in mice treated with PBS or DTX administration. GFP (green), and TH (red). d–f SDH astrocytic Ca2+ responses by formalin in mice with ablation of descending LC-NAergic neurons (d), conditional knockout of α1A-ARs in Hes5+ astrocytes (Adra1a-cKO; Hes5-CreERT2;Adra1aflox/flox) compared with control mice (control; Adra1aflox/flox) (e), and pretreatment intrathecally with PBS or silodosin (3 nmol) (f). Percentage of responding astrocytes (d control, n = 6 mice; ablated, n = 6 mice; e: control, n = 5 mice; Adra1a-cKO, n = 5 mice; f PBS, n = 6 mice; silodosin, n = 6 mice, *P < 0.05, **P < 0.01, ****P < 0.0001, unpaired t-test); averaged trace and AUC (d control, n = 255 ROIs; ablated, n = 263 ROIs; e control, n = 253 ROIs; Adra1a-cKO, n = 224 ROIs; f PBS, n = 364 ROIs; silodosin, n = 296 ROIs, ****P < 0.0001, Mann–Whitney U test). Data show the mean ± SEM

Full size image

We previously identified α1A-AR as an astrocyte-expressing receptor necessary for Ca2+ responses evoked by intraplantar capsaicin [7]. Consistent with our previous study [7], immunohistochemical analysis confirmed that 96.5 ± 1.9% of SDH astrocytes expressed α1A-ARs (Additional file 1: Figure S2). Thus, we examined the role of α1A-AR using Hes5-CreERT2;Adra1aflox/flox mice (treated with tamoxifen) that lack this receptor in SDH astrocytes, especially localized in superficial laminae [7]. The number of SDH astrocytes with increased [Ca2+]i by formalin in Adra1aflox/flox control mice (Control) was dramatically decreased in Hes5-CreERT2;Adra1aflox/flox mice (Adra1a-cKO) (Fig. 1e). The average trace and AUC for Ca2+ responses were also lower in Adra1a-cKO mice than in control mice. Adra1a-cKO mice treated with tamoxifen also lack α1A-AR expression in brain Hes5+ astrocytes [7]. To determine the importance of α1A-ARs in the spinal cord, we intrathecally administered the α1A-AR-specific antagonist, silodosin, before formalin injection. Silodosin-pretreated mice also showed marked inhibition of the formalin-induced astrocytic Ca2+ responses (the percentage of SDH astrocytes with [Ca2+]i increases, the average trace of Ca2+ responses, and their AUCs) (Fig. 1f). Taken together, the Ca2+ responses in SDH astrocytes following formalin injection are mediated by the activation of α1A-ARs through descending LC-NAergic signals.

In this study, we demonstrate for the first time that intraplantar injection of the noxious irritant, formalin, activates SDH astrocytes (especially the Hes5+ subset) via α1A-ARs stimulated by descending LC-NAergic signaling. Previous data showing induction of the neuronal activity marker c-FOS in LC-NAergic neurons [9] supports our findings. Given that astrocytic Ca2+ responses in the SDH after intraplantar capsaicin are mediated by α1A-AR-mediated descending LC-NAergic signaling [7], this raises the possibility that this signaling pathway from the LC-NAergic neurons to SDH astrocytes is a common mechanism for astrocytic Ca2+ responses in the SDH evoked by noxious chemical irritants. However, the decrease in the number of responding astrocytes was slightly lower in mice with LC-NAergic neuron ablation than in mice with conditional α1A-AR-knockout and silodosin pretreatment. This could be due to incomplete ablation of LC-NAergic neurons projecting to the 4th lumbar SDH where astrocytic Ca2+ responses were monitored or the involvement of other descending NAergic pathways, for example, from regions A5 and A7 (although the LC is the main source of NA in the SDH [10]). In addition, considering the residual astrocytic Ca2+ responses observed in mice either with genetic knockout or pharmacological blockade of α1A-ARs, it seems that other neurotransmitters, such as glutamate, GABA, and ATP, which are known to cause astrocytic Ca2+ elevations [11], may also be involved. Nevertheless, our findings indicate that α1A-AR-mediated descending LC-NAergic signals are a primary driver of Ca2+ responses in SDH astrocytes evoked by noxious stimuli.

In this study, there were different patterns of the average traces of Ca2+ responses after intraplantar formalin injection among experiments. The reason for this difference remains unclear. Nevertheless, Ca2+ responses during several minutes after the injection are commonly observed and are consistent with our previous data [8]. However, Ca2+ responses in Adra1aflox/flox and Hes5-CreERT2;Adra1aflox/flox mice were different from others. It may involve a genetic factor (and/or tamoxifen treatment) because the genetic background of Adra1aflox/flox mice was derived from BDF1 [(C57BL/6 × DBA/2)F1] strain and these mice were not fully backcrossed on the C57BL/6 background, while other experiments used C57BL/6 mice.

Formalin is used as a model for acute and persistent inflammatory pain associated with peripheral tissue injury. The role of spinal NAergic signals in formalin-induced pain has been examined in many studies [12, 13], but it remains controversial. For example, intrathecal treatment with α2-AR agonists reduces formalin pain [12, 14], intrathecal treatment with anti-dopamine-β-hydroxylase antibody-conjugated saporin, which kills SDH-projecting NAergic neurons, attenuates formalin pain [15]. An explanation for this discrepancy may be partly associated with the action of NA in SDH astrocytes. It should be noted that we measured the astrocytic Ca2+ responses for the first 10 min after formalin injection, a time period that corresponds to acute phase of formalin-induced nociceptive behavior. Further investigations using a tool to manipulate Ca2+ responses specifically in Hes5+ SDH astrocytes will uncover their in vivo role in nociceptive information processing and behaviors evoked by formalin.