The cation-chloride cotransporters play a key role in maintaining the intracellular Cl⁻ concentration (Kaila et al., 2014). The sodium-potassium-chloride (Na⁺-K⁺-Cl⁻) cotransporter 1 (NKCC1), which is expressed in neurons as well as astrocytes (Yan et al., 2001a), oligodendrocytes (Wang et al., 2003), and microglia (Tóth et al., 2022), relies on the plasmalemmal Na⁺ gradient to transport two Cl⁻ ions into the cell along with one Na⁺ and one K⁺ ion, whereas the neuron-specific (Payne et al., 1996) potassium-chloride (K⁺-Cl⁻) cotransporter 2 (KCC2) relies on the plasmalemmal K⁺ gradient to transport one Cl⁻ and one K⁺ ion out of the cell in an electroneutral manner (Kaila et al., 2014). While there have been conflicting results regarding the developmental expression of NKCC1 in neurons (Virtanen et al., 2020), there is evidence for a developmental increase in KCC2 expression in humans as well as in rats (Dzhala et al., 2005). The expression of the Cl⁻ extruder KCC2 is necessary for GABA_A receptor-mediated hyperpolarization, which explains why GABA has a depolarizing effect in neonatal rat neurons (Rivera et al., 1999). The clinical relevance of this developmental shift is supported by the observations that KCC2 and NKCC1 expression levels are altered in several neurodevelopmental disorders associated with epilepsy (Aronica et al., 2007) and that mutations in the SLC12A5 gene, which encodes KCC2, are associated with autism spectrum disorder and schizophrenia (Merner et al., 2015), two disorders of neuronal development (see also Cherubini et al., 2022).

Changes in KCC2 or NKCC1 expression have also been observed in several different animal models of CNS and PNS injury: KCC2 protein or Kcc2 mRNA is downregulated following peripheral nerve injury (Coull et al., 2003), spinal cord injury (Boulenguez et al., 2010), fluid percussion injury (Bonislawski et al., 2007), axotomy (Nabekura et al., 2002; Shulga et al., 2008), ischemia/stroke (Papp et al., 2008; Jaenisch et al., 2010), and deafferentation (Tighilet et al., 2016), while NKCC1 is upregulated following traumatic brain injury (Lu et al., 2008; Hui et al., 2016; Zhang et al., 2016), spinal cord injury (Yan et al., 2018) and ischemia (Yan et al., 2001b). However, so far, only general, spatially unresolved lesion-induced changes of KCC2 and NKCC1 have been studied. It has remained unclear whether a precise, anatomically well-defined lesion can evoke spatially restricted, layer-specific alterations in KCC2 or NKCC1 expression.

Therefore, we combined laser microdissection with microarray analysis and reverse transcription-quantitative polymerase chain reaction (RT-qPCR) to explore layer-specific changes in KCC2 and NKCC1 expression at different time points following entorhinal denervation. This approach has been widely used to study both the direct effects of axonal injury on the neurons in the entorhinal cortex as well as the secondary effects of the deafferentation on the neurons in the hippocampus, which are located far away from the lesion site and therefore unlikely to suffer direct damage (Deller and Frotscher, 1997; Perederiy and Westbrook, 2013). Moreover, the layered arrangement of the different synaptic inputs to the dentate gyrus granule cells allows one to differentiate denervated from non-denervated dendritic segments and analyze layer-specific morphological, molecular, and functional changes following entorhinal denervation (Deller et al., 2001, 2006). Our analysis revealed that entorhinal denervation induced layer-specific changes in the expression levels of KCC2 and NKCC1 in the dentate gyrus, which might reflect homeostatic adaptations to the loss of synaptic input.

Entorhinal denervation is a well-established model for studying the neuronal reorganization processes following deafferentation (Deller and Frotscher, 1997; Deller et al., 2007; Perederiy and Westbrook, 2013). Using this model in combination with laser microdissection of defined cell layers, we detected cell- and layer-specific changes in the expression levels of the cation-chloride cotransporters KCC2 and NKCC1 following entorhinal denervation. Compared to non-denervated control mice, we found that Kcc2 mRNA was strongly downregulated in the gcl of the dentate gyrus 7 dpl, while Nkcc1 mRNA was upregulated in the oml/mml at this time point. Subsequent immunohistochemical staining revealed that KCC2 protein expression was reduced in the denervated oml/mml, where the dendrites of the granule cells are located, whereas NKCC1 protein was highly expressed in reactive astrocytes in the oml/mml of the denervated dentate gyrus. Together, these results suggest that KCC2 expression is selectively and transiently downregulated in granule cells while the increased NKCC1 expression is primarily mediated by reactive astrocytes in the denervated molecular layer (Figure 5).

The use of laser microdissection allowed us, for the first time, to pinpoint the expression changes of the two cation-chloride cotransporters in specific cell layers. Compared to controls, Kcc2 mRNA was downregulated in the gcl. Since KCC2 is almost exclusively expressed in neurons (Kaila et al., 2014) and the cell bodies of the granule cells make up the majority of cells in the gcl, we concluded that Kcc2 mRNA was selectively downregulated in granule cells. In situ hybridization in mice confirms that Kcc2 mRNA is mainly confined to the gcl of the dentate gyrus (Allen Brain Atlas²). In contrast, KCC2 protein levels are highest in the molecular layer of the dentate gyrus, i.e., in the dendrites of the dentate granule cells, while the granule cell somata and axon initial segments show a comparatively weak expression (Gulyás et al., 2001; Báldi et al., 2010). The absence of Kcc2 mRNA downregulation in the oml/mml at 7 dpl could indicate the presence of Kcc2 mRNA in other cell types, such as inhibitory interneurons. Indeed, our immunostainings show that KCC2-expressing neurons are present in both the oml/mml of the denervated (ipsilateral) and the non-denervated (contralateral) hemispheres. However, compared to the ipsilateral iml and the contralateral oml/mml, the fluorescence intensity in the ipsilateral oml/mml was reduced, suggesting that KCC2 is primarily lost from the denervated dendritic segments of the granule cells but not from the denervated dendrites of interneurons. Interestingly, a recent study showed that a selective permanent loss of KCC2 in interneurons leads to cortical seizures due to alterations in the distribution of interneuron subtypes (Zavalin et al., 2022). That suggests that a granule cell-specific, transient reduction of KCC2 in the denervated dentate gyrus with preserved KCC2 in interneurons would not cause such a dramatic excitation-inhibition imbalance. Indeed, while electrolytic entorhinal cortex lesions were shown to induce hippocampal seizures (Dasheiff and McNamara, 1982; Kelley and Steward, 1996), the wire knife approach has not been associated with seizures (Dasheiff and McNamara, 1982). Taken together, our data show that entorhinal denervation leads to a reduction of Kcc2 mRNA in granule cells as well as a selective downregulation of KCC2 protein expression in the denervated dendrites in the oml/mml.

The downregulation of KCC2 protein expression in the denervated granule cells would be expected to increase the intracellular Cl⁻ concentration and reverse the polarity of GABAergic currents, leading to an increase in neuronal excitability. KCC2 has been shown to be downregulated following different types of nervous system injury, such as chronic peripheral nerve constriction (Coull et al., 2003), spinal cord transection (Boulenguez et al., 2010), axotomy (Nabekura et al., 2002; Shulga et al., 2008), middle cerebral artery occlusion (Jaenisch et al., 2010), and deafferentation (Tighilet et al., 2016), indicating that the downregulation of KCC2 may be a common response to neuronal injury (Wu et al., 2016). Some authors have concluded that KCC2 downregulation is pathological, with depolarizing GABA responses promoting hyperexcitability (Boulenguez et al., 2010), excitotoxicity (Nabekura et al., 2002) and cell death (Kontou et al., 2021). Other authors have argued that reduced KCC2 expression may be beneficial by increasing the survival rate (Shulga et al., 2008) and promoting the reorganization of injured neurons (Jaenisch et al., 2010). It has even been speculated that the downregulation of KCC2 in the cochlear nucleus following unilateral cochlear neurectomy in cats may represent a homeostatic adaptation to the loss of synaptic input and preserve the activity levels in the deafferented neurons (Tighilet et al., 2016). Thus, the downregulation in KCC2 we observed in the denervated granule cells might reflect a homeostatic, rather than a pathological, response to deafferentation. However, depolarizing GABAergic inputs can also mediate inhibition by shunting excitatory inputs (Kilb, 2021). Therefore, it would be necessary to measure neuronal activity while monitoring KCC2 expression and chloride concentration in granule cells to determine whether KCC2 downregulation increases neuronal excitability. Since KCC2 agonists have been proposed for the treatment of spinal cord injuries (Hegarty and Stanicka, 2022), the transient downregulation of KCC2 in a denervated brain area suggests that treatment with KCC2 agonists could not only affect reorganization processes at injury sites but also brain rewiring in connected brain regions. Although currently hypothetical, our results and data from others (reviewed in Kaila et al., 2014; Hegarty and Stanicka, 2022) suggest that it may be worthwhile to investigate potential therapeutic effects.

Intriguingly, in addition to its chloride transport function, KCC2 protein has been shown to contribute to the regulation of excitatory and inhibitory synaptic function (Gauvain et al., 2011; Chamma et al., 2013; Chevy et al., 2015; Al Awabdh et al., 2022) and to the formation and maturation of spines (Li et al., 2007; Fiumelli et al., 2013; Awad et al., 2018). In line with this, KCC2 is expressed close to excitatory (Gulyás et al., 2001; Gauvain et al., 2011) and inhibitory synapses (Gulyás et al., 2001; Chamma et al., 2013). In this context, it is interesting to compare the time course of changes in spine density and KCC2 expression after denervation of the dentate gyrus. Entorhinal denervation leads to a transient loss of spines during the first week and a full recovery after several weeks (Parnavelas et al., 1974; Caceres and Steward, 1983; Vuksic et al., 2011). Therefore, it is tempting to speculate that the downregulation and recovery of KCC2 are associated with the denervation-induced reduction and recovery of spine density. However, future studies are needed to establish a mechanistic link between changes in KCC2 and spine remodeling triggered by entorhinal denervation.

In contrast to KCC2, both Nkcc1 mRNA and NKCC1 protein expression were upregulated in the oml/mml following entorhinal denervation. Previous studies have similarly reported an increase in NKCC1 protein and/or mRNA levels in the brain and the spinal cord following injury (Yan et al., 2001a, 2018; Lu et al., 2008; Hui et al., 2016; Zhang et al., 2016). However, since NKCC1 is expressed in many different cell types (Kurki et al., 2022), we additionally performed immunohistochemical staining to validate the expression changes at the single-cell level. Based on the microarray and qPCR data showing a selective upregulation in the oml/mml, but not the gcl, we concluded that Nkcc1 mRNA levels in the granule cells were not significantly affected by the lesion. That was consistent also with the expression of Nkcc1 mRNA in the dentate gcl (Allen Brain Atlas³). Instead, we hypothesized that Nkcc1 mRNA upregulation might be mediated by an increased expression in reactive astrocytes, which are induced by entorhinal denervation (Del Turco et al., 2003; Burbach et al., 2004). Therefore, we performed immunohistochemical staining for GFAP, which is upregulated in reactive astrocytes (Pekny and Pekna, 2014). The GFAP fluorescence was strongly increased in the denervated oml/mml and highly colocalized with NKCC1, which was also upregulated compared to the contralateral oml/mml. It was previously shown that reactive astrocytosis (quantified by GFAP immunofluorescence in the dentate gyrus) reaches a peak between 4 and 8 days and persists until about 12 days after aspiration of the entorhinal cortex in mice (Steward and Trimmer, 1997), which fits our microarray data showing a significant increase in Nkcc1 mRNA between 3 and 14 dpl.

The increased expression of NKCC1 in reactive astrocytes might be related to a “dedifferentiation” to an immature, proliferative state (Buffo et al., 2008). It has been known for a long time that NKCC1 is involved in the regulation of cell proliferation in mouse fibroblasts (Panet et al., 2000), and inhibition of NKCC1 reduces the proliferation of mouse glioma cells (Luo et al., 2020). Following entorhinal cortex lesion in rats, the number of proliferating astrocytes is increased in the denervated cell layers (Hailer et al., 1999). However, increased NKCC1 expression in individual astrocytes and/or microglia, which also express functional NKCC1 (Tóth et al., 2022), could also account for the increased NKCC1 fluorescence in the oml/mml following entorhinal denervation. Interestingly, the electrophysiological characteristics of astrocytes also undergo a process of dedifferentiation following entorhinal cortex lesion in rats: the resting membrane potential and the input resistance are increased, which has been attributed to a decrease in inward-rectifying potassium current (Schröder et al., 1999). Buffering of the extracellular K⁺ concentration is an important astrocytic function, which is normally accomplished by the Na⁺/K⁺ pump and the inward-rectifying potassium channel Kir4.1 (Larsen et al., 2014). However, under pathological conditions, such as traumatic brain injury, NKCC1 activity or expression is upregulated and contributes to astrocytic swelling in vitro (Jayakumar et al., 2011) and in vivo (Zhang et al., 2016). Furthermore, NKCC1-deficient astrocytes do not swell in response to high extracellular K⁺ concentrations (Su et al., 2002), indicating that NKCC1 is essential for volume regulation during extreme conditions. Thus, it is possible that NKCC1 takes over the function of Kir4.1 in regulating the astrocytic volume and extracellular K⁺ concentration following entorhinal denervation, leading to an increased fluorescence signal. Understanding the cell-specific expression patterns of both KCC2 and NKCC1 following CNS and PNS injury is essential for developing effective treatments. For example, while several animal studies have reported that a blockade of NKCC1 upregulation following CNS injury improved clinical outcomes (Savardi et al., 2021), the situation appears to be different in the PNS. One study showed that increased activation of NKCC1 was necessary for the regeneration of myelinated sensory neurons in the dorsal root ganglion of rats that had undergone controlled sciatic nerve injury (Mòdol et al., 2015). Thus, NKCC1 upregulation may be pathological or beneficial depending on the cell type. Increasing KCC2 expression pharmacologically in mice with staggered spinal cord lesions improved the clinical outcome, but failed to do so in completely lesioned mice, indicating that the functional recovery was mediated by the reactivation of the remaining connections from the brain to the spinal cord that had been spared in the staggered lesion model (Chen et al., 2018). This reactivation of the spared circuitry was mainly accomplished by inhibitory interneurons and appears to be specific to KCC2 since increasing inhibition with GABA receptor agonists did not promote recovery (Chen et al., 2018). Therefore, it appears likely that some of the other functions of KCC2, such as the regulation of synaptic transmission, are even more important for the functional recovery after spinal cord injury than its chloride transport function.

In conclusion, we provide evidence for possibly homeostatic changes in KCC2 and NKCC1 expression in dentate granule cells and reactive astrocytes following entorhinal denervation in vivo. The temporary downregulation of KCC2 in the denervated granule cells might be associated with spine loss and contribute to a stabilization of the network activity by increasing GABAergic depolarization and neuronal excitability before other homeostatic mechanisms, such as dendritic retraction (Vuksic et al., 2011) and commissural sprouting by mossy cell axons (Del Turco et al., 2003) take effect, while the upregulation of NKCC1 in astrocytes might serve specific functions, such as enhancing proliferation and/or regulating the cell volume.

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

The animal study was reviewed and approved by Regierungspräsidium Darmstadt (Germany) and by the Ethics Committee of the University of Zagreb School of Medicine (Croatia).

DD, JS, MP, JM, FB, and MV acquired and analyzed the data. DD, PJ, and TD conceived and supervised the study. All authors were involved in data interpretation, critically revising the manuscript, read, and approved the final manuscript.

This work was supported by Deutsche Forschungsgemeinschaft (SFB 1080 to TD; DFG No. 467764793 and JE 528/10–1 to PJ).

TD received a honorarium from Novartis for a lecture on human neuroanatomy.

The remaining authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.