Tachykinins represent a neuropeptide family widely distributed in the body. The first members of this family were substance P (SP), neurokinin A (NKA), derived from the preprotachykinin A/Tac1 gene, and neurokinin B (NKB), encoded by the preprotachykinin B/Tac3 gene (Borbély and Helyes, 2017). Other members of this family have also been identified: neuropeptide K (NPK) (Roth et al., 1985), neuropeptide γ (NPγ) (Kage et al., 1988), hemokinin-1 (HK-1), and endokinins (EKs) (Kurtz et al., 2002). HK-1 encoded by the preprotachykinin C/Tac4 gene was discovered 20 years ago in B lymphocytes (Zhang et al., 2000), and similar to SP, it was shown in several immune cell types, including myeloid (polymorphonuclear granulocytes and eosinophils), lymphoid, dendritic cells, neurons, and microglia (Borbély and Helyes, 2017). This widespread expression of HK-1 in diverse cell types might be related to a broad range of physiological and pathophysiological functions by activating a multitude of signaling pathways, which are currently not understood. HK-1 induces a variety of physiological and pathophysiological functions in the immune and hematopoietic systems, gastrointestinal tract, airways, cardiovascular, endocrine, and neural systems, bone, joints, and cancer. The role of HK-1 in pain is contradictory and likely to be concentration/dose dependent (Borbély and Helyes, 2017). In mice, centrally administered HK-1 caused scratching (Borbély and Helyes, 2017), licking, and biting at low doses (Watanabe et al., 2010). Intrathecally injected HK-1 was analgesic in nanomolar concentrations in an NK1- or opioid receptor-dependent manner, but hyperalgesic in picomolar concentrations (Fu et al., 2005).

Tachykinins possess a differential affinity for the three tachykinin receptors. HK and EKs have the highest affinity for NK1R, similar to SP (Satake et al., 2013; Steinhoff et al., 2014). They are all G protein-coupled receptors of the rhodopsin family. Signaling through the NK receptors is complex with multiple directions: (1) G_q-related: activation of phospholipase C, resulting in inositol trisphosphate and diacylglycerol formation, mobilization of Ca²⁺ from intracellular stores, and activation of protein kinase C; (2) G_s-related activation of adenylyl cyclase, resulting in the cAMP formation and protein kinase A (PKA) activation; or (3) activation of phospholipase A2 and arachidonic acid production (Steinhoff et al., 2014; Garcia-Recio and Gascón, 2015). The existence of a specific receptor for HK-1 is currently intensively investigated (Duffy et al., 2003). It was shown that an NK1R antagonist did not inhibit pain caused by HK-1 in different concentrations, suggesting a specific HK-1 target/receptor (Watanabe et al., 2010). NK1R antagonists were developed as analgesic drug candidates, but they failed in human pain conditions. This might be due to different NK1R splice variants linked to distinct binding and activation mechanisms and/or a currently unidentified receptor (Zhang et al., 2000; Borbély and Helyes, 2017; Hunyady et al., 2019; Borbély et al., 2020). Our research group earlier showed that HK-1 mediated chronic adjuvant-induced inflammation and related pain (Borbély et al., 2013) and neuropathic and neurogenic inflammatory hyperalgesia via NK1 receptor activation (Hunyady et al., 2019). Interestingly, we recently demonstrated that unlike SP, HK-1 activates primary sensory neurons by inducing calcium influx via an NK1R-independent mechanism (Borbély et al., 2020). NK1R was shown to interact with the Mas-related G protein receptors (Mrgr) in certain pain-related mechanisms (Zhou et al., 2015). The human Mrgprx2, which is an ortholog of the rat Mrgprb5, was described as a potential target for HK-1 in mast cells to cause degranulation (Manorak et al., 2018). However, naturally occurring variants of Mrgprx2 lose HK-1 binding ability, showing the high importance of receptor variants (Alkanfari et al., 2018).

Based on a broad range of data demonstrating proinflammatory and pronociceptive functions of HK-1 in different organs partially independently of NK1 receptor activation (Hunyady et al., 2019; Borbély et al., 2020), our main aim was to identify the molecular mechanisms responsible for these actions. Our recent results demonstrated NK1R-independent calcium influx in cultured trigeminal ganglion cells induced by 1 μM, but not 500 nM HK-1 (Borbély et al., 2020). In the present study, we investigated the concentration- and treatment duration-dependent intracellular signaling mechanisms and pathways in these primary sensory neurons using an unbiased transcriptomic approach.

We demonstrate here the first transcriptomic data on the signaling pathways of HK-1 in rat primary sensory neurons, focusing on potential HK-1 targets, mechanisms of action, and DE genes involved in pain transmission and inflammation. The effect of HK-1 is complex, concentration dependent, and time dependent. Our findings support the well-described nociceptive actions of HK-1 and might explain the divergent neuronal activation processes. Since HK-1 is likely to act both on the primary sensory neurons and satellite glia cells, in this study we extract knowledge on these interactions to mimic in vivo conditions (Sakai et al., 2012; Theoharides et al., 2019).

Peripheral inflammatory and neuroinflammatory mechanisms resulting from complex interactions of immune cells, glial cells, and neurons play crucial roles in several diseases, including chronic pain (Siiskonen and Harvima, 2019). Itga4 and Antxr2 were upregulated for both concentrations, Tenm3 for 1 μM and Cxcl9 was downregulated after 500 nM at both timepoints. Itga4 encodes CD49d (alpha 4), one chain of very late antigen 4 (VLA-4), which belongs to adhesive molecules that activate the inflammatory process by facilitating the migration of immune cells into the central nervous system. The role of VLA-4 in genetic predisposition to chronic neuroinflammatory diseases has been demonstrated by several studies (Andreoli et al., 2007; Odoherty et al., 2007; Correia et al., 2009), supporting our findings in sensory neurons. Anthrax toxin components binding to ANTXR2, highly expressed on primary sensory neurons, were shown to inhibit inflammatory and neuropathic mechanical and thermal hyperalgesia (Yang et al., 2022). Tenm3 encodes a transmembrane protein also related to pain (Rouillard et al., 2016). Chemokines (CXCL4, CXCL9, CXCL10, and CXCL11) are important nociceptive mediators produced by neurons, microglia, and/or astroglia (Colvin et al., 2004). The CXCR3 receptor is expressed predominantly on T cells activated by CXCL9, CXCL10, and CXCL11 (Ransohoff, 2009). A CXCR3 receptor antagonist was recently shown to inhibit glia activation and neuropathic pain and enhance the effectiveness of morphine (Piotrowska et al., 2018). We showed the upregulation of Prss12, Mal, and Mag, but the downregulation of the Na_v beta subunit 4 (Scn4b) 6 h after both HK-1 concentrations. Motopsyn/Prss12 can activate the PAR receptor in astrocytes and trigger glutamate release to activate neuronal NMDA receptors (Lee et al., 2007; Wójtowicz et al., 2015). MAL is a protein predominantly expressed by oligodendrocytes and Schwann cells and inhibits peripheral nerve myelination (Buser et al., 2009). Voltage-gated sodium channels (Na_v) initiate the action potential in excitable cells, and their mutations are implicated in chronic pain (Namadurai et al., 2015).

Figure 5 represents a summary of DE genes in different conditions. Both HK-1 concentrations resulted in more downregulated than upregulated genes at 24 h. Rph3a, Gabra2, Ryr2, Mag, and Scn1a were downregulated, while Hacd2 was upregulated. Hacd2 plays an important role in long-chain fatty acid biogenesis involved in neuronal membrane functions. Tachykinin receptors interact with Ryr2 in the endoplasmic reticulum, releasing calcium (Lin et al., 2005). The interaction of Rph3A with the NMDA receptor in hippocampal neurons plays a crucial role in synaptic retention and long-term potentiation. Moreover, Rph3A can also interact with AMPA receptors (Franchini et al., 2022). Notably, we have found remarkable upregulation of Fzd1 expressed by astrocytes and involved in their cross-talks (L'Episcopo et al., 2011, 2018). Increased levels of myelin-associated proteins, including Mag and Mbp proteins, were demonstrated through Activin A oligodendroglial ACVR1B-mediated white matter remyelination after ischemic stroke in mice (Zheng et al., 2021).

One of the most interesting issues is assigning potential targets for HK-1. Interestingly, the Tacr genes encoding tachykinin receptors were expressed around the detection limit in the primary sensory neuron cultures. These results are similar to the dorsal root ganglia data from Linnarson and co-workers (http://linnarssonlab.org/drg/). In human primary sensory neurons, Tac receptors were also detected at a low level (LaPaglia et al., 2018). Tachykinin receptors with three NK1R isoforms are expressed in primary sensory neurons of the rat trigeminal system (Beaujouan et al., 1999, 2002; Page, 2005; Garcia-Recio and Gascón, 2015; Edvinsson et al., 2021), which we could not detect by our sequencing technique in the cultures derived from neonatal rats. The lack of a functionally active NK1 receptor in our system is supported by no calcium influx in response to SP treatment (Borbély et al., 2020).

HK-1 at 1 μM induced calcium signaling-related transcriptomic alterations 6 h later, which is supported by our earlier fluorescent calcium influx data, which was not inhibited either by the NK1 receptor antagonist CP99994 or NK1R deletion (Borbély et al., 2020). In calcium signaling processes, F2r and EGFR receptors were DEs (Figures 6, 7). F2r/PAR1 was downregulated, while the EGFR was upregulated. NK1R signaling by SP can activate tyrosine kinase receptors such as EGFR, and PAR1 can activate EGFR (Castagliuolo et al., 2000; Arora et al., 2008). This is supported by our earlier in vivo findings that in HK-1-deficient mice, the PAR agonist mast cell tryptase-induced arthritic pain was decreased (Borbély et al., 2020). Mast cells were shown to be involved in fibromyalgia-related pain (Theoharides et al., 2019). Interestingly, Nr4a1 was downregulated 6 h after 1 μM and upregulated 24 h after 500 nM HK-1. Nr4a1 is an orphan nuclear receptor (Safe et al., 2016) involved in innate and adaptive immune responses in leukocytes (Shaked et al., 2015), which is downregulated and inactivated during neuroinflammation (Palumbo-Zerr et al., 2015; Shaked et al., 2015; Xiong et al., 2020). NR4a1 suppresses inflammatory responses driven by interferon (IFN) and nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) signaling (Freire and Conneely, 2018) and microglia polarization (Rothe et al., 2017). Calcium influx might inhibit CXCL9 production and consequently decrease T cell activation (Mishra and Lal, 2021). Downregulated Slc25a5 might lead to mitochondrial dysfunction (Babenko et al., 2018), which is involved in several chronic pain conditions (Meeus et al., 2013; van den Ameele et al., 2020; Yousuf et al., 2020) related to reactive oxygen species generation and nociceptor sensitization (Salvemini et al., 2011). Intracellular calcium homeostasis, regulated mainly by the mitochondria, endoplasmic reticulum, and nucleus, determines neuronal excitability. Nuclear calcium sensor 1 (NCS-1) was suggested to modulate the expression of mitochondrial genes (Simons et al., 2019) together with the ATP sensor transporter Slc25a5 (Zhang et al., 2009). Dysfunction of Grin1, an NMDA receptor subunit, has been associated with several neurological disorders (Intson et al., 2019). Gria 2 is an AMPA receptor subunit having a major role in calcium permeation and voltage rectification; its abnormal functions also play a role in neural pathophysiologies (Salpietro et al., 2019). Downregulation of Grin1 and Gria 2 might decrease calcium concentration after a longer time, influencing neuronal survival (Guo and Ma, 2021). Several TRP channels are involved in sensory functions (Nilius et al., 2003; Voets et al., 2004, 2005; Nilius and Sage, 2005), and their expression or functional changes alter neuronal responsiveness (Nilius and Owsianik, 2011). TRPM 3,7,8 were downregulated 24 h after 500 nM HK-1, which is the opposite of what we earlier found in response to another sensory neuropeptide, PACAP (Takács-Lovász et al., 2022). Notably, PACAP was downregulated 6 h after 500 nM HK-1, suggesting a potential link between the effects of these two peptides on TRP channel expression. The mitochondrial complex I Ndufb6 subunit was significantly downregulated, similar to PACAP (Takács-Lovász et al., 2022). Figure 6 demonstrates these findings in network.

Our previous in vivo data demonstrated the pain-mediating role of HK-1 in several acute and chronic orofacial, neuropathic, and inflammatory pain models (partial sciatic nerve ligation, acute, and chronic CFA and mast cell tryptase-induced, as well as K/BxN serum-transfer arthritis models) (Borbély et al., 2013, 2020; Aczél et al., 2018; Hunyady et al., 2019). Although it is difficult to directly compare the present results with in vivo conditions, several similarities can be determined between gene expressions in the mouse TRG after adjuvant-induced orofacial inflammation and the present TRG culture experiment following HK-1 treatment. We earlier showed that the HK-1 encoding Tac4 gene was upregulated in the TRG in the adjuvant-induced model. Meanwhile, fibroblast growth factor 13 (Fgf13, also termed as glia activating factor) was downregulated in wild-type mice, but upregulated in HK-1-deficient ones (Aczél et al., 2020). This well correlates with the Fgf9 downregulation we observed in the present study in response to 1 μM 6 h later. Fgfs are involved in neuron–glia interactions and glia proliferation, mediating neuroinflammatory mechanisms (Stork et al., 2014). Furthermore, an orphan G protein-coupled receptor, Gpr62, playing a role in axo-myelinic signaling, was upregulated in the TRG of adjuvant-treated wild-type mice (Hay et al., 2021) similar to Gpr108 upregulation after 1 μM HK-1 in the present experiment to inhibit Toll-like receptor-triggered inflammatory responses (Dong et al., 2018). The voltage-gated K⁺ channel-interacting protein 9 (Kcnj9) was also downregulated in the TRG after adjuvant injection (Aczél et al., 2020), which supports our findings on the Kcnip4 downregulations in the present model. Inward-rectifying K+ currents have been shown to change in the trigeminal ganglia during peripheral inflammation (Takeda et al., 2011). Transmembrane protein 100 (Tmem100) was upregulated in the TRG in the in vivo mouse model, while in our cell system the expressions of related members of this protein family, Tmem128, significantly increased. Integrin subunit alpha 7 (Itga7) involved in glia proliferation (Tan et al., 2022), was also downregulated both in the mouse model (Aczél et al., 2020) and in the present cell culture study Itgav both are expressed on glial cells (Mapps et al., 2022). RNA-Seq literature data demonstrated Kcnj9, F2rl2 (PAR3) upregulation in the DRG in a neuropathic pain model similar to what we found for F2r and PAR1 (Stevens et al., 2020). Validation of the RNA sequencing data with qPCR showed a good correlation for all investigated genes except Fgfr1. However, such discrepancies are not unusual based on the literature data; others described that these two techniques provide similar results in 87.9% (Protasio et al., 2013) and 95.2% of genes (Liu et al., 2014), which is in agreement with the present findings.

In conclusion, this is the first approach to determining transcriptomic alterations induced by HK-1 in primary sensory neurons that might be related to their activation mechanisms and pain. We showed concentration- and time-dependent actions and identified potential pain-related processes such as microglial–neuron or GPCR interactions, myelin-associated gene expression, and orphan GPCRs having roles in GABAergic pathways by modifying NMDA and AMPA receptor subunits, demonstrating that not just neurons, but mast cells and glia cells might be affected by HK-1. The effects specific to the 1 μM (sodium channel activity, neurotransmitter receptor activity, cytokine activity) were potentially calcium influx-dependent, as we earlier showed that the 1 μM, but not the 500 nM concentration, induced a calcium signal in this experimental paradigm. Although the target(s) of HK-1 cannot be determined on the basis of the present results, it is not likely to be the NK1 tachykinin receptor. Identifying the receptor might open novel perspectives in analgesic research.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be at: https://www.ebi.ac.uk/ena, PRJEB47291.

The animal studies were approved by the Ethical Committee on Use of Laboratory Animals at the University of Pécs (permission no: BA02/2000-51/2017). The studies were conducted in accordance with the local legislation and institutional requirements. Written informed consent was obtained from the owners for the participation of their animals in this study.

ÉB, ÉS, KB, and ZH: conceptualization. TA and KB: methodology. JK: software. JK, KT-L, LC, and PU: validation. KT-L and JK: formal analysis, data curation, and visualization. ZH: investigation and supervision. AG and ZH: resources and funding acquisition. KT-L: writing—original draft preparation. KT-L, ÉB, ÉS, KB, and ZH: writing—review and editing. All authors contributed to the article and approved the submitted version.