Title page
Central relaxin-3 receptor (RXFP3) activation increases ERK phosphorylation in
septal cholinergic neurons and impairs spatial working memory.
Héctor Albert-Gascó
1
, Álvaro García-Avilés
1
, Salma Moustafa
1
, Sandra Sánchez-Sarasua
1
,
Andrew L. Gundlach
2,3
, Francisco E. Olucha-Bordonau
1
*, Ana M. Sánchez-Pérez
1
*
1
Department of Medicine, School of Medical Sciences, University Jaume I, 12071 Castellón de la
Plana, Spain.
2
The Florey Institute of Neuroscience and Mental Health, Parkville, Victoria 3052, Australia
3
Florey Department of Neuroscience and Mental Health, The University of Melbourne, Victoria
3010, Australia
* Equal contribution
AMSP tel +34964387459; email: sanchean@uji.es
FEOB tel: +34964387460; email: folucha@uji.es
2
ABSTRACT
The medial septum/diagonal band (MS/DB) is a relay region connecting the hypothalamus and
brainstem with the hippocampus, and both the MS/DB and dorsal/ventral hippocampus receive
strong topographic GABA/peptidergic projections from the nucleus incertus of the pontine
tegmentum. The neuropeptide relaxin-3, released by these neurons, is the cognate ligand for a
G
i/o
-protein-coupled receptor, RXFP3, which is highly expressed within the MS/DB, and both
cholinergic and GABAergic neurons in this region of rat brain receive relaxin-3 positive
terminals/boutons. Comprehensive in vitro studies have demonstrated that a range of cell
signaling pathways can be altered by RXFP3 stimulation, including inhibition of forskolin-
activated cAMP levels and activation of ERK phosphorylation. In this study we investigated
whether intracerebroventricular (icv) injection of RXFP3-A2, a selective relaxin-3 receptor
agonist, altered ERK phosphorylation levels in the MS/DB of adult male rats. In addition, we
assessed the neurochemical phenotype of phosphorylated (p) ERK-positive neurons in MS/DB
after RXFP3-A2 administration by dual-label immunostaining for pERK and key neuronal
markers. RXFP3-A2 injection significantly increased pERK levels in MS/DB, compared to
vehicle at 20 and 90 min post-injection. In addition, icv injection of RXFP3-A2 increased the
number of cells expressing pERK in the MS/DB after 90 min, with increases detected in
cholinergic, but not GABAergic neurons. Moreover, we found that septal cholinergic neurons
express RXFP3 and that icv infusions of RXFP3-A2 impaired alternation in a spatial working
memory behavioral paradigm. The presence of the receptor and the specific RXFP3-related
activation of the MAPK/ERK pathway in MS/DB cholinergic neurons identifies them as a key
target of ascending relaxin-3 projections with implications for the acute and chronic inhibition of
cholinergic neuron activity/function by relaxin-3/RXFP3 signaling.
Key words: Calcium-binding proteins, Choline acetyltransferase, GABA neurons, MAPK/ERK
pathway, Nucleus incertus, Septum, working spatial memory, RXFP3 labeling
3
INTRODUCTION
The septal area is involved in the regulation of behavioural processes of arousal attention and
spatial navigation/exploration particularly via connections from the medial septum/diagonal band
(MS/DB) to the hippocampus (Vertes and Kocsis 1997). Functionally, attention, arousal, and
locomotion related to navigation and exploration and mnemonic processes in humans (Morales et
al. 1971; Bannerman et al. 2004) are associated with hippocampal theta rhythm, a synchronous 4-
12 Hz oscillation of primarily principal neurons and with place cell configuration (Kemp and
Kaada 1975; O’Keefe 1976; Leung and Yim 1986; Raghavachari et al 2001; Hasselmo 2005;
Buzsáki and Moser 2013; Fuhrmann et al 2015). Notably, hippocampal theta rhythm can be
generated and modulated from the septum (Morales et al 1971; Bannerman et al 2004), and
different types of neurons within the MS/DB participate in this process (Sotty et al 2003).
Septal cholinergic neurons are slow-firing neurons that promote hippocampal theta rhythm (Sotty
et al 2003; Yoder and Pang 2005; Vandecasteele et al 2014) and are responsible for transient
arousal states and hippocampal activation (Zhang et al 2011). Septal GABAergic parvalbumin
(PV)-positive neurons are fast-firing (Sotty et al 2003; Yoder and Pang 2005; Vandecasteele et al
2014) and inhibit inhibitory hippocampal interneurons (Freund and Antal 1988; Toth et al 1997;
Freund and Gulyas 1997; Hangya et al 2009). Resultant disinhibition of hippocampal granular
and pyramidal cells promotes and facilitates synchronicity in theta frequency (Freund and Antal
1988; Tóth et al. 1997; Freund and Gulyás 1997; Hangya et al. 2009b). In addition, the majority
of septal glutamatergic neurons are also fast firing and have been reported to drive hippocampal
pyramidal cells (Huh et al 2010). Furthermore, septal glutamatergic neurons were recently
reported to excite interneurons at the CA1 stratum oriens/ alveus border in hippocampus, which
regulate feedforward inhibition of Schaffer collateral and perforant path input to CA1 pyramidal
neurons in a locomotion-dependent manner (Fuhrmann et al 2015).
Various strategies have been used to study the role of the MS/DB and its effect in memory
processes. Most of studies are centered on activation or deactivation of specific cell types within
this region. For instance, time dependent increases in acetylcholine levels have been observed in
the hippocampus after acquisition of spatial memory tasks such as the T-maze or spontaneous
4
alternation task (Hepler et al 1985; Ragozzino et al 1996; Fadda et al 1996). In contrast, lesion
studies of the MS/DB have reported to disrupt spatial working memory but not nonspatial
working memory (Kelsey and Vargas 1993). On no case did the lesions impair working memory
on reference memory visual discrimination tasks or simultaneous conditional discrimination task
(Thomas and Gash 1986). In addition, when lesions are not electrolytic but with specific
excitotoxins which target cholinergic neurons, an impairment in a variety of spatial working
memory tasks is produced (Johnson et al 2002; Gibbs and Johnson 2007; Fitz et al 2008).
The MS/DB receives strong connections from mesencephalic and brainstem areas, including the
posterior hypothalamus and supramammillary nucleus and brainstem, particularly from the
nucleus incertus (NI) Goto et al. 2001; Olucha-Bordonau et al. 2003; Ryan et al. 2011; Olucha-
Bordonau et al. 2012). Previous studies have shown that activation of these areas alter
hippocampal theta rhythm and locomotion. For example, in urethane-anesthetized rats, electrical
stimulation of the NI induces an increase in hippocampal theta rhythm and electrolytic lesion of
the NI abolishes the hippocampal theta induced by stimulation of the reticularis pontis oralis
(RPO) region (Nunez et al 2006). Moreover, infusion of R3/I5, a specific RLN3 agonist, into the
medial septum promotes theta rhythm (Ma et al 2009). In addition, it has been proposed that the
nucleus incertus may relay a general stress response over the telencephalon centers involved in
memory processes (Rajkumar et al 2016) and feeding behavior (Calvez et al 2016).
The majority of NI neurons in the rat synthesize and release GABA (Ford et al 1995; Olucha-
Bordonau et al 2003; Ma et al 2007) and a significant population of these neurons express the
neuropeptide, relaxin-3, a member of the insulin/relaxin superfamily (Bathgate et al 2002;
Burazin et al 2002; Ma et al 2007). The largest number of relaxin-3 neurons is located in the NI,
but they are also present in the ventral periaqueductal grey, the pontine raphe nucleus and an area
dorsal to the lateral substantia nigra (Tanaka et al 2005; Ma et al 2007). Relaxin-3 is the single
cognate of the G
i/o
-protein-coupled receptor, RXFP3 (Liu et al 2003), and RXFP3 mRNA and
binding sites are strongly expressed in the brain in a topographical distribution that aligns with
that of relaxin-3 containing axons and nerve terminals in rat (Sutton et al 2004; Ma et al 2007)
and mouse (Smith et al 2010) brain.
5
In vitro studies in various cell lines (Chinese hamster ovary (CHO) cells and human embryonic
kidney (HEK) 293 cells) stably expressing human RXFP3 and SN56 cholinergic neuroblastoma
cells, which display endogenous RXFP3 expression, demonstrate that RXFP3 stimulation can
consistently decrease forskolin-stimulated cAMP levels and activate the MAPK/ERK pathway, as
reflected by changes in ERK1/2 phosphorylation and activation of immediate early
gene/transcription factor expression (e.g. activator protein 1, AP-1, nuclear factor-κB, NF-κB and
serum response element, SRE). However, while central relaxin-3 injections have been associated
with changes in immediate early gene expression and there are preliminary reports of RXFP3
activation producing hyperphosphorylation of putative RXFP3-positive neurons in rat brain slices
in vitro (Kania et al 2014) consistent with inhibition of cellular cAMP levels by Gi/o coupled
receptors, no reports exist on the effect of central RXFP3 activation in vivo on relevant cell
signaling pathways.
Therefore, in this study, we investigated the effect of RXFP3 stimulation in rat brain by central
administration of the selective relaxin-3 receptor agonist, RXFP3-A2 (Shabanpoor et al 2012) on
MAPK/ERK pathway-related signaling in the MS/DB. Cerebellum, which lacks RXFP3
expression (Sutton et al 2004), was used as a negative control tissue. Using immunoblotting, we
quantified phosphorylated ERK (pERK) and total ERK levels, as described in earlier in vitro
studies (van der Westhuizen et al 2007), to assess the impact of RXFP3 activation on overall
septal MAPK/ERK activation. Subsequently, we used immunofluorescence staining to
characterize the neurochemical phenotype of pERK-positive neurons by co-localizing pERK
immunoreactivity with choline acetyltransferase as a marker for cholinergic neurons and the
calcium-binding proteins (CaBP), parvalbumin, calretinin and calbindin for different populations
of septal GABAergic neurons (Olucha-Bordonau et al 2012). In addition, we have assessed the
neuronal target of RLN3 by double immunofluorescence of RXFP3 and medial septal markers.
Finally, we have studied the behavioral effect of A2 icv administration using a working spatial
memory task.
Our data demonstrate that cholinergic neurons in the MS/DB express RLN3 receptor, RXFP3.
Activation of RXFP3 by agonist infusion increased ERK phosphorylation in cholinergic neurons