Available online at www.sciencedirect.com

www.elsevier.com/locate/brainres

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0

0006-8993/$ - see frohttp://dx.doi.org/10

☆This is an openWorks License, whisource are credited

nCorrespondenceNational University

E-mail address: g

Research Report

Selective lesioning of nucleus incertus withcorticotropin releasing factor-saporin conjugate$

Liying Corinne Leea,b,c, Ramamoorthy Rajkumara,b,c,Gavin Stewart Dawea,b,c,n

aDepartment of Pharmacology, Yong Loo Lin School of Medicine, National University Health System, 117597, SingaporebNeurobiology and Ageing Programme, Life Sciences Institute, National University of Singapore, 117456, SingaporecSingapore Institute for Neurotechnology (SINAPSE), 117456, Singapore

a r t i c l e i n f o

Article history:

Accepted 19 November 2013

The nucleus incertus (NI), a brainstem nucleus found in the pontine periventricular grey, is

the primary source of the neuropeptide relaxin-3 in the mammalian brain. The NI neurons

Available online 25 November 2013

Keywords:

Nucleus incertus

Relaxin-3

CRF–saporin

Lesion

Rat

Fear conditioning

nt matter & 2013 The Au.1016/j.brainres.2013.11.02

-access article distributech permits non-commer.to: Department of Pharmof Singapore, 117456, Siavin_dawe@nuhs.edu.sg

a b s t r a c t

have also been previously reported to express several receptors and neurotransmitters,

including corticotropin releasing hormone receptor 1 (CRF1) and gamma-aminobutyric acid

(GABA). The NI projects widely to putative neural correlates of stress, anxiety, depression,

feeding behaviour, arousal and cognition leading to speculation that it might be involved in

several neuropsychiatric conditions. On the premise that relaxin-3 expressing neurons in

the NI predominantly co-express CRF1 receptors, a novel method for selective ablation of

the rat brain NI neurons using corticotropin releasing factor (CRF)–saporin conjugate is

described. In addition to a behavioural deficit in the fear conditioning paradigm, reverse

transcriptase polymerase chain reaction (RT-PCR), western blotting (WB) and immuno-

fluorescence labelling (IF) techniques were used to confirm the NI lesion. We observed a

selective and significant loss of CRF1 expressing cells, together with a consistent decrease

in relaxin-3 and GAD65 expression. The significant ablation of relaxin-3 positive neurons of

the NI achieved by this lesioning approach is a promising model to explore the

neuropsychopharmacological implications of NI/relaxin-3 in behavioural neuroscience.

& 2013 The Authors. Published by Elsevier B.V. All rights reserved.

thors. Published by Elsevier B.V. All rights reserved.1

d under the terms of the Creative Commons Attribution-NonCommercial-No Derivativecial use, distribution, and reproduction in any medium, provided the original author and

acology, Yong Loo Lin School of Medicine, Centre for Life Sciences, Level 5, 28 Medical Drive,ngapore.(G.S. Dawe).

1. Introduction

The nucleus incertus (NI), better known for its status asthe principal source of neuronal relaxin-3, is a brainstemnucleus located ventral and medial to the posterodorsal

tegmental nucleus (PDTg) in the pontine periventricular grey(Burazin et al., 2002; Goto et al., 2001; Olucha-Bordonau et al.,2003). The NI neurons were initially shown to express theinhibitory neurotransmitter, gamma-aminobutyric acid(GABA) (Ford et al., 1995) and have later been revealed to

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0180

co-express the recently discovered neuropeptide, relaxin-3(Ma et al., 2007). This small and distinct group of neurons hassparked off renewed interest due to its high expression of thecorticotropin releasing hormone receptor type 1 (CRF1)(Bittencourt and Sawchenko, 2000; Potter et al., 1994), whichsuggests a role of the NI in the stress response. Subsequentstudies reported the presence of other receptors in the NI, e.g.5-hydroxytryptamine (5-HT or serotonin) receptor subtype 1A(5-HT1A) (Miyamoto et al., 2008), and relaxin/insulin-likefamily peptide receptor 3 (RXFP3) (Sutton et al., 2004). Otherneuropeptides expressed by nucleus incertus include neuro-medin B (Chronwall et al., 1985; Wada et al., 1990) andcholecystokinin (Kubota et al., 1983; Olucha-Bordonau et al.,2003). Of particular interest, expression of relaxin-3 is highlyspecific and most abundant in the NI and only a smallnumber of neurons in other nearby regions such as thepontine raphe nucleus, periaqueductal grey and dorsal sub-stantia nigra express relaxin-3 (Tanaka et al., 2005). As such,the NI – serves as a key point of regulation for functioning ofthe relaxin-3 neural circuitry.

The NI possibly exerts its actions through numerousprojections to several parts of the brain, including theprefrontal cortex, medial septum (MS), hippocampus,amygdala, hypothalamus and the raphe nuclei (Gotoet al., 2001) – areas that are implicated in various psychia-tric conditions. Relaxin-3 binds to the RXFP3, whichbelongs to the family of relaxin peptide receptors thatwas recently elevated from orphan receptor status (vander Westhuizen et al., 2008). The precise functions of the

Fig. 1 – Selection of CRF–saporin dose for ablation of NI neuronscells in the NI of (A) naïve control (n¼3), (B) 21.5 ng/site CRF–sapo(D) 86 ng/site CRF–SAP injected (n¼3) rats. The 86 ng/site (172 ngCRF1 expressing cells in the NI. Scale bars are 100 μm.

NI and relaxin-3 remain largely undetermined but intri-guing reports have been published in the past decadedemonstrating the alleged role of the NI/relaxin-3/RXFP3system in feeding behaviour (McGowan et al., 2006; McGowanet al., 2005), stress (Burazin et al., 2002; Smith et al., 2012),anxiety (Watanabe et al., 2011) and cognition (Cervera-Ferriet al., 2011; Farooq et al., 2013; Ma et al., 2009). Recent studiesshow c-Fos expression in the nucleus incertus occurs inresponse to acute antipsychotic treatments in rats (Rajkumaret al., 2013) and polymorphisms in human relaxin-3 and RXFP3associated with metabolic disturbances in patients with schizo-phrenia treated with antipsychotic drugs (Munro et al., 2012).Thus the study of the NI and relaxin-3 is an exciting newfrontier in behavioural neuroscience.

A strategy to achieve potent and selective lesioning oftarget brain structures has been to utilise cell-surface proteinbinding peptides or antibodies conjugated with saporin, amonomeric ribosomal inactivating protein (Heckers et al.,1994; Li et al., 2008; Thorpe et al., 1985; Waite et al., 1994).Selectivity is achieved because, as a ribosomal toxin, thesaporin is only toxic when internalised by the correspondingreceptor. The corticotropin releasing factor (CRF)–saporinconjugate toxin, used in the present study, is expected toselectively ablate CRF1 expressing cells (Hummel et al., 2010;Maciejewski-Lenoir et al., 2000). On the premise that relaxin-3expressing neurons in the NI predominantly co-express CRF1receptors (Tanaka et al., 2005), the present investigationattempted to establish a method for selective ablation ofthe NI using the CRF–saporin conjugate.

. Images show representative examples of the CRF1 positiverin injected (n¼3), (C) 43 ng/site CRF–SAP injected (n¼3) andin total) dose of CRF–SAP produced the greatest reduction in

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0 181

2. Results

Out of the total of 76 rats that underwent the surgicalprocedure, 43 receiving CRF–saporin and 33 serving as variouscontrols, no mortality attributable to the CRF–saporin lesion

Fig. 2 – Specificity of the CRF RI/II antibody in the NI. (A) Represstaining of the NI in a rat (n¼3). (B) Representative example of imthe CRF RI/II antibody with a 10� relative concentration of CRFpeptide blocks the immunofluorescence staining demonstrating

Fig. 3 – RT-PCR and real-time PCR analysis of CRF–saporin lesionof blank saporin, while lesions of the NI were produced by injecthe decreased expression of CRF1, relaxin-3 (RLN3) and GAD65 in(n¼4). Equivalent expression of TPH2 indicated that CRF–SAP tadensitometry analysis of the DNA gel relative to beta-actin. (C) Rand TPH2 mRNA expression changes between sham- (n¼4) andexpression decreased significantly after CRF-saporin targeted lelesion. Data are expressed as mean7SEM. *po0.05,**po0.01, ***

was observed. Two rats were euthanised under veterinaryadvice because of an unrelated infection and a case ofmalocclusion of the incisors. In one experiment, post-surgical weight gain was monitored daily over 14 days butthere was no significant difference in weight gain betweenthe sham- and NI-lesioned rats (n¼8 per group, n.s.).

entative example of CRF RI/II antibody immunofluorescencemunofluorescence staining after overnight preabsorption ofblocking peptide (n¼3). The preabsorption with blockingspecificity of the antibody. Scale bars are 100 μm.

ing of NI neurons. Sham lesions were produced by injectiontion of CRF–saporin. (A) Representative DNA bands showingthe NI-lesioned rats (n¼4) compared to sham-lesioned rats

rgeted only the CRF1 positive cells. (B) Semi-quantitativeeal-time RT-PCR analysis of CRF1, relaxin-3 (RLN3), GAD65NI-lesioned (n¼4) groups. CRF1, relaxin-3 and GAD65 mRNAsioning. TPH2 mRNA expression was not affected by thepo0.001.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0182

2.1. 172 ng of CRF–saporin brings about a loss in CRF1expressing cells

To determine an appropriate dose of CRF–saporin, 40� , 20�and 10� dilutions of the original stock solution of CRF–saporin were infused separately into the NI of rats. CRF1immunofluorescence staining results showed that infusion of172 ng of CRF–saporin was sufficient to bring about a loss inCRF1 expressing cells in the NI (Fig. 1A–D). This dose istherefore used for the subsequent experiments. The specifi-city of the CRF RI/II antibody was assessed by preabsorptionof the antibody with the CRF blocking peptide, which abol-ished CRF1 staining in the NI of naïve rats (Fig. 2A–B).

2.2. CRF–saporin selectively targets and ablates CRF1positive cells

RT-PCR analysis showed that the NI-lesioned rats had asignificant reduction in the expression of CRF1 receptorscompared to the sham-lesioned group. As hypothesised,corresponding decreases in the expression of relaxin-3 andGAD65 were also observed in the NI-lesioned rats (Fig. 3A).TPH2 expression was unaltered in both the sham andNI-lesioned group as seen in the densitometry analysis ofthe PCR bands (Fig. 3B). In a separate group of animals, a real-time PCR analysis showed that the CRF1, relaxin-3 and GAD65mRNA expression in NI-lesioned rats was 0.004-, 0.02- and0.07 fold relative to the sham-lesioned controls while TPH2mRNA levels remained almost the same in both sham- andNI-lesioned rats (Fig. 3C).

The protein expression assessed by western blotting ofCRF1, pro-relaxin-3, GAD65 and TPH2 was similar among thenaïve, saline injected, true sham (surgery but no infusion)and sham-lesioned (blank saporin infusion) groups (Fig. 4Aand B). However, protein levels of CRF1 receptor was con-siderably reduced in the NI-lesioned rats. There was also aconsistent decrease in the levels of pro-relaxin-3 in the NI-lesioned rats as compared to the sham-lesioned rats (Fig. 4A).The same set of samples was also checked for GAD65 andTPH2 protein levels. A significant decrease in the expressionof GAD65 was also observed in the NI-lesioned rats whileTPH2 protein levels remained unchanged. Densitometry ana-lysis of the western blot demonstrated a statistically signifi-cant reduction in CRF1 (approximately 54%), pro-relaxin-3(approximately 53%) and GAD65 (approximately 64%) proteinexpression (Fig. 4B).

Further confirmation of the lesion through immunofluor-escence labelling verified the loss of CRF1 receptor positivecells in the NI of the NI-lesioned rats as compared to thesham-lesioned rats (Fig. 5A and B). Similarly, relaxin-3expressing cells also decreased considerably in the NI-lesioned rats (Fig. 5D). Examination of the TPH2 expressionrevealed that TPH2 positive cells lined the midline of the NIand were unaffected by the lesioning procedure (Fig. 5E and F).Positive CRF1 staining of the nearby locus coeruleus (LC) inboth sham- and NI-lesioned rats was also detected (Fig. 5Gand H). Immunostaining for the glial cell marker, glial fibrillaryacidic protein (GFAP), showed up-regulation of glial cells in theNI of both sham-lesioned and NI-lesioned rats at 14 days aftersurgery (Fig. 6).

2.3. CRF–saporin in the NI decreases levels of relaxin-3in the medial septum

The levels of pro-relaxin-3 in the MS in naïve, saline, truesham, sham- and NI-lesioned rats were assessed by westernblotting. A decrease in relaxin-3 levels was observed in theMS of NI-lesioned rats (Fig. 7A). Densitometry analysis of theblots showed an approximately 90% decrease in pro-relaxin-3levels (Fig. 7B).

2.4. NI-lesioned rats displayed increased freezing in cuedfear conditioning

The NI-lesioned and sham-lesioned rats were tested in a cuedfear conditioning paradigm. The rats were first exposed totone-shock pairing and subsequently freezing behaviour inresponse to the tone was assessed 24 h later. To determine ifthe NI-lesioned rats had any locomotor deficits, the totaldistance travelled by the sham- and NI-lesioned rats weremeasured during the 15 min habituation phase. There was aslight but insignificant decrease in the distance covered bythe NI-lesioned rats (Fig. 8A) possibly due to the increasedperiods of freezing observed during this phase (Fig. 8B).No significant difference between the percentage of freezingbetween sham- and NI-lesioned rats was observed during thetone-shock training phase (Fig. 8C), for 2 min before (Fig. 8D)and during the 30 s tone at the 24 h test phase (Fig. 8D).However, NI-lesioned rats displayed significantly greaterfreezing behaviour than the sham-lesioned rats in the 5 minperiod following the tone during the test phase (Fig. 8D).

3. Discussion

We describe a method to ablate the NI neurons. CRF-saporin,lesioning neither caused mortality nor alteration in feed andwater consumption, which is in agreement with the findingson relaxin-3 knockout mice (Smith et al., 2012; Smith et al.,2009; Watanabe et al., 2011). Our results show that infusion of172 ng of CRF–saporin was sufficient to bring about a sig-nificant loss in CRF1 expressing cells in the NI. This was inaccordance to the findings by Pascual's group where 1–2 mg ofCRF–saporin injected ICV resulted in a significant loss in CRF1positive cells in the fundus of the striatum (FS) and lateralseptum (LS) (Pascual and Heinrich, 2007), suggesting thatCRF–saporin was able to target and permanently silenceCRF1 expressing cells in the brain. Unconjugated saporin didnot affect expression of CRF1 as seen in the sham-lesion ratgroup, which was in agreement with the notion that thesaporin protein alone does not bind to any receptors andcannot be taken in by the cells (Stirpe et al., 1983;Maciejewski-Lenoir et al., 2000).

Previous evidence has shown that all relaxin-3 expressingcells in the NI co-express CRF1 and can be activated by ICVadministration of CRF (Tanaka et al., 2005; Banerjee et al.,2010), thus this study investigated the expression of relaxin-3in the NI cells after the CRF-saporin targeted lesion. Theconsistent decrease in relaxin-3 expression corresponded tothe findings that the NI cells express both CRF1 and relaxin-3.

Fig. 4 – Western blot analysis of CRF–saporin lesioning of NI neurons. The naïve group (n¼3) did not undergo any surgicalprocedures. Saline rats (n¼3) received bilateral injections of 0.2 ll of saline. True sham lesions (n¼3) were produced byinserting the needle containing CRF-SAP into the NI without infusion. Sham lesions were produced by injection of blanksaporin (n¼6) while lesions of the NI (n¼7) were produced by injection of CRF–saporin. (A) Representative western blot bandsshowing the decreased levels of CRF1, pro-relaxin-3 (Pro-RLN3) and GAD65 in the NI-lesioned rats. The absence of change inthe expression of TPH2 indicated that CRF–saporin targeted only the CRF1 positive cells. (B) Semi-quantitative densitometryanalysis of the western blot relative to actin. Data are represented as mean7SEM. *po0.05, **po0.01, ***po0.001.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0 183

Together with the resultant decrease in relaxin-3 levels inone of the known projection targets of the NI, the MS, thesedata indicated that this lesion model is a possible tool for thestudy of relaxin-3 circuitry in the brain. Our lesion model alsodemonstrated a compelling decrease in GAD65 expression, aknown indicator of GABAergic neurons. Relaxin-3 neurons inthe NI were known to co-express GAD65 (Ma et al., 2007), animportant indication that the neurotransmission from the NIis inhibitory. Thus the resulting loss in GAD65 expressionreinforced the findings that NI neurons are GABAergic andalso provides an additional verification that CRF–saporinlesions the NI neurons.

The present method did not completely lesion the NI but wassufficient to produce a clear behavioural deficit. It might bepossible to produce lesions of the NI to a greater extent byinjecting greater volumes or concentrations of CRF–saporin.However, CRF1 receptors are also present in other nearbystructures such as the LC and there is a risk that the selectivityof the lesion will be compromised. Moreover, the NI spans onlyaround 700 μm in the anterior–posterior aspect and hence

multiple injections can cause physical injury to the cells, whichis undesired. The NI and pontine raphe nucleus are both foundcaudal to the 5HT-neurons of the dorsal raphe nucleus.It was previously reported that distributions of 5HT-positiveneurons and relaxin 3-expressing neurons intermingled but didnot overlap (Tanaka et al., 2005). In our lesion model, expressionof TPH2 was unaffected by CRF–saporin infusions and TPH2 cellslining the midline of the NI were intact even at 14 days after theprocedure. This finding reiterates the specificity of CRF–saporinin targeting only the cells that express the CRF1 receptor.A recent paper reported that electrolytic lesion of the nucleusincertus retards extinction of auditory conditioned fear (Pereiraet al., 2013). However, electrolytic lesions lack cellular selectivityand may damage fibres of passage. Here we demonstrated thatCRF–saporin selectively targets and ablates CRF1 positive cellswhile leaving cells without the receptor unharmed, indicatingthe specificity. Moreover, earlier reports showed that CRF–saporin had a greater binding affinity to CRF1 as compared toCRF2α receptors (Maciejewski-Lenoir et al., 2000), rationalising theuse of CRF–saporin to selectively lesion the NI. Moreover, while

Fig. 5 – Immunofluorescence histochemical analysis of CRF–saporin lesioning of NI neurons. Sham lesions were produced byinjection of blank saporin (n¼5) while lesions of the NI (n¼7) were produced by injection of CRF–saporin. Representativephotomicrographs show reduction of (B) CRF1 receptors and (D) relaxin-3 (RLN3) positive cells in NI-lesioned rats compared tosham-lesioned rats (A) and (C), respectively. Presence of TPH2 positive cells in both (E) sham- and (F) NI-lesioned groupsindicated that CRF–saporin was specific to CRF1 receptor positive cells. Positive CRF1 staining of the nearby locus coeruleus(LC) cells in both (G) sham- and (H) NI-lesioned groups showed that the lesion was restricted to the NI. The presence of theadjacent mesencephalic trigeminal nucleus (Me5) confirms the anatomical localisation of LC. Scale bars are 100 μm.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0184

Fig. 6 – Representative photomicrographs of GFAP staining in the NI of (A) naïve (n¼3), (B) sham-lesioned (n¼3) and (C) NI-lesioned rats (n¼3) 14 days after the lesion. There is up-regulation of GFAP in both sham- (B) and NI-lesioned rats (C). Scalebars are 100 μm.

Fig. 7 – Relaxin-3 expression in medial septum wasdetermined by western blotting following CRF–saporintargeted lesioning. The naïve group (n¼3) did not undergoany surgical procedures. Saline rats (n¼3) received bilateralinjections of 0.2 ll of saline. True sham lesions wereproduced by inserting the needle containing CRF–saporininto the NI without infusion. Sham lesions were producedby injection of blank saporin (n¼3) while lesions of the NI(n¼3) were produced by injection of CRF–saporin.(A) Representative western blot bands showing thedecreased expression of pro-relaxin-3 (Pro-RLN3) in the MSof the NI-lesioned rats. (B) Semi-quantitative densitometryanalysis of the western blot relative to actin. Data arerepresented as mean7SEM. *po0.05.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0 185

the NI strongly expresses CRF1 in the rat (Potter et al., 1994;Bittencourt and Sawchenko, 2000; Van Pett et al., 2000) there isno expression of CRF2 (Van Pett et al., 2000).

Although, the NI has been predicted to be involved in avariety of psychiatrically relevant conditions andmanifestations,including stress, anxiety, depression, feeding behaviour, arousaland cognition (Ryan et al., 2011; Smith et al., 2011) thesespeculations are largely founded on the studies that indirectly

infer from RXFP3 distribution in rodent brain (Sutton et al., 2004),relaxin-3-like immunoreactivity, ([125I] R3/I5) binding (Ma et al.,2007; Sutton et al., 2004) and anatomical tract tracing of theafferent and efferent connexions of the NI (Goto et al., 2001;Hoover and Vertes, 2007; Olucha-Bordonau et al., 2003). Currentmethods of studying the functions of the NI include electricalstimulation (Farooq et al., 2013; Nunez et al., 2006), pharmaco-logical activation with CRF (Farooq et al., 2013; Tanaka et al.,2005) and electrolytic lesioning (Pereira et al., 2013) of the NI todetermine its putative modulatory role on mnemonic processingand behaviour. As the NI consists mostly of relaxin-3 positiveneurons, numerous studies also use H3, relaxin-3, relaxin-3agonist/antagonist, relaxin-3 knockout mouse models and silen-cing relaxin-3 in NI neurons to study the various postulatedfunctions of the NI (Callander et al., 2012; Ma et al., 2009; Smithet al., 2012; Smith et al., 2009; Watanabe et al., 2011). The presentmethod to perturb the NI/relaxin-3 system using CRF-saporin isexpected to open an additional approach for research to under-stand relevance of the NI/relaxin-3 system in behaviouralneuroscience.

4. Experimental procedure

4.1. Animals and surgery

Male Sprague–Dawley rats (270–290 g) were procured from theCentre for Animal Resources (CARE), National University ofSingapore, were maintained under standard housing condi-tions (2172 1C, 12 h light-dark cycle and ad libitum food andwater). All experimental procedures were approved by theInstitutional Animal Care and Use Committee (IACUC),National University of Singapore, and were in accordancewith the guidelines of the National Advisory Committee forLaboratory Animal Research (NACLAR), Singapore, and theGuide for the Care and Use of Laboratory Animals, NationalResearch Council of the National Academies, USA. Rats wereanaesthetised with an intraperitoneal injection of a ketamine(75 mg/kg) and xylazine (10 mg/kg) cocktail, placed in astereotaxic frame and burr holes were drilled on the skull atthe coordinate corresponding to the NI (AP: 9.7 mm and ML:0-0.1 mm) (Paxinos and Watson, 2007) calculated from thebregma. Bilateral injections of 0.2 ml/site made 7.5 mm ventralto the surface of the skull delivered 21.5 ng, 43 ng or 86 ng/site

Fig. 8 – Effect of NI lesioning on behaviour of rats in a cued fear conditioning paradigm. (A) Total distance travelled by sham-lesioned (n¼10) and NI-lesioned (n¼8) rats in the fear conditioning chambers and (B) freezing behaviour during the four daysof 15 min habituation. (C) Shows the mean percentage freezing response during the training phase, while (D) and (E) show themean percentage freezing of rats for two minutes before and during the 30 s tone of the test phase (24h after the training). Themean percentage freezing time following the tone during the test phase (F) showed that NI-lesioned rats exhibit a longerfreezing time compared to sham-lesioned rats. All data are represented as mean7SEM, *po0.05, ***po0.001.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0186

of CRF–saporin or blank saporin (Advanced targeting Sys-tems, USA) over 5 min. The needle was left in place for 5 moreminutes before withdrawal. The scalp was sutured and therat was allowed a rehabilitation period of 14 days before anyexperiments were carried out. Saline rats (n=3) receivedbilateral injections of 0.2 ml of saline. True sham lesions wereproduced by inserting the needle containing CRF–saporin intothe NI without infusion. Sham lesions were produced byinjection of blank saporin (n¼6) while lesions of the NI (n¼7)were produced by injection of CRF–saporin. Subsequently, thebrains were freshly harvested (for RT-PCR, real-time PCR or

western blot) or harvested after transcardial perfusion (forimmunofluorescence studies on free floating sections) tocheck for the extent of the lesion. To determine if the lesionof the NI had an effect on behaviour, a separate group ofsham-lesioned (blank saporin) and NI-lesioned (CRF–saporin)rats were subjected to a fear conditioning paradigm.

4.2. Perfusion fixation and fresh tissue harvest

Rats were anaesthetised with an overdose of pentobarbitalprior to transcardial perfusion with 0.9% saline, followed by

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0 187

4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Thebrain was removed immediately and post-fixed overnight at4 1C and then saturated with 30% sucrose in phosphate-buffered saline (PBS). Free floating sections (30 mm) wereobtained with a vibratome (Leica Microsystems, Germany).For qPCR and western blot analysis, the brains were removedimmediately following anaesthesia and 500 mm sections col-lected using a rat brain matrix (Roboz Surgical, USA). Theposition of the NI and MS were confirmed under lightmicroscope, and then collected with a Harris Uni-CoreTM

1 mm micro-punch (Ted Pella Inc, USA) for further analysis.

4.3. Preparation of mouse anti-relaxin-3 antibody

To prepare the mouse anti-relaxin-3 antibody, HK4-144-10cells (Kizawa et al., 2003) were obtained from the Interna-tional Patent Organism Depository (IPOD), National Instituteof Advanced Industrial Science and Technology (AIST), Japan,and first cultured in an antibiotic free GIT medium (WakoPure Chemicals Industries Ltd., Japan). The recovered hybri-doma cells were then allowed to grow vigorously to a logphase. The viability of the cells (490%) was determinedbefore harvesting and freezing for ascities production. Micewere first inoculated intraperitoneally with monoclonalantibody-producing hybridoma cells; thereafter, the ascitesfluid was collected and purified (1st Base, Singapore, ProSciIncorporated, USA). Further purification with Protein A agar-ose (Thermo Scientific, USA) and elution with 0.1 M citratebuffer containing 0.05% sodium azide were carried out. Theeluate was concentrated in a centrifugal filter unit (Millipore,USA), washed with sterile phosphate buffered saline (PBS)and stored as a 1 mg/ml stock containing 0.05% sodium azideand 0.1% glycerol.

4.4. Immunofluorescence histochemistry

Sections were immunostained with a primary antibody solutionwhich contained goat anti-CRF RI/II antibody (SC1757, 1:1000,Santa Cruz, USA), mouse anti-relaxin-3 antibody (1:1000), rabbitanti-tryptophan hydroxylase 2 (TPH2) antibody (AB5572, 1:1000,Chemicon, USA) or rabbit anti-glial fibrillary acidic protein (GFAP)antibody (Z0334, 1:1000, Dako, Denmark). Stained sections weresubsequently visualised with donkey anti-goat Alexa Fluor 555(1:400), donkey anti-mouse Alexa Fluor 488 (1:400) or goat anti-rabbit Alexa Fluor 488 (1:400) secondary antibodies, after whichslides were washed and mounted using Prolong Antifade withDAPI (Invitrogen, Singapore) and viewed with a fluorescencemicroscope. The CRF RI/II polyclonal antibody used here wasraised against the C-terminus of human CRF1, which is con-served across rat and mouse CRF1. Its specificity has beendemonstrated previously by Chen et al. (2000) in experimentsin which pre-incubation with the antigenic peptide abolishedCRF1 signals in western blots and staining in mouse brainsections. In mouse heart sections known to express only CRF2,no staining was observed (Chen et al., 2000). To evaluate thespecificity of this antibody in the NI neurons, a 10� relativeconcentration of the CRF blocking peptide (C-20P, Santa Cruz,USA) was pre-incubated with a 1:1000 dilution of the antibodyovernight at 4 1C. NI sections were then incubated in this

solution overnight and then further processed for CRF RI/IIstaining.

4.5. Reverse transcriptase and real-time polymerase chainreaction

Total RNA was extracted from the tissue and purified accord-ing to the manufacturer's instructions for PureLink RNA minikit (Invitrogen, Singapore). The amount of RNA was quanti-fied with a NanoDrop UV–vis spectrophotometer (ThermoScientific, USA). Approximately 1 mg of total RNA was reversetranscribed with oligo(dT) primers using ImProm-IITM ReverseTranscription system (Promega, USA). The PCR reaction wasperformed using the following sets of primers: CRF1 forwardprimer 5'-ACC GAC CGT CTG CGC AAG TG-3', reverse primer5'-AGC GGA GCG GAC CTC ACT GT-3' (453 bp expected bandsize); relaxin-3 forward primer 5'-CTG CTG CTG GCT TCC TGGGCT CTC CTC GGG GCT CT-3', reverse primer 5'-TCG CTG GCCAGG TGG TCT GTA TTG GCT TCT CCA TCA-3' (218 bpexpected band size) (Chen et al., 2005); GAD65 forward primer5'-GAA CAG CGA GAG CCT GGT-3', reverse primer 5'-CTC TTGAAG AAG CTC ATT GG-3' (362 bp expected band size); TPH2forward primer 5'–TAA ATA CTG GGC CAG GAG AGG-3',reverse primer 5'-GAA GTG TCT TTG CCG CTT CTC-3'(132 bp expected band size); and β-actin forward primer5'-ATC CTG AAA GAC CTC TAT GC-3', reverse primer5'-AAC GCA GCT CAG TAA CAG TC-3' (287 bp expected bandsize). The PCR conditions were conducted at 94 1C for 5 min,30 cycles of 94 1C for 30 s, 60–65 1C for 1 min, 72 1C for 1 minfollowed by 10 min at 72 1C. The PCR products were thenseparated in 1.5% agarose gel, stained with ethidium bro-mide, and then visualised under UV irradiation.

Real-time RT-PCR amplification on a separate group ofsham- (n¼4) and NI-lesioned (n¼4) was carried out using theViiA 7 Real-time RT-PCR system (Applied Biosystems, USA)with SYBR green PCR master mix (Applied Biosystems, USA)with the following gene-specific primers: CRF1 forward primer5'-ACG ACA AAC AAT GGC TAC CG-3', reverse primer 5'-GCAGTG ACC CAGGTA GTT GA-3'; relaxin-3 forward primer 5'–CCCTAT GGG GTG AAG CTC TG-3', reverse primer 5'-CCA GGTGGT CTG TAT TGG CT-3'; GAD65 forward primer 5'-GGG GTGGAG GGT TAC TGA TG-3', reverse primer 5'-ACC CAT CATCTT GTG GGG AT-3'; TPH2 forward primer 5'-GCC TTT GCAAGC AAG AAG GT-3', reverse primer 5'-CGC CTT GTC AGAAAG AGC AT-3'; and β-actin forward primer 5'-ATC CTG AAAGAC CTC TAT GC-3', reverse primer 5'-AAC GCA GCT CAGTAA CAG TC-3'. β-actin was used as an internal control. ThePCR conditions were an initial incubation of 50 1C for 2 minand 95 1C for 10 min followed by 40 cycles of 95 1C for 15 s and59 1C for 1 min.

4.6. Western blot

Fresh NI/MS tissue was dissected and lysed with a radio-immunoprecipitation assay (RIPA) buffer (150 mM sodiumchloride, 1.0% Triton X-100, 0.5% sodium deoxycholate, 0.1%sodium dodecyl sulphate, 50 mM Tris, pH 8.0). Proteins werefirst quantified with a Pierce bicinchoninic acid assay (BCA)kit (Thermo Scientific, USA), separated on a 12% sodiumdodecyl sulphate (SDS)-PAGE and then transferred onto a

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0188

polyvinylidene fluoride (PVDF) membrane. The membranewas blocked with 1% skim milk and incubated with a goatanti-CRF RI/II (1:1000, Santa Cruz, USA), rabbit anti-GAD65(AB5082, 1:1000, Millipore, USA), rabbit anti-TPH2 (1:1000,Chemicon, USA) and rabbit anti-actin (A2066, 1:10 000,Sigma-Aldrich, USA) or 5% Bovine Serum Albumin (BSA)solution and incubated with mouse anti-RLX3 (1:2500) over-night at 4 1C. The blot was then washed and incubated with aHRP-conjugated anti-goat (1:5000), anti-rabbit IgG (1:5000) orgoat anti-mouse Alexa Fluor 488 (1:5000) for 1 h at roomtemperature with agitation. The proteins were then detectedwith a chemiluminescence kit. Protein expression levels werenormalised to actin.

4.7. Fear conditioning

Sham- and NI-lesioned rats were subjected to a fear con-ditioning paradigm (King and Williams, 2009) 14 days aftersurgery. The rats were first acclimatised to the room for 2 hbefore the start of each session. Habituation to the fearconditioning chambers (TSE Systems Inc., USA) was carriedout for 15 mins for 4 days before training. The training phaseconsisted of a 2 min pause, and 6 repeats of 30 s tone with a0.8 mA shock presented in the last 2 s of the 30 s tone. Duringthe test phase 24 h later, the tone was sounded for 30 swithout the shock. The time spent in freezing was recordedfor 5 min following the tone. Freezing was detected by anarray of infrared light beam sensors mounted 14 mm apart tomonitor the position and movement of the animal inside thechamber (TSE Systems Inc., USA). All recordings were donevia the TSE Systems software. This programme uses anaveraging procedure to define the centre of gravity of therat. An instance of freezing was defined as the animal notmoving for more than a threshold duration of 5 s. Percentagefreezing was calculated as the cumulative duration of freez-ing as a percentage of total time.

4.8. Analysis and statistics

RT-PCR gel and western blot images were analysed withImageJ. The density of each band was measured and normal-ised to its corresponding actin band (RT-PCR: n¼4 each forsham and NI-lesioned; western blot: n¼3 for naïve, n¼3 forsaline, n¼3 for true sham, n¼6 for sham-lesioned and n¼7for NI-lesioned). The densitometry data was statisticallyanalysed with unpaired t tests (GraphPad Prism, USA) com-paring sham-lesioned and NI-lesioned groups for each pro-tein. For the real-time PCR (n¼4 each for sham and NI-lesioned), the amplified transcripts were quantified usingthe comparative CT method (Livak and Schmittgen, 2001),with the formula for relative fold change¼2–ΔΔCT. The meanswere analysed with unpaired t tests (GraphPad Prism, USA).Freezing behaviour (threshold set at 3 s) of the sham-lesioned(10) and NI-lesioned (n¼8) rats was recorded in 30 s epochsfor a total of 5 min. Total percentage freezing time wascalculated and analysed with an unpaired t test (GraphPadPrism, USA). The data were expressed as mean7SEM.

Acknowledgements

This work was funded by the Biomedical Research Council(07/1/21/19/512 and 10/1/21/19/645) and the National MedicalResearch Council (IRG10Nov104), Singapore. The authorswish to thank Ms. Lim Zhining for executing pilot studies;Ms. Hong Jia Mei and Dr. Tan Chee Kuan, Francis, for experttechnical advice and support; and Mr. Ho Woon Fei forexcellent technical and administrative assistance.

r e f e r e n c e s

Banerjee, A., Shen, P.J., Ma, S., Bathgate, R.A.D., Gundlach, A.L.,

2010. Swim stress excitation of nucleus incertus and rapid

induction of relaxin-3 expression via CRF1 activation.

Neuropharmacology 58 (1), 145–155.Bittencourt, J.C., Sawchenko, P.E., 2000. Do centrally administered

neuropeptides access cognate receptors? An analysis in the

central corticotropin-releasing factor system. J. Neurosci. 20

(3), 1142–1156.Burazin, T.C., Bathgate, R.A., Macris, M., Layfield, S.,

Gundlach, A.L., Tregear, G.W., 2002. Restricted, but

abundant, expression of the novel rat gene-3 (R3) relaxin in

the dorsal tegmental region of brain. J. Neurochem. 82 (6),

1553–1557.Callander, G.E., Ma, S., Ganella, D.E., Wimmer, V.C.,

Gundlach, A.L., Thomas, W.G., Bathgate, R.A., 2012. Silencing

relaxin-3 in nucleus incertus of adult rodents: a viral

vector-based approach to investigate neuropeptide function.

PLoS One 7 (8), e42300, http://dx.doi.org/10.1371/journal.

pone.0042300.Cervera-Ferri, A., Guerrero-Martinez, J., Bataller-Mompean, M.,

Taberner-Cortes, A., Martinez-Ricos, J., Ruiz-Torner, A., Teruel-

Marti, V., 2011. Theta synchronization between the

hippocampus and the nucleus incertus in urethane-

anaesthetised rats. Exp. Brain Res. 211 (2), 177–192, http://dx.

doi.org/10.1007/s00221-011-2666-3.Chen, J., Kuei, C., Sutton, S.W., Bonaventure, P., Nepomuceno, D.,

Eriste, E., Sillard, R., Lovenberg, T.W., Liu, C., 2005.

Pharmacological characterization of relaxin-3/INSL7 receptors

GPCR135 and GPCR142 from different mammalian species.

J. Pharmacol. Exp. Ther. 312 (1), 83–95.Chen, Y., Brunson, K.L., Muller, M.B., Cariaga, W., Baram, T.Z.,

2000. Immunocytochemical distribution of corticotropin-

releasing hormone receptor type-1 (CRF(1))-like

immunoreactivity in the mouse brain: light microscopt

analysis using an antibody directed against the C-terminus.

J. Comp. Neurol. 420 (3), 305–323.Chronwall, B.M., Skirboll, L.R., O’Donohue, T.L., 1985.

Demonstration of a pontine-hippocampal projection

containing a ranatensin-like peptide. Neurosci. Lett. 53 (1),

109–114.Farooq, U., Rajkumar, R., Sukumaran, S., Wu, Y.,

Tan, W.H., & Dawe, G.S. 2013. Corticotropin-releasing

factor infusion into nucleus incertus suppresses medial

prefrontal cortical activity and hippocampo-medial prefrontal

cortical long-term potentiation. Eur. J. Neurosci. 38 (4), 2013,

2516–2525, http://dx.doi.org/10.1111/ejn.12242.Ford, B., Holmes, C.J., Mainville, L., Jones, B.E., 1995. GABAergic

neurons in the rat pontomesencephalic tegmentum:

codistribution with cholinergic and other tegmental neurons

projecting to the posterior lateral hypothalamus. J. Comp.

Neurol. 363 (2), 177–196.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0 189

Goto, M., Swanson, L.W., Canteras, N.S., 2001. Connexions of thenucleus incertus. J. Comp. Neurol. 438 (1), 86–122.

Heckers, S., Ohtake, T., Wiley, R.G., Lappi, D.A., Geula, C.,Mesulam, M.M., 1994. Complete and selective cholinergicdenervation of rat neocortex and hippocampus but notamygdala by an immunotoxin against the p75 NGF receptor.J. Neurosci. 14 (3 Pt 1), 1271–1289.

Hoover, W.B., Vertes, R.P., 2007. Anatomical analysis of afferentprojections to the medial prefrontal cortex in the rat. BrainStruct. Funct. 212 (2), 149–179, http://dx.doi.org/10.1007/s00429-007-0150-4.

Hummel, M., Cummons, T., Lu, P., Mark, L., Harrison, J.E.,Kennedy, J.D., Whiteside, G.T., 2010. Pain is a salient “stressor”that is mediated by corticotropin-releasing factor-1 receptors.Neuropharmacology 59 (3), 160–166, http://dx.doi.org/10.1016/j.neuropharm.2010.05.001.

King 2nd, S.O., Williams, C.L., 2009. Novelty-induced arousalenhances memory for cued classical fear conditioning:interactions between peripheral adrenergic and brainstemglutamatergic systems. Learn. Mem. 16 (10), 625–634, http://dx.doi.org/10.1101/lm.1513109.

Kizawa, H., Nishi, K., Ishibashi, Y., Harada, M., Asano, T., Ito, Y.,Fujino, M., 2003. Production of recombinant human relaxin 3in AtT20 cells. Regul. Pept. 113 (1–3), 79–84.

Kubota, Y., Inagaki, S., Shiosaka, S., Cho, H.J., Tateishi, K.,Hashimura, E., Tohyama, M., 1983. The distribution ofcholecystokinin octapeptide-like structures in the lower brainstem of the rat: an immunohistochemical analysis.Neuroscience 9 (3), 587–604.

Li, A.J., Dinh, T.T., Ritter, S., 2008. Hyperphagia and obesityproduced by arcuate injection of NPY-saporin do not requireupregulation of lateral hypothalamic orexigenic peptidegenes. Peptides 29 (10), 1732–1739, http://dx.doi.org/10.1016/j.peptides.2008.05.026.

Livak, K.J., Schmittgen, T.D., 2001. Analysis of relative geneexpression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 25 (4), 402–408.

Ma, S., Bonaventure, P., Ferraro, T., Shen, P.J., Burazin, T.C.,Bathgate, R.A., Gundlach, A.L., 2007. Relaxin-3 in GABAprojection neurons of nucleus incertus suggests widespreadinfluence on forebrain circuits via G-protein-coupled receptor-135 in the rat. Neuroscience 144 (1), 165–190.

Ma, S., Olucha-Bordonau, F.E., Hossain, M.A., Lin, F., Kuei, C., Liu,C., Gundlach, A.L., 2009. Modulation of hippocampal thetaoscillations and spatial memory by relaxin-3 neurons of thenucleus incertus. Learn. Mem. 16 (11), 730–742.

Maciejewski-Lenoir, D., Heinrichs, S.C., Liu, X.J., Ling, N., Tucker,A., Xie, Q., Grigoriadis, D.E., 2000. Selective impairment ofcorticotropin-releasing factor1 (CRF1) receptor-mediatedfunction using CRF coupled to saporin. Endocrinology 141 (2),498–504.

McGowan, B.M., Stanley, S.A., Smith, K.L., Minnion, J.S., Donovan,J., Thompson, E.L., Bloom, S.R., 2006. Effects of acute andchronic relaxin-3 on food intake and energy expenditure inrats. Regul. Pept. 136 (1-3), 72–77.

McGowan, B.M., Stanley, S.A., Smith, K.L., White, N.E., Connolly,M.M., Thompson, E.L., Bloom, S.R., 2005. Central relaxin-3administration causes hyperphagia in male Wistar rats.Endocrinology 146 (8), 3295–3300.

Miyamoto, Y., Watanabe, Y., Tanaka, M., 2008. Developmentalexpression and serotonergic regulation of relaxin 3/INSL7 inthe nucleus incertus of rat brain. Regul. Pept. 145 (1-3), 54–59.

Munro, J., Skrobot, O., Sanyoura, M., Kay, V., Susce, M.T., Glaser, P.E., Arranz, M.J., 2012. Relaxin polymorphisms associated withmetabolic disturbance in patients treated with antipsychotics.J. Psychopharmacol. 26 (3), 374–379.

Nunez, A., Cervera-Ferri, A., Olucha-Bordonau, F., Ruiz-Torner,A., Teruel, V., 2006. Nucleus incertus contribution to

hippocampal theta rhythm generation. Eur. J. Neurosci. 23

(10), 2731–2738.Olucha-Bordonau, F.E., Teruel, V., Barcia-Gonzalez, J., Ruiz-

Torner, A., Valverde-Navarro, A.A., Martinez-Soriano, F.,

2003. Cytoarchitecture and efferent projections of

the nucleus incertus of the rat. J. Comp. Neurol. 464 (1),

62–97.Pascual, J., Heinrich, S.C., 2007. Olfactory neophobia and seizure

susceptibility phenotypes in an animal model of epilepsy are

normalised by impairment of brain corticotropin releasing

factor. Epilepsia 48 (4), 827–833.Paxinos, G., Watson, C., 2007. The Rat Brain in Stereotaxic

Coordinates, 6th ed. Elsevier, Singapore.Pereira, C.W., Santos, F.N., Sanchez-Perez, A.M., Otero-Garcia, M.,

Marchioro, M., Ma, S., Olucha-Bordonau, F.E., 2013. Electrolytic

lesion of the nucleus incertus retards extinction of auditory

conditioned fear. Behav. Brain Res. 247, 201–210.Potter, E., Sutton, S., Donaldson, C., Chen, R., Perrin, M., Lewis, K.,

Vale, W., 1994. Distribution of corticotropin-releasing factor

receptor mRNA expression in the rat brain and pituitary. Proc.

Natl. Acad. Sci. USA 91 (19), 8777–8781.Rajkumar, R., See, L.K., Dawe, G.S., 2013. Acute antipsychotic

treatments induce distinct c-Fos expression patterns in

appetite-related neuronal structures of the rat brain. Brain

Res. 1508, 34–43.Ryan, P.J., Ma, S., Olucha-Bordonau, F.E., Gundlach, A.L., 2011.

Nucleus incertus – an emerging modulatory role in

arousal, stress and memory. Neurosci. Biobehav. Rev. 35 (6),

1326–1341.Smith, C.M., Hosken, I.T., Sutton, S.W., Lawrence, A.J.,

Gundlach, A.L., 2012. Relaxin-3 null mutation mice display a

circadian hypoactivity phenotype. Genes Brain Behav. 11 (1),

94–104.Smith, C.M., Lawrence, A.J., Sutton, S.W., Gundlach, A.L., 2009.

Behavioral phenotyping of mixed background (129S5:B6)

relaxin-3 knockout mice. Ann. N.Y. Acad. Sci. 1160, 236–241.Smith, C.M., Ryan, P.J., Hosken, I.T., Ma, S., Gundlach, A.L., 2011.

Relaxin-3 systems in the brain – the first 10 years. J. Chem.

Neuroanat. 42 (4), 262–275.Stirpe, F., Gasperi-Campani, A., Barbieri, L., Falasca, A.,

Abbondanza, A., Stevens, W.A., 1983. Ribosome-inactivating

proteins from the seeds of Saponaria officinalis L. (soapwort),

of Agrostemma githago L. (corn cockle) and of Asparagus

officinalis L. (asparagus), and from the latex of Hura crepitans

L. (sandbox tree). Biochem. J. 16 (3), 617–625.Sutton, S.W., Bonaventure, P., Kuei, C., Roland, B.,

Chen, J., Nepomuceno, D., Liu, C., 2004. Distribution

of G-protein-coupled receptor (GPCR)135 binding sites and

receptor mRNA in the rat brain suggests a role for

relaxin-3 in neuroendocrine and sensory

processing. Neuroendocrinology 80 (5), 298–307.Tanaka, M., Iijima, N., Miyamoto, Y., Fukusumi, S., Itoh, Y., Ozawa,

H., Ibata, Y., 2005. Neurons expressing relaxin 3/INSL 7 in the

nucleus incertus respond to stress. Eur. J. Neurosci. 21 (6),

1659–1670.Thorpe, P.E., Brown, A.N., Bremner Jr., J.A., Foxwell, B.M., Stirpe, F.,

1985. An immunotoxin composed of monoclonal anti-Thy 1.1

antibody and a ribosome-inactivating protein from Saponaria

officinalis: potent antitumor effects in vitro and in vivo. J. Natl.

Cancer Inst. 75 (1), 151–159.van der Westhuizen, E.T., Halls, M.L., Samuel, C.S., Bathgate, R.A.,

Unemori, E.N., Sutton, S.W., Summers, R.J., 2008. Relaxin

family peptide receptors – from orphans to therapeutic

targets. Drug Discov. Today 13 (15-16), 640–651.Van Pett, K., Viau, V., Bittencourt, J.C., 2000. Distribution of

mRNAs encoding CRF receptors in brain and pituitary of rat

and mouse. J. Comp. Neurol. 428 (2), 191–212.

b r a i n r e s e a r c h 1 5 4 3 ( 2 0 1 4 ) 1 7 9 – 1 9 0190

Wada, E., Way, J., Lebacq-Verheyden, A.M., Battey, J.F., 1990.Neuromedin B and gastrin-releasing peptide mRNAs aredifferentially distributed in the rat nervous system. J.Neurosci. 10 (9), 2917–2930.

Waite, J.J., Wardlow, M.L., Chen, A.C., Lappi, D.A., Wiley, R.G., Thal,L.J., 1994. Time course of cholinergic and monoaminergic

changes in rat brain after immunolesioning with 192 IgG-saporin. Neurosci. Lett. 169 (1–2), 154–158.

Watanabe, Y., Tsujimura, A., Takao, K., Nishi, K., Ito, Y.,Yasuhara, Y., Tanaka, M., 2011. Relaxin-3-deficient miceshowed slight alteration in anxiety-related behavior. Front.Behav. Neurosci. 5, 50.