Neuroscience 151 (2008) 242–254
MORPHOLOGICAL CHARACTERIZATION OF RAT Mas-RELATED
G-PROTEIN-COUPLED RECEPTOR C AND FUNCTIONAL ANALYSIS
OF AGONISTS
U. A. HAGER,a1 A. HEIN,a1 J. K. M. LENNERZ,a,b
K. ZIMMERMANN,a W. L. NEUHUBER,b
AND P. W. REEHa*
a
Department of Physiology and Pathophysiology, Friedrich-Alexander-
University Erlangen-Nuremberg, Universitaetsstrasse 17, D-91054
Erlangen, Germany
b against pharmacological interventions and does not involve
Department of Anatomy, Friedrich-Alexander-University Erlangen-
Nuremberg, Krankenhausstrasse 9, D-91054 Erlangen, Germany
Abstract—A recently described family of “orphan” receptors,
called Mas-related G-protein-coupled receptors (Mrg), is pref-
erentially expressed in small nociceptive neurons of the ro-
dent and human dorsal root ganglia (DRG). Mrg are activated
by high affinity peptide fragments derived from the proen-
kephalin A gene, e.g. BAM22 (bovine adrenal medullary). To
study the histological distribution and functional properties
of these receptors, we combined confocal immunohisto-
chemistry in rat DRG and dermis whole mounts, using new
antibodies against the rat Mas-related G-protein-coupled re-
ceptor C (MrgC), with single-fiber recordings and neurochem-
ical experiments using isolated hind-paw skin and sciatic
nerve. In lumbar DRG we found cytoplasmic MrgC labeling
mainly in small- and medium-sized neurons; coexpression
with isolectin B4 (46%) and transient receptor potential va-
nilloid receptor 1 channel protein (TRPV1) (52%) occurred
frequently, whereas calcitonin gene-related peptide (CGRP)
was rarely colocalized with MrgC in DRG (11%) and dermal
nerve fibers (6%). One of the MrgC agonists, BAM22, more
than doubled the heat-induced cutaneous CGRP release from
rat and mouse skin. The effect of BAM22, also known to
activate opioid receptors, was further enhanced by combina-
tion with naloxone that had no effect on its own. This sensi-
tizing effect proved to be independent of secondary prosta-
glandin formation, mast cell degranulation, protein kinase C
(PKC) activation and independent of TRPV1. Nonetheless, the
capsaicin-induced CGRP release was also sensitized. Recep-
tive fields of 26 mechano-heat sensitive C-fibers were treated
with MrgC agonists. Only one unit was strongly and repeat-
1
These authors contributed equally to the work.
*Corresponding author. Tel: ⫹49-9131-8522228; fax: ⫹49-9131-
8522497.
E-mail address: metzner@physiologie1.uni-erlangen.de (P. W. Reeh).
Abbreviations: BAM, bovine adrenal medullary (proenkephalin peptide
fragments; BAM22, BAM8-22); BSA, bovine serum albumin; CGRP,
calcitonin gene-related peptide; CMH, C-fiber mechano- and heat-
sensitive; DRG, dorsal root ganglia; EIA, enzyme immunoassay;
GFAP, glial fibrillary acidic protein; Gö-6976, protein kinase C inhibitor;
GPCR, G-protein-coupled receptor; HEK, human embryonic kidney;
IB4, isolectin B4; ir, immunoreactive/immunoreactivity; Mrg, Mas-re-
lated G-protein-coupled receptor; MSH, melanocyte stimulating hor-
mone; PKC, protein kinase C; PMA, phorbol myristate acetate; SIF,
synthetic interstitial fluid; SNSR, sensory neuron specific receptor; SP,
substance P; TBS, Tris-buffered saline; TRPV1, transient receptor
potential vanilloid receptor 1 channel protein; WT, wild type;
cytoplasmic staining/colocalization; ⬎, satellite cell staining; –, mono-
nuclear cells; *, absence of immunostaining; ¡, axon/process/fiber.
0306-4522/08$32.00⫹0.00 © 2008 Published by Elsevier Ltd on behalf of IBRO.
doi:10.1016/j.neuroscience.2007.09.085
edly excited and showed a profound sensitization to heat
upon BAM22ⴙnaloxone. Two other established MrgC ago-
nists (␥2-melanocyte stimulating hormone and BAM8-22)
were ineffective. Thus, BAM22 sensitizes the capsaicin- and
heat-induced CGRP release in an apparently MrgC-unrelated
way. The sensitization to heat appears unusually resistant
TRPV1. © 2008 Published by Elsevier Ltd on behalf of IBRO.
Key words: Mrg, SNSR, TRPV1, CGRP, BAM22, noxious heat.
Recently, a large family of “orphan” G-protein-coupled re-
ceptors (GPCRs) has been identified and shown to be
selectively expressed in small diameter sensory neurons
of trigeminal and dorsal root ganglia (DRG) (Dong et al.,
2001). These GPCRs are related to the Mas oncogene
(Young et al., 1986) and therefore the encoding genes are
called Mas-related genes (Dong et al., 2001) while the
receptors were termed Mas-related G-protein-coupled re-
ceptors (Mrg). Due to their selective expression in DRG
they have also been named sensory neuron specific re-
ceptors (SNSRs; Lembo et al., 2002), and evolutionary
studies demonstrated that Mrg have experienced strong
positive selection likely directed to nociception (Choi and
Lahn, 2003). Numerous members of the Mrg family have
been identified in the mouse, rat and human genomes and
can be subdivided into four major subfamilies, including
human MrgX and rodent MrgA, MrgB and rat MrgC. Due to
atypical genomic expansion (Zylka et al., 2003) the differ-
ent subfamilies of Mrg contain a variable number of genes,
depending on species. While in mice there are 22 MrgA
and 14 MrgC genes, the rat has only one single represen-
tative of each gene. The rat MrgA gene has been charac-
terized as the adenine receptor (Bender et al., 2002), while
the rat MrgC has been described as rat SNSR (Lembo et
al., 2002; Grazzini et al., 2004).
Further characterization revealed that the rat MrgC is
predominantly, but not exclusively expressed in isolectin
B4 (IB4) positive neurons, and approximately half of them
express the transient receptor potential vanilloid receptor 1
channel protein (TRPV1). In contrast, only a small number
of MrgC positive cells show coexpression with substance P
(SP) or calcitonin gene-related peptide (CGRP) (Lembo et
al., 2002). Rat MrgC is activated by a series of peptides,
most of which belong to the family of endogenous opioid
peptides, suggesting a role of MrgC in nociception. Bovine
adrenal medullary (BAM) peptide (BAM22), a preproen-
,
kephalin product that also contains the N-terminal opioid
sequence, is a potent agonist at human and rat Mrg
242
 U. Hager et al. / Neuroscience 151 (2008) 242–254
Table 1. Primary antibodies
Antigen or lectin Host or plant
MrgC Rabbit, polyclonal
MrgX1 Rabbit, polyclonal
Neurofilament 200 (NF200) Mouse, monoclonal
Biotinylated IB4 Griffonia simplicifolia
SP Guinea pig, polyclonal
TRPV1 Rabbit, polyclonal
GFAP Mouse, monoclonal
Protein gene product 9.5 (PGP 9.5) Guinea pig, polyclonal
S100 protein (S100) Mouse, monoclonal
Calcitonin gene-related peptide (CGRP) Goat, polyclonal
(Lembo et al., 2002; Grazzini et al., 2004). While BAM22
also binds with high affinity to opioid receptors, its C-
terminal fragment, BAM8-22, which is devoid of the opioid
motif, appears to activate the Mrg selectively without losing
potency. Furthermore, ␥2-melanocyte stimulating hor-
mone (MSH), a proopiomelanocortin-derived peptide, was
discovered to be the most potent agonist on rat MrgC in
transfected human embryonic kidney (HEK) 293s cells
(Grazzini et al., 2004). Behavioral studies in rats, using
intradermal and intrathecal injection of MrgC agonists,
showed pain-related behavior and enhanced sensitivity to
noxious heat (Grazzini et al., 2004), further supporting a
role of this receptor in nociception. The function of the
different BAM cleavage products is at this point unknown
and given the sparse data on MrgC expression in DRG
(Lembo et al., 2002; Zylka et al., 2003), we felt that a
combined functional and morphological investigation is
mandatory.
In this study, we report (1) the immunohistochemical
distribution of rat MrgC in DRG and dermis whole mounts
using newly raised antibodies, (2) the influence of MrgC
agonists on stimulated CGRP/histamine release from iso-
lated skin and sciatic nerve preparations, (3) the question
of TRPV1 involvement in the sensitization to heat, and (4)
the effect of MrgC agonists in single-fiber activity and heat
response.
A preliminary account of the present data has been
reported in abstract form (Hager et al., 2005).
EXPERIMENTAL PROCEDURES
Immunohistochemistry
Fixation and tissue processing. Seven male and female
adult Wistar rats (250 –550 g) bred in the Institute of Physiology at
Erlangen were used. Rats were killed by 100% CO2 and DRG
from Th8 to S2 as well as skin flaps from hind paws were removed
on both sides. Epidermis and whole mounts of dermis were pre-
pared from skin flaps by separating these layers along the junc-
tional zone. Epidermis was fixed and paraffin embedded using
standard protocols. Whole mounts of dermis and DRG were im-
mersion fixed in 4% paraformaldehyde in 0.1 M phosphate buffer
(pH 7.4) at 4 °C for 4 – 6 h and 2– 4 h, respectively, and transferred
afterward to 15% sucrose phosphate buffer (pH 7.4). DRG were
mounted on Tissue-Tek (GSV1, Slee Technik, Mainz, Germany),
rapidly frozen in methyl butane at ⫺70 °C and stored at ⫺20 °C.
In a cryostat 10 –12 ␮m thick sections were cut, mounted on
243
Dilution Source
1:100 Eurogentec, Seraing, Belgium
1:100 Novus Biologicals, Littleton, CO, USA
1:200 Sigma, Deisenhofen, Germany
1:200 Vector Laboratories, Burlingame, CA, USA
1:200 Novus Biologicals, Littleton, CO, USA
1:500 Biotrend, Cologne, Germany
1:600 Sigma, Deisenhofen, Germany
1:500 Chemicon, Temecula, CA, USA
1:500 Biotrend, Cologne, Germany
1:200 Biotrend, Cologne, Germany
poly-L-lysine-coated slides and dried for 1 h at room temperature.
Whole mounts of dermis were processed free floating.
Immunofluorescence staining. Slides with mounted sec-
tions were incubated with 5% normal donkey serum (Dako, Ham-
burg, Germany) containing 0.5% Triton X-100 (Merck, Darmstadt,
Germany) and 1% bovine serum albumin (BSA, Roth, Karlsruhe,
Germany) in Tris-buffered saline (TBS, Roth; pH 7.4) for 1 h at
room temperature. Sections were incubated overnight for immu-
nolabeling with primary antibodies (for details and working dilu-
tions see Table 1). We used a new rabbit-anti-rat polyclonal
MrgC-antibody against an N-terminal 16-amino-acid peptide (Eu-
rogentec, Seraing, Belgium); the antibody was purified and vali-
dated (Western blot not shown). All primary antibodies were di-
luted in TBS, containing 1% BSA and 0.5% Triton X-100 at room
temperature. After washing in TBS, sections were incubated for
1 h at room temperature with donkey anti-guinea-pig secondary
antibodies (Cy3-conjugated, 1:400; Dianova, Hamburg, Ger-
many), or donkey anti-mouse secondary antibody (Cy5-conju-
gated, 1:100; Dianova; Alexa Fluor 555-conjugated, 1:1000; Mo-
lecular Probes), or donkey anti-rabbit secondary antibodies (Alexa
Fluor 555/488-conjugated, 1:1000), or donkey anti-goat second-
ary antibodies (Alexa Fluor 555/647-conjugated, 1:1000; Molecu-
lar Probes), or streptavidin (Cy3-conjugated, 1:1000; Dianova). All
secondary antibodies contained 0.5% Triton X-100 and 1% BSA in
TBS. Between each incubation sections were rinsed thoroughly in
TBS. Sections were finally rinsed and coverslipped with TBS–
glycerol (1:1; pH 8.6) for microscopy.
Whole mounts of dermis were preincubated with 1% BSA,
0.5% Triton X-100 and 5% serum (corresponding to the secondary
antibodies previously listed) for 1 h at room temperature. After
rinsing in TBS (pH 7.4), specimens were incubated with primary
antibody in TBS containing 1% BSA for 3 days at 4 °C free
floating. After three washes with TBS for 12 h binding sites were
visualized by incubation with secondary antibodies in TBS for 3 h.
After several rinses in TBS, tissues were mounted and cover-
slipped in TBS– glycerol (1:1; pH 8.6).
Controls. Specificity of the immunohistochemical reactions
(sections and whole mounts) was tested by omission of primary
and secondary antibodies and replacing them with TBS or normal
serum, respectively. In the case of the MrgC-antibody, the spec-
ificity was assessed using three different controls: we performed
1) preabsorption controls (Fig. 1B) with the MrgC primary antibody
(1:100) and the MrgC-peptide (1:10) against which the antibody
was raised; 2) negative controls (Fig. 1H) where the secondary
antibody was used without the primary antibody in the first incu-
bation; and 3) normal serum controls (not shown) where the first
antibody was replaced by normal (rabbit) serum using 1:10 and
1:50 concentrations.
In addition, the staining pattern of the newly raised MrgC
antibody was compared with the staining seen with an antibody
244 U. Hager et al. / Neuroscience 151 (2008) 242–254
(Fig. 1C, I) against the N-terminal extracellular terminus of the
human MrgX1 (MrgX1) (Novus Biologicals, Littleton, CO, USA).
Due to the identical staining profile of the MrgX1 and MrgC anti-
bodies we concluded a similar specificity (Fig. 1A, C).
Microscopy. After initial analysis of all stainings using an
epifluorescence microscope (Leica, Bensheim, Germany) con-
nected to a digital camera (Visitron Systems GmbH, Puchheim,
Germany), confocal images were obtained using BioRad
MRC1000 confocal-laser scanning system (Bio-Rad, Hemel
Hempstead, UK) equipped with a three-line krypton–argon laser
(American Laser Corporation, Salt Lake City, UT, USA) and a
Nikon Diaphot 300 inverted microscope (Nikon, Düsseldorf, Ger-
many). Images were adjusted for contrast and brightness using
Adobe Photoshop 7.0 software (Adobe, San Jose, CA, USA).
Quantification. To quantify the proportion of co-labeling in
the various immunolabeled neuron populations, DRG (n⫽8) from
segments L3, L4, and L5 of three rats were used. All DRG were
completely cryotomized and labeled simultaneously for the follow-
ing marker combinations: MrgC and CGRP, MrgC and IB4, MrgC
and SP. Separate staining was performed on consecutive sec-
tions for TRPV1 and MrgC (species identity) and colocalization
was achieved after matching of corresponding regions in subse-
quent z-levels (Fig. 2). To avoid counting the same cell twice
within each group of marker combinations, consecutive sections
of each DRG were split over five slides, each representing the
three co-labeling combinations and two separated single label-
ings, respectively. Neurons with visible nuclei or strong cytoplas-
mic staining were counted on printouts of individual channels and
a merged image.
Protein search. The protein sequence of the GPCRs MrgC
(Rattus norvegicus), accession number AAQ08317 (Zylka et al.,
2003), and MrgX1 (Homo sapiens), accession number AAK91804
(Dong et al., 2001), was obtained. PSI-BLAST protein searches
for the 31 N-terminal extracellular amino acids were performed
using BLASTP 2.2.15 [Oct. 15, 2006] and the non-redundant
SwissProt database [Feb. 27, 2007] of 237,037 protein sequences
(Altschul et al., 1997; Schaffer et al., 2001).
CGRP and histamine release from isolated skin
and nerve
Animals. Homebred male and female Wistar rats weighing
between 70 and 90 g and littermates of wild type (WT) and
TRPV1⫺/⫺ (KO) mice of both sexes weighing between 25 and
29 g were killed by 100% CO2. TRPV1 transgenic mice had been
continuously backcrossed to C57BL/6/J mice and were congenic
since the year 2001. Every single animal was conventionally
genotyped using customized primers (Metabion, Martinsried, Ger-
many). Both, the WT and KO alleles were identified with two
genomic 5= primers. The sequences were CAT GGC CAG TGA
GAA CAC CAT GG and AGC CTT TTG TTC TTG GCT TCT CCT
for the WT allele, and CCG GTG CCC TGA ATG AAC T and AAG
ACC GGC TTC CAT CCG A for the KO allele.
Release experiments. Both rat/mouse hind paws were
skinned and intact skin flaps were washed for 30 min at 32 °C,
then passed through a series of five 5-min-incubation steps to
assess basal (at 32 °C) and heat-stimulated (47 °C in rats, 46 °C
in mice) CGRP release using enzyme immunoassay (EIA) as
previously described (Zimmermann et al., 2005). If used, the
following agents were present in the second and third incubation
steps: BAM22 (Biotrend, Cologne, Germany; Bachem, Weil am
Rhein, Germany), BAM8-22 (Biotrend), ␥2-MSH (Bachem), nal-
oxone (Sigma, Deisenhofen, Germany) all 10 ␮M, phorbol myris-
tate acetate (PMA) 10 ␮M (Sigma), the protein kinase C inhibitors
Gö-6976 0.5 ␮M (Sigma) and calphostin C 0.5 ␮M (Sigma), the
TRPV1 antagonist capsazepine 3 ␮M (Sigma). The heat stimula-
tion at 46 °C/47 °C was always applied during the third incubation.
For detailed protocols see Fig. 5–7. Both rat sciatic nerves were
excised and desheathed as previously described (Fischer et al.,
2003) and then processed as described above. In separate
CGRP-release experiments we used a KCl concentration of
60 mM as unspecific depolarizing stimulus during the third incu-
bation step or a capsaicin (Sigma) concentration of 0.31 ␮M
(instead of heat stimulation); the remaining features of the protocol
were not changed. In general, preparations from one leg were
used for control stimulation, from the other leg for stimulation and
the influence of compounds. To elucidate the role of prostaglan-
dins a series of experiments was carried out in synthetic interstitial
fluid (SIF) containing S(⫹)-flurbiprofen 1 ␮M (personal gift from
Prof. G. Geisslinger, Pharmazentrum, University Frankfurt/Main,
Germany) to block prostaglandin synthesis, and eight instead of
five samples were collected. In this case, the heat stimulation
(47 °C) was performed later in the sixth incubation step, to ensure
a complete prostaglandin deprivation. To measure the peptide-
stimulated cutaneous histamine release skin flaps were passed
through two 5-min-incubation steps containing SIF in the first step
and one of the following agents in the second step: BAM22,
BAM8-22, ␥2-MSH, all 10 ␮M, and the mast cell degranulator
compound 48/80 2.5 ␮g/ml (Sigma). The temperature of the so-
lutions was maintained at 32 °C. CGRP release values refer to pg
CGRP per ml incubation fluid, whereas histamine release values
refer to ng histamine per ml incubation fluid. All data are given as
mean⫾S.E.M.
Single-fiber electrophysiology
Outbred male and female Wistar rats weighing 220 –350 g were
killed by exposure to 100% CO2 (n⫽18). The isolated skin–sa-
phenous nerve preparation and single-fiber recording technique
were used (Reeh, 1986). The skin was kept under laminar super-
fusion of carbogen gassed SIF. Receptive fields of identified C-
fibers were searched by mechanical probing and further charac-
terized using feedback-controlled radiant heat (32 °C to 48 °C in
20 s), SIF at 4 °C and von Frey filaments. The criterion for the heat
threshold was the temperature at which the second spike of the
heat response was discharged during ramp-shape heating of the
receptive field. For superfusion with BAM22 (Biotrend, Bachem),
␥2-MSH (Bachem) and naloxone (Sigma), the receptive field was
isolated from the surrounding fluid with a 2 mm thick walled metal
ring (volume 400 ␮l). For protocols see Fig. 8A.
In a separate series of experiments we recorded from addi-
tional C-fibers, applying prolonged 5 min heat stimuli through a
homemade flow-through heater with a flow rate of 6 ml/min. The
stimulus consisted of heating the receptive field within 10 s up to
47 °C and maintaining the temperature at 47 °C over the 5 min
period.
Data analysis
Data evaluation and statistical procedures are described together
with the results where appropriate. Figures show means⫾S.E.M.;
P-values ⬍0.05 were considered significant and symbolized by an
asterisk.
RESULTS
MrgC-immunoreactive (ir) neurons and satellite cells
within the rat DRG
Anatomical distribution of MrgC was studied using a newly
raised antibody against the N-terminus of the rat MrgC.
Immunofluorescence staining of sections through the rat
lumbar DRG appeared predominantly in small- and me-
dium-sized neurons (Fig. 1A), as described previously
 U. Hager et al. / Neuroscience 151 (2008) 242–254 245

Fig. 1. (A) MrgC-ir (arrowheads) and immunonegative (asterisks) neurons were identified. The staining was distributed mainly cytoplasmic in small-
and medium-sized neurons. In addition, a dense rim of MrgC-ir around some of the neurons (open arrowheads) was identified (Fig. 2D–F). (B, E)
Preabsorption control (PA); no specific staining is seen in the preabsorption with the N-terminal peptide (antigen) against which the MrgC antibody was
raised. (C) MrgX1-ir was found within the cytoplasm of medium-sized rat DRG neurons (arrowhead). A dense circumferential rim of MrgX1-ir was also
seen around some of the neurons (open arrowhead). (D, E) MrgC-ir in peripheral processes of DRG neurons (D, arrows), extinguished in the
preabsorption control (E). (F, I) All in-focus projection of four single optical sections (1 ␮m) in the corium tissue of rat dermis whole mounts,
immunolabeled for MrgC (F) and MrgX1 (I), showing ir fibers of different caliber (arrows in F and I). Note the small (⬃6 –10 ␮m) mononuclear cells
(open arrows in F and I) in close proximity to the nerve fibers. (G) Zoom-in on MrgC-ir neuron illustrating the granular quality of the cytoplasmic staining
with dense perinuclear -ir. (H) Negative control (negC); no specific staining using the secondary antibody without the primary antibody. (J) Alignment
of the N-terminal residues of the human MrgX1 and rat MrgC. The 31 N-terminal extracellular amino acids of the human MrgX1 were searched against
the SwissProt database for the species Rattus norvegicus. One protein, rat MrgC, was identified and the alignment (bold) shows 56.5% identity (middle
line) or, including similar residues (⫹), 65% similarity. Scale bars⫽50 ␮m A–F, H–I, G 20 ␮m.

(Lembo et al., 2002). The staining was mainly cytoplasmic
with a slightly granular pattern in the perinuclear region.
Individual neurons also showed nuclear staining. Neurons
that showed no immunoreactivity in the cytoplasm were
assigned as MrgC-negative. In addition to the neuronal
staining we observed strong immunoreactivity in satellite
cells surrounding MrgC-ir as well as immunonegative neu-
rons. The satellite cells could be identified by their size,
position and in most cases by their complete circumferen-
tial staining (Fig. 1). The specificity of the MrgC staining
was assessed by negative, rabbit normal serum (not
shown) and preabsorption controls (Fig. 1B, E), all show-
ing no immunoreactivity.
To confirm the intraneuronal staining, double immuno-
labeling for MrgC and the pan-neuronal marker PGP9.5
was used (Fig. 2A–C). This led to a high degree of colo-
calization among the DRG neurons. In contrast, the satel-
lite cell staining showed no colocalization with PGP9.5 as
illustrated by a sharp MrgC-ir rim around the MrgC and
PGP9.5-ir neurons (Fig. 2C). To ensure MrgC-ir in satellite
cells, the colocalization of MrgC with PGP9.5 was comple-
mented by the double immunostaining for the glial marker
glial fibrillary acidic protein (GFAP). The staining for GFAP
showed only satellite cells surrounding unstained neurons.
The double labeling clearly revealed MrgC-ir satellite cells,
often completely surrounding MrgC-ir and immunonega-
tive neurons (Fig. 2D–F).
Immunostaining and N-terminal sequence homology
of MrgC and MrgX1
To compare the pattern of immunoreactivity seen with the
newly raised rabbit–anti-rat MrgC antibody, we used a
rabbit–anti-human MrgX1 antibody in rat. The latter, also
raised against a synthetic peptide, corresponding to the
penultimate N-terminal extracellular component of the hu-
man MrgX1 (Table 1), has not been evaluated in rat. Both
antibodies showed reactivity in similar substructures, in
brief, predominantly small neurons, satellite cells and ax-
ons (Fig. 1A, C).
To further characterize and quantify the amino-acid
sequence homology of the antigen, and to identify possible
other homologous cross-reacting antigens in rat, we per-
formed two PSI BLAST protein searches. The initial search
consisted of the N-terminal 31 residues from the human
MrgX1 against the non-redundant SwissProt database
species restricted to Rattus norvegicus. We identified one
entry with 56% sequence identity, the rat sensory neuron-
246 U. Hager et al. / Neuroscience 151 (2008) 242–254
Fig. 2. (A–R) Rat lumbar DRG, double labeled for MrgC and PGP9.5,
GFAP, IB4, CGRP, SP and TRPV1, respectively. (A–C) The co-
labeling with a pan-neuronal marker (PGP9.5, A) shows strong intra-
neuronal cytoplasmic MrgC staining (B), resulting in the mixed color
yellow in the merge (C). (D–F) Immunolabeling with a glial marker
(GFAP, D) shows a rim surrounding neurons, some of which are
non-reactive for MrgC (asterisks). The pericytoplasmic rim is also ir for
MrgC (E), resulting in the mixed color yellow in the merge (F). This
staining pattern was interpreted as satellite cell staining (open arrow-
head; see also Fig. 1A, C). (G–O) Labeling with IB4 (G), CGRP (J) and
SP (M) showed colocalization with MrgC (H, K, N) resulting in the
mixed color yellow in the merges (I, L, O), respectively. (P–R) Single
optical sections of alternating slices of rat DRG immunolabeled for
TRPV1 (P) and MrgC (Q). “Last” z-level of the TRPV1 immunostain
was merged with the “first” z-level section of the MrgC immunostain. Of
the 112 MrgC-ir neurons (P), 58 showed also cytoplasmic TRPV1-ir,
resulting in the mixed color yellow in the merge (R). (S) Modified Venn
diagrams illustrate the colocalization of MrgC with TRPV1, IB4, CGRP
Fig. 3. (A–C) PGP9.5-ir within the rat lumbar DRG (A), double labeled
for MrgC (B), showed many co-reactive axons resulting in the mixed
color yellow in the merge. Scale bar⫽50 ␮m. (D–F) In single optical
sections of peripheral incoming processes (small arrow) of oriented
(long broken arrow, pointing toward spinal nerve) lumbar DRG (inter-
rupted line), we found strong immunoreactivity for CGRP (D), colocal-
ized with MrgC-ir (E), resulting in the mixed color yellow in the merge
(F). The ventral root (VR), in contrast, was weakly stained or non-
reactive for both markers. Scale bar⫽250 ␮m. For interpretation of the
references to color in this figure legend, the reader is referred to the
Web version of this article.
specific GPCR 1, equivalent to MrgC (Fig. 1J). In reverse,
the N-terminal 31 residues from the rat MrgC searched
against the non-species restricted SwissProt database
identified two proteins, the 100% aligned rat MrgC and the
human MrgX1 (not shown).
The absence of other homologous rat proteins to the
human extracellular N-terminal residues was taken as an
argument that the antibody raised against human MrgX1,
binds to the N-terminal extracellular domain of MrgC in rat
and thereby provides a staining compatible with that of the
anti-rat MrgC. Based on the absence of staining in the
three controls for the anti-rat antibody, and the similarity of
the staining with the anti-human MrgX1, including appro-
priate controls, the immunostaining for MrgC was consid-
ered specific.
Colocalization of MrgC with IB4, CGRP, SP
and TRPV1
Previous in situ hybridization studies revealed that MrgC is
associated preferentially with the IB4 class of nociceptors
whereas only a small proportion of MrgC positive neurons
contain CGRP or SP. Moreover, approximately half of
MrgC-mRNA expressing cells have been reported to colo-
calize with the heat and capsaicin receptor TRPV1 (Lembo
et al., 2002). Therefore, sections of the DRG were double
labeled for MrgC, IB4 and CGRP (Fig. 2G–L). This resulted
in marked colocalization of MrgC with IB4 as well as
CGRP. The different degree of MrgC colocalization with
IB4 and CGRP was quantified and complemented by dou-
ble labeling for SP and reconstructive colocalization with
and SP, respectively, in DRG neurons. A total (⌺) of 4920 neurons
were counted and the sizes of the different areas depicted (colors
corresponding to G–O) are proportional to number of counted cells (n)
in each group. Note that over 50% of the MrgC-ir reactive neurons
showed co-immunoreactivity with TRPV1. Scale bars⫽50 ␮m. For
interpretation of the references to color in this figure legend, the reader
is referred to the Web version of this article.
 U. Hager et al. / Neuroscience 151 (2008) 242–254 247

bar DRG together with the incoming spinal nerves and

double-labeled the sections for MrgC and PGP9.5. The
peripheral processes showed marked colocalization (Fig.
3A–C). We also noted CGRP and IB4 colocalization with
MrgC in the nerve fibers (not shown). By complete sec-
tioning of one of the DRG with the adjacent spinal nerve
stump we were able to reconstruct the fusion of the ventral
root with the spinal nerve (Fig. 3D–F). Double staining for
MrgC and CGRP showed almost no MrgC-ir in the ventral
root but colocalization of MrgC and CGRP in the incoming
peripheral processes of DRG neurons. Likewise, MrgC-ir
was found in the spinal nerve proper.

MrgC immunoreactivity in cutaneous nerve fibers

and subepidermal nerve endings
Fig. 4. (A–F) Single optical sections of MrgC immunostained rat der- The strong immunoreactivity in the peripheral processes of
mis whole mounts with PGP9.5 and CGRP. (A–C) The PGP9.5-ir fiber DRG neurons and in the spinal nerve pointed toward the
bundle (A) shows costaining for MrgC (B), resulting in the mixed color feasibility of staining for MrgC in the periphery. Therefore
yellow in the merge (C). (D–F) We also identified several thin CGRP-ir
fibers (D) in the dermis whole mounts; colocalization with MrgC (E)
we compared the MrgC immunostaining in rat dermis
resulted in the mixed color yellow (F). (G–I) Single optical sections of whole mounts with the MrgX1 antibody and again found
(paraffin embedded) cross-sectioned skin of the rat hind paw. A thin identical patterns (Fig. 1F, I). Ir fibers within thicker nerve
subepidermal PGP9.5-ir (G) ending (inset; box⫽14 ␮m) approaches
the dermo-epidermal junction (arrows). The ending also shows
MrgC-ir (H) resulting in the mixed color yellow in the merge (I). No
intraepithelial process was identified on additional sections (not
shown). Scale bars⫽50 ␮m. For interpretation of the references to
color in this figure legend, the reader is referred to the Web version of
this article.

TRPV1 (Fig. 2M–R). The lumbar segmental DRG L3, L4,

L5 were separately counted for neurons coexpressing
MrgC with either CGRP, SP or IB4 binding, but no major
differences between segments occurred and the data were
therefore pooled.
We counted 4920 ir neurons. A majority of the MrgC-ir
neurons was able to bind IB4 (46%), whereas only a small
proportion showed co-staining for CGRP (11%) or SP
(10%) (Fig. 2S). Triple-labeling identified some neurons
with both, IB4 and CGRP reactivity (not shown), consistent
with the fact, that in the rat a small subset of IB4 positive
cells is known to express CGRP (Averill et al., 1995).
The examination of TRPV1 colocalization could not be
performed in a direct way because of the species identity
of the primary antibodies. Therefore we reconstructed pic-
tures taken from consecutive sections through the rat lum-
bar DRG that had been separately stained. The pictures
were taken in regions of the DRG where anatomical land-
marks were prominent enough to ensure later matching
(Fig. 2P–R). It is noteworthy that the satellite cell staining
could be used for better merging. Using this consecutive
section method, we counted 112 neurons ir for MrgC, and
58 (52%) of the 112 showed immunoreactivity for TRPV1. Fig. 5. Noxious heat-induced CGRP release from isolated hairy skin
Thus, about half of the MrgC-ir cells showed expression of and desheathed sciatic nerve of the rat. (A, B) The isolated skin/nerve
the TRPV1. specimens were conditioned twice for 5 min (gray box) with different
chemicals (all 10 ␮M), and noxious heat stimulation was applied
during the second 5-min period (bottom trace). All heat-induced in-
MrgC within incoming spinal nerve processes
creases of CGRP release over baseline were significant (Wilcoxon
As reported above we found MrgC within small- and me- test). (A) Rat skin: BAM22 significantly facilitates the heat response,
significantly more so in presence of naloxone (nal; U test). BAM8-22
dium-sized DRG neurons. In addition, the processes and ␥2-MSH induced no significant change compared with the control
emerging from the neurons showed MrgC-ir as well (Fig. heat stimulus. (B) Rat desheathed sciatic nerve: BAM22⫹nal signifi-
1D). To further support this impression we dissected lum- cantly facilitates the axonal heat response (Wilcoxon test).
248 U. Hager et al. / Neuroscience 151 (2008) 242–254
Fig. 6. Noxious heat-induced CGRP and histamine release from iso-
lated hairy skin of the rat. (A) No effect of S(⫹)-flurbiprofen 1 ␮M on the
heat-induced CGRP release from isolated rat skin conditioned with
BAM22⫹naloxone (gray box). S(⫹)-Flurbiprofen was present in all
eight incubation steps but did not alter the heat response facilitated by
BAM22⫹naloxone. The heat-induced increases of CGRP release over
baseline were significant (Wilcoxon test). (B) Basal and stimulated
histamine release from isolated rat skin; two incubation steps of 5 min
duration. None of the peptides (all 10 ␮M) present in the second step
(black column) was able to increase histamine release significantly, in
contrast to compound 48/80, 2.5 ␮g/ml (Wilcoxon test).
bundles as well as very thin-labeled fibers could be visu-
alized. Double labeling for MrgC and the neuronal markers
PGP9.5 (Fig. 4A–C) and NF200 (not shown) confirmed the
colocalization of MrgC within nervous elements in dermis
whole mounts. In addition we found mononuclear cells that
were ir for MrgC and MrgX1. The nuclei of these cells were
never ir and the cytoplasm revealed a granular distribution
of the MrgC/MrgX1-ir. This was interpreted as a possible
mast cell staining. Some vessels could be identified by
their faint MrgC/MrgX1-ir in endothelial and smooth muscle
layers (not shown), which has previously been reported
from intestinal blood vessels, staining for MrgX2 (Robas
et al., 2003).
Double-labeling with IB4 and CGRP showed that the
majority of MrgC-ir nerve fibers also showed IB4 co-posi-
tivity (not shown), whereas only 6% (6 of 98) of the fibers
colocalized with CGRP (Fig. 4D–F). This quantification in
the dermis showed even less colocalization of MrgC and
CGRP than in the DRG, which was attributed to the low
number of nerve fibers counted in the dermis whole
mounts. In one case we found a fiber bundle that exhibited
IB4, CGRP and MrgC-ir (not shown), resembling the colo-
calization of IB4 and CGRP within the DRG.
Double-labeled paraffin sections of rat hind-paw epi-
dermis for MrgC with PGP9.5 (Fig. 4G–I), NF200 (not
shown) and CGRP (not shown) revealed subepidermal
nerve endings that showed MrgC-ir. We also observed
MrgC positive nerve fibers that traversed over several
100 –200 ␮m underneath the epidermal junction, while
others seemed to terminate subepidermally. We found no
intraepidermal nerve fibers that exhibited MrgC-ir.
BAM22 increases heat-induced CGRP release from
isolated rat hind-paw skin and desheathed sciatic
nerve
In the skin nerves CGRP, the main representative of the
“sensory” neuropeptides that further comprise SP and neu-
rokinin A, is exclusively expressed in nociceptive sensory
neurons. The noxious heat-induced CGRP release can
thus be considered as an index for the overall activation of
a large nociceptor subpopulation. Although the immuno-
histochemistry revealed only a small proportion of colocal-
ization for MrgC and CGRP, we tested if MrgC agonists
were able to alter the heat-induced CGRP release from the
isolated rat hind-paw skin and, in separate experiments,
from the desheathed sciatic nerve.
The mean basal release of CGRP from the isolated
skin was 12.7⫾0.7 pg/ml (n⫽70; Fig. 5A). Noxious heat
stimulation of 47 °C, applied for 5 min, increased the
CGRP level 8.9-fold (⫾0.95, n⫽19) to 124⫾12 pg/ml.
BAM22, its C-terminal fragment BAM8-22, devoid of the
opioid motif, and ␥2-MSH have all been shown to activate
MrgC in transfected HEK293s cells (Grazzini et al., 2004).
Exposure of the isolated skin to BAM22 10 ␮M one
incubation step prior to heat stimulation had no effect on
basal release, but significantly augmented the subse-
quent heat-induced CGRP release to 201⫾15.6 pg/ml
(n⫽16, P⬍0.01, Wilcoxon; Fig. 5A). To prevent potential
opioidergic effects of BAM22, which contains an opioid
moiety (amino acids 1–7), we added naloxone 10 ␮M, the
opioid receptor antagonist, that did not alter the CGRP
release on its own (n⫽4, data not shown). The combina-
tion of BAM22 with naloxone further increased the CGRP
release to 281⫾22.7 pg/ml (n⫽12), which was significantly
more in comparison to plain BAM22 (P⬍0.01, U test) and
represented a 2.3-fold augmentation of the control heat
response.
The two other MrgC agonists, BAM8-22 10 ␮M (n⫽12)
and ␥2-MSH 10 ␮M (n⫽7), described to be more potent
than BAM22 (Grazzini et al., 2004) were unable to alter the
heat-induced CGRP release from the rat skin (Fig. 5A).
Although only 11% of identified MrgC-ir neurons in rat
DRG colocalized with CGRP, BAM22 (10 ␮M, in combina-
tion with naloxone 10 ␮M) also caused a 3.8-fold enhance-
ment of the heat-induced CGRP release from the isolated
desheathed rat sciatic nerve (n⫽7, P⬍0.01, Wilcoxon; Fig.
5B). At this point, it remained to be elucidated if the facil-
itation of heat-induced CGRP release is receptor mediated
 U. Hager et al. / Neuroscience 151 (2008) 242–254 249

Fig. 7. Noxious heat- and capsaicin-induced CGRP release from isolated hairy skin of rat and mouse. (A–D) The skin was conditioned for 10 min with
different chemicals (gray box). All increases of CGRP release over baseline were significant (Wilcoxon test). The bottom trace illustrates the heat
stimulus [47 °C in case of rat skin (A, C), 46 °C in case of mouse skin (D)]. (A, C) No effects of Gö6976 (0.5 ␮M, A) and calphostin C (0.5 ␮M, C)
on the heat-induced CGRP release facilitated by BAM22⫹naloxone (nal; both 10 ␮M). None of the compounds was able to alter the heat response
in comparison to the skin treated with plain BAM22⫹nal. (B) BAM22⫹nal (both 10 ␮M) significantly increases the CGRP release induced by capsaicin
0.31 ␮M (Wilcoxon test). The black box indicates the capsaicin stimulus. (D) Effect of BAM22⫹nal (both 10 ␮M) on heat-induced CGRP release in
TRPV1 ⫹/⫹ (WT) and TRPV1 ⫺/⫺ (KO) mice. BAM22⫹nal significantly increased the heat response in WT as well as KO mice (Wilcoxon test).

or depends on secondary effects (e.g. mast cells) of
BAM22.
KCl-induced CGRP release is not altered
by BAM22ⴙnaloxone
GPCR-mediated signal transduction could affect the long
cascade reaction leading to heat-induced CGRP release at
various sites: for example, sensitize the process of sensory
transduction, e.g. through heat-activated ion channels
such as TRPV1, or facilitate the neurosecretory chain of
events leading to vesicular exocytosis of CGRP. To ac-
count for the Ca2⫹-dependent basal neurosecretory mech-
anism, CGRP release was induced by unspecific depolar-
ization using KCl (60 mM). However, conditioning with
BAM22 10 ␮M⫹naloxone 10 ␮M did not significantly alter
the KCl-induced CGRP release (97⫾18 pg/ml, n⫽6) in
comparison to the control side (108⫾30 pg/ml, Wilcoxon,
data not shown). On the other hand, phosphorylation of the
neurosecretory machinery by protein kinase C (PKC)
should enhance the vesicular exocytosis (Lou et al., 2005).
Thus, as a positive control, PMA (10 ␮M) was used to
activate PKC which indeed significantly facilitated the KCl-
induced CGRP release (146⫾29 pg/ml, n⫽6) in compari-
son to the control side (119⫾22 pg/ml, n⫽6, P⬍0.05 Wil-
coxon, data not shown).
BAM22 effects do not result from mast cell release
of histamine or secondary prostaglandin formation
As reported above, the immunohistochemistry in dermis
whole mounts revealed MrgC-ir mononuclear cells that
resembled mast cells and/or other cellular elements of the
connective tissue, e.g. fibroblasts. Both cell-types are
known to release prostaglandins which in turn are known
to sensitize polymodal nociceptors to heat (Petho et al.,
2001). Therefore, we treated the isolated skin with S(⫹)-
flurbiprofen 1 ␮M, a nonselective COX1/2 inhibitor, to de-
prive it from prostaglandins. S(⫹)-Flurbiprofen has re-
cently been found to have no effect on the heat-induced
CGRP release by itself (Derow et al., 2007). As the action
of this compound is time-dependent (Muramatsu et al.,
1984), three more incubation steps before heat stimulation
were required. The heat-induced CGRP release on the
250 U. Hager et al. / Neuroscience 151 (2008) 242–254
Fig. 8. Single-fiber recordings from saphenous nerve C-fibers with
receptive fields in isolated rat hairy skin. (A) Original recording show-
ing instantaneous discharge rates in response to repeated noxious
radiant heat stimuli (32– 48 °C, 20 s, gray columns) and to application
of BAM22⫹naloxone (both 10 ␮M). The CMH unit (0.47 m/s, 90.5 mN)
responded twice poorly to heat, was vividly excited by BAM22⫹
naloxone and sensitized to heat (lower threshold, increased peak and
overall discharge). After washout the BAM22 effect was partially
control side (110⫾14 pg/ml, n⫽4) was virtually identical
with that of the flurbiprofen-treated side (115⫾13 pg/ml,
n⫽4; Fig. 6A) when treated with BAM22⫹naloxone, thus,
no secondary effects of prostaglandins are involved in the
heat-sensitizing effect of BAM22. This series of experi-
ments was performed at a later time point, and a different
charge of EIA kits was used. This resulted in generally
lower CGRP baselines and peak values. Nonetheless, due
to the low variance of the data we abstain from normalizing
the results.
To account for a possible mast cell degranulation by
MrgC agonists, basal and peptide-induced histamine re-
lease was measured from isolated skin. None of the pep-
tides (BAM22, n⫽8; BAM8-22, n⫽7; ␥2-MSH, n⫽7; all
10 ␮M) had significant influence on the histamine release.
In contrast, compound 48/80 2.5 ␮g/ml, which served as
positive control, strongly enhanced histamine release by a
factor of 4.4 (n⫽6; P⬍0.05, Wilcoxon; Fig. 6B).
These data suggest that the sensitizing BAM22 effect
is directly related to the nociceptive nerve endings and
does not depend on secondary mechanisms related e.g. to
cells in their close vicinity.
PKC inhibition does not prevent sensitization to heat
by BAM22ⴙnaloxone
The MrgC appears to exert effects through coupling to
Gq/11 (Han et al., 2002; Grazzini et al., 2004) whose signal
transduction finally activates PKC. We pursued the ques-
tion, whether inhibition of PKC was able to prevent the
BAM22-mediated effects on heat-induced CGRP release.
Two different PKC blockers were tested: Gö-6976 0.5 ␮M,
which particularly inhibits the ␣ and ␤I isoforms of PKC,
and calphostin C 0.5 ␮M, a general PKC blocker inhibiting
all isoforms (Kobayashi et al., 1989). Coadministration of
these blockers together with BAM22 and naloxone led to a
heat-induced CGRP release of 132⫾11 pg/dl (n⫽11) for
Gö-6976 and 171⫾18 pg/dl (n⫽4) for calphostin C,
whereas the CGRP levels with plain BAM22⫹naloxone
were 152⫾13 pg/dl (n⫽11) and 148⫾18 pg/dl (n⫽4) in the
Gö-6976 and calphostin C control experiments, respec-
tively (Fig. 7A, C). Thus, neither Gö-6976 nor calphostin C
was able to alter the sensitizing BAM22 effect significantly.
The sensitizing effect of BAM22 partly depends
on TRPV1
The heat-induced CGRP release from skin partly depends
on TRPV1 and is reduced in TRPV1 deficient mice (Zim-
mermann et al., 2005). As our reconstructive colocaliza-
tion studies showed that approximately 50% of MrgC-ir
neurons in the lumbar DRG show colocalization with
TRPV1, we examined whether the sensitizing BAM22
reversed, and the excitatory effect with BAM22⫹naloxone could be
reproduced including sensitization to heat. (B) Averaged histogram of
heat responses of 11 CMH units (smoothed by three-point running
average procedure). On average, there was no change in the heat
response due to BAM22⫹naloxone (both 10 ␮M) applied for 5 min.
The trace below illustrates the heat stimulus.
 U. Hager et al. / Neuroscience 151 (2008) 242–254
effect affects TRPV1 channels. Accordingly, the CGRP
release induced by capsaicin 0.31 ␮M (190⫾26 pg/ml,
n⫽10) could significantly be facilitated by application of
BAM22 10 ␮M⫹naloxone 10 ␮M (295⫾17 pg/ml, n⫽10,
P⬍0.05 Wilcoxon; Fig. 7B).
Furthermore, we tested whether the sensitizing effect
of BAM22 on heat-induced CGRP release could be
blocked by the TRPV1 antagonist capsazepine. From pre-
vious work capsazepine is known to prevent TRPV1 sen-
sitization to heat as well as to proton stimulation without
altering the basal response capacities (Vyklicky et al.,
1998; Fischer et al., 2003; Petho et al., 2004). In this work,
the isolated rat skin of one side was preconditioned with
BAM22 10 ␮M and naloxone 10 ␮M together with capsaz-
epine 3 ␮M, the contralateral side served as control and
was treated with plain BAM22 10 ␮M⫹naloxone 10 ␮M.
The BAM22 effect on the heat-induced CGRP release
reached 196⫾38 pg/dl (n⫽8) on the capsazepine treated
and 188⫾32 pg/dl (n⫽8) on the control side; the difference
was not significant (not shown). Capsazepine is a compet-
itive partial antagonist of the TRPV1 receptor that blocks
the intracellular binding site for capsaicin and endogenous
agonists. Therefore, other sensitizing mechanisms could
not be excluded.
To further check for TRPV1 dependence, we tested
whether the BAM22 effect can be reproduced in the mouse
where most MrgC expressing neurons are reported to lack
the TRPV1 receptor (Dong et al., 2001; Zylka et al., 2003,
2005). To this end, we compared heat-induced CGRP
release in TRPV1-deficient and WT mice. As in the rat,
BAM22⫹naloxone significantly increased in WT mice the
heat-stimulated (46 °C) CGRP release 2.2-fold, reaching
61⫾8 pg/ml (n⫽6) on the treated side and 28⫾7 pg/ml
(n⫽6) on the control side (Fig. 7D; P⬍0.05, Wilcoxon). In
TRPV1 deficient mice the CGRP release on the control
side reached only 22⫾4 pg/ml (n⫽6), but BAM22 still
significantly augmented the release 2.1-fold to 45⫾13
pg/ml (Fig. 7D; P⬍0.05, n⫽6, Wilcoxon). These data indi-
cate that the heat sensitizing mechanism of BAM22 is
independent of TRPV1, although the capsaicin response is
facilitated by BAM22.
Single-fiber recording
Altogether 26 slow conducting single-fiber units expressing
mechanical and heat sensitivity (CMH) were tested with
MrgC agonists. A first population of 11 C-fibers was treated
with BAM22 10 ␮M in combination with naloxone 10 ␮M.
All of these single-fibers were polymodal (CMH) and had
an average conduction velocity of 0.44 m/s (range 0.33–
0.58 m/s). The von Frey thresholds ranged from 2 to 128
mN. In 10 cases receptive field superfusion (for 5 min) of
BAM22 with naloxone had no excitatory effect by itself and
no influence on the subsequent heat response. However,
one single fiber showed a very impressive effect of
BAM22⫹naloxone (Fig. 8A). This CMH unit (conduction
velocity: 0.47 m/s, von Frey threshold: 90.5 mN) re-
sponded poorly to two initial heat stimuli with five and
seven spikes/stimulus above a threshold of 48 °C and
42 °C, respectively. The application of BAM22 with nalox-
251
one onto the receptive field caused an abrupt vivid excita-
tion with instantaneous discharge rates beyond 7/s that
lasted, with some adaptation, for the entire 5 min incuba-
tion period. The subsequent heat stimulus yielded a 4.5-
fold increased response of 27 spikes, a lowered heat
threshold just above 34 °C, and an increase of the peak
discharge rate from 2.3–10.9 spikes/s. Besides the excita-
tory effect on application, this corresponds to a major
sensitization to heat caused by BAM22 with naloxone.
During the subsequent 5 min washout phase, the CMH unit
remained almost silent, and the next heat stimulus showed
a partially reversed BAM22 effect (19 spikes/stimulus,
threshold 43 °C). The repeated superfusion of BAM22 with
naloxone again resulted in vivid excitation and marked
sensitization to heat (37 spikes/stimulus, threshold 36 °C),
followed again by partial reversal during washout (16
spikes/stimulus, threshold 42 °C).
On average, the heat responses of the 11 CMH units
reached 9.0⫾2 spikes/stimulus before and 13.7⫾3 spikes/
stimulus after pretreatment with BAM22 and naloxone; the
difference was not significant (Fig. 8B). Application of ␥2-
MSH 10 ␮M (n⫽6, data not shown) resulted in an average
heat response of 23.5⫾10 spikes/stimulus before and
15.8⫾5 spikes/stimulus after pretreatment; the difference
was again not significant.
To bridge the methodical gap between the electro-
physiologically employed heat stimuli (of 20 s ramp dura-
tion) and the ones used to induce CGRP release (5 min,
47 °C constant), we recorded from nine CMH units that
were pretreated with BAM22 10 ␮M and naloxone 10 ␮M
for 5 min and then exposed to a 5 min constant heat
stimulus of 47 °C (not shown). The same stimulus was
applied to five CMH fibers serving as the control (no treat-
ment with BAM22 and naloxone). In both groups there was
desensitizing discharge that ended between 30 and 90 s
after stimulus onset. The averaged peak discharge of
about six spikes/s in the BAM22⫹naloxone-treated group
versus nine spikes/s in the control group was not signifi-
cantly different. The same applied to the total discharge of
48.7⫾8 spikes/stimulus in the treated group versus
109.6⫾26 spikes/stimulus in the control group.
DISCUSSION
In this study we describe the immunohistochemical ex-
pression of the rat MrgC receptor in small sensory neu-
rons, satellite cells, and probably mast cells, as well as its
co-expression with markers of nociception, often including
IB4 and TRPV1 but rarely CGRP and SP. Nonetheless, the
opioid peptide agonist BAM22 caused a major sensitiza-
tion of heat-, but not KCl-, induced CGRP release from
isolated skin and peripheral nerve that was further en-
hanced by blocking the opioid effect of BAM22. Other
peptide agonists like ␥2-MSH were ineffective, and neither
histamine or prostaglandin release, nor PKC activation,
was involved in BAM22-induced sensitization to heat, that
was independent of TRPV1, although the capsaicin re-
sponse was facilitated by BAM22. Together this indicates a
252 U. Hager et al. / Neuroscience 151 (2008) 242–254
very unusual novel mechanism of nociceptor sensitization
of potential importance for neurogenic inflammation.
Immunofluorescence staining comparison of the novel
rat MrgC antibodies with the human MrgX1 antibodies
resulted in an identical staining pattern in rat DRG, show-
ing mainly small- and medium-sized neurons [as previ-
ously reported (Lembo et al., 2002)] that can be consid-
ered to be potential nociceptors. In addition, we noticed
strong satellite cell staining that colocalized with the glial
cell marker GFAP.
Our colocalization studies demonstrate that MrgC are
associated with the IB4 and TRPV1 classes of nociceptors,
whereas only a small proportion of MrgC-ir neurons co-
stains for CGRP or SP. These findings on the protein level
agree in principle with those previously described on the
mRNA level, using in situ hybridization (Lembo et al.,
2002). Our findings are in accord with the presence of at
least two phenotypically distinct groups of nociceptive neu-
rons. It has recently been shown that the presence or
absence of the transcription factor RunX1 determines if a
neuron differentiates into the Ret receptor/IB4 positive or
the TrkA receptor/CGRP-type, respectively (Chen et al.,
2006). The expression of Mrg as well as TRPV1 was also
markedly reduced in the absence of Runx1 (Chen et al.,
2006). While these findings reveal a transcriptional switch
and the previous in situ hybridization/co-localization stud-
ies for MrgC with TRPV1 resulted in open contradiction
(Lembo et al., 2002 vs. Zylka et al., 2003), our results
demonstrate an almost equally frequent association of
MrgC with IB4/Ret and TRPV1 in the 50% range of ir DRG
neurons. Finally it is the protein that conveys function, and
its level depends not only on transcriptional regulation/
gene expression, but also on trafficking into and translo-
cation in peripheral terminals.
The newly raised antibodies allowed colocalization of
MrgC with CGRP and other neuronal markers in the spinal
nerve axons leaving the DRG, as well as in the nerve
endings of the dermis and in the subepidermal layer of the
skin. This proved axonal trafficking of MrgC to the periph-
ery and coexpression—to a low degree—with the secre-
tory neuropeptide CGRP. Both results encouraged func-
tional investigations.
Noxious heat stimulation is a validated way to induce
graded CGRP release from isolated rat skin in the temper-
ature range of 42–50 °C in a calcium-dependent manner
(Kessler et al., 1999), and heat-induced CGRP release has
proven sensitive to facilitation by inflammatory mediators
as well as to desensitization through muscarinic GPCR
activation (Kessler et al., 1999; Bernardini et al., 2001,
2002). The isolated rat skin was conditioned with several
established MrgC agonists. In accord with the low degree
of colocalization of Mrg with CGRP, BAM8-22 and ␥2-MSH
had no effect on the heat-induced CGRP release from
isolated skin. Nonetheless, BAM22 induced a significant
sensitization, even more in combination with naloxone.
BAM8-22 and ␥2-MSH have both been described to be
more potent MrgC agonists in HEK293 cells than BAM22
(Grazzini et al., 2004) and both peptides evoked heat
hyperalgesia when injected into rat hind paws, although in
millimolar concentrations (Grazzini et al., 2004). Thus, in
our experiments, the sensitizing BAM22 effect on CGRP
release is not likely mediated through the MrgC receptor.
Although BAM22 binds with high affinity to heterolo-
gously expressed MrgC in HEK293 cells, it also contains
an opioid motif and therefore may activate opioid receptors
expressed in and inhibiting the nociceptive nerve terminals
(Lembo et al., 2002). Indeed, blocking opioid receptors by
naloxone caused an additional increase of the BAM22-
facilitated CGRP release. Correspondingly, activation of
␬-opioid receptors has previously been shown to diminish
the stimulated CGRP release from rat skin (Averbeck et
al., 2001). The effect of naloxone in combination with
BAM22 can therefore be understood as a disinhibition by
opioid receptor blocking, although naltrexone has also
been reported to reduce the inositol phosphates produc-
tion stimulated by BAM22 in HEK293 cells transfected with
the ␦-opioid receptor and the human MrgX1 (Breit et al.,
2006). As BAM22 with naloxone did not alter the CGRP
releasing effect induced by unspecific depolarization using
KCl, the BAM22 effect can be considered as selective for
the sensory transduction involved in heat-induced CGRP
release. Thus, BAM22 does not appear to affect the Ca2⫹-
dependent basal neurosecretory mechanisms of vesicular
exocytosis that are a prerequisite for using stimulated
CGRP release as an index of sensory activation.
The localization, distribution and density of MrgC-ir in
dermal mononuclear cells are suggestive of mast cells.
This finding raised the possibility of secondary BAM-pep-
tide effects induced by mechanisms not directly related
to the nociceptive nerve ending, but depending on cells
in their close vicinity. MrgX are expressed in human mast
cells activated by basic secretagogues, leading to mast
cell degranulation (Tatemoto et al., 2006) and histamine
release, known to induce nociceptor sensitization to heat
(Mizumura et al., 1994). Mast cells and fibroblasts are an
ample source of prostaglandins that are as well involved in
nociceptor sensitization to heat (Petho et al., 2001). We
could exclude both secondary mechanisms by blocking the
cyclooxygenases with flurbiprofen and measuring the cu-
taneous histamine release. This suggests that the sensi-
tizing effect of BAM22 is directly related to the nociceptive
nerve ending. In support of this conclusion, BAM22 in-
duced an equivalent facilitation of stimulated CGRP re-
lease from the axons of the desheathed sciatic nerve.
Several studies have previously been performed to
investigate the coupling of Mrg to G-proteins. Some of the
human (MrgX) receptors seem to exert their effects in a
pertussis toxin–sensitive way indicating a coupling to Gi/o-
proteins, others appear to be pertussis toxin insensitive
(Chen and Ikeda, 2004; Burstein et al., 2006). Han et al.
(2002) could show that the mouse MrgA1 and MrgC11, the
closest homologues to the rat MrgC, are coupled to Gq/11.
As the rat MrgC appears to be PTX-insensitive in HEK
cells (Grazzini et al., 2004), MrgC could be coupled in the
same way as in the mouse. Gq/11 activates phospholipase
C liberating diacylglycerol which stimulates PKC to phos-
phorylate the effector proteins, e.g. heat-activated ion
channels (Kessler et al., 1999). However, inhibition of PKC
 U. Hager et al. / Neuroscience 151 (2008) 242–254
by Gö6976 as well as calphostin C did not affect the
sensitizing effect of BAM22 on heat-induced cutaneous
CGRP release, indicating that the PKC pathway is not
involved.
The high degree of colocalization of MrgC with TRPV1
in the rat DRG suggested a possible role of this heat-
activated ion channel in the heat-sensitizing mechanism of
BAM22. BAM22 significantly facilitated the CGRP release
induced by the TRPV1 agonist capsaicin. On the other
hand, blocking TRPV1 by capsazepine had no effect on
the BAM22-sensitized heat-induced CGRP release. Cap-
sazepine is a competitive partial antagonist of the TRPV1
receptor that only blocks the intracellular binding site for
capsaicin and endogenous agonists. Other, physical, acti-
vating mechanisms are not affected (Sauer et al., 2001;
Fischer et al., 2003). Thus, we examined the effect of
BAM22 in TRPV1⫺/⫺ mice lacking this receptor. As pre-
viously described (Zimmermann et al., 2005), the heat-
induced CGRP release was diminished in TRPV1 deficient
versus WT mice, but the knockouts still exhibited the same
proportional increase of the heat-induced CGRP release
upon application of BAM22 with naloxone compared with
the WT mice. Thus, the heat-sensitizing mechanism of
BAM22 seems to be independent of TRPV1, perhaps af-
fecting other heat-activated ion channels that may induce
CGRP release (Leffler et al., 2007). On the other hand, the
TRPV1 activating mechanism of BAM22 seems to be un-
related to the intracellular binding site of capsaicin.
Our immunohistochemical studies showed MrgC-ir in
nerve endings, justifying electrophysiological single-fiber
recordings and testing MrgC agonists on noxious heat
responsiveness. This approach has previously been
proven valuable to elucidate GPCR-mediated sensitizing
(Liang et al., 2001; Petho et al., 2001) and desensitizing
(Bernardini et al., 2002) mechanisms. Unfortunately we
found only 1 of 26 polymodal nociceptors that responded,
though vigorously, to BAM22 (10 ␮M) with naloxone in the
rat skin, whereby naloxone is considered unlikely in excit-
ing nociceptors. A previous behavioral study in rats re-
ported nocifensive reactions (flinching, licking etc.) in re-
sponse to intradermal injection of MrgC agonists, although
millimolar concentrations were used, which restricts com-
parability (Grazzini et al., 2004). The one outstanding
nerve fiber in our recordings also showed a major sensiti-
zation to heat stimulation that would correspond to the heat
hyperalgesia reported from the behavioral experiments.
Both BAM effects (excitatory and sensitizing) were revers-
ible upon washout and could be reproduced without obvi-
ous tachyphylaxis in this one fiber. The discrepancy be-
tween the singular phenomenon of BAM22-induced exci-
tation/sensitization to heat in the single-fiber recordings
and the clear-cut facilitatory effect on heat-induced CGRP
release (in rat and mouse) may be explained by the fact
that stimulated CGRP release is rather independent of
action potential generation and discharge (Nemeth et al.,
2003) but derives from tonic depolarization or ligand-con-
trolled channel opening inducing Ca2⫹ influx and vesicular
exocytosis. Such a depolarization may be subthreshold for
the generation of action potentials (Sauer et al., 2005), or
253
it may be sustained in contrast to the discharge activity
declining under the influence of adaptation or sodium
channel inactivation. The latter is the case with prolonged
(5 min) KCl-induced depolarization that causes only tran-
sient spiking (Sauer et al., 2005) but sustained, and even
increasing, CGRP release from rat skin (Hoffmann T, Carr
R, Reeh PW, unpublished observations).
CONCLUSION
In conclusion, our immunohistochemical studies in the rat
confirmed that the sensory neuron-specific MrgC is pre-
dominantly, but not exclusively, expressed in the IB4 and
TRPV1 classes of small- and medium-sized neurons of the
DRG, and it was demonstrated for the first time in periph-
eral nerve axons and in nerve endings in the upper dermis.
The MrgC agonist peptide BAM22 (10 ␮M) seemed to
activate and sensitize a single cutaneous nociceptor but
caused pronounced facilitation of the heat- and capsaicin-
induced CGRP release from rat/mouse skin as well as
desheathed sciatic nerve. As other established MrgC ag-
onist peptides failed and only a small subpopulation of
MrgC-ir neurons showed co-staining with CGRP, it is un-
likely that the marked BAM22 effects are mediated through
MrgC. Nevertheless, we could show that the strong heat-
sensitizing mechanism of BAM22 is unrelated to TRPV1 as
well as is independent of PKC activation, prostaglandin
formation, and mast cell degranulation in the skin. Thus,
we conclude that BAM22 uses an uncommon route to
induce nociceptor sensitization of potential importance for
pain and hyperalgesia as well as neurogenic inflammation.
Acknowledgments—The first breeding pairs of TRPV1 ⫹/⫺ mice
were a generous gift from John B. Davis (GSK, Harlow, UK). We
would like to thank Iwona Izydorczyk, Annette Kuhn, Susanne
Haux-Oertel and Anette Wirth-Hücking as well as Karin Loeschner
for excellent technical assistance.
REFERENCES
Altschul SF, Madden TL, Schaffer AA, Zhang J, Zhang Z, Miller W,
Lipman DJ (1997) Gapped BLAST and PSI-BLAST: a new gener-
ation of protein database search programs. Nucleic Acids Res
25:3389 –3402.
Averbeck B, Reeh PW, Michaelis M (2001) Modulation of CGRP and
PGE2 release from isolated rat skin by alpha-adrenoceptors and
kappa-opioid-receptors. Neuroreport 12:2097–2100.
Averill S, McMahon SB, Clary DO, Reichardt LF, Priestley JV (1995)
Immunocytochemical localization of trkA receptors in chemically
identified subgroups of adult rat sensory neurons. Eur J Neurosci
7:1484 –1494.
Bender E, Buist A, Jurzak M, Langlois X, Baggerman G, Verhasselt P,
Ercken M, Guo HQ, Wintmolders C, Van den Wyngaert I, Van Oers
I, Schoofs L, Luyten W (2002) Characterization of an orphan G
protein-coupled receptor localized in the dorsal root ganglia re-
veals adenine as a signaling molecule. Proc Natl Acad Sci U S A
99:8573– 8578.
Bernardini N, Reeh PW, Sauer SK (2001) Muscarinic M2 receptors
inhibit heat-induced CGRP release from isolated rat skin. Neuro-
report 12:2457–2460.
Bernardini N, Roza C, Sauer SK, Gomeza J, Wess J, Reeh PW (2002)
Muscarinic M2 receptors on peripheral nerve endings: a molecular
target of antinociception. J Neurosci 22:RC229.
254 U. Hager et al. / Neuroscience 151 (2008) 242–254
Breit A, Gagnidze K, Devi LA, Lagace M, Bouvier M (2006) Simulta-
neous activation of the delta opioid receptor (deltaOR)/sensory
neuron-specific receptor-4 (SNSR-4) hetero-oligomer by the
mixed bivalent agonist bovine adrenal medulla peptide 22 acti-
vates SNSR-4 but inhibits deltaOR signaling. Mol Pharmacol 70:
686 – 696.
Burstein ES, Ott TR, Feddock M, Ma JN, Fuhs S, Wong S, Schiffer HH,
Brann MR, Nash NR (2006) Characterization of the Mas-related
gene family: structural and functional conservation of human and
rhesus MrgX receptors. Br J Pharmacol 147:73– 82.
Chen CL, Broom DC, Liu Y, de Nooij JC, Li Z, Cen C, Samad OA,
Jessell TM, Woolf CJ, Ma Q (2006) Runx1 determines nociceptive
sensory neuron phenotype and is required for thermal and neuro-
pathic pain. Neuron 49:365–377.
Chen H, Ikeda SR (2004) Modulation of ion channels and synaptic
transmission by a human sensory neuron-specific G-protein-cou-
pled receptor, SNSR4/mrgX1, heterologously expressed in cul-
tured rat neurons. J Neurosci 24:5044 –5053.
Choi SS, Lahn BT (2003) Adaptive evolution of MRG, a neuron-
specific gene family implicated in nociception. Genome Res 13:
2252–2259.
Derow A, Izydorczyk I, Kuhn A, Reeh PW, Petho G (2007) Prostaglan-
din E(2) and I(2) facilitate noxious heat-induced spike discharge
but not iCGRP release from rat cutaneous nociceptors. Life Sci
[Epub ahead of print].
Dong X, Han S, Zylka MJ, Simon MI, Anderson DJ (2001) A diverse
family of GPCRs expressed in specific subsets of nociceptive
sensory neurons. Cell 106:619 – 632.
Fischer MJ, Reeh PW, Sauer SK (2003) Proton-induced calcitonin
gene-related peptide release from rat sciatic nerve axons, in vitro,
involving TRPV1. Eur J Neurosci 18:803– 810.
Grazzini E, Puma C, Roy MO, Yu XH, O’Donnell D, Schmidt R,
Dautrey S, Ducharme J, Perkins M, Panetta R, Laird JM, Ahmad S,
Lembo PM (2004) Sensory neuron-specific receptor activation elic-
its central and peripheral nociceptive effects in rats. Proc Natl Acad
Sci U S A 101:7175–7180.
Hager UA, Hein A, Lennerz JKM, Zimmermann K, Neuhuber WL,
Reeh PW (2005) Sensory neuron-specific receptor sensitizes rat
nociceptors and facilitate CGRP release. Program No. 170.9.2005
Abstract Viewer/Itinerary Planner. Washington, DC: Society for
Neuroscience.
Han SK, Dong X, Hwang JI, Zylka MJ, Anderson DJ, Simon MI (2002)
Orphan G protein-coupled receptors MrgA1 and MrgC11 are dis-
tinctively activated by RF-amide-related peptides through the Gal-
pha q/11 pathway. Proc Natl Acad Sci U S A 99:14740 –14745.
Kessler F, Habelt C, Averbeck B, Reeh PW, Kress M (1999) Heat-
induced release of CGRP from isolated rat skin and effects
of bradykinin and the protein kinase C activator PMA. Pain 83:
289 –295.
Kobayashi E, Nakano H, Morimoto M, Tamaoki T (1989) Calphostin C
(UCN-1028C), a novel microbial compound, is a highly potent and
specific inhibitor of protein kinase C. Biochem Biophys Res Com-
mun 159:548 –553.
Leffler A, Linte RM, Nau C, Reeh P, Babes A (2007) A high-threshold
heat-activated channel in cultured rat dorsal root ganglion neurons
resembles TRPV2 and is blocked by gadolinium. Eur J Neurosci
26:12–22.
Lembo PM, Grazzini E, Groblewski T, O’Donnell D, Roy MO, Zhang J,
Hoffert C, Cao J, Schmidt R, Pelletier M, Labarre M, Gosselin M,
Fortin Y, Banville D, Shen SH, Strom P, Payza K, Dray A, Walker P,
Ahmad S (2002) Proenkephalin A gene products activate a new
family of sensory neuron-specific GPCRs. Nat Neurosci 5:201–209.
(Accepted 24 September 2007)
(Available online 4 November 2007)
Liang YF, Haake B, Reeh PW (2001) Sustained sensitization and
recruitment of rat cutaneous nociceptors by bradykinin and a novel
theory of its excitatory action. J Physiol 532:229 –239.
Lou X, Scheuss V, Schneggenburger R (2005) Allosteric modulation of
the presynaptic Ca2⫹ sensor for vesicle fusion. Nature 435:497–501.
Mizumura K, Minagawa M, Koda H, Kumazawa T (1994) Histamine-
induced sensitization of the heat response of canine visceral poly-
modal receptors. Neurosci Lett 168:93–96.
Muramatsu M, Tanaka M, Fujita A, Otomo S, Aihara H (1984) Time-
dependent inhibition of prostaglandins synthesis by nonsteroidal
antiinflammatory drugs: time-dependent alteration of inhibitory ef-
fect and characteristics of its active site. Jpn J Pharmacol 35:
37– 46.
Nemeth J, Helyes Z, Oroszi G, Jakab B, Pinter E, Szilvassy Z,
Szolcsanyi J (2003) Role of voltage-gated cation channels and
axon reflexes in the release of sensory neuropeptides by capsaicin
from isolated rat trachea. Eur J Pharmacol 458:313–318.
Petho G, Derow A, Reeh PW (2001) Bradykinin-induced nociceptor
sensitization to heat is mediated by cyclooxygenase products in
isolated rat skin. Eur J Neurosci 14:210 –218.
Petho G, Izydorczyk I, Reeh PW (2004) Effects of TRPV1 receptor
antagonists on stimulated iCGRP release from isolated skin of rats
and TRPV1 mutant mice. Pain 109:284 –290.
Reeh PW (1986) Sensory receptors in mammalian skin in an in vitro
preparation. Neurosci Lett 66:141–146.
Robas N, Mead E, Fidock M (2003) MrgX2 is a high potency cortista-
tin receptor expressed in dorsal root ganglion. J Biol Chem 278:
44400 – 44404.
Sauer SK, Reeh PW, Bove GM (2001) Noxious heat-induced CGRP
release from rat sciatic nerve axons in vitro. Eur J Neurosci
14:1203–1208.
Sauer SK, Weidner C, Carr RW, Averbeck B, Nesnidal U, Reeh PW,
Handwerker HO (2005) Can receptor potentials be detected with
threshold tracking in rat cutaneous nociceptive terminals? J Neu-
rophysiol 94:219 –225.
Schaffer AA, Aravind L, Madden TL, Shavirin S, Spouge JL, Wolf
YI, Koonin EV, Altschul SF (2001) Improving the accuracy of
PSI-BLAST protein database searches with composition-based
statistics and other refinements. Nucleic Acids Res 29:2994 –
3005.
Tatemoto K, Nozaki Y, Tsuda R, Konno S, Tomura K, Furuno M,
Ogasawara H, Edamura K, Takagi H, Iwamura H, Noguchi M,
Naito T (2006) Immunoglobulin E-independent activation of mast
cell is mediated by Mrg receptors. Biochem Biophys Res Commun
349:1322–1328.
Vyklicky L, Knotkova-Urbancova H, Vitaskova Z, Vlachova V, Kress M,
Reeh PW (1998) Inflammatory mediators at acidic pH activate
capsaicin receptors in cultured sensory neurons from newborn
rats. J Neurophysiol 79:670 – 676.
Young D, Waitches G, Birchmeier C, Fasano O, Wigler M (1986)
Isolation and characterization of a new cellular oncogene encoding
a protein with multiple potential transmembrane domains. Cell
45:711–719.
Zimmermann K, Leffler A, Fischer MM, Messlinger K, Nau C, Reeh
PW (2005) The TRPV1/2/3 activator 2-aminoethoxydiphenyl bo-
rate sensitizes native nociceptive neurons to heat in wildtype but
not TRPV1 deficient mice. Neuroscience 135:1277–1284.
Zylka MJ, Dong X, Southwell AL, Anderson DJ (2003) Atypical expan-
sion in mice of the sensory neuron-specific Mrg G protein-coupled
receptor family. Proc Natl Acad Sci U S A 100:10043–10048.
Zylka MJ, Rice FL, Anderson DJ (2005) Topographically distinct epi-
dermal nociceptive circuits revealed by axonal tracers targeted to
Mrgprd. Neuron 45:17–25.