Regulation of T cell activation, anxiety, and male
aggression by RGS2
Antonio J. Oliveira-dos-Santos*, Goichi Matsumoto*, Bryan E. Snow*, Donglin Bai†
, Frank P. Houston‡
,
Ian Q. Whishaw§
, Sanjeev Mariathasan¶
, Takehiko Sasaki*, Andrew Wakeham*, Pamela S. Ohashi¶
,
John C. Roder†
, Carol A. Barnes‡
, David P. Siderovskiʈ
, and Josef M. Penninger*,
**
*Amgen Institute and ¶Ontario Cancer Institute, Departments of Medical Biophysics and Immunology, University of Toronto, 620 University Avenue,
Toronto, ON M5G 2C1 Canada; †Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto, ON M5S1A8 Canada; ‡Arizona Research
Laboratories, Division of Neural Systems, Memory, and Aging, University of Arizona, Tucson, AZ 85724; §Department of Psychology, University
of Lethbridge, AB T1K3 M4 Canada; and ʈDepartment of Pharmacology, University of North Carolina, Chapel Hill, NC 27599
Communicated by Alfred G. Gilman, University of Texas Southwestern Medical Center, Dallas, TX, August 29, 2000 (received for review June 20, 2000)
Regulators of G protein signaling (RGS) proteins accelerate the
GTPase activity of G␣ protein subunits in vitro, negatively regulating
G protein-coupled receptor signaling. The physiological role
of mammalian RGS proteins is largely unknown. The RGS family
member rgs2 was cloned as an immediate early response gene
up-regulated in T lymphocytes after activation. To investigate the
role of RGS2 in vivo, we generated rgs2-deficient mice. We show
that targeted mutation of rgs2 in mice leads to reduced T cell
proliferation and IL-2 production, which translates in an impaired
antiviral immunity in vivo. Interestingly, rgs2؊/؊ mice also display
increased anxiety responses and decreased male aggression in the
absence of cognitive or motor deficits. RGS2 also controls synaptic
development and basal electrical activity in hippocampal CA1
neurons. Thus, RGS2 plays an important role in T cell activation,
synapse development in the hippocampus, and emotive behaviors.
G protein-coupled receptors (GPCR) control fundamental
cellular processes ranging from embryogenesis and neurotransmitter
release to cell survival, proliferation, and differentiation.
Heterotrimeric G proteins are composed of ␣, ␤, and ␥
subunits that relay GPCR stimulation to downstream signaling
pathways (1). On ligand binding, GPCR undergoes conformation
changes of its intracellular loops that promote the exchange
of GDP for GTP on G␣ and the subsequent dissociation of G␣
and G␤␥. Both GTP-bound G␣ and free G␤␥ propagate the
signal through interactions with downstream effectors. Deactivation
of GPCR signaling depends on the rate of G␣ GTP
hydrolysis.
Studies of GPCR desensitization have led to the discovery of
a large and evolutionarily conserved family of GTP-activating
(GAP) factors for G␣ termed ‘‘regulators of G protein signaling’’
(RGS). RGS proteins share an RGS domain (Ϸ120 amino acids)
capable of accelerating G␣-dependent GTP hydrolysis and the
conversion of active GTP-bound G␣ to its inactive GDP-bound
form (2–7). The first RGS identified was the yeast protein Sst2,
isolated in a screen for mutants that failed to down-regulate the
GPCR-mediated response to mating pheromone (8–9). Over 20
mammalian RGS proteins have been identified (7, 10). Recently
it has been shown that RGS9–1 controls the photosensitization
response in the mouse eye (11). Axin, which contains a RGS
domain, negatively regulates the Wnt-signaling pathway, and
axin-mutant mice display defects in embryonic axis formation
(12). However, Axin has no detectable GAP activity for G␣ (10).
The physiological roles of other mammalian RGS proteins are
yet to be established.
The RGS family member rgs2 was cloned as an early response
gene up-regulated in activated T cells (13–14). RGS2 is also
up-regulated in central nervous system (CNS) neurons after
stimuli that evoke long-term neuronal plasticity (15–16). In vitro,
RGS2 interacts with G␣q/11 and stimulates its GTPase activity
(17–18), and RGS2 overexpression in cells inhibits G␣i- (15) and
G␣s-dependent (19) signaling. To investigate the role of RGS2
in vivo, we generated rgs2-deficient mice. rgs2 deficiency in mice
leads to impaired T cell activation in vitro and in vivo. Interestingly,
mice lacking RGS2 also exhibit reduced male aggressive
behavior and increased anxiety, whereas morphological and
electrophysiological analyses indicate a role for RGS2 in hippocampal
CA1 neuron function.
Materials and Methods
rgs2 Mutant Mice. RGS2 exons 4 and 5 encoding the RGS
domain were replaced with PGK-NeoR
in antisense orientation.
G418-resistant E14K embryonic stem (ES) cells were
screened by PCR and Southern blot for homologous
recombination. Targeted ES cell clones were injected into
C57BL͞6 blastocysts and heterozygous rgs-2ϩ/Ϫ
mice obtained
by mating chimeras with C57BL͞6J mice. PCR primers: wildtype
allele (5Ј-CCGAGTTCTGTGAAGAAAACATTG3Ј;
5ЈGGGACTCCTGGTCTCATGTAGCAT3Ј) and mutant allele
(5Ј-GCTAAAGCGCATGCTCCAGAC-3Ј; 5Ј-GGCCCACATTTACACGAACC-3Ј).
All experiments were carried out
by using littermate mice on a C57BL͞6J background (F5
generation), and mice were maintained in accordance with
institutional guidelines.
Immunocytometry. Single-cell suspensions of thymi, lymph nodes,
and spleens were stained with FITC-, phycoerythrin-, or biotinconjugated
Abs reactive to CD3␧, TCR␣␤, CD4, CD8, CD25,
CD44, CD28, CD69, CD11b (Mac-1), CD11c, B220, CD43,
sIgM, sIgD, or Gr-1. Biotinylated antibodies were visualized by
using streptavidin-RED670. Samples were analyzed by flow
cytometry by using a FACScan (Becton Dickinson).
T and B Cell Assays. For proliferation assays, freshly isolated T cells
were activated with PMA (10 ng͞ml, phorbol myristate acetate)
plus Caϩ2
ionophore (100 ng͞ml, ionomycin), soluble anti-CD3␧
(0.1–1 ␮g͞ml, clone 145–2C11), anti-CD28 (0.02–0.2 ␮g͞ml,
clone 37.51), or recombinant mouse IL-2 (50 units͞ml) in
triplicate for 12, 24, 48, 72, or 120 h, followed by a 12-h pulse with
[3
H]thymidine (1 ␮ Ci͞ml). Cytokine production was determined
by ELISA. For B cell analysis, splenic B lymphocytes were
purified (90–95% B220ϩ
cells) by negative magnetic sorting and
activated with various stimuli. Basal serum Ig levels (IgM, IgG1,
IgG2a, IgG2b, IgG3, and IgA) were determined by ELISA (20).
For lymphocytic choriomeningitis virus (LCMV) infections,
Abbreviations: GPCR, G protein-coupled receptors; GAP, GTP-activating; RGS, regulators of
G protein signaling; CNS, central nervous system; ES, embryonic stem; LCMV, lymphocytic
choriomeningitis virus; EPSP, excitatory postsynaptic potentials.
**To whom reprint requests should be addressed. E-mail: jpenning@amgen.com.
The publication costs of this article were defrayed in part by page charge payment. This
article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C.
§1734 solely to indicate this fact.
Article published online before print: Proc. Natl. Acad. Sci. USA, 10.1073͞pnas.220414397.
Article and publication date are at www.pnas.org͞cgi͞doi͞10.1073͞pnas.220414397
12272–12277 ͉ PNAS ͉ October 24, 2000 ͉ vol. 97 ͉ no. 22
mice were infected with LCMV (700 plaque-forming units, s.c.)
into the hind footpad. Cellular infiltration was measured daily as
described (21).
Behavior Studies. Four- to five-month-old sex-matched littermate
mice were examined for neuromotor functions (22): vibrissae,
eyes, rearing͞support͞standing, muscle tone, tail hanginginduced
turns and limbs͞digits extension, ear reflex, eye blink
and pupillary response to light, sweet͞bitter tasting acceptance͞
rejection, and orientation toward sound. Circadian and openfield
activities (i.e., cage crosses, total distance, vertical movements)
were measured in cages fitted with infrared sensors.
Exploration test was performed by putting the mice in the center
of a white disk (183-cm diameter), with or without objects
distributed in four quadrants. The mice were video tracked
(seven trials). Fecal pellets and rears were counted. Motor
coordination was analyzed by a rotating rod task. For a balance
task, a mouse was held by the tail and allowed to grasp a
suspended horizontal wooden rod (3 mm Ø) with its hind limbs,
with recording of the time for the mouse to pull itself into
balance and the number of falls.
Mice were studied in a spatial water maze for 7 days with a
hidden platform in the southwest quadrant (22). On day 7, a
final probe trial (60 s in the pool without platform) was
applied. From day 8 to 16, tests were reassumed with the
platform moved to the northeast quadrant (reversal, 2 trials
per day). Time to find the platform and total distance were
measured. In the T-maze test, on day 1 the mice were placed
in an empty maze for 10 min with access only to its central and
right arms. Several objects were placed as cues around the
maze. On day 2, each subject was again placed in the T-maze
and given access to the left arm of the maze. The mice were
observed for 5 min, with recording of time spent in the new
area and first choice. Step-down avoidance test was also
carried out, in which the mice were required to remember that
prior stepping off a safe platform resulted in a foot shock (1
s, 0.12 mA). Latency on the platform was recorded in each
trial. Aggression was tested by using displacement from a
plastic tube (30 cm long and 3.05 cm Ø). A rgs2Ϫ/Ϫ
and
littermate control mouse were released into either end of the
tube, with the first mouse to exit being deemed the ‘‘loser,’’
that is, this mouse was deemed to be the less aggressive of the
two. For light͞dark preference tests, cages were divided into
dark and light halves and the time spent by each mouse in each
half measured.
Electrophysiology. Electophysiology studies were carried out as
described (23). Single hippocampal slices (400 ␮m) were
superfused with artificial cerebrospinal fluid (2 ml͞min; saturated
with 95% O2 and 5% CO2 at 30°C). Field potentials
were sampled at 10 kHz and analyzed with PCLAMP6 software
(Axon Instruments, Foster City, CA). Field excitatory postsynaptic
potentials (EPSP) were recorded with a micropipette
(2–4 M⍀) placed in the stratum radiatum of CA1 area.
Stimulation was delivered at Schaffer collateral. Test stimulus
frequency was 0.1 Hz. The tetanic stimulation consisted of four
trains of 100-Hz stimulation lasting 500 ms at 10-s intervals.
EPSP slope was calculated from linear regression of the rising
phase 10–65% of the peak response. The baseline value of
EPSP slope was that from the first 5-min period and was
defined as 100%. Input͞output relationship was constructed
by plotting the amplitude of presynaptic volley (input) vs.
EPSP slope (output).
Results and Discussion
rgs2 Gene Targeting. The rgs2 locus was inactivated in ES cells by
replacing exons 4 and 5 corresponding to the RGS domain (Fig.
1A). Chimeric mice were bred to obtain germ-line transmission
of the mutated rgs2 allele (Fig. 1 B and C). The rgs2 mutation
proved to be null, as shown by Northern blot analyses of brain
tissue (Fig. 1D). rgs2Ϫ/Ϫ
mice were born at the expected Mendelian
frequency, were fertile, and displayed normal growth. No
anatomical or histological abnormalities were observed. Differentiation
of all hematopoietic lineages, including neutrophils,
macrophages, erythrocytes, and platelets, was normal in rgs2Ϫ/Ϫ
mice.
Fig. 1. Generation of rgs2-deficient mice. (A) rgs2 targeting vector. (B)
Southern blot (see 3Ј flanking probe in A) from rgs2 mutant ES cell clones.
Wild-type (WT) and mutant (KO) bands are shown. (C) Genotyping of rgs2
mice. WT and KO PCR primers are shown in A (arrows). Loss of the wild-type
rgs2 allele was confirmed by genomic Southern blot. (D) Northern blot analysis
of brain mRNA from rgs2 WT (ϩ͞ϩ), heterozygous (ϩ͞Ϫ), and KO (Ϫ͞Ϫ) mice
(full-length rgs2 cDNA probe).
Oliveira-dos-Santos et al. PNAS ͉ October 24, 2000 ͉ vol. 97 ͉ no. 22 ͉ 12273
MEDICALSCIENCES
Defective Proliferation and IL-2 Production by rgs2؊/؊ T Cells. Because
RGS2 expression is induced in activated T cells after TCR
engagement or stimulation by mitogens such as ConA or PMA
plus Ca2ϩ
ionophore (PMA ϩ Iono) (ref. 14 and data not
shown), we determined whether RGS2 plays a role in T cell
activation. rgs2Ϫ/Ϫ
mice displayed normal numbers and subsets
of CD4ϩ
and CD8ϩ
T cells and B cells in lymphoid organs (Fig.
2A). The surface levels of TCR␣␤, CD3, CD4, CD8, CD28,
CD45, CD5, H2Kb
, CD44, LFA1, CD25, and CD69 on both
splenic and lymph node peripheral CD4ϩ
and CD8ϩ
T cells were
comparable in rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
mice (not shown). To examine
peripheral T cell function, T cells from the lymph nodes of
rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
littermates were stimulated with anti-CD3␧
mAb, anti-CD3␧ plus anti-CD28 mAbs, or PMA plus Ca2ϩ
ionophore.
The proliferation in vitro of rgs2Ϫ/Ϫ
T cells was
significantly reduced in response to TCR engagement, a deficit
that could not be overcome by CD28 costimulation (Fig. 2B).
Consistent with the identification of RGS2 as an early response
gene in activated T cells (13–14), the proliferative defect was
most obvious at 24 h of stimulation. Induction and kinetics of
apoptosis were comparable among rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
T cells
(not shown). Thus, RGS2 acts as an immediate early gene
downstream of TCR activation in T cells.
In addition to their proliferative defect, rgs2Ϫ/Ϫ
T cells produced
significantly lower levels of the T cell growth factor IL-2
compared with controls (Fig. 2C). IL-2 insufficiency per se does
not explain the proliferative deficit because IL-2 supplementation
did not restore proliferation to normal (Fig. 2B). The
high-affinity IL-2 receptor ␣ chain (CD25) and the activation
markers CD69, CD44, ICAM1, and CTLA4 were up-regulated
to a normal extent and with normal kinetics in rgs2Ϫ/Ϫ
T cells in
response to TCR engagement or mitogen treatment (not shown).
The functional defect in T cell proliferation (Fig. 2B) and
cytokine production (Fig. 2C) can also be observed in rgs2Ϫ/Ϫ
T
cells treated with PMA͞Ca2ϩ
ionophore, a stimulus that bypasses
the proximal TCR signal. Consistent with this finding, the
kinetics and extent of early TCR-mediated signaling events,
including Ca2ϩ
mobilization, overall tyrosyl phosphorylation,
and activation of MAPK and SAPK͞JNKs, were normal in
rgs2Ϫ/Ϫ
lymph node T cells (not shown), suggesting that their
proximal TCR signaling pathways are intact.
With regards to B cells, rgs2Ϫ/Ϫ
mice displayed normal numbers
and differentiation (not shown). In vitro proliferation of
rgs2Ϫ/Ϫ
B cells in response to treatment with anti-IgM Ab, the
F(abЈ)2 fragment of the anti-IgM Ab, anti-CD40, or LPS was
similar to that of rgs2ϩ/Ϫ
B cells (Fig. 2D and data not shown).
Furthermore, baseline serum Ig levels were equivalent in rgs2Ϫ/Ϫ
and rgs2ϩ/Ϫ
mice. Thus, RGS2 is not required for B cell
activation. Our results show that antigen receptor-triggered
RGS2 expression in T cells is required for proliferation and IL-2
production.
Chemokines are critical regulators of T cell function, signaling
via GPCRs expressed on the surface of resting and activated T
cells (24). It has also been shown that overexpression of RGSfamily
members in cell lines can modulate chemokine receptor
signaling (18). However, no differences were observed between
both resting and preactivated (6-h activation with anti-CD3␧
Fig. 2. Impaired proliferation and IL-2 production by rgs2Ϫ/Ϫ T cells. (A)
Normal populations of CD4ϩ and CD8ϩ T cells and B220ϩ B cells in lymph nodes
of rgs2Ϫ/Ϫ mice. Numbers in each quadrant represent percentages of each
subset. (B) Proliferation. Lymph node T cells (2 ϫ 105͞well) from rgs2Ϫ/Ϫ and
rgs2ϩ/Ϫ mice were activated by using PMA [10 ng͞ml]ϩCaϩ2-ionophore [100
ng͞ml], TCR stimulation [anti-CD3␣ mAb (0.1 ␮g͞ml)] with and without CD28
costimulation [anti-CD28 mAb (0.02 ␮g͞ml)] or IL-2 [50 units͞ml]. Results are
shown as mean [3H]thymidine uptake Ϯ SD. Difference in proliferation between
rgs2ϩ/Ϫ and rgs2Ϫ/Ϫ T cells was statistically significant (t test, P Ͻ 0.05).
(C) IL-2 production. Lymph node T cells were activated as above, and IL-2 was
measured by ELISA at 24 and 48 h poststimulation. Mean values of IL-2
production Ϯ SD are shown. (D) Activation of B cells. Purified splenic B cells
(1 ϫ 105͞well) were incubated in triplicate for 24 and 48 h in medium alone
(None) or medium containing anti-IgM (20 ␮g͞ml), anti-IgM (FabЈ)2 (15 ␮g͞ml)
with or without anti-CD40 (5 ␮g͞ml). Results are shown as mean [3H]thymidine
uptake Ϯ SD.
12274 ͉ www.pnas.org Oliveira-dos-Santos et al.
before chemokine treatment) rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
CD4ϩ
or
CD8ϩ
T cells in either proliferation or cell migration in response
to the chemokines SDF-1␣, RANTES, or MIP-1␣ (not shown).
These results do not preclude regulation by RGS2 of T cell
responsiveness to signals mediated by other chemokine receptors.
Furthermore, rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
T lymphocytes showed
comparable responses to the GPCR agonists lysophosphatidic
acid, histamine, and adenosine (not shown). The GPCRs whose
functions are under RGS2 control in T cells remain to be
identified.
Impaired Immunity in rgs2؊/؊ Mice. To determine whether RGS2
plays a role in T cell responses in vivo, rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
mice
were injected in the footpad with LCMV (21). rgs2ϩ/Ϫ
mice
developed an effective LCMV-induced footpad swelling reaction
starting from day 6 postinfection (Fig. 3). In contrast, the
footpad swelling reaction was significantly decreased and delayed
in rgs2Ϫ/Ϫ
mice, indicating that RGS2 expression is required
for an efficient antiviral response in vivo. Thus, the in vitro
deficit in proliferation and IL-2 production by rgs2Ϫ/Ϫ
T cells
translates into an impaired anti-virus response in vivo, indicating
that RGS2 expression is physiologically relevant for normal T
cell function. These results provide the first genetic evidence that
RGS proteins play a role in the activation of lymphocyte.
RGS2 Controls Male Aggression and Anxiety. During our evaluation
of the mutant mice, we noted unusual fighting between male
rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
littermates that sometimes resulted in severe
injuries. Stimuli triggering long-term plasticity in the CNS induce
the up-regulation of RGS2 in the hippocampus, cortex, caudate
putamen, striatum, and amygdala (15–16). Because these regions
regulate a variety of behavioral pathways, we investigated
whether RGS2 plays a role in CNS function. We first determined
whether RGS2 deficiency in mice leads to alterations in motor
functions. Both rgs2Ϫ/Ϫ
and rgs2ϩ/Ϫ
littermate mice showed
normal motor responses (tests listed in Materials and Methods).
Circadian activity was normal in the mutant mice (Fig. 4A), as
was exploratory behavior during two independent open field and
exploration tests (not shown). rgs2Ϫ/Ϫ
mice also exhibited appropriate
performances in rotating rod (Fig. 4B) and balance
tests (not shown), confirming the integrity of the neurological
pathways responsible for motor coordination and balance. Spatial
and conditional learning, as determined by the hidden
platform, water maze, and passive avoidance tests, were comparable
between rgs2Ϫ/Ϫ
and rgs2ϩ/Ϫ
littermates. Long-term
potentiation in the Schaffer collateral CA1 pathway of individual
hippocampus slices was also equivalent in rgs2Ϫ/Ϫ
and rgs2ϩ/Ϫ
mice (Fig. 4C). These data show that RGS2 deficiency has no
apparent effect on motor responses, circadian activity, exploratory
behavior, motor coordination, or spatial learning and
memory.
Social dominance and territorial aggression are wellcharacterized
behaviors in male mice (25). Intriguingly, rgs2Ϫ/Ϫ
male mice exhibited significantly reduced aggressive behavior
compared with their rgs2ϩ/Ϫ
littermates, whereas the aggressive
behavior of rgs2Ϫ/Ϫ
female mice was normal (Fig. 5A). Because
aggression correlates with anxiety (25), we subjected rgs2Ϫ/Ϫ
male mice to the light͞dark preference test. rgs2Ϫ/Ϫ
mice showed
a greater preference for the dark compared with rgs2ϩ/Ϫ
control
littermates (Fig. 5B), indicating increased anxiety. Consistent
with this behavioral phenotype, rgs2Ϫ/Ϫ
mice were found to drop
significantly more fecal pellets than rgs2ϩ/Ϫ
littermates during
the light͞dark preference test (2.3 Ϯ 0.6 vs. 0.8 Ϯ 0.3 pellets) and
Fig. 3. Reduced LCMV-induced footpad swelling in rgs2Ϫ/Ϫ mice. Mice were
infected in the footpad with LCMV and swelling measured. Average footpad
swelling (percentage increase compared with footpads thickness before infection)
of six animals per group.
Fig. 4. Normal circadian activity, motor functions, and long-term potentiation
(LTP) in rgs2Ϫ/Ϫ mice. (A) Circadian activity was monitored during a 24-h
period (12-h light͞dark cycles). Mean crosses accumulated per 30-min intervals
Ϯ SD are shown. (B) Normal motor function in a rotarod test. Mean values
of retention times (latency) Ϯ SD are shown. Trials indicate a stepwise increase
in rotarod speed (10–40 rpm, 5-rpm increase per trial). (C) Tetanus-induced
LTP in CA1 neurons. Averaged normalized field EPSP (fEPSP) slope measurements
obtained from 11 hippocampus slices of 4 rgs2ϩ/Ϫ and 4 rgs2Ϫ/Ϫ mice.
Oliveira-dos-Santos et al. PNAS ͉ October 24, 2000 ͉ vol. 97 ͉ no. 22 ͉ 12275
MEDICALSCIENCES
exploration test (3.5 Ϯ 0.6 vs. 1.3 Ϯ 0.6 pellets). In addition,
rgs2Ϫ/Ϫ
mice exhibited an increased response to acoustic startling.
These data show that RGS2 controls neuronal circuits for
aggression and anxiety.
RGS2 Role in Synapse Development and Basal Electrical Activity of
Hippocampal CA1 Neurons. Fear and anxiety usually involve the
amygdala, and certain components of fear can be affected by
hippocampal inputs into the amygdala (26). Neurons from the
CA1 region of the hippocampus project to the amygdala and
removal of the hippocampus affects fear-based behavior (27).
Moreover, it has also been proposed that RGS2 can regulate the
activity of the G protein-coupled metabotropic glutamate reFig.
5. RGS2 regulates male aggression and anxiety. (A) Reduced male
aggressive behavior. Aggression was measured in paired groups of male or
female rgs2Ϫ/Ϫ and rgs2ϩ/Ϫ mice by using a replacement test. Results represented
as percentages of displacement per group (‘‘wins’’). Eighteen trials
were carried out for each group. (B) Increased anxiety. The anxiety response
was measured by using a dark͞light preference test. Note the increased
preference for the dark environment in rgs2Ϫ/Ϫ mice. The difference in mean
dark͞light latency Ϯ SD was statistically significant (t test; P Ͻ 0.05).
Fig. 6. Reduced spine density in CA1 hippocampal neurons of rgs2Ϫ/Ϫ mice. (A–D) Histology of hippocampus of rgs2ϩ/Ϫ (A and B) and rgs2Ϫ/Ϫ (C and D) mice.
Blue lines demarcate the CA1 neuron region. Normal structure of the CA1 region in rgs2Ϫ/Ϫ mice. Golgi–Cox staining, ϫ4 and ϫ200. (E) Reduced numbers of spines
in rgs2Ϫ/Ϫ CA1 neurons. Densities (spines͞10 ␮m Ϯ SD) of apical and basilar spines were determined by camera lucida. Differences in apical and basilar spine
densities were statistically significant (t test, P Ͻ 0.005).
Fig. 7. Impaired basal electric activity in CA1 neurons of rgs2Ϫ/Ϫ mice. (A and
B) Reduction of input͞output relationship in CA1 pathway in rgs2Ϫ/Ϫ mice.
Superimposed recordings of CA1 field excitatory postsynaptic potentials
(fEPSP) evoked by electrical stimulation of Schaffer collateral input. Input:
peak amplitude of presynaptic volley. Output: field EPSP slope measured at
10–65% rising phase of EPSP. Data are from seven hippocampus slices of four
rgs2ϩ/Ϫ and four rgs2Ϫ/Ϫ mice.
12276 ͉ www.pnas.org Oliveira-dos-Santos et al.
ceptor 1, a receptor implicated in long-term synaptic plasticity in
the hippocampus (28). Detailed examination of rgs2Ϫ/Ϫ
mice
revealed that the number of neurons in the hippocampus (Fig.
6 A–D), in other brain nuclei, the brainstem, cerebellum, and
cortex (not shown) were not affected by RGS2 deficiency.
Interestingly, whereas dendritic morphology and branching were
comparable among cortical neurons of rgs2ϩ/Ϫ
and rgs2Ϫ/Ϫ
mice,
hippocampal CA1 neurons of rgs2Ϫ/Ϫ
mice showed a significant
decrease in the density of apical and basilar spines compared
with those of rgs2ϩ/Ϫ
control littermates (Fig. 6E). The number
of spines correlates with synapse numbers and neuronal plasticity
(29). Importantly, the morphological deficit in spine numbers
of rgs2Ϫ/Ϫ
CA1 neurons correlated with significantly reduced
electrical input͞output relationship in these cells (Fig. 7 A and
B). These data show that RGS2 has a role in synaptic development
and basal electrical activity of hippocampal CA1 neurons.
Alterations in aggressive behavior and fear responses are
important manifestations in highly prevalent psychiatric disorders
such as depression or addictive behaviors. Because these
disorders are responsible for a large social-economic burden, it
is critical to identify gene products potentially involved in the
initiation and͞or progression of these behavioral ‘‘traits’’ into
clinically manifest ‘‘disease.’’ Multiple G protein-coupled receptors
can regulate these behaviors, and pharmacological drugs
that modulate these receptors have been developed to mitigate
anxiety disorders and aggression. Our study provides genetic
evidence that loss of a molecular regulators of G-protein signaling,
RGS2, leads to impaired synaptic development in the
hippocampus, reduced male aggressive behavior, and increased
fear. Thus, RGS-family member such as RGS2 could be involved
in the genetic predisposition to human neuropsychiatric disorders
and could be potential targets for the generation of novel
neuropharmacological drugs. rgs2Ϫ/Ϫ
mice also provide a useful
animal model to elucidate molecular mechanisms of fear and
male aggression.
G protein-coupled receptors regulate fundamental processes
of development, cell activation, and behavior. The present study
shows that RGS2, a member of the RGS family, has important
roles in T cell proliferation and IL-2 production. The T cell
activation deficit observed in vitro translates into an impaired
anti-virus response in vivo, indicating that RGS2 is physiologically
relevant for normal T cell function and the establishment
of an effective antiviral immune response. In addition, mice
lacking RGS2 exhibit significantly reduced aggressive behavior
and increased anxiety in the absence of any measured cognitive
and motor deficits. Furthermore, morphological and electrophysiological
analyses indicate that RGS2 expression is required
for the generation of synaptic plasticity in hippocampal CA1
neurons. Thus, RGS2 is a critical regulator of T cell activation,
neuronal function, and emotive fear, and aggression in vivo.
We are most grateful to K. Bachmaier, Y.-Y. Kong, and I. Kozieradzki
for help and to M. Saunders for editing of the manuscript. J.M.P. and J.R.
are supported by grants from the Medical Research Council of Canada.
D.P.S. is supported by National Institutes of Health Grant AA11605.
1. Gilman, A. G. (1987) Annu. Rev. Biochem. 56, 615–649.
2. Berman, D. M., Wilkie, T. M. & Gilman, A. G. (1996) Cell 86, 445–452.
3. Watson, N., Linder, M. E., Druey, K. M., Kehrl, J. H. & Blumer, K. J. (1996)
Nature (London) 383, 172–175.
4. Hunt, T. W., Fields, T. A., Casey, P. J. & Peralta, E. G. (1996) Nature (London)
383, 175–177.
5. Snow, B. E., Hall, R. A., Krumins, A. M., Brothers, G. M., Bouchard, D.,
Brothers, C. A., Chung, S., Mangion, J., Gilman, A. G., et al. (1998) J. Biol.
Chem. 273, 17749–17755.
6. Koelle, M. R. & Horvitz, H. R. (1996) Cell 84, 115–125.
7. Berman, D. M. & Gilman, A. G. (1998) J. Biol. Chem. 273, 1269–1272.
8. Chan, R. K. & Otte, C. A. (1982) Mol. Cell. Biol. 2, 11–20.
9. Dohlman, H. G., Song, J., Ma, D., Courchesne, W. E. & Thorner, J. (1996) Mol.
Cell. Biol. 16, 5194–5209.
10. Siderovski, D. P., Strockbine, B. & Behe, C. I. (1999) Crit. Rev. Biochem. Mol.
Biol. 34, 215–251.
11. Chen, C. K., Burns, M. E., He, W., Wensel, T. G., Baylor, D. A. & Simon, M. I.
(2000) Nature (London) 403, 557–560.
12. Zeng, L, Fagotto, F., Zhang, T., Hsu, W., Vasicek, T. J., Perry 3rd, W. L., Lee,
J. J., Gumbiner, B. M. & Costantini, F. (1997) Cell 90, 181–192.
13. Siderovski, D. P., Heximer, S. P. & Forsdyke, D. R. (1994) DNA Cell Biol. 13,
125–147.
14. Heximer, S. P., Cristillo, A. D. & Forsdyke, D. R. (1997) DNA Cell Biol. 16,
589–598.
15. Ingi, T., Krumins, A. M., Chidiac, P., Brothers, G. M., Chung, S., Snow, B. E.,
Barnes, C. A., Lanahan, A. A., Siderovski, D. P., Ross, E. M., et al. (1998)
J. Neurosci. 18, 7178–7188.
16. Burchett, S. A., Volk, M. L., Bannon, M. J. & Granneman, J. G. (1998)
J. Neurochem. 70, 2216–2219.
17. Heximer, S. P., Watson, N., Linder, M. E., Blumer, K. J. & Hepler, J. R. (1997)
Proc. Natl. Acad. Sci. USA 94, 14389–14393.
18. Beadling, C., Druey, K., Richter, G., Kehrl, J. & Smith, K. A. (1999) J. Immunol.
162, 2677–2682.
19. Tseng, C-C. & Zhang, X. Y. (1998) Endocrinology 139, 4470–4475.
20. Liu, Q., Oliveira-Dos-Santos, A. J., Mariathasan, S., Bouchard, D., Jones, J.,
Sarao, R., Kozieradzki, I., Ohashi, P., Penninger, J. M. & Dumont, D. (1998)
J. Exp. Med. 188, 1333–1342.
21. Penninger, J. M., Fischer, K. D., Sasaki, T., Kozieradzki, I., Le, J., Tedford, K.,
Bachmaier, K., Ohashi, P. S. & Bachmann, M. F. (1999) Eur. J. Immunol. 29,
1709–1718.
22. Whishaw, I. Q., Haun, F. & Kolb, B. (1999) in Modern Techniques in
Neuroscience, eds. Windhorst, U. & Johansson, H. (Springer, New York).
23. Lu, Y. M., Roder, J. C., Davidow, J. & Salter, M. W. (1998) Science 279, 1363–1367.
24. Kehrl, J. H. (1998) Immunity 8, 1–10.
25. Parmigiani, S., Ferrari, P. F. & Palanza, P. (1998) Neurosci. Biobehav. Rev. 23,
143–153.
26. Rogan, M. T. & LeDoux, J. E. (1996) Cell 85, 469–475.
27. White, N. M. & McDonald, R. J. (1993) Behav. Brain. Res. 55, 269–281.
28. Kammermeier, P. J. & Ikeda, S. R. (1999) Neuron 22, 819–829.
29. Moser, M. B. (1999) Cell. Mol. Life Sci. 55, 593–600.
Oliveira-dos-Santos et al. PNAS ͉ October 24, 2000 ͉ vol. 97 ͉ no. 22 ͉ 12277
MEDICALSCIENCES