PirB Restricts Ocular-Dominance Plasticity in Visual Cortex

See allHide authors and affiliations

Science  22 Sep 2006:
Vol. 313, Issue 5794, pp. 1795-1800
DOI: 10.1126/science.1128232


Experience can alter synaptic connectivity throughout life, but the degree of plasticity present at each age is regulated by mechanisms that remain largely unknown. Here, we demonstrate that Paired-immunoglobulin–like receptor B (PirB), a major histocompatibility complex class I (MHCI) receptor, is expressed in subsets of neurons throughout the brain. Neuronal PirB protein is associated with synapses and forms complexes with the phosphatases Shp-1 and Shp-2. Soluble PirB fusion protein binds to cortical neurons in an MHCI-dependent manner. In mutant mice lacking functional PirB, cortical ocular-dominance plasticity is more robust at all ages. Thus, an MHCI receptor is expressed in central nervous system neurons and functions to limit the extent of experience-dependent plasticity in the visual cortex throughout life. PirB is also expressed in many other regions of the central nervous system, suggesting that it may function broadly to stabilize neural circuits.

Plasticity of connections during development is thought to be driven by cellular processes that strengthen or weaken existing synapses in response to neuronal activity, followed by long-term structural alterations to circuits. The cellular and molecular machinery responsible for synaptic plasticity are well studied (1, 2), but the mechanisms and molecules that couple short-term synaptic changes to long-term structural remodeling are less understood.

One family of proteins that is important for activity-dependent structural remodeling of neural circuits during development is MHCI (3). This family of transmembrane cell surface proteins was thought to act exclusively in cellular recognition by the immune system. However, MHCI genes are now known to be expressed in neurons, where they are regulated by neuronal activity (3, 4), and by adenosine 3′,5′-monophosphate response element–binding protein (CREB) (5) and are essential for normal synaptic plasticity (3). MHCI proteins function through interactions with a variety of transmembrane receptors on immune system cells (6, 7). These interactions are the means by which normal cells are distinguished from abnormal or foreign cells. In the nervous system, the mechanisms by which neuronal MHCI modulates synaptic development are not understood. One hypothesis inspired by examples from the immune system is that neuronal MHCI functions by engaging transmembrane MHCI receptors expressed on other neurons. These interactions could generate intracellular signals that ultimately alter synaptic strength, neuronal morphology, and circuit properties. One such MHCI receptor, known from its role in regulating immune cell activation, is PirB (813).

To examine whether PirB is expressed in the brain, we performed in situ hybridization using PirB-specific probes on mouse brain sections of various ages. Specific mRNA signal was detected throughout the brain at all ages tested, with strong expression in the cerebral cortex, hippocampus, cerebellum, and olfactory bulb (Fig. 1, A to C). These brain regions are also known to express MHCI mRNA and protein (3, 4). PirB protein is also expressed in the brain, as revealed by immunostaining sections with antibodies specific for the cytoplasmic domain of PirB (Fig. 1, D and E). Protein is detected on subsets of neuronal cell bodies (such as those located in cortical layers 5 and 6), on hippocampal and cerebellar neurons, on axonal pathways, and within neuropil.

Fig. 1.

PirB mRNA and protein are expressed in neurons. (A to C) 35S labeled PirB-specific probes representing the 3′ region of PirB mRNA were used to detect PirB mRNA in sections from mouse brains of various ages. In situ hybridizations are shown in darkfield optics (silver grains appear white). (A) P14 sagittal section. (B) P29 sagittal section. (C) Adult sagittal section (top) plus sense control (bottom). (D to J) Immunohistochemistry using PirB-specific antibodies. (D) P18 coronal section and (E) adult sagittal section stained with the A20 antibody to PirB. Scale bars in (A) to (E), 1 mm. (F) Growth cone of a cortical neuron 3 days in vitro (DIV) immunostained with anti-PirB 1477 (red), phalloidin (blue), and anti-synapsin (green). (G) Cortical neuron 18 DIV stained with anti-PirB (red), postsynaptic marker postsynaptic density protein 95 (PSD-95) (green), and Hoechst (blue). (H) Cortical neuron 14 DIV stained with anti-PirB (red), presynaptic protein synaptophysin (green), and Hoechst (blue). (I) Cortical neuron 14 DIV stained with anti-PirB (red), presynaptic protein synapsin (green), and Hoechst (blue). (J) Higher magnification of (I). Scale bars in (F) to (J), 10 μm.

To confirm that PirB protein is expressed in neurons, neocortex from embryonic day 15 (E15) or postnatal day 0 (P0) mice was dissociated, grown in vitro, and immunostained with one of three different PirB-specific antibodies, in conjunction with neuron-specific markers (Fig. 1, F to J). In these cultures, about 20 to 50% of the neurons are immunoreactive for PirB. Immunostaining is present in axonal growth cones, localized behind the actin-rich leading edge and the zone of synapsin-immunostained vesicles (Fig. 1F). PirB protein is enriched in neuronal processes (Fig. 1, G to J), where it appears as puncta. PirB immunostaining is often very close to, but rarely overlaps with, the presynaptic proteins synaptophysin or synapsin (Fig. 1, H to J), suggesting that PirB is localized at or near synapses. Together, these experiments demonstrate that PirB is expressed by central nervous system neurons both in vivo and in vitro.

Because PirB binds to MHCI in immune cells (10, 11), we investigated whether PirB protein can bind to MHCI on neurons. For these experiments, we generated a soluble recombinant fusion protein consisting of the extracellular domain of PirB fused to alkaline phosphatase (PirB-AP). First, we verified that this reagent binds to cells in an MHCI-dependent manner using cultured mouse embryo fibroblasts (MEFs) derived from wild-type mice or from mice with deleted β2-microglobulin and Tap1 genes (β2m/Tap–/–), in which surface expression of MHCI proteins is reduced (14). PirB-AP binds to wild-type MEFs, and this binding is markedly reduced on β2m/Tap–/– cells (Fig. 2A), indicating that PirB-AP does indeed bind to MHCI. Next, we tested whether PirB-AP binds to cultured cortical neurons derived from brains of wild-type or β2m/Tap–/– mice. As with the MEFs, PirB-AP binds to neurons, and again this binding is reduced on neurons with low surface levels of MHCI (Fig. 2B), indicating that soluble PirB binds to neurons in an MHCI-dependent manner. This binding is saturable (Fig. 2C), with relatively low affinity (Kd ∼ 1.3 μM, Fig. 2C), consistent with observations of PirB-MHCI interactions in the immune system (10), as well as with other known MHCI-receptor interactions (15). In sections of the cerebral cortex, the PirB-AP fusion protein binds along the cell bodies and dendrites of cortical pyramidal neurons (Fig. 2D), consistent with the observation that panspecific antibodies to MHCI also immunostain the dendrites of neurons in the cortex and hippocampus (4).

Fig. 2.

Soluble PirB binds to neurons. Soluble PirB-AP binds to MEFs (A) and to cultured cortical neurons (B). Binding is dependent on surface expression of MHCI protein, given that binding is reduced significantly in cultures derived from β2m–/–/Tap1–/– mice (P < 0.01 for MEFs and P = 0.03 for neurons). a.u., arbitrary units. (C) Binding of PirB-AP to neurons is saturable (top). Scatchard analysis (bottom) predicts a dissociation constant of ∼1.3 μM. (D) PirB-AP binds to pyramidal neurons in sections of cortex. Numbers indicate cortical layers. Scale bars, 250 μm (left and middle panels); 50 μm (right panel). Error bars in (A), (B), and (C) = 1 SEM.

Immunoprecipitation of PirB directly from the brain demonstrates that PirB exists primarily as a ∼130-kD glycosylated protein (Fig. 3); this has also been observed in the immune system (13, 16). Immunoprecipitation and Western blots were performed using different antibodies to ensure specificity. PirB can be detected in all brain regions at all ages tested (Fig. 3, A and B). We know that PirB from the brain is glycosylated because it is sensitive to deglycosylation by PNGaseF, which removes N-linked oligosaccharides (Fig. 3C); PirB is insensitive to EndoH, which cleaves a more restricted subset of oligosaccharides. Preparation of synaptosomes from mouse brain indicates that a substantial portion of PirB protein fractionates with synaptosomal plasma membranes but is distinct from the purified synaptic vesicle fraction (Fig. 3D), consistent with immunolocalization (Fig. 1, H to J) and suggesting that PirB may function at or near synapses.

Fig. 3.

PirB protein is expressed throughout the brain and forms complexes with Shp-1 and Shp-2. (A) PirB was immunoprecipitated using 6C1 monoclonal antibody and detected by Western blot using C19 polyclonal antibody from whole brain at P5, P12, P20, P28, or adult (Ad). (B) PirB protein immunoprecipitated from brain stem, thalamus and striatum (thal. striat), cerebellum, and cortex/hippocampus (ctx. hipp) derived from equal amounts of protein from P5 or P20 mouse brains. Control immunoprecipitation was nonspecific rat immunoglobulin in lysate from cerebellum. (C) PirB was immunoprecipitated from whole brain at P7 and treated with EndoH or PNGaseF glycosidases. (D) PirB was immunoprecipitated from synaptosomal fractions (30). PLT100, light membranes; SPM, synaptic plasma membranes; VES, synaptic vesicle fraction; SUP100, soluble fraction. Synapsin and Synaptophysin synaptic vesicle proteins were used as fractionation markers. (E) Subcellular fractionation from wild-type (WT) and PirBTM mouse brains. In wild-type mice, 130-kD mature PirB fractionates with heavy membranes (P10 fraction) and light membranes (P100 fraction), but none is detected in soluble S100 fraction. In contrast, mutant PirBTM is smaller and exhibits increased solubility; thus, a large proportion appears in the soluble (S100) fraction. (F) Anti-phosphotyrosine immunoprecipitation (I.P.) followed by anti-PirB Western blot reveals that PirB is phosphorylated, and no signal is detected in PirBTM mice. (G) Anti-PirB immunoprecipitation followed by anti-PirB Western blot, which is then stripped and probed for anti-phosphotyrosine. Only wild-type PirB is phosphorylated; no phosphorylated PirBTM is detected. (H) Anti–Shp-1 immunoprecipitation, followed by Shp-1 and PirB Western blots, demonstrates that PirB interacts with Shp-1 in brain. (I) Anti–Shp-2 immunoprecipitation followed by Shp-2 and PirB Western blots from P5 or P12 brains demonstrates a complex of PirB and Shp-2 in the brain. (J) The complex of PirBTM and Shp-2 is absent in PirBTM brains. Ig, immunoglobulin.

The discovery of PirB expression in central nervous system neurons raises the question of neuronal PirB function. Therefore, we created a mutant mouse in which we removed four of the exons that encode the transmembrane domain and part of the PirB intracellular domain, rendering PirB unable to convey signals across the plasma membrane (fig. S1). We refer to the resulting mutant mouse, which is completely viable, and the mutant protein as PirBTM. To assess this mutation, PirB protein from the PirBTM mouse brain was examined. Figure 3E shows subcellular fractionation of wild-type and PirBTM brain, followed by immunoprecipitation of PirB and the shorter PirBTM mutant proteins, respectively. Wild-type PirB fractionates with both heavy and light membranes, with no signal detected in the soluble fraction. The mutant PirBTM protein is smaller and fractionates with light membranes and cytosolic fractions. Thus, the loss of the transmembrane domain has altered the solubility of PirB.

The cytoplasmic domain of PirB contains four immunoreceptor tyrosine-based inhibitory motifs (ITIMs). Phosphorylation of these sites is known to recruit Shp-1 and Shp-2 phosphatases to PirB, which in turn modulates signal transduction pathways in the immune system (1618). When immunoprecipitation was performed with antibodies to phosphotyrosine from wild-type or PirBTM brains, followed by anti-PirB Western blot, we found that PirB was phosphorylated only in wild-type brains; no tyrosine-phosphorylated PirBTM protein was detected (Fig. 3F). This result was expected, given that PirBTM is not a transmembrane protein and thus is unable to engage ligand, which normally leads to phosphorylation (10, 11). Similarly, when PirB or PirBTM is immunoprecipitated directly and analyzed for PirB (Fig. 3G, top panel), followed by phosphotyrosine Western blot (Fig. 3G, bottom), only wild-type PirB is phosphorylated. In immune cells, phosphorylated PirB recruits and signals primarily through Shp-1 and Shp-2 phosphatases (8, 16); neuronal PirB also associates with these phosphatases (Fig. 3, H to J), suggesting that components of PirB-dependent signaling mechanisms are conserved between the immune and nervous systems. As expected, PirBTM does not coimmunoprecipitate with Shp-2 (Fig. 3J). Together, these observations demonstrate that the mutant PirBTM fails to transduce signals by means of phosphorylation of its remaining immunoreceptor tyrosine-based inhibitory motifs, which is the primary means by which PirB and other proteins of this class signal (7). Thus, we conclude that signaling through PirB in the brain of PirBTM mice is abrogated.

The search for neuronal MHCI receptors was inspired by the discovery that mice defective for MHCI surface expression have grossly normal brains but have specific abnormalities in synaptic connectivity and plasticity (3). In the visual system of β2m/Tap–/– mice, the ipsilateral projection from the retina to the lateral geniculate nucleus (LGN) is larger than normal, consistent with a defect in developmental synapse elimination. In the hippocampus, long-term potentiation is enhanced and long-term depression is absent, again consistent with a shift in synaptic plasticity toward strengthening at the expense of synaptic weakening (3). Initial examination of the PirBTM mouse also revealed no obvious phenotype. Gross brain histology was normal as assessed by Nissl staining (fig. S2A). In contrast to MHCI mutant mice, however, in PirBTM mice the gross pattern of connections from the retina to the LGN is indistinguishable from the wild-type mouse, with well-defined eye-specific domains (fig. S2B). This phenotypic difference likely reflects the fact that MHCI mutants are deficient in both the β2-microglobulin and the TAP1 genes, resulting in reduced surface expression of virtually all MHCI proteins [more than 50 in this large multigene family (14)]. Thus, MHCI mutant mice represent an extreme loss of function for all potential MHCI receptors expressed in the brain, including but not limited to PirB, as well as for possible receptor-independent MHCI functions such as those reported in the pheromone signaling system (19).

PirB protein is highly expressed in the cerebral cortex (Fig. 1). To examine whether PirB function is required in the cortex, the development of ocular dominance (OD) was assessed in the primary visual cortex of wild-type and PirBTM mice. The adult mouse visual cortex receives functional inputs from both eyes (Fig. 4A). Most of the visual cortex consists of a large monocular zone where neurons are visually driven exclusively by the contralateral eye. A more restricted region, the binocular zone (BZ), receives functional inputs from both ipsilateral and contralateral eyes (20, 21). During development however, neurons across a wider region of visual cortex receive functional inputs from the ipsilateral eye; by the fourth postnatal week, this region becomes restricted by activity-dependent mechanisms to the adult BZ (2224).

Fig. 4.

Enhanced OD plasticity in visual cortex of PirBTM mice, as shown by Arc mRNA induction [(A) to (J)] and transneuronal autoradiography [(K) to (M)]. (A) Schematic of visual system showing connections from the retina to the LGN to the visual cortex. The small BZ receives visual inputs through the LGN from both eyes. (B to D) Developmental restriction of ipsilateral eye representation proceeds normally in PirBTM mice. (B) Arc mRNA induced by visual stimulation of P34 mice, detected by in situ hybridization in cortex ipsilateral to stimulated eye. Yellow arrowheads delineate zone of Arc induction, which corresponds to the BZ. In normally reared mice at P34, the width of Arc induction in wild-type or in PirBTM mice appears indistinguishable. (C) Histograms represent average widths of Arc induction in layer 4. P19: wild-type n = 9, PirBTM n = 7; P34: wild-type n = 7, PirBTM n = 7; three to four sections per animal. (D) Averaged line scans from all wild-type (blue) or PirBTM (red) sections at P34. Scans were made blind to genotype and were aligned at left border of BZ (black vertical line, left). Blue or red vertical line indicates right border of Arc induction. (E to G) OD plasticity after ME is enhanced in PirBTM mice. (E) OD plasticity after ME from P22 to P31 as assessed by Arc induction. The width of in situ hybridization pattern expanded after ME in wild-type mice [compare (E) top with (B) top]. Zone of Arc induction after ME is even more extensive in PirBTM mice [(E) bottom] than in wild-type mice [(E) top]. (F) Consistently enhanced expansion in width of Arc induction in PirBTM versus wild-type mice after periods of ME from P19 to P25, P22 to P31, P31 to P36, and P100 to P110. Histograms represent average width of Arc induction. ME from P19 to P25: wild-type n = 5, PirBTM n = 9; P22 to P31: wild-type n = 6, PirBTM n = 6; P31 to P36: wild-type n = 9, PirBTM n = 9; P100 to P110: wild-type n = 5, PirBTM n = 5; three to four sections per animal. (G) Averaged line scans of layer 4 Arc signal in all wild-type (blue) or PirBTM (red) sections after ME from P22 to P31. Scans aligned at left border of BZ (vertical line, left). Blue or red vertical lines indicate right border. The width is larger in PirBTM mice. (H to J) Enhanced OD plasticity after MD by eyelid suture from P19 to P25 in PirBTM mice: wild-type n = 13, PirBTM n = 13, three to four sections per animal. (K to M) Transneuronal autoradiography reveals an increase in width of anatomical connections between LGN neurons representing ipsilateral eye and layer 4 of cortex after ME (P25 to P40) in PirBTM mice. (K) Darkfield autoradiographs showing increased width of transneuronally transported radioactive label representing input from the ipsilateral eye in layer 4 of cortex in PirBTM (bottom) versus wild-type (top) mice. (L) Histograms are averages from all mice (wild-type n = 7, PirBTM n = 6). (M) Averaged line scans from all wild-type or PirBTM sections. Scans were aligned at left border of BZ (black vertical line, left). Blue or red vertical lines indicate width of thalamocortical projection in wild-type or PirBTM mice. Error bars in (C), (F), (I), and (L) = 1 SEM. *P < 0.05; **P < 0.01. Scale bars, 500 μm.

The state of OD was assessed in the visual cortex by means of the activity-regulated immediate-early gene Arc. A brief (30-min) exposure of one eye to visual stimulation rapidly induces Arc mRNA exclusively in visual cortical neurons, revealing the extent and laminar distribution of cortical neurons functionally connected to the stimulated eye (24). In the hemisphere ipsilateral to the stimulated eye, a circumscribed zone of Arc induction (Fig. 4B) that coincides with the BZ (Fig. 4A) is present, allowing for quantitative high-resolution measurements of refinement and/or plasticity of the ipsilateral eye representation in the mouse (as well as the cat) visual cortex (24). This technique has the advantage, compared with other techniques, of not requiring anesthesia, which masks some forms of cortical plasticity (25). We compared the representation of the ipsilateral eye within the BZ of wild-type and PirBTM mice by means of Arc mRNA induction. All experiments and analyses were performed blind to genotype.

Stimulation of one eye induces Arc mRNA in cortical layers 2 to 4 and 6 [Arc is not expressed in layer 5 (24)]. In the hemisphere ipsilateral to the stimulated eye at P34, Arc induction is restricted to the BZ, revealing the adult pattern of OD (Fig. 4B) (24). The pattern of Arc induction in PirBTM mice at P34 is indistinguishable from that of wild-type mice at the same age (Fig. 4B). To quantify these observations, serial line scans were made through layer 4 across visual cortex, and the width of the zone of Arc induction was measured: The width in wild-type and PirBTM mice at P34 is identical (Fig. 4, C and D). It is conceivable that the BZ in P34 PirBTM mice arises from an earlier representation that is different from normal. However, the pattern of Arc induction at P19 is also indistinguishable between PirBTM and wild-type mice (Fig. 4C). In addition, the width of the zone of Arc induction is larger at P19 than at P34 in both genotypes (Fig. 4C). Together, these observations indicate that PirB function is not required for the normal developmental restriction of the ipsilateral eye representation within the BZ of visual cortex.

The OD of cortical neurons can be shifted readily by altering the relative amounts of activity between the two eyes; this is referred to as OD plasticity. The degree of OD plasticity is extensive during a critical period of development (26) and is far more limited at older ages (2325, 2729). During the critical period, closing or removing one eye for several days shifts OD markedly toward the open, remaining eye. This shift in OD can be assessed directly by means of Arc mRNA induction in the visual cortex ipsilateral to the remaining eye; the Arc mRNA signal expands to occupy a wider-thannormal zone across layer 4 (Fig. 4E) (24), as well as other cortical layers. Unexpectedly, after a period of monocular enucleation (ME), OD plasticity in PirBTM mice was significantly enhanced compared with that of wild-type mice (Fig. 4, E to G) during the critical period. The difference held true for ME from P19 to P25, which overlaps the peak of the critical period; ME from P31 to P36, at the end of the critical period; and ME from P22 to P31, which spans the peak of OD plasticity. In addition, ME from P100 to P110, which is considered adult and well beyond the critical period, produced a greater expansion of Arc induction in PirBTM than in wild-type mice. At the ages examined, the width of Arc induction in layer 4 of PirBTM mice increased from 22% (P31 to P36; P < 0.01) to as much as 54% (P19 to P25; P < 0.01) of that in wild-type mice (Fig. 4F).

It is possible that the notable enhancement of OD plasticity in PirBTM mice is caused by the deleterious effects of eye removal itself, rather than by deprivation or visual experience. Therefore, the OD plasticity experiment was repeated using monocular deprivation (MD) by means of eyelid suture from P19 to P25. In wild-type animals, this technique is known to produce less plasticity than does ME (24); nevertheless, in PirBTM mice, the width of Arc induction in layer 4 was still 17% greater than in wild-type mice (Fig. 4, H to J). Together, these experiments imply that PirB limits OD plasticity induced by either ME or MD.

The induction of Arc mRNA in the BZ of the visual cortex is a functional measure of OD, delineating the area of cortex containing neurons responding to visual stimulation of the ipsilateral eye. Anatomically, the expanded Arc signal present in PirBTM visual cortex after a period of ME may be due to increased horizontal connectivity within the visual cortex, or it may be due to an increase in the spread of thalamocortical axon terminals from the LGN representing the ipsilateral eye (Fig. 4A), or both. To examine whether LGN axons increased their territory within layer 4, we performed ME on wild-type and PirBTM mice at P25; then, at P40, the LGN input from the remaining eye to layer 4 was assessed by means of transneuronal transport after an intraocular injection of 3H-proline (Fig. 4, K to M) (30). The transneuronally transported label within layer 4 of PirBTM mice is 18% wider than that of wild-type mice, indicating that at least part of the expansion in width of Arc mRNA induction observed in the PirBTM visual cortex is due to an anatomical increase in the area of layer 4 receiving inputs from LGN axons.

The enhanced OD plasticity seen in PirBTM mice both at the structural (thalamocortical axons) and at the functional levels (Arc induction in the postsynaptic cortical neurons) indicates that PirB normally functions to restrict the extent of cortical OD plasticity not only during but also after the critical period. In contrast, the normal developmental restriction of functional inputs from the ipsilateral eye to the BZ of the visual cortex is intact in PirBTM mice, as is the normal developmental remodeling of eye-specific projections from the retina to the LGN. Together, these experiments show that PirB is needed to restrict the ability of neural circuits to readjust synaptic connections in response to alterations in activity levels or balance of inputs. In the absence of PirB function, OD plasticity in visual cortex is enhanced at all ages, with a particularly strong effect seen during the critical period. Thus, in addition to mechanisms that enable synaptic plasticity, our results show that there are other mechanisms, such as those that involve PirB, that limit the extent of synaptic plasticity, which must then be determined by a balance of signaling in both pathways.

In the immune system, PirB acts through Shp-1 and Shp-2 phosphatases to inhibit signals that could lead to inappropriate and dangerous activation of immune cells against normal, healthy cells. PirB is thought to regulate cytoskeletal dynamics, cell motility, and adhesion, acting downstream of Src family kinases to modulate integrin signaling, Ca++ signaling, and kinase cascades (18, 31, 32). We report here that in neurons, PirB also recruits both Shp-1 and Shp-2. Thus, PirB may have analogous functions in restricting the response of neurons to activity-dependent or Ca++-dependent signaling, thereby limiting aspects of synaptic plasticity. It is possible that cellular mechanisms that normally regulate and limit the selective strengthening or stabilization of synapses, the formation of new synapses, or even the outgrowth of new neurites after perturbations of sensory input are altered without functional PirB. This may result in exuberant and abnormal growth or strengthening of connections, as reflected in the increase seen here both in the width of thalamocortical connections representing the ipsilateral eye and in the width of Arc mRNA induction reflecting the functional activation of cortical neurons driven by the ipsilateral, open eye. In this context, important similarities emerge between the phenotypes of PirBTM mice and those observed previously in mice mutant for MHCI cell surface expression (3): In both strains of mutant mice, there is an enhancement of mechanisms that favor strengthening of synaptic connections.

It has long been known that there are critical periods during brain development when experience can rapidly alter circuits by changing synaptic connections, both by altering the structure of connections and by changing their strength functionally (26). Such periods are thought to be terminated by progressive developmental changes leading to an adult state in which plasticity is far more limited if it occurs at all. Here, we show that both during and after the critical period, the extent of plasticity is actively constrained by PirB. In the immune system, PirB is known to regulate integrin-dependent cytoskeletal dynamics (31, 33), and integrin activation in neurons can affect several aspects of synaptic function and plasticity (3437). In the visual system, studies show that the tissue plasminogen activator/plasmin extracellular proteolytic cascade is essential for structural plasticity in the visual cortex (38) and cleaves extracellular proteins including the integrin ligands laminin and fibronectin (39). Thus, PirB may restrict neuronal plasticity by affecting the ability of activated integrins to engage the neuronal cytoskeleton. Upon closure of the critical period, factors thought to constrain plasticity include Nogo Receptor (40) and extracellular matrix molecules chondroitin sulfate proteoglycans (41). How PirB function interacts with these factors is not known, but our experiments demonstrate a role for PirB in limiting the extent of synaptic plasticity in the visual cortex and may provide insight into mechanisms that are needed to stabilize neural circuits.

Supporting Online Material

Materials and Methods

Figs. S1 and S2


References and Notes

View Abstract

Stay Connected to Science

Navigate This Article