Expression of immediate-early genes reveals functional compartments within ocular dominance columns after brief monocular inactivation

Abstract

Visual inputs from the 2 eyes in most primates activate alternating bands of cortex in layer 4C of primary visual cortex, thereby forming the well-studied ocular dominance columns (ODCs). In addition, the enzymatic reactivity of cytochrome oxidase (CO) reveals “blob” structures within the supragranular layers of ODCs. Here, we present evidence for compartments within ODCs that have not been clearly defined previously. These compartments are revealed by the activity-dependent mRNA expression of immediate-early genes (IEGs), zif268 and c-fos, after brief periods of monocular inactivation (MI). After a 1–3-h period of MI produced by an injection of tetrodotoxin, IEGs were expressed in a patchy pattern that included infragranular layers, as well as supragranular layers, where they corresponded to the CO blobs. In addition, the expressions of IEGs in layer 4C were especially high in narrow zones along boundaries of ODCs, referred to here as the “border strips” of the ODCs. After longer periods of MI (>5 h), the border strips were no longer apparent. When either eyelid was sutured, changes in IEG expression were not evident in layer 4C; however, the patchy pattern of the expression in the infragranular and supragranular layers remained. These changes of IEG expression after MI indicate that cortical circuits involving the CO blobs of the supragranular layers include aligned groups of neurons in the infragranular layers and that the border strip neurons of layer 4C are highly active for a 3-h period after MI.

Several important features of the anatomical and physiological organization of primary visual (striate) cortex (V1) of macaque monkeys have been known since the early studies of Hubel and Wiesel (1, 2). Visual afferents from the 2 eyes have segregated inputs into V1, forming periodical “ocular dominance columns” (ODCs) perpendicular to the pial surface. Geniculocortical afferents related to each eye terminate separately in ODCs within layer 4C, which, in turn, project in a less restricted manner into ODCs of supragranular layers. Enzymatic reactivity of cytochrome oxidase (CO) in the mitochondria reflects chronic neuronal activity, and its pattern in the supragranular layers is patchy, revealing the so-called “CO puffs” or “blobs” (3). The oval blobs are located in the centers of ODCs, and they have distinct patterns of connectivity and neuron response properties associated with the color processing streams of the koniocellular and parvocellular pathways (4).

Long-term (24 h or more) inactivation of 1 eye by enucleation, tetrodotoxin (TTX) injection, or deprivation by eyelid suture significantly reduces neuronal activity in the ODCs that mainly receive projections from the inactivated eye, and CO activity is decreased in the deprived columns to demarcate the shape of ODCs (5). The expression of immediate-early genes (IEGs), such as Zif268 and c-Fos, is also strongly dependent on neuronal activity, and histochemical staining for these genes reveals ODCs after monocular inactivation (MI) or monocular deprivation (MD) (5, 6). Here, we report that the expression patterns of zif268 and c-fos in ODCs after a brief period (within 3 h) of MI are quite different from those that follow a longer period of MI in V1 of adult macaque monkeys and that this pattern reveals a compartmental organization within ODCs that has not been obvious before. Based on these data, we propose a previously undescribed model of V1 organization that accounts for the changes in neuronal activity that follow MI produced by TTX injection or MD produced by eyelid suture.

Results

The pattern of mRNA expression of the IEGs zif268 and c-fos in V1 of an adult macaque with normal binocular vision (monkey 0) is presented in Fig. 1A. Although the mRNA signals were intense, especially in layers 2/3, 4A, 4Cβ, and 6, ODCs and blob structures were not revealed. In contrast, after 1 h of inactivation of retinal activity by monocular TTX injection (monkey 1), the patterns of mRNA expression changed significantly. In coronal sections, periodical ODCs were observed as a result of reductions of mRNA expression in the deprived columns (Fig. 1B, filled and open circles). In addition, the IEG mRNA signals revealed further compartments within the ODCs. First, very intense signals were observed along the edges of ODCs in layer 4Cβ (Fig. 1B, ↓). Second, the intense mRNA signals in the supra- and infragranular layers were thinner in the nondeprived columns than the width of ODCs (Fig. 1B, black brackets). Furthermore, thin column-like structures were seen in the center of the deprived columns in IEG mRNA expression (Fig. 1B, open brackets). These compartments were more fully revealed in in situ hybridization (ISH) sections cut tangentially to the pial surface (Fig. 1 C and D). These sections revealed that the shapes of the deprived and nondeprived ODCs as IEG mRNA signals were overall less intense in the deprived columns. In the superficial layers, the blob structure was more evident than the shapes of ODCs because of the less intense expression of IEG mRNAs in interblob regions, even in the nondeprived columns. To our surprise, although distinct CO blobs in macaques have been described for only the supragranular layers, the blob-like pattern by IEG mRNA signals was also observed in infragranular layers. The IEG mRNA expressions in the infragranular layers corresponded to the locations of blobs of superficial layers observed in serial sections (Fig. 1D, black arrowheads). In tangential sections through layer 4C, edges of ODCs were quite clear because the IEG mRNA signals were especially intense along the boundaries of ODCs. The boundary signals resemble the “border strips” previously proposed by Horton and Hocking (5), and that term is used here. Although the expression patterns were quite similar between zif268 and c-fos, the border strips were more visible in ISH for zif268 and the infragranular blob structure was more visible in ISH for c-fos.

Fig. 1.

Coronal sections of ISH for IEGs, zif268 and c-fos, in V1 of monkey 0 (A, normal) and monkey 1 (B, 1-h TTX). (B) IEG mRNA expression was low in the deprived columns (filled circles) after MI. Note that the widths of strong IEG mRNA expression (black brackets) are thinner than the actual widths of nondeprived columns (open circles) and that the IEG mRNA expression remains in the center of deprived columns (open brackets). In addition, strong IEG mRNA signals were observed around the border of ODCs in layer 4Cβ (↓). (Scale bars = 500 μm in A and 1 mm in B.) (C) Tangential sections of ISH for IEGs in V1 of monkey 1. The depth of the section from the pial surface is indicated on each panel. Sections are arranged by increasing depth. (Scale bar = 1 mm.) (D) High-magnification views of rectangular windows in C (a–c) and their adjacent section of superficial layer (270 μm away from the pia) stained for CO reactivity (d). Broken lines in a, c, and d indicate the boundary of ODCs with reference to the patterns in b. Black arrowheads on the photomicrographs indicate the locations of CO blobs. Compartments of border strips (BSs), dark columns (DCs), and pale columns (PCs) are indicated in b. (b, Inset) Representation of a magnified view of the cortex within the rectangular area at a single-cell resolution. (Scale bar = 500 μm.)

A similar pattern of IEG expression was also obtained after 2.5 h of MI by TTX injection (monkey 2) (Fig. 2A), although the appearances of both blobs and border strips were less conspicuous than those after 1 h of MI. After 5 h of MI (monkey 3), blobs and border strips were only faintly revealed (Fig. S1). Twenty-four hours after the TTX injection (monkey 4), blobs and border strips were hardly visible, although the expression pattern of IEGs clearly demarcated ODCs with almost no expression in the deprived columns and intense expression in the nondeprived columns (Fig. 2B), similar to the expression pattern of IEGs after a long period of MI (6). We measured the widths of border strips, dark columns, and pale columns, as shown in Fig. 1D-b. The results from monkeys 1 and 2 are shown in Table 1. Because the widths of the dark columns, including the border strips, were only slightly greater than those of the pale columns, the border strips appeared to be mostly parts of the nondeprived columns, although they likely extended slightly into the deprived columns.

Fig. 2.

Coronal (Left) and tangential (Center and Right) sections of ISH for zif268 in V1 of monkey 2 (A, 2.5-h TTX) and monkey 4 (B, 24-h TTX) and those of ISH for c-fos in monkey 6 (C, 3-h eyelid suture). The depth of the section from the pial surface is indicated on each panel. (Scale bar = 1 mm.) (D) Tangential section of ISH for c-fos in layer 6 (840 μm away from the pia) in monkey 6 (Left) and a serial section in layers 2/3 (270 μm away from the pia) stained for CO reactivity (Right). Broken lines indicate the shapes of ODCs as revealed by the IEG mRNA expression. Black arrowheads on the photomicrographs indicate the locations of CO blobs. (Scale bar = 1 mm.) (E) Coronal section of ISH for c-fos in the dLGN ipsilateral to the deprived eye in monkey 1. The level of c-fos mRNA expression was lower in layers 2, 3, and 5 than in layers 1, 4, and 6 of dLGN. (Scale bar = 1 mm.)

Table 1.

Widths of each column in monkeys 1 and 2

	Monkey 1	Monkey 2
Pale column, μm	382.6 ± 10.4	366.2 ± 14.1
Dark column, μm	250.6 ± 12.8	273.4 ± 12.8
Border strip, μm	77.3 ± 1.8	93.5 ± 2.5
BS × 2 + DC	405.2	460.4

Measured widths of pale columns (PCs), dark columns (DCs), and border strips (BSs) as shown in Fig. 1D-b. Presumably, DCs correspond to the deprived columns and PCs correspond to nondeprived columns. Given that the widths of the DCs are shorter than those of the PCs, the BSs likely belong to nondeprived columns, although part of the BS likely overlaps the deprived column, because the value of BS × 2 + DC is larger than the widths of PCs.

MD by eyelid suture produced less pronounced changes of IEG mRNA expression than TTX injections. After 40 min of eyelid suture (monkey 5), the shapes of the ODC were not distinct. Only supra- and infragranular blob-like structures were observed (Fig. S1). After 3 h of eyelid suture (monkey 6), a significant reduction of IEG mRNA expression was observed in the deprived columns outside of layer 4C and blob-like structures were distinct in the centers of the deprived columns, especially in infragranular layers (Fig. 2C). This expression was closely aligned with the CO blobs in serial sections through the superficial layers (Fig. 2D). However, changes in IEG mRNA expression were only weakly apparent in layer 4C, and the ODCs were not clearly revealed. Similar results were obtained in cases with longer MD by eyelid suture (6 and 24 h: monkeys 7 and 8, respectively), but the blob-like structures were less conspicuous after these longer MD periods (Fig. S1).

Finally, a reduction of IEG mRNA expression was also observed even after a brief period of MI or MD in the dorsal lateral geniculate nucleus (dLGN) layers that receive inputs from the TTX-treated or sutured eye, but a border pattern of increased or decreased IEG mRNA signals was not seen along any of the dLGN layers in any monkey examined in this study (Fig. 2E and Fig. S2).

Discussion

Functional Architecture of V1.

From these findings, we propose a model of the functional architecture of macaque V1 (Fig. 3A). The functional unit that is seen in CO preparation in the supragranular layers extends to layer 6, with the exception of layer 4C. In addition, border strips reside in the vicinity of the boundaries of ODCs in layer 4C. As for the ODCs, the normally cryptic border strips are revealed by IEG mRNA expression after MI. Border strips of the left and right eye columns adjoin each other with the boundary of ODCs between them. Previously, Horton and Hubel (3) described CO blobs that included infragranular layers in macaques, and the existence of dim CO blobs in infragranular layers has been suggested for squirrel monkeys and prosimian galagos (7, 8). Horton and his colleagues have also suggested the existence of border strips along ODCs in layer 4C in terms of a gap between anterograde tracer signals from the open eye and CO enzymatic activity after long-term MI (5) and pale stripes in layer 4C in CO staining in monkeys with experimental strabismus (9). Although there has been no subsequent evidence for this architecture, our data strongly support their model except for their suggested absence of blob structure in layer 4A, because our data indicate that the blob structure also includes layer 4A.

Fig. 3.

(A) Our schematic model of V1 architecture. Blob structure extends from layer 2 to layer 6, with the exception of layer 4C. In addition, there is border strip structure in the vicinity of boundaries of ODCs in layer 4C. Although we did not detect strong IEG mRNA signals in layers 4B and 4Cα, previous reports can be referred to for the structure (8, 25). (B) Our schematic view of changes of IEG expression showing the distribution of somas of active neurons after MI by TTX or MD by eyelid suture (in this case, the left eye). In MI, IEG mRNA expression decreases in interblob regions in both columns, especially in deprived columns, and increases in border strips in nondeprived columns. After a period, IEG mRNA expression eventually levels out within each column. This sequential pattern is mostly similar in MD by eyelid suture, except that IEG mRNA expression hardly changes in layer 4C. L, projection from left eye; R, projection from right eye.

Changes of IEG Expression After MI.

IEG expression has been examined in V1 after MI by plenty of researchers (6, 10–12); however, the characteristic pattern of IEG expression shown here has been overlooked in the past. Here, our high-resolution ISH methods in combination with short MI revealed the patterns of IEG expression that likely reflect progressive changes in neuronal activity after the disruption of binocular vision.

Transcription of c-fos is triggered by intracellular calcium and cAMP, which are enhanced through the activation of NMDA receptors, L-type calcium channels, and/or trkB receptors, and it is thought that the expression is enhanced by the neuronal activity where action potential is coincident with synaptic activity (13). The turnover rate of IEG transcripts is remarkably fast: their accumulation peaks at ≈30 min after the onset of neuronal activity and decreases to background level ≈30 min after the offset of neuronal activity (14).

Previously, Horton et al. (12) reported that the Zif268 immunoreactivity is occasionally higher in the columns related to the enucleated eye than in the columns responsive to the intact eye, whereas the opposite pattern occurs in other cases with monocular enucleation. We have also observed this paradoxical reversal of the expected patterns in 2 of the 6 cases with longer periods (5–21 days) of MI (Fig. S3). However, the unexpected increases in expression were limited in layer 4C, and other layers in the columns related to the inactivated eye had less zif268 labeling than those layers in the columns related to the normal eye, as revealed by decreases in mRNA expression of occ1, 5-HT1B, and 5-HT2A (15, 16) and in CO activity. A reversed pattern of layer 4C labeling of zif268 was not observed in the 4 short-term MI cases reported here, and the decreases of IEG matched those of the V1-selective genes (Fig. S4). Thus, the decreases of IEG in the present cases occurred in the deprived columns related to the inactivated eye. The occasional reversed pattern of activation in layer 4C after longer periods of MI is yet to be explained.

As illustrated in Fig. 3B, the pattern of change in activity after MI or MD, which is revealed by IEG mRNA expression, is rather complex. In the acute phase immediately after the TTX injection, activity in the interblob regions appears to decrease rapidly in both deprived and nondeprived columns, whereas activity is maintained in the supra- and infragranular blob regions. Conversely, neurons in border strips of layer 4C in the nondeprived columns may become more responsive. Activity changes develop more slowly in the core regions of this layer. In later phases, the activity changes spread throughout the whole ODCs, such that activity is low in the entire deprived columns and high in the entire nondeprived columns, and the activity differences in the border strips and the blob regions disappear. More studies are needed to explain this changing pattern. A possible interpretation is that interblob regions and layer 4C border strips receive binocular inputs; when lateral inhibition from the inactivated eye is reduced, activity levels increase. Alternatively, binocular activations of the interblob regions may be reduced relative to blob regions. The evidence here that the blob regions in the deprived ODCs have higher levels of initial activity after MI runs counter to other evidence (3) that blobs are highly monocular centers within the ODCs. In either case, once the retinal activity of 1 eye is completely inactivated, the balance of binocular interactions in V1 is suddenly disrupted, causing rapid increases or decreases of neuronal activity in the binocular portions. The later slower changes may be associated with adaptations for the monocular vision, accompanied by long-term depression or potentiation of neurons, and possibly by an activity-dependent remodeling of neuronal circuits after MI. It should be noted that MI effects on gene expression in V1 were not likely caused by any effects of TTX other than the effect on blocking of activity, because changes in gene expression on cortex occurred within 1 h when there was no possible transport of TTX from the injected eye to V1.

IEG Expression After Eyelid Suture.

The pattern of activity change in V1 after MD by eyelid suture was similar to that produced in MI in layers 2/3 and 5/6. However, different from MI, persistent expression of IEGs was observed in the deprived portion of layer 4C in the MD cases. This result is consistent with the CO staining pattern and physiological properties of V1. CO activity is hardly decreased in layer 4C after several months of monocular eyelid suture, whereas a decrease in CO activity is evident in layers 2/3 (5). After eyelid suture, spontaneous activity remains high in retinal ganglion cells, whereas it is almost completely suppressed by TTX. The preserved retinal activity after eyelid suture is likely relayed to layer 4C of V1, where enough neuronal activity is maintained to prevent a significant loss of IEG mRNA expression. Because layer 2/3 neurons integrate visual information from layer 4C neurons to process information about edges, illumination, color, and motion (1), eyelid suture abolishes such meaningful visual information and decreases neuronal activity of layer 2/3 neurons. Such laminar differences in neuronal activity after eyelid suture may explain why neurons in extragranular layers precede layer 4 neurons in ocular dominance plasticity in juvenile animals after eyelid suture (17) and why layer 2/3 neurons and layer 4 neurons appear to have parallel mechanisms for ocular dominance plasticity (18).

Future Perspective.

The histological patterns shown in this paper are expected to reflect underlying differences in neuron connections, morphologies, and physiological properties. The short-term period of high activity in the aligned supragranular and infragranular components may be explained with the evidence that they form integrated circuits for functions that differ from those of the interblob regions (19, 20). Yet, little is known about the circuitry that maintains higher levels of activity in the blob relative to the interblob subdivision of V1 in both monocularly deprived and nondeprived columns. Likewise, the high levels of activity in the border strip after MI need to be explained in terms of physiological and morphological differences in cellular and circuit components. Neurons in the border strip regions are already known to be highly binocular (21). Finally, the muting of such activity patterns, over hours after MI, needs explanation in terms of mechanisms of cortical plasticity that possibly include activity-dependent regulation of inhibitory receptors and transmitters (22, 23).

Materials and Methods

Animals and Sample Preparation.

Nine macaque monkeys (Macaca fuscata, M. fascicularis, and M. mulatta) were studied. The details of the experimental groups are provided in Table S1. Four of the macaques were subjected to MI by TTX injection as follows: 10 μL of TTX (4.7 mM in saline) was slowly microinjected into the left eyeball while the microsyringe was kept in place at least for 20 min under anesthesia from a mixture of ketamine and medetomidine (refer to Table S1 for concentrations here and elsewhere). The pupil of the injected eye was dilated after the injection. Four of the macaques were monocularly deprived by eyelid suture. In these cases, the left eyelids were sutured and glued together under anesthesia. The operated animals were then immediately brought back to their normal cage and injected with atipamezole to antagonize the medetomidine. They awoke within 5–30 min (mostly within 10 min) after atipamezole treatment and were allowed normal behavior. Neither particular dark adaptation nor light stimulation for the induction of IEG expression was administered. Later, the animals were injected with ketamine (and/or medetomidine) again, and we defined the MI or MD time as the duration between atipamezole injection and the second ketamine (or medetomidine) injection. Finally, the monkeys were administered an overdose of pentobarbital and perfused intracardially with 4% paraformaldehyde in 0.1 M phosphate buffer (PB). The brain was then removed from the skull, postfixed for 3–6 h at room temperature in the same fixative, cut into several blocks, and cryoprotected in 30% sucrose in 0.1 M PB at 4° for more than a week. The block samples were frozen and stored at −80°. The frozen blocks were cut on a sliding microtome into 25–35-μm-thick sections. All the sections shown here were derived from the dorsolateral part of V1.

The protocols used in this study were approved by the Animal Research Committee of the National Institute for Basic Biology, National Institute for Physiological Sciences, and National Institute of Advanced Industrial Science and Technology, Japan; they were in accordance with the animal care guidelines of the U.S. National Institutes of Health.

ISH.

For colorimetric ISH, digoxigenin (DIG)-labeled antisense and sense riboprobes were prepared using a DIG-dUTP labeling kit (Roche Diagnostics). We used riboprobes for zif268 and c-fos as in a previous study (24). We prepared another c-fos probe from the cDNA library of monkeys by RT-PCR and conventional TA cloning techniques (target position is 970-1537 in XM_001098940, including primers) and used a mixture of the 2 c-fos probes to enhance the signals. The sense probes did not detect signals stronger than the background signal. ISH was carried out as described previously (15, 24). Images of the ISH sections were captured with an E800M microscope (Nikon) using a 3-CCD color digital camera, DXM1200M (Nikon) and were processed using Photoshop CS3 Extended (Adobe). The scale bars in the figures are corrected for shrinkage caused by ISH.

CO Staining.

To reveal CO enzymatic activity, free-floating sections were immersed into 4.5% sucrose/PB (pH 7.4). Sections were then reacted with 0.33 mg/mL cytochrome C type III (Sigma) and 0.54 mg/mL 3′, 3-diaminobenzidine in sucrose/PB at 37° overnight.

Statistical Analysis.

To measure the widths of border strips, dark columns, and pale columns in monkeys 1 and 2, we took photographs of layer 4C tangential ISH sections for zif268. As shown in Fig. 1D-b, a straight line was placed perpendicular to the ODCs, boundaries were determined, and the widths were measured. The widths were indicated as the mean ± SEM for each animal. Data were corrected from more than 40 places for each number in 5 sections from each animal. Because of the limited number of monkeys in the present study, we did not evaluate the observed increases and decreases in IEG expression statistically in comparison to baseline measures from binocular monkeys.