Precocious deposition of perineuronal nets on Parvalbumin inhibitory neurons transplanted into adult visual cortex

The end of the critical period for primary visual cortex (V1) coincides with the deposition of perineuronal nets (PNN) onto Parvalbumin (PV) inhibitory neurons. Recently, we found that transplantation of embryonic inhibitory neurons into adult V1 reinstates a new critical period. Here we used Wisteria Floribunda Agglutinin (WFA) staining to compare the deposition of PNNs onto neurons during normal development and following transplantation at equivalent cell ages. In accord with previous findings, PV and PNN expression increases from negligible levels at postnatal day 14 (P14) to mature levels by P70. In contrast to P14, PNNs are found on transplanted PV neurons by 21 days after transplantation and persist to 105 days after transplantation. This precocious deposition was specific to PV neurons and excluded transplanted neurons expressing Somatostatin. Notably, the onset of PV expression in transplanted inhibitory neurons follows the timing of PV expression in juvenile V1. Moreover, transplantation has no discernible effect on host PNNs. The precocious deposition of PNNs onto transplanted PV neurons suggests that PNN expression identified by WFA does not reflect neuronal maturity and may be an inaccurate marker for transplant-induced plasticity of cortical circuits.

The critical period for binocular vision is a time of heightened experience-dependent plasticity in primary visual cortex 1,2 . The maturation of GABAergic Parvalbumin-expressing (PV) inhibitory neurons in primary visual cortex has been linked to the onset of the critical period [3][4][5][6][7] . During postnatal development, perineuronal nets (PNNs) appear on PV inhibitory neurons and reach mature levels by the end of the critical period [8][9][10][11][12][13] . The disruption of PNNs and associated signaling in adult animals has been shown to restore visual cortical plasticity 9,10,14-18 . These findings suggest that the deposition of PNNs onto PV inhibitory neurons applies the brakes to critical period plasticity.
Recently, we and others have shown that the transplantation of embryonic inhibitory neurons from the medial ganglionic eminence opens a new critical period plasticity in both juvenile mice 19,20 as well as adult recipients 21,22 . Transplanted inhibitory neurons from the medial ganglionic eminence reactivate critical period plasticity whereas inhibitory cells from the caudal ganglionic eminence do not [21][22][23] . In addition to reactivating ocular dominance plasticity, inhibitory neuron transplantation reverses visual acuity deficits brought on by early life visual deprivation 21 .
Transplanted inhibitory neurons recapitulate several developmental programs tied to cellular age. Cell death in transplanted inhibitory neurons follows a chronological developmental program 24 . Similarly, orientation selectivity in transplanted PV inhibitory neurons matures at the same cell age as for normally developing inhibitory neurons 25 . Moreover, transplantation creates a new critical period that occurs at a time when the donor animal would have had its critical period 21 . However, some developmental programs such as the maturation of membrane excitability have been found to mature more rapidly in transplanted cells than their host counterparts 26 .
Inhibitory neuron transplantation gives us a unique opportunity to assess the role of PNN deposition in closing the critical period for visual cortex. In this study, we examined the deposition of PNNs onto transplanted inhibitory neurons at three intervals relative to the transplant-induced critical period: before (21 days after

Transplanted MGE cells disperse in adult visual cortex. Transplantation of inhibitory neurons into
adult visual cortex from the medial ganglionic eminence (MGE) introduces a new window of critical period plasticity [19][20][21][22] . If the deposition of PNNs terminate critical period plasticity, then more transplanted inhibitory neurons should carry perineuronal nets after the transplant-induced critical period 21 has subsided, between 35 and 105 days after transplantation (DAT; Fig. 2A). To test this prediction, we harvested embryonic day 13.5 (E13.5) tissue from the MGE, a major source of cortical inhibitory neurons 33,34 . We then transplanted MGE precursors into sites bordering the binocular primary visual cortex using a physiological mapping procedure 21 . Transplanted inhibitory neurons were identified by fluorescence and appear to disperse throughout all layers of the adult primary visual cortex as previously reported 21 . Precocious deposition of PNNs on transplanted inhibitory neurons. We quantified the percentage of transplanted inhibitory neurons surrounded by PNNs before (21 DAT), during (35 DAT), and after the transplant-induced critical period (105 DAT). The youngest host age during which transplantation co-labeling was assessed is P90, well after the normal critical period ( Fig. 2A). Surprisingly, by 21 DAT we found that deposition of PNNs on transplanted neurons is already 33.87% (Fig. 2C), comparable to those observed in non-transplanted adults (34.1%; Fig. 1C). More importantly, the percent of transplanted inhibitory neurons with PNNs is similar before (33.87%), during (34.27%), and after (35.27%) the transplant induced critical period (Fig. 2C). Moreover, the density of transplanted inhibitory neurons could not explain the percentage of transplanted cells carrying PNNs. These results reveal the precocious development of PNNs on transplanted inhibitory neurons.

PV expression in transplanted cells is largely independent of host cell age.
Over the course of the normal critical period, we observed that PV expression more than doubles (Fig. 1D). To determine how transplanted inhibitory neurons develop PV expression we examined PV and PNN co-expression 21, 35, and 105 days after transplantation (Fig. 3). Like in normal development, PV expression is significantly lower in the transplanted inhibitory (VGAT) cell population before the induced critical period at 21 DAT (20.84%, Fig. 3D). Contrary to normal development, the number of transplanted cells expressing PV reaches adult levels by 35 DAT (48.71%) and remains stable through 105 DAT (48.62%). Nonetheless, the percentage of transplanted PV neurons at 21 DAT and 35 DAT are comparable with values observed at P14 and P28, respectively 11,28 . We did not find this increase in PV expression to be correlated to the density of transplanted inhibitory (VGAT) cells. Together, our findings suggest that PV maturation is largely cell-intrinsic and independent of host age.
Transplanted cells acquiring PNNs are fated to express PV. In our developmental study, we observed a strong specificity of PNNs for PV inhibitory neurons (Fig. 1). To assess whether this specificity was present in transplanted neurons, we quantified the percent of transplanted PV neurons that carried PNNs at 21, 35, and 105 DAT (Fig. 3E). Surprisingly, the percentage of transplanted PV neurons bound to PNNs had already reached adult levels by 21 DAT (80.61%; Fig. 3E). This was in remarkable contrast to the almost non-existent percentage of PV inhibitory neurons bound to PNNs at P14 observed by our group (1.63%) and others 9,11 . The percentage of   (Fig. 3). We found that the fraction host PV cells bound to PNNs in the transplanted hemisphere remains at adult levels (21 DAT = 77.7%, 35 DAT = 68.97%, 105 DAT = 74.5%) and is similar to the non-transplanted hemisphere (72.54%) (Fig. 3). These results suggest MGE transplantation does not degrade host PNNs.
It was possible that the precocious PNNs we observed were ill-formed and immature. We investigated this possibility by first examining the intensity of WFA for PNNs in normal development (Fig. 4A). It has been observed that the intensity of WFA staining increases as PNNs mature 36 . We quantified the WFA fluorescence intensity of the PNNs surrounding the cell body of PV inhibitory neurons and normalized it to the average intensity value of an adjacent unstained area. As expected, we found that WFA intensity around PV neurons significantly increases across postnatal age from P14 (2.08) to P28 (2.54) and to P74 (3.11), increasing by more than 50% over the course of the juvenile critical period.
Next, we tested whether the PNNs observed on transplanted PV inhibitory neurons were mature (Fig. 4B). We found that WFA staining intensity surrounding transplanted PV neurons remains consistent across the cell ages tested (21 DAT = 2.98, 35 DAT = 2.69, 105 DAT = 2.76) (Fig. 5B). We observed comparable levels of WFA staining in host PV neurons in the transplanted hemisphere regardless of the time after transplantation (21 DAT = 2.76, 35 DAT = 2.96, 105 DAT = 2.91) (Fig. 4B). The consistency of WFA staining intensity across groups further supports our observation that PNN deposition does not depend upon the age of the transplanted inhibitory cells.

Transplanted Somatostatin cells do not acquire PNNs. It has been shown that transplanted
Somatostatin (SOM) cells alone reactivate plasticity in visual cortex 20 . Therefore, it may be that PNNs are deposited onto transplanted SOM cells and terminate the transplant-induced period of cortical plasticity. To test this hypothesis, we stained transplanted tissue 21 and 105 DAT for SOM and PNNs (Fig. 5A). During normal development, the number of cells expressing SOM reaches adult levels by P14 28 . Similarly, we found that transplanted cells labeled with SOM 21 DAT (30.23%) is not different from 105 DAT (20.23%) (Fig. 5B). More importantly, we did not find SOM cells surrounded by PNNs in either transplanted cells (21 DAT = 0%, 10 DAT = 0%) or in host cells of the non-transplanted hemisphere (1.22%) (Fig. 5C and Table 1).

Discussion
Recently we found that the transplantation of embryonic inhibitory neurons creates a new period of heightened plasticity in adult visual cortex 21 . The end of critical period plasticity has been linked to the deposition of perineuronal nets (PNNs) onto Parvalbumin-expressing (PV) neurons [8][9][10][11][12][13]37 . Therefore, we might predict that PNNs are deposited onto transplanted PV inhibitory neurons when transplant-induced plasticity subsides, more than 35 days after transplantation (35 DAT) 19,21 . Alternatively, transplantation could reactivate plasticity by degrading mature PNNs on host PV inhibitory neurons 35 . In this study, we discovered that PV inhibitory neurons transplanted into adult visual cortex acquire PNNs by 21 DAT, much sooner than expected. The level of PNN expression that we found on transplanted inhibitory neurons is comparable to adult levels and remains constant up to 105 DAT.
We also found that transplantation did not disturb the presence of PNNs on host neurons at any of the timepoints studied. A recent study found that inhibitory neurons transplanted into the adult basolateral amygdala induces a new critical period for fear erasure 35 . In contrast to our study of transplantation into adult visual cortex, Yang et al. show that transplantation reduces PNNs surrounding host neurons in the amygdala. The apparent discrepancy suggests that the mechanisms of critical period reactivation may be specific to the host circuit into which these cells integrate.
The maturation of PV inhibitory neurons plays a key role in critical period plasticity 2,3 . It is well known that PV expression in the primary visual cortex of mice increases dramatically between P14 and P28 [27][28][29][30] . In contrast to the precocious deposition of PNNs on transplanted inhibitory neurons, we find that PV expression in transplanted inhibitory neurons more than doubles between 21 and 35 DAT, suggesting that PV expression is determined by cell age. The developmental timing of PV expression in transplanted inhibitory neurons agrees with other findings of cell-intrinsic programs in these cells such as developmental apoptosis 24 , orientation selectivity 25 , and reactivation of cortical plasticity 21 .
Previous studies have shown that PNNs preferentially bind to PV-expressing inhibitory neurons 5,11,31,32 . In our study, we find that PNNs associate with transplanted PV inhibitory neurons and not SOM neurons, respecting this close association. Nonetheless, the deposition reaches adult levels much earlier than expected suggesting that PNN deposition is not a cell-intrinsic process.
PNNs surrounding PV neurons may gate critical period plasticity by capturing key signaling factors 15,38 . The transcription factor Otx2 has been identified to associate with PNNs and regulate critical period plasticity in a positive feedback loop 12,15,29,39 . Perhaps transplanted PV inhibitory neurons acquire PNNs more rapidly than To visualize PNNs, we used Wisteria floribunda agglutinin (WFA) staining. The maturation of PNNs has been associated with the intensity of WFA staining 36 . We found that the intensity of WFA staining on PV neurons increased from P14 to P28. In contrast, the intensity of WFA staining on transplanted PV neurons reached adult levels by 21 DAT, suggesting the precocious maturation of PNNs on these cells. Furthermore, the intensity of WFA staining on host PV neurons in the transplanted and non-transplanted hemisphere suggests that transplantation did not affect the maturity of PNNs in the host circuitry.
It is possible that the maturational state of PNNs are not adequately captured in our study by WFA staining. PNNs are composed of a multitude of proteoglycans including aggrecan and tenascin-R 38,41,42 . Although aggrecan and tenascin-R staining have similar developmental profiles as WFA staining 11 , other PNN components may better reflect the maturity of PNNs. Future studies on specific changes to the configuration of proteoglycans in PNNs surrounding transplanted neurons may reconcile our findings with the normal development of PNNs.
In this study, we find that for transplant-induced plasticity, PNN expression is an inaccurate marker for the plasticity of cortical circuits. Similar to normal juvenile plasticity, transplantation-induced plasticity is limited to a brief critical period, primarily affects deprived-eye responses, and shapes the spatial acuity of cortical responses [19][20][21][22] . Nonetheless, the mechanisms regulating transplant-induced plasticity may be distinct from the juvenile critical period. Future studies on how transplantation alters the cortical circuit may reveal plasticity mechanisms independent of PNNs 43-46 .

Methods
All experiments were approved by the Institutional Animal Care and Use Committee at the University of California, Irvine (2011-2994) and were conducted according to the NIH guide for the Care and Use of Animals.
Animals. Embryonic donor tissue and mice used to characterize postnatal development were produced by crossing the Cre-dependent tdTomato reporter line (Ai14, JAX 007914) with mice that express Cre recombinase Intrinsic Signal Imaging. As described previously 21 , intrinsic signal optical imaging was used to find the binocular visual cortex in host P69-78 day-old mice. Briefly, anesthesia was induced with 2.5% isofluorane and the dose was reduced 0.8% for intrinsic signal mapping. Mice were given intraperitoneal chlorprothixene injections. Mice were presented with a spatiotemporal noise stimulus that swept from −18° to 36° visual field elevation with a periodicity of 0.1 Hz. Fourier analysis of the red light reflection from brain revealed a retinotopic map of binocular visual cortex.
Retinotopic Map-Guided Cell Transplantation. The retinotopic map obtained using intrinsic signal imaging was used to guide skull incisions medial and lateral to the binocular visual cortex. Donor cells were injected at a 45° angle to the cortical surface and advanced approximately 700 µm into the cortex. Three 15-20 nL injections were made in each of the two slits. After transplantation, the scalp was sutured, anesthesia was terminated, and the animal was placed on a warm heating pad for recovery.

Histological Preparation and Cell
Counting. Animals were transcardially perfused (4% in 1 × PBS) either 21 days (n = 4), 35 days (n = 3), or 105 days after transplantation (n = 4). Brains were removed, post-fixed, and cryopreserved in 30% sucrose. A freezing sliding microtome was used to slice the brains into 50 μm thick   coronal sections (Microm HM450). The free-floating slices of tissue were stained and blocked for one hour at room temperature with 0.5% Triton-X (Sigma T8787) and 10% BSA (Fisher BP1600-100) in 1X PBS. Slices were incubated overnight at 4 °C with the primary antibodies rabbit-anti-RFP (tdTomato 1:1000, Abcam ab62341) and mouse-anti-PV (1:1000, Sigma P3088). The sections were washed three times in 1X PBS for five minutes. They were then incubated for two hours in 594 goat-anti-rabbit IgG (1:1000, Invitrogen) for VGAT cells, 647 goat-anti-mouse IgG1 (1:1000, Invitrogen) for PV cells, and Fluorescein labeled Wisteria Floribunda Agglutinin (WFA), Vector Labs) for PNN + cells. The stained tissue was mounted on glass slides with Fluoroshield with DAPI. The sections were imaged with a confocal microscope (Leica SP8, 63X objective, N.A 1.4). ImageJ was used to count the cells expressing VGAT, PV, and WFA in brain sections. The investigator was blinded to groups for PNN identification and counting. A PNN was positively identified if it appeared to robustly surround at least three-fourths of a cell body.
ImageJ was also used to assess the intensity of PNN staining (Image J ROI manager, version 1.51j, NIH). 5-pixel thick regions of interest were drawn on a randomly selected subset of identified PNNs. PNN intensity was normalized by dividing the average intensity value of the net by the average intensity value of the background. Background regions of interest were defined using the PNN's region of interest applied to the dimmest neighboring background.
Statistical Analyses. Kruskal-Wallis ANOVA was used as an omnibus test for significant differences among groups. A Mann-Whitney test with Bonferroni correction was used for pairwise comparisons. Statistical tests were performed using Prism version 7.02 (Graphpad).
Materials, data, and associated protocols are available upon request from the corresponding author.