Chondroitin sulfate expression around spinal motoneurons during postnatal development in rats

Perineuronal nets are extracellular matrix structures that surround neuronal cell bodies and their proximal dendrites in the central nervous system. Chondroitin sulfate proteoglycans, which contain chondroitin sulfates (CSs) are major components of perineuronal nets. CSs are considered to have inhibitory roles in neural plasticity, although the effects differ according to their sulfation pattern. In the present study, we investigated the expression of the CS subtypes CS-A and CS-C surrounding spinal motoneurons in different postnatal periods to explore the potential influence of altered CS sulfation patterns on spinal development. CS-A-positive structures were observed around motoneurons in the cervical, thoracic, and lumbar segments as early as postnatal day (P) 5. Most motoneurons were covered with CS-A-positive structures during the first 2 postnatal weeks. The percentage of motoneurons covered with CS-A-positive structures decreased after P20, becoming lower than 70% in the cervical, and lumber segments after P35. CS-C-positive structures were occasionally observed around motoneurons during the first 2 postnatal weeks. The percentage of motoneurons covered with CS-C-positive structures increased after P20, becoming significantly higher after P25 than before P20. The expression pattern of Wisteria Floribunda agglutinin-positive structures around motoneurons was similar to that of the CS-C-positive structures. The present findings revealed that CS-A and CS-C are differentially expressed in the extracellular matrix surrounding motoneurons. The altered sulfation pattern with increased CS-C expression is associated with the maturation of perineuronal nets and might lead to changes in the motoneuron plasticity.


Introduction
Perineuronal nets (PNNs) are unique lattice-like extracellular matrix (ECM) structures that surround cell bodies and proximal dendrites of some neurons in the central nervous system. PNNs are formed relatively late in the developmental process as mature synaptic connections are established. PNNs are implicated to have a major role in synaptic stabilization and maturation, i.e., restricting the formation of new connections with advancing axons (Berardi et al., 2004;Karetko and Skangiel-Kramska, 2009;Dzyubenko et al., 2016). The formation of PNNs is considered to coincide with the closure of the critical period for plasticity (Wang and Fawcett, 2012).
A major component of PNNs is chondroitin sulfate proteoglycans (CSPGs), among which the sulfated glycosaminoglycan (GAG) side chain is considered to contribute to neural plasticity and development (Caterson, 2012;Dyck and Karimi-Abdolrezaee, 2015;Soleman et al., 2013). There are 5 sulfation patterns in the repeating disaccharide unit of GAG, i.e., chondroitin sulfate (CS)-O, CS-A, CS-C, CS-D, and CS-E. In the spinal cord, CS-A and CS-C, also termed chondroitin-4-sulfate and chondroitin-6 sulfate, respectively, are expressed in the PNNs (Mikami and Kitagawa, 2013). CSs are considered to have inhibitory roles in neural plasticity, although the effects differ according to their sulfation pattern (Properzi et al., 2005;Wang et al., 2008;Butterfield et al., 2010;Lin et al., 2011;Swarup et al., 2013). In previous studies, PNNs were usually visualized by immunohistochemistry using antibodies against CSPG core proteins, tenascin-R or CS-56, a pan marker for CSs, as well as by staining with the plant lectin Wisteria Floribunda agglutinin (WFA; Irvine and Kwok, 2018;Miyata et al., 2012;Sigal et al., 2019). Few studies, however, have examined the expression of CS by classifying the subtypes, such as CS-A and CS-C.
A previous study using immunohistochemistry for CSPG core proteins showed that the formation of PNNs in the rat spinal cord begins in the second postnatal week (Galtrey et al., 2008). In rats, the corticospinal tract (CST) develops from late embryonic days to the first 2 weeks after birth. The CST axons reach the white matter at the cervical level at birth (postnatal day 0 [P0]), the thoracic level at P3-P5, and the lumbar level at P7, and then enter the gray matter 2-3 days later to complete functional synapse formation after P16 (Donatelle, 1977;Schreyer and Jones, 1982;Gribnau et al., 1986). PNN formation, therefore, is coincident with the formation of synaptic connections between CST axons and spinal neurons. Although disaccharide units with different sulfation patterns might also have important roles in synaptic formation in the spinal cord (Dyck and Karimi-Abdolrezaee, 2015), the expression patterns of CS-A and CS-C in PNNs during the postnatal period are not well elucidated.
In the present study, therefore, to explore the potential significance of the differences in CS sulfation patterns in the PNNs during postnatal development, we investigated the percentage of motoneurons that were surrounded by CS-A-and CS-C-positive structures, as well as WFApositive structures in various spinal segments.

Results
The postnatal development of WFA-, CS-A− , and CS-C-positive structures surrounding motoneurons in the cervical (C7), thoracic (T8), and lumbar (L5) segments of rat spinal cord was analyzed in this study. The spinal regions that were analyzed are shown in Fig. 1. Higher magnification images are shown in Figs. 2-4.
CS-A immunoreactivity was also examined using a different monoclonal antibody (Clone LY111) in the cervical segment sections. Positive structures surrounding ChAT-positive neurons were observed at different postnatal days (P10, 84.05 ± 3.32%; P20, 73.26 ± 2.36%; P35, 70.95 ± 3.37%; P40, 77.28 ± 3.40), as shown in Suppl. 1. Blue bars indicate the percentage of motoneurons with greater than 80% WFA-positive surrounding structures, and red bars indicate between 20% and 80% WFA-positive surrounding structures. Data are expressed as the mean ± SEM (>80% and >20%). Statistical analyses of percentage of motoneurons with WFA-positive structures were performed using one-way ANOVAs, followed by Tukey-Kramer test. Significant differences from P5, P10, and P15 are expressed as † †: P < 0.01. Other significant differences are expressed as *: P < 0.05 and **: P < 0.01. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
The WFA-staining combined with double-labeling immunohistochemistry for CS-C and ChAT showed that many ChAT-positive neurons were covered with both WFA-positive surrounding structures and CS-C-positive structures in the cervical segment at P30. The percentages of ChAT-positive neurons with both WFA-positive and CS-C-positive  (Fig. 5).

Discussion
The present study revealed differential expression patterns of CS-A and CS-C in the ECM structures surrounding motoneurons in the spinal cord during postnatal development in rats. CS-A is highly expressed in structures surrounding the motoneurons in all spinal segments at early stages from P5 to P15, with expression decreasing after P35. Unlike CS-A, CS-C-positive surrounding structures were rarely observed until P15, but increased thereafter. Although the expression patterns seemed similar among the 3 spinal segments, CS-C-positive surrounding structures were observed earlier in the cervical segment than in the thoracic and lumbar segments from P5 to P15.
In the present study, we used the monoclonal antibody clone 2H6 to detect CS-A. Although this antibody effectively binds CS-A (Shimazaki et al., 2005), it recognizes multiple sequences containing CS-A units in CS chains (Deepa et al., 2007). Therefore, we compared another antibody to CS-A (LY111) in some sections and observed similar expression patterns of positive structures surrounding motoneurons. We used the monoclonal antibody clone 3B3 to detect CS-C. This antibody was raised against CS-C neoepitopes on CS-GAG chains that were pre-digested with either chondroitinase ABC or chondroitinase ACII, and efficiently recognizes CS-C stubs on the CS-GAG chains (Caterson, 2012). Therefore, before immunostaining using this antibody, we pretreated the tissue with chondroitinase ABC.
WFA lectin staining is a conventional marker for PNNs, and the staining intensity corresponds to the maturation of PNNs (Härtig et al., 1992;Slaker et al., 2016). The present study revealed that WFA-positive structures surrounding motoneurons rapidly increased after P15, and were preserved after P25, suggesting that the formation of PNNs around Blue bars indicate the percentage of motoneurons with greater than 80% CS-C-positive surrounding structures surroundings, and red bars indicate 20%-80% CS-C-positive surrounding structures. Data are expressed as the mean ± SEM (>80% and >20%). Statistical analyses were performed using oneway ANOVAs, followed by Tukey-Kramer test. Significant differences from P5, P10, P15, and P20 are expressed as † †: P < 0.01. Other significant differences are expressed as *: P < 0.05 and **: P < 0.01. motoneurons is most active between P15 and P25. These periods occur shortly after complementation of functional connections between CST axons and spinal neurons. Thus, the timing of PNN maturation appears to coincide with changes in the motoneuron plasticity, leading to closure of the critical period of synaptic formation with CST. The present study also revealed some differences in the expression patterns among the 3 spinal segments from P5 to P15. The WFA-positive surrounding structures in the cervical and lumbar segments gradually increased, but there were no WFA-positive surrounding structures in the thoracic segment.
The CS-C expression pattern appears to be identical to the WFA staining pattern from P5 to P40. In fact, approximately 70% of motoneurons with WFA-positive surrounding structures were also positive for CS-C, and approximately 85% motoneurons with CS-C-positive surrounding structures were also positive for WFA. These findings suggest that that CS-C expression coincides with the maturation of PNNs surrounding motoneurons. The CS-A expression pattern, on the other hand, was consistent with the CS-56 antibody pattern in a previous report (Wang et al., 2017), showing that CSPGs are highly expressed in the spinal cord at late embryonic and neonatal stages in mice. These findings might indicate that the CS components of PNNs change drastically from CS-A dominant to CS-C dominant in the spinal cord during postnatal development.
The functional significance of CS-A and CS-C is controversial. Both positive and negative effects on axonal outgrowth are observed with CS-A and CS-C (Properzi et al., 2005;Wang et al., 2008;Butterfield et al., 2010;Lin et al., 2011;Swarup et al., 2013). A previous study in goldfish showed that the descending axons regenerate to preferentially terminate on neurons not covered with CS-C-positive PNNs after spinal cord injury, suggesting that CS-C in the PNNs inhibits new contact formation between descending axons and spinal neurons (Takeda et al., 2018). The previous study in goldfish also showed that CS-A in the ECM or PNNs does not inhibit for regenerating axons after spinal cord injury (Takeda et al., 2017(Takeda et al., , 2018. The high CS-A expression in the ECM during the first postnatal 2 weeks, therefore, might be involved in the development of motoneuron connectivity. These results appear to be consistent with the present findings that expression of CS-C, but not CS-A might be associated with the maturation of PNNs, leading to closure of the critical period of plasticity. Further experimental studies are necessary to determine whether the altered CS sulfation pattern around motoneurons in the postnatal period is associated with the maturation of PNNs, leading to changes in motoneuron plasticity, and with the development of motor function for different spinal segments.

Materials
Wistar rats from P5 to P40 were purchased from Japan SLC (Hamamatsu, Japan) and used in the present study. All experimental procedures were approved by the Institutional Animal Care and Use Committee of the Animal Research Center of Yokohama City University.

Immunohistochemistry
The slides were dried for 1 h and washed 3 times in phosphatebuffered saline (PBS) for 5 min each. The sections were then incubated in PBS including 0.5% Tween 20 (PBST) for 30 min. For CS-C staining, the sections were pre-incubated with chondroitinase ABC (0.2 U/mL, Millipore Sigma, St. Louis, MO, USA) in 50 mM Tris-HCl (pH 7.5) containing 50 mM acetic acid at 37 • C for 1 h. After the blocking procedure with Block Ace (5%, UK-B80, DS Pharma Biomedical Co., Ltd), the sections were incubated in a moist chamber overnight at 4 • C with primary antibodies as follows: (1)  Antibody specificity was verified by incubation with 0.5% normal mouse serum (Jackson ImmunoResearch Laboratories) or 0.5% normal goat serum (Jackson ImmunoResearch Laboratories) instead of the primary antibodies.

Image acquisition and analysis
Sections including C7, T8, and L5 were digitally photographed using a Keyence BIOREVO All-in-One Fluorescence Microscope (BZ9000, Keyence, Osaka, Japan) and transferred to Adobe Photoshop CS (Adobe, San Jose, CA, USA) to generate the figures. Contrast and brightness were adjusted with Adobe Photoshop Software. Three sections from each spinal segment were observed. ChAT-positive motoneurons in the lateral part of the ventral horn at C7 and L5, and in the ventral horn at T8 level were analyzed.
We analyzed WFA-, CS-A-, and CS-C-positive ECM structures surrounding ChAT-positive motoneurons. We identified ChAT-positive neurons for which 20% or more of the circumference was surrounded by CS-A-positive structures as positive for CS-A. We further divided ChAT-positive neurons into 3 groups: (1) those for which more than 80% of the circumference was surrounded by CS-A-positive structures, (2) those for which 20% to 80% of the circumference was surrounded by CS-A-positive structures, and (3) those for which less than 20% of the circumference was surrounded by CS-A-positive structures. Similar criteria were applied to the CS-C-positive structures, and WFA-positive structures. We counted the number of ChAT-positive motoneurons for each group with the researchers blinded to group conditions. We then calculated the percentage of ChAT-positive neurons surrounded by WFA-positive, CS-A-positive, or CS-C-positive structures among all ChAT-positive neurons for each of the 3 groups. This process repeated in all rats. Statistical analyses were performed using SPSS. One-way ANOVAs followed by Tukey-Kramer test were used to analyze differences.