Maternal thyroid hormones (THs) are essential for normal offspring's neurodevelopment even after onset of fetal thyroid function. This is particularly relevant for preterm children who are deprived of maternal THs following birth, are at risk of suffering hypothyroxinemia, and develop attention-deficit/hyperactivity disorder. Expression of neocortical Ca2+/calmodulin kinase IV (Camk4), a genomic target of thyroid hormone, and nuclear receptor-related 1 protein (Nurr1), a postnatal marker of cortical subplate (SP) cells, was studied in euthyroid fetuses and in pups born to dams thyroidectomized in late gestation (LMH group, a model of prematurity), and compared with control and developmentally hypothyroid pups (C and MMI groups, respectively). In LMH pups, the extinction of heavy Camk4 expression in an SP was 1–2 days delayed postnatally compared with C pups. The heavy Camk4 and Nurr1 expression in the SP was prolonged in MMI pups, whereas heavy Camk4 and Nurr1 expression in layer VIb remains at P60. The abnormal expression of Camk4 in the cortical SP and in layer VIb might cause altered cortical connectivity affecting neocortical function.
Extremely low-birth-weight (ELBW) preterm children suffer cognitive and motor maturation delay (O'Callaghan et al. 1995; Samara et al. 2008; Volpe 2009) and have an increased risk of suffering schizophrenia (Jones et al. 1998; Smith et al. 2001). Neurological disorders of ELBW preterm children are caused by several etiological factors that include transient hypothyroxinemia of prematurity (THOP), because ELBW preterm children are hypothyroxinemic during an important period of brain development (Morreale de Escobar and Ares 1998; van Wassenaer and Kok 2004). THOP is due to the lack of maternal thyroid hormones (THs), owing to the importance of maternal thyroxine (T4) as a substrate for the generation of local cerebral 3–5-3′-triiodothyronine (T3; Kester et al. 2004; Morreale de Escobar et al. 2008). In preterm children, cerebral T3 concentrations are not comparable with those of children born at term, until normalization of their thyroid function which may take up to several weeks after birth depending on the birthdate. THOP has been associated with several neurological disorders, including poor cognitive development, attention-deficit/hyperactivity disorder (Johnson 2007; Simic et al. 2009), and disabling cerebral palsy (den Ouden et al. 1996; Reuss et al. 1996). Epidemiological studies show that preterm births occur in 12% of all pregnancies, and that THOP can affect 35–50% of preterm neonates, who are at high risk of suffering neurodevelopmental alterations (Rovet and Simic 2008; Williams and Hume 2008). These findings stress the importance of maternal TH for the fetus during the second half of human pregnancy (Ares et al. 1997; Williams et al. 2005).
We have used an experimental model to study the effect of maternal TH during gestation from the beginning of fetal thyroid function until term, in which fetuses lack maternal TH during the last 5–6 days of gestation (LMH group; Berbel et al. 2010). Previous results showed that 40-day-old LMH pups had altered cortical migration resulting in heterotopic neurons in the subcortical white matter (wm) and abnormal cortical lamination (Berbel et al. 2010). LHM pups had lower step-down latencies than control (C) pups, indicating altered memory consolidation. We also observed a dramatic 40–50% reduction in the ratios of phosphorylated cAMP-responsive element-binding protein (pCREB) to p-active transcription factor-1 (pATF1), pCREB to CREB, phosphorylated extracellular signal-regulated kinase 1 (pERK1) to ERK2, and pERK2 to ERK2 in the hippocampus on LMH pups (Berbel et al. 2010). In agreement with this, in a recent study of gene expression in LMH pups at E21, using microarrays, it was observed that a large fraction of the differentially expressed genes are downstream targets of Ca2+/calmodulin kinase IV (Camk4; Morte et al. 2010). This study revealed a prominent role of the Camk4–CREB pathway as a target of TH in the fetal cortex (Fig. 1). Camk4 expression was previously also shown to be induced by T3 in rat fetal telencephalic cultures (Krebs et al. 1996) and in mouse embryonic stem cells (Liu and Brent 2002). The spatial distribution of Camk4 expressing cells in C rats is poorly known. A study by Wang et al. (2001) described the expression of the Camk4 gene during murine embryogenesis in different tissues, organs, and systems, including the central nervous system (CNS). However, data concerning Camk4 expression in the neocortex are scarce.
The influence of the subplate (SP) in the outcome of plasticity and functional maturation of the neocortex has been reported in the past years (Friauf et al. 1990; Kanold and Shatz 2006; Kanold et al. 2009). SP and wm alterations are particularly relevant because one of the most frequent lesions of encephalopathy in premature infants, especially in those born with an ELBW, is periventricular leukomalacia (Volpe 2009; Kinney et al. 2012), and to some extent alterations of the SP have also been related to the pathogenesis of brain developmental disorders, including schizophrenia (Eastwood and Harrison 2003). Nurr1 (NR4a2) is a member of the NR4a family of orphan nuclear receptors, which also include Nurr77 (NR4a1) and Nor1 (NR4a3). Although our interest in Nurr1 in the present work is limited to its use as a marker of the SP, it is worth to mention that altered expression of Nurr1 occurs in schizophrenia (Xing et al. 2006) and the interrelations between Camk4, NR4a, and T3 receptor (TR) in the regulation of transcription. Camk4 potentiates the transcriptional activity of the TR, which might be due to direct phosphorylation of coactivators or by changing the equilibrium between the coactivators and the silencing mediator for retinoid and thyroid hormone receptors (SMRT; Kuno-Murata et al. 2000; McKenzie et al. 2005). The latter mechanism is also involved in the regulation of NR4a transcriptional activity by Camk4. For Nurr77, Camk4 shifts the equilibrium from the repressed state to the active state by promoting the cytoplasmic translocation of SMRT and the recruitment of the activating signal cointegrator-2 coactivator (Sohn et al. 2001).
Our aim was to study the expression of Camk4 during cerebral cortex development and maturation in LMH fetuses and at postnatal ages, focusing our attention in the organization of infragranular layers and the subcortical wm. The results were compared with C and developmentally hypothyroid (MMI) pups. The present study provides novel data on neocortex Camk4 expression during the fetal and postnatal development of C, LMH, and MMI rats and also helps to identify structural changes that might underlie the neurobehavioral alterations of preterm children as a consequence of premature deprivation of maternal TH.
Materials and Methods
Animals and Treatments
Wistar rats were housed in temperature-controlled (22–24°C) animal quarters, with automatic light and dark cycles of 14 and 10 h, respectively. All surgical interventions were under anesthesia by inhalation of 1.5–2% isoflurane (Laboratorios Dr Esteve, S.A., Barcelona, Spain) in O2 (0.9 l O2/min). Care of the animals, drugs administration, and surgical procedures were performed under veterinarian control according to the European Union guidelines after the approval of the Institutional Ethics Committees.
Young adult females, weighing approximately 250–300 g, were mated at E0. Pregnant rats were surgically thyroidectomized at E16 as previously described (Berbel et al. 2010). From the day of thyroidectomy until delivery at P0 (E22), they were infused with rat 1–84 parathormone (4 µg/100 g body weight/day; H3086; Bachem GmbH, Weil am Rhein, Germany) and rat calcitonin (1 µg/100 g body weight/day, H3072; Bachem GmbH) diluted in 0.1M acetate buffer (pH 4.0). Both hormones were simultaneously infused via osmotic mini-pumps (ALZET, model 2001, Alza Corporation, Mountain View, CA, USA delivery ratio of 1 µL/h per day) placed under the dorsal skin. In addition, rats were supplemented by adding to the drinking water with 0.16% Ibercal-D® (Merck S.L., Barcelona, Spain) in the drinking water, which resulted in a final concentration of 9.5 IU Vitamin D3 and 11.8 mg calcium pidolate per 100 mL drinking water (Fig. 2). Immediately after birth (P0), LMH pups were transferred to normal rats for lactation and nursing. These normal rats were mated on the same day as the experimental ones and on the day of delivery, their pups were sacrificed just before the transfer of the experimental pups (Fig. 2).
Two additional MMI and C groups were also studied. An MMI group consisted of fetuses and pups born to rats treated with 0.02% methimazole and 1% KClO4 in the drinking water from E10 until the day of sacrifice (a complete list of the number of pups and litters used per experimental groups at the different ages is indicated in Supplementary Table 1). The C group consisted of fetuses and pups born to sham-operated normal rats. Fetuses were extracted from C, LMH, and MMI rats by cesarean section at the indicated embryonic ages (Fig. 2 and Supplementary Table 1).
Determination of Circulating T3 and T4 Concentrations
Heparinized blood from all dams (∼6 mL) was obtained under ether anesthesia from the cardiac ventricle on the day of delivery and from their pups on P60. The plasma was spun off and kept at –20°C for the determination of circulating T4 and T3 concentrations using very sensitive and highly specific radioimmunoassays (Morreale de Escobar et al. 1985). Levels of circulating TH hormones (Table 1) show that both LMH and MMI dams at P0 were hypothyroid, with significantly low levels of circulating TH compared with C dams. In contrast, LMH pups had similar levels of circulating TH compared with C pups, while MMI pups were hypothyroid.
|Dams at P0||Pups at P40|
|T4 (µg/dL)||T3 (ng/dL)||T4 (µg/dL)||T3 (ng/dL)|
|C||4.04 ± 0.95||56.2 ± 13.5||3.83 ± 1.02||45.1 ± 8.1|
|LMH||0.10 ± 0.02a||6.8 ± 0.8a||4.26 ± 1.63||41.3 ± 7.4|
|MMI||0.07 ± 0.04a||4.8 ± 1.0a||0.19 ± 0.10b||6.6 ± 2.2b|
|Dams at P0||Pups at P40|
|T4 (µg/dL)||T3 (ng/dL)||T4 (µg/dL)||T3 (ng/dL)|
|C||4.04 ± 0.95||56.2 ± 13.5||3.83 ± 1.02||45.1 ± 8.1|
|LMH||0.10 ± 0.02a||6.8 ± 0.8a||4.26 ± 1.63||41.3 ± 7.4|
|MMI||0.07 ± 0.04a||4.8 ± 1.0a||0.19 ± 0.10b||6.6 ± 2.2b|
Results are mean values ± SD.
aP < 0.001 for LMH and MMI compared with C.
bP < 0.001 for MMI compared with LMH and C.
Conventional Histology and Immunohistochemistry
Fetuses (from E17 to E21; Supplementary Table 1) and pups (from P0 to P60; Supplementary Table 1) were weighed, anesthetized, and perfused with 50 mL of saline, followed by 200 mL of 4% paraformaldehyde and 0.002% CaCl2 in 10 mM phosphate buffer (PB; pH 7.3–7.4). The brains were postfixed by immersion in perfusion fixative at room temperature for 4 h and stored at 4°C in PB containing 0.5% sodium azide. Parallel series of coronal 100-µm thick sections (embryos and pups under P5 were cut at 50 µm) containing the parietal cortex were cut with Vibratome and collected in PB-saline (PBS). Series 1 was immunostained with mouse antimature neurons neuronal nuclei (NeuN) monoclonal antibody (mAb; 1:400; Millipore, Temecula, CA, USA) and horse antimouse biotinylated polyclonal antibody (Ab; 1:150; Vector Laboratories, Inc., Burlingame, CA, USA). Series 2 was immunostained with rabbit anti-Camk4 Ab (1:2000; Sigma-Aldrich, Inc., St. Louis, MO, USA) and biotinilated goat antirabbit Ab (1:150, Vector Laboratories). Then, in both cases, with Vectastain ABC kit (1:200, Vector Laboratories) and 0.05% 3,3′-diaminobenzidine (Sigma-Aldrich Co.), mounted on gelatinized slides, air dried during 24 h, dehydrated in ethanol, cleared in xylol and coverslipped. Series 4, 5, and 6 were double immunostained first with rabbit anti-Camk4 Ab (1:2000; Sigma-Aldrich Co.) and then with one of the following: Mouse antinuclear receptor-related 1 protein (Nurr1) mAb (1:250; Abcam, Inc., Cambridge, MA, USA), mouse anti-NeuN mAb (1:250; Milipore), and mouse antiglutamic acid decarboxylase-67 (GAD67) mAb (1:2000; Millipore), followed with goat antirabbit rodhamine-labeled Ab (1:150, Invitrogen, Eugene, OR, USA) for Camk4 red fluorescence, and with horse antimouse biotinylated Ab (1:150, Vector Laboratories), followed with Avidin-BODIPY FL (1:1000, Invitrogen) for Nurr1, NeuN, and GAD67 green fluorescence. Series 7, 8, and 9 were double immunostained first with rabbit anti-Nurr1 Ab (1:250; Santa Cruz Biotechnology) and then with one of the following: Mouse anti-NeuN mAb (1:400; Milipore), mouse anti-GAD67 mAb (1:2000; Millipore), and mouse anti-2′,3′-cyclic nucleotide 3′-phosphodiesterase (CNPase) mAb (1:100; Milipore). Then, with goat antirabbit Alexa 488-labeled Ab (1:150; Invitrogen) for Nurr1 green fluorescence, and with antimouse biotinylated Ab (1:150; Vector Laboratories), followed with avidin–rodhamine (1:1000; Invitrogen) for NeuN, GAD67, and CNPase red fluorescence. Series 10 was double immunostained with mouse antiNurr1 mAb (1:250, overnight 4°C; Abcam Inc.) and rabbit antiglial fibrilary acidic protein (GFAP) Ab (1:1000; Sigma-Aldrich Co.). Then with horse antimouse biotinylated Ab (1:150; Vector Laboratories), followed with Avidin-BODIPY FL (1:1000; Invitrogen) for Nurr1 green fluorescence, and with goat antirabbit rodhamine-labeled (1:150; Invitrogen) for GFAP red fluorescence. All the primary antibodies were incubated overnight at 4°C, except GAD67 that was incubated during 48 h at 4°C, while secondary antibodies were incubated during 1.5 h, at room temperature. Fluorescent sections were mounted with ProLong (Invitrogen), studied and photographed with a confocal laser fluorescence microscope, and processed using the LCS Lite software.
In Situ Hybridization
Camk4 expression was analyzed by in situ hybridization using digoxigenin-labeled cRNA probes. Fetuses and pups at different ages (Supplementary Table 1) were perfused for 2 min with sterile 0.1 M PBS followed by 4% paraformaldehyde and 0.002% CaCl2 in 10 mM PB, and brains were then postfixed overnight at 4°C. Twenty-micrometer-thick coronal sections were obtained with Vibratome and collected in 0.4% paraformaldehyde in 0.1 M PBS. After overnight hybridization with denatured Camk4 digoxigenin-labeled probes, sections were rinsed in 0.1 M saline sodium maleic buffer (pH 7.0) and incubated overnight at room temperature in alkaline phosphatase-conjugated antidigoxigenin Ab (1/3500; Roche Diagnostics GmbH, Mannheim, Germany). The sections were then rinsed and reacted with 0.0022% 5-bromo-4-indoyl phosphate/nitroblue tetrazolium chloride (Sigma-Aldrich Co.) in 0.12 M Tris saline (pH 9.5; Sigma-Aldrich Co.). The sections were rinsed in 0.1 M PBS, dehydrated in ethanol, cleared in xylol, and coverslipped.
Several P19 sections (Supplementary Table 1) were double immunolabeled with sheep antidigoxigenin Ab (1:100, overnight at room temperature; Roche Diagnostics, Barcelona, Spain) and antisheep biotinylated Ab (1:150, 1 h at room temperature; Vector Laboratories) with Avidine-BODIPY FL (1:1000, 3 h at room temperature; Invitrogen) for Camk4-mRNA green fluorescence. Then, with rabbit anti-Camk4 Ab (1:2000, overnight at room temperature; Sigma-Aldrich Co.) and goat antirabbit rodhamine (1:150, 1 h at room temperature; Invitrogen) for Camk4 red fluorescence. Fluorescent sections were mounted with ProLong (Invitrogen), studied and photographed with a confocal laser fluorescence microscope, and processed using the LCS Lite software.
Quantification of Double-Immunolabeled Cells
Double-immunolabeled cells were plotted and counted in square fields (300-µm wide) from confocal photomicrographs. In total 16 fields, from 4 pups (2 pups/litter) in C, LMH and MMI groups were placed at random over SP, layer VIb, and wm of the parietal cortex. Plots and counts of labeled cells were obtained using the Cellgraph system (Microptic, S.L., Barcelona, Spain).
The Systat statistical software (Systat, Inc., Evanston, IL) was used. For concentrations of circulating hormones, we used 1-way analysis of variance (ANOVA). Frequency distributions of immunoreactive cell density were analyzed using 2-way ANOVAs, factors being ages and experimental groups. Significant differences between means (P ≤ 0.05 and ≤0.001, respectively) were identified by Tukey's test.
Camk4 and Nurr1 Immunolabeling
On E17, Camk4 labeling was found in the cortical plate (CP) and SP of C and MMI fetuses (Fig. 3A,B). This labeling pattern was maintained during embryonic development in C fetuses. However, the labeling increased in a narrow band in the inner CP (E19C; Fig. 3C; asterisk), which moved with age to more superficial locations, being located in E21C fetuses in a superficial band located in the immature layer V (Fig. 4D; asterisk). This pattern was maintained in LMH fetuses (Fig. 4E, asterisk). In contrast, both the radial pattern and the intensity of the immunolabeling were different in MMI pups. The SP labeling was weaker in E19MMI than in C fetuses, and the border between the SP and wm was blurred in E21MMI pups (Fig. 4F) compared with E21C (Fig. 4D) and E21LMH (Fig. 4E) pups. In addition, the characteristic band of immunolabeling in the CP was not present due to a more homogeneous distribution of the immunostaining to the CP (Figs 3D and 4F). During fetal development (Figs 3 and 4), no Camk4 labeling was observed in the intermediate zone (IZ) and in the subventricular and ventricular zones (SVZ–VZ, respectively) in any of the 3 groups studied.
The radial distribution of Camk4 labeling was similar between E21 and P0 pups (Fig. 5A1–C3). As in E21, the CP labeling in P0MMI pups (Fig. 5C1) was weaker than in P0C pups (Fig. 5A1). Between P0 and P10, the density of Camk4-immunoreactive (ir) neurons in all groups (all Camk4-ir cells were NeuN-ir; see below in Characterization of Camk4-ir and Nurr1-ir Cells) increased in the upper tier of CP, corresponding to upper layer V and layers II–IV (Fig. 5D1–E1). Heavily labeled Camk4-ir neurons in layer VIb (remnant of the SP) were scarce in P10C pups (Fig. 5D1), but instead, their density increased in P10LMH (Fig. 5E1; arrow) and P10MMI (Fig. 5F1; arrow) pups (see also Supplementary Fig. 2).
As already described (Hoerder-Suabedissen et al. 2009) in C pups, Nurr1-ir cells, as Camk4-ir neurons, were also found in SP on P0. They were also found in the CP, but in contrast to Camk4-ir neurons, they were more uniformly distributed through immature layers V and VIa (Fig. 5A2–C2). The density of Nurr1-ir cells in CP decreased with age and on P10, they were scarce and found in layer VIa. From P9 onwards, a diffuse immunoreactivity that labeled the barrels in layer IV was seen. Nurr1-ir cells in SP were seen in all groups, but they were more heavily labeled in P10MMI pups (Fig. 5F2; arrow). From P8 onwards, 2 types of Nurr1 immunolabeling were found. One set of Nurr1-ir cells was labeled inside the nucleus (Fig. 5H2; arrows), while the other set was weakly labeled in the cytoplasm (Fig. 5G2,H2; arrowheads). Nuclear labeling was either heavy or weak. In all groups, double-immunolabeled Camk4-ir/Nurr1-ir neurons were seen in layers II–VI, including SP (Fig. 5G1–G3,H1–H3). In LMH and MMI, almost all heterotopic Camk4-ir neurons in the wm were also Nurr1-ir (Fig. 5H3; arrows).
Between P11 and P20, as on P10 (Fig. 5), most of Camk4-ir neurons in the parietal cortex were located in layers II–IV and upper layer V (Fig. 6A1–C1) in all groups. On P11–12, Camk4-ir neurons in layer VIb (emerging from the SP) were still present in LMH and MMI pups. Their number was dramatically reduced thereafter in LMH pups, but they were still present in P20MMI pups (Fig. 6C1: arrow). Camk4-ir neurons in layer VIb were also seen in P60MMI pups (Fig. 7E,F; arrows).
In all groups, the density of Nurr1-ir cells in the parietal cortex decreased from P10 onwards and was progressively restricted to lateral layer VIa, although they remained in layer VIb (Fig. 6A2–C2; arrow).
Camk4 In Situ Hybridization
The radial distribution of Camk4 mRNA paralleled Camk4 immunolabeling. It was mostly found in both SP and CP. However, in E21LMH and E21MMI fetuses (Fig. 4B,C), the labeling in CP was weaker than in E21C fetuses (Fig. 4A), and the intermediate band of heavy immunolabeling observed in C and LMH pups was not present in the in situ preparations (Fig. 4A–C).
On P8–P10, Camk4 mRNA increased in cortical layers II–IV in all groups (Fig. 8A–C). In layer VIb (remnant of SP), decreased progressively in C pups and in P10C was no longer visible (Fig. 8A). In contrast, it was still present in P10LMH and P10MMI pups (Fig. 8B,C), but it was not seen in P11LMH pups. No differences in Camk4 expression between C and LMH pups were seen from P11 onwards. Heavy Camk4 expression in layer VIb (Fig. 8E) and in heterotopic cells of the wm (Fig. 8F) was still seen in P19MMI pups.
Characterization of Camk4-ir and Nurr1-ir Cells
Double immunolabeling for Camk4 and NeuN showed that from P8 onwards all Camk4-ir cells were NeuN-ir (Fig. 6E1–E3,F1–F3 and Supplementary Figs 2–4), except at P0, where 2.7% in P0C, 4.2% in P0LMH, and 2.1% in P0MMI were NeuN-immunonegative (Supplementary Fig. 1). In contrast, the percentage of NeuN-ir cells that were Camk4-immunonegative (NeuN/Camk4(−) in Fig. 9A) ranged from 9.7% in P20C to 22.8% in P10MMI (P < 0.05). In C and MMI pups, no statistically differences were found between means among ages; on average, 11.7% in C (P = 0.463) and 21.3% in MMI pups (P = 0.621). In LMH pups, the percentages significantly decreased with age; ranged from 19.9% in P0 to 10.6% in P20 (P < 0.05). At P0, were similar to P0MMI pups (P = 0.721), whereas at P15 and P20 reached C values (P = 0.678; Fig. 9A).
The percentage of Camk4-ir neurons that were GAD67-ir in P8C was 12.7% and only 3.5% in P8MMI pups (P < 0.05; Fig. 10A). Whereas 44.5% of GAD67-ir cells were Camk4-ir in P8C and 22.2% in P8MMI pups (P < 0.001; Fig. 10B). All GAD67-ir neurons showed weak Camk4 labeling in the cytoplasm (Fig. 6G1–G3). No differences between P8C and P8LMH were seen.
As mentioned before, from P8 onwards, nuclear and cytoplasmic Nurr1-ir cells were seen. The percentage of cytoplasmic Nurr1-ir neurons changed with age; it ranged from 2.7% of the total of Nurr1-ir cells in P10C to 34.6% in P20LMH pups (Supplementary Fig. 5). No cytoplasmic Nurr1-ir cells were seen before P8. A few nuclear labeled Nurr1-ir cells (2.2%) were NeuN-immunonegative (Fig. 6H1–H3). The percentage of NeuN-ir that were Nurr1-ir was similar in P20C (10.9%) and P20MMI (13.5%) pups (P = 0.433).
In C and in less extent in LMH pups, the percentages of Camk4-ir cells that were Nurr1-ir (Camk4(+)/Nurr1(+) in Fig. 9B) peaked at P10; 67.4% in C and 55.9% LMH pups. Then, these percentages dropped to 34.8% and 34.5% at P20, respectively. No significant difference among ages was found in MMI pups; 50.8% at P0MMI and 62.8% at P15MMI (on average 55.8%; Fig. 9B; P = 0.102). The percentage of Nurr1-ir cells that were Camk4-immunonegative (Nurr1(+)/Camk4(−) in Fig. 9C) was very low among ages. It ranged from 3.6% in P10MMI to 13.8% in P10C. No statistically differences between groups were found in P0 (on average, 4.9%; P = 0.855) and P20 (on average, 11.2%; P = 0.749) pups (Fig. 9C).
In P20C pups, 25.8% Nurr1-ir cells were GAD67-ir and this percentage dropped to 10.2% in P20MMI pups (P < 0.001; Fig. 10C). In P20C pups, 70.4% GAD67-ir were Nurr1-ir, while it was 50.9% in P20MMI pups (P < 0.001; Fig. 10D). All Nurr1-ir labeling in GAD67-ir neurons observed were weakly labeled in the cytoplasm (Fig. 6I1–I3). No differences between P20C and P20LMH pups were seen.
In P20C pups, 47.8% Nurr1-ir cells were CNPase-ir and only 16.3% in P20MMI pups (P < 0.001; Fig. 10C). In P20C pups, 65.0% CNPase-ir were Nurr1-ir and 100.0% in P20MMI pups (P < 0.001; Fig. 10C). The Nurr1-ir labeling in CNPase-ir cells was weak in the cytoplasm (Fig. 6J1,J2; arrows). The percentage of double-labeled Nurr1 and GFAP astrocytes was very low in both groups, about 1.5% in P20C pups (2 of 137 GFAP-ir astrocytes were double immunolabeled) and none in P20MMI pups (Fig. 6K1–K3). No differences between P20C and P20LMH pups were seen.
To our knowledge, this is the first study describing the radial distribution of Camk4 protein and mRNA in the parietal cortex of C rats and the alterations caused by temporal and chronic TH deficits. Both in LMH and MMI pups, Camk4 immunohistochemistry and in situ hybridization showed abnormal lamination and protracted expression in the SP and ectopic cells in the wm, suggesting altered migration during corticogenesis. Protracted Camk4/Nurr1 expression in the SP ceased on P11–12 in LMH pups (1–2 days later than in C pups) and remained in layer VIb of P60MMI rats. Double immunolabeling of Camk4-ir/NeuN-ir and Camk4-ir/Nurr1-ir revealed that, from P8 to P20, all Camk4 labeled cells are neurons and that nuclear Nurr1-ir cells in SP and wm are a subset of Camk4-ir neurons. Camk4- and Nurr1-immunolabeled cells were characterized by using neuronal (NeuN and GAD-67) and glial (GFAP and CNPase) markers, showing differences between C, LMH, and MMI pups. These results show that Camk4 is a novel marker for SP neurons.
Ontogeny and Function of SP Cells
The study of the ontogeny of SP and layer VIb in vertebrates is fundamental to understand the development and functional maturation of the neocortex (Wang et al. 2011). However, the equivalence of rodent SP with that of primates or carnivores (Kostovic and Rakic 1980; Luskin and Shatz 1985a, 1985b) has not been established yet. The SP emerges from the primordial plexiform zone or preplate (Marin-Padilla 1971; De Carlos and O'Leary 1992), but its size varies in different vertebrate species (Kostovic and Rakic 1990; Wang et al. 2011). In rodents, SP neurons are born around E11 (Price et al. 1997) and begin to extend axons toward the thalamus by E13 (De Carlos and O'Leary 1992; Molnár et al. 1998). Although the fate of SP neurons is controversial, recent studies have shown that SP cells, labeled with specific markers, remain in adulthood and become the adult layer VIb (Hoerder-Suabedissen et al. 2009; Hoerder-Suabedissen and Molnár 2013).
SP cells form a heterogeneous population (Kostovic and Rakic 1990; Hanganu et al. 2001; Watakabe et al. 2007; Hoerder-Suabedissen et al. 2009; Hoerder-Suabedissen and Molnár 2013), but the cellular diversity and function of the SP are not well known yet despite that a number of SP markers are available (Chun and Shatz 1989; Allendoerfer and Shatz 1994; McQuillen and Ferriero 2005; Kostovic and Judas 2007; Watakabe et al. 2007; Bayatti et al. 2008; Hoerder-Suabedissen et al. 2009; Oeschger et al. 2012). There is strong evidence that SP neurons play an important role in thalamo-cortical axon path finding (Molnár et al. 2012). During development, SP neurons may fire action potentials (Torres-Reveron and Friedlander 2007). They are necessary for the establishment of ocular dominance and orientation columns (Kanold et al. 2003; Kanold and Shatz 2006) and for the maturation of inhibitory circuits in layer IV (Kanold and Shatz 2006). The dendritic morphology of SP neurons change by P7, according to the flow of sensory information (Hoerder-Suabedissen and Molnár 2012), and might explain the rearrangement of neurites seen in the Golli-tau-eGFP mouse (Piñón et al. 2009). These plastic changes seen in the Golli-tau-eGFP mouse contribute to the dynamic integration of SP neurons into the cortical barrel field circuitry during postnatal development, which plays a key role to establish the cytoarchitectonic pattern in layer IV and to refine layer IV circuitry (Piñón et al. 2009). SP and wm abnormalities have been related to the pathogenesis of various brain developmental disorders, including periventricular leukomalacia, autism, schizophrenia, and cerebral palsy (Volpe 1996; Eastwood and Harrison 2003; McQuillen and Ferriero 2005; Volpe 2009; Stolp et al. 2012). For this reason, crucial importance for a better understanding of the role of SP neurons in cortical development is the identification of SP cells subpopulations, which may have very different roles in various pathologies.
In this study, we have found that Camk4 is a novel marker for SP neurons as shown with Nurr1 double immunolabeling. Camk4 labeling in the SP and layer VI in the parietal cortex decreases in P10C pups, while it was still visible in P12LMH and P60MMI pups in similar antero-posterior and rostro-caudal levels. Whether or not this altered expression indicates a delay in the maturation of the SP in LMH and MMI pups (e.g., caused by delay in the maturation of SP to become layer VIb) warrants further study.
Decreased density of GABAergic neurons by 50–80% was observed in postmortem wm sections of premature infants with wm lesions born at 25–32 weeks of gestation, compared with premature infants without wm lesions, depending on birthdate suggesting altered tangential migration (Robinson et al. 2006). We have observed that the percentage of GAD67-ir/Camk4-ir neurons in P8MMI (22.2%) compared with P8C (44.5%) pups was reduced by 50.1%, and that of GAD67-ir/Nurr1-ir neurons in P20MMI (50.9%) compared with P20C (70.4%) was reduced by 27.7% (Fig. 9B,D). This reduction might reflect a decreased number of GAD67-ir GABAergic neurons in infragranular layers, including SP, as a result of either reduced neurogenesis ratio or altered tangential migration. The later was observed in organotipic cultures of hypothyroid embryos (Cuevas et al. 2005). Another possibility, which does not exclude previous ones, is that as a result of TH deficiency, GAD67 expression would be altered in both Camk4 and Nurr1 neurons of LMH and MMI pups. In fact, this is not unlikely because in a previous study decreased parvalbumin immunoreactivity has been observed in hypothyroid rats, although their radial distribution was similar to C rats (Berbel et al. 1996). Decreased cortical GABAergic neurons might cause increased audiogenic seizures reported in perinatal hypothyroid (van Middlesworth and Norris 1980) and in transiently hypothyroxinemic (Ausó et al. 2004) rats. However, further studies are needed to quantify the possible loss of GABAergic neurons in the neocortex (including SP) and to clarify the organization of inhibitory circuits in LMH and MMI rats.
Possible Effects of Altered Camk4 and Nurr1 Expression on the Development and Maturation of the Neocortex, and Implications for Late Maternal Hypothyroidism in Man
The functional role of Camk4 expression in neurons during development and the possible implication of Camk4 in neuronal pathologies are under study. Therefore, a direct relationship between the alterations found in our study and human neural diseases is still very speculative, but some cues can be obtained from the available data. Camk4 is an activator of several transcription factors, such as CREB, and in proteins involved in the CREB pathway (Chawla et al. 1998). It is well known that TH affects many functional roles in the CNS, impacting neurogenesis, neuronal differentiation and maturation, synaptogenesis, and behavior (Zoeller and Rovet 2004; Berbel et al. 2007). A prominent role of the Camk4–CREB pathway has been proposed to be involved in these developmental processes (Morte et al. 2010) and of CREB in neuropsychiatric disorders (Carlezon et al. 2005). There is strong evidence for the action of the Camk4–CREB pathway in the expression of FMR1 gene, encoding fragile X mental retardation protein (FMRP; Wang et al. 2009, 2012). The lack of FMRP causes the fragile X syndrome, which is the most common cause of inherited mental retardation and autism spectrum disorders (Krueger and Bear 2011). The later also frequent in preterm children (Johnson 2007; Simic et al. 2009). Camk4 and many components of its downstream pathways are under TH control in the fetal cortex (Morte et al. 2010). In LMH pups, reduction in the number of synapses in CA3, reduced ERK and CREB expression in the hippocampus and decreased step-down latency has been observed (Berbel et al. 2010). In addition, Opazo et al. (2008) have found that transient maternal and fetal hypothyroxinemia in pregnant rats at the beginning of fetal corticogenesis affects spatial learning and the synaptic nature and function in the hippocampal CA1 of their offspring. In the CNS, Camk4–CREB pathway seems to be active in neurons since Camk4 has not been found expressed in glial cells (Watterson et al. 2001; Murray et al. 2009). Our data support these findings because we have not found double Camk4 immunolabeling neither with GFAP nor with CNPase. Therefore, an alteration the Camk4–CREB pathway in neocortical neurons most likely will affect neocortical circuitry, unbalancing the ratio of excitatory to inhibitory inputs.
Our experimental model shows alterations in euthyroid fetuses born rats made hypothyroid after onset of fetal thyroid function and in part, it mimics the situation of human preterm neonates who must complete their neurodevelopment without a maternal TH supply (Berbel et al. 2010). Although the postnatal development and maturation of the CNS are comparatively longer in humans than in rats, similarities may be established when onset of fetal thyroid gland secretion is taken as the reference point (Berbel et al. 2007).
The changes in Camk4 and Nurr1 expression found here may help to explain the decreased mental development described in children born prematurely (O'Callaghan et al. 1995; Kester et al. 2004; Rovet and Simic 2008; Williams and Hume 2008). Altered Camk4 and Nurr1 expression might not only affect SP development and maturation, but also cortical connectivity. Altered Camk4 and Nurr1 expression in the neocortex may help to understand the decreased mental development described in children born prematurely (O'Callaghan et al. 1995; Kester et al. 2004; Rovet and Simic 2008; Williams and Hume 2008), and other neuropathologies associated with prematurity such as schizophrenia (Harrison and Eastwood 2001; Lewis and Levitt 2002,Harrison and Law 2006; Addington et al. 2007) or autism spectrum disorders (Johnson 2007; Simic et al. 2009).
Our results show that, during development, TH have selective and permanent effects on SP maturation, affecting the cytoarchitectonic organization of the cerebral cortex. Any situation resulting in a decreased availability of T4 to the fetal and postnatal brain is potentially adverse for its development and maturation.
This work was supported by the Spanish Ministerio de Ciencia e Innovación (SAF2009-10689 to P.B. and SAF2011-25608 to J.B.).
We thank Prof. M.J. Obregón for her help with the TH determinations. Conflict of Interest: None declared.