Behavioral tests

All experiments were approved by the Chancellor’s Animal Research Committee at UCLA and conducted according to the National Institute of Health guidelines. Earlier studies of Reln or dab1 mice found that their responses did not differ by sex (Villeda et al., 2006; Akopians et al., 2008).

Tissue preparation and immunohistochemistry

One hour after noxious stimulation, dab1^lacZ mice were re-anesthetized (Sodium Pentobarbital, 100 mg/kg) and perfused transcardially with 4% paraformaldehyde, and post-fixed for 1 hour (4°C) in the same fixative. Reln and Reln^rl–Orl;GAD67^GFP mice received the same fixative but with a 1–3 hour post-fix. Following an overnight wash with 0.12M phosphate buffer (PB), spinal cords were dissected and cryoprotected in 30% sucrose/PB for 2–3 days. Upper cervical (C1-3) and lumbar (L4-5) segments were blocked and frozen in Optimum Cutting Temperature (Sakura) and stored at −80°C.

To evaluate Fos expression, we immunostained 5–6 40 μm coronal cryostat sections per mouse lumbar levels 4–5. We used rabbit anti-Fos (1:10,000; Calbiochem, PC38) with 0.1M Tris buffer with 1.4% NaCl and 0.1% bovine serum albumin (TBS). Standard avidin-biotin techniques combined with Nickle-intensified diaminobenzidine were used to visualize Fos expression. Sections were mounted, dried, dehydrated, cleared and coverslipped. Coronal or sagittal sections (25–30 μm) were used for double immunofluorescent experiments.

Reduced mechanical sensitivity and Fos expression characterize dab1^lacZ/lacZ mice

Results from the Chaplan up-down test of mechanical sensitivity found no differences in the 50% withdrawal latencies between dab1^+/+ and dab1^lacZ/+ mice (Fig. 1F). The dab1^lacZ/lacZ mice, however, had much higher mechanical thresholds (Fig. 1F). After noxious mechanical stimulation the number of Fos-labeled cells did not differ between the dab1^+/+ and dab1^lacZ/+ dorsal horns (Fig. 1G), whereas the dab1^lacZ/lacZ mice had fewer Fos-expressing cells in laminae I–II and LSN (Fig. 1G–J, short arrows). Again, Fos-labeled neurons in dab1^+/+ and dab1^lacZ/+ lateral reticulated area were greatly reduced in dab1^lacZ/lacZ mice (Fig. 1H–J, long arrows). These results support our previous findings on reduced mechanical pain sensitivity in Reln and dab1 mutants (Villeda et al., 2006; Akopians et al., 2008; Wang et al., 2012).

Dab1 expression in the adult dorsal horn

As mutations in the Reelin pathway cause abnormalities in neuronal positioning, we then characterized Dab1 expression in adult dorsal horn using a combination of β-gal histochemistry and Dab1 immunohistochemistry. In the superficial dorsal horn, it is difficult to distinguish Dab1-expressing neurons due to extensive terminal staining. Dab1 expression in dab1^+/+ appeared as diffuse reaction product and a few discernible neurons (small arrowheads in Fig. 2A). Several Dab1 cells are detected in laminae III–IV, while the lateral reticulated area of lamina V has a number of medium and a few large Dab1-labeled neurons (large arrowheads in Fig. 2A). The LSN contained strong Dab1 reaction product and a few labeled cells.

Open in a separate window

Figure 2

Dab1 and β-galactosidase are co-expressed in the dorsal horn

A, Dab1 immunoreactivity (brown) concentrated in dab1^+/+ laminae I–II, lateral reticulated area of lamina V, and the lateral spinal nucleus (LSN). Relatively small Dab1 neurons (small arrowheads) are in laminae I–II whereas medium and large Dab1 neurons (large arrowheads) are found in the lateral reticulated area. B, The dab1^lacZ/+ dorsal horn contains both brown Dab1 and blue β-gal precipitate. The colocalization seen in the boxed lateral reticulated area is enlarged in the inset. C, The dab1^lacZ/lacZ dorsal horn contains a pattern of blue β-gal reaction product similar to Dab1 immunoreactivity shown in A and B. Scale bar: A–C, 200 μm; Inset in B, 50 μm.

In the dab1^lacZ/+ dorsal horn, the β-gal reaction product and Dab1 immunoreactivity are found within the same areas and co-localize within cells of the lateral reticulated area (Fig. 2B and inset), confirming that both methods identify the same neurons. As expected, the dab1^lacZ/lacZ dorsal horn contained only β-gal deposits which were concentrated in laminae I–II and the lateral reticulated area (Fig. 2C), similar to the protein expression pattern in dab1^+/+ dorsal horn (Fig. 2A). Because β-gal deposits are not found in axons of Dab1 neurons (Abadesco et al., 2014), the heavy β-gal product in the superficial dorsal horn suggests that there are many more Dab1-labeled neurons than we previously reported (Villeda et al., 2006; Akopians et al., 2008). The heterogeneity in size and regional distribution suggests that Dab1 is widely expressed in several distinct dorsal horn populations.

Superficial dorsal horn: most Dab1 cells co-express the transcription factor Lmx1b

Next we sought to identify the phenotype of the small Dab1 neurons in laminae I–II that resemble the classically defined nociceptive-specific dorsal horn neurons. Because the majority of interneurons in this area are excitatory (Todd, 2010), we asked whether they were glutamatergic neurons. There are two homeobox proteins, Lmx1b and Tlx3 (T-Cell Leukemia Homeobox 3), that mark a late-generated dorsal neuron population (dIL^B) that migrates to the superficial laminae of the dorsal horn and differentiates into an excitatory glutamatergic phenotype (Gross et al., 2002; Müller et al., 2002; Cheng et al., 2004; Cheng et al., 2005; Dai et al., 2008; Rebelo et al., 2010). Because Lmx1b expression is maintained in adults (Dunston et al., 2005; Dai et al., 2008), we asked if Dab1 dorsal horn neurons express Lmx1b. The β-gal product was detected in the same dab1^lacZ/+ and dab1^lacZ/lacZ dorsal horn areas as Lmx1b (data not shown). To better identify the Dab1-expressing neurons, we conducted double immunofluorescence experiments, confocal imaging and extensive analyses on Reln^+/+ and Reln^−/− dorsal horns and confirmed Dab1 and Lmx1b co-localization (Fig. 3A–F). Analyses of 3 μm confocal images revealed that 70% of the small to medium Dab1-expressing superficial dorsal horn neurons co-express Lmx1b (yellow arrowheads in Fig. 3B, B1-2, E, E1-2; Reln^+/+ 71±2%; Reln^−/− 70±2%). Compared to Reln^+/+, there are a few medium and large-sized Dab1-Lmx1b neurons in Reln^−/− laminae I–II (Fig. 3F; compare Fig. 3A to D and Fig. 3G to H) and dorsal funiculus (data not shown). Based on previous studies (Cheng et al., 2004; Cheng et al., 2005; Dai et al., 2008), the Dab1-Lmx1b neurons are likely to be glutamatergic.

Open in a separate window

Figure 3

Superficial dorsal horn: 70% of Dab1 cells co-express Lmx1b and are mispositioned within the Reln^−/− IB4 terminal zone

A–F, Confocal 3 μm thick images of Reln^+/+ (A–C) and Reln^−/− (D–F) lumbar dorsal horns show that many cells express both Dab1 (red cytoplasm) and Lmx1b (green nucleus), including small interneurons in laminae I–II (small yellow arrowheads) and large neurons in laminae I–III (large yellow arrowheads). Examples of single-labeled Dab1 cells are marked by white arrows. Boxes in A and D are enlarged in B, C, E, and F. Individual channels of B and E are labeled B1-2 and E1-2. G–H, Lumbar confocal images show the distribution of Dab1 (red) and Lmx1b (cyan) within and above the IB4 (green) band of terminals. Although the area of lamina II marked by IB4 afferents does not vary by genotype, the area of laminae I–IIouter, located dorsal to the IB4 terminals, is larger in Reln^+/+ than in Reln^−/− mice. An arrowhead marks a large ectopic Dab1-Lmx1b neuron in Reln^−/− lamina I. I, The number of Dab1 cells in laminae I–II outer, i.e., above the IB4 area, do not vary by genotype, but more Dab1, Dab1-Lmx1b, and total Dab1 neurons are found within the Reln^−/− compared to the Reln^+/+ IB4 band (n=4 mice/genotype). Scale bars: A, D; G–H, 100 μm; B, E; C, F; B1-2, E1-2, 50 μm.

We also tested other common markers of excitatory superficial dorsal horn neurons (Antal et al., 1991; Todd, 2010; Gutierrez-Mecinas et al., 2016). Although many Protein Kinase C gamma (PKCγ)-positive interneurons express Lmx1b, none of the Dab1-labeled neurons co-expressed PKCγ (data not shown). The Dab1 neurons also did not express Calretinin, but a few Dab1-Lmx1b neurons did contain Calbindin and Somatostatin (data not shown). Additionally, our previous studies showed that a few Dab1 dorsal horn neurons expressed the Neurokinin-1 receptor that binds substance P and were mispositioned in Reelin-signaling pathway mutant dorsal horns (Villeda et al., 2006; Akopians et al., 2008). We conclude that most of the Dab1 neurons in the superficial dorsal horn are excitatory glutamatergic neurons.

Superficial dorsal horn: Dab1-Lmx1b-expressing neurons are mispositioned in Reln^−/− mice

As mispositioned neurons in Reln^−/− mice express high levels of Dab1, we compared the distribution of Dab1 and Dab1-Lmx1b neurons in Reln^+/+ and Reln^−/− superficial dorsal horn. Initial analyses divided laminae I–II into equal-sized bins in both the mediolateral and dorsoventral divisions, but the number of Dab1 neurons per bin did not differ between genotypes.

Because the superficial dorsal horn lamination pattern is precisely defined by primary afferent terminations, we next focused our analysis on lamina IIinner, which is marked by isolectin B4 (IB4) and receives nonpeptidergic nociceptive afferents (Basbaum et al., 2009; Todd, 2010). Although the area identified by IB4-positive afferents did not vary by genotype, we found fewer Dab1-only (Reln^+/+ 4±0.4; Reln^−/− 8±1; p=0.002, Fig. 3I) and Dab1-Lmx1b neurons (Reln^+/+ 13±2; Reln^−/− 18±2; p=0.043, Fig. 3I) in the Reln^+/+ than Reln^−/− IB4 region. Additionally, the IB4 band in Reln^−/− lamina IIinner appeared shifted dorsally and perhaps laterally (Fig. 3G–H). Consistent with the more dorsal location of IB4-positive terminals, the area of laminae I–II outer was larger in Reln^+/+ than in Reln^−/− mice (Reln^+/+ 17,444± 1,357 μm²; Reln^−/− 11,264± 978 μm²; p=0.002, Fig. 3G–H). Despite the reduced area in Reln^−/− laminae I–II outer, the number of Dab1 cells in this area did not vary by genotype (Fig. 3I). To verify that the Dab1 cells did not die selectively in Reln^−/−, we compared total Dab1 neurons per 3 μm hemisections of Reln^+/+ (65±14 neurons) and Reln^−/− (77±11 neurons) mice and found no significant difference (p=0.15). Thus the Dab1- and Dab1-Lmx1b-labeled laminae I–II neurons, and the IB4-positive terminals all appear to be mispositioned in Reln^−/− superficial dorsal horn.

Superficial dorsal horn: several Dab1 neurons are GABAergic

After finding that 70% of the Dab1 superficial dorsal horn neurons are excitatory, we asked if the other Dab1 neurons in laminae I–II might be inhibitory and express glutamic acid decarboxylase (GAD67). To best answer this question, we used Reln^rl–Orl mice interbred with a well-characterized GAD67^GFP line that expresses GFP under the GAD67 promoter (Tamamaki et al., 2003; Abadesco et al., 2014). The majority of GAD67^GFP-labeled neurons did not co-localize with Dab1 in Reln^rl−Orl/+ or Reln^{rl/Orl/rl−Orl} superficial dorsal horn (Fig. 4A–D). A few double-labeled neurons were present, however, in laminae I–II (yellow arrowheads in Fig. 4B, B1-2, D, D1-2). We also observed Dab1 and Dab1-GAD67^GFP neurons in the white matter dorsal to lamina I in Reln-Orl^+/+ (Fig. 4E1-3) and Reln^{rl/Orl/rl−Orl} (Fig. 4F1-3) sagittal sections. More large Dab1-immunoreactive neurons appeared to be located in Reln^{rl/Orl/rl−Orl} than in Reln-Orl^+/+ superficial dorsal horn (Fig. 4F1, 3 vs 4E1, 3). Markers of subsets of GABAergic dorsal horn interneurons, such as those that express neuronal nitric acid synthase, Parvalbumin, and ChAT also were tested and did not co-localize with Dab1 cells (Fig. 4G–H; data not shown; Laing et al., 1994). We conclude that relatively few Dab1 neurons are inhibitory.

Open in a separate window

Figure 4

A small number of Dab1 neurons in laminae I–II are GABAergic and in the lateral reticulated area are cholinergic

A–D, Confocal images of Reln^rl−Orl/+; GAD67^GFP/+ (A–B) and Reln^{rl−Orl/rl−Orl}; GAD67^GFP/+ (C–D) lumbar dorsal horns. Boxes in A and C are enlarged in B and D, with individual channels in B1-2 and D1-2. A few Dab1 (red) and GAD67^GFP neurons appear co-localized (yellow arrowheads in B and D). E–F, Sagittal lumbar dorsal horn sections from Reln-Orl^+/+;GAD67^GFP/+ (E1-3) and Reln^{rl−Orl/rl−Orl}; GAD67^GFP/+ (F1-3) mice illustrate ectopic cells in the white matter dorsal to lamina I. Dab1 (white arrows in E1, 3) and Dab1-GAD67^GFP-labeled (yellow arrowheads in E1-3, F1-3) neurons are frequently found dorsal to lamina I (top of image). G–H, Reln^+/+ and Reln^−/− dorsal horns show that Dab1 (red) and ChAT-positive neurons (green) do not co-localize in laminae I–IV, but double-labeled cells boxed in G and H (yellow arrowheads in G1-3, H1-3) are found in the lateral reticulated area of lamina V. Single-labeled Dab1 and ChAT neurons are marked with white arrows. Scale bars: A, C; G–H, 100 μm; B, D; B1-2, D1-2; E1-3, F1-3; G1-3, H1-3, 50 μm.

Lateral reticulated area and LSN: Dab1-Lmx1b neurons are mispositioned in Reln^−/− mice

Previously we found Dab1 neurons in the lateral reticulated area of lamina V, an important nociceptive area containing projection neurons for which few markers have been identified. Here we asked if the Dab1 neurons in the lateral reticulated area also express Lmx1b and found that 67% of Dab1 neurons co-express Lmx1b in both genotypes (Fig. 5A, yellow arrowheads in Fig. 5C, C1-2, E, E1-2). We then analyzed both Dab1- and Dab1-Lmx1b-labeled neurons to look for evidence of neuronal positioning errors in Reln^−/− mice. Greater numbers of Dab1-Lmx1b neurons are found in the Reln^+/+ (4±0.4) than in Reln^−/− lateral reticulated area (2±0.3; p=0.0004, Fig. 5G), whereas fewer single-labeled Dab1 neurons are present in Reln^+/+ (2±0.3) than Reln^−/− mice (4±0.6; p=0.002, Fig. 5G). Several cholinergic neurons in the lateral reticulated area (Phelps et al., 1984) also expressed Dab1 in Reln^+/+ and Reln^−/− mice (yellow arrowheads in Fig. 4G1-3 and 4H1-3), but due to the small number of double-labeled cells, positioning errors were not evaluated.

Open in a separate window

Figure 5

Lateral reticulated area and LSN: Dab1-Lmx1b-labeled cells are mispositioned in Reln mutants

A–F, Lumbar confocal images show that Dab1 (red) and Lmx1b (green) are co-expressed in Reln^+/+ (A–D) and Reln^−/− (E–F) lateral reticulated area of lamina V (A, C, E, yellow arrowheads) and LSN (B, D, F, yellow arrowheads). Close to 70% of Dab1 neurons in both areas co-express Lmx1b. Single-labeled Dab1 cells are marked with white arrows. A, C, E, Boxed area in A is enlarged in C and shows a more organized array of Dab1-Lmx1b neurons in Reln^+/+ than Reln^−/− lateral reticulated area (E). B, D, F, Boxed area of Reln^+/+ LSN is enlarged in D and contains more Dab1-Lmx1b-labeled cells than in Reln^−/− LSN (F). G, Fewer Dab1 only and more Dab1-Lmx1b-labeled cells are in Reln^+/+ than in Reln^−/− lateral reticulated area. H, More Dab1-Lmx1b-labeled and total Dab1 cells are seen in Reln^+/+ than Reln^−/− LSN (n=5 mice per genotype). Scale bars: A, B, 100 μm; C–F; C1-2, D1-2, E1-2, F1-2, 50 μm.

As with other Reelin pathway mutants, we recorded about 50% more NeuN-labeled neurons in the LSN of dab1^+/+ and dab1^lacZ/+ than found in dab1^lacZ/lacZ mice (dab1^+/+ 11±1; dab1^lacZ/+ 11±1; dab1^lacZ/lacZ 5±1; dab1^+/+ vs dab1^lacZ/lacZ p=0.003; dab1^lacZ/+ vs dab1^lacZ/lacZ p=0.004; Villeda et al., 2006; Akopians et al., 2008). Here we report that almost 70% of the Dab1 neurons in the LSN co-express Lmx1b (Fig. 5B, yellow arrowheads in Fig. 5D, D1-2, F, F1-2), with almost twice as many Dab1-Lmx1b neurons in Reln^+/+ (3.2±0.5) than Reln^−/− LSN (1.7±0.2; p=0.008, Fig. 5H). The total Dab1-positive neurons in Reln^+/+ (5±1) is also greater than in Reln^−/− LSN (3±0.3; p=0.02, Fig. 5H). Thus Dab1 cells from the lateral reticulated area and the LSN are incorrectly positioned in Reln^−/− mice, including those that co-localize with Lmx1b.

Dab1-Lmx1b-positive neurons participate in nociceptive circuits

Having determined that the majority of the Dab1-expressing neurons in the superficial dorsal horn, lateral reticulated area, and LSN co-express Lmx1b, we next asked if the Dab1-Lmx1b neurons are activated by noxious thermal or mechanical stimulation. We found clear examples of Fos protein expression in Dab1-Lmx1b-positive neurons in dab1^+/+ and Reln^+/+ superficial dorsal horn (Fig. 6A, A1-4, C, C1-4). Triple-labeled cells (white arrowheads) in the wild-type lateral reticulated area (Fig. 6B, B1-4, D, D1-4) also were detected after thermal or mechanical stimulation. Dab1-Fos- (magenta arrows) and Lmx1b-Fos- (cyan arrows) expressing neurons also were evident (colored arrows in Fig. 6A, A1-3, B, B1-3, C, C1-3, D, D1-3). Thus Dab1-, Lmx1b-, and Dab1-Lmx1b neurons participate in both noxious heat and mechanical circuits.

Open in a separate window

Figure 6

Noxious stimulation induced Fos expression in Dab1-Lmx1b neurons

Triple TSA immunofluorescence labeling showed that after both thermal (A–B) and mechanical stimulation (C–D), some Dab1 (red) and Lmx1b (green)-positive neurons express Fos (cyan) in dab1^+/+ (A–B) and Reln^+/+ (C–D) superficial dorsal horns (A, A1-4, C, C1-4) and in the lateral reticulated area (B, B1-4, D, D1-4). Boxed areas in A4–D4 are enlarged in A–D, respectively. Triple-labeled cells are marked with white arrowheads. Double-labeled Dab1-Fos (magenta), Lmx1b-Fos (cyan), and Dab1-Lmx1b (yellow) cells are marked with color-matched arrows. Heat and mechanical stimulation induced Fos in medium to large-sized Dab1-Lmx1b neurons in the superficial dorsal horn and the lateral reticulated area. More Lmx1b neurons appeared to express Fos after mechanical stimulation especially in the superficial dorsal horn. Dab1-Lmx1b neurons in the lateral reticulated area appear as an organized array in wild-type mice (B, D). Scale bars: A4; B4; C4; D4, 100 μm; A; A1-3; B; B1-3; C; C1-3; D; D1-3, 50 μm.

Lateral cervical nucleus: Dab1 neurons sustain positioning errors in Reelin-signaling pathway mutants

The neurons of the lateral cervical nucleus (LCN; Fig. 7A) are surrounded by axons of the dorsolateral funiculus, but only in cervical levels 1–3 (Kajander and Giesler, 1987). The areas of LCN and LSN cannot be differentiated in these upper cervical levels and thus are analyzed together (Burstein et al., 1990). The LCN neurons receive nociceptive information from neurons throughout the spinal cord and convey it rostrally as part of the spinocervicothalamic pathway (Giesler et al., 1979; Kajander and Giesler, 1987; Burstein et al., 1990). Most LCN neurons project to contralateral thalamus and respond to a wide-range of innocuous and noxious thermal and mechanical stimuli (Giesler et al., 1979; Kajander and Giesler, 1987).

Open in a separate window

Figure 7

Lateral cervical nucleus: Neurons are mispositioned in Reelin pathway mutants

A, Schematic of the lateral cervical nucleus (LCN) found at levels C1-3. As there is no delineation between the LCN and LSN, they were analyzed together. B–C, Dab1 expression in Reln^+/+ (B) and Reln^−/− (C) cervical dorsal horn is concentrated in the superficial dorsal horn and LCN/LSN. D–F, β-gal histochemistry followed by Dab1 immunoreaction reveals brown Dab1 protein (dab1^+/+, dab1^lacZ/+) and blue β-gal (dab1^lacZ/+, dab1^lacZ/lacZ) in the superficial dorsal horn and LCN/LSN. G–I, The dab1^+/+ and dab1^lacZ/+ LCN/LSN contain about twice as many NeuN-labeled cells as dab1^lacZ/lacZ mice (n=5 mice/genotype). J–L, Neurons and their processes that express the Neurokinin-1 receptor show a compression in the dab1^lacZ/lacZ (L) compared to dab1^+/+ (J) and dab1^lacZ/+ (K) LCN/LSN (n=3 mice/genotype). Scale bars: B–F, 200 μm; G–L, 100 μm.

Due to the importance of this area in relaying nociceptive information, we first examined the LCN to determine if neuronal mispositioning in this region might contribute to the sensory abnormalities in Reelin pathway mutants. The Dab1 expression in the C1-3 segments had strong immunoreactivity in neurons of laminae I–II and in the combined LCN/LSN (Fig. 7B–D). The dab1^lacZ/+ LCN/LSN contains neurons that are double-labeled with β-gal and Dab1 (brown, Fig. 7E), whereas the dab1^lacZ/lacZ has only β-gal concentrated in laminae I–II, the lateral reticulated area, and in the LCN/LSN area (Fig. 7F). This demonstrates that the pattern of Dab1 expression at high cervical levels resembles that in lumbar spinal cord (Fig. 2).

We next asked if LCN neurons are mispositioned in dab1^lacZ/lacZ mice. The dab1^+/+ LCN/LSN contained an average of 90±6 neurons/dorsal horn and did not differ from dab1^lacZ/+ LCN/LSN with 87±5 neurons. The dab1^lacZ/lacZ LCN/LSN contained 46±2 NeuN-labeled neurons, 49% and 47% fewer neurons than in dab1^+/+ and dab1^lacZ/+ mice, respectively (dab1^+/+ vs dab1^lacZ/lacZ p=0.00002; dab1^lacZ/+ vs dab1^lacZ/lacZ p=0.00004; Fig. 7G–I). We also analyzed Reln^+/+ and Reln^−/− LCN/LSN neurons and found that on average, Reln^+/+ LCN/LSN contained 78±7 neurons/dorsal horn while Reln^−/− mice had 41±5 neurons, a 48% reduction in LCN/LSN neurons (p=0.007). Interestingly, nearly 50% of both dab1^lacZ/lacZ and Reln^−/− LCN/LSN neurons are displaced, findings which support our contention that this disruption is due to the loss of the canonical Reelin-Dab1 signaling pathway.

The loss of Reelin in the spinal cord affects Dab1 neurons differently

Both somatic and preganglionic motor neurons in the spinal cord express Dab1 and respond to Reelin signaling, yet in Reln^−/− and dab1^−/− mice, somatic motor neurons have rather subtle positioning errors compared to the extensively mispositioned sympathetic and parasympathetic preganglionic neurons (Yip et al., 2000; Phelps et al., 2002; Palmesino et al., 2010; Abadesco et al., 2014). Dab1-labeled dorsal horn neurons also show differences in the extent of their positioning errors. Disruptions in the cellular organization of the smaller Dab1 interneurons in Reln^−/− laminae I–II were difficult to identify, whereas the larger Dab1 projection neurons in Reln^−/− lateral reticulated area, LSN, and LCN/LSN were clearly missing from their respective areas. The final locations of these mispositioned Dab1 projection neurons remain unclear but likely include the extra Dab1 and Dab1-Lmx1b neurons in the IB4 area of laminae IIinner and large ectopic cells detected along the outer edge of lamina I.

Based on previous studies, the migratory pathway of Dab1 and the late-born Lmx1b (dIL^B) neurons overlap extensively, both temporally and spatially (Müller et al., 2002; Dunston et al., 2005; Villeda et al., 2006; Rebelo et al., 2010). If Reelin functions to terminate the migration of the laterally located Dab1 neurons in the lateral reticulated area, LSN and LCN as reported with the sympathetic preganglionic neurons, then the absence of Reelin would cause these dorsal horn projection neurons to fail to stop laterally and instead migrate past their normal locations or return to the midline (Yip et al., 2000; Phelps et al., 2002; Krüger et al., 2010). On the other hand, the Dab1 neurons in the Reln^−/− superficial dorsal horn seem to remain near their correct laminae but are somewhat out of place compared to Reln^+/+ mice. Because the number of Dab1 neurons and the percentage of Dab1 neurons co-expressing Lmx1b do not differ between Reln^+/+ and Reln^−/− superficial dorsal horns, there is no evidence of selective cell death in our model. Furthermore, based on previous studies (Caviness and Rakic, 1978; Rice and Curran, 2001), neuronal generation is usually normal in Reelin-signaling pathway mutants, but their migration pathways are altered.

Dab1 and Lmx1b both identify large neurons in the lateral reticulated area which express Fos following mechanical or thermal stimulation. Based on their size, location, and response to stimulation, these Dab1- and Dab1-Lmx1b cells likely correspond to the so-called wide-dynamic-range neurons that relay nociceptive information to rostral targets (Menétrey et al., 1980). Similarly, Dab1 and Lmx1b also identify large neurons in the LCN/LSN, many of which are likely to be projection neurons that relay nociceptive information to the thalamus (Giesler et al., 1979; Kajander and Giesler, 1987). The significant positioning errors sustained by these neurons imply that Reelin regulates their migration during development. Although generally the correct afferents still contact mispositioned neurons, it is unclear whether or not normal functional connections are formed (Yip et al., 2003; Pascual et al., 2004), especially since Reelin signaling was found to play a role in regulating hippocampal synaptic function (Qiu et al., 2006; Pujadas et al., 2010). In addition, Reelin-Dab1 signaling has been implicated in dendrite development in the hippocampus and the cerebral cortex (Niu et al., 2004; Olson et al., 2006; Matsuki et al., 2008). Reln^−/− and dab1^−/− somatic motor neurons also have stunted dendrites compared to wild-type mice (Phelps et al., 2002; Abadesco et al., 2014). Thus without Reelin, the Dab1 dorsal horn neurons also may have abnormal dendrites, in addition to being mispositioned, which would further impair their neuronal connections.

We thank Dr. Allan Basbaum for valuable guidance and suggestions on the manuscript, Dr. Brian Howell for providing the dab1^lacZ mice and Dab1 antibody, Drs. Carmen Birchmeier and Thomas Müller for providing Lmx1b antibody, Dr. Alin Akopians for helpful comments on the manuscript, Frank Lee for assistance with Reln^rl–Orl; GAD67^GFP data, and Aly Mulji for mouse colony care and completing the NK-1R data analyses. This study acknowledges support from the National Science Foundation (IOB-0924143 to PEP) and the Microscopy Core of the IDDRC from the NICHD (P30HD004612 and U54HD087101).