Tetraspanin, CD151, is required for maintenance of trigeminal placode identity

Authors


Address correspondence and reprint requests to Marianne Bronner, Division of Biology 139-74, California Institute of Technology, Pasadena, CA 91125, USA. E-mail: mbronner@caltech.edu

Abstract

J. Neurochem. (2011) 117, 221–230.

Abstract

The trigeminal ganglion is the largest of the cranial ganglia and responsible for transmitting sensory information for much of the face. The cell surface glycoprotein CD151 is an early marker of the trigeminal placode, the precursor to the ganglion. Here, we investigate the role of CD151 during specification of trigeminal placode cells in the developing chicken embryo. Expression of the transcription factor Pax3, the earliest known marker of the trigeminal placode, briefly precedes that of CD151, but they then subsequently overlap in the trigeminal placode. Loss of CD151 protein dramatically decreases the number of Pax3+ placode cells in Stage 13–14 embryos, leading to loss of ophthalmic trigeminal neurons by Stages 16 and 17. Although the initial size of the Pax3 population is similar to that in controls, the number of Pax3+ cells decreases with time without alterations in cell death or proliferation. This suggests a role for CD151 in maintenance of the specification state in the trigeminal placode, uncovering the first known role for a tetraspanin in a developmental system.

Abbreviations used
DAPI

4′,6-diamidino-2-phenylindole

GFP

green fluorescent protein

MO

morpholino

PBS

phosphate-buffered saline

PH3

phosphohistone H3

PTW

0.1% Tween-20 in phosphate buffered saline

The trigeminal ganglion is the largest of the cranial ganglia and consists of neurons from two separate sources, the cranial neural crest and the trigeminal placode. Although the role of the cranial neural crest in formation of the peripheral nervous system has been well studied, much less is known about contributions from ectodermal placodes. In the case of the trigeminal placode, several signaling families, including platelet derived growth factors and Wingless and Ints (McCabe et al. 2007; Canning et al. 2008; McCabe and Bronner-Fraser 2008), have been implicated in initial steps of its induction. However, the mechanisms whereby these signals translate to specification events remain unclear. Here, we define specification as the time at which a tissue has the ability to assume a trigeminal fate even when removed from its native state and put into a neutral environment (Slack 1991).

To address the question of how trigeminal placode cells make the transition between induction and specification, we performed a subtractive library screen to reveal genes up-regulated during placode induction (McCabe et al. 2004). One candidate identified in this screen as downstream of inductive interactions was the cell surface glycoprotein CD151, part of the tetraspanin web that provides a scaffold for cell surface signaling (reviewed by Hemler 2005). CD151 is best known as a direct binding partner of integrin α3β1 (Serru et al. 1999; Berditchevski et al. 2001). Additionally, CD151 can facilitate signal transduction through interactions with secondary signaling molecules, including protein kinase C, type II phosphatidylinositol 4-kinase, protein kinase B/conditionally active protein kinase B, Rac, and Cdc42 (Berditchevski and Odintsova 1999; Yauch and Hemler 2000; Yauch et al. 2000; Zhang et al. 2001; Johnson et al. 2009). CD151 plays a role in migration of various cells types including normal leukocytes (Barreiro et al. 2005), endothelial cells (Sincock et al. 1999) as well in several metastatic cancers (Winterwood et al. 2006; Yang et al. 2008; Zijlstra et al. 2008; Hirano et al. 2009; Johnson et al. 2009). Knock-out mouse studies of CD151 show no obvious phenotypes in the developing embryo (Wright et al. 2004; Takeda et al. 2007), but exhibit kidney failure, reduced platelet function, extended bleeding times, greater blood loss, and pathological angiogenesis (Lau et al. 2004; Wright et al. 2004; Sachs et al. 2006; Takeda et al. 2007; Geary et al. 2008). Interestingly, mutations in the human CD151 can lead to defects including kidney failure, skin blistering, and sensorineural deafness (Karamatic Crew et al. 2004). In other cell types, CD151 has been shown to increase cell adhesion (Fitter et al. 1999; Chattopadhyay et al. 2003; Sawada et al. 2003; Shigeta et al. 2003; Nishiuchi et al. 2005; Johnson et al. 2009). CD151 functions in a cell context dependent manner as a molecular facilitator of differentiation, cell shape change, and other processes (reviewed by Tarrant et al. 2003; Hemler 2005; Zoller 2009). However, nothing is known about the role of CD151 during development. Because of its diverse functions and importance in context specific roles in other systems, CD151 presents an interesting candidate for further exploration in the developing trigeminal placode.

In this study, we investigate the role of CD151 in the process of specification of trigeminal placode cells in the developing chicken embryo. Onset of CD151 expression follows but then overlaps with that of Pax3, the earliest known marker of trigeminal placode specification. Furthermore, down-regulation of CD151 protein by antisense oligonucleotide morpholinos (MOs) dramatically decreases the number of Pax3+ placode cells and subsequently ophthalmic trigeminal neurons. The results reveal for the first time an important developmental role for CD151, suggesting that it may be important for the maintenance of ophthalmic trigeminal placode specification.

Materials and methods

In situ hybridization

Fertilized chicken (Gallus gallus domesticus) embryos were fixed overnight in 4% paraformaldehyde. Antisense digoxigenin-labeled RNA probes were produced according to manufacturer’s directions (Roche Molecular Biochemicals, Indianapolis, IN, USA). Whole mount in situ hybridization was performed as previously described (Kee and Bronner-Fraser 2001) with Nitro blue tetrazolium chloride/5-bromo-chloro-3-indolyl phosphate, toluidine salt (Roche) for color detection. Whole mount embryos were subsequently cryoprotected, embedded in gelatin, cryosectioned at 20 μm and digitally photographed on a Zeiss Axioskop2 microscope with an Axiocam HRc camera (Carl Zeiss Inc., USA).

Morpholinos

Translation-blocking antisense oligonucleotide MOs were designed to the 5′-UTR of CD151 (Geneprobes) with 3′-end modification of carboxyfluoroscein. CD151MO and CD151MO2 were designed to the area in parentheses with the translation start site in bold:

CD151MO: 5′-GAGTGCTTTGGTAGGGAAAGGAATTTGCT(AA GCCAGAAAGATGCGTGAATATAC)TGAGAAGAAAG-3′

CD151MO2: 5′-(GAGTGCTTTGGTAGGGAAAGGAATT)TGCT AAGCCAGAAAGATGCGTGAATATACTGAGAAGAAAG-3′. Notice that CD151(2) MO does not bind to the start site (bold), making it a weaker MO (personal communication with manufacturer).

Standard control MO with 3′-end modification of carboxyfluoroscein was utilized: 5′-CCTCTTACCTCAGTTACAATTTATA-3′.

Morpholinos (1 mm) along with carrier or rescue DNA (2 μg/μL), were dissolved with sterile 10 mm Tris pH8.0, with the exceptions of the CD151(2) experiments when 1.5 mm was used. Carrier DNA has been shown to improve the transfection efficiency of FITC-labeled MOs (Voiculescu et al. 2008). For carrier DNA, the DNA-encoding green fluorescent protein (GFP) was removed from the pCIG expression construct (Megason and McMahon 2002). For rescue experiments, full length CD151 cDNA that was mutated around the start site so that it was resistant to CD151 MO (lower case mutated base pairs, Acc ATG aGa Gag TAc). CD151 (255–1013 bp, accession number NM_001006472) was then subcloned with pCIH2B-GFP whose expression was driven by the chicken β-actin promoter and the cytomegalovirus enhancer (generous gift from Dr Sauka-Spengler). GFP-H2B was driven from an internal ribosome entry site sequence downstream of coding sequence which allowed detection of transfected nuclei. Cells that were transfected with MO alone may have punctate FITC, whereas those cells that are co-transfected with CD151-pCIH2B-GFP will have strong nuclear GFP in addition to the punctate FITC from the MO.

In ovo electroporations

Embryos were incubated at 37°C until desired stage (3–7ss). A mixture of MO and carrier DNA was injected underneath the vitelline membrane but on top of the ectoderm using a glass needle to target the presumptive trigeminal placode at Stage 8 to middle Stage 9 (3–7ss). Embryos were then electroporated vertically with five pulses at 8 V for 50 ms at 100 ms intervals as previously described (Shiau et al. 2008). Embryos were then resealed and incubated at 37°C until desired stage (Stages 10–16). Embryos fixed overnight in 4% paraformaldehyde. Electroporation efficiency of the embryos was visualized by using dissecting fluorescent Zeiss Stereo microscope and embryos were photographed using the Axiocam mRm camera.

Immunostaining sections and cell counting

Embryos were cryoprotected, embedded in gelatin and cryosectioned at 10 μm. Gelatin was removed from sections by incubating in phosphate-buffered saline (PBS) at 42°C for 10 min. Sections were blocked at room temperature (25°C) with 5% Sheep serum in 0.1% Tween-20 in phosphate buffered saline (PTW) for 1 h. Primary antibodies in blocking solution were incubated overnight at 4°C at the following concentrations: mouse Pax3 (Developmental Hybridoma Bank, Iowa City, IA, USA) 1 : 10; mouse Islet1 supernatant (Developmental Hybridoma Bank) 1 : 100; rabbit FITC (Invitrogen, Carlsbad, CA, USA) 1 : 100; rabbit Phosphohistone H3 (PH3; Upstate Biotechnology, Lake Placid, NY, USA) 1 : 250. Sections were then rinsed in PBS three times for at least 15 min each, and blocked for at least 30 min prior to secondary antibody application. Secondary antibodies in blocking solution were incubated for 1.5 h at 25°C. Secondary antibodies (goat anti-mouse IgG Alexa 568, goat anti-rabbit IgG Alexa 488, goat anti-rabbit IgG Alexa 647) were used at 1 : 1000 except 1 : 250 for Alexa 647 (Molecular Probes, Eugene, OR, USA). Sections were rinsed in PBS three times for at least 15 min each. Sections were then exposed 4′,6-diamidino-2-phenylindole (DAPI) (2 μg/mL) to reveal nuclei and rinsed in PBS before coverslipping.

Trigeminal placode cells were counted as previously described (McCabe and Bronner-Fraser 2008) from the rostral midbrain where the notochord begins to the rostral hindbrain where the neural tube changes from elongated to more round. Only those embryos were analyzed that had high electroporation efficiency over the presumptive trigeminal placode region in whole mount view as determined visually on a dissection fluorescent microscope (for example, Fig. 1a, f, and k). As a consequence, the vast majority of cells had high levels of FITC and all Pax3+ cells were counted. The FITC-labeled MO was visible throughout the cytoplasm, making it difficult to resolve individual cells. In cross-sections from the rostral midbrain level ectoderm to the rostral hindbrain level ectoderm, all Pax3+ cells were counted on one side of the embryo, regardless of lateral extent of the placode, which generally extended to the midpoint of the dorsoventral axis. The average number of cells per sections was calculated for individual embryos. All control values were then normalized to 100% and CD151 morphant cell counts were expressed as a percentage of control. Raw cell counts are provided in the text as a supplement to the normalized data presented in graphs. Student’s t-test were performed when comparing two groups. When three or more groups were compared, anova was performed with the Bonferonni-Dunn post hoc test. A p-value of < 0.05 was considered statistically significant.

Figure 1.

 CD151 is required for Pax3 and maintenance of trigeminal placode specificiation. (a–e) Control morpholino (MO) electroporated in midbrain level trigeminal placode at Stage13/14. (f–j) CD151 MO decreases Pax3 expression (arrowhead) in trigeminal level ectoderm at Stage13/14, but the morpholino does not block all Pax3+ cells (arrow). (k–o) CD151 MO reduction of Pax3 can be rescued by over-expressing resistant CD151 plasmid. Whole mount view of electroporated embryo (a, f, k). Pax3 expression in trigeminal placode cross-section of midbrain level ectoderm (arrow, red, b, g, l). FITC-morpholino detection (green, c, h, m). DAPI nuclei detection (d, i, n). Merge of Pax3, FITC, and DAPI (e, j, o).

Immunostaining whole mount and cell counting

Embryos were blocked using 5% Sheep serum in PTW for 3 h at 25°C. Primary antibodies in blocking solution were incubated overnight at 4°C (concentrations above). Embryos were then rinsed in PBS for three times for at least 1 h each. Embryos were blocked using 5% Sheep serum in PTW for 1 h at 257°C. Secondary antibodies (described before) in blocking solution were incubated overnight at 4°C at a concentration of 1 : 250. Embryos were then rinsed in PBS for three times for at least 1 h each. Embryos were photographed prior to confocalling using a Zeiss Axioskop2 microscope using the Axiocam mRm camera. Embryos were subsequently analyzed by confocal microscopy using a Zeiss Pascal Exciter with the Multitime macro. To capture the entire trigeminal ganglion, images were collect with a 20× objective and tiled with at least two 20× fields. Z stacks of 3.75 μm with a 5 μm pinhole were captured and aligned by Zeiss LSM software. The ophthalmic lobe trigeminal ganglion cell number was analyzed using Imaris imaging software (version 5.7.2; Bitplane Scientific Software, Saint Paul, MN, USA). Both transfected and untransfected Islet + cells were counted.

Results

CD151 is expressed after trigeminal placode marker, Pax3

CD151 was identified in a subtractive library screen looking for candidate genes up-regulated in response to trigeminal placode induction (McCabe et al. 2004). As a first step in establishing its possible function in downstream events, such as specification, migration, or neurogenesis, we have previously examined its distribution pattern in the head of the developing chick embryo. CD151 mRNA is first detected at low levels in Stage 9 embryos in the ophthalmic (opV) portion of the trigeminal placode, increasing to high levels by Stage 10 (McCabe et al. 2004). Interestingly, the timing of CD151 expression in the trigeminal placode slightly follows onset of the trigeminal placode specification marker Pax3, which can be detected in a few trigeminal placode cells beginning at early Stage 9, and is clearly detectable throughout the placode by late Stage 9 (Baker et al. 1999). At Stage 10, CD151 and Pax3 expressions overlap in trigeminal placode cells (McCabe et al. 2004; McCabe and Bronner-Fraser 2008). Note that CD151 is specific to trigeminal placode cells in the midbrain level ectoderm, whereas Pax3 mRNA expression can also be detected in the dorsal neural tube and the early migrating neural crest. As the trigeminal placode cells become more prominently spaced apart at Stage 11, CD151 and Pax3 mRNA expression overlap. This co-expression pattern is maintained in the placode until at least Stage 15, where Pax3 can be seen in migrating trigeminal placode cells in the mesenchyme, whereas CD151 is only expressed in the placode (McCabe et al. 2004). From the expression patterns and timing of CD151 and Pax3 mRNAs, we conclude that CD151 is a specific trigeminal placode marker at the midbrain level ectoderm that is expressed slightly after the transcription factor Pax3.

CD151 is required for Pax3 expression in the trigeminal placode

To determine its role in the developing trigeminal placode, antisense MO oligonucleotides were designed against the translation start site of CD151. Embryos were electroporated prior to the onset of CD151 expression, at Stage 8 to middle Stage 9, and collected at Stages 13 and 14 to determine effects on the trigeminal placode formation. Interestingly, we observed an obvious loss of Pax3+ expression in trigeminal placode cells in CD151 morphant knock-down. Embryos were analyzed for Pax3 expression in the trigeminal placode. In cross-sections of the midbrain level ectoderm, a dramatic reduction in Pax3+ cells can be detected in CD151 morphant embryos (Fig. 1g–j) compared with controls (Fig. 1b–e). The number of Pax3+ cells in the trigeminal placode at the midbrain level were counted and then normalized to control embryos. There was a 40% reduction in the number of Pax3 cells compared with a control MO (Fig. 2a note data presented in graph is expressed as a normalized percentage of control) (Control MO: 33.13 ± 2.6 cells/section, CD151: 21.09 ± 2.1). This result was verified by testing a second CD151 MO designed against another region 5′-UTR of CD151 that does not overlap with the translation start site or the original CD151 MO. Similar to the original CD151 MO, this second MO resulted in a loss of Pax3 expression, albeit less efficiently; it yielded a 20% reduction in the numbers of Pax3 expressing cells compared with control MO-treated embryos (Control: 100.0% ± 7.545, N = 8; CD151MO(2): 79.87% ± 3.820, N = 7; p = 0.0405).

Figure 2.

 CD151 morpholino (MO) blocks Pax3 expression in trigeminal placode. (a) CD151 MO results in 40% decrease in number of Pax3+ cells when electroporated at Stage 8 and collected at Stage 13/14 (Control MO: N = 9, CD151 MO: N = 8; p = 0.0002, Student’s t-test). (b) CD151 MO-resistant plasmid can rescue loss of Pax3 in presence of CD151 MO (Control MO: N = 11, CD151: N = 12, CD151 rescue N = 10, p < 0.0001 for Control MO vs. CD151 MO and CD151 MO vs. CD151 rescue, no significant difference between Control MO and CD151 rescue, anova with Bonferonni-Dunn post hoc test). (c) Pax3+ cells are generated in presence of CD151 MO if collected at Stage 10 (Control MO: N = 7, CD151 MO: N = 5, p = 0.4601). All error bars are standard error of the mean. *Significant p-value (a, p = 0.0002; b, p < 0.0001). TGP, trigeminal placode.

As an additional control for the specificity of the CD151 MO, a CD151 expression construct was generated to rescue the effect of the CD151 MO in a separate set of experiments (Figs 1l–o, 2b). The 5′-UTR upstream of the start site was mutated so that the CD151 MO could not block translation of the exogenous gene product. Over-expression of the resistant CD151 plasmid resulted in near control levels of Pax3 expression in the trigeminal placode (Fig. 1l–o). When quantified, CD151 rescue was returned to 87% of control values, and not statistically different from controls (p > 0.05, anova with Bonferonni-Dunn post hoc test). As with the previous experiments, CD151 MO resulted in approximately 40% reduction in Pax3 expression (Fig. 2b) (Control: 46.02 ± 2.5 cells/section, N = 11 embryos; CD151: 28.84 ± 2.1, N = 12; CD151 rescue: 39.03 ± 1.7, N = 10 embryos). These results suggest that CD151 plays a critical role in trigeminal placode specification.

CD151 is required for maintenance of trigeminal placode induction

To determine when the loss of Pax3+ placodal cells occurs, embryos were electroporated at Stages 8 and 9 and collected at various intermediate stages between Stages 10 and 12. Interestingly, at Stage 10, concomitant with the onset of strong CD151 mRNA expression, we observed no significant difference in the number of Pax3+ trigeminal placode cells between control and CD151 morphant embryos (Fig. 2c) (Control: 19.44 ± 2.4 cells/section, N = 7 embryos; CD151: 22.53 ± 3.3, N = 5 embryos). By intermediate Stages of 11 and 12, there is about a 25% reduction, albeit not statistically significant, in the numbers of Pax3+ cells (data not shown). Thus, the phenotype begins to manifest around the onset of the first migrating trigeminal placode cells at Stage 11, by which time the majority of trigeminal placode cells have been specified (Baker et al. 1999). Significant differences only become readily apparent at Stages 13 and 14.

Loss of CD151 does not alter cell death or proliferation rates

Several possible explanations could account for the observed 40% decrease in numbers of Pax3+ expressing trigeminal placode cells in CD151 morphant embryos between Stages 10 and 14. Pax3 expressing cells may fail to be initially specified but subsequently may die. Alternatively, they may initially express Pax3, but then fail to maintain its expression. In addition, they may fail to proliferate, but this seems the least likely since previous experiments suggest that the majority of Pax3 expressing cells are already post-mitotic neuroblasts (McCabe et al. 2009).

To distinguish between these possibilities and address what accounts for the decrease in Pax3+ cells, we examined whether there were alterations in cell proliferation and/or cell death. To quantify the amount of cell death, we performed TUNEL staining on control and CD151 morphant embryos. The number of Pax3+ cells and total number of cells by DAPI were counted within the MO-electroporated ectoderm and data were expressed as the percentage of dying cells within the ectoderm. The results failed to reveal a significant difference in the number of dead cells between control and CD151 morphant embryos, making cell death unlikely to account for the loss of Pax3+ cells (Fig. 3a) (Control MO: 1.96 ± 0.37 cell/section, CD151 MO: 2.34 ± 0.20). Similarly, there was no significant difference in the number of dividing cells as determined by PH3 protein expression or total number of cells transfected between the two groups, as revealed by DAPI/FITC (Fig. 3b–c) (3B PH3: Control MO – 1.07 ± 0.15 cells/section, CD151 MO – 1.035 ± 0.11; 3C DAPI: Control MO – 49.80 ± 3.6 cells/section, CD151 MO – 55.06 ± 3.1). Therefore, the loss of Pax3+ cells after CD151 MO treatment is likely because of a loss of maintenance of specification of the trigeminal placode cells rather than an increase of cell death or a decrease in proliferation of trigeminal placode progenitors.

Figure 3.

 CD151 morpholino (MO) does not alter cell death or proliferation rates. (a) CD151 does not significantly increase the percentage of Tunel+/transfected (FITC) in embryos electroporated at Stage 8 and collected at Stage 13/14 (Control: N = 10, CD151 MO: N = 8, p = 0.7733). (b) Proliferation rates as determined by Phosphohistone H3 (PH3) are not altered in the presence of CD151 morpholino (Control MO: N = 8, CD151: N = 8, p = 0.8448). (c) Total cell number (DAPI) does not change in the presence of CD151 MO (Control MO: N = 10, CD151: N = 8, p = 0.1602).

CD151 MO reduces neurons in trigeminal ganglion

We next tested whether the reduction in the number of Pax3+ cells in CD151 morphant embryos affects subsequent neurogenesis of the trigeminal opV ganglion formation. To this end, embryos were electroporated at Stages 8 and 9 with control or CD151 MOs and incubated for 48 h to Stage 16 and 17. To assess the degree of neurogenesis in the opV branch of the trigeminal ganglion, morphant embryos were immunostained in whole mount using Islet1 to mark the nuclei of trigeminal placode-derived neurons. At this stage, Islet1 is specific to placode-derived trigeminal neurons, as the neural crest-derived portion of trigeminal neurons do not differentiate and express neuronal markers until after Stage 18 (D’Amico-Martel and Noden, 1980). Embryos were imaged by confocal microscopy in which the opV lobe was captured by performing tiling of multiple 20× fields and z-sectioned so that the entire opV lobe could be digitally reconstructed. To analyze the total number of opV neurons, the optically reconstructed opV lobe was analyzed using Imaris imaging software to count the total number of Islet+ cells and Islet+/ MO+ cells.

The results reveal a dramatic decrease of opV neurons in CD151 morphant embryos compared with controls (Fig. 4a–f), seen both by traditional fluorescent microscopy (Fig. 4a–f) and high resolution confocal microscopy plus digital reconstruction (Fig. 4g–n, Figure S1a–b). The reduction of CD151 protein by the MO reduces the number of opV neurons that are Islet/FITC+ by approximately 47% of control (Fig. 4i) (Control MO: 218.2 ± 22.78 Islet/FITC+ cells/opV ganglion, CD151 MO: 120.4 ± 20.53). There is a 33% decrease in the number of Islet+ neurons, indicating that there is a small, but not significant difference in the percentage of transfected neurons (Fig. 5b) (Control: 100.0% ± 5.427, N = 9; CD151: 80.8% ± 7.514, N = 9; p = 0.0548). Given that the CD151 MO-treated embryos fail to maintain Pax3 expression in the placodal ectoderm, it is likely that the failure in maintenance of specification leads to failure of neuronal differentiation and a resultant reduction in the number of Islet+/CD151MO+ cells. These data support the idea that CD151 acts a molecular facilitator of signal transduction such that signaling necessary for specification is significantly less efficient in its absence.

Figure 4.

 CD151 morpholino (MO) blocks neurogenesis in forming trigeminal ganglion. (a–c) Whole mount view of control embryo has normal formation of ophthalmic lobe of trigeminal ganglion as visualized by neuronal marker Islet1 (red) and transfected with Control MO (green). (d–f) Whole mount view of CD151 MO treated embryo has far fewer neurons (Islet, red) electroporated with FITC-MO (green). (g–i) Confocal projection of entire control opV ganglion illustrates normal formation. (j–l) Confocal project of entire CD151 opV ganglion depicts an opV lobe that has fewer Islet+ cells, appears thinner, and less organized compared with controls. (m) Single confocal cross-section take at middle depth shows the core of the opV in a control MO embryo. (n) Single confocal cross-section taken at the same depth as (m) indicate that there are fewer Islet+ cells and is unorganized when CD151 MO is present. Embryos were electroporated at Stage 8 and collected at Stage16/17.

Figure 5.

 Neurogenesis is reduced by CD151 morpholino. (a) CD151 MO results in approximate 47% reduction in morpholino transfected Islet+ neurons (Control: N = 9, CD151 MO: N = 9, p < 0.0001). (b) A slight, but non-significant decrease in transfection efficiency of Islet1+ neurons was seen with CD151 MO (Control MO: N = 9, CD151 MO: N = 9, p = 0.0548). All error bars are standard error of the mean. *p < 0.0001. opV, ophthalmic trigeminal lobe of ganglion.

Discussion

In vertebrates, an intact and functional peripheral nervous system is critical for survival. At cranial levels, the trigeminal ganglion is responsible for transmitting sensory information about temperature, pain, and touch for much of the face (reviewed by Baker and Bronner-Fraser 2001). This in turn protects the animal from injuring the eyes, mouth and facial skin. Understanding initial formation of the peripheral nervous system is a key for replacing/repairing missing or malfunctioning neurons in the case of traumatic injuries as well as in trigeminal neuralgia.

The work presented helps form a bridge in the understanding of the processes that occur between induction and specification of trigeminal placode cells. Developmental biological processes and the molecules that promote them are generally studied from the perspective of individual molecules. However, in the developing embryo, the various molecules and developmental steps occur in a multidimensional fashion to orchestrate the formation of a functioning embryo. Therefore, it is critical to find the connection points between individual genes and single signaling pathways. We believe that CD151 may be an example of an important facilitator in signal transduction during formation of the trigeminal ganglion. After MO-mediated knock-down of CD151 protein, the trigeminal placode marker Pax3 is dramatically reduced as a function of time as is the subsequent differentiation of Islet1+ neurons in the ophthalmic lobe of the trigeminal ganglion. However, Pax3+ cells are generated properly at initial stages (Stage 10), but then many stop expressing Pax3. Because the CD151-morpholinated cells do not appear to be dying or proliferating at a slower rate, the results suggest that CD151 is critical for the maintenance of specification of the trigeminal placode. The data suggest that CD151 plays an important role in maintenance of Pax3 expression in some, but perhaps not all, the placodal cells.

Specification as defined by Slack (1991) has occurred in a tissue when the fate of the cells does not change after transplantation from its native state in the embryo into a neutral environment. The earliest known marker for ophthalmic trigeminal (opV) placode is the paired homeodomain transcription factor, Pax3, whose mRNA can be detected as early at Stage 8 and protein detected at Stage 9 (Stark et al. 1997; Baker et al. 1999). Baker et al. (1999) have shown that the specification of the ophthalmic trigeminal (opV) placode correlates with expression of Pax3. When explants of trigeminal level ectoderm were removed from embryos at 3–13ss (Stages 8–11) and cultured in a collagen gel in the absence of serum, approximately half of the explants contained Pax3+ at the 8–9ss (Stages 9 and 10), and 100% by the 10–13ss embryos. Furthermore, commitment to trigeminal placode fate appears to occur after 8ss (Stage 9). Although the majority of opV placode cells are specified by Stage 10, trigeminal placode cells continue to be born and ingress into the mesenchyme until Stage 21 (D’Amico-Martel and Noden 1983). Taken together, these previous results suggest that once Pax3 is expressed in trigeminal level ectodermal placode cells, they become specified. However, in CD151 morphant embryos, there is a failure to maintain Pax3 expression, and consequent inability to differentiate into Islet1+ cells in the opV lobe.

The tetraspanin super family member CD151 was isolated as a candidate for involvement in placode development from a library screen looking for genes up-regulated after trigeminal placode induction. CD151 is a cell surface glycoprotein that is part of the tetraspanin web that facilitates signal transduction. Because specification of the trigeminal placode is likely to be a multi-factorial process, CD151 has the potential to mediate signal transduction of several growth factor signaling families to facilitate specification. This could occur directly by a previously unknown interaction between CD151 and the various growth factor receptors or through known interactions with protein kinase C, type II phosphatidylinositol 4-kinase, protein kinase B/conditionally active protein kinase B, Rac, and Cdc42 (Berditchevski and Odintsova 1999; Yauch and Hemler 2000; Yauch et al. 2000; Zhang et al. 2001; Johnson et al. 2009). Utilizing an RT-PCR screen for receptors of well-known growth factors expressed during specification (McCabe et al. 2007), we find that placode cells express receptors for fibroblast growth factor (FGF), insulin-like growth factor, transforming growth factor beta, Sonic Hedgehog, and Wnt families. To date, both Wnt and FGFs have been implicated in the specification of the trigeminal placode (Lassiter et al. 2007, 2009; Canning et al. 2008). Interestingly, when Wnt signaling is blocked by using a dominant negative T cell factor construct, the trigeminal placode cells were either not formed or not maintained (Lassiter et al. 2007). Canning et al. (2008) argue that FGF and Wnt are both required for the specification trigeminal placode cells and subsequent neurogenesis. It is intriguing to speculate that CD151 may play a role in regulating the signaling of Wingless and Ints and FGFs and/or other signaling pathways during the specification of the trigeminal placode.

Beside their well-known function as molecular facilitators, an additional and important function of tetraspanin family members recently has been discovered in intracellular communication through the secretion of exosomes. Exosomes are small vesicles that contain internalized membrane proteins that are subsequently released extracellularly under normal and pathological conditions. The exosomes frequently contain proteins, mRNA and miRNA which can change the protein production of the receiving cell (reviewed by Zoller 2009). Tetraspanin–tetraspanin complexes are known to be components of exosomes (Hemler 2003; Pols and Klumperman 2009). In particular, CD151 has been shown to be critical for internalization of another tetraspanin which can increase migration in cancer cell lines (Rana et al. 2010). This new potential role of CD151 in exosomes leads to the intriguing idea that the role of CD151 in the maintenance of Pax3 expression may be because of coordination of extracellular signals in the tetraspanin web. Alternatively, it may play a role via exosomes, of changing the protein production of nearby cells. For the present results, we cannot distinguish between possible cell autonomous effects through the tetraspanin web or intracellularly through exosomes. Future experiments involving the careful molecular dissection of these potential roles are required to distinguish between these possibilities.

CD151 protein–protein interactions have been extensively studied in normal and tumor cell lines (reviewed by Hemler 2005, 2008). The experiments presented here represent a much needed link between the functional actions of the molecule CD151 described in culture and the role it plays in coordinating signaling in a definable and measurable process in a developing embryo. At least two signaling families, Wnt and FGF receptors, are immediate candidates for interacting with CD151.

Acknowledgements

We would like to thank Samuel Ki for his technical assistance. This work was supported by DE16459 to MBF. The authors have no competing interests.

Ancillary