Disruption of Trip11 in cranial neural crest cells is associated with increased ER and Golgi stress contributing to skull defects in mice

Abstract

Background:

Absence of Golgi microtubule-associated protein 210 (GMAP210), encoded by the TRIP11 gene, results in achondrogenesis. Although TRIP11 is thought to be specifically required for chondrogenesis, human fetuses with the mutation of TRIP11 also display bony skull defects where chondrocytes are usually not present. This raises an important question of how TRIP11 functions in bony skull development.

Results:

We disrupted Trip11 in neural crest-derived cell populations, which are critical for developing skull in mice. In Trip11 mutant skulls, expression levels of ER stress markers were increased compared to controls. Morphological analysis of electron microscopy data revealed swollen ER in Trip11 mutant skulls. Unexpectedly, we also found that Golgi stress increased in Trip11 mutant skulls, suggesting that both ER and Golgi stress-induced cell death may lead to osteopenia-like phenotypes in Trip11 mutant skulls. These data suggest that Trip11 plays pivotal roles in the regulation of ER and Golgi stress, which are critical for osteogenic cell survival.

Conclusion:

We have recently reported that the molecular complex of ciliary protein and GMAP210 is required for collagen trafficking. In this paper, we further characterized the important role of Trip11 being possibly involved in the regulation of ER and Golgi stress during skull development.

Introduction

Lethal skeletal dysplasia is known as one of the most severe skeletal defects^1,2. Mutations of Golgi microtubule-associated protein 210 (GMAP210), encoded by the TRIP11 gene, result in lethal skeletal dysplasia³. A recent mouse phenotypic analysis shows that GMAP210 is essential for chondrogenesis, but less essential in other skeletogenic cells such as osteoblasts and osteoclasts⁴, suggesting that GMAP210 functions specifically in chondrocytes. However, human fetuses with lethal skeletal dysplasia due to GMAP210 mutations also display severe skeletal defects in the skull vault, where chondrocytes are not present³. This observation suggests that GMAP210 may also play a critical role in osteogenesis during skull development.

The skull consists of several bony plates critical for protecting sensory organs and brain^5,6. The vertebrate skull, consisting of the neurocranium (e.g., skull vault and base) and the viscerocranium (e.g., jaws), is formed from skeletogenic mesenchyme derived from either mesodermal cells or cranial neural crest cells (CNCCs)^7,8. CNCCs are multipotent stem cells that generate skeletogenic precursor cells in a part of the skull vault. Causes of particular craniofacial skeletal malformations can be traced back to problems occurring during specific phases of CNCC development^9,10. Therefore, it is critical to understand details of the molecular mechanism and function of CNCCs in skull formation.

We recently examined the function of GMAP210 in CNCCs¹¹. We found that the molecular complex of ciliary protein intraflagellar transport 20 (IFT20) and GMAP210 regulates skull development¹¹. Importantly, the molecular complex of IFT20-GMAP210 plays a pivotal role in trafficking of type I collagen. Thus, this cellular mechanism is required for smooth collagen trafficking and secreting adequate amounts of collagen matrix essential for skull maturation¹¹. However, it remains unclear whether GMAP210 is solely involved in the regulation of collagen trafficking in osteogenesis during craniofacial bone development.

In this study, we report an additional role of Trip11 in CNCCs derived skull bone. Our study suggests that Trip11 may play a pivotal role in the regulation of ER and Golgi stress during skull vault development.

Results and Discussion

GMAP210 is expressed in cranial neural crest derivatives

It has been reported that GMAP210 is expressed in all cell types such as cartilage and bone in the trunk region³. However, it remains unclear how GMAP210 functions in cranial neural crest cells (CNCCs), a critical cell population for developing craniofacial skeletal structure. To examine GMAP210 expression in CNCCs, we crossed Wnt1-Cre mice with Rosa26-EGFP reporter mice to label CNCCs and performed immunohistochemistry. We found expression of GMAP210 in migratory and postmigratory CNCCs (Fig. 1A). Expression of GMAP210 was also detected in CNCC-derivatives including frontal bone, cartilage primordium of ala orbitalis, nasal septum cartilage, and maxillary and mandibular bones (Fig. 1B). These results suggest that GMAP210 may function to regulate craniofacial bone development in mice.

Fig. 1. GMAP210 expression pattern in neural crest cells.

(A) Whole-mount and immunohistological analysis of Rosa26EGFP:Wnt1-Cre embryo at E9.0. Yellow dashed line in the left panel indicates the orientation of immunohistological analysis. Asterisks indicate the dorsal side. Right panels show the localization of GMAP210 (magenta) in the boxed areas. Right upper panel (a) shows migratory cranial neural crest cells (CNCCs) from dorsal surface of neural tube into the first branchial arch. Right lower panel (b) shows postmigratory CNCCs in the first branchial arch. ba1, first branchial arch; ba2, second branchial arch; e, eye; nt, neural tube. (B) Immunohistological analysis of GMAP210 (magenta) in the wild-type’s frontal bone (a: OF and b: Non-OF position), c: cartilage primordium of ala orbitalis, d: nasal septum cartilage, e: maxillary bone, and f: mandibular bone at E18.5. White dashed lines indicate skull tissue. These are representative images of n = 3 individual embryos in each group. OF, osteogenic front.

CNC-specific disruption of Trip11 results in craniofacial bone defects.

To examine the role of GMAP210, we disrupted Trip11 in a CNC-specific manner using Wnt1-Cre mice (hereafter Trip11 cKO mice). Although the majority of Trip11 cKO mice died after birth¹¹, some survived up to the weaning stage (Fig. 2A). Microcomputed tomography (micro-CT) analysis revealed that CNC-derived bones (e.g., nasal and frontal bones) showed defects of mineralization in Trip11 cKO mice (Fig. 2B). Skeletal staining analysis confirmed that formation of craniofacial bones was severely attenuated in Trip11 cKO mice (Fig. 2C). Superimposition of the lateral photographs further showed the attenuation of craniofacial bone structure in Trip11 cKO mice (Fig. 2D). Next, we wished to examine craniofacial defects during embryogenesis. Whole mount skeletal staining showed that while body size was comparable between wild-type and Trip11 cKO embryos (Fig. 3A), face size was smaller in Trip11 cKO embryos compared with wild-types (Fig. 3B, C). Cephalometric analysis revealed that CNC-derived bones, including nasal-frontal, maxillary and mandibular bones were defective, which was consistent with our experimental strategy to delete Trip11 in a CNC-specific manner (Fig. 3D, E). At this time, the cause of neonatal death observed in the majority of Trip11 cKO mice remains unclear. One possibility is that Trip11 cKO mice may not be able to move their jaw and tongue appropriately due to defective hyoid bones and condylar cartilage formation (Fig. 3F, G). In this case, Trip11 cKO mice presumably could not take milk and/or could not properly breathe (Fig. 3H). Consistent with previous findings^3,4, cartilage components such as cartilage primordium of ala orbitalis, nasal capsule cartilage, nasal septum cartilage and Meckel’s cartilage were also defective (Fig. 4A). Quantification analysis of Safranin O revealed that they were less maturated in Trip11 cKO embryos (Fig. 4B). We confirmed that the amount of type I collagen matrix in frontal bone was decreased in Trip11 cKO embryos compared with controls (Fig. 4C). Together, these results suggest that Trip11 is essential for craniofacial bone and cartilage development.

Fig. 2. Loss of Trip11 in neural crest cells results in craniofacial skeletal defects at postnatal stages.

(A, B) Whole-mount and micro-CT images of control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) mice at postnatal day (P) 20. Arrows indicate the defected frontal bone area. (C) Alcian blue- and Alizarin red-staining of WT and Trip11 cKO mice at P8. Lower panels are high magnification images of upper panels. Arrows indicate defected frontal bone area. Black dashed lines indicate osteogenic front (OF). (D) Superimposition of the lateral photographs orientated by a horizontal plane extending from the lower point of the globe of the eye to upper margin of ear canal, showing the facial profile of WT and Trip11 cKO at P8. cs, coronal suture; fb, frontal bone; ib, interparietal bone; jb, jugal bone; md, mandible; mx, maxilla; nb, nasal bone; pb, parietal bone; OF, osteogenic front.

Fig. 3. Loss of Trip11 in neural crest cells results in craniofacial skeletal defects.

(A) Whole-mount view of Alcian blue- and Alizarin red-staining of control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) embryos at E18.5. (B) Lateral and frontal view of faces at E18.5. (C) Superimposition of the lateral photographs oriented by a horizontal plane extending from the lower point of the globe of the eye to upper margin of ear canal at E18.5. (D) Cephalometric landmarks of lateral and transverse view. Abbreviations: see Table 1. (E) Comparison of craniofacial measurements between WT and Trip11 cKO at E18.5. Abbreviations: see Table 2. (F) Alcian blue- and Alizarin red-staining of WT and Trip11 cKO at birth (P0). Lower panels show high magnification images of the boxed areas in upper panels. Arrow indicates defected hyoid bone in Trip11 cKO mouse. (G) Skeletal staining analysis of the mandibular bone of WT and Trip11 cKO mice at P0. Arrow indicates defective condylar cartilage. (H) Whole-mount view of WT and Trip11 cKO at P0. Right panels show high magnification images of the boxed areas. Arrow indicates a milk band in WT mouse. These are representative images of n=3 individual embryos in each group, and data E are as mean ± SD, n = 3 individual embryos in each group. bo, basioccipital; bs, basisphenoid; eo, exoccipital; hb, hyoid bone; md, mandible; mx, maxilla; p, palatine; pmx, premaxilla; ppmx, palatal process of maxilla; ppp, palatal process of palatine; ptg, pterygoid; tr, tympanic ring.

Fig. 4. Loss of Trip11 induces defective craniofacial bone and cartilage formation.

(A) Safranin O staining of control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) at E17.5. Bottom six panels are high magnification images of the boxed areas in upper panels. ao, cartilage primordium of ala orbitalis; mc, Meckel’s cartilage; nc, nasal capsule cartilage; ns, nasal septum cartilage. (B) Quantification of the ratio of Safranin-O staining positive area from cartilage primordium of ala orbitalis, nasal capsule cartilage, nasal septum cartilage and Meckel’s cartilage (ROI: 0.10mm²). (C) Picrosirius red staining of WT and Trip11 cKO at E18.5. Lower panels are high magnification images of the boxed area in the upper panels with polarized lens. White dashed lines indicate skull tissue. These are representative images of n=3 individual embryos in each group, and data B are as mean ± SD, n = 3 individual embryos in each group.

Loss of Trip11 increases apoptosis in skull tissues.

We have recently reported that proliferation and differentiation activities in CNC-derived osteoblasts are normal in Trip11 cKO mice¹¹. This suggests that the initial process of osteogenesis may not be affected by deletion of Trip11. However, it remains unclear whether disruption of Trip11 affects cell survival and thus causes craniofacial skeletal defects. To examine this possibility, we performed TUNEL analysis at mid-late gestation. While the number of apoptosis positive cells in CNC-derived bone and cartilage were comparable at E15.5 (Fig. 5A), it was increased in the area of CNC-derived frontal bone of Trip11 cKO mice at E18.5 (Fig. 5B). Our quantification analysis showed that increased cell death was observed in the non-osteogenic front area of CNC-derived frontal bone (Fig. 5C). These results suggest that defective cell survival is associated with skull abnormalities in Trip11 cKO mice.

Fig. 5. Loss of Trip11 induces apoptosis during skull development.

TUNEL assay (magenta) and corresponding quantification of coronal sections from control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) embryos at (A) E15.5 and (B) E18.5. Upper panels show low magnification images with anatomical landmarks. Lower ten panels show the high magnification images of the boxed areas in the upper panel focusing on OF position of frontal bone (a, f, k and p; OF position), Non-OF position of frontal bone (b, g, l and q; Non-OF position), cartilage primordium of ala orbitalis (b, g, l and q; ao), maxillary bone (c, h, m and r; mx), mandibular bone (d, i, n and s; md) and nasal septum cartilage (e, j, o and t; ns). White dashed lines indicate the outline of each tissue. (C) Graphs show the quantification of the ratio of TUNEL positive cells in each area (ROI: 0.25mm²) at E15.5 (upper graph) and E18.5 (lower graph). These are representative images of n=3 individual embryos in each group, and data C are as mean ± SD, n = 3 individual embryos in each group. ao, cartilage primordium of ala orbitalis; mc, Meckel’s cartilage; md, mandibular bone; mx, maxillary bone; ns, nasal septum cartilage; OF, osteogenic front.

Loss of Trip11 is involved in the elevation of ER stress in skull tissues.

The endoplasmic reticulum (ER) is a critical cellular organelle required for folding a diversity of proteins. Unfolded and misfolded proteins can accumulate in the ER, causing elevated levels of ER stress^12,13. To avoid excessive accumulation of unfolded proteins, ER stress response machinery is used to promote cell survival^14,15 and alteration of these ER stress responses can cause multiple skeletal defects in humans^{12,13,16–19}. However, it remains unclear whether GMAP210 is involved in the regulation of ER stress response in skull tissues. To address this question, expression of several ER stress markers was examined using control and Trip11 cKO skull tissues. Among tested ER stress markers, sXbp1 expression was moderately reduced and the expression levels of Atf4, Bip and Chop were significantly increased in Trip11 cKO skulls compared with controls (Fig. 6A). Immunohistochemical analysis further confirmed that the number of Bip and CHOP positive cells increased in CNC-derived Trip11 cKO frontal bone (Fig. 6B, C). These are important findings because it has been well characterized that ER stress response genes are frequently upregulated when the ER stress-apoptosis pathway is activated^20–22. Because upregulation of ER stress frequently enlarges the physical space of ER cisternae in order to maintain cellular homeostasis^17,18,23, we examined whether disruption of Trip11 alters ER cisternae structures in the skull. Transmission electron microscopy (TEM) analysis revealed that the size of ER cisternae in Trip11 cKO skull tissues was larger than in controls (Fig. 6D, E). These data demonstrate that Trip11 may be involved in the regulation of ER stress response during skull development.

Fig. 6. Loss of Trip11 is involved in abnormal ER stress in skull tissue.

(A) Quantitative RT-PCR analysis of ER stress markers in control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) using frontal bones at E18.5. (B, C) Immunohistological analysis for BiP or CHOP of WT and Trip11 cKO frontal bones at E18.5. Arrows indicate BiP or CHOP positive cells. White dashed lines indicate skull tissue. Quantification of BiP or CHOP positive cell ratio in the frontal bones at E18.5 (ROI: 0.25mm²). ao, cartilage primordium of ala orbitalis. (D) Transmission Electron Microscopy (TEM) images of frontal bones from WT and Trip11 cKO mice at E18.5. Lower panels show higher magnification images of the boxes in upper panels. Red arrowheads and black arrows indicate swollen ER and Golgi cisternae, respectively. n, nucleus; cy, cytoplasm. (E) Quantification of ER cisternae width, which is measured using TEM images of D. The average width is quantified from 20 cells/200 ER lumens in per skull. ER width is defined as the shorter dimension of the ER cisternal space in TEM images. These are representative images of n=3 individual embryos in each group, and data A, B, C and E are as mean ± SD, n = 3 individual embryos in each group.

Loss of Trip11 is associated with aberrant Golgi stress during skull development.

The Golgi complex plays a pivotal role in regulating transportation of proteins via the selective secretory pathway and some studies demonstrate that the Golgi complex also functions as a sensor of cellular stress^24,25. Although the signal transduction cascade activated by ER stress is well characterized, little is known about whether aberrant Golgi stress response is associated with skeletal defects. To examine this possibility, expression of several Golgi stress markers was examined using control and Trip11 cKO skull tissues. Among analyzed Golgi stress markers^24,26–28, the expression levels of Golga2, Golgb1, Acbd3, Arf4, Cth, Slc7a1 and Slc7a11 were significantly increased in Trip11 cKO skulls compared with controls (Fig. 7A). Western blot analysis further confirmed that the levels of GM130 (Golgi matrix protein 130, part of a cis-Golgi matrix protein) known as a marker of Golgi stress²⁹, was increased in Trip11 cKO compared with wild-type skull (Fig. 7B, C). Because recent studies suggest that increased Golgi stress induces the fragmentation of Golgi and cell death^24,26,27, we hypothesized that aberrant Golgi stress may cause smaller sizes of cis-Golgi in Trip11 cKO skulls. To examine this possibility, the size of cis-Golgi was examined by confocal microscope. The Golgi labeled by GM130, a marker of cis-Golgi, was smaller in Trip11 cKO osteoblasts compared with controls (Fig. 7D, E). Profile plot analysis of cis-Golgi^29,30 confirmed that the size of cis-Golgi was reduced in Trip11 cKO osteoblasts (Fig. 7F, G). Finally, TEM analysis revealed that Golgi stack structures were severely attenuated and swollen Golgi cisternae were presented in Trip11 cKO skull (Fig. 7H). These observations prompt us to hypothesize that GMAP210 may have two cellular functions in skull development. As we have recently reported, GMAP210 is critical for regulating collagen trafficking in the area of the osteogenic front (OF)¹¹. In addition to this unique function, GMAP210 is also required for osteoblast cell survival via orchestrating both ER and Golgi stress. Therefore, loss of Trip11 in CNC-derived osteoblasts may result in both less collagen secretion and increased cell death, leading to skull defects in mice (Fig. 8).

Fig. 7. Loss of Trip11 is associated with aberrant Golgi stress in skull tissue.

(A) Quantitative RT-PCR analysis of Golgi stress markers in control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) frontal bones at E18.5. (B) Western blot analysis for GMAP210 and GM130 using WT and Trip11 cKO frontal bones at E18.5. Skull lysates were prepared from three individual WT and Trip11 cKO embryos (#1, #2, and #3). (C) Quantification of relative protein level of GM130. (D) Immunofluorescent staining for GM130 of control (WT) and Trip11:Wnt1-Cre (Trip11 cKO) osteoblasts isolated from nasal and frontal bones at E18.5. Lower panels show high magnification representative images. (E) Quantification the cis-Golgi size in osteoblasts. The average size of cis-Golgi in each sample is measured from at least 100 osteoblasts, and the measurement is repeated three times using other osteoblasts isolated from 3 individual skulls in each group. (F) Profile plots analysis of GM130 (red), and DAPI (blue) fluorescence. Yellow lines in the lower panels of D indicate where fluorescent intensity is quantified. “G” is defined as cis-Golgi width. (G) Quantification of cis-Golgi width. The average Golgi width is quantified by 30 osteoblasts in each sample, and the measurement is repeated three times using other osteoblast isolated from 3 individual skulls in each group. (H) Transmission Electron Microscopy (TEM) images of frontal bones from WT and Trip11 cKO mice at E18.5. Yellow dashed lines indicate Golgi shape and red arrowheads indicate swollen cisternae. These are representative images of n=3 individual embryos in each group, and data A, C, E and G are as mean ± SD, n = 3 individual embryos in each group.

Fig. 8. A Schematic model of this study.

We previously reported that GMAP210 is critical for regulating collagen trafficking in the area of the OF. In this paper, we demonstrate that GMAP210 is also required for osteoblast cell survival via orchestrating both ER and Golgi stress. Loss of Trip11 in CNC-derived osteoblasts results in less collagen secretion (a) and increased cell death (b), thus leading to skull defects in mice.

Conclusions

Mutation of TRIP11 causes achondrogenesis in humans. These patients also display a lack of mineralization in the skull vault where chondrocytes are not present; however, its molecular etiology remains unclear. Recently, we have reported that lack of Trip11 in CNCCs results in altered collagen trafficking in skull development. In this paper, we further characterized the skull defects observed in CNC-specific Trip11 mutants. While we don’t exclude other causative events which may be involved in the etiology of Trip11 mutants, our results suggest that Trip11 may play a pivotal role in the regulation of ER and Golgi stress, which may be required for proper skull development.

Experimental procedures

Animals

Trip11 floxed mice were generated as previously described⁴. Wnt1-Cre mice were obtained from The Jackson Laboratory³¹. Mice were housed under controlled conditions, namely 21–22°C (+/−0.5°C), 30–75% (+/−10%) relative humidity and 12:12 light-dark cycle with lights on at 7:00 am. Food and water were available ad libitum for all animals. IACUC number was “AWC-18-0137”. Experimental protocol was reviewed and approved by the Animal Welfare Committee and the Institutional Animal Care and Use Committee of the McGovern Medical School at UTHealth. We crossed Trip11^flox/+:Wnt1-Cre (male) with Trip11^flox/flox (female) mice to obtain Trip11 mutants. Through this mating, over 300 embryos and pups were produced and a total of 75 mutants (Trip11^flox/flox:Wnt1-Cre) were harvested at embryonic day (E) 15.5, 17.5, 18.5, newborn, P8 and P20. Because we mainly focused on embryonic stages, sex differences were not tested. Trip11 mutants were harvested from at least 40 different litters.

Skeletal preparations, histological analysis, and fluorescence imaging

Staining of craniofacial tissues with Alizarin red and Alcian blue was carried out as previously described^11,32,33. Cephalometric analysis was performed using the lateral and dorsoventral pictures after skeletal staining. Definitions of cephalometric landmarks and measurements are described in Table 1, 2^34,35. Immunofluorescent staining, H&E staining, Safranin-O staining, and Picrosirius red staining were performed as previously described^33,36,37. TUNEL assay was performed using Click-iT™ Plus TUNEL Assay for In Situ Apoptosis Detection kit (Invitrogen; C10618). Primary antibodies used in immunohistochemistry staining were as follows: BiP (1:100, Santacruz Biotechnology; Sc-13968), CHOP (1:100, Santacruz Biotechnology; Sc-575), EGFP (1:1,000, abcam; ab13970), and GMAP210 (1:200, LS Biosciences; LS-C20059). Stained slides were examined with an Olympus FluoView FV1000 laser scanning confocal microscope using the software FV10-ASW Viewer (version 4.2). For histological and immunohistological analysis, we used coronal sections containing eye, which enabled us to examine CNC-derived frontal bone. A minimum of ten to twelve sections were analysed for each genotyped sample.

Table 1.

Definitions of cephalometric landmarks

N		The most anterior point on the nasal bone
E		The intersection of the frontal bone and floor of anterior cranial fossa
Po		The most posterior and superior point on the skull
Ba		The most posterior and inferior point on the occipital condyle
Co		The most posterior and superior point on the mandibular condyle
Go		The most posterior point on the mandibular ramus
Mn		The most concave portion of the concavity on the inferior border of the mandibular corpus
Me		The most inferior and anterior point of the lower border of the mandible
LI		The most anterior and superior point on the alveolar bone of the mandibular incisor
Mi		The junction of the alveolar bone and the mesial surface of the first mandibular molar
Mu		The junction of the alveolar bone and the mesial surface of the first maxillary molar
Go1		The points on the angle of the mandible that produce the widest width
Go2		The right points on the angle of the mandible that produce the widest width
P1		The most anterior and medial points on the left within the temporal fossae that produce the most narrow palatal width
P2		The most anterior and medial points on the right within the temporal fossae that produce the most narrow palatal width
C1		The points on the cranium on the left that produce the widest cranial width
C2		The points on the cranium on the right that produce the widest cranial width

Table 2.

Cephalometric measurements of the craniofacial skeleton

Neurocranium	Po-N		Total skull length
	Po-E		Cranial vault length
	Ba-E		Total cranial base length
	Po-Ba		Posterior neurocranium height
Viscerocranium	E-N		Nasal length
	E-Mu		Viscerocranial height
Mandible	Co-LI		Total mandibular length
	Co-Me		Length from condylar head to Me
	Go-Mn		Posterior corpus length
	Mi-LI		Anterior corpus length
Transverse mesurements	Go1-Go2		Bigonial width
	C1-C2		Maximum cranial width
	P1-P2		Palatal width

Difinitions of cephalometric landmarks

N The most anterior point on the nasal bone

E The intersection of the frontal bone and floor of anterior cranial fossa

Po The most posterior and superior point on the skull

Micro-computed tomography (μCT) analysis

After fixation, skull samples were scanned with a Scano Viva CT 40 (Scanco Medical, Wangen-Bruttisellen, Switzerland). Voxel size was 15μm×15μm×15μm, X-ray energy was 23μAs and 70kVp, slice thickness was 15μm, and 3-dimensional (3D) reconstructions were performed by IMARIS software (version 9.6).

Isolation of primary osteoblast, and immunocytochemistry

We collected CNC-derived osteoblasts from embryos at E18.5. Nasal and frontal bones were subjected to five sequential digestions with an enzyme mixture containing 1 mg/ml collagenase type I (Sigma-Aldrich, C0130) and 1 mg/ml collagenase type II (Sigma-Aldrich, C6885). Cell fractions (from two to five of the sequential digestions) were collected and cultured in growth medium [alpha-MEM supplemented with 10% (vol/vol) fetal bovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin]. Immunocytochemistry was performed as previously described^11,36. The primary antibody GM130 (1:500, BD Biosciences, 610822) was used. Plot profile was performed³⁰ to measure the size of cis-Golgi using ImageJ software.

Transmission Electron Microscopy (TEM) analysis

Skull tissues were fixed for at least 24 h in Karnovsky’s fixative, decalcified in 10% EDTA for one week, and then refixed for at least 24 h in Karnovsky’s fixative. Samples were then secondarily fixed in 1% osmium tetroxide for one hour, dehydrated in increasing concentrations of ethanol, embedded in epoxy resin, sectioned at 100 nm thickness using an ultramicrotome, stained with uranyl acetate and lead citrate, and imaged at 80 kV using a JEOL JEM-1230 TEM equipped with a Gatan digital camera.

Quantitative real-time RT-PCR

Total RNA was extracted from CNC-derived nasal and frontal bones at E18.5 (n=3 per group) using TRIzol Reagent (Thermo Fisher Scientific; 15596-026). Quantitative RT-PCR was carried out using the SsoAdvanced Universal SYBR Green Supermix (Bio-Rad; 1725274). Conditions for qRT-PCR were 95°C for 2 min, 95°C for 5 sec, and 60°C for 30 sec, for 40 cycles. Sequences of each primer are listed in Table 3. Data were normalized to Gapdh levels and quantified using the 2^−ΔΔCT method.

Table 3.

Primer for sequences for quantitative RT-PCR

Gene names	Protein names	Forward primer (5’-3’)	Reverse primers (5’-3’)
Gapdh	GAPDH	CGTCCCGTAGACAAAATGGT	TCAATGAAGGGGTCGTTGAT
Trip11	GMAP210	GTCTCCAAGCAAAGAGATGAGACT	GTGGTTTTGCTTTTCTAATTCAGC
sXbp1	XBP-1 (spliced)	CTGAGTCCGAATCAGGTGCAG	GTCCATGGGAAGATGTTCTGG
usXbp1	XBP-1 (unspliced)	CAGCACTCAGACTATGTGCA	GTCCATGGGAAGATGTTCTGG
Xbp1	XBP-1 (total)	TGGCCGGGTCTGCTGAGTCCG	GTCCATGGGAAGATGTTCTGG
Atf4	ATF4	GGGTTCTGTCTTCCACTCCA	AAGCAGCAGAGTCAGGCTTTC
Atf6	ATF6	TCGCCTTTTAGTCCGGTTCTT	GGCTCCATAGGTCTGACTCC
Chop	CHOP	CCACCACACCTGAAAGCAGAA	AGGTGAAAGGCAGGGACTCA
Bip	BiP	TTCAGCCAATTATCAGCAAACTCT	TTTTCTGATGTATCCTCTTCACCAGT
Edem	EDEM	CTACCTGCGAAGAGGCCG	GTTCATGAGCTGCCCACTGA
TFE3	TFE3	TGCGTCAGCAGCTTATGAGG	AGACACGCCAATCACAGAGAT
Golga2	GM130	TGTCACAGAACCGAGAGCTGA	GCTGCTCCTGTAGACCCTG
Golgb1	Giantin	GCCTTCACTAAGAGCATGTCAT	GCTGATCCTTTAGAGCAATGCAG
Acbd3	GCP60	GAGGAGCTTTACGGCCTGG	CTTATGCAGTGCCACGAACTT
Arf4	ARF4	CTGGCAAGACGACAATTCTGT	CCACAAAAATGAGACCCTGGGTA
Cth	CSE	TTCCTGCCTAGTTTCCAGCAT	GGAAGTCCTGCTTAAATGTGGTG
Slc7a1	xCT	CTGCCTCAACACCTATGACCT	GAGAGCAGCAATCAAGAAGGAG
Slc7a11	Cat1	GGCACCGTCATCGGATCAG	CTCCACAGGCAGACCAGAAAA

Western blot analysis

CNC-derived osteoblast lysates or skull tissue lysates were prepared in RIPA buffer. After centrifugation at 15,000 g, supernatants were separated by SDS/PAGE, blotted onto a PVDF membrane, analyzed with specific antibodies, and visualized with enhanced chemiluminescence. Antibodies used were as follows: GM130 (1:500, BD Biosciences; 610822), GMAP210 (1:500, LS Biosciences, LS-C20059), and α-tubulin (1:5,000, Abcam, ab7291). The Clarity Max ECL Substrate (Bio-Rad) was used for chemiluminescent detection, and signals were quantified with Image Lab Version 5.0 (Bio-Rad).

Statistical analysis

Statistical analyses were performed using Student’s t tests. P values of less than 0.05 were considered statistically significant. P values were showed in each graph.