Microstructure changes and miRNA-mRNA network in a developmental dysplasia of the hip rat model

Summary

MicroRNAs (miRNAs) interact with mRNAs in various pathophysiological processes. In developmental dysplasia of the hip (DDH), the miRNA-mRNA pairs affecting acetabular cartilage (AC) development remain unknown. We investigated dynamic microstructure changes and mRNA and miRNA expression profiles in the AC proliferative zone in a DDH rat model. Abnormal chondrocyte proliferation was observed, and several differentially expressed mRNAs and miRNAs were identified. Downregulated mRNAs and target genes of upregulated miRNAs were primarily enriched in bone and cartilage development. Six hub genes were identified using the predicted miRNA-mRNA interaction network and gene expression pattern analysis. The expression levels of these hub genes and paired miRNAs aligned with our predictions, and most of the pairs were significantly negatively correlated. Excessive chondrocyte proliferation in the AC proliferative zone can delay AC ossification, which might be crucial to DDH development. Specific miRNA-mRNA interaction pairs may serve as diagnostic biomarkers and therapeutic targets.

Graphical abstract

Highlights

• •Excessive proliferation of chondrocytes in the proliferative zone can lead to DDH
• •Abnormal chondrocyte proliferation is associated with multiple genes and pathways
• •The hub genes in DDH are potentially regulated by miRNAs in an epigenetic manner

Cell biology; Developmental biology; Transcriptomics; Model organism

Introduction

Developmental dysplasia of the hip (DDH), formerly known as “congenital dislocation of the hip,” is a primary cause of early-onset osteoarthritis in adults.¹ DDH is among the most common skeletal abnormalities in children, affecting 1.5–20 in 1,000 newborns, with the exact prevalence differing slightly among various geographical locations.^2,3 As a complex three-dimensional deformity, DDH presents as an abnormal relationship between the femoral head and acetabulum, including a broad spectrum of anatomical malformations ranging from mildly dysplastic to irreducible dislocation.⁴ Although the pathogenesis of DDH is not fully understood, known risk factors include genetic predisposition, breech presentation, female sex, high birth weight, and oligohydramnios.⁵ Moreover, several DDH-associated genes have been identified through genetic association studies and mutation screening experiments^6,7,8; however, the concordance rates between identical and dizygotic twins are 41% and 3%, respectively, highlighting the roles of epigenetic and environmental factors in the pathogenesis of DDH.⁹ Nevertheless, the epigenetic changes caused by environmental factors contributing to DDH remain unclear. Consequently, exploring the influence of environmental factors on epigenetic variation may provide valuable insights into unraveling the pathophysiological mechanisms and molecular processes underlying DDH.

After birth, the cup-shaped rim of the acetabulum develops superiorly from the ilium hyaline cartilage, posteriorly from the ischium, and anteriorly from the pubis. The hyaline cartilage comprises the growth plate cartilage on the bony pelvic side and the articular cartilage on the femoral head side.¹⁰ The growth plate cartilage plays an important role in endochondral ossification, and delayed ossification in the acetabular cartilage has been associated with DDH development.¹¹ The growth plate can be functionally divided into resting, proliferative, pre-hypertrophic, and hypertrophic zones; the proliferative zone height is related to the growth rate.¹²

Swaddling models, widely employed to investigate the molecular biological changes in DDH,^13,14 are induced by environmental factors and provide substantial insights. Given the similar arrangement of chondrocytes in the acetabular and growth plate cartilages, we hypothesized that chondrocytes in the proliferative zone of the acetabular cartilage play a crucial role in the development of DDH. Although some of the genes and signaling pathways necessary for DDH development have been identified,^11,15 a global examination of chondrocyte gene expression changes in the acetabular cartilage is still lacking. Therefore, precisely dissecting the changes occurring in the chondrocytes from those occurring at the acetabular cartilage proliferative zone is important for studying their unique expression profiles and improving our fundamental understanding of DDH development.

Laser capture microdissection (LCM) is a technique that targets the extraction of cell populations or single cells for downstream analyses such as real-time PCR or RNA sequencing (RNA-seq).¹⁶ Recently, Lui et al.¹⁷ generated a microanatomical atlas of gene expression in different chondrocyte zones of normal growth plates using RNA-seq based on LCM. Broad-based approaches for the global examination of gene expression profiles have been instrumental in deciphering the molecular mechanisms underlying growth plate development and identifying novel signaling pathways; however, similar studies under pathological conditions, including DDH, are lacking.

RNA-seq is a sensitive and precise approach for studying gene function in almost all eukaryotes. RNA-seq can accurately quantify gene transcript levels and their isoforms across a dynamic range and is particularly useful for detecting low-abundance transcripts.¹⁸ Thus, an increasing number of bone and cartilage development studies have focused on the gene expression profiles in chondrocytes in growth plates,^17,19 paving the way for further molecular characterization and functional and pathological analyses. MicroRNAs (miRNAs) comprise a highly conserved class of non-coding RNAs that function by annealing primarily to the 3′-untranslated regions of target mRNAs to silence or suppress gene expression at the post-transcriptional level, which is an important epigenetic mechanism.²⁰ Moreover, miRNAs are intrinsic regulators in chondrocyte proliferation, differentiation, and endochondral ossification.^21,22

Based on this background, the aim of the present study was to explore the mechanisms underlying delayed endochondral ossification in the acetabular cartilage in DDH using a rat swaddling-induced DDH model, which has been widely employed to investigate the molecular biological changes induced by environmental factors, providing substantial insights into the pathogenesis of DDH. Chondrocytes were obtained using LCM from the proliferative zones of the acetabular cartilage in the rat DDH model, which were subjected to genome-wide mRNA and miRNA sequencing with the goal of identifying core differentially expressed mRNAs and miRNAs.

Results

The chondrocytes in the proliferative zone of the acetabular cartilage in DDH rats show increased proliferation and decreased hypertrophic differentiation

In the DDH group, gross anatomical observations revealed gradual thickening of the acetabular cartilage. No substantial gross morphological differences were observed in the acetabulum five days after modeling. However, the acetabular fossa became shallow, and the femoral head completely lost contact with the acetabulum and formed a secondary acetabulum (Figures 1A1–D1) 10 days after modeling.

Figure 1.

Histological observations and laser capture microdissection days after modeling

The developmental dysplasia of the hip (DDH) model was generated using newborn Wistar rats, and changes between normally developed (A1–4; C1–4) and DDH (B1–4; D1–4) rats were compared after 5 (A1–B4) and 10 (C1–D4) days.

(A1–D1) General appearance of the gross specimen, ilium (Il), ischium (Is), and femoral head (Fh). A dysplastic true acetabulum (white arrow) and false acetabulum (red arrow) appeared on day 10.

(A2–D2) Low-magnification and (A3–D3) high-magnification images; part 3 is a magnified view of the red box in part 2 (safranin O-fast green stain). Clear differences can be observed in the labro-acetabular complex (black oval) between (A2) normal and (B2) DDH groups on day 5.

(A3–D3) Clear differences can be seen in the width of the proliferative zone (white round) and hypertrophic zone (black round) among the four groups.

(A4–D4) Microscopic observations after laser capture microdissections. Scale bars, 500 μm in black and 100 μm in blue.

The coronal sections of the acetabular cartilage of the DDH group showed noticeable histological changes compared to those of the normal group. Observations under low magnification showed that the acetabular roof in the DDH group became markedly narrow and thick by days 5 and 10 after modeling, with small and deformed femoral heads. Additionally, the labro-acetabular complex in the DDH group became thick after 5 days and appeared distorted after 10 days. Fibrous tissue proliferation was observed in the true acetabulum after 10 days (Figures 1A2–D2). Microscopic observation under high magnification revealed that the chondrocytes in the junction between the cartilaginous and osteal acetabulum had a configuration similar to that of the growth plate, with clearly distinguishable proliferative and hypertrophic zones. In the DDH group, the number of chondrocytes increased in the proliferative zone and decreased in the hypertrophic zone, particularly at the edge of the acetabulum (Figures 1A3–D3).

Identification and functional enrichment analysis of differentially expressed genes

Chondrocyte mRNA expression levels in the proliferative zone were compared between the DDH and normal groups 5 and 10 days after modeling. After 5 days, we identified 1,454 differentially expressed mRNAs in the DDH group, of which 986 were upregulated and 468 were downregulated. After 10 days, we identified 871 differentially expressed mRNAs, of which 655 were upregulated and 216 were downregulated (Figures 2A and 2B). Figures 2C and 2D illustrate the differentially expressed genes (DEGs) identified in the two groups after 5 and 10 days, respectively.

Figure 2.

Transcriptomic analyses of mRNA in the proliferative zone of the acetabular cartilage by RNA sequencing

(A and B) Volcano plots of mRNA sequence transcriptome data comparing gene expression levels between developmental dysplasia of the hip (DDH) and normal groups at (A) 5 and (B) 10 days after modeling. Significant differentially expressed genes (adjusted p-value <0.05 and fold change >1.5) are highlighted in red (upregulated) or blue (downregulated); non-differentially expressed genes are shown in gray.

(C) Heatmap of significantly upregulated (n = 986) and downregulated (n = 468) mRNAs in the DDH group compared to those in the normal group 5 days after modeling.

(D) Heatmap of significantly upregulated (n = 655) and downregulated (n = 216) mRNAs in the DDH group compared to those in the normal group 10 days after modeling.

(E–H) Gene Ontology (GO) annotation and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses of significantly upregulated (E, F) and downregulated (G, H) mRNAs in the DDH group compared with those in the normal group at 5 (E, G) and 10 (F, H) days after modeling. BP: Biological processes; CC: Cellular components; MF: Molecular functions. DDH group at 5 days after modeling: D51, D52, D53; DDH group at 10 days after modeling: D101, D102, D103; Normal group at 5 days after modeling: N51, N52, N53; Normal group at 10 days after modeling: N101, N102, N103.

Figures 2E–2H present the Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis results of the DEGs, and Table 1 lists the high-ranking GO or KEGG terms that are marked in the corresponding figure. GO analyses showed that the downregulated DEGs identified at 5 and 10 days after modeling were primarily enriched in biological processes of chondrocyte differentiation, regulation of cartilage development, and skeletal system morphogenesis and in cellular components related to collagen and the extracellular matrix. KEGG analysis showed that the downregulated DEGs after 5 and 10 days were involved in parathyroid hormone synthesis, secretion, and action and the Hedgehog signaling pathways, which are the canonical pathways involved in bone and cartilage development (Figures 2G and 2H). Conversely, the upregulated DEGs after 5 and 10 days were not enriched in bone or cartilage development-related terms in either the GO or KEGG analyses (Figures 2E and 2F), suggesting that the downregulated DEGs in the DDH group are more critical for DDH development. Therefore, the remainder of our analyses focused on the downregulated DEGs.

Table 1.

Top results of Gene Ontology and Kyoto Encyclopedia of Genes and Genome analyses

ID	Description	P. adjust	Z score
GO:0009141	Nucleoside triphosphate metabolic process	0.00	7.35
GO:0044455	Mitochondrial membrane part	0.00	7.28
GO:0003735	Structural constituent of ribosome	0.00	6.48
rno00190	Oxidative phosphorylation	0.00	5.83
GO:0030099	Myeloid cell differentiation	0.00	6.71
GO:0044815	DNA packaging complex	0.00	5.48
GO:0016209	Antioxidant activity	0.00	3.87
rno04613	Neutrophil extracellular trap formation	0.00	6.48
GO:0001503	Ossification	0.00	−4.90
GO:0070167	Regulation of biomineral tissue development	0.02	−3.16
GO:0031214	Biomineral tissue development	0.02	−3.61
GO:0030500	Regulation of bone mineralization	0.02	−3.61
GO:0062023	Collagen-containing extracellular matrix	0.02	−3.74
GO:0031012	Extracellular matrix	0.02	−4.36
rno04928	Parathyroid hormone synthesis, secretion, and action	0.05	−2.65
GO:0048705	Skeletal system morphogenesis	0.00	−3.74
GO:0061035	Regulation of cartilage development	0.00	−2.83
GO:0032330	Regulation of chondrocyte differentiation	0.00	−2.65
GO:0051216	Cartilage development	0.00	−3.46
GO:0002062	Chondrocyte differentiation	0.00	−3.00
GO:0060350	Endochondral bone morphogenesis	0.00	−2.64
GO:0062023	Collagen-containing extracellular matrix	0.07	−2.83
rno04340	Hedgehog signaling pathway	0.00	−2.24

Identification and functional enrichment analysis of differentially expressed miRNAs

In the DDH group, 55 miRNAs were differentially expressed on day 5, of which 17 were upregulated and 38 were downregulated (Figure 3A). On day 10, 89 miRNAs were differentially expressed in the DDH group, of which 88 were upregulated and 1 was downregulated (Figure 3B). Figures 3C and 3D illustrate the differentially expressed miRNAs (DEMs) between the two groups after 5 and 10 days.

Figure 3.

MicroRNA (miRNA) transcriptome analyses in the proliferative zone of the acetabular cartilage by RNA sequencing

(A and B) Volcano plots of miRNA sequence transcriptome data comparing miRNA expression between the developmental dysplasia of the hip (DDH) and normal groups (A) 5 and (B) 10 days after modeling. Significant differentially expressed miRNAs (adjusted p-value <0.05 and fold change >1.5) are highlighted in red (upregulated) or blue (downregulated); non-differentially expressed miRNAs are shown in gray.

(C) Heatmap of significantly upregulated (n = 17) and downregulated (n = 38) miRNAs 5 days after modeling in the DDH group compared with those in the normal group.

(D) Heatmap of significantly upregulated (n = 88) and downregulated (n = 1) miRNAs 10 days after modeling in the DDH group compared with those in the normal group.

(E and F) Top three enriched Gene Ontology (GO) terms for biological processes (BP), cellular components (CC), and molecular functions (MF), in addition to top three Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways based on target genes predicted from upregulated miRNAs (E) 5 and (F) 10 days after modeling. ECM: Extracellular matrix. DDH group at 5 days after modeling: D51, D52, D53; DDH group at 10 days after modeling: D101, D102, D103; Normal group at 5 days after modeling: N51, N52, N53; Normal group at 10 days after modeling: N101, N102, N103.

miRNAs guide the silencing complex to degrade mRNA or block its translation by pairing with the mRNA bases of target genes. Therefore, we predicted the target genes of the upregulated DEMs based on this negative regulatory relationship. Figures 3E and 3F present the enrichment analysis results based on the observed target genes. Terms associated with cartilage development, such as collagen synthesis and extracellular matrix-related terms, were enriched in both the GO and KEGG analyses, corresponding with the results of the enrichment analyses of the downregulated DEGs.

mRNA-miRNA network construction

Table 2 demonstrates a significant difference in the distribution of differential and non-differential genes between 5 and 10 days after modeling. The Venn diagram in Figure 4A shows that there were 34 common genes among the downregulated DEGs identified at 5 days (468 genes) and 10 days (216 genes) after modeling. The targeted relationship between mRNA and miRNA prediction results and 34 common genes were combined to generate a 38-node network (Figure 4B), in which 28 genes had potential interactions and 10 miRNAs had potential negative regulatory relationships with these genes. Because all genes were downregulated and all miRNAs were upregulated in DDH, proteins are represented as blue-colored nodes (fold change [FC] < 0) and miRNAs are presented as red-colored nodes (FC > 0) in the network. Solid lines indicate protein interactions, while dotted lines suggest potential regulatory relationships between mRNAs and miRNAs. Using Cytoscape tools for visualization, the MCODE plugin highlighted six core nodes (indicated with purple borders). Building on our prior functional enrichment analysis, we identified two gene modules with functions related to bone and cartilage development, indicated by ovals with green and orange backgrounds. Given that these genes were all downregulated in the DDH group, we further utilized gene expression pattern analysis to identify key hub genes.

Table 2.

Comparison of the distribution of differential and non-differential genes between 5 and 10 days after modeling (chi-square with Fisher’s exact test)

	5 days	10 days	χ2	p
Significantly downregulated	468	216	77.57	<0.001
Non-significantly downregulated	5,875	5,621

Figure 4.

mRNA and microRNA (miRNA) network

(A) Common genes between downregulated differentially expressed genes (DEGs) 5 and 10 days after modeling in the developmental dysplasia of the hip (DDH) group compared with those in the normal group.

(B) Network between downregulated DEGs (blue nodes) and their corresponding upregulated differentially expressed miRNAs (red nodes). Purple borders indicate core gene module nodes identified via MCODE; ovals with green and orange backgrounds represent two gene modules in which functional enrichment results were related to bone and cartilage development.

Gene expression patterns and hub gene identification

Two patterns were identified in the gene expression pattern analysis (Figures 5A and 5B): pattern 0, comprising 3,696 genes with decreasing expression after modeling (Figure 5D), and pattern 1, comprising 1,654 genes with increasing expression after modeling (Figure 5E). Since the p value of pattern 0 was less than 0.05, the 3,696 genes in pattern 0 are notably concentrated in the downward expression trend. Among the 34 common genes identified, six showed a linear decrease in gene expression after modeling and were thus classified as hub genes. Figure S1 shows a workflow of the hub genes selection process. These hub genes included microtubule-associated protein 2 (Map2); zinc finger CCCH-type, RNA-binding motif and serine/arginine-rich 2 (Zrsr2); CD248 molecule (Cd248); estrogen-related receptor gamma (Esrrg); protocadherin gamma subfamily A1 (Pcdhga1); and keratocan (Kera) (Figure 5C).

Figure 5.

Expression pattern analyses and hub gene identification

(A and B) Expression profile models displaying the model profile number (upper left corner) and p values (lower left corner).

(C) Six common genes between the protein-protein interaction network (blue) and those with consistently decreasing expression (yellow) identified as the hub genes (green), encompassing Map2, Zrsr2, Cd248, Esrrg, Pcdhga1, and Kera.

(D) Gene expression of Profile No. 0 in the developmental dysplasia of the hip (DDH) group from days 5 to 10 after modeling; the expression levels of 3,696 genes consistently decreased with time.

(E) Gene expression of Profile No. 1 in the DDH group from days 5 to 10 after modeling; the expression levels of 1,654 genes consistently increased with time. x axis: time; y axis: time series of gene expression levels for each gene after log-normalized transformation.

Quantitative real-time PCR verification of hub genes and miRNAs

To verify the accuracy of our transcriptome sequencing results, we used quantitative real-time PCR to verify the expression differences between the six hub genes and six miRNAs with potential targeting relationships. The gene IDs or sequences of the hub genes and miRNAs are presented in Table S2. The hub gene verification analysis showed that the expression of Map2, Zrsr2, Cd248, Esrrg, Pcdhga1, and Kera was markedly downregulated in the DDH group compared to the corresponding expression levels in the normal group (Figures 6A–6F), consistent with the transcriptomic sequencing results. The miRNA verification analysis showed that the expression of novel_miR_769, novel_miR_451, rno-miR-672-5p, novel_miR_1051, novel_miR_241, and novel_miR_885 was markedly upregulated in the DDH group compared to the corresponding expression levels in the normal group, consistent with the transcriptome sequencing results (Figures 6G–6L). Pearson’s correlation analysis showed that the expression levels of novel_miR_241 (r = −0.677, p = 0.016) and novel_miR_885 (r = −0.826, p = 0.001) were significantly negatively correlated with the expression level of Cd248 (Figures 7A and 7B); novel_miR_885 expression was significantly negatively correlated with Esrrg expression (r = −0.790, p = 0.002, Figure 7C); novel_miR_241 (r = −0.819, p = 0.001), novel_miR_451 (r = −0.853, p = 0.001), and rno-miR-672-5p (r = −0.891, p = 0.001) expression levels were significantly negatively correlated with Map2 expression (Figures 7D–7F); and novel_miR_769 (r = −0.632, p = 0.027) and novel_miR_885 (r = −0.788, p = 0.002) expression levels were significantly negatively correlated with Pcdhga1 expression (Figures 7G and 7H). No significant negative correlation was observed between the expression levels of novel_miR_1051 (r = −0.450, p = 0.142) and Zrsr2 (Figure 7I).

Figure 6.

Quantitative real-time PCR verification of hub genes and microRNAs (miRNAs)

(A–F) mRNA expression levels 5 and 10 days after modeling are significantly lower in the developmental dysplasia of the hip (DDH) group than those in the normal group (n = 3).

(G–L) miRNA expression 5 and 10 days after modeling significantly higher in the DDH group than in the normal group (n = 3 biological replicate samples); Student’s t test, ∗p < 0.05, ∗∗p < 0.01 vs. the normal group. D5: DDH group 5 days after modeling; N5: Normal group 5 days after modeling; D10: DDH group 10 days after modeling; N10: Normal group 10 days after modeling; Cd248: CD248 molecule; Esrrg: estrogen-related receptor gamma; Kera: keratocan; Map2: microtubule-associated protein 2; Pcdhga1: protocadherin gamma subfamily A1; Zrsr2: zinc finger CCCH-type, RNA binding motif and serine/arginine-rich 2.

Figure 7.

Pearson’s correlation analyses evaluating correlations between hub genes and miRNAs (n = 12)

(A and B) novel_miR_241 and novel_miR_885 expression levels showing significant negative correlations with Cd248 expression.

(C) novel_miR_885 expression shows a significant negative correlation with Esrrg expression.

(D–F) novel_miR_241, novel_miR_451, and rno-miR-672-5p expression levels show significant negative correlations with Map2 expression.

(G and H) novel_miR_769 and novel_miR_885 expression levels show significant negative correlations with Pcdhga1 expression.

(I) No correlation between novel_miR_1051 and Zrsr2 expression. Cd248: CD248 molecule; Esrrg: estrogen-related receptor gamma; Kera: keratocan; Map2: microtubule-associated protein 2; Pcdhga1: protocadherin gamma subfamily A1; Zrsr2: zinc finger CCCH-type, RNA binding motif and serine/arginine-rich 2.

Targeted relationship validation of novel_miR_769/novel_miR_885 and Pcdhga1/Esrrg

The expression of selected mRNAs and miRNAs was confirmed using quantitative real-time PCR; however, luciferase assays can provide more definitive evidence of direct regulatory interactions. Therefore, we used the RNAhybrid (https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid) tool to predict the binding sites for the eight miRNA-mRNA pairs predicted to have regulatory relationships. Because multiple potential binding sites were predicted for each miRNA-mRNA pair, we selected the binding site with the lowest minimum free energy from the prediction results for the dual-luciferase reporter assay to reduce false positives. The luciferase reporter assay indicated that the novel_miR_769/novel_miR_885 mimics significantly reduced the activity of LUC-Pcdhga1/Esrrg wild type in 293T cells (p < 0.01), thereby validating the target relationship between novel_miR_769/novel_miR_885 and Pcdhga1/Esrrg (Figure 8).

Figure 8.

Pcdhga1/Esrrg is targeted by novel_miR_769/novel_miR_885

(A) Predicted novel_miR_769 binding site on the Pcdhga1 3′-untranslated region (UTR) and dual-luciferase plasmid design. The bases in the yellow background connected by the short vertical lines are the predicted binding sites.

(B) Dual-luciferase reporter assay validates the direct target relationship of novel_miR_769 and the Pcdhga1 3′-UTR.

(C) Predicted novel_miR_885 binding site on the Esrrg 3′-UTR and dual-luciferase plasmid design.

(D) Dual-luciferase reporter assay validates the direct target relationship of novel_miR_885 and the Esrrg 3′-UTR. Results are presented as the mean ± standard deviation luciferase activity. ∗p < 0.05, Student’s t test; (n = 3 technical replicates).

Discussion

DDH is a complex, multifactorial pathological process; proper diagnoses and early interventions may optimize clinical outcomes and reduce the need for invasive surgical treatments. However, a sensitive, specific, and cost-effective test for DDH has not yet been established in pediatric orthopedic medicine. Several studies have reported an association between increased thickness of the acetabular roof cartilage, which is caused by delayed endochondral ossification, and DDH, based on analyses of both clinical samples and animal models.^11,23,24 Our findings in the present study using a DDH rat model confirmed these reported associations. Further histological observations showed that the increase in cartilage thickness was caused by excessive chondrocyte proliferation in the proliferative zone of the cartilaginous and osteal acetabulum junction. Considering that acetabular development is precisely controlled by the proliferative zone in the growth plate of the acetabular cartilage,¹² we performed a transcriptomic analysis of this area, which revealed dysregulation of previously unreported genes or RNAs possibly required for DDH development.

To obtain new insights into the molecular mechanisms underlying DDH, we characterized the mRNA and miRNA expression profiles in the proliferative zone of the acetabular cartilage in rats with DDH and healthy controls 5 and 10 days after modeling. Bioinformatic analyses revealed many differentially expressed mRNAs and miRNAs in the DDH group. Additionally, GO and KEGG enrichment analyses revealed that the DEGs and target genes of the upregulated DEMs were mainly enriched in functions and pathways associated with bone and cartilage development, suggesting their involvement in DDH development. Furthermore, we investigated the regulatory relationships between miRNAs and mRNAs to better understand the pathogenesis of DDH and to inform the development of diagnosis and treatment strategies by constructing an miRNA-mRNA regulatory network. In combination with a gene expression pattern analysis, the regulatory network enabled the identification of continuously downregulated mRNAs classified as hub genes. Moreover, the results of the RNA-seq analysis were validated using quantitative real-time PCR, confirming the abnormal expression of the six hub genes in DDH.

The six hub genes related to DDH development identified in rats were Pcdhga1, Map2, Zrsr2, Esrrg, Cd248, and Kera. PCDHGA1 is a member of the protocadherin gamma gene cluster and an important tissue morphoregulator.²⁵ Transforming growth factor β (TGF-β) superfamily members have been suggested to play a role in regulating the expression of various cadherins and chondrogenesis by modifying cell-cell or cell-matrix interactions,²⁶ indicating the potential participation of PCDHGA1 in DDH development; however, this requires further investigation. MAP2 is the predominant cytoskeletal regulator influencing microtubule dynamics and microtubule-actin interactions.²⁷ Chondrocyte cilia primarily consist of microtubules. The direction of the cilia affects the growth direction on the growth plate, and chondrocytes with cilium defects in the skeletal system can cause numerous human diseases.²⁸ Therefore, MAP2 may be a promising target for clinical interventions in DDH. ZRSR2 is an essential factor that recognizes 30 splicing sites during pre-mRNA splicing. Minor spliceosome malfunctions can affect bone growth in childhood and possibly bone metabolism in adulthood.²⁹ However, the function of ZRSR2 is not yet fully understood. We speculate that malfunctions of the minor spliceosomes caused by low ZRSR2 expression levels affect endochondral ossification in the acetabular cartilage in DDH. ESRRG is widely expressed in tissues where estrogen has important physiologic functions and is involved in various metabolic processes.³⁰ For example, decreased chondrocyte proliferation is caused by cartilage-specific ESRRG overexpression.³¹ Notably, we observed low Esrrg expression with increased chondrocyte proliferation in the proliferative zone of the acetabular cartilage in the DDH group, confirming that ESRRG negatively regulates chondrocyte proliferation from another perspective. KERA encodes the core protein keratocan—a class II member of the small leucine-rich repeat proteoglycan (SLRP) gene family. Proteoglycans are the main components of the cartilage extracellular matrix, of which SLRP family members participate in embryonic bone development³² and interact with TGF-β and epidermal growth factor receptors to suppress cell growth.³³ CD248 is a type I transmembrane glycoprotein highly expressed during embryogenesis.³⁴ CD248 expression on stromal cells plays an important role in vascular homeostasis, cell proliferation, and bone formation via platelet-derived growth factor (PDGF)-induced signaling.³⁵ Angiogenesis is a critical process in endochondral ossification, and a lack of CD248-mediated PDGF signaling can cause disorganized vessel formation,³⁶ which may explain the downregulation of Cd248 expression identified in the DDH model in the present study.

The six identified hub genes may jointly contribute to DDH development; however, confirmation through future study is needed. Nevertheless, these results demonstrate the reliability of our method of hub gene identification. Although numerous genes and loci have been reported to be associated with DDH susceptibility to date,^6,7,8 no firm correlations between genotypes and DDH phenotypes have yet been established. Furthermore, environmental and mechanical factors considerably contribute to the development of DDH given its multifactorial etiology; thus, the role of epigenetics in DDH development should not be ignored. Considering that the inhibitory effects of miRNAs on their target genes are an important part of epigenetics, we constructed a miRNA-mRNA network to identify the critical miRNAs involved in DDH development. miRNAs play critical roles in chondrocyte proliferation, differentiation, and endochondral ossification^21,37; however, the association between miRNAs and DDH remains unclear. We identified several miRNAs with potential regulatory effects on the hub genes, as verified by quantitative real-time PCR. The predicted miRNA expression levels differed considerably between the DDH and control groups, and most of the miRNAs were negatively correlated with the hub genes. Notably, the two key miRNAs identified in previous studies based on a DDH rabbit model did not overlap with the candidate miRNAs identified in this study using a rat DDH model,^11,23 which may be due to differences between the species and sample selections.

To the best of our knowledge, this is the first study to explore the molecular pathogenesis of acetabular cartilage development in DDH using genome-wide trans-omics of miRNA and mRNA expression profiles. This study assessed the dynamic changes in differentially expressed mRNAs and miRNAs in the proliferative zone of the acetabular cartilage in normal and DDH rats at two time points during cartilage development (days 5 and 10). The aberrant Pcdhga1, Map2, Zrsr2, Esrrg, Cd248, and Kera expression during this period suggests that these genes may be involved in DDH development and are regulated by miRNAs. These findings highlight potential targets for the development of diagnostic biomarkers and therapies for DDH. However, these results are preliminary, and further in vivo and in vitro investigations are required to identify potential target genes that may contribute to DDH pathogenesis. In addition, we found that the numbers of hypertrophic chondrocytes in the acetabular cartilage decreased due to the excessive proliferation of proliferative chondrocytes, which may explain the ossification delay in the DDH acetabular cartilage. Therefore, this study provides valuable insights that may inform future studies on DDH.

Limitations of the study

The present study has some notable limitations. The animal model employed in this study replicated pathological changes akin to DDH by mimicking the swaddling position as an environmental factor. However, this model cannot be considered to be physiologically representative of DDH in humans induced by other factors. Further, the sample size was small and the quantity of chondrocytes obtained by LCM was low, which may have led to incomplete RNA-seq data. Furthermore, the mRNA-miRNA regulation network constructed in this study relied on prediction algorithms and large-scale data analysis. Consequently, it is essential to note that the identification of miRNA-mRNA interactions does not necessarily imply functional target regulation, and the results should be interpreted with caution. Substantiating these findings necessitates functional validation experiments and verification in human samples.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
DMEM	Servicebio, Wuhan, CHN	G4511
Fetal bovine serum	Servicebio, Wuhan, CHN	G8002
Lipo6000™ transfection reagent	Beyotime, Shanghai, CHN	C0526
Critical commercial assays
Evo M-MLV RT Kit with gDNA Clean	Accurate Biotechnology, Co., Ltd.	AG11711
Agilent RNA 6000 Nano Kit	Agilent Technologies, CA, USA	5067–1511
Rib-Zero rRNA Remival Kit	Epicentre, Madison, WI, USA	RZNB1056
NEBNext Ultra Ⅱ Directional RNA Library Prep Kit for Illumina	NEB, USA	E7760S
Dual-Glo® Luciferase Assay System	Promega, Beijing, CHN	E2920
Deposited data
Raw and analyzed data	This paper	GEO: GSE240299
Experimental models: Cell lines
293T	Cell Resource Center, Peking Union Medical College (PCRC), Beijing, CHN	1101HUM-PUMC000091
Experimental models: Organisms/strains
Wistar	Charles River Laboratories, Beijing, CHN	102
Oligonucleotides
Primers for quantitative real-time PCR, see Table S1	This paper	N/A
The gene id or sequence of the hub genes and miRNAs, see Table S2	This paper	N/A
Software and algorithms
TargetScan	Lewis et al.39	https://www.targetscan.org/vert_72/
miRanda	Betel et al.40	http://www.microrna.org/microrna/home.do
Cytoscape	Shannon et al.41	https://cytoscape.org/
Short Time-series Expression Miner	Ernst et al.42	http://www.sb.cs.cmu.edu/stem/
RNAhybrid	Krüger et al.43	https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid
Other
Workflow of the hub genes selection process, see Figure S1	This paper	N/A

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Qiang Shu (shuqiang@zju.edu.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

• •Raw and processed microarray data supporting the results of this study have been deposited at the Gene Expression Omnibus and are publicly available as of the date of publication. The accession number is listed in the key resources table. Microscopy data reported in this paper will be shared by the lead contact upon request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

Animals and establishment of the DDH model

All animal experiments were performed following the NIH Animal Research Advisory Committee guidelines and were approved by the Laboratory Animal Welfare and Ethics Committee of Zhejiang University (approval number: ZJU20220211). Wistar rats were provided by the Laboratory Animal Center of Zhejiang University and housed in specific pathogen-free cages. The rats were fed a standard diet and housed in individual cages under constant humidity (50 ± 5%) and temperature (22 ± 1°C) with a 12-h light/dark cycle. Mature male (3 months old, 350 ± 20g) and female (3 months old, 230 ± 20g) Wistar rats were mated to obtain female newborn rats, which were randomly allocated to the DDH and control groups; newborn rats from different litters were used as biological replicates. Due to the association between the incidence of DDH and gender in the population, we have chosen to focus on a single gender in order to minimize systemic errors. DDH was established in newborn rats as previously described.¹³ In brief, DDH was induced by fixing the rat’s hind legs in a hip adduction and extension position with medical tape; the newborn rats in the normal control group did not undergo any intervention. All newborn rats were fed by their mothers and exhibited normal growth and nutritional status. The newborn rats in the DDH group were released from the medical tape for 30 min per day. Five days after birth, half of the newborn rats in both groups were randomly selected and euthanized by decapitation; the other half were euthanized 10 days after birth. The hip joints were isolated and harvested for subsequent experiments.

Method details

Histological observations

The hip joint samples were fixed overnight in 4% paraformaldehyde and subsequently decalcified with 10% ethylenediaminetetraacetic acid in phosphate-buffered saline for two weeks. The samples were then dehydrated in a series of alcohol dilutions (70%, 80%, 90%, 95%, and 100%) and embedded in 100% paraffin. Serial sections were cut to a thickness of 5 μm and stained with safranin O-fast green. The slides were examined and viewed using a diagnostic scanner (3DHISTECH Ltd., Budapest, Hungary) and the corresponding CaseViewer 2.4 software.

LCM

The excised hip joints were trimmed to reveal as much of the bone composition as possible and to conserve the cartilage composition. The hip joint samples were embedded in optimum cutting temperature compound, frozen on dry ice, and stored at −80°C. Frozen blocks of the hip joints were sectioned (10 μm) onto RNAse-free 0.17 polyethylene-naphthalene membrane slides (Cark Zeiss, PALM Microlaser Technologies, Bernried, Germany) and maintained at −80°C for short-term storage. The acetabular cartilage proliferative zone was defined as the proliferative column between the resting and hypertrophic zones, which were microdissected using the previously described cut-and-capture method³⁸ in a PALM Laser-MicroBeam system (PALM Cark Zeiss, Berbrued, Germany) (see Figures 1A4–D4). The tissues dissected from 80 sections from a single animal were pooled for each sample.

RNA-seq

Triplicate biological replicates of samples from the proliferative zone of the acetabular cartilage were obtained from both the DDH and normal groups at 5 and 10 days after modeling. Total RNA from the proliferative zone of the acetabular cartilage was isolated using TRIzol Reagent (Life Technologies, USA) according to the manufacturer’s instructions. RNA concentration and purity were measured using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA). RNA integrity was assessed using the RNA Nano 6000 Assay Kit of the Agilent Bioanalyzer 2100 System (Agilent Technologies, CA, USA). All RNA samples were found to have a 260/280 nm ratio between 1.8 and 2.1, an RNA concentration of ≥200 ng/μL, and RNA integrity values ≥8.

A total of 1.5 μg RNA per sample was used as input material for ribosomal RNA (rRNA) removal with the Ribo-Zero rRNA Removal Kit (Epicentre, Madison, WI, USA). Sequencing libraries were generated using NEBNextRUltra Directional RNA Library Prep Kit for IlluminaR (NEB, USA) following the manufacturer’s recommendations, and index codes were added to attribute sequences to each sample. PCR products were purified (AMPure XP system), and library quality was assessed on the Agilent Bioanalyzer 2100 and by quantitative PCR. Clustering of the index-coded samples was performed on a cBot Cluster Generation System using TruSeq PE Cluster Kitv3-cBot-HS (Illumina) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina platform and reads were generated.

Raw sequencing reads in FASTQ format were first processed through in-house Perl scripts. In this step, clean reads were obtained by removing reads containing adapters, reads containing poly-N, and low-quality reads from the raw data. The miRNA reads were also trimmed and cleaned by removing sequences smaller than 15 nucleotides or longer than 35 nucleotides. Simultaneously, the Q20, Q30, GC-content, and sequence duplication levels of the clean data were calculated. To identify known miRNAs, reads linked to the reference genome were compared with the mature sequences of known miRNAs in the miRbase (v22.1, http://www.mirbase.org/) database and their upstream 2-nucleotide and downstream 5-nucleotide ranges, allowing at most one mismatch. Thus, the identified reads were considered known miRNAs. All downstream analyses were conducted based on high-quality, clean data.

Differential expression and enrichment analyses

Differential expression analyses of the mRNAs and miRNAs in the DDH and normal groups were performed using the “DESeq” R package (R-4.2.1, R Core Team, Vienna, Austria) on days 5 and 10 after modeling. The resulting p-values were adjusted using the Benjamini–Hochberg approach to adjust for the false discovery rate. The mRNAs and miRNAs with an adjusted p-value of <0.05 and a fold change of >1.5 were considered DEGs and DEMs, respectively, and were visualized in volcano plots and heatmaps generated using the “ggplot2” R package.

GO and KEGG enrichment analyses of the DEGs and target genes of the DEMs were performed using the “clusterProfiler” R package and visualized using a bubble chart drawn with the “GOplot” R package.

Construction of protein–protein interaction (PPI) and miRNA–mRNA networks

Venn diagrams were used to identify common DEGs between days 5 and 10. The STRING database (https://string-db.org) was used to predict the PPIs among the common DEGs, with a minimum required interaction score of 0.15. TargetScan³⁹ and miRanda⁴⁰ were used to select miRNAs in the DEMs with potential regulatory relationships with the common DEGs. Cytoscape 3.9.1, an open-source tool, was used to visualize the miRNA–mRNA network.⁴¹ We used the MCODE plugin in Cytoscape to identify the core functional modules in the network. The parameters were as follows: degree cutoff = 2, node score cutoff = 0.2, K-core = 2, and maximum depth = 100. Of the identified modules, we focused on those strongly correlated with the enrichment results and labeled them in the network diagram accordingly.

Gene expression pattern analysis and hub gene identification

A gene expression pattern analysis was performed using the Short Time-series Expression Miner software,⁴² which was used to identify consistently downregulated mRNAs from days 5 and 10. The parameters were set as follows: (1) maximum unit change in model profiles between time points = 1; (2) maximum number of output profiles = 20, with similar profiles being merged; (3) minimum ratio of fold change for DEGs >2.0. Venn diagrams were used to identify consistently downregulated mRNAs in the common DEGs, which were classified as hub genes.

Quantitative RT-PCR

The expression levels of the RNAs identified to be dysregulated in DDH in the RNA-seq analysis were validated using quantitative RT-PCR; the same RNA samples were used for deep sequencing. Six hub genes and six core miRNAs in the network were selected for verification to determine the reliability of the data. The Evo M-MLV RT Kit with gDNA Clean (AG11711; Accurate Biotechnology, Co., Ltd., Hunan, China) was used to reverse-transcribe mRNA and miRNA into complementary DNA. The stem-loop method was used to detect miRNAs. RT-PCR was performed on an ABI 7500 detection system (Applied Biosystems, Waltham, MA, USA) using the following program: 50°C for 2 min, 95°C for 10 min, and 40 cycles at 95°C for 5 s and 60°C for 30 s. GAPDH and U6 were used as internal mRNA and miRNA controls, respectively. Three RNA samples from each ploidy group were validated using RT-PCR and each sample was analyzed three times.

Binding site prediction and dual-luciferase reporter assays

The RNAhybrid (https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid) tool⁴³ was used to predict the binding sites for the miRNA–mRNA pairs predicted to have potential regulatory relationships, and a dual-luciferase reporter system was then used to detect the target relationship. 293T cell line was obtained from the Cell Resource Center, Peking Union Medical College (Beijing, China) and cultured in Dulbecco’s modified Eagle medium (DME H-21) with glucose (4.5 g/L) and 10% fetal bovine serum at 37°C in an atmosphere of 5% CO2. Hanbio Biotechnology (Shanghai, China) supplied the novel_miR_769/novel_miR_885 mimic, inhibitor, and their respective negative controls. The wild-type (WT) and mutant (MUT) 3′-UTR of Pcdhga1/Esrrg containing the predicted novel_miR_769/novel_miR_885 binding site was amplified. The psi-CHECK-Pcdhga1/Esrrg-3′UTR-WT/MUT plasmids (Hanbio Biotechnology, Shanghai, China) and novel_miR_769/novel_miR_885 and negative control (NC) mimics were co-transfected into 293T cells with the aid of Lipofectamine 3000 (Invitrogen, Carlsbad, CA, USA). The luciferase activity levels of firefly and Renilla were measured at 48 h after transfection using the Promega Dual-Luciferase System (E1910; Promega, Madison, WI, USA), according to the manufacturer’s instructions.

Quantification and statistical analysis

The cycle threshold (Ct) values from quantitative RT-PCR were measured, and statistical analyses were performed using the 2^−ΔΔCt relative quantification method. All other statistical analyses were performed using SPSS 21.0 (IBM Corp., Armonk, NY, USA). Fisher’s exact test was performed to evaluate the randomness of the significantly and non-significantly downregulated genes at the two time points. The quantitative RT-PCR data were analyzed using one-way analysis of variance to detect the statistical differences among multiple groups, followed by Student’s t test to determine the statistical differences between specific groups. Correlations between the mRNA and miRNA levels were assessed using Pearson’s correlation coefficients. All data are presented as the mean ± standard deviation. The minimum significance level was set at p < 0.05.

Published: March 7, 2024