Up-regulated IL-17 and Tnf signaling in bone marrow cells of young male osteogenesis imperfecta mice

Highlight

1. There were 117 shared differentially expressed genes in bone marrow (BM) cells isolated from two types of OI young male murine models.

2. The upregulated IL-17 signaling, Tnf signaling and osteoclast differentiation were significantly enriched in OI BM cells.

3. These dysregulated DEGs and pathways in BM cells might be associated with the excessive bone resorption of the OI mice.

Introduction

Osteogenesis imperfecta (OI) is a congenital disorder characterized by bone fragility (Forlino et al., 2011). The patients generally suffer from recurrent bone fractures and deformities. Up to 21 genes have been associated with OI, but most of the total cases are caused by heterozygous mutation in either of the genes coding for the type I collagen alpha chains, COL1A1 or COL1A2 (Forlino et al., 2011). The pathogenic variants of the two genes often cause either insufficient synthesis or abnormal structure of skeletal collagen, which is correlated with mild OI type I and more severe type II–IV respectively (Saito & Marumo, 2015; Nijhuis et al., 2019).

Multiple reports have highlighted the impaired bone formation and excessive bone absorption in defective collagen-related OI (Takeyari et al., 2021; Iwamoto, Takeda & Ichimura, 2002), which are the main targets in OI treatments. Type I collagen fibrils are produced by bone-forming osteoblasts, making the dysfunctional osteoblasts and their weakened mineralization the main focus of OI etiological research. Conversely, the mechanism of enhanced bone destruction in OI bones has not been well understood, although anti-resorptive drugs have been widely applied in clinical interventions.

Some recent studies have indicated the contribution of inflammation and inflammatory factors in OI bone phenotype. Increased transformation growth factor-beta (TGF β) and excessive TGF signaling have been regarded as a promising target to enhance bone formation in some OI models and patients (Grafe et al., 2014; Marom, Rabenhorst & Morello, 2020). Also, the elevated serum level of interferon (IFN) and tumor necrosis factor (TNFα) have been found in patients and mouse models separately, suggesting the chronic inflammation state in OI (Zhytnik et al., 2020). Inflammation has been closely linked to bone biochemical changes and bone loss in rheumatoid arthritis (RA) and osteoporosis (Weyand & Goronzy, 2021; Tilg et al., 2008). The progenitor cells of osteoblasts and osteoclasts derived from hematopoietic stem cell (HSC) lineage are both situated in the same bone marrow (BM) niche (Ono & Nakashima, 2018). The inflammatory alterations of the BM niche that have not been well explored may directly or indirectly act on the bone manifestations of OI.

In the process of bone marrow development, the transformation of red bone marrow to yellow bone marrow is most obvious in infancy and childhood (Berg, 2021). Compared with the old bone marrow, the level of ROS in the young bone marrow is lower, and the cells are less affected by aging, which can better reflect the initial bone marrow state under genetic defects (Yao et al., 2021). Also, the skeletal symptoms of OI are most obvious in childhood and can be relieved in adulthood (Paterson, McAllion & Stellman, 1984). Studying the bone marrow of young mice will help to understand the cause and process of OI. Here, we performed RNA-sequencing (RNA-seq) and differential gene expression analyses in femoral BM cells isolated from young mice of two types of OI mouse models and wild-type mice to explore the changes in the BM niche. Both the Col1a1^+/−365 mice and heterozygous oim mice (Col1a2^oim/+) display osteogenesis deficiency accompanied by an increased number of osteoclasts. The differentially expressed genes (DEGs) shared by the two OI models were involved in IL-17 signaling pathway, Tnf signaling pathway and osteoclast differentiation. The DEGs related to the three pathways were all upregulated in OI BM cells when compared with normal cells, suggesting the activation of these signalings. These dysregulated genes and pathways in mutant BM cells are likely to play an important role in the pathological changes of OI bones.

Materials and Methods

Results

Differential gene expression analysis

RNA-seq was performed in BM cells of the three genotypes of mice. The statistical power (depth = 100, cv = 0.1, effect = 2, α = 0.05, power = 0.9) of this experimental design, calculated in RNASeqPower was 0.0875. Compared with BM^wt, a total of 279 DEGs were identified in BM^+/−365 and 219 of which were upregulated and 60 were downregulated (Fig. 1). A total of 291 DEGs were differentially expressed in BM^oim/+ when compared with BM^wt, 237 of which were upregulated and 54 were downregulated (Fig. 1). There were 117 DEGs shared by the two types of OI mice (Fig. 1). A total of 102 of the common DEGs were concordantly upregulated and 12 were concordantly downregulated, suggesting their consistency of alteration (Data not shown).

The top 20 upregulated shared DEGs in BM^+/−365 or BM^oim/+ were listed in Tables 2 and 3 respectively. A total of 17 of them were the same genes that contained several Ifn signaling-related genes (e.g., Lfit1, Lfit3, Lfi44 and Irf7) (Tables 2 and 3). Notably, the transcription of Coch, most abundantly expressed in the inner ear, was obviously increased in the two types of BM cells (Tables 2 and 3). The abnormal Coch expression might be associated with OI hearing loss.

Table 2:

The top 20 upregulated shared DEGs in BM^+/−365 when compared with BM^wt.

Gene	Description	Log2 (Fold chang)	Q-value
Fgf23	fibroblast growth factor 23	5.83	0.0002360794650486
Ptgds	prostaglandin D2 synthase (brain)	5.54	0.0021177353272204
Oas1g	2’–5’ oligoadenylate synthetase 1G	4.9	0.0009431000835065
Apol9b	apolipoprotein L 9b	4.6	0.0002133173066225
Ifit1	interferon-induced protein with tetratricopeptide repeats 1	4.28	0.0003668837536278
Ifit3	interferon-induced protein with tetratricopeptide repeats 1	4.25	0.0015924514011703
Ifi44	interferon-induced protein 44	4.17	0.003021888072243
Coch	coagulation factor C homology/cochlin	4.11	0.0260410468387774
Isg15	ISG15 ubiquitin-like modifier	3.98	0.0000106275044397477
Ifit3b	interferon-induced protein with tetratricopeptide repeats 3b	3.91	0.001908267305679
Rsad2	radical S-adenosyl methionine domain containing 2	3.38	1.62699675914253e−13
Usp18	ubiquitin specific peptidase 18	3.24	0.0010491995647924
Oas1a	2’–5’ oligoadenylate synthetase 1A	3.05	0.0016396718614034
Oasl2	2’–5’ oligoadenylate synthetase 2	3	0.0031321485213441
Gbp6	guanylate binding protein 6	3	0.0010429030626467
Irf7	interferon regulatory factor 7	2.98	0.002419157920696
Cfb	complement factor B	2.97	0.0133696378926878
Gbp10	guanylate-binding protein 10	2.97	0.0013662770771023
Zbp1	Z-DNA binding protein 1	2.9	0.0088547659475085
Rtp4	receptor transporter protein 4	2.85	0.0013424962435546

DOI: 10.7717/peerj.13963/table-2

Table 3:

The top 20 upregulated shared DEGs in BM^oim/+ when compared with BM^wt.

Gene	Description	Log2 (Fold chang)	Q-value
Ptgds	prostaglandin D2 synthase (brain)	7.5	6.17204589560265e−9
Coch	coagulation factor C homology/cochlin	6.02	2.94335300566472e−8
Oas1g	2’–5’ oligoadenylate synthetase 1G	5.18	9.681071247414061e−14
Fgf23	fibroblast growth factor 23	4.99	0.000606927029448
Ifi44	interferon-induced protein 44	4.46	1.23002041345624e−21
Ifit3	interferon-induced protein with tetratricopeptide repeats 3	4.44	1.08339825790861e−9
Ifit1	interferon-induced protein with tetratricopeptide repeats 1	4.22	5.155872680064089e−11
Ifit3b	interferon-induced protein with tetratricopeptide repeats 3b	4.2	9.503797111990189e−11
Gbp6	guanylate binding protein 6	3.63	1.1575697141787101e−74
Apol9b	apolipoprotein L 9b	3.52	0.0012727812836212
Cfb	complement factor B	3.47	9.1080396056311e−39
Isg15	ISG15 ubiquitin-like modifier	3.37	7.51494012126258e−18
Tgtp1	T cell specific GTPase 1	3.36	1.28520549551154e−10
Oasl2	2’–5’ oligoadenylate synthetase 2	3.19	1.83115647109809e−26
Gbp10	guanylate-binding protein 10	3.19	3.41807441735383e−13
Oas1a	2’–5’ oligoadenylate synthetase 1A	3.15	3.8026707877775104e−15
Usp18	ubiquitin specific peptidase 18	3.14	1.29677254526075e−14
Iigp1	interferon inducible GTPase 1	3.14	0.000013763932948896
Irf7	interferon regulatory factor 7	3.12	6.5309598268511e−36
Ly6i	lymphocyte antigen 6 complex, locus I	3.1	3.91587391622785e−21

DOI: 10.7717/peerj.13963/table-3

Gene Ontology biology process analysis

Gene Ontology (GE) biology process analysis indicated the enrichment of multiple immune-related genes in such as cellular response to interferon-beta (Ifnβ), immune system process, and immune response, and cellular response to interferon-alpha (Ifnα) in defective BM cells (Fig. 2A). We tested the level of Lfit1 and Irf7, and data showed that both of the two genes were significantly upregulated in mutant BM cells (Figs. 2B and 2C), suggesting the activated Ifn signaling in OI BM.

KEGG pathway enrichment and GSEA analysis

The significantly enriched pathways by KEGG analysis of the 117 shared DEGs included the IL-17 signaling pathway, Tnf signaling pathway and osteoclast differentiation in OI BM cells (Fig. 3A). There were 11 enriched DEGs involved in the aforementioned pathways (Table 4). GSEA analysis demonstrated that all of the three pathways were markedly upregulated in OI-derived BM cells (Figs. 3B–3D). Besides, the sequencing data demonstrated that both I117a and Tnf increased more than two-fold in BM^oim/+ (Table 5). The Tnf expression also got significantly elevated in Col1a1^+/365 mice-derived BM cells (Table 5). In addition, the mRNA level of Rankl (Tnfsf11) rose in defective BM cells (Table 5). It has been reported that upregulated IL-17 and Tnf signalings contribute to the excessive bone resorption in multiple inflammatory bone loss diseases (Tsukasaki & Takayanagi, 2019; Weitzmann, 2017), suggesting the potential role of overactivated IL-17 and Tnf pathways in an increased number of osteoclasts under the OI background.

Table 4:

The shared DEGs involved in ‘IL-17 signaling pathway’, ‘Tnf signaling pathway’ and ‘osteoclast differentiation’ of BM^+/−365 and BM^oim/+ when compared with BM^wt.

Gene	Description	Log2 (Fold chang)		Q-value
		BM+/−365	BMoim/+	BM+/−365	BMoim/+
Ccl2	chemokine (C-C motif) ligand 2	2.73	2.92	5.23025917899991e−16	0.00000124536325640583
Gm5431	predicted gene, 45935	1.11	1.47	0.0421644826331762	4.85491962362027e−9
Cxcl10	chemokine (C-X-C motif) ligand 10	2.09	1.57	0.0009539010036805	0.0028754859114639
Lif	leukemia inhibitory factor	1.46	1.70	0.0098724762681209	0.0024271897572989
Ifng	interferon gamma	1.71	2.15	0.0002510762900628	2.01552889198521e−7
Fosb	FBJ osteosarcoma oncogene B	2.40	2.23	0.00000208273094362867	0.0008131293868126
Stat1	signal transducer and activator of transcription 1	1.27	1.47	0.0000710335367185181	4.59990999943035e−23
Jund	jun D proto-oncogene	1.00	1.04	0.0015850803802578	2.41298702546188e−14
Jun	jun proto-oncogene	1.86	1.66	4.2489456147999705e−20	1.75119110664004e−52
Lfi47	interferon-induced protein 47	1.62	1.75	0.0247317221652343	5.43681272346479e−24
Fcgr1	Fc receptor, IgG, high affinity I	2.13	2.63	0.0033203206673691	1.34407141389434e−33

DOI: 10.7717/peerj.13963/table-4

Table 5:

Some key genes that were not shared DEGs involved in ‘IL-17 signaling pathway’, ‘Tnf signaling pathway’ and ‘osteoclast differentiation’ of BM^+/−365 and BM^oim/+ when compared with BM^wt.

Gene	Description	Log2 (Fold Chang)		Q-value
		BM+/−365	BMoim/+	BM+/−365	BMoim/+
Tnf	tumor necrosis factor	0.77	1.21	0.0110345421756999	2.8259877145669397e−13
Il17a	interleukin 17A	2.98	4.62	0.529929360546415	0.0204958961911906
Tnfrsf11a	tumor necrosis factor receptor superfamily, member 11a, NFKB activator/Rank	−0.004	0.20	0.996564355301463	0.791864645628747
Tnfsf11	tumor necrosis factor (ligand) superfamily, member 11/Rankl	1.09	0.95	0.587237549231164	0.486382634245118

DOI: 10.7717/peerj.13963/table-5

The changes of DEGs in BM cells isolated from OI modeled mice

The heatmap as shown in Fig. 4A displayed the differential expression of the 11 DEGs mentioned above. Many of them at least participate in two of the IL-17 signaling pathway, Tnf signaling pathway, and osteoclast differentiation, suggesting their complex crosstalk. RT-PCR validation indicated that Ccl2 (Fig. 4B), Jun (Fig. 4C) Cxcl10 (Fig. 4D), Lif (Fig. 4E) and Stat1 (Fig. 4F) were apparently upregulated in mutant BM cells, which was consistent with the sequencing data. And Fosb (Fig. 4G) and Ifng (Fig. 4H) expression also showed an increasing trend in OI mice-derived BM cells. These results preliminarily identified the dysregulated DEGs, which might play important roles in OI bone erosion alterations.

Besides, the expression of these genes in adult mice was also detected by RT-PCR. Data showed that Tnfα (Fig. S1C), Lif (Fig. S1D), and Ccl2 (Fig. S1E) were still elevated in mutant BM cells when compared to the normal cells, although there were no differences in the expression of Ifit1 (Fig. S1A), Jun (Fig. 1F), Stat1 (Fig. S1G) and Cxcl10 (Fig. S1H). The obviously increased level of Irf7 (Fig. S1B) and Ifng (Fig. S1I) in BM^oim/+ instead of BM^+/−365 could be observed in RT-PCR assay. The IL-17 signaling pathway, Tnf signaling pathway and osteoclast differentiation under OI background remained up-regulated until adulthood.

Discussion

OI is mainly caused by the defective production or assembly of type I collagen, making bone tissue severely affected (Forlino et al., 2011; Hoyer-Kuhn, Netzer & Semler, 2015). OI patients with COL1A1 or COL1A2 mutations can be classified into OI type I-IV phenotypes (Forlino et al., 2011). OI type I due to decreased synthesis of collagen fibrils shows the mildest phenotype, while structurally aberrant type I collagen can give rise to much more severe type II–IV. Despite the subtypes and caused pathogenic variants, high bone turnover represented by hypercellular osteocytes but insufficient mineralization together with enhanced bone absorption generally exist in collagen defect-related OI (Liu et al., 2019; Lopez Franco et al., 2005; McBride, Shapiro & Dunn, 1998; Saban et al., 1996). Many studies have focused on the mechanism underlying the unit bone formation insufficiency, while excessive bone erosion has not been extensively explored.

Bone resorption is mainly mediated by osteoclasts that originate from HSC-derived monocytes (Boyle, Simonet & Lacey, 2003). The BM monocytes can differentiate into mature multinucleated osteoclasts in the presence of macrophage-colony stimulating factor (M-CSF) and receptor activator of nuclear factor kappa B ligand (RANKL) (Boyle, Simonet & Lacey, 2003; Feng, Guo & Li, 2019). It has been proved that inflammatory cytokines such as TNFα can act synergistically with RANKL to promote osteoclastotogenesis (Zhao et al., 2012). Recently, the inflammatory component in OI pathogenesis has been increasingly attractive. The chronic inflammation state of OI murine models and pediatric patients has been suggested by their elevated serum levels of inflammatory cytokines including IFN and TNFα (Zhytnik et al., 2020; Brunetti et al., 2016). Many clues suggest that the cells and cytokines of the BM niche can regulate the maturation and function of osteoclasts through direct contact and the paracrine effect (Herrmann & Jakob, 2019; Ciucci et al., 2015). But the inflammatory changes in BM cells are still unclear.

OI is a type of congenital genetic disease, with the most severe symptoms in childhood (Paterson, McAllion & Stellman, 1984). Therefore, exploration of the molecular changes of bone marrow cells of OI in early childhood is helpful to understand its pathogenesis.

Here, we performed RNA-seq of whole BM cells isolated from two types of OI models that respectively stimulate abnormal quantity and structure collagen-related OI. Data showed that the BM cells from the two young murine models shared 117 DEGs (Q ≤ 0.05, |Log₂FC| ≥ 1) (Fig. 1). A total of 17 of the top 20 upregulated shared DEGs were the same genes, including several Ifn signaling-related genes containing Lfit1, Lfit3, Lfi44 and Irf7 (Tables 2 and 3). Consistent with the findings of Zhytnik’s group, GO biology process analysis also enriched multiple DEGs involved in the cellular response to Ifnβ and Ifnα in defective BM cells (Fig. 2), suggesting the activated Ifn signaling in OI BM. Transcription factor STAT1 is a key effector of the IFN pathway (Michalska et al., 2018). The upregulated Stat1 in OI BM cells (Fig. 4F) further revealed the dysregulated Ifn signaling. IFNs are key cytokines for both innate and adaptive immune responses (Takayanagi et al., 2005). Some previous studies found that IFN-γ can promote osteoclastogenesis under T-cells activation and enhance the multinucleation of myeloid lineage cells in osteoporosis (Biros et al., 2022). Thus the overactivated Ifn signaling might also act on the bone resorption in the OI background.

KEGG and GSEA analysis of the 117 shared DEGs showed that the IL-17 signaling pathway, Tnf signaling pathway and osteoclast differentiation were significantly upregulated in OI BM cells (Fig. 3). A total of 11 enriched DEGs were involved in the three pathways and displayed complex crosstalk (Table 4). Il17a, Tnf and Rankl showed different degrees of upregulation (Table 5), indicating the chronic inflammation and enhanced osteoclastogenesis. RT-PCR assay confirmed the elevated expression of some DEGs in young and adult mice (Fig. 4, Fig. S1). The results of 4-weeks old mice were consistent with the results of RNA-seq (Fig. 4). While only several genes remained elevated in 12-weeks old OI mice (Tnfα, Lif, and Ccl2) (Fig. S1), which might be related to the relieved bone phenotype in adulthood. IL-17A and TNFα have been proved to play an essential role in inflammatory bone erosion by inducing the production of RANKL (Weitzmann, 2017; Zhao et al., 2012). IL-17A can also promote TNFα secretion. Matthews et al. (2017) found the increased serum level of Tnf in homozygous oim mice, however, anti-TNFα therapy failed to reduce bone absorption. These results suggested that blocking Tnf signaling alone is insufficient to effectively reverse the excessive bone resorption in OI bones. Targeting both IL-17 and Tnf signalings may be an efficient strategy for OI treatment.

Conclusions

This study preliminarily revealed the dysregulated biological process of cellular response to Ifn together with upregulated IL-17 signaling, Tnf signaling and osteoclast differentiation in young male OI BM niche by RNA-seq. Either defective collagen production or abnormal collagen assembly shared similar alterations in gene profiles and pathways involving inflammation and osteoclast activation. Data presented here not only contributed to understanding the mechanism of the enhanced bone absorption in the bone of OI, but also provided more evidence to develop potential anti-inflammation therapies.

Supplemental Information