Transforming growth factor-β receptor 1: An intervention target for genetic poor cartilage quality induced by prenatal dexamethasone exposure

Graphical abstract

Highlights

• •We confirmed that prenatal dexamethasone exposure leads to hereditary low cartilage quality;
• •Genetic low cartilage quality induced by dexamethasone exposure during pregnancy is associated with reduced cartilage matrix synthesis mediated by the inhibition of expression of the TGFβ signaling pathway;
• •Glucosamine can improve the hereditary low quality of cartilage caused by dexamethasone during pregnancy;
• •Glucosamine could reverse the poor genetic cartilage quality in offspring induced by PDE via up-regulating SP1 expression and promoting its binding to TGFβR1 promoter.

Abstract

Introduction

Fetal-originated osteoarthritis is relative to poor cartilage quality and may exhibit transgenerational genetic effects. Previous findings revealed prenatal dexamethasone exposure (PDE) induced poor cartilage quality in offspring.

Objectives

This study focused on further exploring molecular mechanism, heritability, and early intervention of fetal-originated osteoarthritis.

Methods

Pregnant rats (F0) were segregated into control and PDE groups depending upon whether dexamethasone was administered on gestational days (GDs) 9–20. Some female offspring were bred with healthy males during postnatal week (PW) 8 to attain the F2 and F3 generations. The F3-generation rats were administrated with glucosamine intragastrically at PW12 for 6 weeks. The knee cartilages of male and female rats at different time points were harvested to assay their morphologies and functions. Furthermore, primary chondrocytes from the F3-generation rats were isolated to confirm the mechanism and intervention target of glucosamine.

Results

Compared with the control, female and male rats in each generation of PDE group showed thinner cartilage thicknesses; shallower and uneven staining; fewer chondrocytes; higher Osteoarthritis Research Society International scores; and lower mRNA and protein expression of SP1, TGFβR1, Smad2, SOX9, ACAN and COL2A1. After F3-generation rats were treated with glucosamine, all of the above changes could be reversed. In primary chondrocytes isolated from the F3-generation rats of PDE group, glucosamine promoted SP1 expression and binding to TGFβR1 promoter to increase the expression of TGFβR1, p-Smad2, SOX9, ACAN and COL2A1, but these were prevented by SB431542 (a potent and selective inhibitor of TGFβR1).

Conclusions

PDE induced chondrodysplasia in offspring and stably inherited in F3-generation rats, which was related to decreased expression of SP1/TGFβR1/Smad2/SOX9 pathway to reduce the cartilage matrix synthesis, without major sex-based variations. Glucosamine could alleviate the poor genetic cartilage quality in offspring induced by PDE by up-regulating SP1/TGFβR1 signaling, which was prevented by a TGFβR1 inhibitor. This study elucidated the molecular mechanism and therapeutic target (TGFβR1) of genetic chondrodysplasia caused by PDE, which provides a research basis for precisely treating fetal-originated osteoarthritis.

Introduction

Osteoarthritis (OA) is a long-term and crippling condition that manifests itself through cartilage destruction, osteophyte formation, together with joint pain. In the 1990s, Barker showed that coronary heart disease, hypertension, and type-2 diabetes in adulthood were linked to underweight births based upon epidemiological investigation, and postulated that multiple conditions manifesting in adulthood may depend on intrauterine-based pathogenesis [1]. With further research, mounting epidemiological evidence suggested that adult OA also has an intrauterine origin. In 2011, Aigner and Richter proposed that age-related OA might be related to fetal cartilage development [2]. Subsequently, Sayer and Jordan also reported that underweight infants at birth eventually bore an exacerbated degree of hand and lumbar OA through epidemiological investigations [3], [4]. Our group’s recently published animal-model studies confirmed that exposure to multiple exogenous substances (such as caffeine, nicotine, and alcohol), together with limited feeding during pregnancy, could induce poor cartilage quality in offspring together with increased vulnerability for developing adult-onset OA [5], [6], [7]. In summary, these findings indicate that OA originates in the fetus.

Dexamethasone, a synthetic glucocorticoid that easily passes the placental barrier, is used clinically to treat premature birth and related pregnancy diseases to prevent neonatal respiratory distress syndrome [8]. However, increasing evidence suggests that dexamethasone can lead to underweight newborns combined with susceptibility to long-term adult-onset conditions such as atherosclerotic heart conditions and hypertension [9], [10]. In vivo data additionally indicated that prenatal dexamethasone exposure (PDE) could drive underweight births, dysplasia within several organs (such as the kidneys, bones and cartilage), and susceptibility to conditions in their children [11], [12], [13], [14], [15], [16], [17], [18]. Our group’s preceding studies highlighted poor cartilage quality in offspring after PDE with different courses of treatment, doses and time windows [19]. In addition, a high-fat diet and prenatal nicotine exposure could also lead to poor cartilage quality and OA susceptibility in adult offspring, which can be inherited by the F2 generation [20], [21]. Zhao et al. also found that the poor cartilage quality in offspring rats caused by prenatal caffeine exposure can be transmitted to the F3 generation, and its mode of action was possibly linked to fetal excessive exposure to intrauterine maternal glucocorticoids [22]. However, it remains unclear whether poor multi-generation genetic cartilage quality is induced by PDE in offspring.

Transforming growth factor β (TGFβ) signaling system could be involved in the genesis, proliferation, differentiation, and functional properties and maintenance of chondrocytes [23], [24], [25]. Beginning on day 10.5 of embryonic growth, mesenchymal stem cells (MSC) differentiate into chondrocytes, during which TGFβ signaling pathway plays an important role [23]. Both transforming growth factor β type I receptor and type II receptor (TGFβR1 and TGFβR2) are members of the TGFβ pathway. TGFβ-superfamily ligands bind to TGFβR2, which recruits and phosphorylates TGFβR1 to form a heteromeric complex [26]. TGFβR1 then phosphorylates and activates the downstream proteins, SMAD family member 2/3 (Smad2/3) and SRY (sex-determining region Y)-box9 (SOX9), which cooperate to control gene-expression levels, including those of aggrecan (ACAN) and collagen type II alpha 1 (COL2A1), among others [27]. Chondrocytes and the extracellular matrix are major constituents in articular cartilage that reflect the articular cartilage quality, whereas poor cartilage quality remains intimately linked with adult incidence and development of OA [28]. A previous study of ours validated that prenatal caffeine consumption could reduce the acetylation of TGFβR1 promoter to induce underweight infant births together with increased vulnerability for adult-onset OA [29]. Therefore, TGFβ signal pathway is crucial for cartilage development, although whether it is involved in PDE-induced poor cartilage quality remains to be further explored.

Glucosamine (GlcN) is a glucose derivative and the main substrate for the synthesis of proteoglycans in cartilage matrix [30]. 90 % of oral glucosamine is absorbed and reaches peak levels in the blood at approximately 1 h after oral administration in rats, after which it is metabolized by the liver and eventually decomposes into carbon dioxide, water and urea [31]. The results of a long-term clinical trial showed that GlcN is significantly beneficial for joint structure and can inhibit the progression of OA. GlcN is also recommended as a drug for long-term systematic treatment of knee OA [32]. However, the mechanism underlying the benefit of GlcN in treating OA is not fully clear. At present, it is generally believed that GlcN only provides a substrate for cartilage matrix synthesis. However, previous findings showed that osteogenic differentiation of dental pulp stem cells increased after GlcN administration by increasing TGFβR1 expression [33]; and another study also showed that GlcN was metabolized into lactic acid after being absorbed, thereby increasing TGFβ peptide, and receptor expression and activity to promote collagen production by tendon cells [30], [34]. However, it remains unclear whether GlcN can improve the synthesis of cartilage collagen to improve the poor cartilage quality caused by PDE via the TGFβ pathway.

In this study, we generated rat models of PDE-induced poor articular cartilage quality, which were transmitted to F3 generation through maternal inheritance. Then, we detected changes in cartilage quality and signals related to cartilage-matrix synthesis and degradation across rats in three generations (F1-F3), and explored the therapeutic target of GlcN intervention in the F3-generation rats. This study was aimed at clarifying the molecular mechanism of poor hereditary cartilage quality caused by adverse intrauterine environments and the target of early drug intervention to establish a research foundation for early and accurate therapy against fetus-originated OA.

Methodology

Materials

d-Glucosamine HCl (purity > 98.5 %, MB1695) was purchased from Dalian Meilun Biotechnology™ (Dalian, China). Dexamethasone (No. D1756) and isoflurane were acquired from Sigma-Aldrich (St. Louis, MO, USA) and Baxter Healthcare™ (Deerfield, IL, USA), respectively. Primary antibody dilution buffer (NO. G2025) was procured through Servicebio ™ (Wuhan, China). Trizol® was procured through Omega Bio-Tek™ (Doraville, GA, USA). The SYBR Green dye and reverse transcription and real-time quantitative polymerase chain reaction (RT-qPCR) kits were procured through TaKaRa Biotechnology™ (Dalian, China). Phospho-SMAD Family Member 2 (p-Smad2) (AP0548) and ACAN (A8536) were procured through Abclonal™ (Wuhan, China); TGFβR1 (ab31013), COL2A1 (ab34712), Smad2 (ab40855), SOX9 (ab185230) antibodies and goat anti-rabbit IgG H&L (FITC) (ab6717) were procured through ABCAM™ (Shanghai, China). Oligonucleotide primers were procured through TIANYIHUIYUAN™ (Guangzhou, China). DMEM/F-12 medium and fetal bovine serums (FBS) were procured through Gibco™ (Carlsbad, California, USA). SB431542 was procured through Merck ™ (Beijing, China). MTS Assay Kit® and ATP Determination Kit® (S0026) were procured through Promega™ (Fitchburg, Wisconsin, USA) and Beyotime™ (Wuhan, China), separately. The remaining materials consisted of analytical-grade reagents/chemicals.

In vivo assay: animal handling and protocol

Specific pathogen-free Wistar rats (No. 2017–0018, certification number: 42000600014526, weighing 220 ± 20 g (females) and 270 ± 20 g (males)) were obtained from the Experimental Animal Research Center of the Hubei Provincial Center for Disease Control and Prevention (Hubei, China), and were raised under standard conditions. Animals were allowed to acclimatize for seven days prior to assay. Two female rats were caged with one male rat overnight for breeding. Gestational day (GD) 0 was set following confirmed intercourse through sperm presence within vaginal smears, and pregnant rats were then reared separately, with free access to food and water. Dams were randomly allocated into groups receiving a hypodermic injection of placebo (diluent, saline) or dexamethasone (0.2 mg/kg∙d) at 6 AM once daily between GD9 and GD20. At GD20, two hours after administration, some of the pregnant rats (n = 12/group) were treated with 2 % isoflurane and euthanized. Knee joints of the hind limbs were removed from both male and female fetuses. Some right hind limbs (n = 5/group) were fixed within 4 % paraformaldehyde for histological analysis, and the remaining samples were stored at −80℃ until required for further analysis. The remaining pregnant rats were maintained until delivery of the F1 offspring. Litter sizes were approximately 12–14 at birth for balanced nutrition and the sex ratio was 1:1. Following birth, the F1-generation rats were divided into PDE and control cohorts. Twelve litters were included in each cohort, fed until postnatal week 8 (PW8). F2-generation rats were obtained by mating female rats in each cohort with healthy male rats, and F3 rats were obtained similarly. At PW8 for the F1, F2 and F3 generations, some of offspring were anesthetized with 2 % isoflurane and euthanized via cervical dislocation, followed by collection of bilateral knee joints from the hind limbs. F3-generation rats were fed for 8 weeks, and two male and two female litters from 12 nests were randomly left in each cohort and housed for 12 weeks. By consulting the literature [31], [35] and considering the treatment effect and side effects of GlcN, as well as the animal model, the above rats were then exposed to GlcN (300 mg/kg∙d, dissolved in saline) by oral gavage once a day for 6 weeks (GlcN and PDE + GlcN cohort) or saline (control and PDE cohort). We obtained cartilage samples at 18 weeks according to the method above.

Ethics statement

All animal experiments were conducted in accordance with the National Institutes of Health’s Guide for the Care and Use of Laboratory Animals (Revised 1996). And animal-based studies were performed in line with ethical guidelines/protocols accepted by Institutional Animal Care and Use Committee (IACUC), Wuhan University Center for Animal Experiment, China (Approval no. WP20210061).

Safranin O-fast green staining and histological scoring

Tissue samples were fixed with 4 % paraformaldehyde for 3 days, decalcified for 21 days in 20 % EDTA (pH 7.4), and then embedded in paraffin. Sagittal-slices (5 μm) were prepared and stained with Safranin O-fast green. After deparaffinized to water (the specimens were sequentially placed in xylene I − xylene II − anhydrous ethanol I − anhydrous ethanol II − 90 % alcohol − 85 % alcohol − 75 % alcohol and then washed with water), sections were stained with Fast Green for 5 min, following staining with Safranin O for 20 s before quickly dehydrated with absolute ethanol. Finally, clear xylene was applied to transparent sections for 5 min before being sealed with neutral gum. The imaging was performed using an AH-2® light microscope (Olympus™, Japan). Post-staining images were analyzed by the mean optical density (MOD) of 6 different visual fields for each sample (n = 5) through ImageJ® (v.1.52q). Briefly, after splitting the images into the single channels and converting to 8-bit format, ImageJ was used to directly calculate MODs in the selected area after uncalibrated OD. This study employed the established Osteoarthritis Research Society International (OARSI) scoring setup [36], whereby: 0 = healthy/undamaged cartilage; 0.5 = loss of Safranin O staining with no visible structural shifts; 1 = minute fibrillation; 2 = vertical damage of cartilage restricted within superficial layer; 3 = vertical damage, ≤25 % cartilage surface; 4 = vertical damage, 25–50 % cartilage surface; 5 = vertical damage, 50–75 % cartilage surface; 6 = vertical damage, ≥75 % cartilage surface. The final scores were means from two independent researchers.

Immunohistochemical (IHC) assays

After sagittal-slices (5 μm) were prepared, the specimens were deparaffinized to water and antigen retrieval with boiling in sodium citrate buffer, and sections were blocked in BSA for 30 min followed by incubation with the primary antibody (1:200 for TGFβR1, 1:50 for Smad2, 1:200 for p-Smad2, 1:250 for SOX9, 1:200 for ACAN and 1:200 for COL2A1) in humidified chamber at 4 °C overnight. Biotinylated secondary antibody was added for 30 min on the second day, followed by an avidin-biotinylated horse radish peroxidase complex. Finally, the peroxidase activity was revealed by immersion in DAB substrate. The images for p-Smad2, ACAN and COL2A1 at GD20 and SOX9 of Fig S3B were scanned using a digital slice-scanner (Leica, Aperio CS2) and additional images were performed using an AH-2® light microscope (Olympus™, Japan). The staining intensities were determined by MODs of 6 different visual fields for each sample (n = 5). Post-staining images were analyzed through ImageJ. After converted into 8-bit format, the images were calibrated with uncalibrated OD mode, and MODs of the default thresholds (selected by ImageJ) were then calculated.

Cellular culturing

Primary articular chondrocytes were extracted from 4-weeks-old rats in F3-generation from control (CON-F3C) and PDE cohorts (PDE-F3C). Briefly, distal femoral/proximal tibial cartilage was separated, cleansed from connective tissue, and exposed to type II collagenase (2 mg/mL; Invitrogen™, USA) (37 °C/six hours). Chondrocytes were harvested and seeded (1x10⁵/mL) within DMEM/F-12 medium augmented by 10 % fetal bovine serum with 100 mg/ml streptomycin, and 100 U/ml penicillin. Following two passages, cultures were exposed to the concentration range (0, 100, 200, 300 and 400 μg/mL) of GlcN for different time [37]. In addition, cultures were exposed to SB431542 [38] based on treatment with GlcN (at the concentration of 300 μg/mL) for 72 h. Such assays were performed on three separate occasions for a minimum of three assay runs/occasion.

MTS analysis

Primary rat chondrocytes were seeded in 96-well-plates and cultured in media containing 0, 25, 50, 100, 200, or 400 μg/mL GlcN for 72 h, or at the concentration of 400 μg/mL for 24, 48 or 72 h. Cellular viability was determined using the MTS Assay Kit® according to the protocol. Absorption levels were determined using an enzyme-linked immunosorbent assay reader (TECAN™, Australia) at 490 nm. A blank control was set to exclude the influence of other factors, and the optical density values obtained from exposed cell cultures were normalized to blank control.

Cellular immunofluorescence(IF)

For the cellular IF, the confocal dish was used for cell culture. The samples were fixed with 4 % paraformaldehyde for 15 min, and 0.5 % Triton X-100 was added for 30 min at room temperature. After rinsing with PBS three times, the samples were blocked with BSA for 1 h. The specimens were then incubated overnight at 4 °C with primary antibody dilution (1:200 for TGFβR1, 1:200 for Smad2, 1:200 for p-Smad2, 1:250 for SOX9, 1:50 for ACAN and 1:200 for COL2A1). After rinsing with TBST for 3 times, the cells were incubated with fluorescent secondary antibody in dark for 1 h. Finally, the samples were stained with DAPI at room temperature for 5 min before rinsing with TBST. The IF staining was observed under a confocal microscope (LCS-SP8-STED®, Leica™). ImageJ was used to assay the MODs of 6 different visual fields for each sample (n = 3). After splitting the images into the single channels and converting them to 8-bit format and corrected with the background, ImageJ was used to directly calculate MODs in the selected area.

RNA extraction & RT-qPCR

Total RNA was extracted from articular cartilage tissues/chondrocytes according to previous protocols [7]. Briefly, total RNA of articular cartilage and chondrocytes was isolated using TRIzol reagent. After samples were transferred into 1.5 ml EP tubes, 1 ml TRIzol and 200 μl chloroform were added, then placed for 10 min on ice after mixed. After centrifugation at 12,000 g for 15 min, the upper liquid was transferred into a new EP tube, and placed for 10 min at room temperature after the same amount of isopropanol was added and mixed. The total RNA was obtained after centrifugation at 12,000 g for 10 min. Samples were washed with 1 ml 75 % precooled ethanol twice and dissolved with free RNA enzyme water, which was reverse transcribed using a First-Strand cDNA Synthesis Kit after the concentration and purity were detected by NanoDrop 2000 micronuclei acid analyzer (Themo Scientific, USA) to ensure that the A260/A280 ratio of all samples was between 1.8 and 2.0. The relative mRNA-expression levels of TGFβR2, TGFβR1, Smad2, SOX9, matrix metallopeptidase 13 (MMP13), ADAM metallopeptidase with thrombospondin type 1 motif 5 (ADAMTS5), COL2A1 and ACAN were normalized against glyceraldehyde 3-phosphatedehydrogenase (GAPDH). The primer sequences and annealing temperature are shown in Table S1. The RT-qPCR conditions were as follow: 50 °C/2 min, 95 °C/10 min; 95 °C/15 s, 62 °C/1 min (40 cycles).

ATP assay and transcription factors binding sites prediction

ATP was detected according to the protocol of the ATP determination Kit. JASPAR (https://jaspar.genereg.net) and PROMO (https://alggen.lsi.upc.es) databases were used to analyze protein and DNA binding sites. The promoter sequence of TGFβR1 was imported into JASPAR and PROMO datasets and default values were selected for other parameters.

Chromatin immunoprecipitation (ChIP) assay

The ChIP assay was described in one of our previous articles [6]. Briefly, cells were treated with paraformaldehyde at a final concentration of 1 % for 10 min, and the crosslinking was terminated with glycine of 0.125 M. Each supernatant was discarded after washing twice with PBS and centrifugation at 2000 g for 5 min. After Lysis Buffer was added, the cells were sonicated for 5 min (30 % power, 2 s on and 1 s off) to break the chromatin. The resulting lysate was then divided into three tubes (for the specific sites, IgG and input). After adding beads, they were rotated overnight at 4℃. The samples were then washed with Washing buffer I, II, III and TE buffer (with different electrolyte concentrations) and centrifuged to remove the supernatant. After Elution Buffer containing proteinase K was added, the samples were incubated in a water bath at 65 °C overnight. Finally, DNA was extracted from the samples according to the instructions of the DNA Purification Kit (TIANGEN™, China), and the samples of ChIP were further used for qPCR. The conditions were as follow: 50 °C/2 min, 95 °C/10 min; 95 °C/15 s, 62 °C/1 min (40 cycles).

Statistical analysis

SPSS 20® (SPSS Science Inc™, USA) and Prism 7® (Graph Pad Software™, USA) were employed for all such analyses and graphical representations. Quantitative datasets reflected mean ± S.E.M. Qualitative data reflected median with quartile range. In vivo experiments, for quantitative data, two-way ANOVA was employed for comparative analyses of mean values as required when sexes were Included before conducting pairwise comparisons with the SNK-q (normal distribution and equal variances), while Friedman test was used for qualitative data and quantitative data with abnormal distribution or unequal variances before pairwise comparison with Nemenyi test. In vitro, one-way ANOVA was employed for comparative analyses of mean values before conducting pairwise comparisons with the SNK-q method (normal distribution and equal variances), while Kruskal-Wallis H test was employed for data with abnormal distribution or unequal variances prior to pairwise comparison with Nemenyi test. The representative statistical analysis of some crucial experimental data was Included in Table S2. *P < 0.05, ^**P < 0.01 vs respective control were deemed to confer a statistically significant outcome.

Results

The influence of PDE upon cartilage qualities in offspring’s multiple generation rats

Effects of PDE on the fetal cartilage quality and TGFβR1/Smad2/SOX9 signaling in F1-generation offspring rats

We first assessed morphological changes in cartilage tissues during the intrauterine period of F1-generation rats to evaluate the quality of cartilage. The results of Safranin O-fast green demonstrated uneven and shallower staining, decreased MOD values, thinner cartilage thickness (C.T.) and reduced chondrocyte counts (C.C.) in female and male rat cartilages of the PDE cohort, versus those in the control cohort (P < 0.05, P < 0.01, Fig. 1A, B; Fig S1A). RT-qPCR transcriptomic expression profiling for genes related to cartilage-matrix synthesis (ACAN and COL2A1) and genes in the TGFβ pathway (TGFβR1, Smad2 and SOX9) decreased in the PDE cohort (P < 0.05, P < 0.01, Fig. 1C), although dysregulated TGFβR2 expression was not observed (Fig. 1C). IHC staining further demonstrated that proteomic expression levels of TGFβR1, SOX9, ACAN and COL2A1 were downregulated as well (P < 0.01, Fig. 1D, E; Fig S1B). In addition to reducing cartilage synthesis, increased degradation of cartilage matrix is also an important contributor to the poor cartilage quality. However, our results demonstrated that ADAMTS5 was downregulated in the PDE cohort (P < 0.05, P < 0.01, Fig. 1C), while MMP13 expression remained unchanged (Fig. 1C).

Fig. 1.

Effects of PDE on synthesis and degradation of rat cartilage matrix within F1-generation rats at GD20. (A) Safranin O-Fast Green stain for knee cartilage, scar bar = 500 μm, n = 5; (B) Cartilage thickness (C.T.) (μm) and cell counts (C.C.) per 100 μm²; (C) ACAN, COL2A1, MMP13, ADAMTS5, TGFβR2, TGFβR1, Smad2, and SOX9 mRNA expression by RT-qPCR, n = 8; (D) Immunohistochemical staining and MOD quantification of aggrecan, scar bar = 25 μm, n = 5; (E) Immunohistochemical staining and MOD quantification of TGFβR1 and SOX9, scar bar = 25 μm, n = 5. Datasets reflected means ± S.E.M. *P < 0.05, ^**P < 0.01 vs respective control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Effects of PDE on cartilage quality and TGFβR1/Smad2/SOX9 signaling in adult rats of the F1-F3 generations

Subsequently, we observed morphological changes in the cartilage of 8-week-old rats (both sexes) in F1-F3 generations under PDE-model. The Safranin O-fast green demonstrated that chondrocytes of F1, F2 and F3 (both female and male rats) were arranged neatly, with the deep and uniform staining, and clear tidemarks in the control, while PDE cohort showed uneven and shallower staining, decreased MODs, thinner cartilages, structural disorder, decreased cell counts, tidemark interruption or even disappearance, and increased OARSI scores (P < 0.05, P < 0.01, Fig. 2A, B; Fig S2A). RT-qPCR showed that transcriptomic expression levels of genes related to cartilage matrix synthesis (ACAN and COL2A1) were severely downregulated across the F1, F2 and F3-generation rats (P < 0.05, P < 0.01, Fig. 2C). And IHC staining also demonstrated that the COL2A1 protein was downregulated (P < 0.05, P < 0.01, Fig S3A). In addition, this study detected mRNA-expression levels related to the cartilage matrix synthesis pathway and found that TGFβR1, Smad2 and SOX9 were severely downregulated in F1-F3 rat cartilages of PDE cohort (both sexes) (P < 0.05, P < 0.01, Fig. 2C), and the protein-expression levels of TGFβR1 and SOX9 (key molecules in TGFβ pathway) were also markedly lower (P < 0.05, P < 0.01, Fig. 2D; Fig S2B, S3B). Finally, RT-qPCR detection of cartilage matrix degradation-linked genes (MMP13 and ADAMTS5) were markedly upregulated within PDE cohort (P < 0.05, P < 0.01, Fig. 2C). To sum up, PDE led to poor cartilage quality in the F1, F2 and F3-generation rats (both genders) at PW8, and the expression of cartilage matrix synthesis relative to TGFβR1/Smad2/SOX9 pathway was decreased while cartilage matrix degradation was increased.

Fig. 2.

Effects of PDE on cartilage quality and expression of TGFβR1/Smad2/SOX9 singling of F1- F3 rats. (A) Safranin O - Fast Green staining of knee cartilage from F2-generation rats, MOD analysis and OARSI scores, scar bar = 50 μm, n = 5; (B) The articular cartilage thickness (μm) of F1, F2 and F3-generation rats; (C) ACAN, COL2A1, MMP13, ADAMTS5, TGFβR1, Smad2, and SOX9 mRNA expression by RT-qPCR, n = 8; (D) Immunohistochemistry staining of TGFβR1 protein together with MOD quantification, scar bar = 50 μm, n = 5. Quantitative data reflected means ± S.E.M. Qualitative data reflected Median with Quartile Range. *P < 0.05, ^**P < 0.01 vs respective control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Influences of GlcN on poor cartilage quality in F3-generation rats

GlcN influenced PDE-dependent poor cartilage quality and TGFβR1/Smad2/SOX9 pathway downregulation

To explore whether GlcN could improve the heritable poor cartilage quality, we determined indexes of articular cartilage tissue in F3-generation rats after GlcN treatment. Safranin O-fast green revealed that the control cohort (F3 generation, both sexes) showed a normal articular cartilage morphology, deep matrix staining, an even thickness, and a smooth and complete surface at PW18. However, in the PDE cohort, the cartilage showed more loss of staining; decreased chondrocytes counts; disordered, even interrupted or disappeared tidemark; uneven surfaces; thinner cartilage; and increased OARSI scores; all of these outcomes could be reversed by treatment with GlcN (P < 0.05, P < 0.01, Fig. 3A, B; Fig S4A). RT-qPCR and IHC staining further demonstrated that the mRNA and protein expression levels of genes related to cartilage matrix synthesis (ACAN and COL2A1) were markedly upregulated within PDE + GlcN cohort (P < 0.05, P < 0.01), but no obvious changes were observed after GlcN treatment alone (Fig. 3C-E; Fig S5A, B).

Fig. 3.

GlcN influenced upon cartilage quality and TGFβR1/Smad2/SOX9 singling activity in the PDE model rats of F3 generation. (A) Safranin O - Fast Green staining of knee cartilage from F3-generation rats treated with GlcN, scar bar = 100 μm, n = 5; (B) OARSI analysis; (C) The immunohistochemistry staining of aggrecan, scar bar = 50 μm, n = 5; (D) MOD analysis of aggrecan; (E)The expression of COL2A1 and ACAN mRNA, n = 8; (F) The expression of TGβR1 and SOX9 mRNA determined through RT-qPCR, n = 8; (G) Immunohistochemistry staining for p-Smad2, scar bar = 25 μm, n = 5; (H) Immunohistochemical optical density analysis of p-Smad2 within rat cartilage (both genders). Quantitative data reflected means ± S.E.M. Qualitative data reflected Median with InterQuartile Range. *P < 0.05,^**P < 0.01 vs respective control. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Previous results showed that GlcN could change the expression of key genes in the TGFβ pathway in dental pulp stem cells [33], thereby affecting cell differentiation. To preliminarily explore the mechanism whereby GlcN improves cartilage quality, considering previous results, we first analyzed the TGFβR1/Smad2/SOX9 pathway expression profile. RT-qPCR and IHC staining showed that transcriptomic and proteomic expressions of TGFβR1, Smad2 and SOX9 in the PDE cohort (F3 generation) at PW18 were still expressed at significantly lower levels than in the control (P < 0.05, P < 0.01, Fig. 3F; Fig S4B, S5C-F). After treatment with GlcN, the mRNA and protein expressions of TGFβR1 and SOX9 in the PDE cohort were reversed (P < 0.05, P < 0.01), although Smad2 did not significantly change (Fig S4B, S5C-F). But, it was further found that GlcN could reverse the low degree of phosphorylation of Smad2 induced by PDE (P < 0.05, P < 0.01, Fig. 3G, H). In essence, GlcN improved poor cartilage quality caused by PDE in F3-generation rats and reversed the low expression and activity of TGFβR1/Smad2/SOX9 pathway, which was related to cartilage matrix synthesis.

GlcN influenced chondrocyte matrix synthesis in F3-generation rats

To further explore the effect of GlcN on chondrocytes, we treated primary chondrocytes from F3-generation rats with GlcN. MTS assays showed that chondrocyte viability was not significantly influenced when the GlcN concentration was<400 μg/mL or when the time was<72 h (Fig. 4A, B). Therefore, this study analyzed the expression levels of cartilage matrix synthesis-related to genes (ACAN and COL2A1) at the concentration of 0, 100, 200, 300 or 400 μg/mL for 24, 48 or 72 h. The results demonstrated that GlcN upregulated ACAN and COL2A1 in concentration- and time-dependent manners, and that the ACAN- and COL2A1-expression levels were highest levels at GlcN concentration of 300 μg/mL for 72 h treatment (P < 0.05, P < 0.01, Fig. 4C-F). Furthermore, IF staining showed that GlcN treatment could reverse low proteomic levels of ACAN and COL2A1 in PDE-F3C (P < 0.01, Fig. 4G, H; Fig S6A, B). These outcomes suggest that GlcN improved the PDE-dependent low expression of genes related to cartilage matrix synthesis (ACAN and COL2A1) in chondrocytes of F3-generation rats.

Fig. 4.

GlcN influenced upon ACAN and COL2A1 expression within F3 rat chondrocytes driven by PDE. (A-B) MTS assays of chondrocytes activity treated with GlcN at the different time and concentrations, n = 8; (C-D) ACAN and COL2A1 transcriptomic expression within rat chondrocytes exposed to GlcN of concentrations as indicated (for 72 h), n = 5; (E-F) ACAN and COL2A1 transcriptomic expression within rat chondrocytes exposed to GlcN of the time as indicated (at the concentration of 300 μg/ml), n = 5; (G) The expression of COL2A1 protein detected by immunofluorescence assay in F3-generation rat chondrocytes treated with GlcN, scar bar = 50 μm, n = 3. (H) Semi-quantitative immunofluorescence assay of COL2A1 protein expression. The values reflected means ± S.E.M. *P < 0.05, ^**P < 0.01 vs respective control.

GlcN influenced PDE-dependent TGFβR1/Smad2/SOX9 signaling in chondrocytes from F3-generation rats

TGFβR1/Smad2/SOX9 signaling involved in the intervention effect of GlcN was studied in chondrocytes. Transcriptomic TGFβR1 and SOX9 expression in PDE-F3C cells treated with GlcN increased in concentration- and time-dependent manners (P < 0.05, P < 0.01, Fig. 5A, B; Fig S7A, B), but it had no obvious influence upon transcriptomic Smad2 expression (Fig S7C, D). Similarly, IF staining demonstrated that the protein expression levels of TGFβR1, p-Smad2 together with SOX9 increased after GlcN treatment (P < 0.05, P < 0.01), without a major change in Smad2 expression (Fig. 5C-F; Fig S8A, B). It is suggested that GlcN promoted the expression of TGFβR1 in chondrocytes of F3-generation rats induced by PDE and reversed the low expression and activity of TGFβR1/Smad2/SOX9 signaling.

Fig. 5.

GlcN influenced upon TGFβR1/Smad2/SOX9 signaling pathway downregulation within chondrocytes. Transcriptomic expression for TGFβR1 (A) and SOX9 (B) within rats chondrocytes treated with GlcN of the concentration as indicated, n = 5; (C, D) The expression of TGFβR1 and Smad2 protein assayed by immunofluorescence and semi-quantitative analysis, scar bar = 50 μm, n = 3; (E, F) Immunofluorescence assay of SOX9 protein and semi-quantitative analysis, scar bar = 50 μm, n = 3; (G) The SP1 mRNA expression detected by RT-qPCR, n = 5; (H) The binding site of SP1 and TGFβR1 promoter detected by ChIP-PCR, n = 3. The values were the means ± S.E.M. *P < 0.05, ^**P < 0.01 vs respective control.

GlcN often changes the expression of genes by affecting ATP levels in cells [39]. To further probe the mechanism of GlcN-driven TGFβR1 dysregulation, we detected the ATP levels in chondrocytes treated with GlcN, but no significant change was found (Fig S7E). Transcription factors are another common cause of gene expression changes. Predictions made with PROMO (Fig S7F) and JASPAR (Fig S7G) revealed meaningful binding sites between the SP1 and TGFβR1 promoter zones (Δ-1810 to −1801, −741 to −732, −634 to −625, −566 to −558, −284 to −276, −278 to −269, −277 to −268, −145 to −138, and −94 to −86) in rat cartilage. Moreover, RT-qPCR demonstrated that GlcN reversed PDE-dependent SP1 downregulation in F3 chondrocytes (P < 0.01, Fig. 5G). ChIP-PCR further demonstrated that GlcN mainly changed the binding to SP1 in the promoter region of TGFβR1 at the site Δ-1831 to −1751 (P < 0.01, Fig. 5H). In vivo, this study identified SP1 expression in fetal/adult rat cartilage tissues across F1-F3 generations, and the outcomes revealed SP1 downregulation in the PDE cohort, which was partially reversed by GlcN (P < 0.01, Fig S7H, I). These data indicate that GlcN reversed TGFβR1 downregulation by increasing SP1 expression, thus promoting its binding to the TGFβR1 promoter.

Influences of SB431542 on GlcN activity

SB431542 is a specific and potent inhibitor of TGFβR1 that can inhibit TGFβR1 phosphorylation, thus inhibiting changes in Smads pathway. After CON-F3C chondrocytes were treated with SB431542, the mRNA expression levels of SOX9, ACAN, and COL2A1 and the protein expression levels of p-Smad2, SOX9, and COL2A1 decreased significantly (P < 0.05, P < 0.01, Fig. 6A-E; Fig S9A-C). Moreover, combined treatment with SB431542 and GlcN markedly reversed the effects of GlcN on the expression of these indexes in the PDE-F3C cells (P < 0.05, P < 0.01, Fig. 6A-E; Fig S9A-C). These indicate that GlcN could reverse the PDE-dependent low expression and activity of TGFβR1/Smad2/SOX9 signaling, thereby improving matrix synthesis in chondrocytes.

Fig. 6.

The effect of SB431542 on the intervention action of GlcN. (A-C) The mRNA expressions of SOX9, ACAN, and COL2A1 mRNA detected by RT-qPCR in F3 PDE chondrocytes treated with SB431542 and GlcN, respectively; (D, E) Immunofluorescence assay of COL2A1 protein in F3 PDE chondrocytes treated respectively with SB431542 and GlcN and semi-quantitative analysis, scar bar = 50 μm, n = 3. The values reflected means ± S.E.M. *P < 0.05, ^**P < 0.01 vs respective control.

Discussion

PDE induced the genetic poor cartilage quality

Dexamethasone is normally administered via intramuscular injection of 6 mg every 12 h for 4 times, and the common application time for treating premature delivery is at gestation weeks 24–34. Therefore, in this study, rats at GD9-20 were subcutaneously injected with a dose of 0.2 mg/kg∙d dexamethasone, which translates to approximately 0.033 mg/kg·d in humans (within the clinical range). In this study, after exposure to dexamethasone during pregnancy, the cartilage of male and female offspring rats showed thinner cartilage thicknesses, disordered structures, and decreased cells and contents in the cartilage matrix during the fetal period and 8 weeks after birth. Furthermore, we found these phenotypes were stably inherited by F2- and F3-generation rats, all of which also showed thinner cartilage thicknesses, uneven and shallower cartilage staining, disordered cartilage structure, tidemark interruption or even disappearance, increased OARSI scores in PDE group, and lower mRNA and protein expression of genes related to cartilage matrix synthesis (ACAN and COL2A1), with no significant sex-based differences. These findings indicate that PDE caused poor cartilage quality in offspring rats, which continued after birth and was heritable.

The genetic effects caused by adverse environmental exposure can be divided into intergenerational or transgenerational inheritance. When the F0 generation is exposed to an adverse environment during pregnancy, the F1 generation and early generated germ cells are directly exposed to this environment; thus, hereditary phenotypic changes can occur in both F1 and F2 generations, a process known as intergeneration inheritance. Phenotypes exceeding the F2 generation (referred to as transgenerational inheritance) are subdivided into endogenous or exogenous transgenerational inheritance [40]. To distinguish between them, Skinner et al. proposed using the term endogenous transgenerational inheritance if the germ cells were not directly exposed to environmental factors; otherwise, they proposed using the exogenous transgenerational inheritance [41]. In this study, PDE changed the cartilage quality of F1 and F2-generation rats with the germ cells directly exposed to dexamethasone and F3-generation rats without direct exposure. Therefore, the poor cartilage quality caused by PDE can be classified as the endogenous transgenerational inheritance.

TGFβ signaling was involved in genetic poor cartilage quality induced by PDE

TGFβ signaling is implicated in important functions of chondrocyte aggregation, proliferation, differentiation, and cartilage matrix synthesis. In this study, it was found that PDE reduced the number of chondrocytes in F1 fetal rats (both sexes) and transcriptomic and proteomic expression of TGFβR1/Smad2/SOX9 signal in F1 to F3-generation rats, but did not significantly affect TGFβR2 expression. Furthermore, the expressions of MMP13 and ADAMTS5 (related to cartilage matrix degradation) did not increase significantly in F1 generation fetal rats, but were upregulated in F1 to F3-generation adult rats. It is suggested that the genetic poor cartilage quality in rats driven through PDE was mainly linked to TGFβR1/Smad2/SOX9 pathway downregulation, whereas the increased expression of genes that promote cartilage matrix degradation in adulthood might be related to the poor quality of adult cartilage originating from fetuses, which was inherently defective and more intolerant to stimulation (thereby leading readily to cartilage degradation). However, it remains unclear how this change leads to the stable transgenerational hereditary of poor cartilage quality by germ cells. Because the most common change associated with fetal diseases is epigenetic modification, we speculate that such transgenerational inheritance could be linked to epigenetic modifications in imprinted genes, resulting in stable changes in TGFβ pathway, but this possibility requires further study for confirmation.

TGFβR1 as an intervention target mediated GlcN to improve the genetic poor cartilage quality induced by PDE

Although the clinical efficacy of GlcN is currently controversial, it can significantly prevent joint destruction, improve cartilage quality and inhibit cartilage inflammation in rodent models of arthritis [42]. In this study, after using GlcN to treat F3-generation rats with PDE-induced poor cartilage quality, the cartilage matrix contents markedly increased in the knee joints, the tidemarks became intact, and no major variations were observed within the chondrocyte populations (when compared to the corresponding findings in the control cohort), indicating that cartilage quality clearly improved. In vitro, treating F3 chondrocytes with GlcN significantly also reversed the PDE-induced downregulation of cartilage matrix-synthesis genes. Therefore, GlcN is an effective intervention drug to significantly improve the genetic poor cartilage quality caused by PDE.

We further examined the effect of GlcN on TGFβR1/Smad2/SOX9 signaling in F3-generation rats following PDE. Dataset outcomes demonstrated that GlcN therapy reversed TGFβR1 transcriptomic and proteomic expression downregulation, and upregulated downstream Smad2 phosphorylation and SOX9 mRNA and protein expression. However, following cotreatment with GlcN and SB431542 [38], a selective and potent inhibitor of TGFβR1, the above effects of GlcN were abolished. It is suggested that GlcN can improve genetic poor cartilage quality by reversing the low expression and activity of TGFβR1/Smad2/SOX9 induced by PDE, where TGFβR1 is the main intervention target, which further indicates that the TGFβ pathway is crucial for PDE-induced genetic cartilage quality deficiency.

Previous research demonstrated that GlcN often alters gene expression by changing cellular ATP levels [39]. However, in this study, GlcN treatment did not significantly change the ATP levels in F3-generation rat chondrocytes. Through bioinformatics predictions and cellular experiments, we found that GlcN increased SP1 expression, thereby promoting its binding to the TGFβR1 promoter at the Δ-1831 to −1751 region, which promoted TGFβR1 expression. Further, we found that SP1 expression was significantly lower in adult rat cartilage tissues (F1 to F3 generations) in the PDE model and that GlcN could reverse this change. In addition, Suh et al. also found that GlcN could enhance SP1 expression within proximal renal tubule cells [43], consistent with the results of this study and further supporting the above conclusion. In conclusion, GlcN can reverse the low expression of PDE-induced TGFβR1/Smad2/SOX9 signaling pathway through SP1/TGFβR1 signaling, thereby improving the genetic poor cartilage quality.

Conclusion

As shown in Fig. 7, in this study, pregnant rats (F0) at GD9-20 were administered with dexamethasone, and the female adult offspring in each group were mated with healthy male rats to obtain F2 and F3-generation rats. The F3-generation rats at PW12 were administrated with GlcN for six weeks. The cartilage quality and related pathways were detected at GD20 from F1-generation, PW8 from F1, F2, F3-generation and PW18 from F3-generation female and male rats. It was found that PDE caused decreased cartilage matrix (P < 0.01), thinner cartilage thicknesses (P = 0.0006 in females, P = 0.0003 in males, Table S2), fewer chondrocytes (P = 0.0123 in females, P = 0.0074 in males, Table S2) and poor cartilage quality in fetal rats of F1-generation; caused thinner cartilage thicknesses, uneven and shallower cartilage staining, disordered cartilage structure, tidemark interruption or even disappearance, increased OARSI scores (P < 0.05, P < 0.01) in adult rats of F1, F2 and F3-generation; decreased mRNA and protein expression levels of genes related to cartilage matrix synthesis (ACAN and COL2A1) and genes in TGFβ pathway (SP1, TGFβR1, Smad2 and SOX9) (P < 0.05, P < 0.01) in F1, F2 and F3-generation female and male fetal and adult rats, although PDE did not increase the expressions levels of genes related to cartilage matrix degradation (MMP13 and ADAMTS5) in F1-generation fetal rats, which were only upregulated in F1 to F3-generation adult rats (P < 0.05, P < 0.01). After F3-generation female and male rats of PDE group were administrated with GlcN, all of the above morphological changes and decreased expression of SP1 (P = 0.00001, Table S2), TGFβR1, p-Smad2, SOX9, ACAN and COL2A1 were reversed. In in vitro experiments, treating the primary chondrocytes isolated from F3-generation rats in PDE cohorts with GlcN significantly reversed the downregulation of cartilage matrix-synthesis genes (ACAN and COL2A1) and the low phosphorylation of Smad2 (P = 0.011, Table S2) induced by PDE via promoting SP1 expression and binding to TGFβR1 promoter at the Δ-1831 to −1751 region (P < 0.01), but these reversals be prevented by a TGFβR1 inhibitor. In conclusion, PDE-induced poor cartilage quality in female and male offspring was stably inherited by the F3-generation rats, and its mechanism (at least in part) was linked to downregulated SP1/TGFβR1/Smad2/SOX9 signaling expression and activity, thus reducing cartilage matrix synthesis. Furthermore, GlcN effectively alleviated the poor genetic cartilage quality induced by PDE by increasing TGFβR1 expression. These findings of the present study contribute to experimental and theoretical foundations concerning early intervention and therapy for OA originating from fetuses.

Fig. 7.

SP1/TGFβR1 signaling mediated the intervention effect of hereditary low cartilage quality induced by PDE.

Compliance with Ethics Requirements

All institutional and National guidelines for the care and use of animals were followed.

This studies were approved by the Institutional Animal Care and Use Committee (IACUC) of the Wuhan University Center for Animal Experiment (WP20210061)

CRediT authorship contribution statement

Liang Liu: Conceptualization, Formal analysis, Investigation, Data curation, Methodology, Visualization, Writing – original draft, Writing – review & editing. Bin Li: Methodology, Investigation, Validation, Writing – review & editing. Qingxian Li: Methodology, Data curation, Formal analysis. Hui Han: Data curation, Formal analysis, Methodology. Siqi Zhou: Data curation, Investigation. Zhixin Wu: Data curation, Formal analysis. Hui Gao: Data curation, Formal analysis. Jiayong Zhu: Methodology. Hanwen Gu: Data curation, Formal analysis. Liaobin Chen: Writing – review & editing, Conceptualization, Funding acquisition. Hui Wang: Methodology.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary material

The following are the Supplementary data to this article: