Abstract

   The osteonecrotic area of steroid-induced avascular necrosis of the
   femoral head (SANFH) is a hypoxic microenvironment that leads to
   apoptosis of transplanted bone marrow mesenchymal stem cells (BMSCs).
   However, the underlying mechanism remains unclear. Here, we explore the
   mechanism of hypoxic-induced apoptosis of BMSCs, and use the mechanism
   to improve the transplantation efficacy of BMSCs. Our results show that
   the long non-coding RNA [52]AABR07053481 (LncAABR07053481) is
   downregulated in BMSCs and closely related to the degree of hypoxia.
   Overexpression of LncAABR07053481 could increase the survival rate of
   BMSCs. Further exploration of the downstream target gene indicates that
   LncAABR07053481 acts as a molecular “sponge” of miR-664-2-5p to relieve
   the silencing effect of miR-664-2-5p on the target gene Notch1.
   Importantly, the survival rate of BMSCs overexpressing LncAABR07053481
   is significantly improved after transplantation, and the repair effect
   of BMSCs in the osteonecrotic area is also improved. This study reveal
   the mechanism by which LncAABR07053481 inhibits hypoxia-induced
   apoptosis of BMSCs by regulating the miR-664-2-5p/Notch1 pathway and
   its therapeutic effect on SANFH.

   Subject terms: Apoptosis, Stem-cell research
     __________________________________________________________________

   Long non-coding RNA [53]AABR07053481 is shown to inhibit
   hypoxia-induced apoptosis of bone marrow mesenchymal stem cells in rats
   by regulating the miR-664-2-5p/Notch1 pathway, promote the repair
   efficacy of BMSCs on early steroid-induced avascular necrosis of the
   femoral head.

Introduction

   Steroid-induced avascular necrosis of the femoral head (SANFH) is a
   common orthopedic complication after high-dose steroid hormone
   treatments^[54]1,[55]2. SANFH ranks first among all types of
   non-traumatic femoral head necrosis and is one of the main diseases
   endangering human health worldwide^[56]3,[57]4. Furthermore, 80% of
   untreated patients are prone to collapse of the femoral head structure
   and subchondral fractures in the late stage, and the disability rate is
   high^[58]5–[59]7. Therefore, early prevention and treatment of SANFH
   are particularly important to reduce hip disability. Preclinical
   studies have shown that bone marrow mesenchymal stem cell (BMSC)
   transplantation may become an ideal treatment for early
   SANFH^[60]8–[61]11. However, the transplanted BMSCs display high levels
   of apoptosis in the osteonecrotic area in vivo, which limits the
   transplantation efficacy^[62]12,[63]13. The apoptosis of transplanted
   BMSCs is mainly related to the local hypoxic microenvironment in the
   necrotic area of the femoral head^[64]14–[65]17. There is still a lack
   of effective intervention methods for hypoxia-induced apoptosis of
   transplanted BMSCs. How to effectively inhibit hypoxia-induced
   apoptosis of transplanted BMSCs and improve the efficacy of BMSC
   transplantation is a challenge for further research to overcome.

   In the research of diseases related to ischemia-hypoxia, many
   non-coding RNAs (ncRNAs) have been found to play key roles in
   regulating hypoxia-induced apoptosis^[66]18–[67]20. Long non-coding RNA
   (LncRNAs) belongs to a class of ncRNAs with transcripts that are
   between 200 nt and 100 kb in length. Many studies have shown that
   LncRNAs are widely involved in various biological processes, such as
   chromosome silencing, chromatin modification, transcriptional
   activation, transcriptional interference, and intranuclear
   transport^[68]21–[69]23. LncRNA expression levels are regulated by
   parameters of the cell growth environment, such as hypoxic conditions.
   For example, under hypoxic conditions, LncEFNA3 can promote the
   proliferation of breast cancer cells from blood vessels to surrounding
   tissues^[70]24. Moreover, LncHOTTIP promotes epithelial-mesenchymal
   transformation under hypoxia^[71]25, and LncMALAT1 enhances
   arsenic-induced glycolysis under hypoxia^[72]26. Furthermore, the
   differential expression of LncRNAs may be closely related to cell
   function under hypoxic conditions. However, it is unclear whether there
   are differentially expressed LncRNAs that regulate hypoxia-induced
   apoptosis of BMSCs transplanted to the necrotic area of the femoral
   head. Therefore, searching for LncRNAs related to hypoxia-induced
   apoptosis of BMSCs and understanding their biological functions and
   regulatory mechanisms will help reveal new anti-apoptotic targets and
   provide a novel method to reduce hypoxia-induced apoptosis of
   transplanted BMSCs.

   In this study, we analyzed the expression profile of LncRNAs in BMSCs
   using microarray data and found that LncAABR07053481 was associated
   with hypoxia-induced apoptosis of BMSCs. The role and mechanism of
   LncAABR07053481 as a competing endogenous RNA (ceRNA) of mir-664-2-5p,
   and its effect on the Notch1 pathway and the inhibition of
   hypoxia-induced apoptosis of BMSCs were determined, and the repair
   efficacy of inhibiting hypoxia-induced apoptosis of BMSCs in early
   SANFH was evaluated. Our findings provide a novel idea and method for
   improving the efficacy of BMSC transplantation.

Results

LncAABR07053481 is downregulated under hypoxia and associated with apoptosis
of BMSCs

   To study the effect of hypoxia on BMSCs, we successfully isolated rat
   BMSCs (Supplementary Fig. [73]1) and constructed an early SANFH rat
   model (Supplementary Fig. [74]2a, b). Consistent with previous
   studies^[75]27,[76]28, our data showed that the oxygen concentration in
   the necrotic area of the femoral head was <0.1% (Supplementary
   Fig. [77]2c). Subsequently, BMSCs were cultured under hypoxia (0% O[2])
   for 48 h to simulate a hypoxic microenvironment, and we found that, in
   the hypoxic environment, the content of reactive oxygen species (ROS)
   in BMSCs was increased (Supplementary Fig. [78]2d, e), the
   mitochondrial membrane potential was decreased (Supplementary
   Fig. [79]2f, g), ATP synthesis was decreased (Supplementary
   Fig. [80]2h), the expression of apoptosis-related proteins was
   increased (Supplementary Fig. [81]2i–l), and there were a large number
   of apoptotic BMSCs (Supplementary Fig. [82]2m–o). This indicates that
   hypoxia can lead to the apoptosis of BMSCs. Studies have found that
   LncRNA plays an important role in the regulation of hypoxia-induced
   apoptosis^[83]29,[84]30. Therefore, to reveal the LncRNAs that regulate
   BMSC apoptosis under hypoxic conditions, BMSCs were cultured under
   normoxia (20% O[2]) or hypoxia (0% O[2]) conditions for 48 h, and the
   LncRNA profiles of BMSCs exposed to normoxia or hypoxia were obtained
   by microarray analysis. Our results showed that 106 LncRNAs were
   upregulated and 173 LncRNAs were downregulated under hypoxic
   conditions, with 33 potential LncRNAs, in particular, being upregulated
   or downregulated by >3-fold after hypoxia, as shown in the heatmap
   (Fig. [85]1a). We focused on a LncRNA transcript that was highly
   downregulated during hypoxia (i.e., LncAABR07053481, transcript ID:
   ENSRNOT00000078988.1, source: Ensembl; the LncRNA transcript also
   matches LOC120101634, Accession: [86]XR_005502377, source: RefSeq).
   LncAABR07053481, located on chromosome 3, is composed of six exons with
   a full-length 2112-nt transcript. Of note, LncAABR07053481 lacks the
   ability to encode protein (Fig. [87]1b). In addition, it does not
   produce any detectable peptides according to the prediction made with
   CPAT and by in vitro translation assay (Fig. [88]1c, d). Subsequently,
   to verify the microarray data, we detected the expression of
   LncAABR07053481 in the BMSC hypoxia model by qPCR, and the results were
   consistent with the microarray data. Specifically, we found that the
   expression of LncAABR07053481 was significantly downregulated under
   hypoxia, and, with the aggravation of hypoxia, the apoptosis rate of
   BMSCs increased, while the expression of LncAABR07053481 was
   continuously downregulated (Fig. [89]1e–g). In vitro, our results
   showed that the expression of LncAABR07053481 in BMSCs was
   downregulated under hypoxia, suggesting that LncAABR07053481 may be
   closely related to hypoxia-induced apoptosis of BMSCs. We also verified
   the expression of LncAABR07053481 in vivo and the relationship between
   LncAABR07053481 and the apoptosis of transplanted BMSCs. In vivo, we
   found that, compared with the control (Normal) group, the expression of
   LncAABR07053481 in the BMSCs transplanted with the model (SANFH) group
   was significantly downregulated (Fig. [90]1h). Moreover, the expression
   of LncAABR07053481 in the BMSC transplantation group was continuously
   downregulated with the prolongation of transplantation time in vivo
   (Fig. [91]1i). Thus, both in vitro and in vivo data showed that the
   differential expression of LncAABR07053481 may be closely related to
   hypoxia-induced apoptosis of BMSCs.

Fig. 1. LncAABR07053481 is downregulated under hypoxia and associated with
apoptosis of BMSCs.

   [92]Fig. 1
   [93]Open in a new tab

   a Cluster analysis of LncRNAs (n = 3). b Location of LncRNA in the
   genome. c Prediction of LncAABR07053481 coding capability by the
   coding-potential assessment tool (CPAT). d The expression of the
   Myc-fused protein was analyzed by immunoblotting with an anti-Myc
   antibody (n = 3). e, f The relationship between the degree of hypoxia
   and apoptosis (n = 4). g Expression of LncAABR07053481 under different
   oxygen concentrations (n = 4). h Detection of LncAABR07053481
   expression in the normal femoral head and the area of osteonecrosis by
   qPCR (n = 6). i Detection of LncAABR07053481 expression in BMSCs at
   different times after transplantation by qPCR (n = 6). Data were shown
   as mean ± SD. ^*P < 0.05; In (h), statistical significance was
   calculated by Student’s t-test; in (f, g, i), statistical significance
   was calculated by one-way ANOVA with Tukey’s post hoc tests.

LncAABR07053481 can inhibit hypoxia-induced apoptosis of BMSCs

   To further clarify the role of LncAABR07053481 differential expression
   in the apoptosis of BMSCs, we transfected BMSCs with
   LncAABR07053481-overexpression lentivirus (Lv-LncAABR07053481) and
   LncAABR07053481-knockdown lentivirus (shLV-LncAABR07053481),
   respectively. The results of qPCR showed that LncAABR07053481 was
   successfully overexpressed or knocked down in the BMSCs (Fig. [94]2a).
   Then, to verify the key role of LncAABR07053481 in the regulation of
   apoptosis in BMSCs, we first knocked down LncAABR07053481 in BMSCs
   under normoxia, and the results showed that, under normoxia, the
   LncAABR07053481-knockdown group (sh [95]AABR07053481) exhibited
   increased BMSC apoptosis compared with the empty vector group (Vec)
   (Fig. [96]2b, c). We next subjected BMSCs to hypoxia (0% O[2]) for
   48 h. Notably, compared with the empty vector group, overexpression of
   LncAABR07053481 in BMSCs (Oe [97]AABR07053481) significantly reduced
   the hypoxia-induced apoptosis of BMSCs (Fig. [98]2d, e). Consistently,
   the protein level of caspase-3 (CASP-3) was also decreased, while the
   expression levels of the anti-apoptosis–related proteins Survivin and
   B-cell lymphoma 2 (Bcl-2) were increased in BMSCs overexpressing
   LncAABR07053481 (Fig. [99]2f–i). Conversely, knockdown of
   LncAABR07053481 in BMSCs (sh [100]AABR07053481) significantly increased
   the level of CASP-3 protein, inhibited the expression of Survivin and
   Bcl-2, and further aggravated the hypoxia-induced apoptosis of BMSCs
   (Fig. [101]2d–i).

Fig. 2. LncAABR07053481 can inhibit hypoxia-induced apoptosis of BMSCs.

   [102]Fig. 2
   [103]Open in a new tab

   a The expression of LncAABR07053481 was detected by qPCR (n = 4). b–e
   Detection of BMSC apoptosis under different conditions by Annexin
   V-fluorescein isothiocyanate (FITC, green)/propidium iodide (PI, red)
   (n = 4). f–i The expression levels of CASP-3, Survivin, and Bcl-2 were
   detected by western blotting (n = 3). j Forty-eight hours after
   surgery, the fluorescence intensity of
   1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyanine iodide (DiR,
   red) in the transplantation area was detected by live imaging of small
   animals (n = 7). k Quantitative analysis of DiR fluorescence intensity
   in the transplanted area (n = 7). l–m At 48 h after surgery,
   TdT-mediated dUTP nick-end labeling (TUNEL, red)
   /4’,6-diamidino-2-phenylindole (DAPI, blue) was used to detect
   apoptosis in the transplantation area (n = 7). Data were shown as mean
    ±  S.D. ^*P < 0.05; in (a, c, e, g–i, k, m), statistical significance
   was calculated by one-way ANOVA with Tukey’s post hoc test.

   To further determine the anti-apoptosis effect of LncAABR07053481 in
   vivo, we overexpressed or knocked down LncAABR07053481 in BMSCs, then
   labeled BMSCs with
   1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyanine iodide (DiR)
   fluorescence to track the viable BMSCs in vivo (i.e., positive
   DiR-labeled cells were viable). Finally, BMSCs were transplanted to
   treat the SANFH model rats, and the survival of BMSCs in vivo was
   detected using the Spectrum IVIS. At 48 h after BMSC transplantation,
   we found that the DiR fluorescence intensity in the operative area of
   the overexpression group (LD/Oe [104]AABR07053481) was significantly
   higher than that in the empty vector group (LD/Vec), while that in the
   knockdown group (LD/sh [105]AABR07053481) was significantly decreased
   (Fig. [106]2j, k). Furthermore, the results of TUNNEL staining showed
   that the LncAABR07053481-overexpression group displayed a significantly
   reduced apoptosis rate of transplanted BMSCs, while LncAABR07053481
   knockdown increased the apoptosis of transplanted BMSCs in the
   transplantation area (Fig. [107]2l, m). Together, these findings
   suggest that LncAABR07053481 can inhibit hypoxia-induced apoptosis of
   transplanted BMSCs in vitro and in vivo.

LncAABR07053481 inhibits hypoxia-induced apoptosis of BMSCs by regulating the
Notch1 pathway

   To establish the mechanism by which LncAABR07053481 inhibits
   hypoxia-induced apoptosis of BMSC, we performed microarray analysis on
   BMSCs overexpressing LncAABR07053481. The results showed that the
   gene-expression profile differed between the
   LncAABR07053481-overexpression group and the empty virus group
   (Fig. [108]3a). KEGG pathway analysis and GSEA demonstrated that these
   differentially expressed genes were significantly enriched in the
   Notch1 pathway. Moreover, the Notch1 pathway had the highest enrichment
   scores, in which Notch1 mRNA was significantly upregulated after
   LncAABR07053481 overexpression (Fig. [109]3b, c). We next verified the
   microarray data using qPCR and western blot analysis, and the results
   showed that the expression of Notch1 and Notch1 intracellular domain
   (NICD1) in BMSCs overexpressing LncAABR07053481 were significantly
   upregulated under hypoxia (Fig. [110]3d–g). Previous studies have shown
   that the activation of the Notch1 signaling triggers the cleavage of
   Notch1 receptors, causing NICD1 (the intracellular active fragment of
   Notch1) to release from the cell membrane and into the nucleus, which
   can inhibit apoptosis by upregulating the transcription and expression
   of Notch1 target genes (Survivin and Bcl-2)^[111]31–[112]34. Therefore,
   these data suggest that LncAABR07053481 promotes the expression of
   anti-apoptotic proteins such as Survivin and Bcl-2 by activating the
   Notch1 pathway and finally inhibits the hypoxia-induced apoptosis of
   BMSCs.

Fig. 3. LncAABR07053481 inhibits hypoxia-induced apoptosis of BMSCs by
regulating the Notch1 pathway.

   [113]Fig. 3
   [114]Open in a new tab

   a mRNA clustering analysis (n = 3). b Kyoto Encyclopedia of Genes and
   Genomes (KEGG) pathway enrichment analysis of differentially expressed
   genes. c Gene set enrichment analysis (GSEA). d The expression of
   Notch1 was detected by qPCR (n = 3). e–g The expression levels of
   Notch1 and NICD1 were detected by western blotting (n = 3). h, i
   Detection of BMSC apoptosis by Annexin V-FITC/PI (n = 3). j–n The
   expression levels of Notch1, NICD1, Survivin, and Bcl-2 were detected
   by western blotting (n = 3). Data were shown as mean  ±  SD.
   ^*P < 0.05; In (d, f, g), statistical significance was calculated by
   Student’s t-test; In (i, k–n), statistical significance was calculated
   by one-way ANOVA with Tukey’s post hoc test.

   Subsequently, we verified the role of the Notch1 pathway in the
   regulation of hypoxia-induced apoptosis of BMSCs. The results showed
   that under hypoxia, overexpression of LncAABR07053481 in BMSCs
   significantly increased the levels of Notch1 and NICD1, upregulated the
   expression of Survivin and Bcl-2, and reduced the apoptosis of BMSCs
   (Fig. [115]3h–n). Knockdown of the Notch1 gene in BMSCs overexpressing
   LncAABR07053481 blocked the anti-apoptosis of LncAABR07053481,
   downregulated the expression of Survivin and Bcl-2, and increased the
   apoptosis of BMSC (Fig. [116]3h–n). By contrast, in the overexpression
   of NICD1 on the basis of LncAABR07053481 knockdown to activate the
   Notch1 pathway, hypoxia-induced apoptosis of BMSCs was significantly
   inhibited, and the expression levels of Survivin and Bcl-2 were
   significantly upregulated (Fig. [117]3h–n). These findings suggest that
   LncAABR07053481 can inhibit hypoxia-induced apoptosis of BMSCs by
   regulating the Notch1 pathway.

LncAABR07053481 functions as a ceRNA and sponges miR-664-2-5p in BMSCs

   To investigate the molecular mechanism by which LncAABR07053481
   regulates the Notch1 pathway, we first confirmed that LncAABR07053481
   was mainly located in the cytoplasm by RNA-FISH (Fig. [118]4a).
   Increasing evidence has shown that classic regulation of LncRNA in the
   cytoplasm occurs as a result of ceRNAs; this removes the silencing
   effect of miRNA on target genes by adsorbing miRNA and finally promotes
   the expression of target genes^[119]35,[120]36. As a result, we
   considered whether LncAABR07053481 regulates and activates Notch1
   through similar mechanisms. To test this hypothesis, we used
   bioinformatics tools, such as Targetscan and RegRNA, to predict the
   possible binding miRNAs of the untranslated region at the 3’ end of
   Notch1 mRNA (Notch1 mRNA 3’UTR) and LncAABR07053481. The results show
   that miR-664-2-5p is a potential target for the combination of
   LncAABR07053481 and Notch1 mRNA 3’UTR (Fig. [121]4b). In addition,
   prediction by bioinformatics tools demonstrated that LncAABR07053481
   and Notch1 mRNA 3′UTR share a consensus binding site on miR-664-2-5p
   (Fig. [122]4c). Therefore, we speculated that LncAABR07053481 acts as
   the ceRNA of miR-664-2-5p.

Fig. 4. LncAABR07053481 functions as a ceRNA and sponges miR-664-2-5p in
BMSCs.

   [123]Fig. 4
   [124]Open in a new tab

   a RNA fluorescence in situ hybridization (RNA-FISH) detection of the
   subcellular localization of LncAABR07053481, 18S, and U6 were used as
   positive control. (n = 4). b Bioinformatics tools predict the miRNAs
   that may bind to the untranslated region at the 3’ end of Notch1 mRNA
   (Notch1 mRNA 3′UTR) and LncAABR07053481. c Binding sites of
   miR-664-2-5p with LncAABR07053481 and Notch1 mRNA 3′UTR. d, e
   Luciferase reporter assay shows that miR-664-2-5p binds to
   LncAABR07053481 and Notch1 mRNA (n = 5). f–h RNA immunoprecipitation
   (RIP) detection of LncAABR07053481 enrichment to miRNA
   ribonucleoprotein complexes (miRNPs) (n = 4). g–i RIP detection of
   enrichment of Notch1 mRNA to miRNPs (n = 4). Data were shown as mean
    ±  SD. ^*P < 0.05; In (d–i), statistical significance was calculated
   by one-way ANOVA with Tukey’s post hoc test.

   Next, we applied a luciferase assay to detect the specific binding
   sites between the LncAABR07053481/Notch1 mRNA 3′UTR and miR-664-2-5p.
   We mutated miR-664-2-5p putative binding sites in LncAABR07053481, and
   then constructed dual-luciferase reporter vectors that contained the
   LncAABR07053481 mutant (LncAABR07053481-Mut) or wild-type
   (LncAABR07053481-Wt). Figure [125]4d shows that the relative luciferase
   activity of LncAABR07053481-Wt in BMSCs was obviously reduced after
   co-transfection of miR-664-2-5p, but did not change the activity of the
   mutant vector, suggesting that miR-664-2-5p is a direct target of
   LncAABR07053481. Similarly, we constructed a luciferase reporter vector
   that contained wild-type or mutant miR-664-2-5p putative binding sites
   in the Notch1 mRNA 3′UTR (Notch1 mRNA 3′UTR-Wt and Notch1 mRNA
   3’UTR-Mut). The alignment of the complementary binding of miR-664-2-5p
   and the 3′UTR of Notch1 was confirmed by the luciferase assay
   (Fig. [126]4e). These results confirmed that miR-664-2-5p could
   specifically bind to LncAABR07053481 and the Notch1 mRNA 3′UTR at the
   predicted binding sites.

   To confirm the competitive relationship between LncAABR07053481 and
   Notch1 mRNA, RIP experiments were conducted to detect the enrichment of
   LncAABR07053481 and Notch1 mRNA to miRNA ribonucleoprotein complexes
   (miRNPs) in BMSC extracts using an AGO2 antibody. The results showed
   that both LncAABR07053481 and Notch1 mRNA were specifically enriched in
   AGO2 antibody-associated complexes, but not in control IgG
   (Fig. [127]4f, g). Moreover, the enrichment of LncAABR07053481 and
   Notch1 mRNA to miRNPs increased after the overexpression of
   miR-664-2-5p, while the enrichment of LncAABR07053481 and Notch1 mRNA
   decreased following the knockdown of miR-664-2-5p (Fig. [128]4f, g). We
   also found that the overexpression of LncAABR07053481 in BMSCs
   significantly increased the enrichment of LncAABR07053481 to miRNPs
   (Fig. [129]4h), but significantly decreased the enrichment of Notch1
   mRNA to miRNPs (Fig. [130]4i). By contrast, overexpression of the 3′UTR
   of Notch1 mRNA increased the enrichment of Notch1 mRNA to miRNPs
   (Fig. [131]4i), while the enrichment of LncAABR07053481 to miRNPs was
   decreased significantly (Fig. [132]4h). Taken together, these data show
   that LncAABR07053481 functions as a ceRNA for miR-664-2-5p in the
   regulation of hypoxia-induced BMSC apoptosis.

The LncAABR07053481/miR-664-2-5p/Notch1 axis is involved in the regulation of
hypoxia-induced apoptosis of BMSCs

   To further verify the regulatory role of the
   LncAABR07053481/miR-664-2-5p/Notch1 axis in hypoxia-induced apoptosis
   of BMSCs, we overexpressed LncAABR07053481 in BMSCs (Fig. [133]5a),
   then subjected the cells to hypoxia for 48 h. The results showed that,
   compared with the empty vector group, the upregulation of
   LncAABR07053481 significantly increased the expression of both Notch1
   and NICD1 (Fig. [134]5d–f) and reduced the hypoxia-induced apoptosis of
   BMSCs (Fig. [135]5g, h). Next, we upregulated the expression of
   miR-664-2-5p again in the BMSCs that overexpressed LncAABR07053481
   (Fig. [136]5b). The results showed that, under hypoxic conditions, the
   expression levels of Notch1 and NICD1 in BMSCs were significantly
   downregulated (Fig. [137]5d–f), while the apoptosis of BMSCs was
   increased (Fig. [138]5g, h). This trend may be because the upregulation
   of miR-664-2-5p increased the enrichment of LncAABR07053481 and Notch1
   to miRNPs, strengthened the silencing of Notch1 mRNA by miR-664-2-5p,
   and significantly inhibited the anti-hypoxia-induced apoptosis of
   LncAABR07053481. Subsequently, we continued to upregulate the
   expression of NICD1 in BMSCs that overexpressed both LncAABR07053481
   and miR-664-2-5p to activate the Notch1 pathway (Fig. [139]5c–f). As a
   result, we found that the apoptosis of BMSCs was decreased
   significantly following the upregulation of NICD1 (Fig. [140]5g, h).
   These results reveal that the upregulation of NICD1 promotes the
   survival of BMSCs under hypoxic conditions. Overall, our data suggest
   that the LncAABR07053481/miR-664-2-5p/Notch1 axis regulates the
   hypoxia-induced apoptosis of BMSCs.

Fig. 5. The LncAABR07053481/miR-664-2-5p/Notch1 axis is involved in the
regulation of hypoxia-induced apoptosis of BMSCs.

   [141]Fig. 5
   [142]Open in a new tab

   a–c The expressions of LncAABR07053481, miR-664-2-5p, and NICD1 were
   detected by qPCR (n = 4). d–f The expression levels of Notch1 and NICD1
   were detected by western blotting (n = 3). g, h Detection of BMSC
   apoptosis under different conditions by TUNEL/DAPI (n = 3). Data were
   shown as mean  ± SD. ^*P < 0.05; In (a–c, e, f, h), statistical
   significance was calculated by one-way ANOVA with Tukey’s post hoc
   test.

LncAABR07053481-overexpressed BMSCs promote the repair of early SANFH

   Determining novel anti-apoptotic targets and inhibiting hypoxia-induced
   apoptosis of BMSCs will help to improve the repair effect of
   transplanted BMSCs in early SANFH. Our data indicate that
   LncAABR07053481 inhibits hypoxia-induced apoptosis of BMSCs in vitro
   and in vivo. However, it is unclear whether the inhibition of the
   hypoxia-induced apoptosis of BMSCs by LncAABR07053481 can improve the
   repair effect in early SANFH. Therefore, we transplanted BMSCs with
   overexpression or knockdown of LncAABR07053481 to repair early SANFH in
   a rat model, then further evaluated the efficacy of
   LncAABR07053481-modified BMSCs on early SANFH. At 12 weeks after BMSC
   transplantation, we evaluated the repair of the osteonecrotic area by
   micro-CT examination. Ultimately, the results showed that, compared
   with the Normal group, there were different degrees of bone defects in
   the lesion debridement group (LD), empty vector group (LD/Vec), and
   knockdown group (LD/sh [143]AABR07053481), and the defect area had not
   been completely repaired. In the LncAABR07053481-overexpression group
   (LD/Oe [144]AABR07053481), the bone defect area had been completely
   filled with bone tissue, and the defect area had been completely
   repaired, with no significant difference compared with the normal group
   (Fig. [145]6a). Similarly, compared with the normal group, the number
   of trabeculae, the trabecular thickness, and the new bone volume
   fraction in the osteonecrotic area of the LD group, LD/Vec group, and
   LD/sh [146]AABR07053481 group were significantly decreased. By
   contrast, compared with the LD group, LD/Vec group, and LD/sh
   [147]AABR07053481 group, the LncAABR07053481-overexpression group
   demonstrated a significantly increased number of trabeculae, trabecular
   thickness, and new bone volume fraction in the osteonecrotic area
   (Fig. [148]6b–d). In addition, the results of H&E and Masson’s staining
   were consistent with those of the micro-CT examination. In the
   LncAABR07053481-overexpression group, the osteonecrotic area was
   completely repaired, and the bone tissue tended to mature
   (Fig. [149]6e). Immunohistochemical detection of osteogenic markers
   showed that, compared with the Normal group, the levels of runt-related
   transcription factor 2 (Runx2) and osteopontin (OPN) were decreased in
   the LD group, LD/Vec group, and LD/sh [150]AABR07053481 group. The
   LncAABR07053481-overexpression group also demonstrated significantly
   increased levels of Runx2 and OPN in the osteonecrotic area compared
   with those of the LD group, LD/Vec group, and LD/sh [151]AABR07053481
   group (Fig. [152]6f–h). These results confirm that LncAABR07053481
   improves the efficacy of transplanted BMSCs in early SANFH.

Fig. 6. LncAABR07053481-overexpressed BMSCs promote the repair of early
SANFH.

   [153]Fig. 6
   [154]Open in a new tab

   a At 12 weeks after BMSC transplantation, micro-CT was used to analyze
   the repair of the necrotic area of the femoral head (n = 6). b–d
   Quantitative analysis of the number of trabeculae, the trabecular
   thickness, and the volume fraction of new bone tissue (n = 6). e At 12
   weeks after BMSC transplantation, H&E and Masson staining were used to
   evaluate the repair of the necrotic area of the femoral head (n = 6). f
   At 12 weeks after BMSC transplantation, immunohistochemistry was used
   to detect osteogenic markers of runt-related transcription factor 2
   (Runx2, Cy3 label, red) and osteopontin (OPN, FITC label, green)
   (n = 6). g Quantitative analysis of Runx2 expression (n = 6). h
   Quantitative analysis of OPN expression (n = 6). Data were shown as
   mean ± SD. ^*P < 0.05; In (b–d, g, h), statistical significance was
   calculated by one-way ANOVA with Tukey’s post hoc test.

Discussion

   BMSC transplantation to repair osteonecrosis or bone defect is a new
   method in the field of orthopedic research, and it brings new hope to
   patients with disabilities such as SANFH^[155]37–[156]39. However, the
   osteonecrotic area is a hypoxic environment, which brings many
   challenges to the survival of BMSCs and is also an important obstacle
   to be solved in BMSC transplantation for
   osteonecrosis^[157]12,[158]27,[159]28,[160]40. It has been found that
   various LncRNAs are abnormally expressed in cells in hypoxic
   conditions, and play important roles in the regulation of cell
   function^[161]29,[162]30. In this study, we found that the expression
   of LncAABR07053481 was significantly downregulated in BMSCs under
   hypoxia, which was closely associated with the hypoxia-induced
   apoptosis of BMSCs. Importantly, when we upregulated LncAABR07053481,
   it was confirmed that LncAABR07053481 inhibited hypoxia-induced
   apoptosis of BMSCs both in vitro and in vivo. Moreover, the
   upregulation of LncAABR07053481 led to an elevated expression of Notch1
   by competitively binding to miR-664-2-5p. Notch1 is an important
   signaling pathway, which affects cell fate and plays important roles in
   regulating biological processes such as
   apoptosis^[163]31–[164]34,[165]41,[166]42. Our findings not only
   suggest an important role of LncAABR07053481 acting as an miRNA sponge
   during the regulation of hypoxia-induced BMSC apoptosis but also
   highlight the therapeutic potential of BMSC transplantation with
   overexpression of LncAABR07053481 in the repair of early SANFH.

   The human genome is pervasively transcribed, but >80% of RNA
   transcripts are not translated into proteins, and these are
   collectively referred to as non-coding genes, of which LncRNAs are the
   main constituent element^[167]43,[168]44. These LncRNAs are not only
   diverse but also complex in function, with roles in many biological
   processes, including apoptosis, proliferation, differentiation, and
   metabolism^[169]45–[170]47. For example, LncMT1DP reduces Nrf2 by
   stabilizing miR-365, a negative regulator of Nrf2, and plays a key role
   in the Cd‐induced apoptosis of hepatocytes^[171]48. Moreover, SNHG12
   acts as a molecular sponge of microRNA to promote tumor proliferation,
   metastasis, and epithelial-mesenchymal transition^[172]49. In addition,
   lncMAR1 has been shown to act as a miR‐487b sponge to regulate Wnt5a
   protein, resulting in muscle differentiation and regeneration^[173]50.
   Similarly, in this study, we discovered a key LncRNA in the BMSC
   hypoxia model, LncAABR07053481, through microarray analysis. Further
   study showed that the expression of LncAABR07053481 was significantly
   downregulated in hypoxia compared with normal BMSCs; moreover, its
   expression was continuously downregulated with the aggravation of
   hypoxia and the concomitantly-increased apoptotic rate. As such, we
   hypothesized that LncAABR07053481 plays a key regulatory role in the
   hypoxia-induced apoptosis of BMSCs. Functional studies revealed that
   overexpression of LncAABR07053481 significantly improved the survival
   rate of BMSCs under hypoxia and inhibited hypoxia-induced apoptosis of
   BMSCs both in vitro and in vivo.

   We next performed a microarray analysis to further study the downstream
   regulation mechanism of LncAABR07053481. The results showed that the
   Notch1 pathway was significantly activated after LncAABR07053481
   overexpression and that LncAABR07053481 was positively correlated with
   the expression of Notch1 and its target genes Survivin and Bcl-2.
   Furthermore, through functional experiments, we verified that
   LncAABR07053481 could inhibit hypoxia-induced apoptosis of BMSCs by
   regulating the expression of Notch1. As typical non-coding genes,
   LncRNAs have complex and diverse biological functions. An increasing
   number of studies have argued that the different biological functions
   of LncRNAs largely depend on their distinct subcellular
   localization^[174]51,[175]52. Accumulated evidence has also shown that
   LncRNAs located in the cytoplasm can participate in gene regulation at
   the posttranscriptional level, including by acting as ceRNAs and
   protecting the target mRNAs from repression^[176]35,[177]53,[178]54. We
   used RNA-FISH assays to show that LncAABR07053481 was preferentially
   localized in the cytoplasm, indicating its potential function as an
   miRNA sponge. Bioinformatics analysis indicated that miR-664-2-5p is a
   target of LncAABR07053481 and the Notch1 mRNA 3′UTR. Expectedly, the
   results of the luciferase assay confirmed that miR-664-2-5p could
   specifically bind to LncAABR07053481 and the Notch1 mRNA 3′UTR at the
   predicted binding sites. We further confirmed the competitive
   relationship between LncAABR07053481 and Notch1 mRNA by RIP assay. Our
   results also showed that the overexpression of LncAABR07053481
   inhibited hypoxia-induced apoptosis of BMSCs in vitro by upregulating
   Notch1 as a ceRNA to sponge miR-664-2-5p. Collectively, our results
   indicate that a strategy targeting LncAABR07053481 may serve as a valid
   therapeutic regimen to inhibit hypoxia-induced apoptosis of BMSCs.

   SANFH is a complex pathological process in which various factors lead
   to intramedullary microvascular lesions, affecting the blood and oxygen
   supply of the femoral head and ultimately causing hypoxia in the
   necrotic area^[179]14–[180]17. The hypoxia apoptosis of transplanted
   BMSCs in the osteonecrotic area limits the transplantation efficacy of
   BMSCs^[181]12,[182]13. Therefore, exploring the intervention targets to
   inhibit the hypoxia-induced apoptosis of BMSCs is key to improving
   transplantation efficacy. In this study, we confirmed that
   LncAABR07053481 can effectively inhibit hypoxia-induced apoptosis of
   BMSCs and improve the survival rate of BMSC transplantation both in
   vitro and in vivo. In view of this, we further evaluated the effect of
   BMSCs overexpressing LncAABR07053481 on early SANFH in rats. The
   results confirmed that LncAABR07053481 overexpression significantly
   inhibited the hypoxia-induced apoptosis of BMSCs after transplantation
   and significantly improved the repair efficacy of BMSCs on early SANFH.
   Therefore, our results show that the LncAABR07053481 intervention
   strategy may represent a novel therapeutic target to effectively
   improve the efficacy of BMSCs in the repair of early SANFH.

   In summary, we found that LncAABR07053481, a significantly
   downregulated LncRNA under hypoxia, can inhibit hypoxia-induced
   apoptosis of BMSCs. LncAABR07053481 acts as a sponge for miR-664-2-5p
   to attenuate its repressive effect on Notch1 and activates the Notch1
   pathway to inhibit hypoxia-induced apoptosis of BMSCs by promoting the
   expression of anti-apoptotic genes. We also systematically evaluated
   the therapeutic effect of LncAABR07053481-overexpressing BMSC
   transplantation on early SANFH, which provides an effective molecular
   strategy and a novel target for improving the transplantation efficacy
   of BMSCs (Fig. [183]7).

Fig. 7. Effect and mechanism of LncAABR07053481 on the repair of early SANFH
by regulating the Notch1 pathway to inhibit hypoxia-induced apoptosis of
BMSCs: LncAABR07053481 acts as a sponge for miR-664-2-5p to attenuate its
repressive effect on Notch1 and activates the Notch1 pathway and releases
NICD1 (the intracellular active fragment of Notch1) from the cell membrane.

   [184]Fig. 7
   [185]Open in a new tab

   NICD1 localizes to the nucleus, where it forms a transcriptionally
   active complex with the DNA-binding protein CSL and the coactivator
   Mastermind (MAM) to upregulate transcription of Notch target genes
   (Survivin and Bcl-2). Following which, hypoxia-induced apoptosis of
   BMSCs in the area of osteonecrosis is inhibited, improving the
   transplantation effect of BMSCs on early SANFH.

Methods

Animals

   All the animals used in this study were provided by the Laboratory
   Animal Center of Guizhou Medical University (Grant No.2100733) and
   approved by the Experimental Animal Bioethics Committee of Guizhou
   Medical University, Guiyang, China. We extracted BMSCs from young
   Sprague-Dawley (SD) rats (n = 18; body weight, 20–30 g), regardless of
   sex, and adult male SD rats (n = 228; body weight, 500–600 g) were used
   to construct SANFH models. All procedures were carried out in strict
   accordance with the guidelines for the Care and Use of Experimental
   Animals issued by the National Institutes of Health (NIH publication
   No. 85-23, 1996 revised).

Cell culture

   The bilateral tibias and femurs of SD rats (n = 18; body weight,
   20–30 g) were obtained under anesthesia. The bone marrow cavity was
   washed with complete l-glutamine Dulbecco’s Modified Eagle Medium
   (L-DMEM) (Gibco, USA) containing 10% fetal bovine serum (Gibco),
   100 U/mL of penicillin, and 100 μg/mL of streptomycin (Hyclone, USA) to
   obtain the mixture containing bone marrow. The mixed solution was
   transferred to a 15-mL tube and centrifuged at 1000 rpm for 5 min. The
   supernatant was discarded to obtain single-cell precipitation, which
   was resuspended with 5 mL of complete L-DMEM, and then transferred to a
   25-cm^2 cell culture bottle. The cells were cultured at 37 °C and 5%
   CO[2]. When the cell confluence reached 90%, the cells were digested
   with 1 mL of 0.25% trypsin–0.02% ethylenediaminetetraacetic acid (EDTA)
   (Gibco) at 37 °C and passaged at a ratio of 1:3.

Differentiation identification of BMSCs

   Third-generation BMSCs were inoculated into six-well plates. When the
   cell fusion degree reached 60,100, and 60% according to the
   instructions of the BMSC osteogenesis induction kit (Cyagen
   Biosciences, USA), the BMSC adipogenic induction kit (Cyagen
   Biosciences) and the BMSC chondrogenic induction kit (Cyagen
   Biosciences), respectively, BMSC differentiation was induced, while the
   control group continued to be cultured with complete L-DMEM. The BMSCs
   were identified by staining at 2, 3, and 4 weeks after differentiation.
   ALP staining was used to identify osteogenic differentiation, oil red O
   staining was used to identify adipogenic differentiation, and Alixin
   blue staining was used to identify chondrogenic differentiation.

Surface antigen identification of BMSCs

   Third-generation BMSCs were obtained, and the cell concentration was
   adjusted to 2.0 × 10^7 cells/mL. The 50 μL cell suspension was fed into
   the flow tube, and 5 μL surface antigen antibodies (Anti-CD29/AF647,
   Anti-CD90/PE-CyTM7, Anti-CD73/PE, Anti-CD45/FITC, and Anti-CD11b/V450)
   (BD, USA) were added in order. Next, 45 μL buffer (Hyclon) was added to
   each tube and incubated at room temperature for 30 min. Following
   incubation, the cells were washed twice with phosphate-buffered saline
   (PBS) (Hyclon), and 500 μL of buffer was added to each tube for
   detection by flow cytometry (Beckman, USA).

Hypoxia cell model

   The proportion of mixed gas in the cell incubator (Thermo, USA) was
   adjusted to an oxygen concentration of 0%, nitrogen concentration of
   95%, and carbon dioxide concentration of 5%. Then, BMSCs were placed in
   the mixed gas for continuous treatment for 48 h to induce hypoxia.

GeneChip microarray assay

   Total RNA in cell samples from both groups was extracted with Trizol
   reagent (Invitrogen, USA). The microarray and data collection were
   conducted by KangChen Biotech (Shanghai, China), as mentioned
   previously^[186]55. The Agilent Feature Extraction software program
   (version 11.0.1.1; Agilent Technologies, USA) was used to analyze
   acquired array images. Quantile normalization and subsequent data
   processing were performed using the GeneSpring GX v12.1 software
   package (Agilent Technologies). After quantile normalization of the raw
   data, LncRNAs and messenger RNAs (mRNAs), among which ≥3 out of nine
   samples had flags in the normoxia or hypoxia group, were chosen for
   further data analysis. A paired t-test was employed to assess the
   differentially expressed LncRNAs and mRNAs between the experimental and
   control groups. Differentially expressed LncRNAs and mRNAs between the
   two groups were identified through fold change filtering (fold change
   >2.0). Differentially expressed LncRNAs and mRNAs with statistical
   significance between the two groups were identified through P value
   filtering (P < 0.05). Data visualization of LncRNA and mRNA expression
   levels was achieved via a heatmap using the pheatmap R package.

Bioinformatics analysis

   MicroRNA (miRNA) binding sites on LncRNAs or mRNAs were predicted using
   the web-based program RNAhydrid
   ([187]https://bibiserv.cebitec.uni-bielefeld.de/rnahybrid/). The Gene
   Set Enrichment Analysis (GSEA) version 4.1.0 tool (Broad Institute,
   USA; [188]http://www.gsea-msigdb.org/gsea/index.jsp) was used for
   pathway analysis of gene-expression data to find enriched Kyoto
   Encyclopedia of Genes and Genomes (KEGG) pathways. GSEA signatures were
   assessed using the Molecular Signatures Database version 6.2, and all
   processing parameters used were the defaults adopted in GSEA version
   4.1.0.

LncRNA coding capacity prediction

   The Coding-Potential Assessment Tool (CPAT)
   ([189]http://lilab.research.bcm.edu/cpat/) was used to assess the
   coding capacity of LncAABR07053481. The CPAT conducted a logistic
   regression model using the sequence features of open reading frame
   coverage, open reading frame size, hexamer usage bias, and Fickett
   TESTCODE statistic. The CPAT selected 0.364 as a cutoff for the coding
   probability (CP); a CP of <0.364 suggests a non-coding sequence,
   whereas that of ≥0.364 indicates a coding sequence^[190]56.

In vitro translation assay

   To validate that LncAABR07053481 is indeed a non-coding RNA, we
   performed in vitro translation experiments. Briefly, the
   LncAABR07053481 transcript was cloned into pcDNA4-myc-His-plasmid to
   obtain the recombinant, and pcDNA4/LncAABR07053481-myc-His-plasmid was
   transfected into BMSCs using eukaryotic transduction kit (Thermo) for
   48 h. Meanwhile, pcDNA4/KLF4-myc-His was used as a coding protein
   control. The expression of Myc-fused protein was analyzed by
   immunoblotting with an anti-Myc antibody (Abcam ab32072, 1:1000).

Lentiviral (LV) vector production and cell transfection

   The LV vector that knocked down or overexpressed LncAABR07053481 was
   purchased from Genechem (Shanghai, China). Briefly, the LV (sh
   [191]AABR07053481 or Oe [192]AABR07053481) constructs were generated
   based on different regions of the Rattus LncAABR07053481 sequence (NCBI
   accession [193]XR_005502377.1). The same LV vector containing an insert
   of nonspecific RNA oligonucleotide (sh-NC or Oe-NC) was used as a
   negative control. Lentiviruses containing these plasmids were generated
   by transfecting the 293 T packaging cell line with the shRNA vector
   (GU6-LncAABR07053481-GFP-puro) and overexpression vector
   (Ubi-LncAABR07053481-GFP-puro). Cell transfection: third-generation
   BMSCs were divided into groups according to transfection conditions. LV
   vectors (MOI = 90) were added into the BMSCs culture medium. After
   12 h, the culture medium was changed to complete L-DMEM. Three days
   following the infection, the expression of the green fluorescent
   protein was observed using an inverted fluorescence microscope, and the
   transfection efficiency was calculated. Five days after transfection,
   the cells were continuously selected with a complete medium containing
   2 μg/mL of puromycin (Solarbio, Beijing, China), and the efficiency of
   target gene knockdown or overexpression was detected by quantitative
   real-time polymerase chain reaction (qPCR).

Reverse transcription and qPCR

   The total RNA was isolated from bone tissue or cells with the TRIzol
   reagent. Tissue samples: bone forceps crush frozen bone tissue and
   place it in a mortar, then add liquid nitrogen and grind the pieces
   into a powder. RNA was isolated from tissue powder frozen on liquid
   nitrogen by homogenization in Trizol Reagent. Cell samples: RNA was
   isolated by direct lysis of cells by Trizol Reagent. RNA extraction was
   performed using an RNAeasy RNA extraction kit (Sangon Biotech,
   Shanghai). RNA was quantified using a Nanodrop instrument. cDNAs were
   synthesized by PrimeScriptTM RT reagent kit (Sangon Biotech). qPCR was
   performed using the SYBR Premix Ex Taq II kit (Sangon Biotech) and the
   Applied Biosystems 7500 Fluorescent Quantitative PCR system (Applied
   Biosystems Life Technologies, USA). GAPDH was used to normalize the
   expression levels of LncAABR07053481, Notch1, and miR-664-2-5p. The
   transcript levels were analyzed by the 2^−ΔΔCt method. Primers (Sangon
   Biotech) were as follows: LncAABR07053481-F: CATGGAAGGGCTTGCTCCAT;
   LncAABR07053481-R: CCAGCTGAACCACCTCGAAT; Notch1-F: TGAAGGAACGAGCCTGGGT;
   Notch1-R: CATTCAGGCAGGTCCCACTT; miR-664-2-5p-RT:
   CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAGCCCAATC;
   miR-664-2-5p-F:TCGGCAGGCTGGCTGGGGAAAA;
   miR-664-2-5p-R:CGTGGAGTCGGCAATTCAGTTGA.

Western blot analysis

   Cellular protein was extracted with 1 × cell lysis buffer (Beyotime,
   China). Equal amounts of proteins obtained from different types of cell
   lysates were separated by 10 or 15% SDS-PAGE (Beyotime), transferred to
   PVDF membranes, and subjected to western blotting using an ECL
   chemiluminescence reagent (Merck Millipore). Antibodies for P53 (Abcam
   ab131442, 1:1000), Bcl-2 (Abcam ab196495, 1:1000), Bid (Abcam ab272880,
   1:2000), Cleaved-CASP-3 (Abcam ab13847, 1:500), Survivin (Abcam
   ab134170, 1:1000), myc (Abcam ab32072, 1:1000), β-actin (Abcam ab8227,
   1:3000), and IgG (Abcam ab6795, 1:2000) were purchased from Abcam
   (Cambridge, USA). Antibodies for Notch1 (Cell Signaling Technology
   3608, 1:1000) and NICD1 (Cell Signaling Technology 4147, 1:1000) were
   purchased from Cell Signaling Technology (Boston, USA).

RNA fluorescence in situ hybridization (RNA-FISH)

   Cells were seeded on coverslips, fixed with 4% paraformaldehyde
   (Solarbio) for 10 min at room temperature, and permeabilized in 70%
   ethyl alcohol (Solarbio) at 4 °C at least 1 h. Hybridization was
   performed using the LncRNA Fish Probe Mix (Ribobio, China) in a moist
   chamber at 37 °C for 12–16 h. As positive controls, 18S and U6
   (Ribobio) were used for the cytoplasm and nucleus, respectively. Image
   sections were acquired using laser confocal microscopy (Leica SP5,
   Heidelberg, Germany).

Luciferase assay

   As mentioned previously, after BMSCs were transfected with lentiviruses
   containing the luciferase reporter gene and lysed with cell lysate
   according to the instructions of the luciferase assay kit (Solarbio),
   the renilla luciferase reaction substrate was added. Finally, the
   luciferase activity was measured using a Dual-Luciferase reporter assay
   system (Promega, USA).

RNA immunoprecipitation (RIP) assay

   A Magna RIP RNA-Binding Protein Immunoprecipitation Kit (Millipore,
   USA) was used to determine the relationship between LncAABR07053481 or
   Notch1 mRNA and miR-664-2-5p. Briefly, the cells were harvested, lysed,
   and then reacted with RIP buffer containing magnetic beads conjugated
   with anti-Argonaute2 (AGO2) and negative control immunoglobulin G (IgG)
   (Millipore). Following the recovery of the antibody using protein A/G
   beads, the coprecipitated RNAs were used for complementary DNA
   synthesis and evaluated by qPCR.

ROS determination

   Cells grown in confocal Petri dishes were treated according to the
   conditions of each group. A DHE (Sigma, USA) probe was then added to
   the cells and incubated for an additional 30 min. Finally, the
   generation of intracellular ROS was assessed by fluorescence analysis
   using laser confocal microscopy.

Mitochondrial membrane potential detection

   According to the instructions of the JC-1 mitochondrial membrane
   potential detection kit (KeyGen Biotech, China), the treated cells were
   incubated with the dye mixture at 37 °C for 30 min and then washed with
   PBS three times. The cells were observed under a laser confocal
   microscopy and recorded by a random typical visual field. The red:green
   fluorescence ratio was calculated by ImageJ software (version
   1.4.3.67).

ATP levels detection

   The BMSCs were collected according to the experimental groups and the
   instructions of the ATP Assay Kit (Beyotime). An appropriate lysate was
   added to the samples of each group, before centrifuging at 12,000×g/min
   for 5 min at 4 °C. The supernatant was collected as the sample to be
   tested. The standard and test solutions were prepared according to the
   instructions of the kit. Then, 20 μL test or standard samples were
   added successively to each well of a 96-well plate in triplicate. Next,
   100 μL of the detection-working solution was added, mixed well, and
   incubated for 25–30 min at 37 °C or room temperature according to the
   kit instructions. Finally, the absorbance at 532 nm was detected by an
   enzyme-labeling instrument (Biotech, USA). The standard curve was drawn
   according to the absorbance value, and the concentration or activity of
   the sample was calculated using the standard curve.

Flow cytometry assay

   Cells were cultured in a 6-well culture plate, then subjected for 48 h
   to hypoxia or normoxia conditions when the cells grew to 85%. The cells
   were washed with phosphate-buffered saline, and 5 μL of Annexin
   V-fluorescein isothiocyanate (FITC) and 5 μL of propidium iodide (PI)
   were directly added according to the instructions of the Annexin V-FITC
   apoptosis detection kit (BD). The cells were gently vortexed and
   incubated at room temperature in the dark for 15 min, then detected by
   flow cytometry.

TUNNEL/DAPI detection of apoptosis

   The treated cells were digested and centrifuged, and the cell density
   was adjusted to 2.5 × 10^4/mL before their inoculation in a confocal
   Petri dish. The cells of each group were treated strictly in accordance
   with the instructions of the TUNEL (Beyotime) and DAPI (Solarbio)
   detection kits. The samples were observed under a confocal microscope,
   and images were collected.

SANFH animal model

   SANFH models were established as previously described in ref. ^[194]13.
   Briefly, adult male SD rats (n = 228; body weight, 500–600 g) were used
   to induce steroid-associated osteonecrosis with sequential injections
   of lipopolysaccharides and methylprednisolone. The rats were
   intravenously injected twice with lipopolysaccharides (2 mg/kg; Sigma)
   (the injection time was >30 min) once a day. From day 2, after the
   second injection of lipopolysaccharides, methylprednisolone (60 mg/kg;
   Pfizer, USA) was injected into the gluteal muscle of rats for 7
   continuous days. Rats in the control group were injected with saline.
   Eight weeks after modeling, magnetic resonance imaging (MRI) was used
   to evaluate the necrosis of the femoral head area, and the successful
   models were used for in vivo experiments.

Construction and transplantation of tissue-engineered bone

   The xenogeneic antigen-extracted cancellous bone (XACB) (Yapeng
   Biological Technology, Shanghai, China) was made into a small cylinder
   with a diameter of 2 mm and a height of 2 mm, before soaking in a
   complete L-DMEM medium for 24 h. After that, 500 μL of BMSC suspension
   (1 × 10^7 cells/mL) was dripped into XACB and cultured in a 37 °C and
   5% CO[2] incubator for 3 h. Following incubation, 3–4 mL of complete
   L-DMEM was added to Petri dishes containing XACB before proceeding with
   the culture and changing the medium once every 2–3 days. On the fifth
   day of culture, the growth of BMSCs on the surface of XACB was observed
   by scanning electron microscopy (Hitachi, Japan). In the same way,
   BMSCs overexpressing or with knockdown of LncAABR07053481 were
   co-cultured with XACB to construct tissue-engineered bone. One side of
   the femoral head of the anesthetized rat model was exposed under
   sterile conditions. At the junction of bone and cartilage, a sterile
   spherical drill with a diameter of 2 mm was used to grind from the
   posterolateral to the anterior medial side at a depth of ~3 mm. Then,
   the necrotic bone tissue was scraped off, and the constructed
   tissue-engineered bone was implanted into the bone defect area. There
   were ten rats in each group; the normal rats (n = 28), rats with simple
   necrotic bone curettage (n = 40), and rats implanted with
   tissue-engineered bone with empty vector (n = 40) were used as
   controls.

Oxygen concentration detection

   The oxygen concentration was measured using a needle-type oxygen
   electrode (Hansatech Instruments, Norfolk, UK) and analyzed with the
   OxygraphPlus software program (Hansatech Instruments), as previously
   described in ref. ^[195]57. Briefly, the oxygen electrode was
   calibrated at 37 °C with 0 and 100% air saturation (i.e., 20.9% oxygen
   of all dissolved gas by volume). Measurements were performed according
   to the manufacturer’s instructions. Oxygen measurements were made by
   penetrating the bone tissue with the oxygen electrode tip. Upon
   introduction into the tissue, the microelectrodes responded with a time
   constant that was estimated to be of the order of 10 s. A stable
   reading was obtained within 30 s, and upon reaching the current plateau
   value, the electrode was extracted from the bone tissue and the tip was
   maintained within the suffusing saline solution.

Live imaging of animals

   Before BMSC transplantation, we labeled live-BMSCs with DiR fluorescent
   dye (YEASEN, Shanghai, China) and co-cultured them with XACB. At 48 h
   after BMSC transplantation, the rats were anesthetized by an
   intraperitoneal injection of chloral hydrate (10%, 3 mL/kg) (Leagene,
   Beijing, China); then, non-invasive in vivo multispectral fluorescence
   imaging was performed using the Spectrum in vivo imaging system (IVIS)
   (PerkinElmer, USA) with the Living Image software (version 4.4).

Hematoxylin and eosin (H&E) and Masson staining

   Whole intact femurs were dissected from rats, fixed in
   phosphate-buffered paraformaldehyde (Solarbio) for 48 h, then
   decalcified for 2 weeks in 10% EDTA (Solarbio). Subsequently, the
   samples of each group were dehydrated with ethanol (Solarbio)
   solutions, adopting gradient concentrations of 50, 70, 80, 90, 95, and
   100%. Finally, the femur tissue samples were embedded in paraffin and
   cut into 4-μm-thick sections using an ultra-thin microtome (Leica,
   Germany). The sections were stained according to the instructions of
   the H&E staining kit (Solarbio) and the Masson staining kit (Solarbio).
   Following staining, the sections were observed under the microscope and
   imaged.

Micro-computed tomography (CT) analysis

   The femurs were dissected from rats and fixed with 4% paraformaldehyde
   for 24 h, before being scanned and analyzed by high-resolution micro-CT
   (Skyscan 1172, Germany). We used the NRecon image reconstruction
   software version 1.6, CTAn data analysis software version 1.9, and
   CTVol 3D model visualization software version 2.0 to analyze the
   parameters of the trabecular bone in the transplantation area. The
   scanner was set at 50 kVp and 201 μA, with a resolution of 6 μm. The
   region of interest was analyzed to determine the trabecular number
   (Tb.N), trabecular thickness (Tb.Th), new bone volume fraction (BV/TV),
   trabecular number/trabecular separation, and thickness per tissue
   volume.

Immunofluorescence (IF) assay

   After treatment, femoral head tissue was fixed in 4% formaldehyde,
   dehydrated by ethanol gradient (Solarbio), decalcified by decalcified
   solution (Solarbio), embedded in paraffin, and finally cut into
   4-μm-thick sections. Then, the tissues were incubated with anti-Runx2
   or anti-OPN overnight at 4 °C. The next day, the cells were incubated
   with secondary antibodies conjugated to FITC and Cy3 for 30 min at
   37 °C. For nuclear staining, the cells were incubated with DAPI for
   10 min. Fluorescence was visualized using laser confocal microscopy.

Statistics and reproducibility

   All statistical analyses were conducted using SPSS version 20.0 (IBM
   Corporation, USA) or GraphPad Prism version 7 (GraphPad Software, USA).
   Each experiment was performed at least in triplicate. Parametric or
   non-parametric statistical tests were selected by determining first if
   the data were normally distributed using the Kolmogorov–Smirnov
   normality test. If the data followed a normal distribution and their
   variance was uniform, then the data were presented as mean ± standard
   deviation values (SD), and statistical significance was determined
   using two-tailed unpaired Student’s t-tests for comparison between two
   groups or one-way ANOVA followed by Tukey’s post hoc test for multiple
   comparisons (≥3 groups). If the data did not follow a normal
   distribution pattern, they were described by M(P25 and P75) values, and
   the Kruskal–Wallis rank-sum test was used for comparisons between
   groups. If the difference between groups was statistically significant,
   the Dwass–Steel–Critchlow–Fligner method was further used for multiple
   comparisons. P < 0.05 was considered to be statistically significant.

Reporting summary

   Further information on research design is available in the [196]Nature
   Portfolio Reporting Summary linked to this article.

Supplementary information

   [197]Peer Review File^ (1.8MB, pdf)
   [198]Supplementary Information^ (9.3MB, pdf)
   [199]42003_2023_4661_MOESM3_ESM.pdf^ (68.2KB, pdf)

   Description of Additional Supplementary Files
   [200]Reporting Summary^ (2.1MB, pdf)
   [201]Supplementary Data^ (41.6KB, xlsx)

Acknowledgements