Abstract

   Calcification of autologous pathological vessels and tissue engineering
   blood vessels (TEBVs) is a thorny problem in clinic. However, there is
   no effective and noninvasive treatment that is available against the
   calcification of TEBVs and autologous pathological vessels. Gli1^+
   cells are progenitors of smooth muscle cells (SMCs) and can
   differentiate into osteoblast-like cells, leading to vascular
   calcification. Our results showed that the spatiotemporal distribution
   of Gli1^+ cells in TEBVs was positively correlated with the degree of
   TEBV calcification. An anticalcification approach was designed
   consisting of exosomes derived from mesenchymal stem cells delivering
   lncRNA-ANCR to construct the engineered exosome-Ancr/E7-EXO. The
   results showed that Ancr/E7-EXO effectively targeted Gli1^+ cells,
   promoting rapid SMC reconstruction and markedly inhibiting Gli1^+ cell
   differentiation into osteoblast-like cells. Moreover, Ancr/E7-EXO
   significantly inhibited vascular calcification caused by chronic kidney
   disease. Therefore, Ancr/E7-EXO reprogrammed Gli1^+ cells to prevent
   calcification of vascular graft and autologous pathological vessel,
   providing unique insights for an effective anticalcification.
     __________________________________________________________________

   Engineered exosomes reprogrammed Gli1^+ cells to inhibit calcification
   in TEVBs and autologous vessels.

INTRODUCTION

   The morbidity and mortality of cardiovascular diseases are increasing
   year by year worldwide ([50]1). Vascular calcification, as an
   independent risk factor of cardiovascular death, is an important
   predictor of cardiovascular disease incidence and mortality, which
   often leads to vascular wall sclerosis, lumen stenosis, hemodynamic
   damage, and, lastly, circulatory obstruction ([51]2). Vascular
   calcification is the most common complication and common pathological
   change of diabetes, chronic kidney disease (CKD), and atherosclerosis
   ([52]3), because the related metabolic and inflammatory factors of
   chronic disease can accelerate vascular calcification ([53]4). For the
   early vascular calcification in vivo, there is no effective reversal
   intervention in clinic ([54]5). With the aggravation of vascular
   calcification, patients often need to undergo vascular replacement, and
   tissue engineering blood vessels (TEBVs) are a promising biological
   vascular substitute ([55]6). However, when TEBVs are implanted into
   patients with underlying diseases such as diabetes, calcification is
   also easy to occur in grafts and leads to implantation failure ([56]7).
   For TEBVs, on the one hand, it is necessary to promote the rapid and
   normal reconstruction of SMCs to resist the occurrence of aneurysms
   ([57]8); on the other hand, it is necessary to inhibit calcification to
   maintain long-term homeostasis and vascular patency ([58]3, [59]9). At
   present, there is no noninvasive treatment for vascular calcification
   under the above pathological conditions. Therefore, exploring effective
   strategies to inhibit calcification is substantial to autologous
   vascular calcification caused by chronic diseases and vascular grafts.

   More and more studies have shown that endogenous vascular progenitor
   cells play a role in the occurrence and development of vascular
   remodeling diseases under pathological conditions ([60]3). They may
   serve as repositories of primitive progenitor cells and may be involved
   in vascular calcification and chondrogenic metaplasia in vessels
   ([61]10). Gli1^+ cells have mesenchymal stem cell (MSC) morphology and
   typical surface markers, have the ability of three-lineage
   differentiation, form an extensive perivascular network in tissue, and
   exist in adventitia of vessels ([62]11, [63]12). Studies have shown
   that Gli1^+ cells promote the formation of regenerative type H vessels
   for tissue regeneration and repair ([64]13). Recent studies have shown
   that Gli1^+ cells are the progenitors of vascular smooth muscle cells
   (VSMCs) and differentiate into highly differentiated SMCs in acute
   vascular injury, while, in chronic vascular injury diseases, Gli1^+
   cells can differentiate into osteoblast-like cells, which may cause
   vascular calcification ([65]14). Therefore, targeting Gli1^+ cells and
   reprogramming in vivo may be an interesting way to treat vascular
   calcification.

   Exosomes are extracellular vesicles with a diameter of 30 to 100 nm
   that can carry RNA and affect cellular function by transmitting
   information between cells and regulating intercellular communication
   ([66]15). Exosomes derived from MSCs promote vascular repair after
   injury by secreting cytokines and inducing cell differentiation and are
   widely used in cardiovascular diseases ([67]16, [68]17). At the same
   time, exosomes are very promising nanovesicles in regenerative medicine
   by carrying drugs or nucleic acids for the regulation of tissue injury
   repair and regeneration, showing safety and effectiveness in vivo
   ([69]15). LncRNA-ANCR has shown the ability to inhibit osteoblast
   differentiation in previous studies. During the differentiation of
   human osteoblasts, the expression of lncRNA-ANCR decreased
   significantly, while its expression inhibits the differentiation of
   human osteoblasts by inhibiting Runt-related transcription factor2
   (Runx2) expression through the recruitment of the enhancer EZH2
   ([70]18). LncRNA-ANCR silencing can promote osteogenesis of
   postmenopausal osteoblasts both in vitro and in vivo ([71]19).

   Inspired by the potential of exosomes to load and deliver signal
   molecules, we hope to combine exosome therapy with long noncoding RNA
   (lncRNA) therapy through engineering method and regulate the
   differentiation trend of Gli1^+ cells by loading lncRNA-ANCR and
   targeting Gli1^+ cells. The surface modification of exosomes can
   achieve targeting in vivo. Previous studies have shown that E7 peptide
   has a special affinity for MSCs and can promote MSC homing ([72]20).
   Therefore, in this study, we designed and constructed Lamp2b-E7
   plasmid, transfected it into human bone marrow MSCs (HuBM-MSCs), and
   introduced lncRNA-ANCR into HuBM-MSC–derived E7-EXO by electroporation
   to construct an engineering exosome-Ancr/E7-EXO. Through tail vein
   injection, Ancr/E7-EXO can recruit the site of vascular calcification
   in CKD rats and then inhibit the osteogenic differentiation of Gli1^+
   cells. The modification of Ancr/E7-EXO on TEBVs by electrostatic
   self-assembly can regulate the differentiation of Gli1^+ cells into
   contractile SMCs in a high-glucose environment, promote the
   reconstruction of SMCs in TEBVs, and effectively inhibit calcification
   in vivo.

   The scheme of engineered exosome-Ancr/E7-EXO reprogramed Gli1^+ cells
   to prevent calcification of vascular grafts and autologous pathological
   vessels in vivo. Construction of the engineered exosomes was loaded
   with ANCR and surface-modified E7 peptide Ancr/E7-EXO ([73]Fig. 1A).
   Ancr/E7-EXO was modified to TEBVs by self-assembly, regulating the
   differentiation of Gli1^+ cells into contractile SMCs and inhibiting
   vascular grafts calcification in vivo ([74]Fig. 1B). Ancr/E7-EXO
   injected through the tail vein of CKD rats accumulated in the
   calcification sites due to Gli1^+ cells, reversed the osteogenic
   differentiation of Gli1^+ cells, and inhibited autologous vascular
   calcification ([75]Fig. 1C).

Fig. 1. Scheme of engineered exosome-Ancr/E7-EXO reprogramming Gli1^+ cells
to prevent calcification of vascular grafts and autologous pathological
vessels in vivo.

   [76]Fig. 1.
   [77]Open in a new tab

   (A) Construction of the engineered exosomes loaded with ANCR and
   surface-modified E7 peptide Ancr/E7-EXO. (B) Ancr/E7-EXO was modified
   to TEBVs by self-assembly, regulating the differentiation of Gli1^+
   cells into contractile SMCs and inhibiting vascular grafts
   calcification in vivo. (C) Ancr/E7-EXO injected through the tail vein
   of CKD rats accumulated in the calcification sites due to Gli1^+ cells,
   reversed the osteogenic differentiation of Gli1^+ cells, and inhibited
   autologous vascular calcification.

RESULTS

Tracing of the spatiotemporal distribution of Gli1^+ cells in TEBVs and
autologous pathological vessels

   TEBVs are prone to calcification after implantation, which is
   accelerated in high-glucose conditions. Therefore, Alizarin red and von
   Kossa staining of TEBVs were performed at different times after TEBV
   implantation in normal and diabetic rats. The staining revealed that
   the calcified nodules were present in two groups, and a full-layer
   calcification appeared in diabetic rats, with severe lumen stenosis at
   the sixth weeks (fig. S1, A and B). These results suggested that TEBVs
   were prone to calcification in both normal and pathological conditions.
   Gli1^+ cells exist in the adventitia of artery and are considered to be
   the progenitor cells of VSMCs, but it is not clear whether they can
   dedifferentiate into osteoblasts under pathological conditions in
   TEBVs. Therefore, we tracked the spatial distribution of Gli1^+ cells
   in TEBVs at each time point after implantation. Gli1^+ cells in TEBVs
   at each time point after implantation increased with time, but the
   diabetes group had more Gli1^+ cells than normal group ([78]Fig. 2A).
   The total number of Gli1^+ cells in the diabetic group at 14 days was
   approximately 2.65-fold than that in the normal group ([79]Fig. 2B).
   The amount of Gli1^+ cells in different spatial locations still showed
   a similar trend ([80]Fig. 2, C to E). The number of Gli1^+ cells in the
   media of TEBVs in the diabetic group was 1.38 times of that in the
   normal group ([81]Fig. 2D). Gli1^+ cells in the intima of TEBVs in the
   diabetic group were about fourfold than that in the normal group. We
   also observed that there were more Gli1^+ and osteopontin (OPN) signal
   coexpression (fig. S2A) in the adventitia of TEBVs at the early stage
   of implantation, and the coexpression signal increased with the
   development of diabetes (fig. S2, B and C). Therefore, the above
   results showed the spatiotemporal distribution of Gli1^+ cells in TEBVs
   under normal physiological and pathological conditions, which was
   positively correlated with the degree of TEBV calcification, suggesting
   that they might be the main cellular source of calcification in TEBVs,
   as confirmed. In addition, the role of Gli1^+ cells was also evaluated
   in the model of chronic vascular calcification caused by CKD. The
   development of CKD due to the loss of kidney tissue was monitored by
   detecting the increased level of blood urea nitrogen (BUN)
   ([82]Fig. 2H). Fluorescent reflection imaging (FRI) showed that
   hydroxyapatite formation was detected in aortic and thoracic aorta
   ([83]Fig. 2F); the fluorescence intensity in the CKD group was 2.14
   times higher than that in the sham group ([84]Fig. 2G). The
   cross-sectional images of the aorta showed that the amount of Gli1^+
   cells in the CKD group was eightfold than that in the sham group
   ([85]Fig. 2, I and J). A positive correlation between the fluorescence
   intensity of hydroxyapatite and Gli1^+ cells was found. These results
   suggested that Gli1^+ cells might be the main cellular source of
   calcification in chronic vascular injury diseases in vivo.

Fig. 2. Spatiotemporal distribution of Gli1^+ cells in TEBVs and CKD-induced
calcified vessels.

   [86]Fig. 2.
   [87]Open in a new tab

   (A) Immunofluorescence staining in the cross-sectional area of TEBVs in
   each group after TEBV implantation [4′,6-diamidino-2-phenylindole
   (DAPI), blue; Gli1^+, red]. The lines divide the vascular adventitia
   (A), media (M), and intima (I). Scale bars, 100 μm. Higher
   magnification of the yellow line border. Scale bars, 20 μm. (B to E)
   Number of total Gli1^+ cells, as well as intima, media, and adventitia
   Gli1^+cells in TEBVs (n = 3). (F and G) Representative images and
   quantification of hydroxyapatite fluorescence intensity of the aortic
   segment in CKD and sham group (n = 3) at 13 weeks after subtotal
   nephrectomy (5/6Nx). (H) Blood urea nitrogen (BUN) stratified for the
   sham and CDK group. (I and J) Representative image and quantification
   of CKD-induced calcified vessels in the cross-sectional area of the
   aorta in rats by immunofluorescence staining (DAPI, blue; Gli1^+, red).
   Results are expressed as means ± SD by two-tailed Student’s t test. *P
   < 0.05, **P < 0.01, ***P <0.001, and ****P <0.0001. wk, weeks.

Construction and characterization of Ancr/E7-EXO targeting Gli1^+ cells

   The exosomes from HuBM-MSCs were isolated by ultracentrifugation,
   called MSC-EXO, to obtain engineered exosomes with the ability to
   target Gli1^+ cells. They were characterized by typical saucer-like
   structure, surface markers, and particle size (fig. S3, A to C).
   MSC-EXO was modified with the following schematic modification
   ([88]Fig. 3A). First, the MSC affinity peptide-E7 was conjugated with
   lamp-2b to avoid being cut off and to increase its expression on the
   exosome membrane after E7 plasmid transfected into HuBM-MSCs. An amino
   acid sequence (MLKPLEQ) was also designed with the same amino acid as
   E7 but in a scrambled order, called M7 peptide, and used as a negative
   control for E7. After the E7 and M7 plasmid were transfected HuBM-MSCs,
   Western blot indicated that the expression of Lamp2b on the membrane of
   E7-EXO and M7-EXO and in HuBM-MSCs was increased ([89]Fig. 3C).
   Real-time polymerase chain reaction (PCR) also suggested that E7
   plasmid was successfully transfected into HuBM-MSCs and expressed
   ([90]Fig. 3D). Scanning electron microscopy (SEM) image showed a
   typical oval structure of E7-EXO ([91]Fig. 3B). Next, MSC-EXO, E7-EXO,
   and M7-EXO were labeled with PKH67 and incubated with Gli1^+ cells at
   different times to determine the affinity of E7-EXO for Gli1^+ cells.
   The results revealed that E7-EXO was efficiently internalized into
   Gli1^+ cells (fig. S3, D and E). After carotid artery replacement,
   TEBVs come into direct contact with blood, which induces neutrophil
   recruitment and subsequently recruits proinflammatory macrophage
   infiltration. Although the E7-EXO has the affinity of Gli1^+ cells, it
   is impossible to predict its ability to deal with inflammatory cells,
   so we carried out coculture experiments ([92]Fig. 3I). The positive
   rate of E7-EXO phagocytosis by Gli1^+ cells was approximately 3.43
   times higher than that of macrophages ([93]Fig. 3, I to K). In
   addition, we also cocultured Gli1^+ cells with endothelial cells (ECs),
   SMCs, and neutrophils (fig. S3, F to K). The above results demonstrated
   that E7-EXO had a specific affinity with Gli1^+ cells. The ANCR plasmid
   was loaded into E7-EXO through electroporation at 100 V and 50 μF,
   making the ANCR plasmid retained to the maximum extent. The E7-EXO
   after successful electroporation was called Ancr/E7-EXO. The loading
   efficacy was tested by quantitative PCR (qPCR), and the results showed
   that the Ancr/E7-EXO loaded 2.95 times more ANCR than MSC-EXO
   ([94]Fig. 3F). Approximately 1% of ANCR plasmids were loaded into
   Ancr/E7-EXO, meaning that each Ancr/E7-EXO loaded 100 ANCR plasmids.
   The SEM images and nanoparticle size of Ancr/E7-EXO were similar to
   those of MSC-EXO ([95]Fig. 3, E, G, and H). These results demonstrated
   that the electroporation did not change the physical properties and
   stability of E7-EXO, proving the successful construction of the
   Ancr/E7-EXO targeting Gli1^+ cells.

Fig. 3. Synthesis and characterization of Ancr/E7-EXO.

   [96]Fig. 3.
   [97]Open in a new tab

   (A) Scheme representing Ancr/E7-EXO construction. (B) Scanning electron
   microscopy (SEM) image of E7-EXO. Scale bar, 100 nm. (C) Western blot
   analysis of CD63 and Lamp2b of transfected HuBM-MSCs and
   HuBM-MSC–derived exosomes. GAPDH, glyceraldehyde-3-phosphate
   dehydrogenase. (D) The quantitative expression of Lamp2b-E7 in
   transfected HuBM-MSCs was compared with that in HuBM-MSCs. (E) SEM
   images of Ancr/E7-EXO. Scale bar: 100 nm. (F) qPCR showing ANCR
   expression in the Ancr/E7-EXO. (G and H) Zeta potential (G) and
   particle size (H) of Ancr/E7-EXO and MSC-EXO. (I to K)
   Immunofluorescence staining and quantification of the phagocytic
   positive rate in the coculture experiment (n = 5). Results are
   expressed as means ± SD by two-tailed Student’s t test. **P < 0.01 and
   ****P < 0.0001.

The promotion of Ancr/E7-EXO on the differentiation of Gli1^+ cells into
contractile SMCs in high-glucose conditions

   On the basis of previous studies ([98]21), MSCs were isolated from rat
   bone marrow. The culture of MSCs revealed that they had osteogenic,
   adipogenic, and chondrogenic differentiation abilities (fig. S4A).
   Gli1^+ cell populations were positive for CD44, CD29, Sca1^+, and
   negative for CD34 and CD45 and were screened by flow cytometry from
   MSCs ([99]Fig. 4A and fig. S4B). We cultured the isolated Gli1^+ cells
   and carried out the follow-up experiments. Immunofluorescence staining
   showed that Gli1^+ protein was located in the cytoplasm and nucleus
   (fig. S4C). Further in vitro cell culture revealed that Gli1^+ cells
   were able to differentiate into three phenotypes of SMCs (fig. S4D).
   Besides, the number of Gli1^+ cells, which were extracted from the bone
   marrow of diabetic rats, increased in a time-dependent manner (fig. S4,
   E and F). The number of Gli1^+ cells in diabetic rats’ bone marrow at
   day 60 was 3.35 times higher than that in healthy rats. Wound-healing
   assay showed that Ancr/E7-EXO markedly increased Gli1^+ cell migration
   ([100]Fig. 4, B and C). The dose of 200 μg of Ancr/E7-EXO was best in
   promoting the differentiation of Gli1^+ cells into contractile SMCs and
   in inhibiting the synthetic SMCs induced by high glucose (fig. S5, A to
   C). It is generally believed that MYH11 and thrombospondin are both the
   differentiation markers of SMCs, while MYH11 represents the contractile
   phenotype of SMCs and thrombospondin represents the synthesis phenotype
   of SMCs ([101]22, [102]23). In high-glucose medium, both MSC-EXO and
   Ancr/E7-EXO increased the differentiation of Gli1^+ cells into
   contractile SMCs [Myosin heavy chain 11 (MYH11) and smooth muscle actin
   alpha (α-SMA)], but Ancr/E7-EXO showed a stronger ability obviously
   ([103]Fig. 4, D to F). The expression of MYH11 in the Ancr/E7-EXO group
   was 1.07 times higher than that in the MSC-EXO group ([104]Fig. 4E).
   The synthetic SMCs (thrombospondin) were observed in the high-glucose
   and MSC-EXO groups, while no thrombospondin was observed in the
   Ancr/E7-EXO group ([105]Fig. 4D). In addition, Ancr/E7-EXO notably
   reduced the expression of thrombospondin compared with the effect of
   MSC-EXO ([106]Fig. 4G). The qPCR results showed that the expression of
   MYH11 in the Ancr/E7-EXO group was 1.48 times higher than that in the
   high-glucose group, while the expression of thrombospondin decreased by
   29% ([107]Fig. 4, H and J). Furthermore, the expression of MYH11 in the
   Ancr/E7-EXO group was 1.32 times higher than that in the MSC-EXO group,
   while the expression of thrombospondin decreased by 22% ([108]Fig. 4, H
   and J). The expression trend of α-SMA was similar to that of MYH11
   ([109]Fig. 4I). Hence, Ancr/E7-EXO also had a remarkable effect on
   promoting Gli1^+ cell differentiation into contractile SMCs compared
   with the effect of MSC-EXO. Western blot was used to further clarify
   the effect of ANCR plasmid ([110]Fig. 4K). The transfection of ANCR
   plasmid into Gli1^+ cells did not affect the expression of Gli1^+,
   while α-SMA increased by 2.25 times and thrombospondin decreased by 45%
   ([111]Fig. 4, L to N). Therefore, the above results proved that
   Ancr/E7-EXO promoted the differentiation of Gli1^+ cells into
   contractile SMCs and inhibited the differentiation into synthetic SMCs
   in high-glucose conditions.

Fig. 4. The effect of Ancr/E7-EXO on SMC phenotype.

   [112]Fig. 4.
   [113]Open in a new tab

   (A) Gli1^+ cells were isolated from rat bone marrow MSCs by
   fluorescence-activated cell sorting. (B and C) Ancr/E7-EXO–induced
   Gli1^+ cell migration was evaluated using wound-healing assay. Scale
   bar, 500 μm. (D) Immunofluorescence staining was used to assess the
   effect of 200 μg of Ancr/E7-EXO on the differentiation of Gli1^+ cells
   in high-glucose medium at day 7: MYH11 (red) and α-SMA (green)
   represented contractile SMCs, and thrombospondin (red) represented
   synthetic SMCs. (E to G) Quantification of the immunostaining intensity
   in each group (n = 3). (H to J) Phenotypic changes of Gli1^+ cells
   under different conditions quantitatively analyzed by qPCR. (K to N)
   Western blot analysis was used to detect the differentiation and
   quantification of ANCR transfected Gli1^+ cells after 72 hours. Results
   are expressed as means ± SD by two-tailed Student’s t test. *P < 0.05,
   **P < 0.01, ***P < 0.001, and ****P < 0.0001. ns, not significant.

The rescue of Ancr/E7-EXO on the differentiation of Gli1^+ cells into
osteoblast-like cells in high-glucose conditions

   High-glucose environment will lead MSCs to differentiate into
   osteoblast-like cells and whether this process can be inhibited by
   Ancr/E7-EXO. Ancr/E7-EXO (200 μg) was added to Gli1^+ cells in
   high-glucose medium, the number of calcified nodules was reduced at day
   7, and the result was more obvious at day 14 (fig. S5, D to G). The
   potential ability of Ancr/E7-EXO to reverse the osteogenic
   differentiation of Gli1^+ cells was further evaluated by dividing
   Gli1^+ cells in high-glucose mediums into the control group, the
   osteogenic induction medium group (OI), the osteogenic induction medium
   + MSC-EXO group (OI + MSC-EXO), and the osteogenic induction medium +
   Ancr/E7-EXO group (OI + Ancr/E7-EXO). The Alizarin red staining
   suggested that Ancr/E7-EXO significantly reduced the mineralized
   nodules compared with their number in the control and OI groups at day
   7 (fig. S6A). Moreover, the number of calcium nodules in OI +
   Ancr/E7-EXO was 0.85 times that of OI + MSC-EXO, indicating that
   Ancr/E7-EXO inhibited the formation of calcium nodules more efficiently
   than MSC-EXO (fig. S6B). The fluorescence intensity of Runx2 and OPN in
   the OI + Ancr/E7-EXO group was significantly lower than that in the
   control group (fig. S6, C to E). In addition, qPCR also showed that the
   expression of bone morphogenetic protein-2 (BMP-2), Runx2, and OPN in
   the OI + Ancr/E7-EXO group was significantly lower than that in the
   control group (fig. S6, F to H). The osteogenic induction culture was
   carried out for 14 days to better prove the effect of Ancr/E7-EXO.
   Alizarin red staining showed that the number of calcified nodules in
   the control group was 37.11 times higher than that in OI + Ancr/E7-EXO
   group ([114]Fig. 5, A and B). The fluorescence intensity of OPN in the
   OI + Ancr/E7-EXO group was markedly lower than that in other groups,
   which was only 36% of that in the control group, and Runx2 was only 45%
   of that in the control group ([115]Fig. 5, C to E). In addition, the
   expression of osteoblast-like cell markers in each group demonstrated
   that the inhibitory effect of the OI + Ancr/E7-EXO group was markedly
   higher compared to that of the control group ([116]Fig. 5, F to H).
   qPCR showed that the expression of OPN in the OI + Ancr/E7-EXO group
   was only 26% of that in the high-glucose group, while the expression of
   Runx2 was 34%. Besides, the expression of OPN in the OI + MSC-EXO group
   was 1.33 times higher than that in the OI + Ancr/E7-EXO group, while
   Runx2 was 1.63 times higher. After 7 and 14 days of osteogenic
   induction, Ancr/E7-EXO exhibited excellent osteogenic inhibition in
   vitro.

Fig. 5. Preventive effect of Ancr/E7-EXO on the differentiation of Gli1^+
cells into osteoblast-like cells.

   [117]Fig. 5.
   [118]Open in a new tab

   (A and B) Osteogenic differentiation of Gli1^+ cells in each group
   detected by Alizarin red staining at day 14. The number of calcified
   nodules (red) was calculated and quantified by ImageJ. Scale bar, 250
   pixel (px). (C to E) Gli1^+ cells were labeled with anti-Runx2 (red)
   and anti-OPN (red) antibodies to distinguish the osteogenic
   differentiation of each group, and the fluorescence signals of each
   group were quantified (n = 5). (F to H) qPCR of the markers of the
   osteogenic differentiation of Gli1^+ cells at day 14 under different
   conditions. Results are expressed as means ± SD by two-tailed Student’s
   t test. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Phagocytosis of Ancr/E7-EXO by Gli1^+ cells in vivo and enrichment of
Ancr/E7-EXO to calcified sites in CKD

   Ancr/E7-EXO–modified TEBVs were used to evaluate the inhibitory effect
   of Ancr/E7-EXO on vascular calcification in vivo.
   1-Ethyl-3-(3-dimethylaminopropyl)carbodiimide
   (EDC)/N-hydroxysuccinimide (NHS) can covalently cross-link the exosomes
   on TEBVs by amide reaction, thus improving the fixation strength.
   Polyethyleneimine/Au Nanopraticle (PEI/AuNP) was used as a known method
   to modify exosomes on TEBVs ([119]24). By comparison, we hope to modify
   more exosomes on the media and adventitia of TEBVs (fig. S7A).
   According to the results, EDC/NHS self-assembly strategy was chosen to
   cross-link as much Ancr/E7-EXO as possible on TEBVs to construct the
   Ancr/E7-EXO–modified TEBVs ([120]Fig. 6A). Ancr/E7-EXO was labeled by
   PKH67, a large number of Ancr/E7-EXO cross-linking on the media and
   adventitia of TEBVs compared with that in the acellular vessels
   (without considering the autofluorescence of vessels) ([121]Fig. 6B).
   This result suggested that this simple self-assembly strategy
   successfully modified Ancr/E7-EXO on TEBVs. Then, the Ancr/E7-EXO on
   TEBVs was labeled with DiD and then implanted into rats. The
   colocalization of Gli1^+ cells and Ancr/E7-EXO in the adventitia of
   TEBVs was observed from weeks 1 to 3 ([122]Fig. 6C). It is suggested
   that, at early stage of implantation, Ancr/E7-EXO can be swallowed by
   Gli1^+ cells to play a role. In vivo fluorescence is used to observe
   the metabolism of Ancr/E7-EXO after implantation for a longer time, the
   TEBVs modified by Ancr/E7-EXO labeled with
   1,1-dioctadecyl-3,3,3,3-tetramethylindotricarbocyaine iodide (DIR)
   fluorescence probe were implanted into the carotid artery of rats
   ([123]Fig. 6D). The signals of Ancr/E7-EXO retention were observed in
   the neck of the rats from weeks 1 to 6 after implantation ([124]Fig. 6,
   D and E), indicating that the Ancr/E7-EXO still remained in the TEBVs
   for 6 weeks in vivo, thus providing sufficient time for the regulation
   of Gli1^+ cell differentiation.

Fig. 6. Modification of TEBVs with Ancr/E7-EXO and enrichment of Ancr/E7-EXO
in autologous calcified blood vessels.

   [125]Fig. 6.
   [126]Open in a new tab

   (A) Scheme of Ancr/E7-EXO loaded into the media and adventitia of TEBVs
   by EDC-NHS, and implantation of modified TEBVs. (B) Ancr/E7-EXO was
   labeled with PKH67, and representative images of acellular vessel and
   Ancr/E7-EXO–modified TEBVs by EDC/NHS are shown. (C) Ancr/E7-EXO was
   labeled with DiD, and immunofluorescence staining in the
   cross-sectional TEBVs in each group from 1 to 3 weeks after
   implantation (DAPI, blue; Gli1^+ cells, green; Ancr/E7-EXO, red). Scale
   bars, 100 and 500 μm. (D and E) Bioluminescence images and
   quantification of Ancr/E7-EXO–modified TEBVs obtained at different
   times after TEBV transplantation (n = 4). (F) Scheme of CKD rats
   treated with Ancr/E7-EXO. (G and H) Ancr/M7-EXO and Ancr/E7-EXO were
   injected into the tail vein of CKD rats at week 13 from CKD
   development, and the bioluminescence images and quantification of
   aortic and thoracic aorta were compared (n = 3). Results are expressed
   as means ± SD by two-tailed Student’s t test. *P < 0.05.

   The ANCR/M7-EXO and Ancr/E7-EXO labeled with DIR were injected at week
   13 after subtotal nephrectomy ([127]Fig. 6F). The fluorescence signal
   from the beginning of the aorta to the thoracic aorta in the
   Ancr/E7-EXO group was 1.58 times higher than that in the ANCR/M7-EXO
   group ([128]Fig. 6, G and H), indicating that Ancr/E7-EXO enriched to
   the calcified site in vivo and was phagocytosed by Gli1^+ cells because
   of its special affinity with Gli1^+ cells.

The enhancement of Ancr/E7-EXO on long-term patency and normal mechanical
properties of TEBVs in vivo

   We implanted TEBVs modified by Ancr/E7-EXO into Sprague-Dawley rats.
   Seven days after implantation, diabetic rats were established by
   injection of streptozotocin, and their blood glucose was continuously
   monitored for 60 days ([129]Fig. 7A and fig. S8G). According to the
   type of TEBVs and the presence or absence of diabetes, the rats were
   divided into four groups: group 1: normal rats with TEBVs (N-TEBVs);
   group 2: diabetic rats with TEBVs (D-TEBVs); group 3: normal rats with
   Ancr/E7-EXO–modified TEBVs (N-TEBVs-Ancr/E7-EXO); and group 4: diabetic
   rats with Ancr/E7-EXO–modified TEBVs (D-TEBVs-Ancr/E7-EXO). Groups 1
   and 2 showed more severe intimal hyperplasia and lower blood flow
   velocity at 42 days after implantation (fig. S8, A, C, and E), while
   groups 3 and 4 showed a slighter intimal hyperplasia and collagen
   deposition (fig. S8, A and B). The immunofluorescence staining revealed
   that the expression of α-SMA in the media of groups 3 and 4 was
   significantly higher than that in the control group (fig. S8, D and F).
   The number of contractile SMCs in diabetic rats of group 4 was 17.13
   times higher than that in group 2. These results revealed that
   Ancr/E7-EXO promoted the rapid remodeling of SMCs in TEBVs. Hematoxylin
   and eosin (H&E) staining performed at day 60 after implantation showed
   that vascular stenosis and occlusion occurred after TEBV implantation,
   while Ancr/E7-EXO–modified TEBVs maintained excellent cell regeneration
   and vascular patency in both normal and high-glucose environment
   ([130]Fig. 7B). Masson staining showed that Ancr/E7-EXO–modified TEBVs
   had less collagen fiber deposition in both normal and high-glucose
   environment ([131]Fig. 7C), indicating that Ancr/E7-EXO inhibited
   collagen fiber secretion. The intima-media thickness ratio revealed
   that Ancr/E7-EXO significantly reduced intimal hyperplasia of TEBVs in
   groups 3 and 4 ([132]Fig. 7E). The results of Doppler ultrasound showed
   that the Ancr/E7-EXO group maintained a more stable hemodynamics in
   either normal physiological or high-glucose condition ([133]Fig. 7D).
   The blood flow velocity in group 3 was stable at 81.20 ± 21.51 cm/s,
   while that in group 4 was stable at 70.27 ± 15.73 cm/s, which were much
   higher than those in groups 1 and 2 ([134]Fig. 7F). Micro–computed
   tomography (micro-CT) also showed that the TEBVs in groups 3 and 4
   remained unobstructed and in excellent shape in normal and high-glucose
   conditions, while the control group had no vascular signal
   ([135]Fig. 7G). The above results suggested that Ancr/E7-EXO might
   represent a promising solution for an early and rapid reconstruction of
   VSMCs as well as long-term patency of TEBVs in both normal and
   high-glucose conditions.

Fig. 7. Examination of long-term patency, normal morphology, and mechanical
properties of TEBVs.

   [136]Fig. 7.
   [137]Open in a new tab

   (A) Diabetic rat model established by the injection of streptozotocin
   at day 7 after TEBV implantation. The change in blood glucose in rats
   was measured from days 11 to 14 and then measured once a week. SMC
   remodeling was detected at day 60. (B) Intimal hyperplasia and SMC
   remodeling were observed by H&E staining. (C) Collagen secretion by
   Masson trichrome staining. (D and F) Patency was determined with a
   portable ultrasound scanner (n = 5). (E) Intima-media thickness ratio
   was calculated to evaluate intimal hyperplasia (n = 5). (G) Overall
   patency of TEBVs in each group by micro–computed tomography (micro-CT).
   (H and I) Young’s Modulus of TEBV media in all groups measured by
   atomic force microscopy (AFM) (n = 5). (J and K) Images of the height
   and peak force error showing the arrangement of vascular fibers in each
   group (n = 5). Results are expressed as means ± SD by two-tailed
   Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P <
   0.0001.

   Furthermore, the mechanical properties of the media of TEBVs were
   examined by atomic force microscopy (AFM) ([138]Fig. 7H). At day 60,
   the Young’s modulus of TEBVs and native blood vessels indicated an
   increase in the stiffness in groups 1 and 2, with a value of 3325 and
   3729 kPa, respectively, compared with the values of the native blood
   vessels, while the stiffness in groups 3 and 4 was similar to that of
   the native vessels ([139]Fig. 7I). This result indicated that
   Ancr/E7-EXO could maintain the normal vascular stiffness of TEBVs. The
   above results suggested that, on the one hand, Ancr/E7-EXO inhibited
   the differentiation of Gli1^+ cells into synthesized SMCs and reduced
   collagen deposition; on the other hand, Ancr/E7-EXO inhibited the
   differentiation of Gli1^+ cells into osteoblast-like cells, thereby
   reducing the calcification and maintaining a normal vascular stiffness.
   The morphological characteristics ([140]Fig. 7J) and height
   characteristics ([141]Fig. 7K) showed that groups 3 and 4 had a regular
   fiber arrangement similar to that of the native vessels, while the
   matrix fiber arrangement of the control group was disordered. These
   results provided direct evidence that Ancr/E7-EXO maintained the normal
   vascular morphology and stiffness.

The dual effects of Ancr/E7-EXO on promoting SMC remodeling and inhibiting
calcification of TEBVs in vivo

   To evaluate the cell reconstruction after TEBV implantation, we
   quantified the cell composition of TEBVs in each group at day 42. In
   groups 1 and 2, we observed that osteoblast-like cells (OPN)
   contributed 92.5% and 86.3% of Gli1^+ cells, while only 13.6 and 10.7%
   of cells in groups 3 and 4 expressed OPN (fig. S9, A and C). These
   results indicated that Ancr/E7-EXO significantly inhibited Gli1^+ cells
   differentiated into osteoblasts. Similarly, the composition of SMCs was
   quantified. The proportion of contractile SMCs in group 3 was 63.5%,
   which was higher than that in group 1 (fig. S9B). However, there was no
   difference among all groups for the proportion of α-SMA (fig. S9D). The
   reason is that a large number of SMCs in groups 1 and 2 abnormally
   proliferate and migrate to the intima, leading to serious intima
   hyperplasia. Thus, the expression of contractile SMCs is abnormally
   increased. Furthermore, in the coimmunofluorescence staining performed
   at 60 days after implantation, it showed that few Gli1^+ cells and
   α-SMA were both located in groups 3 and 4, and the contractile SMCs in
   the media of groups 3 and 4 increased significantly compared with that
   in groups 1 and 2 ([142]Fig. 8A). The number of α-SMA positive cells in
   the media of groups 3 and 4 was 53.92 and 8.38 times higher than that
   of groups 1 and 2, respectively ([143]Fig. 8B); meanwhile, the number
   of thrombospondin positive cells decreased by 8.14 and 24.92% compared
   with groups 1 and 2, respectively ([144]Fig. 8C). OPN and Runx2 in
   TEBVs were stained to evaluate the inhibitory effect of Ancr/E7-EXO on
   vascular calcification. Double immunofluorescence staining revealed
   that more Gli1^+ cells and OPN were colocalized in groups 1 and 2,
   while groups 3 and 4 had almost no Gli1^+ cells and OPN signal
   ([145]Fig. 8, A and E). The Runx2 fluorescence intensity in groups 1
   and 2 were 12.86 and 4.22 times higher than that of groups 3 and 4,
   respectively ([146]Fig. 8D). It indicated that Ancr/E7-EXO could
   significantly reduce the calcification of TEBVs, and this result was
   consistent with that of the vascular calcification found by Alizarin
   red and von Kossa staining ([147]Fig. 8, F and G). The above results
   suggested that Ancr/E7-EXO promoted vascular smooth muscle remodeling
   and inhibited the calcification after TEBV implantation, providing a
   promising therapeutic approach for the reconstruction of smooth muscle
   in TEBVs and the resistance to vascular calcification aggravated by
   basic diseases such as diabetes.

Fig. 8. The regulation of Ancr/E7-EXO on SMC remodeling and calcification in
TEBVs.

   [148]Fig. 8.
   [149]Open in a new tab

   (A) Immunofluorescence staining in the cross-sectional TEBVs in each
   group at 60 days after implantation (DAPI, blue; Gli1^+ cells, red;
   α-SMA, green; thrombospondin, green; Runx2, red; OPN, green). Scale
   bars, 20 μm. (B to E) Quantification of α-SMA, thrombospondin, Runx2,
   and OPN in the media of cross-sectional TEBVs (n = 3). (F and G) TEBV
   calcification in each group by Alizarin red and von Kossa staining.
   Results are expressed as means ± SD by two-tailed Student’s t test. *P
   < 0.05, **P < 0.01, and ***P < 0.001.

The therapeutic effect of Ancr/E7-EXO on calcification of autologous vessels

   Ancr/E7-EXO also played the same role in vascular calcification caused
   by CKD. Increased levels of BUN were detected from weeks 1 to 12
   ([150]Fig. 9A). According to FRI of near-infrared spectroscopy, calcium
   tracer showed that the positive signal of hydroxyapatite in the
   treatment group (5/6Nx 13 weeks + Ancr/E7-EXO) was 60.42% of that in
   the control group (5/6Nx 13 weeks), suggesting the degree of vascular
   calcification decreased in CKD rats after Ancr/E7-EXO treatment
   ([151]Fig. 9, B and C). The double immunofluorescence staining showed
   the presence of more Gli1^+ cells and contractile SMCs colocated in the
   media of the treatment group ([152]Fig. 9E). The number of Gli1^+ cells
   expressing α-SMA in the media was 110.4, which was 2.94 times higher
   than that in the control group ([153]Fig. 9F). Many Gli1^+ cells and
   osteoblast-like cells were colocated in the media of the control group,
   and the number of Gli1^+ cells expressing OPN in the media was 49.94,
   which was 6.39 times higher than that in the treatment group
   ([154]Fig. 9, E and G). The above results proved that Ancr/E7-EXO
   inhibited the osteogenic differentiation of Gli1^+ cells in autologous
   vessels and promoted the differentiation of Gli1^+ cells into
   contractile SMCs.

Fig. 9. Protection of Ancr/E7-EXO against autologous vascular calcification
in CKD.

   [155]Fig. 9.
   [156]Open in a new tab

   (A) BUN stratified for the sham and treatment group. (B and C)
   Representative images and quantification of hydroxyapatite fluorescence
   intensity in the sham, control, and treatment groups (n = 3). (D)
   Volcano plot between the vascular samples of the treatment and control
   groups. Red dots, genes significantly down-regulated in treatment
   group; yellow dots, genes significantly up-regulated in treatment
   group; blue dots, genes not differentially expressed. (E)
   Representative images of cross-sectional aorta in the sham, control,
   and treatment groups by immunofluorescence staining (DAPI, blue; Gli1^+
   cells, red; α-SMA, green; OPN, green) (n = 4). (F and G) Quantification
   of Gli1^+ cells expressing α-SMA and OPN in the media of
   cross-sectional aorta (n = 4). (H) Heatmap of calcification-related
   gene expression. (I) Bubble map of the top 10 biological process terms
   in the Gene Ontology (GO) enrichment analysis. (J) Top 30 of Kyoto
   Encyclopedia of Genes and Genome (KEGG) pathways in the enrichment
   analysis. cAMP, cyclic adenosine 3′,5′-monophosphate; CoA, coenzyme A;
   IL-17, interleukin-17; MAPK, mitogen-activated protein kinase; TGF-β,
   transforming growth factor–β. (K) Gene set enrichment analysis of
   muscle contraction. Results are expressed as means ± SD by two-tailed
   Student’s t test. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P <
   0.0001. wk, weeks.

   Furthermore, bioinformatic technique was used to analyze the
   differentially expressed genes in aorta and thoracic aorta induced by
   CKD between the Ancr/E7-EXO and control groups. We observed that the
   expression of Ptgs2 decreased, while the expression of Myocd, Nos3,
   Cd163, and Hspb7 increased ([157]Fig. 9D). Ptgs2, also known as COX-2,
   could induce the increased mRNA expression of osteogenic genes OPN and
   Runx2 and lead to calcification of aortic valve and aorta ([158]25,
   [159]26). Ptgs2 also played a pivotal role in the process of
   atherosclerosis, and its expression could be used as a biomarker of the
   degree of atherosclerosis ([160]27). In this study, the expression of
   Ptgs2 decreased after Ancr/E7-EXO treatment, indicating that
   Ancr/E7-EXO could inhibit vascular calcification–related genes. Myocd
   can regulate the expression of genes encoding SMC contractile proteins
   to maintain the contractile phenotype of SMCs ([161]28). In this study,
   Myocd was up-regulated, indicating that Ancr/E7-EXO could also maintain
   the contractile phenotype of SMCs. In addition, up-regulated expression
   of Nos3, Hspb7, and Cd163 indicated that Ancr/E7-EXO can play an active
   role in the maintenance and regulation of vascular immune
   microenvironment ([162]29–[163]31). At the same time, by comparing the
   expression changes of vascular calcification–related genes, the
   expression of calcification-related genes decreased significantly in
   the Ancr/E7-EXO group ([164]Fig. 9H). These results suggested that
   Ancr/E7-EXO can up-regulate the expression of contraction-related genes
   and down-regulate the expression of vascular calcification–related
   genes in SMCs. Moreover, Gene Ontology (GO) analyzed the most abundant
   biological processes, including myocardial contraction (P < 0.00001)
   ([165]Fig. 9I). Here, our sample tissues were aorta and thoracic aorta,
   and aortic smooth muscle and myocardium may have the same developmental
   origin, so we thought that enriched myocardial contraction is the SMC
   contraction. At the same time, by Kyoto Encyclopedia of Genes and
   Genomes (KEGG) pathway enrichment analysis, myocardial contractile
   signal pathway was detected in top 30 of the enrichment pathways (P <
   0.02) ([166]Fig. 9J). In addition, after the gene set enrichment
   analysis of the overall expression forms of muscle contraction pathway,
   the related gene set was up-regulated in the Ancr/E7-EXO group
   ([167]Fig. 9K). These results suggested that Ancr/E7-EXO can promote
   vascular SMCs differentiation into contractile SMCs, thus maintaining
   vascular contractile function.

DISCUSSION

   Vascular calcification is a chronic vascular disease, which is not only
   a common cause of vascular graft failure but also a common complication
   of diabetes, CKD, and other diseases. Previous studies on small
   diameter vascular grafts focused on the evaluation of the efficacy of
   TEBVs on neointimal hyperplasia and thrombosis in the early graft (≤3
   months), which are early complications of grafts that are prone to
   vascular stenosis or blockage ([168]32). Calcification is an important
   risk factor for late graft failure in patients who underwent
   bioengineered cardiovascular grafts ([169]6) suffering from underlying
   diseases such as diabetes ([170]33), which is a problem rarely
   evaluated and managed. After TEBV implantation, on the one hand, the
   reconstruction of SMCs in vivo is difficult; on the other hand, TEBVs
   are more likely to calcify in combination with injury and pathological
   environment. Therefore, how to promote the normal reconstruction of the
   smooth muscle layer of TEBVs and inhibit calcification is still a
   technical challenge. Moreover, clinical study found that medial
   arterial calcification is detected in 17 to 42% of patients with type 2
   diabetes mellitus, in 27 to 40% of patients with advanced CKD, and in
   up to 72% of patients with threatening chronic limb ischemia ([171]2).
   Severe vascular calcification can increase arterial hardness, can lead
   to loss of vascular function and even vascular rupture, and can lead to
   subsequent incidence and mortality of cardiovascular disease ([172]34).
   The incidence and degree of vascular calcification in patients with CKD
   are higher than those in the general population, and the incidence of
   vascular calcification increases with the decline of the renal function
   ([173]35). There is no noninvasive and effective solution for vascular
   calcification caused by TEBVs and CKD ([174]33).

   Endogenous vascular progenitor cells are an important source of cells
   for repair and regeneration after vascular injury, and their abnormal
   differentiation will also lead to abnormal repair after injury
   ([175]36). Gli1 is considered to be a specific marker of MSCs in
   vascular adventitia and has the differentiation ability of SMCs and
   osteoblast-like cells in the process of injury and repair ([176]14). As
   endogenous vascular progenitor cells, the fate and contribution of
   Gli1^+ cells after TEBV implantation are still unclear. Therefore, we
   evaluated the spatiotemporal distribution of Gli1^+ cells after TEBV
   implantation. The results showed that Gli1^+ cells continued to
   increase in both media and adventitia after TEBV implantation in normal
   and diabetic rats. Gli1^+ cells were significantly present at day 3
   after TEBV implanted in diabetic rats. The Alizarin red and von Kossa
   staining of TEBVs revealed the formation of calcification foci as early
   as 2 weeks after implantation. These results indicated that the
   spatiotemporal distribution of Gli1^+ cells in TEBVs was positively
   correlated with the degree of calcification under normal physiological
   and pathological conditions. Through Gli1^+ costaining with OPN, more
   Gli1^+ cells had osteogenic phenotype in both normal and diabetic
   groups, indicating that Gli1^+ cells are also the key source of TEBV
   calcification, which is the first time to find that calcification of
   TEBVs is driven by Gli1^+ cells. Besides, we also localized Gli1^+
   cells during autologous vascular calcification. After 13 weeks of
   induction of CKD in rats, we observed that Gli1^+ cells were mainly
   distributed in the media of blood vessels. Immunofluorescence staining
   showed that the presence of Gli1^+ cells was eight times higher than
   that of the control group, which was also consistent with the degree of
   aortic calcification shown by calcium tracer. Therefore, we also
   verified that the calcification of autologous blood vessels is caused
   by the differentiation of Gli1^+ cells into osteoblasts. Then, how to
   deal with TEBV and autologous vascular calcification caused by the
   differentiation of Gli1^+ into osteoblasts is the key target for us to
   solve the problem of vascular calcification.

   Effectively targeting and regulating the differentiation of Gli1^+
   cells is the key goal of this study. Previous studies focused on the
   regulation of graft environment, such as modifying antioxidant
   substances ([177]32) or nitric oxide (NO)–releasing molecules on TEBVs
   ([178]9), achieving anticalcification through antioxidant effect, or
   using the immunomodulatory effect of MSC-EXO to prevent calcification
   ([179]37), but these studies did not eliminate the symptoms of vascular
   calcification from the source. Exosome is the key bridge of
   intercellular communication, which can be used for signal transmission
   such as RNA. Therefore, we designed Ancr/E7-EXO–engineered exosomes
   that can target Gli1^+ cells. The E7 peptide attached to the membrane
   surface has a special affinity for Gli1^+ cells, while the lncRNA-ANCR
   loaded inside can inhibit the differentiation of osteoblasts. We
   observed that the exosomes modified with E7 peptide could effectively
   target Gli1^+ cells. In the coculture of macrophages, neutrophils, and
   Gli1^+ cells, E7-EXO still had a high affinity for Gli1^+ cells.
   Therefore, Ancr/E7-EXO can avoid early phagocytosis of macrophages and
   provide time for Ancr/E7-EXO to reprogram Gli1^+ cells in vivo. After
   TEBV implantation, the rapid reconstruction of smooth muscle layer and
   the inhibition of calcification are the main issues of this study,
   which are also an important part of the functional remodeling of TEBVs
   in vivo and play an important role in maintaining vascular contraction
   and mechanical strength. However, the difficulties in SMC
   reconstruction and abnormal proliferation after reconstruction often
   lead to the occurrence of aneurysms and intimal hyperplasia. By
   modifying Ancr/E7-EXO on TEBVs, we verified its function in rat carotid
   artery replacement model. We observed that SMC regeneration appeared at
   the sixth week after operation, and there were abundant and orderly
   vascular smooth muscle layers in the media of TEBVs at the eighth week.
   At the same time, TEBVs modified with Ancr/E7-EXO can avoid intimal
   hyperplasia after implantation, in either normal or diabetic
   environment. The mechanism may be that Ancr/E7-EXO promotes the rapid
   and normal reconstruction of the smooth muscle layer, reduces the
   abnormal proliferation and migration of SMC, and thus inhibits the
   intimal hyperplasia of TEBVs. Therefore, Ancr/E7-EXO can promote the
   rapid and normal reconstruction of TEBV smooth muscle layer, avoid
   intimal hyperplasia, and open the way for the rapid host of TEBVs. We
   also evaluated the late calcification of the grafts. Calcification
   often leads to increased arterial stiffness and decreased vascular
   compliance ([180]38). In the control group, we found that the graft
   stiffness increased significantly at the eighth week. The stiffness of
   TEBVs modified by Ancr/E7-EXO is close to that of normal blood vessels.
   These results suggest that Gli1^+ cells can proliferate and
   differentiate into osteoblast-like cells during the repair of TEBV
   injury, which leads to the increase of graft stiffness, while
   Ancr/E7-EXO can inhibit the differentiation of Gli1^+ cells into
   osteoblast-like cells, thus making TEBV stiffness close to normal blood
   vessels. In addition, a large number of osteogenic signals were found
   in the TEBVs of the control group by immunofluorescence staining and
   flow cytometry. However, the osteoblast-like signal of Ancr/E7-EXO
   group was significantly suppressed, indicating that Ancr/E7-EXO could
   significantly inhibit the calcification of TEBVs. This may be due to
   the inhibitory effect of lncRNA-ANCR on osteoblast-like differentiation
   of Gli1^+ cells. Therefore, the Ancr/E7-EXO that we constructed can
   target and regulate Gli1^+ cells in vivo and overcome the problem of
   TEBV calcification from the source. It not only can promote the rapid
   reconstruction of SMCs but also can effectively inhibit the early
   intimal hyperplasia and long-term calcification of the graft and
   promote the normal reconstruction of the smooth muscle layer of TEBVs.

   Because the main cause of autologous vascular calcification is also the
   differentiation of Gli1^+ cells into osteoblast-like cells during
   chronic injury, the Ancr/E7-EXO that we designed has a special affinity
   for Gli1^+ cells. FRI showed that the exosome signal of aortic
   calcification site was higher than that of the control group,
   indicating that Ancr/E7-EXO was enriched in the calcification site. In
   addition, the autologous vessels treated with Ancr/E7-EXO had more
   phenotypes of SMCs and Gli1^+ cells, indicating that Ancr/E7-EXO
   promoted Gli1^+ cell differentiation into SMCs. In the control group,
   more osteogenic phenotypes were colocated with Gli1^+ cells, and the
   expression of OPN in autologous vessels treated with Ancr/E7-EXO
   decreased, indicating that Ancr/E7-EXO significantly inhibited the
   differentiation of Gli1^+ cells into osteogenic phenotypes. Calcium
   tracer imaging further showed that Ancr/E7-EXO inhibited autologous
   vascular calcification under pathological condition. By bulk RNA
   sequencing (RNA-seq), we observed a significant difference in
   calcification gene expression between the control group and the
   Ancr/E7-EXO treatment group. Ptgs2 could induce the increased
   expression of osteogenic genes OPN and Runx2 ([181]25, [182]26), in
   which the expression decreased after Ancr/E7-EXO treatment. In
   addition, the expression of genes encoding SMC contractile protein also
   increased after Ancr/E7-EXO treatment. The increased expression of
   genes related to the regulation and maintenance of vascular immune
   microenvironment may be the role of exosomes derived from MSCs
   ([183]37). These results further proved that Ancr/E7-EXO reprogrammed
   Gli1^+ cells in vivo, thus achieving the purpose of noninvasive
   treatment of autologous vascular calcification caused by chronic
   diseases. Because autologous vascular calcification has the same
   pathological mechanism in a variety of pathological conditions, the
   Ancr/E7-EXO designed in this study can also be used for autologous
   vascular calcification caused by other chronic diseases. For example,
   in vascular calcification caused by atherosclerosis and chronic
   diabetes, Ancr/E7-EXO also has an anticalcification effect.

   In conclusion, the Ancr/E7-EXO designed in this study can effectively
   capture and reprogram Gli1^+ cells by modifying TEBVs, thus
   accelerating the normal reconstruction of the smooth muscle layer of
   TEBVs and significantly inhibiting the calcification. Ancr/E7-EXO can
   also be infused intravenously, which can enrich in the calcified site
   and significantly inhibit autologous vascular calcification caused by
   CKD. This method provides a promising treatment strategy for vascular
   calcification and other cardiovascular graft calcification under
   complex pathological conditions, having a translational value in
   clinical application.

MATERIALS AND METHODS

Cell culture

   The HuBM-MSCs were purchased from Biospes and cultured in α–modified
   essential medium (α-MEM; HyClone), supplemented with 10%
   exosome-depleted fetal bovine serum (FBS) [Biological Industries (BI)]
   and 2% penicillin-streptomycin (Beyotime Biotechnology), and incubated
   in a humidified environment at 37°C under 5% CO[2]. The plasmid
   expressing E7-Lamp2b fusion protein was transfected into HuBM-MSCs
   using the Entranster-H4000 transfection reagent (Engreen). Gli1^+ cells
   were mechanically harvested from the femurs of Sprague-Dawley rats as
   previously described ([184]39). In brief, 100-g Sprague-Dawley rats
   were selected and euthanized through cervical dislocation. Bilateral
   femur and tibia were removed under sterile conditions and immersed in a
   phosphate-buffered saline (PBS) solution; the middle femur and tibia
   were cut into debris to release the bone marrow from the cavity. The
   cell culture medium containing 2% dual antibody α-MEM was aspirated
   with a 2-ml syringe, and bone marrow was repeatedly rinsed. The
   collected cell suspension was filtered through a 75-μm filter and
   centrifuged at 1000 rpm/min for 5 min. Then, the red blood cells were
   removed from collected cell suspension using red blood cell lysis
   buffer (Solarbio), the cell suspension was centrifuged at 1000 rpm/min
   for 5 min, and the cells were seeded in a flask containing α-MEM cell
   culture medium supplemented with 10% FBS (AusGeneX). The
   differentiation of purified Gli1^+ cells was obtained by culturing them
   until 60 to 70% or 90 to 100% confluence in 24-well plate, and α-MEM
   medium was replaced by osteogenic, chondrogenic, and adipogenic
   differentiation media (Cyagen). Alkaline phosphatase (Cyagen) staining
   and Oil red O (Sigma-Aldrich) staining were performed after 14 days of
   culture. Alcian blue (Cyagen) staining was performed after 21 days of
   chondrogenic induction. The smooth muscle cell differentiation was
   obtained by culturing Gli1^+ cells in α-MEM, containing 10% FBS, 2%
   penicillin-streptomycin, transforming growth factor–β (10 ng/ml;
   PeproTech), and platelet-derived growth factor B (5 ng/ml; PeproTech).
   The osteogenic differentiation was obtained by culturing Gli1^+ cells
   in osteogenic differentiation medium (Procell) supplemented with 30 mM
   glucose solution, and the culture medium was changed every 2 days; the
   cells at three to four passages were used for in vitro experiments.
   Cells cultured for 7 and 14 days were subjected to Alizarin red
   staining (Cyagen).

Flow cytometry

   Rat Bone Marrow (RBM)-MSCs were subjected to an in vitro culture for
   three passages and selected to characterize the surface markers. A
   total of 10^6 cells were suspended in a volume of 400 μl of sterile PBS
   supplemented with 1% FBS. The cells were labeled with an antibody
   cocktails for 40 min on ice: Alexa Fluor 488–conjugated rat anti-mouse
   CD34 (1:100, BioLegend), allophycocyanin-CD45 (1:100, BioLegend),
   phycoerythrin (PE)–CD44 (1:100, eBioscience), PE-Cy7-CD29 (1:100,
   BioLegend), and fluorescein isothiocyanate (FITC)–Sca1^+ (1:100,
   Biorbyt). Cells were washed twice with cold PBS after incubation and
   then separated on a FACS Calibur flow cytometer (BD Biosciences). The
   data were analyzed using FlowJo V10 software (Tree Star Inc., Ashland,
   OR, USA).

   Flow sorting was performed when RBM-MSCs were subcultured to P3
   passage. The cells were resuspended at a density of 10^6 in 2 ml of PBS
   containing 10 μl of PE-CD44 (1:100, BD Bioscience), PE-Cy7-CD29 (1:100,
   BioLegend), FITC-Sca1+ (1:100, Biorbyt), 10% FBS, and 2% dual antibody
   and incubated at 4°C for 30 min. The cells were washed twice with PBS
   and then sorted.

   MSC-EXO was analyzed on a Flow NanoAnalyzer (N30E, China) instrument.
   MSC-EXO pellets were resuspended in 30 μl of PBS. To a 30-μl MSC-EXO
   sample with a particle concentration of approximately 1.67×10^5
   particles/ml, 20 μl of FITC-conjugated antibody against CD9, CD63, and
   CD81 was separately added. The mixture was incubated at 37°C for 30 min
   in the dark. MSC-EXO was washed twice with 1 ml of cold PBS by
   ultracentrifugation at 110,000g for 70 min at 4°C. The pellets were
   resuspended in 50 μl of cold PBS for NanoFCM analysis.

TEBV dissociation and flow cytometry

   A single-cell suspension of TEBVs was prepared following the enzymatic
   digestion protocol described previously ([185]40). Briefly, the
   isolated TEBVs were finely cut and incubated in 1 ml of collagenase
   type II (20 mg/ml) for 1 hour at 37°C. The cell suspension was strained
   through a 40-μm filter and washed twice with PBS. The cells were
   resuspended at a density of 10^6 in 2 ml of PBS containing 10 μl of
   Alexa Fluor 488–conjugated rat anti-mouse GLI1 (1:50, Santa Cruz
   Biotechnology), PE–α-SMA (1:100, Novus Biologicals), Alexa Fluor
   647–conjugated rat anti-mouse OPN (1:50, Santa Cruz Biotechnology), 10%
   FBS, and 2% dual antibody and incubated at 4°C for 30 min. The cells
   were washed twice with PBS and then analyzed.

Exosome isolation

   The supernatant of E7 plasmid–transfected HuBM-MSCs subjected to 48
   hours of culture was collected and centrifuged at 3000g for 10 min at
   4°C. The supernatant was collected and filtered using a 0.22-μm filter
   (Biosharp) to remove the remaining cellular debris. Then, the
   supernatant was ultracentrifuged at 110,000g for 120 min at 4°C using
   an ultrahigh-speed centrifuge (CP100NX, HITACHI, Japan). The exosome
   pellets were resuspended in PBS, and the amount of exosomal protein was
   measured by the BCA Protein Assay kit (Beyotime). The exosomes were
   stored at 80°C.

Loading of E7-EXO with plasmid

   A total of 25 μg of purified exosomes (BCA Protein Assay Kit, Beyotime)
   and 25 μg of ANCR plasmid were gently mixed in 100 μl of
   electroporation buffer to load E7-EXOs with ANCR plasmid.
   Electroporation was carried on an electroporation cuvette using a Gene
   Pulser II Electroporator (Bio-Rad, USA) at 100 V and 50 μF. After
   electroporation, the mixture was immediately placed at 37°C for 30 min
   to completely recover the membrane of the exosomes. The intact
   Ancr/E7-EXO membrane was confirmed by TEM, and the Ancr/E7-EXOs were
   washed twice with cold PBS and centrifuged at 100,000g for 70 min to
   remove the ANCR plasmid that was not incorporated into the exosomes.
   The efficiency of electroporation of the ANCR plasmid was detected by
   qPCR.

Western blotting

   Total proteins were extracted from HuBM-MSCs or exosomes using
   radioimmunoprecipitation assay lysis buffer, and the concentration was
   determined by bicinchoninic acid (BCA) assay (Beyotime). Proteins were
   separated by electrophoresis and transferred to a polyvinylidene
   difluoride membrane. The membrane was incubated with the following
   primary antibodies: rat anti–Lamp-2b (1:400, Santa Cruz Biotechnology);
   rabbit anti-CD63 antibody (1:500, Bioss); rabbit anti–throbospondin 2
   antibody (1:500, Bioss); mouse anti–α-SMA antibody (1:500, Bioss); and
   rabbit anti-GLI1 antibody (1:500, Bioss) at 4°C overnight, followed by
   2 hours of incubation with the corresponding secondary antibody at room
   temperature. The blots were visualized by enhanced chemiluminescence
   (GelDoc 2000 Gel imaging system, Bio-Rad).

Real-time qPCR and qPCR

   Total RNA was extracted from HuBM-MSCs using an miRNeasy Mini kit
   (217004, Qiagen) according to the manufacturer’s instructions.
   Complementary DNA (cDNA) was synthesized using the RevertAid First
   Strand cDNA Synthesis (K1622, Thermo Fisher Scientific). cDNA was
   amplified in a real-time qPCR thermocycler (CFX96, BIO-RAD) using the
   SYBR Green qPCR Master (4943914001-SR, Roche). Cel-miR-39 was used as
   control, and the relative gene expression was calculated using the
   2^−ΔΔCT method. The primers used for reverse transcriptase (RT)–PCR
   were as follows: rat DANCR, (forward) 5′-ACATTTAATGACCCCGCTTG-3′ and
   (reverse) 5′-GATTCCAGGACACCCAGAAA-3′; rLamp2b-E7, (forward)
   5′-TTGGCTAATGGCTCAGCTTT-3′ and (reverse) 5′-ATGGGCACAAGGAAGTTGTC-3′;
   and cel-miR-39, (forward) 5′-agcccgTCACCTGGTGTAAATC-3′ and (reverse)
   5′-CAGTGCAGGGTCCGAGGTAT-3′.

   In qPCR for osteogenic inhibition, Gli1^+ cells were dissolved in
   TRIzol reagent, and total RNA was isolated by a TRIpure total RNA rapid
   extraction kit. cDNA was synthesized using a TSK302S (RT6 cDNA
   Synthesis Kit version 2) and amplified in a RT-qPCR thermocycler
   (FQD-96A, BORI Hangzhou, China) using the 2× T5 Fast qPCR Mix (SYBR
   Green I). The relative gene expression was normalized to the expression
   of actin. The primers used for RT-PCR were as follows: MYH11, (forward)
   5′-GGCCATGAGTGACAGAGTCC-3′ and (reverse) 5′-TTGCCGTGCACTCTCATTCT-3′;
   α-SMA, (forward) 5′-CAGCTATGTGGGGGACGAAG-3′ and (reverse)
   5′-TCCGTTAGCAAGGTCGGATG-3′; thrombospondin 2, (forward)
   5′-CCCAAGGGGACCACACAAAT-3′ and (reverse) 5′-TAGTCATCGTCCCGGTCAGT-3′;
   OPN, (forward) 5′-CCGAGGTGATAGCTTGGCTT-3′ and (reverse)
   5′-GACTCATGGCTGGTCTTCCC-3′; BMP-2, (forward) 5′-GCCAAACACAAACAGCGGAA-3′
   and (reverse) 5′-GGTGATCAGCCAGGGGAAAA-3′; Runx2, (forward)
   5′-ATGGCCGGGAATGATGAGAA-3′ and (reverse) 5′-GGGGAGGATTTGTGAAGACCG-3′;
   and actin, (forward) 5′-AGATCAAGATCATTGCTCCTCCT-3′ and (reverse)
   5′-ACGCAGCTCAGTAACAGTCC-3′.

Gli1^+ cell endocytosis of E7-EXOs

   Gli1^+ cells were prestained with PKH67 (green), and SMCs, ECs,
   neutrophils, and macrophages were stained with CellMasTM Deep Red
   plasma membrane stain (Life Technology) (blue). Gli1^+ cells and SMCs,
   ECs, neutrophils, and macrophages were separately cultured in 24-well
   plate at a density of 2 × 10^5 cells. Meanwhile, a cocktail containing
   both Gli1^+ cells and SMCs, ECs, neutrophils, and macrophages (cell
   density of 1 × 10^5, respectively) was seeded together in 24-well
   plates. E7-EXO and M7-EXO were fluorescently labeled with PKH26 (red)
   membrane dye (Bestbio). After labeling, the E7-EXO and M7-EXO were
   washed with 50 ml of PBS, collected by ultracentrifugation, and
   resuspended in PBS. After cell adherent, 2 μg of E7-EXO or M7-EXO was
   added into 2 × 10^5 recipient cells and incubated for 24 hours.
   Unphagocytic exosomes were washed with PBS, and the nuclei were stained
   with 4′,6-diamidino-2-phenylindole (DAPI) after fixation with 4%
   paraformaldehyde. Olympus laser scanning confocal microscope (Olympus
   SpinSR, Japan) was used to observe the phagocytosis of exosomes by
   Gli1^+ cells and macrophages.

Vascular cell preparation and Ancr/E7-EXO–modified TEBVs

   The carotid artery of Sprague-Dawley rats was isolated, and the blood
   was rinsed away with PBS. The vessels were digested with 20% trypsin at
   4°C for 2 hours and then washed twice with PBS to obtain the vascular
   matrix. The prepared vascular matrix was first incubated with collagen
   solution (4 mg ml^−1) for 24 hours, washed twice with PBS, and then
   incubated with 0.1 M EDC, 0.7 M NHS, and 20 μl of Ancr/E7-EXO (25
   μg/μl) at 4°C for 48 hours to obtain Ancr/E7-EXO–modified TEBVs.

Implantation of TEBVs in Sprague-Dawley rats and establishment of CKD rat
model

   All animal surgeries and experiments were approved by the Animal Ethics
   Committee of the Military Medical University. Male Sprague-Dawley rats
   (230 to 250 g) were treated with an intraperitoneal injection of 1%
   sodium pentobarbital (40 mg/kg). The rats were anesthetized, and the
   two ends of one side of the common carotid artery were removed by
   forceps, cut in the middle, pulled out by a self-made cuff, and
   flipped. Both ends of the prepared TEBVs [1 cm long, 1 mm diameter,
   soaked in heparin sodium (50 U/ml)] were anastomosed to the end of the
   carotid artery. The vascular suture was ligated at the cuff and the
   vascular clamp was released. The diabetic rats were treated with an
   intraperitoneal injection of streptozotocin on day 7 after the
   operation to establish the diabetic rat model, which was performed
   according to the literature ([186]41). A two-step subtotal nephrectomy
   (5/6Nx) was performed according to the literature ([187]14), and a
   right single nephrectomy was performed 1 week after the left subtotal
   nephrectomy. Rats in the sham operation group were fed with normal
   diet. Rats in the Ancr/E7 group were treated with Ancr/E7-EXO (10
   mg/kg) through the tail vein every 5 days. The control group and the
   sham group were treated with the same volume of normal saline for 13
   weeks. Calcification was accelerated by feeding the rats with a
   high-phosphorus diet (1.2% Pi) throughout the study. The CKD model was
   established successfully by increasing the serum urea nitrogen level.
   Rats were euthanized at 13 weeks after surgery.

Assay of Gli1^+ phagocytosis of Ancr/E7-EXO in vivo

   Ancr/E7-EXO was fluorescently labeled with
   1,1'-Diocatadecyl-3,3,3',3'-Tetramethylindodicarbocyanine,4-Chlorobenze
   nesulfonate Salt (DiD) membrane dye (Engineering For Life), and, then,
   Ancr/E7-EXOs were washed twice with 10 ml of PBS by ultracentrifugation
   at 100,000g for 70 min at 4°C. Then, Ancr/E7-EXO was modified to
   vascular matrix in accordance with the above method. The
   Ancr/E7-EXO–modified TEBVs were implanted into rats. On days 7, 14, and
   21, rats were anesthetized, and, then, the TEBVs were harvested and
   postfixed with 4% paraformaldehyde for 3 hours at 4°C and then placed
   in 30% sucrose overnight. TEBVs were embedded by Optimal Cutting
   Temperature Compound (OCT) and then sectioned.

Assay of Ancr/E7-EXO stability in vivo

   Ancr/E7-EXO was fluorescently labeled with DiR membrane dye probe
   (Abcam), and the unbound probe was washed twice with 10 ml of PBS by
   ultracentrifugation at 100,000g for 70 min at 4°C. Then, Ancr/E7-EXO
   stained by probe was modified to vascular matrix in accordance with the
   above method. The Ancr/E7-EXO–modified TEBVs were implanted into rats.
   Rats were anesthetized at various time points to receive skin
   preparation for FRI (NEWTON 7.0 FT-100. VILBER). Filter parameters were
   set to 782-nm emission and 750-nm excitation.

Histological analysis

   Harvested TEBVs were rinsed with saline and cut into two sections. One
   section of TEBVs was directly embedded in a cryogenic agent, cut into
   40-μm-thick sections, and used to measure the Young’s modulus by AFM.
   The other section of TEBVs was immediately fixed in 4%
   paraformaldehyde, embedded in paraffin, and cut into 5-μm-thick
   sections. Paraffin sections were stained with H&E, Masson’s trichrome,
   von Kossa, and Alizarin red. Images were observed under a bright-field
   microscope (Olympus VS200, Japan).

Immunofluorescence staining

   Cells were fixed in 4% paraformaldehyde for 15 min, and the paraffin
   sections of TEBVs were dewaxed. After washing twice with PBS, 0.05%
   Triton X-100 (Solarbio, 30 min, room temperature) and 10% goat serum
   (Solarbio, 40 min, 37°C) in PBS were used for permeation and blocking,
   respectively. The samples were then incubated with antibodies at 4°C
   overnight. The primary antibodies were as follows: mouse anti-Gli
   (1:200, Novus Biologicals), mouse anti-MYH11 (1:200, Abcam), mouse
   anti–α-SMA (1:300, Boster), rabbit anti-thrombospondin (1:200, Abcam),
   rabbit anti-Runx2 (1:200, Bioss), and mouse anti-OPN (1:200, Bioss).
   The secondary antibodies were as follows: Alexa Fluor 488–, Alexa Fluor
   568–, and Alexa Fluor 647–conjugated antibodies (Invitrogen). The
   nuclei were stained with DAPI (Biosharp) for 5 min at room temperature.
   Confocal microscopy images were taken and analyzed by the Olympus laser
   scanning confocal microscope (Olympus SpinSR, Japan) and cellSens
   software, respectively.

Macroscopic fluorescence reflection imaging

   Aortic calcification was detected by a bisphosphonate derivative
   near-infrared fluorescent imaging agent (Osteosense 750 EX,
   PerkinElmer, Boston, MA, USA), which was injected into the rats through
   the tail vein 24 hours before imaging. The rats were euthanized, and
   the aorta was perfused with normal saline. After dissection, FRI was
   used to map the near-infrared fluorescence signal using OsteoSense 750
   EX fluorescent imaging agent (excitation/emission: 745/800 nm).

Ultrasound and micro-CT detection of TEBVs in vivo

   The rats were anesthetized with 1% sodium pentobarbital, and the neck
   skin was prepared. The neck of the rats was smeared with ultrasonic
   gel. The color Doppler ultrasound (Esaote) frequency was 13.0 MHz, the
   depth was 2.5 cm, the mechanical index was M10.4, and the gray scale
   was 46. After the detection of several consecutive cardiac cycles, the
   Doppler mode was selected to obtain the common carotid artery
   ultrasound spectrum. The images were saved, and the carotid blood flow
   velocity of the grafted TEBVs was measured. After anesthesia, the
   thoracic cavity of the rats was opened, and the angiography agent
   iohexol (300 mg/ml; Sigma-Aldrich) was injected from the left
   ventricle. The patency of TEBVs in each group was characterized by CT
   images using Bruker micro-CT of the carotid artery in small animals.

AFM assay

   An AFM (JPK NanoWizard NanoOptics, Germany) was used to measure vessel
   stiffness and to scan vessel media. The optical microscopy mirror
   (Eclipse-Ti-S, Nikon, Japan) combined with AFM was able to control the
   tip and sample positioning. All measurements were performed with the
   same soft cantilever (MLCT Bio, Bruker; nominal spring constant, 0.01
   N/m). Four points were measured for each blood vessel. A total of 50 to
   60 points were measured, and the average Young’s modulus was
   calculated. AFM force mapping of the media was performed in the range
   of 30 μm.

Bulk RNA-seq analysis of aorta and thoracic aorta of CKD rats

   Harvested aorta and thoracic aorta of CKD rats were rinsed with
   precooled PBS and immediately flash-frozen in liquid nitrogen. The raw
   RNA-seq data were collected by Berry Genomics Corporation, Beijing,
   China. R package DESeq2 v1.28.1 was used to perform differential gene
   expression analysis of bulk RNA-seq data. The resulting P values were
   adjusted using the Benjamini and Hochberg’s approach for controlling
   the false discovery rate. Genes with |log[2] (fold change)| > 1 and q
   value < 0.05 were assigned as differentially expressed. GO and KEGG
   enrichment analysis of differentially expressed gene sets were
   implemented by the top G
   ([188]www.bioconductor.org/packages/release/bioc/html/topGO.html) and
   KOBAS ([189]42) package, respectively.

Statistical analysis

   Statistical analysis was performed using GraphPad Prism 8.0 (San Diego,
   CA, USA) software. All experimental results were presented as means ±
   SD. Two groups were compared using two-tailed Student’s t test. A value
   of P < 0.05 was considered statistically significant.

Acknowledgments