Abstract

Background

   Emerging evidence has highlighted the therapeutic potential of human
   umbilical cord mesenchymal stem cells (UC-MSCs) in chemotherapy-induced
   premature ovarian failure (POF). This study was designed to investigate
   the appropriate timing and molecular mechanism of UC-MSCs treatment for
   chemotherapy-induced POF.

Methods

   Ovarian structure and function of mice were assessed every 3 days after
   injections with cyclophosphamide (CTX) and busulfan (BUS). UC-MSCs and
   UC-MSCs-derived extracellular vesicles (EVs) were infused into mice via
   the tail vein, respectively. Ovarian function was analyzed by follicle
   counts, the serum levels of hormones and ovarian morphology. The
   apoptosis and proliferation of ovarian granulosa cells were analyzed in
   vitro and in vivo. Label-free quantitative proteomics was used to
   detect the differentially expressed proteins in UC-MSC-derived EVs.

Results

   After CTX/BUS injection, we observed that the ovarian function of POF
   mice was significantly deteriorated on day 9 after CTX/BUS infusion.
   TUNEL assay indicated that the number of apoptotic cells in the ovaries
   of POF mice was significantly higher than that in normal mice on day 3
   after CTX/BUS injection. Transplantation of UC-MSCs on day 6 after
   CTX/BUS injection significantly improved ovarian function, enhanced
   proliferation and inhibited apoptosis of ovarian granulosa cells,
   whereas the therapeutic effect of UC-MSCs transplantation decreased on
   day 9, or day 12 after CTX/BUS injection. Moreover, EVs derived from
   UC-MSCs exerted similar therapeutic effects on POF. UC-MSCs-derived EVs
   could activate the PI3K/AKT signaling pathway and reduce ovarian
   granulosa cell apoptosis. Quantitative proteomics analysis revealed
   that clusterin (CLU) was highly expressed in the EVs of UC-MSCs. The
   supplementation of CLU proteins prevented ovarian granulosa cells from
   chemotherapy-induced apoptosis. Further mechanistic analysis showed
   that CLU-knockdown blocked the PI3K/AKT signaling and reversed the
   protective effects of UC-MSCs-derived EVs.

Conclusions

   Administration of UC-MSCs and UC-MSCs-derived EVs on day 6 of CTX/BUS
   injection could effectively improve the ovarian function of POF mice.
   UC-MSCs-derived EVs carrying CLU promoted proliferation and inhibited
   apoptosis of ovarian granulosa cells through activating the PI3K/AKT
   pathway. This study identifies a previously unrecognized molecular
   mechanism of UC-MSCs-mediated protective effects on POF, which pave the
   way for the use of cell-free therapeutic approach for POF.

Graphical Abstract

   [48]graphic file with name 13287_2024_3926_Figa_HTML.jpg

Supplementary Information

   The online version contains supplementary material available at
   10.1186/s13287-024-03926-7.

   Keywords: POF, UC-MSCs, EVs, Clusterin, PI3K/AKT

Introduction

   Premature ovarian failure (POF) is an ovarian defect characterized by
   markedly reduced ovarian reserves in women aged less than 40 years
   [[49]1, [50]2]. The incidence of POF has been reported to be 1% in
   females under the age of 40 years [[51]3]. Chemotherapy and/or
   radiation therapy can lead to apoptosis of normal cells and damage the
   reproductive system, which have become common and important causes of
   POF in young cancer patients [[52]4]. Hormone replacement therapy (HRT)
   is the primary treatment for POF-related symptoms including either
   estrogen therapy or a combination of estrogen and progesterone therapy
   [[53]5]. However, HRT can only relieve menopausal symptoms, slow
   deterioration, fail to repair damaged ovarian tissue [[54]6–[55]8].

   During the past few years, mesenchymal stem cells (MSCs) therapy has
   attracted increasingly more attention due to their potential to restore
   ovarian structure and function [[56]9–[57]11]. Several reports have
   shown that human umbilical cord MSCs (UC-MSCs) can be used for
   improving ovarian function and restoring serum sex hormone levels, as
   well as inhibiting apoptosis of ovarian tissue [[58]12, [59]13].
   Meanwhile, increasing evidence suggests that paracrine signaling plays
   a role in the therapeutic efficacy of UC-MSCs. Extracellular vesicles
   (EVs), membranous structures secreted by cells, have emerged as major
   paracrine effectors. EVs include exosomes (with a diameter ranging from
   30 to 150 nm), microcapsules (200–500 nm in diameter), and apoptotic
   bodies (exceeding 1000 nm in diameter) [[60]14]. In addition, EVs have
   been identified as important messengers between cells, delivering
   functional molecules such as microRNAs, long noncoding RNAs, and
   proteins to recipient cells [[61]15, [62]16]. Several studies have
   indicated that EVs promoted angiogenesis, attenuated cell apoptosis and
   restore ovarian function via transferring functional miRNAs or proteins
   [[63]4, [64]17, [65]18]. Nevertheless, the efficacy of UC-MSCs in
   preclinical and clinical studies of POF is not always satisfactory,
   with results varying dependent on the timing of administration in
   different studies. Meanwhile, the potential mechanisms of the
   protection provided by UC-MSCs and UC-MSCs-derived EVs in POF have not
   been elucidated. Based on the above findings, we speculate that UC-MSCs
   may restore the ovarian function of POF mice by transferring
   EVs-containing proteins.

   Here, we aimed to identify the pathological changes in the ovaries of
   mice with chemotherapeutic drug-induced POF over time and further
   explore the appropriate timing of UC-MSCs or UC-MSCs-derived EVs
   therapy. Meanwhile, we characterized the proteome profiles of
   UC-MSCs-derived EVs and explored the underlying molecular mechanism of
   UC-MSCs-derived EVs in restoring ovarian function.

Materials and methods

Isolation, culture, and identification of UC-MSCs

   The utilization of human umbilical cord tissue was endorsed by the
   Institutional Ethics Review Board of Renmin Hospital of Wuhan
   University (Permit Number: WDRY2019-G001). Informed consent was
   obtained from all participants. The umbilical cord tissue was
   sterilized with 75% alcohol, followed by multiple rinses with
   phosphate-buffered saline (PBS). After removing the umbilical vein and
   artery, the tissue was cut into small pieces and evenly tiled on the
   sterile culture dishes. The dishes were supplemented with serum-free
   stem cell culture medium (YoCon, Beijing, China), and placed in a
   37 °C, 5% CO[2] incubator. The culture medium was renewed at intervals
   of 2–3 days. Cells were trypsinized and passaged until they reached
   80–90% confluence. After the fourth generation of cells was cultured
   for 72 h, the supernatant was collected as UC-MSC-conditioned medium
   (MSC-CM) for subsequent experiments. GW4869 has been widely used for
   inhibiting the secretion of EVs. When the fourth generation of cells
   reached a confluence level of 70–80%, the culture medium was replaced
   with fresh medium containing GW4869 (10 μM), and the supernatant was
   collected after 24 h as GW4869-CM for subsequent experiments.

   UC-MSCs were analyzed for surface markers using the Human MSC Analysis
   Kit (562,245, BD Biosciences, NJ, USA). Lipogenic, osteogenic, and
   chondrogenic differentiation of UC-MSCs were performed to verify the
   differentiation potential of the MSCs using specialized inducers
   (6114531, 6114541, and 6114551, DAKEWE, Shenzhen, China). Alizarin Red,
   Alcian Blue and Oil Red O stainings were used to identify osteoblasts,
   adipocytes, and chondrocytes, respectively.

Preparation and identification of UC-MSCs-derived EVs

   The MSC-CM was sequentially centrifuged at 300 g, 2000 g, and 10,000 g
   to remove large particles and cell debris. The supernatant was
   transferred to a 100 kDa MWCO ultrafiltration tube (UFC9050, Merck,
   USA) and centrifuged at 5000 g for 30 min to collect the concentrated
   solution above the membrane. After that, the collected solution was
   centrifuged at 100,000 g for 70 min to remove the supernatant. The
   centrifugation process was repeated once and the pellet resuspended in
   an appropriate volume of PBS. The entire experiment was conducted at
   4 °C.

   The size and particle concentration of EVs were analyzed using
   nanoparticle tracking analysis (NTA). Additionally, the morphology of
   EVs was evaluated using transmission electron microscopy (TEM). Western
   blotting was employed to confirm the expression of related markers
   (CD9, Alix) and negative control (Calnexin). The dose of EVs derived
   from UC-MSCs was quantitatively determined by bicinchoninic acid assay
   (BCA) protein assay.

Animals

   Seven-week-old female specific pathogen-free (SPF level) C57/BL6 mice
   (Hubei Provincial Center for Disease Control and Prevention, Wuhan)
   were used in this study. All animal experiments were approved by the
   Committee of Animal Care and Use of Hubei Provincial Center for Food
   and Drug Safety Evaluation and Animal Experiment (Permit Number:
   202210022) and the Bestcell Model Biological Center Notification of the
   Result for Ethical Approval for Research Involving Animals (The permit
   Number: BSMS 2022-06-27A). All animal experiments were complied with
   the guidelines for the care and use of laboratory animals. All the
   animals were allowed free access to distilled water, standard food and
   were housed under standard feeding conditions (temperature 20 ± 2 °C;
   humidity 45–55%; 12 h light & dark cycle). Mice were randomly divided
   into 4–5 mice per cage. The animals were excluded if the animal died
   prematurely, preventing the collection of behavioral and histological
   data.

Animal model

   POF was induced in female C57BL/6 mice by intraperitoneal injections
   with busulfan (BUS, 30 mg/kg, MedChemexpress, NJ, USA) and
   cyclophosphamide (CTX, 120 mg/kg, Sigma-Aldrich, MO, USA). Mice were
   randomly divided into two groups (34 mice/group), with four or five
   mice per group (NC and POF groups) were sacrificed at 3, 6, 9, 12, 15,
   18, and 21 days after injection.

   The POF mice receiving UC-MSCs (5 per group) were injected with 200 μL
   PBS containing UC-MSCs (2 × 10^7 cells/kg) via tail vein on days 6, 9,
   and 12 after CTX and BUS treatment, respectively. Mice (5 per group)
   undergoing EVs treatment were injected with 200 μL of PBS containing
   75 μg EVs by tail vein on day 6. All mice were examined for estrous
   cycles 3 weeks after CTX/BUS injection and were sacrificed when they
   were in diestrus phase. The test time was between 09.30 and 11.30 am
   and testing order was randomized daily. Mice serum was sampled by
   extracting the eyeball blood after mice were anesthetized with 1%
   pentobarbital sodium. After that, all mice were euthanized by cervical
   dislocation. The animals were included in the study if they underwent
   successful UC-MSCs or UC-MSCs-derived EVs injection. All study
   personnel were blind to treatment allocation and had no way of
   influencing whether an animal would receive PBS, UC-MSCs or EVs derived
   from UC-MSCs.

Ovarian morphology analysis and follicle counts

   The ovaries were collected and fixed with tissue fixation solution
   (B0038, Powerful biology). The fixed ovaries were dehydrated, embedded
   in paraffin, and cut into 5 μm sections for hematoxylin and eosin (HE)
   staining and immunofluorescence analysis. The ovarian tissue morphology
   was observed using an optical microscope (Olympus Corporation, Tokyo,
   Japan), and the number of follicles at different stages, including
   primordial, primary, secondary, and atretic, were counted [[66]19,
   [67]20]. Three sections were analyzed for each tissue sample from
   n = 4–5 mice per group.

Immunofluorescence assay

   Ovarian tissue sections were fixed and incubated with 3% bovine serum
   albumin (BSA) for 30 min at 37 °C. Subsequently, the sections were
   incubated with primary antibodies against cleaved-caspase3 (1:400)
   (#9661, CST, MA, USA) or proliferating cell nuclear antigen (PCNA)
   (1:200) (GB12010, Servicebio, Wuhan, China). After washing with PBS,
   the sections were incubated with secondary antibodies at 37 °C for 1 h.
   DAPI staining was performed (Servicebio, Wuhan, China), and images were
   detected using a laser scanning confocal microscope (Nikon Corporation,
   Tokyo, Japan).

TdT mediated dUTP nick end labeling (TUNEL) assay

   According to the manufacturer's instructions, the TUNEL apoptosis assay
   kit (Servicebio, Wuhan, China) was used to detect apoptotic cells in
   ovarian tissue sections from different groups, followed by quantitative
   analysis.

Hormone assay

   Blood samples were collected from mice and centrifuged at 2000 rpm for
   20 min at 4 °C. Serum levels of E2, and FSH were measured using ELISA
   kits following the manufacturer's instructions.

Cell culture and treatment

   Human ovarian granulosa cells (KGN cell line) and human renal
   epithelial cells (HEK293T cell line) were purchased from the China
   Center for Type Culture Collection (Wuhan China). KGN cells were
   cultured in Dulbecco’s modified Eagle’s medium (DMEM, Gibco, NY, USA)
   containing 5% fetal bovine serum (FBS, Gibco, NY, USA), while HEK293T
   cells were cultured in DMEM medium supplemented with 10% FBS at 37 °C
   and 5% CO[2] concentration in a humidified atmosphere. Following 72 h
   of incubation, when the cell fusion efficiency reached 80–90%, the
   supernatant was collected as KGN-CM and 293 T-CM for subsequent
   experiments.

   An ovarian granulosa cell injury model was established by treating KGN
   cells with 5 μg/mL nitrogen mustard (NM) (mechlorethamine
   hydrochloride, Macklin, Shanghai, China) for 48 h in vitro. KGN cells
   were seeded at 3 × 10^4 cells in 6-well plates. After 24 h, these cells
   were exposed to media with 50% various supernatant, respectively.
   Subsequently, co-cultures were incubated for 48 h with NM.

Annexin-V-FITC and propidium iodide staining

   According to the manufacturer's instructions, cells from different
   groups were collected, washed twice with PBS, and resuspended. Cells
   were then stained with Annexin V-FITC/PI Apoptosis Detection Kit
   (C1062, Beyotime Biotechnology, China) for 20 min. Finally, the
   apoptotic cell rate was determined using flow cytometry.

CCK-8 assay

   KGN cells were seeded at 3000 cells in 96-well plates. After replacing
   the medium, NM was added to induce the cells for 24 h, and their growth
   was measured after adding 10 μL CCK-8 into the culture medium for
   1.5 h. The optical density (OD) value at 450 nm was determined.

siRNA transfection

   Cells were seeded into a 6-well plate at a density of 1 × 10^5 cells
   per well. For small interfering RNA (siRNA)-mediated gene knockdown,
   100 nM of siRNA was transfected into the cells using an EL (TransGen
   Biotech, China) transfection reagent, according to the manufacturer's
   protocols. The culture medium was changed after 6 h. The supernatant
   was harvested as siNC-CM and siCLU-CM 48 h post-transfection and
   filtered through a 0.22 um sterile filter. The siRNAs were synthesized
   by the Ribobio (Guangzhou, China). siCLU: 5′-GCATCATAGACGAGCTCTT-3′.
   siNC: 5′-TTCTCCGAACGTGTCACGT-3′.

Western blotting

   Protein was obtained through lysis of cells using
   radioimmunoprecipitation assay buffer (P0013C, Beyotime, China),
   supplemented with a mixture of protease inhibitors and phosphatase
   inhibitors. Protein concentration was analyzed using the bicinchoninic
   acid assay (BCA, P0009, Beyotime Biotechnology, China). The proteins
   were then separated by 10% sodium dodecyl sulfate–polyacrylamide gel
   electrophoresis (SDS-PAGE) and transferred onto a polyvinylidene
   difluoride (PVDF) membrane. The membrane was incubated with 5% BSA in
   Tris-buffered saline with 0.5% Tween-20 to block nonspecific binding.
   Primary antibodies against BAX (50,599–2-Ig, Proteintech, USA), BCL-XL
   (2764 s, CST, USA), cleaved-caspase 3 (9661 s, CST, USA), P-AKT
   (4060 s, CST, USA), AKT (46,917, CST, USA), P-PI3K (ab278545, abcam,
   UK), PI3K (4257, CST, USA), and β-actin (81,115–1-RR, Proteintech, USA)
   were incubated with the membrane overnight at 4 °C at an appropriate
   dilution (1:1000). After washing three times, secondary antibodies were
   added to the membrane and incubated at RT for 1 h. The chemiluminescent
   signal was detected using an enhanced chemiluminescence kit (Absin,
   CN). The results were analyzed using Image J software.

RNA extraction and real-time PCR analysis

   Total RNA was extracted from cultured cells with Trizol (15596-018,
   Invitrogen, USA), reverse-transcribed into cDNA with a reagent kit
   (RR047B, Takara, China), and amplified by qRT-PCR (CFX connect,
   Bio-Rad, USA) according to manufacturer's instructions. H-CLU-F:
   5′-AAACGAAGAGCGCAAGACAC-3′. H-CLU-R: 5′- TGTTTCAGGCAGGGCTTACA-3′.

Quantitative proteomics analysis

   EVs were subjected to Label-free quantification protein profiling, with
   the culture medium (Control) was designed as the negative control. All
   protein samples were prepared in triplicates for each biological
   replicate. The experiment was performed at Wuhan Biobank company
   (China). Mass spectrometry data were analyzed using MaxQuant software
   (V1.6.6) with the Andromeda database search algorithm (Cox and Mann
   2008; Tyanova, Temu, and Cox 2016). The protein identification was
   performed against the human protein reference database in Uniprot.
   Functional annotation and enrichment analysis were conducted using Gene
   Ontology (GO) ([68]https://www.geneontology.org/), while protein
   analysis related to metabolic pathways was performed using the Kyoto
   Encyclopedia of Genes and Genomes (KEGG) database
   ([69]https://www.genome.jp/KEGG).

Statistical analysis

   All data were analyzed using GraphPad Prism 9 software (La Jolla, CA,
   USA) and were presented as the mean ± standard deviation (SD).
   Differences between group means were determined with Student's t-test
   or one-way analysis of variance. P < 0.05 was considered statistical
   significance.

Results

Characterization of UC-MSCs and UC-MSC-derived EVs

   The staining results showed that UC-MSCs displayed osteogenic,
   chondrogenic, and adipogenic differentiation abilities in vitro
   (Fig. [70]1A–C). Flow cytometry analysis revealed that UC-MSCs were
   positive for CD90, CD105 and CD73 but negative for CD11b, CD34, CD19,
   CD45 and HLA-DR (Fig. [71]1D). NTA showed that the mean particle
   diameter of UC-MSC-derived EVs was 121.3 nm (Fig. [72]1E). Western
   blotting analysis revealed that the EVs were positive for molecular
   markers, including Alix, CD9 and TSG101, and negative for Calnexin
   (Fig. [73]1F). TEM analysis showed that EVs displayed round
   nanoparticles ranging from 50 to 200 nm (Fig. [74]1G).

Fig. 1.

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

   Identification of human UC-MSCs and UC-MSCs-derived EVs. A The
   adipogenic differentiation was defined by Oil red O staining. Scale
   bar: 20 μm. B The Osteogenic differentiation was determined by Alizarin
   red staining. Scale bar: 20 μm. C The chondrogenic differentiation was
   defined by Alcian blue staining. Scale bar: 4 μm. D Human UC-MSCs were
   positive for CD73, CD90 and CD105 (more than 99%) but negative for
   CD11b, CD34, CD19, CD45 and HLA-DR (less than 1%) by flow cytometry. E
   NTA particle size distribution analysis of EVs. F Western blot analysis
   of EVs- specific markers (Alix, Tsg101, and CD9) and cellular marker
   (Calnexin). The culture medium sample of non-cultured human UC-MSCs
   were used as a negative control. All full-length blots are presented in
   Additional file [77]2: Figure 1F. G The morphology and size of EVs
   under TEM. Scale bar: 200 nm

The structure and function of the ovaries were significantly impaired on day
9 after CTX/BUS injection

   After injection of CTX/BUS, the POF mice developed ovarian atrophy over
   time (Fig. [78]2A). The changes in the number of follicles showed a
   clear time-dependent pattern. Compared to that in the control mice, the
   number of primordial follicles (NC vs. POF, 9.3 ± 0.4 vs. 6.1 ± 0.8) in
   POF mice decreased significantly after 6 days of CTX/BUS injection.
   After 9 days of injection, the number of primary follicles (NC vs. POF,
   7.8 ± 0.7 vs. 5.0 ± 1.6) and secondary follicles (NC vs. POF,
   13.8 ± 5.0 vs. 7.1 ± 1.3) were also significantly decreased. In
   addition, the number of atretic follicles (NC vs. POF, 0.4 ± 0.1 vs.
   1.6 ± 0.1) in POF mice increased significantly on day 3 and remained at
   high levels (Fig. [79]2B). Moreover, we found that the E2 concentration
   (NC vs. POF, 125.7 ± 17.5 pg/mL vs. 86.1 ± 26.9 pg/mL) in POF mice was
   significantly lower than that of the control mice on day 9
   (Fig. [80]2C), whereas the FSH concentration in POF mice did not show
   significant change compared with the controls (Fig. [81]2D). These
   results suggested that ovarian architecture and function were
   dramatically impaired in POF mice after 9 days of CTX/BUS injection.

Fig. 2.

   [82]Fig. 2
   [83]Open in a new tab

   Time-dependent changes in the number of follicles and hormone levels
   after CTX/BUS injection. A Representative images of HE staining of
   ovarian tissues. Scale bar: 200 μm. B Number of primordial follicles,
   primary follicles, secondary follicles and atretic follicles at
   different time points. C ELISA for serum levels of E2. D ELISA for
   serum levels of FSH. *p < 0.05, **p < 0.01, ***p < 0.001. n = 4–5 per
   group for all experiments

Apoptosis of ovarian granulosa cells is one of the key phenomena in
CTX/BUS-induced POF

   In comparison to that in the control mice, TUNEL positive cells (NC vs.
   POF, 10.1 ± 3.4/mm^2 vs. 32.8 ± 15.0/mm^2) were markedly increased from
   day 3 in POF mice (Fig. [84]3A and C). Notably, we observed that
   TUNEL-positive cells were primarily ovarian granulosa cells. To
   determine whether CTX/BUS causes ovarian fibrosis, the fibrosis areas
   of ovarian tissue were examined using Masson staining. The Masson
   positive areas of ovaries in POF mice were not significantly different
   from that of the control mice (Fig. [85]3B and D).

Fig. 3.

   [86]Fig. 3
   [87]Open in a new tab

   Changes in apoptosis and fibrosis in ovarian tissue. A Representative
   images of TUNEL staining. Scale bar: 50 μm. B Representative images of
   Masson staining. Scale bar: 50 μm. C The quantification of TUNEL assay.
   D The quantification of Masson staining. *p < 0.05, **p < 0.01. n = 4–5
   per group for all experiments

The therapeutic effect of UC-MSCs treatment on day 6 of CTX/BUS injection was
significantly better than that on days 9 or 12

   Given that the ovarian function was declined significantly on day 9 of
   CTX/BUS injection, we explored the therapeutic timing of UC-MSCs for
   POF. The mice were injected intravenously with UC-MSCs (2 × 10^7
   cells/kg) on days 6, 9, and 12 of CTX/BUS injection, respectively. HE
   staining showed that ovarian atrophy was inhibited in mice treated with
   UC-MSCs on days 6 and 9. In contrast, no efficacy was observed in POF
   mice with UC-MSCs on day 12 (Fig. [88]4A). The number of primordial
   follicles (POF vs. POF + UC-MSC, 5.0 ± 0.8 vs. 9.5 ± 1.0), primary
   follicles (POF vs. POF + UC-MSC, 6.8 ± 1.8 vs. 11.1 ± 1.4), secondary
   follicles (POF vs. POF + UC-MSC, 4.5 ± 1.7 vs. 9.7 ± 1.8) and atretic
   follicles (POF vs. POF + UC-MSC, 1.3 ± 0.7 vs. 0.3 ± 0.3) was
   significantly restored by UC-MSCs treatment on day 6 (Fig. [89]4B).
   When mice were transplanted with UC-MSCs on day 9, the number of
   primordial follicles (POF vs. POF + UC-MSC, 5.0 ± 0.8 vs. 7.6 ± 1.7)
   and secondary follicles (POF vs. POF + UC-MSC, 4.5 ± 1.7 vs. 6.4 ± 0.7)
   was significantly increased compared to that in the POF mice, however,
   no difference was observed in the UC-MSCs-treated mice on day 12
   (Fig. [90]4B). Moreover, the E2 concentration was significantly
   enhanced (POF vs. POF + UC-MSC (Day 6) vs. POF + UC-MSC (Day 9) vs.
   POF + UC-MSC (Day 12), 47.3 ± 24.6 pg/mL vs. 94.1 ± 47.7 pg/mL vs.
   89.6 ± 40.6 pg/mL vs. 134.4 ± 91.9 pg/mL) in the all UC-MSCs-treated
   mice when compared to the POF mice (Fig. [91]4C). FSH concentration
   displayed a decreasing trend in the mice treated with UC-MSCs on day 6
   (Fig. [92]4D). Additionally, the ovarian weight (POF vs. POF + UC-MSC,
   6.8 ± 0.4 mg vs. 7.6 ± 0.6 mg) was significantly increased in mice
   treated with UC-MSCs on day 6 (Fig. [93]4E).

Fig. 4.

   [94]Fig. 4
   [95]Open in a new tab

   Therapeutic effects of UC-MSCs in POF mice. A Representative images of
   HE staining. Scale bar: 200 μm. B The number of different types of
   ovarian follicles. (C) ELISA for serum levels of E2. D ELISA for serum
   levels of FSH. E The weight of ovaries. *p < 0.05, **p < 0.01,
   ***p < 0.001. n = 5 per group for all experiments

   Meanwhile, TUNEL staining revealed that transplantation of UC-MSCs on
   day 6 markedly reduced the number of TUNEL-positive cells (POF vs.
   POF + UC-MSC, 33.1 ± 8.7/mm^2 vs. 17.7 ± 3.3/mm^2) in the ovaries,
   whereas no notable changes were observed in mice treated on days 9 or
   12 (Fig. [96]5A and C). To investigate whether UC-MSCs promoted the
   proliferation of ovarian cells, we detected proliferative cells by PCNA
   immunostaining. The results suggested that treatment with UC-MSCs on
   day 6 significantly increased the number of PCNA positive cells (POF
   vs. POF + UC-MSC, 2416 ± 415.4/mm^2 vs. 3321 ± 825.1/mm^2) in POF mice.
   The same effect was not seen in POF mice treated with UC-MSCs on days 9
   or 12 (Fig. [97]5B and D). Therefore, these findings indicated that
   UC-MSCs infusion on day 6 inhibited apoptosis and promoted
   proliferation of ovarian cells in POF mice. Collectively, these results
   demonstrated that the therapeutic effect of UC-MSCs treatment on day 6
   of CTX/BUS injection was significantly better than that on days 9 or
   12.

Fig. 5.

   [98]Fig. 5
   [99]Open in a new tab

   The anti-apoptotic and pro-proliferative effects of UC-MSC in POF. A
   Representative images of TUNEL staining. Scale bar: 50 μm. B
   Representative images of PCNA immunofluorescence staining. Scale bar:
   50 μm. C The quantification of TUNEL assay. D The quantification of
   PCNA immunofluorescence. **p < 0.01. n = 5 per group for all
   experiments

UC-MSC-derived EVs improved ovarian structure and function in POF mice

   EVs from UC-MSCs have been reported to have the same therapeutic role
   as UC-MSCs in disease treatment [[100]15]. We evaluated the role of
   injection of EVs derived from UC-MSCs in normal and POF mice on the day
   6 after CTX/BUS injection. HE staining showed the size of the ovaries
   was restored by treatment with EVs (Fig. [101]6A). Consistent with the
   improved effect of UC-MSCs treatment, EVs derived from UC-MSCs
   significantly increased the number of primordial follicles (POF vs.
   POF + EVs, 5.4 ± 0.8 vs. 8.4 ± 1.2), primary follicles (POF vs.
   POF + EVs, 3.2 ± 0.6 vs. 5.4 ± 1.1), and secondary follicles (POF vs.
   POF + EVs, 2.5 ± 1.0 vs. 5.0 ± 1.7) and decreased the number of atretic
   follicles (POF vs. POF + EVs, 1.0 ± 0.5 vs. 0.4 ± 0.2) in POF mice
   (Fig. [102]6B). Similarly, UC-MSC-derived EVs promoted the recovery of
   hormones (E2, POF vs. POF + EVs, 56.9 ± 25.5 pg/mL vs.
   86.9 ± 14.0 pg/mL; FSH, POF vs. POF + EVs, 32.5 ± 9.3 mIU/mL vs.
   16.1 ± 5.6 mIU/mL) and ovarian weight (POF vs. POF + EVs, 5.3 ± 0.9 mg
   vs. 7.1 ± 0.6 mg) (Fig. [103]6C–E). Furthermore, UC-MSC-derived EVs did
   not confer any toxicity to normal mice.

Fig. 6.

   [104]Fig. 6
   [105]Open in a new tab

   Therapeutic effects of EVs derived from UC-MSCs in POF mice. A
   Representative images of HE staining. Scale bar: 200 μm. B The number
   of different types of ovarian follicles. C The weight of ovaries. D
   ELISA for serum levels of E2. E ELISA for serum levels of FSH.
   *p < 0.05, **p < 0.01, ***p < 0.001. n = 5 per group for all
   experiments

UC-MSC-derived EVs attenuated apoptosis and promoted proliferation in ovarian
granulosa cells

   We further explored whether UC-MSC-derived EVs protected ovarian
   granulosa cells against CTX/BUS-induced apoptosis. Compared with those
   in the POF mice, the administration of EVs led to decreasing the number
   of TUNEL-positive cells (POF vs. POF + EVs, 107 ± 16.7/mm^2 vs.
   37.5 ± 11.5/mm^2) (Fig. [106]7A). Meanwhile, cleaved-caspase3
   expression in the ovaries of POF mice was significantly inhibited by
   the treatment of EVs (POF vs. POF + EVs, 75.9 ± 8.2/mm^2 vs.
   18.8 ± 6.9/mm^2) (Fig. [107]7B). Subsequently, we also examined the
   effect of EVs derived from UC-MSCs on the proliferation of ovarian
   granulosa cells. An increased number of proliferating cells in the
   ovaries of mice treated with EVs was observed through PCNA assay
   analysis (POF vs. POF + EVs, 2281 ± 604.9/mm^2 vs. 3659 ± 333.8/mm^2)
   (Fig. [108]7C). These results showed that EVs derived from UC-MSCs
   exerted an anti-apoptosis and pro-proliferative effect in the ovaries
   of POF mice, which was similar to that of UC-MSCs.

Fig. 7.

   [109]Fig. 7
   [110]Open in a new tab

   The anti-apoptotic and pro-proliferative effects of EVs derived from
   UC-MSCs in POF. A Representative images and quantification of TUNEL
   staining. Scale bar: 50 μm. B Representative images and quantification
   of cleaved-caspase3 immunofluorescence staining. Scale bar: 50 μm. C
   Representative images and quantification of PCNA immunofluorescence
   staining. Scale bar: 50 μm. *p < 0.05, **p < 0.01, ***p < 0.001. n = 3
   per group for all experiments

UC-MSC-derived EVs inhibited apoptosis and promoted proliferation in KGN
cells

   Flow cytometry showed that MSC-CM exerted an anti-apoptosis effect in
   KGN cells (NM vs. NM + MSC-CM, 40.4 ± 3.2% vs. 20.6 ± 2.0%), but
   293 T-CM had no effect (Fig. [111]8A). In addition, Western Blot
   analysis indicated that MSC-CM significantly up-regulated the relative
   protein expression levels of the anti-apoptotic protein BCL-XL and
   down-regulated pro-apoptotic protein cleaved-caspase3 (Fig. [112]8C).
   Activation of the PI3K/AKT signaling pathway is associated with cell
   proliferative and apoptosis. To investigate whether MSC-CM can activate
   the PI3K/AKT signaling pathway, we detected the relative protein
   expression levels of p-PI3K and p-AKT in KGN cells. The inclusion of
   MSC-CM led to an increase in the protein expression levels of p-PI3K
   and p-AKT in KGN cells, compared to the treatment with 293 T-CM
   (Fig. [113]8C). Furthermore, CCK8 assay indicated that MSC-CM
   significantly enhanced proliferation of KGN cells (NM vs. NM + MSC-CM,
   26.8 ± 6.2% vs. 43.6 ± 3.4%) (Fig. [114]8E).

Fig. 8.

   [115]Fig. 8
   [116]Open in a new tab

   EVs from UC-MSCs inhibited apoptosis and promoted proliferation of KGN
   cells. A, B The KGN cells apoptosis detected by flow cytometry. C, D
   The expression of apoptosis-related proteins and proteins related to
   the PI3K/AKT signaling pathway in KGN cells detected by Western blot
   analysis. All full-length blots are presented in Additional file
   [117]2: Figure 8C and D. E, F CCK8 assay for proliferation of KGN
   cells. *p < 0.05, **p < 0.01, ***p < 0.001

   We used the EVs inhibitor (GW4869) to inhibit the secretion of
   UC-MSC-derived EVs. The anti-apoptotic effect of MSC-CM on KGN cells
   was reversed by GW4869 pretreatment (NM + MSC-CM vs. NM + GW4869-CM,
   23.5 ± 3.1% vs. 52.4 ± 5.1%) (Fig. [118]8B and D). Meanwhile, after
   inhibition of EVs with GW4869, the activation of the PI3K/AKT signaling
   pathway in KGN cells induced by MSC-CM was suppressed (Fig. [119]8D).
   Furthermore, the inhibition of EVs significantly impaired the
   proliferative effect of MSC-CM on KGN cells (NM + MSC-CM vs.
   NM + GW4869-CM, 47.1 ± 3.5% vs. 31.6 ± 5.1%) (Fig. [120]8F). Our
   findings suggested that UC-MSC-derived EVs not only inhibit NM-induced
   apoptosis but also promote the proliferation of KGN cells, and that the
   PI3K/AKT signaling pathway might be involved in these processes.

Proteomic analysis of UC-MSC-derived EVs

   The quantitative proteomics analysis showed that 728 proteins were
   identified in the proteomes of control medium samples and EVs from
   human UC-MSCs, among which 522 proteins were significantly upregulated
   and 2 proteins were significantly downregulated in EVs (Fig. [121]9A
   and B). GO enrichment analysis annotated biological process in which
   the identified proteins were involved. Notably, the regulation of
   apoptotic signaling pathway was closely related to the efficacy of EVs
   on POF (Fig. [122]9C). Moreover, the KEGG analysis indicated
   significant enrichment of the PI3K/AKT signaling pathway, which was
   associated with 38 proteins (Fig. [123]9D). The five anti-apoptotic
   proteins that differed significantly from the control medium, including
   CLU, Gelsolin (GSN), Guanine nucleotide-binding protein G(i) subunit
   alpha-2 (GNAI2), Peroxiredoxin-2 (PRDX2) and 14-3-3 protein gamma
   (YWHAG), may play key regulatory roles in the inhibition of apoptosis
   by EVs (Fig. [124]9E).

Fig. 9.

   [125]Fig. 9
   [126]Open in a new tab

   Label-free quantitative proteomics comparison of UC-MSCs-derived EVs
   and NC medium samples. A Heat map of differential expression proteins
   in EVs and control medium. B Volcano plot of differential expression
   proteins in EVs and control medium. The red points indicated higher
   protein expression in the EVs and blue points indicated lower protein
   expression in the EVs compared to the control medium. C GO enrichment
   analysis of biological process. D KEGG pathway enrichment analysis. E
   Log2 proteins abundance of GSN, CLU, GNAI2, PRDX2 and YWHAG.
   **p < 0.01, ***p < 0.001

CLU of UC-MSCs-derived EVs inhibited apoptosis and promoted proliferation in
KGN cells by activating the PI3K/AKT pathway

   The protein expression of CLU, GSN, GNAI2, PRDX2 and YWHAG in UC-MSCs
   was inhibited significantly by siRNA transfection (Fig. [127]S1A). Flow
   analysis results indicated that the downregulation of CLU expression by
   siRNA significantly impaired the anti-apoptotic effect of MSC-CM
   (NM + siNC-CM vs. NM + siCLU-CM, 19.4 ± 0.33% vs. 27.0 ± 1.6%)
   (Fig. [128]10A). Consistently, treatment with exogenous CLU proteins
   also significantly inhibited NM-induced apoptosis in KGN cells (NM vs.
   NM + CLU-100 ng/mL, 30.1 ± 2.9% vs. 19.7 ± 1.2%) (Fig. [129]10B). By
   contrast, the downregulation of GSN, GNAI2, PRDX2 and YWHAG expression
   in UC-MSCs did not affect NM-induced apoptosis of KGN cells (Fig.
   [130]S1B). When CLU was down-regulated in UC-MSCs, the ability of
   MSC-CM to regulate BCL-XL and cleaved-caspase3 in KGN cells was
   suppressed (Fig. [131]10C). We next investigated whether CLU of
   UC-MSCs-EVs had anti-apoptotic effect on KGN cells through the PI3K/AKT
   pathway. The relative protein expression levels of p-PI3K and p-AKT
   were decreased with CLU knockdown (Fig. [132]10C). CCK8 results showed
   that knockdown of CLU downregulated the proliferative effect of MSC-CM
   on KGN cells (NM + siNC-CM vs. NM + siCLU-CM, 70.6 ± 7.0% vs.
   56.5 ± 7.5%) (Fig. [133]10D). Likewise, exogenous CLU proteins had a
   pro-proliferative effect over KGN cells (NM vs. NM + CLU-100 ng/mL,
   35.4 ± 13.0% vs. 61.3 ± 9.2%) (Fig. [134]10E). Taken together, our
   findings demonstrated that CLU from UC-MSCs-derived EVs inhibited
   apoptosis and promoted proliferation of ovarian granulosa cells by
   activating the PI3K/AKT pathway.

Fig. 10.

   [135]Fig. 10
   [136]Open in a new tab

   CLU of UC-MSCs-derived EVs inhibited apoptosis and promoted
   proliferation of KGN cells. A, B The effect of siCLU-CM and exogenous
   CLU proteins on KGN cell apoptosis were detected by flow cytometry. C
   Western blot analysis was performed to detect the effects of MSC-CM
   after knockdown of CLU on the expression of apoptosis-related proteins
   and proteins related to the PI3K/AKT signaling pathway in KGN cells.
   All full-length blots are presented in Additional file [137]2:
   Figure 10C. D, E CCK8 assay to detect the effect of MSC-CM after
   knockdown of CLU and exogenous CLU proteins on KGN cell proliferation.
   *p < 0.05, **p < 0.01

Discussion

   In this study, we found that mice exhibit ovarian failure after 9 days
   of intraperitoneal injections with CTX/BUS. Treatment of POF mice with
   human UC-MSCs and EVs derived from UC-MSCs on day 6 was able to
   significantly improve ovarian function in chemotherapy-induced ovarian
   damage. Meanwhile, human UC-MSCs and EVs had anti-apoptotic and
   pro-proliferative effects on ovarian granulosa cells in vivo and in
   vitro. Further mechanistic studies showed that CLU of EVs derived from
   human UC-MSCs inhibited apoptosis and promoted proliferation of ovarian
   granulosa cell through activating the PI3K/AKT pathway.

   POF is a disorder with multiple causative factors, including genetic
   factors, autoimmune disorders, chemotherapy and/or radiotherapy, as
   well as environmental pollutants [[138]16, [139]21, [140]22]. POF is
   one of the sequelae of cytotoxic chemotherapy among young female with
   cancer. CTX and BUS are the most commonly used alkylating agent
   antitumor drugs. Apoptosis of ovarian cells, ovarian cortex fibrosis
   and depletion of ovarian reserve have been reported to be involved in
   chemotherapy induced POF [[141]18, [142]23, [143]24]. To investigate
   the pathological changes of POF induced by chemotherapeutic agents, we
   examined the number of follicles, hormone levels, cell apoptosis and
   fibrosis in POF and NC mice every three days after injection of CTX
   with BUS until day 21. Our results revealed that the mice treated with
   CTX/BUS exhibited a significant loss of primordial follicles on day 6,
   and a significant reduction in the number of follicles at all stages on
   day 9, compared with the control group. On the other hand, the number
   of atretic follicles in POF mice was significantly increased on day 3.
   The previous studies reported that the mechanism of follicle depletion
   involves increased recruitment of primordial follicles into the growing
   pool and induction of apoptotic death in growing follicles [[144]25].
   Here, we observed that the primordial follicle depletion was earlier
   than primary and secondary follicles. Our data further support that
   reduction of the primordial pool can arise indirectly via the loss of
   activated and/or growing follicles. In addition, the apoptosis of
   ovarian granulosa cells in POF mice was significantly increased and
   maintained high levels 3 days after CTX/BUS injection, but no signs of
   fibrosis were observed within 21 days. We demonstrate here that on day
   9 after CTX/BUS injection, mice developed significant ovarian failure,
   which was closely associated with the apoptosis of ovarian granulosa
   cells.

   Ovarian tissue cryopreservation with subsequent transplantation is one
   of the means for female fertility preservation [[145]26]. However, this
   treatment carries the risk of carrying malignant cells [[146]23]. HRT
   could suppress menopausal symptoms but barely improve reproductive
   functions of the ovaries [[147]27, [148]28]. In recent years,
   MSCs-based therapies have been reported to be effective in treating POF
   [[149]13, [150]28, [151]29]. UC-MSCs have no ethical difficulties and
   can be easily obtained in comparison to other sources of stem cells.
   Several studies showed that human UC-MSCs transplantation restores
   ovarian function by inhibiting apoptosis, improving lipid metabolism
   and alleviating excessive autophagy [[152]13, [153]18, [154]30].
   However, the timing of human UC-MSCs treatment is a critical factor
   determining the recovery of ovarian function, further studies are
   required to explore the appropriate timing. Based on the pathological
   changes observed in POF, the current study administered human UC-MSCs
   to mice via tail vein injection on crucial time points of days 6, 9,
   and 12 to investigate their potential therapeutic effects. POF mice
   that were injected with UC-MSCs on day 6 significantly increased the
   number of primordial, primary and secondary follicles, and inhibited
   follicle atresia. We observed increased numbers of primordial and
   secondary follicles, but the number of primary and atretic follicles
   did not significantly change in POF mice treated with UC-MSCs on day 9.
   Notably, no significant differences in follicle numbers were observed
   in POF mice treated with UC-MSCs on day 12. Correspond with previous
   studies [[155]12], human UC-MSCs inhibited ovarian cell apoptosis and
   promoted proliferation. However, we further found that there was no
   anti-apoptosis and pro-proliferative effect when UC-MSCs were
   administered on days 9 and 12. These results suggest that the timing of
   UC-MSCs injections is critical for therapeutic intervention.

   Numerous studies have demonstrated that paracrine mechanisms are
   critical in human UC-MSCs-mediated improvement of ovarian function
   [[156]17, [157]31]. Here, we isolated EVs to investigate the role of
   EVs derived from human UC-MSCs in POF. The results showed that EVs
   derived from human UC-MSCs inhibited apoptosis and restored the ovarian
   structures and function in chemotherapy-induced POF mice. Meanwhile, we
   observed that EVs promoted cell proliferation in ovaries. Previously,
   Eslami et al. have shown that injecting POF mice with 400 μg of EVs
   derived from bone marrow MSCs improved ovarian function and restored
   fertility [[158]32]. Another study by Xiao et al. indicated that
   intra-ovarian injection of amniotic fluid stem cell-derived exosomes
   (125 μg) was adequate to alleviate ovarian injury in
   chemotherapy-induced POF mice [[159]33]. In our study and previous
   studies, the timing, doses, and treatment approaches of EVs differed
   significantly among protocols. The doses of EVs were lower and the
   treatment approach was simpler in our study compared to those used in
   previous studies. Meanwhile, our results regarding the optimal timing
   of human UC-MSCs treatment in POF help to ascertain the timing for EVs
   treatment as well. In vitro, our data suggested that the apoptosis of
   NM-damaged KGN cells was significantly inhibited after treatment with
   UC-MSCs supernatant. Meanwhile, the proliferation of KGN was
   significantly enhanced after coculture with human UC-MSCs supernatant.
   However, there was no similar effect when KGNs were cocultured with
   293T-CM. In addition, when exosome secretion of human UC-MSCs was
   suppressed by GW4869, the anti-apoptosis and pro-proliferative effects
   of UC-MSC supernatant was weakened. These results demonstrate that
   UC-MSC-derived EVs have the anti-apoptosis and pro-proliferative
   effects, like UC-MSCs. However, there are still numerous issues that
   need to be resolved for their successful clinical application. There is
   no internationally accepted standardized method for the
   high-throughput, high-purity, and minimal damage separation of EVs from
   cell culture medium [[160]34]. Furthermore, the heterogeneity of EVs
   poses one of the main challenges in current clinical EVs studies.

   EVs can regulate the function of target cells by transferring
   reparative proteins, mRNAs, lncRNAs and miRNAs. Hu et al. proposed that
   human UC-MSCs derived EVs can regulate bone metabolism by transferring
   CLEC11A [[161]35]. Previous studies have shown that exosomal miRNA (eg,
   miR-22-3p and miR126-3p) from human UC-MSCs repair ovarian injury and
   improve ovarian function [[162]17, [163]18]. However, there are few
   studies on bioactive proteins of EVs for the treatment of POF induced
   by chemotherapeutic agents. To test the critical functional molecules,
   label-free quantitative proteomics was performed to detect proteins in
   EVs derived from human UC-MSCs. With the label-free quantitative
   proteomics, 522 cellular proteins levels were significantly upregulated
   in EVs. GO enrichment analysis showed that biological processes of
   these proteins were involved in apoptotic signaling pathways. The CLU
   is a glycosylated protein that involved in a variety of physiological
   processes, including apoptotic cell death, lipid transport and DNA
   repair [[164]36]. Some studies have suggested that CLU can protect
   against cardiomyocytes and podocytes from oxidative stress-induced
   apoptosis [[165]37, [166]38]. You et al. found that CLU expression
   could promote ovarian cancer cell proliferation and resistance to
   cisplatin [[167]39]. To our knowledge, the effect of CLU in POF has not
   been reported so far. In this study, our data indicated that CLU was
   highly enriched in EVs of human UC-MSCs. CLU knockdown was found to
   attenuate the protective effect of UC-MSC-derived EVs on KGN cells
   apoptosis. Meanwhile, the pro-proliferative effect of EVs was arrested
   when CLU was inhibited. We could not determine whether knockdown of CLU
   with siRNA in human UC-MSCs might affect the expression of other genes
   to affect the regulatory effects of EVs. Therefore, the anti-apoptosis
   and pro-proliferative effect of CLU was validated by the use of
   recombinant CLU proteins. Additionally, Hou et al. proposed that CLU
   replenishment may alleviate cell death, which supports our findings in
   the study [[168]40]. Indeed, engineered EVs loaded with additional
   amounts of CLU will further prove our conclusion. This will be
   corroborated in our follow‐up studies.

   It is well known that the PI3K/AKT pathway plays an important role in
   regulating autophagy, apoptosis, proliferation and follicular growth
   [[169]30, [170]41]. It has been reported that exosomes derived from
   human UC-MSCs could restore damaged ovarian function through the
   PI3K/AKT pathway [[171]18]. As expected, our findings provided evidence
   that the supernatants of UC-MSCs enriched with EVs modulated the
   PI3K/AKT pathway in ovarian granulosa cells. Evidence suggested CLU
   could inhibit senescence and inflammation through activation of
   PI3K/AKT pathway [[172]42, [173]43]. Therefore, we hypothesized that
   CLU could play a role in POF by activating the PI3K/AKT signaling
   pathway. In the present study, we revealed that phosphorylation of PI3K
   and AKT was significantly inhibited in KGN cells after knockdown of
   CLU. These results indicated that EVs derived from UC-MSCs could
   alleviate apoptosis and promote proliferation via CLU-mediated PI3K/AKT
   activation. In addition, an in-depth study of the mechanisms of
   CLU-mediated activation of multiple signaling pathways is important for
   enhancing the therapeutic potential of UC-MSCs-derived EVs and the
   translation capacity into clinic. These questions will be given
   priority in our future studies.

Conclusion

   In this study, we explored the appropriate timing of UC-MSCs treatment
   for POF. Our findings suggested that UC-MSCs treatment on day 6
   effectively ameliorated ovarian structure and function in POF mice. The
   EVs derived from UC-MSCs have similar effects as UC-MSCs in improving
   ovarian function in POF mice. Importantly, our preliminary study
   demonstrated for the first time that CLU from UC-MSCs-derived EVs can
   inhibit apoptosis and promote proliferation of ovarian granulosa cells
   through activating the PI3K/AKT pathway. Further studies are required
   to clarify whether CLU-overexpressed UC-MSCs can enhance the efficacy
   of ovarian function improvement in POF, which may pave the way for the
   use of gene-modified cell therapy for POF.

Supplementary Information

   [174]13287_2024_3926_MOESM1_ESM.pdf^ (744.6KB, pdf)

   Additional file 1: Fig. S1 The knockdown efficiency by qPCR assay and
   the effect of these supernatant on KGN apoptosis. (A) The mRNA levels
   of CLU, GSN, GNAI2, PRDX2 and YWHAG after transfection with siRNA in
   UC-MSCs. (B) The effect of supernatant after siRNA transfection of
   UC-MSCs on apoptosis of KGN cells detected by flow cytometry.
   **p < 0.01, ***p < 0.001.
   [175]Additional file 2^ (1.4MB, pdf)

Acknowledgements