Effect of neuron‐derived neurotrophic factor on rejuvenation of human adipose‐derived stem cells for cardiac repair after myocardial infarction

Abstract The decline of cell function caused by ageing directly impacts the therapeutic effects of autologous stem cell transplantation for heart repair. The aim of this study was to investigate whether overexpression of neuron‐derived neurotrophic factor (NDNF) can rejuvenate the adipose‐derived stem cells in the elderly and such rejuvenated stem cells can be used for cardiac repair. Human adipose‐derived stem cells (hADSCs) were obtained from donors age ranged from 17 to 92 years old. The effects of age on the biological characteristics of hADSCs and the expression of ageing‐related genes were investigated. The effects of transplantation of NDNF over‐expression stem cells on heart repair after myocardial infarction (MI) in adult mice were investigated. The proliferation, migration, adipogenic and osteogenic differentiation of hADSCs inversely correlated with age. The mRNA and protein levels of NDNF were significantly decreased in old (>60 years old) compared to young hADSCs (<40 years old). Overexpression of NDNF in old hADSCs significantly improved their proliferation and migration capacity in vitro. Transplantation of NDNF‐overexpressing old hADSCs preserved cardiac function through promoting angiogenesis on MI mice. NDNF rejuvenated the cellular function of aged hADSCs. Implantation of NDNF‐rejuvenated hADSCs improved angiogenesis and cardiac function in infarcted mouse hearts.


| INTRODUC TI ON
Global coronary heart disease, especially acute myocardial infarction (MI), is associated with high morbidity and mortality, and the age of first MI is decreasing. 1 The cost of healthcare continues to rise as the size of the aged population increases, contributing to major social and economic problems worldwide. 2,3 Advancements in clinical treatments including thrombolytic drugs, coronary artery bypass grafts and interventional treatments have greatly improved the prognosis of MI. However, progression to congestive heart failure remains a significant global health crisis, especially in aged populations. Since current therapies focus on the restoration of blood flow to the defect area, the damaged cardiomyocytes are unable to regenerate and are replaced by fibroblasts and collagen matrix (fibrotic scar tissue). This scar tissue becomes thin and inelastic, leading to ventricular chamber dilation and progression to congestive heart failure. 4 In recent years, with the rapid development of tissue engineering and regenerative medicine, stem cell therapy for MI has provided new hope and ideas for cardiac repair. Autologous stem cell transplantation has greatly reduced the problem of immune rejection and has become one of the most promising treatment methods at present. [5][6][7] Currently, stem cells used for tissue repair and regeneration include embryonic stem cells, hematopoietic stem cells, bone marrow mesenchymal stem cells and adipose-derived stem cells (ADSCs). All these cells have entered the clinical trial stage in the fields of soft tissue defects, ischaemic heart disease, diabetes and neural degenerative diseases. 8,9 ADSCs have become more popular cells in tissue repair because of their high abundance and relative ease of extraction. They also induce a low immune response in the host and have multi-directional differentiation and self-renewal abilities. [10][11][12] In the current study, we focus on ADSC transplantation for the treatment of acute MI and investigated its efficacy for cardiac repair and regeneration.
Clinical studies found that the efficacy of stem cell therapy in elderly patients is significantly reduced. 13 Although there are multiple contributing factors, age is the most important. With ageing, the number and quality of stem cells significantly decrease, and their ability to repair and regenerate is compromised. This decline in stem cell function results in the unsatisfactory clinical effect of autologous stem cell transplantation in aged patients. 14 Neuron-derived neurotrophic factor (NDNF, also named C4orf31) is identified as a secretory protein which promotes the growth and migration of neurons, the formation of extracellular matrix and inhibits apoptosis. 15,16 Previously, our group found that NDNF protein rejuvenated aged human multipotent mesenchymal stromal cells (hMSCs) and transplantation of the NDNF-rejuvenated hMSCs improved cardiac function through inhibiting apoptosis and promoting angiogenesis in infarcted mouse hearts. 17 However, the effect of age on the expression level of NDNF in human adipose-derived stem cells (hADSCs) remains unknown and studies on the treatment of MI with NDNF overexpression by hADSCs have not been reported. We have been suggested that NDNF can be selected as a target to rejuvenate aged hADSCs through genetic modification, thereby improving the ability of aged stem cells to facilitate angiogenesis and tissue repair and ultimately lead to the preservation in cardiac function.
We used a lentiviral vector to overexpress NDNF in old hADSCs and the biological function of proliferation and migration of hADSCs was investigated in vitro. The genetically modified aged stem cells were then implanted into heart after MI using an adult mouse model and the effects of NDNF overexpression hADSCs on heart repair were investigated in vivo.

| Human adipose tissue collection, cultivation and identification
Adipose tissue was obtained from hospitalized patients (17-

| Overexpression of NDNF in old hADSCs by gene modification
Cell transduction was carried out using a lentiviral expression vector carrying the NDNF gene (Lenti-Puro-EF1α-NDNF-Homo-IRES-eGFP, Cyagen Biosciences Inc, Santa Clara, CA) according to the manufacturer's instructions. Empty virus (Old) and NDNF (Old + NDNF) were transduced into old hADSCs by lentiviral vector (n = 6, age 72.5 ± 10.52 years). The expression differences of mRNA and protein levels of NDNF after transduction were detected by RT-PCR and Western blotting as described in supplemental methods.
The effect of overexpression of NDNF on cell proliferation and migration was observed by BrdU (5-bromo-2'-deoxyuridine, Sigma, cat no. A2385) pulse chasing and the wound-healing cell migration assay described in supplemental methods.

| Cardiac function measurement
Echocardiograph was used to dynamically record the changes in cardiac function of mice. The left ventricular internal dimension in systole (LVIDs), left ventricular internal dimension-diastole (LVIDd), ejection fraction (EF%) and left ventricular fractional shortening (FS%) of mice were measured before and 7, 14, 21 and 28 days after MI. Twenty-eight days after MI and cell transplantation, the mouse hearts were dissociated from surrounding tissue. After fixation with 10% formalin for 48 hours and dehydration with 75% ethanol, the hearts were cut along the horizontal axis into continuous 1 mm sections, photographed for morphometry. ImageJ software was used to measure the size of the infarcted area (the ratio between the length of the infarct area and the circumference of the entire left ventricle) and the thickness of the infarct respectively.

| Masson's trichrome staining
The heart segments were embedded in paraffin and made into 5-μm thick sections. The level of cardiac fibrosis was detected by Masson's trichrome staining. Briefly, after gradient dewaxing and fixation, the paraffin sections were successively immerged into different colours of dyes. The red coloured area represented viable cardiac tissue and the blue coloured area represented collagen fibres, which was performed to confirm scar tissue in the left ventricular free wall.

| Immunofluorescent staining
Von Willebrand factor (vWF) and α-smooth muscle actin (α-SMA) were detected by immunofluorescence staining to compare differences in the number of new capillaries and small arterioles. After the gradient dewaxing of paraffin sections, slides were incubated in Tris-EDTA by microwave heating for 10 minutes, and subsequently blocked in BSA for 1 hour. Primary antibodies (rabbit anti-vWF, Proteintech, cat no. 11778-1-AP at 1:100 dilution, mouse anti-α-SMA, Sigma, cat no. A5228 at 1:200 dilution) were incubated at 4°C overnight. The next day, incubation with Alexa568 secondary antibody (goat anti-rabbit, Invitrogen, cat no. A11011) and Alexa488 secondary antibody (rabbit antimouse, Invitrogen, cat no. A11029) at 1:2000 dilution was carried out for 1 hour at room temperature.
The nuclei were counter stained with DAPI (Sigma, cat no. D9542) for 10 min. The fluorescent-positive area in three randomly selected high-power fields per slide was photographed using a Nikon fluorescence inverted microscope. ImageJ software was used to calculate the percentage of fluorescent positive areas.

| Data analysis
All data are presented as mean ± SD. Statistical analyses were performed with GraphPad Prism software (v.7.0). Student's t test was F I G U R E 1 Cultivation and identification of hADSCs. A, General morphological observation of human adipose-derived stem cells (hADSCs) cultured for 7 days after isolation from donors (17-92 years old). B, Rate of cell growth was significantly higher in young than old hADSCs, n = 6/group, **P < 0.01 Young vs Old. C, Representative histogram plots of cell surface markers from young and old hADSCs. D, Comparison of % positive cells of hADSCs from young and old donors. n = 3/group F I G U R E 2 The proliferation, migration and differentiation of hADSCs decreased with age. A, Representative micrographs of immunofluorescent staining for 5-bromo-2'-deoxyuridine (BrdU, red) with nuclei stained blue with DAPI. The percentage of BrdU + cells decreased with age. B, Representative images showed the cell migration of human adipose-derived stem cells (hADSCs) using the woundhealing cell migration assay. Migration rate decreased with age. C, Representative micrographs of adipogenic differentiation stained for oil red 'O'. Mature adipose cells were stained in red in both old and young groups. The adipogenic differentiation rate decreased with age. D, Representative micrographs of osteogenic differentiation stained for alizarin red. The osteocytes were stained in red in both old and young groups. The osteogenic differentiation rate decreased with age used for comparisons of means between two groups. Comparisons of parameters among three or more groups were analysed using oneway ANOVA followed by Tukey or two-way ANOVA with repeated measures over time, followed by Bonferroni post-hoc tests. The list of X (age) and Y (proliferation/migration/ differentiation level/mRNA levels of the 7 factors) was analysed using a correlation analysis.
Differences were considered statistically significant at P < 0.05.

| HADSCs were successfully isolated, cultured and identified
Human adipose-derived stem cells were successfully isolated and the cell morphology was identified under a microscope. HADSCs in F I G U R E 3 NDNF expression decreased with age. A, The cell regeneration-related genes, Sirt1 (sirtuin), Sirt2, Sirt6, Bmi1 (Polycomb complex protein BMI-1), and Cbx8 (chromobox homolog 8), in human adipose-derived stem cells (hADSCs) negatively correlated with age. B, The cell senescence-related gene P16 (cyclin-dependent kinase inhibitor 2A) increased with age. C, Neuron-derived neurotrophic factor (NDNF) mRNA expression showed inverse correlation with age. The mRNA (D, n = 5/group) and protein (E, n = 3/group) levels of NDNF in the elderly group was significantly lower than that of the youth group. *P < 0.05, **P < 0.01 F I G U R E 4 NDNF transduction rejuvenated old hADSCs by increasing proliferation and migration. A, Old human adipose-derived stem cells (hADSCs) were transduced with a lentiviral vector overexpressing neuron-derived neurotrophic factor (NDNF) which was also tagged with green fluorescent protein (GFP). Representative micrographs showed the GFP + cells to indicate the transduction efficiency. There was no difference in transduction efficiency between old hADSCs transduced with NDNF (Old + NDNF) or with empty viruses (Old), n = 6/ group. B, NDNF mRNA expression evaluated by RT-PCR was significantly higher in Old + NDNF compared to Old group, n = 3/group. C, NDNF protein expression evaluated by Western blotting was significantly higher in Old + NDNF compared with that of the Old group, n = 3/group. D, Representative micrographs of immunofluorescent staining for 5-bromo-2'-deoxyuridine (BrdU in red) with nuclei stained blue with DAPI. The percentage of BrdU + cells was significantly higher in Old + NDNF compared with that of the Old group, n = 6/group. E, Representative images showed cell migration of hADSCs using the wound-healing cell migration assay. Migration rate was significantly higher in Old + NDNF compared with empty vector-transduced hADSCs (Old), n = 6/group. *P < 0.05, **P < 0.01 the young group were long fusiform, growing radially or vertically around each centre, with uniform shape, size and an orderly arrangement. However, hADSCs in the old group were spindle or polygonal in shape and arranged in a non-uniform way ( Figure 1A). Next, we seeded the cells at the same number and counted at day 2, 4 and 6 after cell plating to determine the growth curve. The growth of young hADSCs was significantly higher than those of the old hAD-SCs starting from day 2 and sustained up to day 6 after cell seeding ( Figure 1B). The hADSCs from both young and old groups were more than 95% positive for the ADSC surface markers (CD90, CD44, CD73 and CD105) as identified by flow cytometry (Figure 1C and   1D). On the other hand, the hADSCs from both groups were negative for the haematopoietic lineage cell markers of CD45, CD11b, CD19 and HLA-DR (Data not shown). These results conform to the criteria of ADSC identification.

| Proliferation, migration, adipogenic and osteogenic differentiation of hADSCs decreased with age
To further confirm the difference in cell growth capacity between the young and old hADSCs, we performed BrdU pulse chasing to label actively proliferating cells. Consistent with the result from cell growth, there was an inverse correlation between the number of BrdU + cell and age showing decreasing number of BrdU + cells with increasing age (Figure 2A). On testing the cell migratory ability by wound-healing cell migration assay, the result revealed that the migration ability of hADSCs decreased with age ( Figure 2B).
Adipogenic and osteogenic differentiation can be induced in hADSCs, and a large number of cells were stained by oil red O and alizarin red, which confirmed that this cell had adipogenic and osteogenic differentiation abilities. However, the adipogenic ( Figure 2C) and osteogenic ( Figure 2D) differentiation abilities gradually declined with age.

| The mRNA and protein levels of NDNF decreased significantly with age
In an attempt to identify the possible factor (factors) responsible for the age-related changes in cellular function, we performed an array of RT-PCR to examine the regeneration-and senescence-related genes. There was a significant negative correlation with age among regeneration-related genes Sirt1 (sirtuin), Sirt2, Sirt6, Cbx8 (chromobox homolog 8) and Bmi1 (Polycomb complex protein BMI-1, Figure 3A) whereas the cell senescence-related gene P16 (cyclin-dependent kinase inhibitor 2A) increased with age ( Figure 3B). Among these genes, the NDNF mRNA showed a strong inverse correlation with age ( Figure 3C). This was further confirmed by the result showing that the mRNA ( Figure 3D) and protein ( Figure 3E) levels of NDNF in the elderly group were significantly lower than that of the youth group. These findings implied that NDNF may be one of the key factors involved in age-related changes in hADSC function.

| Overexpression of NDNF in old hADSCs significantly enhanced cell proliferation and migration
To investigate if restoring NDNF level can restore the proliferative and migratory capabilities in old hADSCs, a lentiviral expression vector carrying the NDNF gene (also tagged with green fluorescent protein, GFP) was used to transduce old hADSCs (Old + NDNF). Old hADSCs transduced with empty vector without the NDNF gene but with GFP served as control (Old). Transduction efficiency when quantified by GFP + cells was comparable between the two groups ( Figure 4A). The expression of NDNF mRNA ( Figure 4B) and protein ( Figure 4C) was significantly greater in NDNF-transduced than in empty vector-transduced old hADSCs. As expected, BrdU pulse chasing revealed that the number of BrdU + cells in Old + NDNF were significantly higher than that of the old group ( Figure 4D). The cell migration test showed that the migration rate in Old + NDNF was significantly greater compared with that of the old group ( Figure 4E).
These data confirmed our hypothesis that restoration of NDNF improved proliferation and migration of old hADSCs in vitro.

| NDNF promoted the repair of cardiac injury in vivo
To further examine the reparative capacity of NDNF-overexpressed  Morphological analysis ( Figure 6A) and Masson's trichrome staining of hearts ( Figure 6B) indicated that the scar size in the Old + NDNF was significantly smaller than that in the medium and the old groups with the young group had the smallest scar size at 28 days post-MI ( Figure 6C). On the other hand, scar thickness was significantly greater in the Old + NDNF than that in the medium and the old groups with the young group had the greatest scar thickness ( Figure 6D). Next, to confirm that implantation of NDNF-overexpressing hADSCs-restored NDNF level in the infarcted hearts in vivo, the protein level of NDNF was examined in the infarct and border areas from the four experimental groups ( Figure 7A). The protein expression of NDNF as measured by Western blotting was significantly higher in the infarcted area of the Old + NDNF than in the medium control and the old groups ( Figure 7B). The protein expression of NDNF in the Old + NDNF was comparable to that of the young group, indicating implantation of NDNF-overexpressing hADSCs effectively restored NDNF levels in the infarcted hearts ( Figure 7B).
To further understand the underlying mechanism responsible for the functional changes, immunofluorescent staining was performed to determine blood vessel density (stained with vWF) and arteriole density (stained with α-SMA). The results showed that blood vessel density ( Figure 7C and D) and arteriole density ( Figure 7E and 7F) of the young and the Old + NDNF groups were significantly greater than those of the medium control group and the old group with empty virus transduction. Collectively, these data suggested that implantation of NDNF-overexpressing old hADSCs promoted cardiac repair and delayed progressive heart failure possibly through restoration of NDNF level and improving angiogenesis.

| D ISCUSS I ON
In the current study, hADSCs were successfully isolated from young  I G U R E 6 In vivo implantation of NDNF-overexpressing old hADSCs decreased scar size and increased scar thickness in infarcted mouse hearts. A, Representative heart sections 28 days after myocardial infarction (MI) in mice that received implantation of control medium (Medium), empty vectortransduced old human adipose-derived stem cells (hADSCs, Old), neuron-derived neurotrophic factor (NDNF)-transduced old hADSCs (Old + NDNF) and untransduced young hADSCs (Young). B, Representative heart slides stained with Masson's trichrome and planimetry-based quantification revealed that the scar size area was larger in the Medium and Old groups than in the Old + NDNF and Young groups at 28 days after MI (C). D, The scar thickness was greater in the Old + NDNF and Young groups than in the Medium and Old groups at 28 days after MI, n = 6/ group are the most promising and potentially valuable source because of their large reserve, ease of access, high abundance, low immunogenicity, multi-directional differentiation and strong self-renewal ability. The International Federative Committee on Adipose Science points out that ADSCs are one of the ideal biological cells. 18 ADSCs were first isolated and extracted from adipose tissue by Zuk et al 12 These cells have the potential for multi-directional differentiation, and the expression of surface markers conform to the criteria of stem cell identification. 19,20 In this study, adipose tissue was col- In vivo implantation of NDNF-overexpressing old hADSCs increased angiogenesis and arteriole genesis. A, Western blotting analysis of the expression of NDNF protein in the infarcted area of mouse hearts that received implantation of control medium (Medium), empty vector-transduced old human adipose-derived stem cells (hADSCs, Old), neuron-derived neurotrophic factor (NDNF)transduced old hADSCs (Old + NDNF) and untransduced young hADSCs (Young). B, NDNF protein levels were significantly higher in the Old + NDNF and Young groups compared with the Old and Medium control groups, n = 3/group. C, Representative micrographs of immunofluorescent staining for von Willebrand factor (vWF, red) with nuclei stained blue with DAPI. D, Capillary density was significantly higher in the Old + NDNF and Young groups compared with the Old and Medium control groups, n = 6/group. E, Representative micrographs of immunofluorescent staining for α-smooth muscle actin (α-SMA, green) with nuclei stained blue with DAPI. F, Arteriole density was significantly higher in the Old + NDNF and Young groups compared with the Old and Medium control groups, n = 6/group the abilities of adipogenic and osteogenic differentiation decreased significantly with age, which was consistent with the result from Ding et al, 25 indicating that age is an important factor affecting the biological characteristics of stem cells. Although the separation and extraction of ADSCs are influenced by a number of factors such as donor sex, body mass index, source location, disease comorbidity, cell passage, cell cryopreservation and resuscitation, all of which may impact the experimental results, [26][27][28][29] age is still one of the key factors affecting the biological function of stem cells.
The mechanisms by which ADSCs treat MI are described to include possible cell differentiation into myocardial cells to repair the damaged heart tissue, but mainly act through paracrine effects. The neurotrophic factor NDNF is a secretory protein found in the nervous system, which is suggested to be involved in regulating nerve development, migration and differentiation, promoting hippocampal neuron migration and axon growth, and increasing neuronal survival. 15,16 Previous studies have shown that NDNF promotes endothelial angiogenesis and capillary regeneration, improves myocardial remodelling and function in mouse ischaemic hind limbs, strengthens the development of collateral circulation vessels, and has a beneficial effect on various ischaemic cardiovascular diseases. 34,35 This is consistent with the mechanism of action of ADSC transplantation for the treatment of MI. Previous research from our group has confirmed that the expression of NDNF in hMSCs was significantly reduced with age. Overexpression of NDNF in old hMSCs inhibited cell apoptosis, promoted angiogenesis and improved cardiac function. 17 The present study found that NDNF expression in hADSCs was also negatively correlated with age. Furthermore, restoration of NDNF in old hADSCs improved biological function and rejuvenated the aged hADSCs. More abundant and better functional autologous hADSCs were obtained by amplification in vitro. After in vivo implantation of NDNF-overexpressing old hADSCs, angiogenesis was facilitated which improved collateral circulation in the infarcted area and eventually led to the preservation in cardiac function and delaying the progression of heart failure.

| CON CLUS ION
The proliferation, migration and differentiation abilities of hADSCs decreased significantly with age. The level of NDNF in aged hAD-SCs was significantly lower than that in the young group, suggesting that NDNF may be used as a target for stem cell rejuvenation.
Overexpression of NDNF in aged hADSCs significantly enhanced proliferation and migration in vitro as well as promoted local angiogenesis, repair and improved cardiac function after implantation into the infarcted mouse hearts in vivo. These findings provide experimental support for the clinical application of hADSCs in elderly patients with ischaemic heart disease.

ACK N OWLED G EM ENTS
The

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

AUTH O R CO NTR I B UTI O N
KY contributed to the collection of data, data analysis and interpretation, article writing; H-FS, SH, W-JY, X-MF, FR, HG, X-YZ, JZ, Z-XP, and G-XX contributed to the collection of data, data analysis and interpretation; JX contributed to financial support, conception and design and manuscript writing; R-KL contributed to the administrative support, final approval of the manuscript.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available on request from the corresponding author.