Telocytes in the human ascending aorta: Characterization and exosome‐related KLF‐4/VEGF‐A expression

Abstract Telocytes (TCs), a novel interstitial cell entity promoting tissue regeneration, have been described in various tissues. Their role in inter‐cellular signalling and tissue remodelling has been reported in almost all human tissues. This study hypothesizes that TC also contributes to tissue remodelling and regeneration of the human thoracic aorta (HTA). The understanding of tissue homeostasis and regenerative potential of the HTA is of high clinical interest as it plays a crucial role in pathogenesis from aortic dilatation to lethal dissection. Therefore, we obtained twenty‐five aortic specimens of heart donors during transplantation. The presence of TCs was detected in different layers of aortic tissue and characterized by immunofluorescence and transmission electron microscopy. Further, we cultivated and isolated TCs in highly differentiated form identified by positive staining for CD34 and c‐kit. Aortic‐derived TC was characterized by the expression of PDGFR‐α, PDGFR‐β, CD29/integrin β‐1 and αSMA and the stem cell markers Nanog and KLF‐4. Moreover, TC exosomes were isolated and characterized for soluble angiogenic factors by Western blot. CD34+/c‐kit+ TCs shed exosomes containing the soluble factors VEGF‐A, KLF‐4 and PDGF‐A. In summary, TC occurs in the aortic wall. Correspondingly, exosomes, derived from aortic TCs, contain vasculogenesis‐relevant proteins. Understanding the regulation of TC‐mediated aortic remodelling may be a crucial step towards designing strategies to promote aortic repair and prevent adverse remodelling.


| INTRODUC TI ON
Telocytes (TCs) are a recently defined interstitial cell type, morphologically characterized by a small oval cell body and a variable number of thin, long branching processes called telopodes (Tps). 1

,2 These
Tps comprise alternating regions of podoms (including mitochondria, endoplasmic reticulum and caveolae) and thin podomeres. 3 TCs have been found in a wide range of tissues 2-8 including the heart 9-11 and the vascular system. [12][13][14][15] TCs were detected in close relation to blood vessels, nerve endings and smooth muscle cells (SMCs) 6,12,16 and participate in inter-cellular signalling, tissue remodelling, renewal and regeneration. 2,9,17 With respect to the peripheral vascular system, TCs were only described in arterioles, venoules and capillaries. 15,18 In mediumsized arteries, the number of TCs was increased in the adventitial layer after vascular injury in rats. 12 Moreover, TCs appeared to be adherent in the endothelial layer of pig coronary arteries. 14 Although Li H. et al. excluded the possibility of TCs in bigger vessels due to their reduced tolerance to turbulent blood flow, 14 they were located in the adventitial layer of the aortic arch of mice. 13 To our knowledge, TCs have not been described in the human aorta in literature up to now. A potential contribution of TC in aortic tissue homeostasis could be of great interest as regeneration of the aorta would be of high clinical relevance. Its misbalance is considered to lead to aortic aneurysms, a potentially fatal disease when resulting in dissection or rupture.
The human ascending thoracic aorta (HTA) differs embryologically and genetically from the aortic arch and plays a crucial role in the development of aortic dilatation-from aneurysm formation to lethal dissection. [19][20][21] Therefore, the presence of TCs in the HTA, with the ability for tissue remodelling and regeneration, should be of specific interest. The results could give vital information for the development of HTA pathologies.
TC is known to have tissue-specific makers. In this line, only the group of Yanyan Li identified specific molecular markers, for vascular TC, which were described in middle-sized arteries. 12 They found the mesenchymal cell marker vimentin and the hematopoietic cell marker CD34 in TC. Little more analyses were done to identify the immunophenotype of aortic TCs.
The aim of this study was to gather first insights into the existence, characteristics and distribution of TCs in the different layers of HTA by transmission electron microscopy and light microscopy.
The study was also designed to investigate the specific cell/stem cell markers of aortic TCs. We also establish the isolation of TCs and characterized TC-derived microparticles.

| Human specimen
Twenty-five human non-pathological aortic tissue samples were obtained during heart transplantation from donors' hearts.
Samples were only taken if the aorta had to be shortened to fit during implantation. After receiving the specimens, aortic tissue was washed three times with sterile PBS (PAA Laboratories, Inc., Austria) to remove blood residues. Tissue was cut into pieces and processed as following: one piece was snap-frozen and stored in liquid nitrogen, one piece was fixed in 4.5% formalin, and another one in 2.5% glutaraldehyde and the remaining part was subjected to cell isolation. This study was approved by the ethical committee of the Medical University of Vienna (EK 1280/2015). Written informed consent was obtained from all patients prior to inclusion in the study. The investigation conformed to the principles that are outlined in the Declaration of Helsinki regarding the use of human tissue.

| Isolation and fluorescence-activated cell sorting (FACS) of aortic telocytes
Aortic TCs were isolated from healthy human ascending aortic tissue obtained from healthy heart donors during heart transplantation as previously described. 11  TCs attached after plating for 45-60 min. After cultivation for 48h, TC proliferation reached the logarithmic growth phase. Following 72 h of primary culture, single adherent cells displayed typical TC morphology. Cells were maintained at 37℃ in humidified atmosphere (5% CO2 in air) until becoming semi-confluent (usually 4 days after plating) when the cells were detached using 0.25% trypsin (Sigma-Aldrich) and 2 nM EDTA (Sigma-Aldrich) and re-plated at the same density of 5 x 4 cells/cm 2 . The morphology of TCs was observed, and cells were imaged with a phase-contrast microscope (Olympus CKX41 with Olympus SC-20 camera, Olympus Life Science, Vienna, Austria).
For CD34/c-kit-specific cell sorting of isolated aortic TCs, cultured TCs were collected in FACS buffer (PBS including 0.1% FBS), and 25mM HEPES was added to the FACS buffer to prevent it from becoming basic and maintain the pH between 7.0 and 8.0, and 1mM-5mM EDTA to the buffer to prevent formation of aggregates.
Cells were stained with 1x of the antibody concentration used for immunocytochemistry, followed by appropriate secondary antibody (see Table 1). Cells were resuspended at a concentration of 2-3x10 7 cell/ml. Immediately before sorting, cells were filtered through a 70µm mesh filter to prevent clogging and collected in HG-DMEM supplemented with 30% FBS afterwards. Collected cells were divided, one part was analysed directly by Western blot and the second part was cultivated in standard culture medium. Cell sorting was performed with the BD FACSAria™III Fusion

| Isolation and cell sorting by Dynabeads of fibroblasts and vSMCs
Cell isolation and sorting of CD90 + fibroblasts and α-SMA + vSMC were done as described above. Except, after filtration the cells were seeded and cultured in in DMEM supplemented with 10% FBS, 1,5 mM HEPES, 100 IU/ml penicillin and 100 IU/ml streptomycin. After reaching cell confluence, cells were harvested and proceeded with FACS method as described above. Purity estimation was done by Western blot.

| Immunocytochemical staining and microscopy
Cells were grown on 8-chamber slides (Nunc ® Lab-Tek ® Chamber Slide™, Sigma-Aldrich) for 2 days, washed with PBS (Thermo Fisher Scientific, Massachusetts, USA) and fixed in 4% paraformaldehyde for 10 min, followed by permeabilization in 0.1% saponin and blocked  (Table 1), followed by incubation with an appropriate secondary antibody and mounted in VECTASHIELD mounting medium including 0.5 µg/ml DAPI (VECTASHIELD; VectorLabs, Burlingame, CA). Negative controls were obtained following the same protocol, but omitting the primary antibodies, and the usage of purified anti-mouse and anti-rabbit IgG (Abcam, Cambridge, UK). For confocal microscopy, we used a LSM700 Meta microscopy laser system, the appropriate filters and a ZEN 2010 microscopy system (Zeiss, Inc. Jena, Germany).

| Immunohistochemical and immunofluorescence staining
Twenty-five aortic tissue samples were fixed in 4% PBS-buffered formaldehyde at 4℃ for a minimum of 24 hours. The samples were embedded in paraffin. Sections were deparaffinized with xylene and rehydrated in a descending series of ethanol (96%, 80%, 70% and 50%).

| Transmission electron microscopy
Samples of the aortic wall of approx. 2 cm 2 were fixed immediately after surgery in 2.5% glutaraldehyde. After 6 hrs, samples were cut into smaller pieces of 1 mm 3 and washed three times in 0.

| Microvesicle and exosome isolation
Microvesicle and exosome isolation was performed as previously described with minor modifications. 22 Briefly, cells were grown in FCS-free culture medium for 24 h. The cell suspension was centrifuged at 480g at 4℃ for 5 min to remove any intact cells, followed by

| RNA isolation and real-time PCR (RT-qPCR)
RNA isolation and RT-qPCR were performed in isolated and CD34 + /

| Histological analysis of aortic tissue
The presence and distribution of TCs was assessed in HTA tissue. For the initial histological analysis of potential presence of TCs in HTA, the surgical specimens were stained by toluidine blue-and H&E as  Figure 1B and Table 3), independent of the donor's age and baseline characteristics.
We classified the three layers of aortic tissue, namely the tunica intima, tunica media and tunica adventitia, to pinpoint the differences in TC presence and features depending on their location in HTA ( Figure 1C and Table 3). Most recently, Billaud et al. described a CD34 + progenitor cell population associated with the vasa vasorum in the human adult aorta, 28 whereby a subset of these cells coexpressed the endothelial cell marker CD31 and where described for the endothelium of vasa vasorum, which should be distinct from CD34 + /CD31 − TCs. Therefore, we performed a CD34/CD31 double staining, as well as a CD31/c-kit double staining, and demonstrate a TC specificity by the lack of endothelial marker CD31 expression ( Figure 1E and 1F). However, the adventitia showed the highest density of CD34 + /CD31 − TC, compared to the medial or intimal layer ( Figure 1E).

| Ultrastructural characteristics of aortic TCs
Telocytes were studied by transmission electronic microscopy In contrast, the second population of TCs showed a very close relation to SMCs ( Figure 2C). These cells appeared to have more and shorter but wider processes than the spindle-shaped cells.
This population showed a high number of vesicles located in and around Tps (Figures 2D and 5B).

| Isolation and identification of aortic TCs
Primary TC cultures exhibit a characteristic morphology and can easily be distinguished from fibroblasts, SMC and epithelial cells by phasecontrast microscopy 11,29 ( Figure 2C). TCs were isolated from healthy HTA tissue as described previously 11

| Identification of specific aortic TC markers by immunofluorescence
We next intended to characterize the molecular markers specific for aortic TCs by immunofluorescence and mRNA expression methods. First, we sorted primary HTA cells based on their expression of CD34 and c-kit by FACS to isolate the TC population ( Figure 4A). The   Figure 4E and 4F). Human fibroblasts and vSMCs were used as control confirmed by Western blot (Figure 4C and 4F). With focus on stem cell markers involved in angiogenesis and vascular repair, we assessed the KLF-4 and Nanog expression in CD34 + /c-kit + -TCs.

| CD34 + /c-kit + TC's microvesicles and exosomes contain angiogenic factors
TCs are known for their regenerative properties. 15 Aortic cell homeostasis, including repair and regeneration, depends on angiogenic factors as VEGF. 28,30 Thus, we tested sorted CD34 + /c-kit + TCs for VEGF-A expression and confirmed very high levels ( Figure 5A). Intercellular communication by releasing extracellular vesicles has been demonstrated to be another function of TCs. 11,31 VEGF-A is a soluble factor 32 ; hence, we went on to test the hypothesis that TCs may be involved in aortic tissue homeostasis by shedding microvesicles (sMVs) and exosomes containing angiogenic factors such as VEGF-A.
TEM analysis showed TCs, Tps and their podoms containing mitochondria, rough endoplasmic reticulum and specific vesicles (Figures 2, 5B, 5C). Putative extracellular vesicles were found throughout the aortic tissue but at much higher density in the immediate vicinity of Tps ( Figure 5B and 5C). We isolated exosomes and MVs by ultra-high centrifugation ( Figure 5D) and analysed the quality and surface markers of isolated vesicles by qNano and Western blot ( Figure 5E and 5F). The identity of exosomes and sMVs was confirmed by the presence of cell surface proteins such as HSP90 and TSG101. 22 Moreover, we found the CD34 and KLF-4 in isolated exosome and microparticles, which correlates with the expression of those proteins in TCs. With respect to angiogenic markers, we detected VEGF-A and PDGF-A in the exosomal fraction. In summary, our study reveals the presence of TCs in the HTA and provides further insights towards the goal of understanding this tissue. We suggest that TCs communicate with their environment via exosomes containing angiogenic factors such as VEGF-A, PDGF-A and KLF-4.

| DISCUSS ION
Our results demonstrate the presence of TCs in the human aortic wall. To date, CD34 and vimentin were identified as markers for vascular TCs in middle-sized arteries. 12 The mesenchymal cell marker vimentin is also expressed in vSMCs and fibroblasts. Similar to TCs, endoneural fibroblasts are also CD34 positive. 15,33 Our study shows that a combination of six marker proteins, CD34, c-kit, PDGFRα/-β,

Previous studies suggested a communication between TCs and
SMCs by gap junctions or by paracrine mechanisms such as exosome release. 1,9,39 Additionally, TCs were shown to not only release exosomes and ectosomes, but also multivesicular cargos. 40 TCs were described to modulate stem cells by secretion of soluble factors. 41 Thus, TCs seem to be able to induce and influence both, healthy homeostasis and pathological changes in aortic cell/ECM composition.
For instance, in case of vascular injury or damage (eg continuous cell stress), SMCs are be modulated (or dedifferentiate) from a mature 'contractile' to a less differentiated 'synthetic' phenotype. 42,43 KLF-4 has been shown to play a crucial role in regulating SMC (de-) differentiation in aortic tissue. [44][45][46] We found that KLF-4 was highly expressed in aortic TCs. This finding provides the first indication of a possible pericyte role in angiogenic processes. The release of exosomes with the stem cell factor KLF-4, as well as the CD34 + /ckit + /vimentin + /αSMA +− TC subtypes, seems to be involved in keeping the balance of a physiological SMCs contractile phenotype. 46 c-kit was recently described to play a crucial role in TCs and their function in tissue regeneration. 2,31,47 It was suggested that the c-kitpositive TC subtype contributes to cell maintenance and homeostasis and may participate in repair processes. 48 We therefore propose a corresponding function of c-kit-positive TCs in the human aorta.
Excessive stem cell factor expression/secretion of TCs could lead to the dedifferentiated synthetic phenotype of vSMCs and pathological ECM remodelling. Moreover, VEGF-A appears to be strongly associated with aortic tissue remodelling, and later, aneurysm rupture. 49