Ex vivo expansion of cord blood-derived endothelial cells using a novel xeno-free culture media

Aim: Endothelial cells (ECs), isolated from peripheral blood (PB), bone marrow (BM) and cord blood (CB), are limited in numbers and expansion has had limited success. We used a novel serum-free medium (EndoGo) to evaluate effects on ex vivo expansion of CB-derived ECs. Materials & methods: Flow cytometry and matrigel were used to determine expansion of ECs and for determination of the EC progenitor cell. Results: EndoGo™-containing cultures demonstrated superior expansion and stimulated proliferation of two distinct subpopulations, CD34+CD31+ and CD34-CD31+, which exhibited different morphology, phenotype and function. EndoGo also expanded the CB endothelial progenitor cells from freshly isolated CB. Conclusion: These findings demonstrate the potential of EndoGo to expand CB ECs, which could generate increased numbers of ECs for therapeutic applications.


EndoGo
EndoGo XF was supplied by Biological Industries (CT, USA). Supplement Mix was added to EndoGo medium. Full EndoGo medium was then supplemented with 10% Human Platelet Lysate (HPL; Mill Creek Life Sciences, MN, USA) and 1% Penicillin-Streptomycin-Glutamine (GPS, Gibco). ECFCs/ECs cultured in EndoGo were removed using recombinant Trypsin Solution which was inhibited using Trypsin inhibitor (both from Biological Industries) according to manufacturer's recommendations. Due to the clotting reactivity when combining serum and EndoGo-HPL with trypsin, all cells cultured with EndoGo were washed and assayed with phosphate-buffered saline (PBS, Gibco) + 2% HPL.
Growth time course CB culture derived ECFCs/ECs were stained with CD34 and CD31 antibodies and sorted using a MoFlo Astrios (Beckman Coulter). Whole CB ECs and each sorted cell subpopulation, CD31 + CD34 + cells and CD31 + CD34cells were seeded in duplicate at 2 × 10 4 in ECM or EndoGo. Cells were harvested with trypsin-EDTA for cells in ECM or recombinant trypsin with trypsin inhibitor for cells in EndoGo and counted on day 7. Harvested cells were stained with CD34, CD45 and CD31 for flow cytometric analysis.
Cytospin/growth slide CB ECFCs/ECs were affixed either by direct cytospin (1600 rpm for 5 min at room temperature) or cultured in ECM for 24 hr in a 37 • C, 5% CO 2 incubator and subsequently fixed and stained. Slides were fixed and stained with Hema3 (Fisher Scientific, NH, USA). Images were obtained using Olympus DP-10 camera (Olympus Imaging, PA, USA).

Matrigel assay
The matrigel assay was performed using Matrigel (BD Biosciences) following manufacturer's recommendations. Briefly, 200 μl Matrigel solution was added to the well and incubated at 37 • C for 30 min before plating cells. Cells were plated at 30,000-40,000/well of a 96-well flat bottom plate at 100 μl volume. ECs or admixed ECs were prepared together before plating onto Matrigel.

Statistics
Data were analyzed for statistical significance using the Student's t-test. Data and statistics were performed using GraphPad Prism 6 software (GraphPad, CA, USA).

Growth of ECs in vitro Ex vivo expansion of CB ECFCs/ECs
Development of clinical EC cellular therapy protocols has been limited largely due to the lack of efficient ex vivo EC expansion with published reports either expressing sufficient clinical levels of ECs via theoretical calculations or lacking complete phenotypic and functional study of the end populations. Numerous groups have attempted and reported EC expansion in vitro using different culture media; however, total EC and EPC proliferation have not been properly studied and translation into clinical protocols has not been demonstrated [8,13,21,27,30]. Since CB has become more readily available, we used this blood source and isolated CD34 + cells and cultured in a standard EC medium (ECM), which is similar in composition to other commercially available EC mediums. From initial seeding, CB CD34 + cells required 2-3 weeks before adherent cell populations were readily visible (data not shown). To determine the effectiveness of ECM in promoting CB EC expansion, we seeded 200,000 cells and cultured for 7 days in ECM ( Figure 1A). ECM-induced fivefold expansion of CB ECFCs/ECs (CD45 -CD31 + ) (1.06 ± 0.22 × 10 6 ). Since our goal was to develop or optimize ex vivo expansion protocols of ECs, we also tested various combinations of VEGF, EGF, bFGF1 and SCF to ECM but all were suboptimal unless combined in total medium (ECM; data not shown). The addition of SCF did not promote further total cell expansion yet preserved the CD31 + CD34 + population which suggests a positive role of SCF on EC progenitors (data not shown). Overall, the lack of significant cell expansion with ECM is suboptimal in promoting EPC expansion and thus prevents subsequent proliferation of the total EC population and potential clinical translation.

Phenotype & morphology of ex vivo expanded CB ECs
Since ECs are phenotypically defined from blood sources using CD45 and CD31 expression and ECs were established from CD34 selected CB, we used this panel to phenotype the ex vivo expanded progeny. Figure 1B shows the representative flow phenotype of CB ECFCs/ECs expanded with ECM and EndoGo. Cells were first gated on CD45, of which CB ECs were contained in the CD45 negative fraction (99-100%). Although future science group www.future-science.com CD31 + CD34 -ECs were significantly expanded with EndoGo (ECM-0.96 ± 0.2 x 10 6 , EndoGo-2.4 ± 0.29 x 10 6 ; Figure 1C), the expansion of the CD31 + CD34 + population was most pronounced in EndoGo when compared with ECM (0.25 ± 0.1 x 10 6 and 6004 ± 2915, respectively) ( Figure 1D). This suggests that EndoGo might interact specifically with the progenitor cell population (CBMNC CD45 -CD34 + ) to induce both proliferation (of total CD31 + ECs) and self-renewal of EPCs (of total CD34 + cells). Of note was the increase in CD31 + CD34 + ECs from the addition of 10% HPL to α-MEM medium with EGF, bFGF and VEGF compared with ECM, suggesting some effect on the EPC though not sufficient to expand the population similar to EndoGo (α-MEM 10%-HPL: 46000 ± 4000 and ECM: 6004 ± 2915). The morphology of in vitro CB ECFCs/ECs cultured with different media also varied ( Figure 1E). Although ECM promoted more fibroblastic, tube-like cells, the addition of serum to EndoGo produced smaller, circular cells that spread out in culture. EndoGo promoted fibroblast-like cells that formed tube-like structures during in vitro culture. It is unclear if growth of ECs in different in vitro cultures promotes expansion of subtypes of ECs that we did not phenotype or if media is altering the attachment of ECs to the culture flask. Overall, this suggests that EndoGo, including platelet lysate and supplement, is superior for optimal expansion of CB ECs.

Phenotypic & functional characteristics of CB ECs expanded in EndoGo Flow characterization of EndoGo expanded CB ECFCs/ECs
Since EndoGo specifically expanded the CB EC CD34 + population compared with ECM, we examined the CD34 + and CD34subpopulations for phenotypic, morphologic or functional differences. In Figure 2A, we used flow cytometry to phenotype for CD117 (KIT), CD133 (PROM1), CD184 (CXCR4), CD144 (CDH5), CD146 (MCAM) and CD42a (GPIX), all of which are markers used to identify ECs obtained from various blood sources. Both the CD34 + and CD34 -EC subpopulations were CD144 + , CD146 + and CD42a -. However, CD133 and CD117 expression was detected on CD34 + ECs while not present on CD34 -ECs consistent with a progenitor-like population. Additionally, CD34 + ECs also expressed elevated levels of CD184, typically found on homing and niche-attached populations.

Morphology, growth characteristics & function of EndoGo expanded CB EC subpopulations
We next performed a cytospin analysis on ECFC/EC subpopulations to examine for morphological differences between the CD34 + and CD34 -CB EC subpopulations cultured in EndoGo. From Figure 2B (left column), the CD34 -EC subpopulation had a slightly larger cytoplasm to nucleus ratio (2:1), suggestive of a more mature subpopulation. The CD34 + EC subpopulation was similar in ratio (1:1). Figure 2B (right column) demonstrates the morphology of each population during in vitro growth and attachment (24 h). The CD34 -EC subpopulation demonstrated more fibroblast-like characteristics with cellular elongation and branching while the CD34 + EC subpopulation did not branch. Together, the smaller sized, larger nucleus:cytoplasm ratio and nonbranching CD34 + EC subpopulation is suggestive of a less mature, more progenitor-like population.
We next tested for functional differences between the CD34 + and CD34 -ECFC/EC subpopulations via tube formation in Matrigel ( Figure 2D). Although robust tube formation occurred with the unsorted ECs, as expected, tube formation only occurred with the CD34subpopulation and was absent with CD34 + ECs ( Figure 2D, top row). Most unique was the lack of Matrigel loop formation with increasing numbers of CD34 + cells. Upon addition of increasing ratios of CD34 + cells into CD34 -Matrigel cultures, tube formation was inhibited ( Figure 2D, bottom row). Quantitatively, CD34matrigel assays contained more total loops per well, 63 ± 1, than CD34 + matrigel assays (0) with decreasing loop formation as more CD34 + cells were added to assay ( Figure 2F). It is unclear why unsorted CB ECs formed more tubes in assay than the CD34population (126 ± 2), though sorting or selection may interfere initially with optimal tube formation.
EC sprouting begins the process of angiogenesis through growth factor signaling and cell activation. Upon closer examination of each subpopulation, CD34 + ECs produced enhanced sprouting while CD34 -ECs did not form sprouts. Together, this suggests that while tube formation is promoted by CD34 -ECs, initiation of angiogenesis is promoted by CD34 + cells. Enhanced microscopic inspection revealed significant sprouting of the individual cells in the CD34 + Matrigel culture while not present in the CD34 -Matrigel culture ( Figure 2F).

Identification of the CB EPC
Although EndoGo significantly expanded CB ECFCs/ECs compared with a standard EC culture medium in ex vivo culturing, our results specifically focused on expansion of the CD34 + EC subpopulation.
To further confirm the effectiveness of CB EPC expansion by EndoGo, we used the CD45 -CD34 + EC lines generated from multiple CB sources to expand cells and phenotype for the EPC ( Figure 3F). Importantly, EndoGo uniquely expanded the CB EPC compared with ECM after only seven days in culture (Input: 11 ± 0; ECM: 59 ± 12 [5-fold]; EndoGo: 603 ± 215 [54-fold]). This further demonstrates the effective expansion properties of EndoGo on CB ECs in vitro and more specifically in targeting expansion of the CB EPC.

Discussion
Ex vivo expansion of ECs could greatly benefit cell therapeutic treatment of many different diseases, including ischemia, stroke and diabetes. Currently, an expansion culture has not been developed that promotes efficient expansion of CB ECs or even CB EPCs at clinically relevant numbers. Here, we demonstrate significant expansion of CB ECFCs/ECs and EPCs using EndoGo medium, with specific addition of HPL, a fact recently established in long-term cultures of PB ECs [30]. Various reports have suggested clinical-scale expansion of ECs in respective cultures but these efforts were based solely on theoretical calculations and not clinical scale experimentation [22,23]. Numerous studies have developed ex vivo cultures to expand EPCs (reviewed in [31]). Senegaglia et al. demonstrated 70-fold expansion of CB EPCs (6.23 x 10 6 total) generated from CD133+ CB cells over 60 days [32]. Lippross et al. reported a 117.6-fold increase in EPCs generated from BM CD34 + (2.94 x 10 6 total) or 147.9-fold increase in EPCs from BM CD133+ (3.7 x 10 6 total) over 30 days [33]. Masuda et al. induced a 52.9-fold increase (0.5 x 10 6 total) in biologically functional EPCs (based on colony formation) over 7 days [34]. In long-term cultures (28 days, Day 0; EC: 25000, EC 34+: 1560) of CB EC cell lines with weekly passages, EndoGo could induce 679-fold expansion of total ECs (17 x 10 6 total), while ECM promoted only 136-fold expansion (3.4 x 10 6 total) (data not shown). Importantly, expansion of the CD34 + fraction in 28-day cultures with EndoGo compared with ECM was significantly higher (EndoGo: 2478-fold [3.9 x 10 6 total]; ECM: 88-fold [0.13 x 10 6 total]). This is significant considering reports citing EPC expansion would demonstrate a loss of CD34 or CD133 expression through long-term cultures [23,32,34]. Future work is being explored to attempt actual clinical scale expansion of CB ECs with EndoGo (data not shown).
EndoGo induced expansion of two separate populations within CB ECs that phenotypically, morphologically and functionally were different, CD34 + CD31 + and CD34 + CD31 -. One recent study has demonstrated induction of CD34 + expression on CD34 -PB ECs in confluent cultures [35]. We seeded CD31 + CD34 -CB ECs at higher concentrations and observed some re-expression of CD34 on the CD34cells, specifically when cultures became confluent (data not shown). Conversely in separate confluent cultures, media changes performed every other day inhibited CB EC growth (data not shown). This is suggestive of a more complicated autocrine and paracrine signaling of subpopulations within CB EC ex vivo cultures that needs to be discovered.
Our studies identified the CD34 + fraction as a more progenitor-like population: flow phenotype of CD117, CD133, CD34 and CXCR4; morphology with higher nuclear:cytoplasmic ratio and ability to derive CB ECs from CD34 + fraction. Earlier studies have also reported expression of CD117 and CD184 (CXCR4) [16] on EPCs and their importance for EC homing and initiation of tissue repair [36,37]. Clinical trials targeting CXCR4 have been used to mobilize EPCs for tissue repair during myocardial infarct and ischemia [38,39]. Various studies have used these markers to phenotype stem and progenitor cell populations of HSCs [40] and cancer stem cells [41]. This is supportive of the notion of EndoGo expansion of the CB EC progenitor cell for use in cell therapy protocols.
The inability of the cultured CD34+ CB ECs to form tubes in the Matrigel assay while displaying increased sprouting depicts a population that may home to injured sites and initiate angiogenesis. We attempted long-term culturing of CD34 + cells on Matrigel but cell viability decreased and cells never formed tubes (data not shown). Although these results are similar to Ferreras et al. [42], an opposite report by Tasev et al. demonstrated that the CD34 + EC fraction could indeed form tubes within a matrix assay [35]. This discrepancy may be due to the different EC culture conditions since both groups isolated ECs from PB samples. However, the use of different sources cannot be ruled out as phenotypes could vary between products, which would require further testing.
The exact phenotype of the CB EPC is currently unknown, although many reports have successfully identified unique markers [12,43]. Conflicting reports demonstrating EPC isolation from both hematopoietic (CD45 + and nonhematopoietic (CD45 -) sources as well as opposing markers (CD34 -) cause inconsistencies in EC biology discussions [12,25,44,45]. Expression of monocyte-macrophage markers on in vitro ECs have been identified further suggesting a hematopoietic link to the EPCs [45,46]. We did not detect CD14 expression on in vitro cultured CB ECs throughout culturing and efforts to isolate ECs from the CD45 + CB MNC population were also not successful (data not shown). Other groups have attempted to define EPCs from PB and BM using CD144, CD42a and CD146, although methods of isolation and culturing were varied [47,48]. Variances in isolation techniques and culture regimens could explain these inconsistencies; more specifically, these markers were not applied in a specific panel to isolate and promote EC growth. Attempts to provide a consensus nomenclature have demonstrated a largely unknown phenotype to clarify or determine specific subsets of ECs and for specific lineage tracing [49]. We explicitly utilized multimarker cell sorting to isolate various subpopulations within CB and seeded these subpopulations into identical cultures to uniquely identify the EPC phenotype and avoid confusion. It is also interesting to note that the CB EPC we identify, derived from the CD45 -CD34 + CD31 + CD144 + CD42apopulation, exhibited higher differential expression of CD34 + and CD31 + . This demonstrates the fluid nature of cell surface markers and requires caution when using markers as absolutes when attempting to identify cell populations through flow cytometry. Although the detection of CD133 on this population validates other studies that isolated ECs using this marker [14,16,18,44], expression of CD133 is not unique to EPCs as expression is also observed on HSCs [18,40]. Furthermore, though CD133 expression was only observed in the CD42afraction, some expression was detected on the CD144population establishing heterogeneity of CD133 on EC subpopulations (data not shown).
To our knowledge, this is the first report demonstrating multiple markers on a single population where CB ECs can be derived. Our CB EPC phenotype represents 0.003-0.004% of total CB MNCs which makes detection and thus assaying difficult. However, upon early proliferation of CB-derived ECs, the effectiveness of EndoGo expansion compared with ECM on the CB EPC enables in-depth study of this population (Figure 3). Future work needs to determine the effectiveness of EndoGo compared with other commercially available EC mediums, specifically regarding the ability to expand large, clinical-scale numbers of total ECFC/ECs, EC progenitor cells and possibly heterogeneous subpopulations that might impact effectiveness of specific targeted cell therapy. Specific expansion of the CB EPC we identified is critical given the heterogeneic phenotype demonstrated in various EC expansion studies. Further, we established expansion of EPCs using the phenotype at the end of culture, which is lacking in many publications suggesting EPC expansion. We discovered that EndoGo could specifically expand the unique phenotype of the CB EPC that we used to freshly isolate from CB, maintaining this population 2-3 weeks ex vivo.
Although mobilization studies of ECs have been reported for treatment of ischemia and stroke, success is limited due to the age and disease status of patients along with potential for further damage to patient tissues from mobilization of other cell populations [50,51]. Ex vivo expansion provides an exciting alternative to generating EC or EPCs for cellular therapy. In this study, EndoGo demonstrates a unique capability of significantly expanding total CB ECFC/EC numbers and promoting the expansion of CB EPCs. Although more studies are needed to effectively determine if EndoGo could produce large-scale, clinically relevant numbers of ECs leading to more effective therapies, EndoGo provides an excellent medium to expand ECFC/EC cell subpopulations, which could be useful in determining whether certain populations are needed to promote specific angiogenesis of arterial, venous or capillary vasculature. Further work needs to address whether these ex vivo expanded CB ECFCs/ECs can be utilized in treating disease models, whether in animal models or clinical-scale translatable expansions into clinical trials. Study and characterization of ECs from different sources also needs to be clarified to better understand EC biology and to develop protocols for cellular therapy.

Conclusion & future perspective
Although further studies are needed to determine how efficiently EndoGo can expand endothelial cells, this study is an important beginning step in developing an effective protocol to expand endothelial cells for clinical translation. This study also builds on the developing groundwork regarding the CB EPC that EndoGo targets for proliferation and expansion. Better understanding of specific subpopulations of endothelial cells may contribute to the improvement of EC expansion protocols for clinical cell therapy. Further detailed studies into the EPC and lineage progression could enable development of better ex vivo expansion protocols to promote larger clinical-scale numbers for infusion into patients undergoing cell therapy. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.
No writing assistance was utilized in the production of this manuscript. The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.

Open access
This work is licensed under the Creative Commons Attribution 4.0 License. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Summary points
• Revascularization of tissues is essential for clinical therapies.
• Large clinical-scale numbers are needed for endothelial cell therapy.
• Insufficient endothelial cell numbers exist in cord blood (CB) products and ex vivo expansion provides an approach to overcome that deficiency. • EndoGo XF significantly expands CB endothelial cells.
• CD34 + CB endothelial cells can be expanded using EndoGo XF.
• Ex vivo expanded CB CD34 + endothelial cells demonstrate less fibroblast-characteristics and lack of tube formation in vitro compared with CD34endothelial cells, suggestive of an angiogenic initiating cell. • EndoGo XF can significantly expand the CB endothelial progenitor cell.