Re-Defining Stem Cell-Cardiomyocyte Interactions: Focusing on the Paracrine Effector Approach.

Stem cell research for treating or curing ischemic heart disease has, till date, culminated in three basic approaches: the use of induced pluripotent stem cell (iPSC) technology; reprogramming cardiac fibroblasts; and cardiovascular progenitor cell regeneration. As each approach has been shown to have its advantages and disadvantages, exploiting the advantages while minimizing the disadvantages has been a challenge. Using human germline pluripotent stem cells (hgPSCs) along with a modified version of a relatively novel cell-expansion culture methodology to induce quick, indefinite expansion of normally slow growing hgPSCs, it was possible to emphasize the advantages of all three approaches. We consistently found that unipotent germline stem cells, when removed from their niche and cultured in the correct medium, expressed endogenously, pluripotency genes, which induced them to become hgPSCs. These cells are then capable of producing cell types from all three germ layers. Upon differentiation into cardiac lineages, our data consistently showed that they not only expressed cardiac genes, but also expressed cardiac-promoting paracrine factors. Taking these data a step further, we found that hgPSC-derived cardiac cells could integrate into cardiac tissue in vivo. Note, while the work presented here was based on testes-derived hgPSCs, data from other laboratories have shown that ovaries contain very similar types of stem cells that can give rise to hgPSCs. As a result, hgPSCs should be considered a viable option for eventual use in patients, male or female, with ischemic heart disease.


Introduction
After a half century of focused research, heart disease remains a leading cause of death in developed nations. Although advancements in research have led to numerous breakthroughs and progress for treating cardiovascular disease, in many respects, we are still only scratching the surface of innovative ideas for reducing cardiac-related deaths. A significant problem remaining to be solved is identifying the parameters and/or inductive signals that are necessary for repairing heart muscle damaged by ischemia. To solve this problem, decades of research have led to three primary areas of investigation: The potential use of induced pluripotent stem cell (iPSCs) technology, direct reprogramming of surviving cardiac fibroblasts, and regeneration of endogenous cardiovascular progenitor cells [1,2] . The first approach using iPSCs has shown promise in vitro; however, problems remain for their use in vivo such as teratoma formation, genetic instability, and accumulation of mitochondrial DNA mutations in iPSCs from elderly patients, which would most likely be the largest demographic needing iPSC-therapy [3,4] . In contrast, the latter two regenerative approaches have recently gained interest due to their paracrine-inducing abilities for restoring cardiac function [1,[5][6][7][8][9] ; however, the methodologies and protocols for inducing cardiac regeneration continue to vary widely.
Clinical trials that test the efficacy of transplanted adult stem cells (ASCs) within ischemic heart tissue have been common practice for nearly a decade; however, outcomes have not provided definitive clinical applications for patients. One explanation for the investigative ambiguity stems from the many different types of ASCs that have been pursued for

R E S E A R C H A R T I C L E
JSRM/Vol.14 No.1, 2018; P10

JSRM Code: 014010300003
Mahapatra S 1 , Martin D 1 , Gallicano GI 1 Stem cell research for treating or curing ischemic heart disease has, till date, culminated in three basic approaches: the use of induced pluripotent stem cell (iPSC) technology; reprogramming cardiac fibroblasts; and cardiovascular progenitor cell regeneration. As each approach has been shown to have its advantages and disadvantages, exploiting the advantages while minimizing the disadvantages has been a challenge. Using human germline pluripotent stem cells (hgPSCs) along with a modified version of a relatively novel cell -expansion culture methodology to induce quick, indefinite expansion of normally slow growing hgPSCs, it was possible to emphasize the advantages of all three approaches. We consistently found that unipotent germline stem cells, when removed from their niche and cultured in the correct medium, expressed endogenously, pluripotency genes, which induced them to become hgPSCs. These cells are then capable of producing cell types from all three germ layers. Upon differentiation into cardiac lineages, our data consistently showed that they not only expressed cardiac genes, but also expressed cardiac-promoting paracrine factors. Taking these data a step further, we found that hgPSC-derived cardiac cells could integrate into cardiac tissue in vivo. Note, while the work presented here was based on testes-derived hgPSCs, data from other laboratories have shown that ovaries contain very similar types of stem cells that can give rise to hgPSCs. As a result, hgPSCs should be considered a viable option for eventual use in patients, male or female, with ischemic heart disease transplantation. For example, adult epidermal stem cells are different than adult mesenchymal bone marrow stem cells with respect to gene expression, physiology, and origin. As a result, when introduced into the cardiac niche, the different types of ASCs present unanticipated variations. Consequently, finding a stem cell population that is the most suitable for treating cardiac ischemia remains an important endeavor.
Some of the more successful stem cell trials have been those that utilize both direct and indirect mechanisms to help induce cardiac repair [2,10] . Stem cells that are differentiated into cardiomyocytes in vitro (or in vivo) can integrate and adhere upon transplantation with endogenous cardiac cells via gap junctions (e.g., connexin 43 gap junction proteins). Cellular adhesion can then exert direct physiological interaction/repair. Some stem cells can also secrete paracrine factors that indirectly affect surrounding tissue to 'regenerate' or inhibit apoptosis [5,9,[11][12][13] . Debate has arisen, though, concerning which of these approaches is best for clinical use. For example, direct physiological interaction where stem cell-derived cardiomyocytes physically 'beat' can result in positive or negative outcomes for patients. If stem cell-derived cardiomyocytes are electrically connected to the heart muscle, ventricular force can be significantly restored [14][15][16] . However, if that electrical connection is not complete, transplanted cardiomyocytes that beat can cause detrimental arrhythmias and decreased ventricular force. Alternatively, stem cells that do not beat, but do secrete paracrine factors that can effect surrounding healthy cardiac tissue, have become a strong investigative mechanism for repairing ischemic tissue.
Our hypothesis is straightforward. We postulate that germline stem cells when removed from their niche begin to express factors redefining their stemness from unipotent (able to make sperm or eggs) to pluripotent. These redefined cells, known as germline pluripotent stem cells (hgPSCs), can then be induced to form paracrine effectoryielding cardiac cells. At first, our data provided constant, clear evidence that hgPSCs could be induced to form cardiomyocytes; however, we encountered a consistent obstacle with our initial approach; hgPSCs grew very slowly. Expansion of cells from dish to dish took months and it was concluded that their growth curve could significantly impede their use in vivo. Thus, to solve this problem, we modified a novel technological cell culture advancement that had been shown to promote certain cell types to acquire or maintain 'stemness' while simultaneously pushing them into the cell cycle [22] . This approach had not been reported in the literature for expanding hgPSCs. Our initial observations after subjecting hgPSCs to this novel culturing technology resulted in data leading us to generate a secondary hypothesis; that hgPSCs could be 'quickly' and indefinitely expanded to enrich their population, followed by their differentiation into cardiomyocytes. After testing our hypotheses, the data revealed that cardiomyocytes generated from hgPSCs acquired the ability to positively influence cardiac tissue.

Methods and Materials
Gene names and abbreviations Table 1 provides a reference list of abbreviations for the genes and proteins that were analyzed.

Isolation and culture of SSCs from human testes
Testes were acquired from the Washington Regional Transplant Community (WRTC) with permission from the next of kin. The tunica albica was removed and the seminiferous tubules were cut into 1g tissue samples and either stored in liquid nitrogen or used fresh [18] .
Frozen tissue samples were transferred to a 120ml container with 40ml ice-cold DMEM/F12 (Life Technologies Cat #11320082) + Antibiotic-Antimycotic (Life Technologies Cat #15240062), and washed twice. After washing in the medium, 2-3ml of the medium was left in the 120ml container (on ice) where the sample tissue is sliced by sterile scissors. The tissue was transferred into a 50ml tube with an additional 40ml ice-cold DMEM/F12 + Antibiotic-Antimycotic. The tissue was allowed to sediment for 2-5 minutes and supernatant is removed. A 10ml enzyme solution of 1x Hank's Balanced Salt Solution (HBSS) (Life Technologies Cat #14025076) was prepared with 2.5 mg/ml Collagenase Type IV (Life Technologies Cat #17104019), 1.25 mg/ml Dispase (Life Technologies Cat #17105041), vortex and filter through 0.22μm syringe filter (MidSci Cat #TP99722). The solution is added to the tissue sediment. The enzyme solution along with the tissue sediment was incubated 30mins in a 37 C water bath with 100rpm shaking. Afterwards, the enzyme was removed and re-suspended in 10ml human embryonic stem cell (hESC) medium (500ml DMEM/F12, 20% knockout serum replacement (Life Technologies Cat #A3181502), 0.1mM beta-mercaptoethanol (Life Technologies Cat #21985023), 5ml Non-essential amino acids (100X) (Life Technologies Cat #11140050), L-glutamine (100X) (Life Technologies Cat #25030081), and Antibiotic-Antimycotic. A 40μm mesh filter (Sigma Aldrich Cat #22363547) was placed atop of a 50ml tube and the supernatant and sample were slowly filtered through it to extract spermatogonial cells.
The filtered tissue sample was then centrifuged (1000rpm/5min). The supernatant was removed, and re-suspended in fresh 6ml hESC medium. The medium and sample were then seeded into a 6-well tissue culture plate along with 3.5μl of 10ng/ml Recombinant human GDNF (Life Technologies Cat #PHC7045) and the plate was placed into a 34 C and 5% CO2 incubator and cultured for 4 days.
Originally, we coated plates with gelatin (Sigma Aldrich Cat #Z707910); however, we later found that hgPSCs could grow virtually the same in uncoated TPP tissue culture plates (Sigma Inc. #92006).

Production of hESLCs from SSCs
After isolation, SSCs were cultured in hESC medium along with 3.5μl of GDNF for 4 days to stimulate growth and colony formation. They were incubated at 37 C and 5% CO2. Media was changed every other day. After the 4 th day of incubation, the hESC medium plus GDNF was replaced with hESC medium supplemented with 10 ng/ml fibroblast growth factor (bFGF) (Peprotech Cat #100-18C-10UG). Colonies must be cultured for at least 10 days to form hgPSCs [18] .

Reestablishing hgPSCs from expansion in GEM
After expansion, GEM was replaced with hESC medium containing bFGF. This medium was replaced every other day. Colonies spontaneously and consistently formed within 5-10 days.

Differentiation of hgPSCs into cardiac lineages
Un-expanded or previously expanded hgPSCs were cultured in hESC medium and 0.25μM Cardiogenol-C Hydrochloride (Sigma Aldrich Cat #C4866-5MG) for 10 days. After differentiation, the media was replaced into complete DMEM medium supplemented with 20% human serum. This medium was considered post-differentiation medium where the cardiac clusters could be cultured for up to 30 days.

Confocal analysis
Colonies were isolated using Dumont #5 forceps and fixed in 4% paraformaldehyde for 1 hour, followed by permeabilization with 1% Triton-X-100 for 30 minutes. After two washes in PBS, areas were blocked with 2% BSA in PBS and 1% Tween 20. Primary (1:100) and secondary (1:800) antibodies were diluted in blocking solution, and colonies were incubated with each antibody for 1 hour at 37˚ C or overnight at 4 o C. In the final step, colonies were moved directly from secondary antibody into a DAPI solution at 2μg/mL. Colonies were then mounted onto slides with anti-fade and sealed with

Refs
Atrial Natriuretic Peptide ANP 204bps Cardiac hormone that regulates blood pressure, vasodilation, natriuresis, and diuresis. [75] NK2 homeobox 5 NKX2.5 215bps Cardiac transcription factor responsible for heart formation and development. [76] Cardiac Troponin-I CTNI 335bps Key regulatory protein associated with the thin filament, inhibits actomyosin interactions at diastolic levels of intracellular Ca2+. [77] Cardiac Troponin-T CTNT 217bps Fixation of troponin complex on the actin filament and also participitates in muscle contraction [78] Myosin heavy chain MHC 542bps Molecular motor of the heart that generates motion by coupling its ATPase activity to its cyclic interaction with actin. [26,79] Myosin light chain 2A MLC2A 270bps Atrial marker expressed during devlopment and adulthood. Also regulates heart contraction along with MLC 2V. [80,81] Myosin light chain 2V MLC2V 380bps Ventricular marker during human heart development and in adulthood [80,81] Myocyte Enhancer Factor 2C ISL-1 127bps a LIM homeodomain transcription factor expressed in majority of cells in both right ventricle and atria of the heart. ISL1 is also responsible for survival, proliferation, and migration of cardiac progenitor cells. Promotes vasculo-and angiogenesis in myocardium and cardiomyocyte proliferation. Also helps regenerates mycoardium. [1,11,39,85,86] Insulin-like growth factor IGF-1 372bps Switch macrophages from pro-inflammatory to antiinflammatory phenotype both in vitro and vivo. Also promotes resident stem mobilization and cardiac lineage commitment. [1,11,39,85,86] Stromal derived factor-1 SDF-1 250bps Promotes repair and regeneration by recruiting circulating progenitor cells to the injured site. Also secreted by cardiac stem cells. [1,11,39,85,86] Connective tiddue growth factor CTGF 237bps Acts as a cofactor for other growth factor that promotes fibrosis and wound healing by enhancing ECM protein synthesis. [

Embryo and heart tube procurement and cell fusion
Female FVB mice were gonadotropin primed. E9.5 Embryos were collected from the uterine horns using Dumont #5 forceps to tease away fetal membranes; a modified version of previously described protocols [24,25] . Embryos were transferred to organ culture dishes (No. 3037, Falcon Inc., Lincoln Park, NJ). Beating hearts were isolated using two Dumont #5 forceps, one to spread the head away from the tail of the embryo, the other forceps was used to snip the heart from the body at the most caudal and rostral ends of the heart. Fetal hearts were then placed into a well of a 96 well plate and incubated at 37 C and 5% CO2. Hearts could beat for 72-96 hrs without distortion of morphology.
Mature Cardiomyocyte colonies were transiently transfected with uncut 5ul of cMHC-GFP DNA for 24 hours in 6-well dishes using Lipofectamine transfection kit (Life Technologies Cat #L3000001). cMHC-GFP (a kind gift from Dr. Eugene Kolossov [26] ) transfected colonies were picked using a 10µl pipette tip and placed into a holding well containing differentiation medium (above). They were then seeded into crevasses of fetal hearts using hand-drawn glass needle attached to a mouth pipette. Excess GFP-positive colonies were laid on top of beating fetal hearts. The 96 well plate was then incubated at 37 C and 5% CO2 for 48hrs.
Fetal hearts and colonies were analyzed first using a Leica stereoscope equipped with a fluorescent light source to globally identify areas of GFP within heart tissue. Those that appeared positive were fixed in 4% paraformaldehyde (Electron Microscopy Sciences Cat# 15714-S) for 2 hrs and then quenched in 1% glycine/PBS for 30min. Fetal hearts were then permeabilized with 1% Triton-X-100 for 2hrs. After three 30min washes in PBS, hearts were blocked with 2% BSA in PBS and 1% Tween 20. Primary CNX43 antibody (Rabbit Cell Signaling Cat #3512) and secondary antibodies (Peroxidase Affinipure Donkey Anti-Rabbit-Cat #711-035-152, Jackson Immunoresearch Labs) were diluted in blocking solution 1:100 and 1:800 respectively, and hearts were incubated with each antibody overnight 4˚ C. In the final step, hearts were moved directly from secondary antibody and incubated in a DAPI (Cat #D1306, ThermoFisher Scientific) solution at 2μg/mL. Hearts were then mounted onto slides with anti-fade and sealed with coverslips. Confocal images of colonies were taken on an Olympus Fluoview 500 Laser Scanning Microscope (See above).

Detection of secreted proteins
Paracrine factors secreted into the medium were detected by two methods. First, ~150 cardiac differentiated colonies were grown in 1ml of differentiation medium (containing serum replacement instead of human serum). Media was then isolated at 12hrs, 24hrs, and 48hrs and frozen. Antibodies to specific paracrine factors were then added to the thawed medium along with 10ml of protease inhibitor cocktail (Sigma Aldrich Cat #P8340) overnight at 4˚ C with rocking. The next day, magnetic beads (10µl) coated with g-protein (Life Technologies Cat #10003D) were added directly to the medium containing the primary antibodies. The slurry mixed for 2 hrs at 4˚ C. Using a magnet, the beads were washed 5x with PBS. Sample buffer was added to the beads, which were then heated at 95 o C for 5 min. 15µl of each time-point sample was added to a lane of a 4-20% polyacrylamide gel and the subjected to SDS-PAGE. The resultant gel was silver-stained (Bio Rad Cat #1610449) to observe the bands.
The second method, ethanol precipitation of proteins [modified from [28][29][30] ], was used to analyze proteins that were either too close to the molecular weight of the antibody heavy chain or was not compatible with Immunoprecipitation (I.P). In this case, 1ml samples of media taken at 12, 24, and 48hrs were subjected to four volumes of 100% ethanol and precipitated overnight at -80 o C. The next day the samples were centrifuged at 14,000rpm for 5 min and the supernatant was discarded. The remaining pellet was allowed to dry for 30min-1hr after which sample buffer was added and heated to 95 o C for 5 min. 15µl of each time-point sample was added to a lane of a 4-20% polyacrylamide gel and the subjected to SDS-PAGE. The gel was then used for Western Blot analysis to detect the paracrine factors.
ELISA TGFB secreted into the medium was detected by Human TGF beta 1 Platinum ELISA kit (Cat # BMS249-4, Life Technologies, Inc). Medium cultured in differentiated cardiomyocytes at 0h, 12h, 24h and 48h were tested along with TGFB standards (31pg/ml-2000pg/ml) at 450nm. A standard curve for TGF beta was plotted P14

Statistical analyses.
Standard error of the means were calculated from N > 3, which were then analyzed by student t tests. A p < 0.05 or better was considered significant.

Identifying the source cell from testis
Recent investigations have demonstrated that when isolated and cultured in the proper medium, germline stem cells can be induced to form cell/tissue from all three germ layers (i.e., ectoderm, mesoderm, and endoderm) [18][19][20]31] .
Identifying the stem cells within the human testes that give rise to hgPSCs has been somewhat elusive. To begin identifying this stem cell population, we first generated single cell suspensions assaying them for clonal expansion, which is accepted as a primary characteristic of stem cells. We used this approach first because, particularly in humans, a molecular or cellular target for identifying the actual stem cell population within the testes remains a debated topic in the literature. Consequently, we initially used cell size as a marker for isolating and identifying this population of stem cells. Figure 1 shows typical examples of isolated single cells, while Figure 2A shows the milieu of different sizes of cells enzymatically isolated from human testes tissue [18] . Using a hand-drawn glass pipette connected to a plastic filtered mouth suctioning tip, small (<5-7µm), medium (~8-12µm), and large (>12-20µm) cells were placed one by one into wells of a 96 well plate ( Figure 1). As a reference, a typical, single mouse ESC cell expanded into a colony very quickly after plating ( Figure 1A). Different cells isolated from the human testis ( Figure 1B) did not all expand in vitro. After 10-14 days of culture, only medium-sized cells grew into colonies capable of being differentiated and expressing markers from the three embryonic germ layers (Figure 1 C-F). Although the table in figure 1G reports that only 32% of medium-sized cells actually grew into distinctive colonies identical to that seen in Figure 1E, we concluded that this methodology was useful for confirming that a specific sub-population of cells within the SSC--enriched fraction could grow clonally and generate cells from all three germ layers.

P15
To apply a more rigorous approach for identifying which cell type could grow clonally into colonies of hgPSCs, we relied on data from many independent laboratories that had previously analyzed candidate markers for SSCs (publications found in [32] ). Different laboratories had previously analyzed candidate genes as markers for the testicular stem cells pinpointing SSEA4, GFR1α and GPR125 as putative SSC markers [publications found in [32] ]. To confirm or refute clonal growth of cells expressing each marker, we isolated immunofluorescently tagged cells from a testicular isolate using the same mouth pipette procedure as above, again placing each cell into its own well in a 96 well plate. Cells were labeled with primary antibodies directed against SSEA4, GFR1α, GPR125, or vimentin [32] , followed by fluorescently labeled secondary antibodies. The insets in Figure 1 show examples of fluorescent single cells within each well. Interestingly, SSEA4+ cells not only formed colonies, they were the same size as the medium cells previously seen to form colonies. Conversely, Gfrα1+ cells were much smaller and did not form colonies while the larger cells were only able to divide a few times in culture and resembled fibroblasts.
Most of the larger cells were vimentin positive suggesting they were Sertoli cells or cells from the lamina propria. As a result, these data confirmed that SSEA4+ cells are most likely the hgPSCs that acquire the ability to not only grow clonally when isolated from the testes and grown in a specified hESC medium, but also form cells/tissues from all three germ layers.

hgPSC colony-derived cardiac-lineage cells express cardioprotective paracrine factors
One approach that is gaining momentum for improving cardiac function after ischemic injury is the delivery of paracrine factors that can reprogram or induce repair of existing cardiac tissue. To identify if hgPSCs could be induced to form cardiac cells that secrete cardioprotective paracrine factors, we subjected them to a protocol used for human ESCs documented by [33] , followed by assaying by gene expression analysis using RT-PCR.
The procedure begins with 1g pieces of human testes tissue. Individual tissue samples are thawed and the SSC-enriched fraction is isolated using the protocol from [18] . Each sample yields ~500 hgPSC colonies. SSCs undergo de-differentiation to form hgPSCs when cultured in hESC medium for 10 days. Distinct hgPSC colonies begin to form after 2-3 days and each Yamanaka factor (Sox2, Oct4, Lin28, Klf4) as well as other stemness factors, nanog and CD73, are expressed ( Figure 2B). CD73 represents a marker for newly described endogenous plastic somatic cells (ePSCs; [34] Figure 2B).

Figure 2. Cardiac lineages can be produced from hgPSCs. A). Cells of the SSC-enriched fraction are cultured in medium containing GDNF for four days. The fraction is full of
All are detected in colonies after 10-14 days in culture. It is important to note that unlike iPSCs, which need exogenous expression of Yamanaka factors, these same factors are endogenously expressed and regulated in hgPSCs. Upon differentiation of hgPSC colonies, expression of cardiac genes is usually detected ~10 days of culture ( Figure 2C). Confocal microscopy verified distinct cardiac protein expression ( Figure 2D-I), while transient transfection with a cardiac MHC-promoter reporter construct confirmed induction of the cardiac lineage ( Figure 2J-L).
To test the ability of hgPSC-derived cardiac cells to express cardioprotective paracrine factors, we utilized the cMHC-GFP construct to pick cardiac-lineage colonies from the surrounding fibroblast-like cells. We then assayed for cardio-protective paracrine factor expression [1,6] . Strong gene expression of each paracrine factor was detected ( Figure 3A), while protein expression for two paracrine proteins in particular, IGF-1 and NRG1, was also very strong ( Figure  3B-D). Expression of IGF-1 and NRG-1 has recently been shown to be important for inducing surrounding healthy cardiac tissue to both proliferate and differentiate into the ischemic region [1,6] . NRG-1 is of special interest because of its potential for translation into the clinic for treating infarcts [1] . These data reveal that expression of paracrine factors is relatively robust within hgPSC-derived cardiac lineage cells after 21 days of differentiation. To our knowledge, this is the first evidence that hgPSC-induced cardiac lineage cells express these factors.

Differentiated hGPSCs show loss of pluripotency and teratoma formation
For quality-control issues, it was important to confirm that hGPSCs lost their ability to form teratomas. First, using RT-PCR, we verified that virtually all of the pluripotency genes were down regulated ( Figure 3E); however, there was weak expression of Nanog and CD73 in cardiomyocytes. Through literature search, we found that Nanog can be expressed in the myocardium of rats [35] and CD73 is needed for successful differentiation into cardiomyocytes [36] ( Figure 3E). Equally as important to identifying loss of pluripotency was identifying molecularly, the ability (or loss of) of these cells to form teratomas. The RT-PCR data in Figure 3F clearly show expression of three genes, DND1 and p19 CDKI and p18 ink4C in hGPSCs-derived cardiomycoytes [37] . The expression of these genes are consistent with the loss of teratoma formation in hGPSCs and cardiomyocytes [37,38] . As a result, cardiomyocytes differentiated from hGPSCs are molecularly incapable of forming teratomas. We also performed a no-RT control on p18 INK4C primers because they sit on one exon and observed that there was no genomic DNA contamination in both hgPSCs and cardiomyocytes (data not shown).

Figure 3. Differentiation of cardiac colonies beyond day10 results in colonies that express pro-cardiac regenerative paracrine factors. A) Rt-PCR shows expression of seven procardiac regenerative paracrine factors. B-D) Immunofluorescent and DIC analyses using antibodies directed against IGF-1 and NRG-1 show colonies staining positive for both paracrine factors. Arrowheads in B and C point to regions of variable staining within colonies. Asterisks highlight fibroblasts that can emanate from colonies. They show no fluorescent staining. E) Expression of most pluripotency genes is below the level of detection in differentiated colonies
. Nanog and CD73 are expressed; however, nanog expression has been reported at low levels by [35] and CD73 expression has been reported as necessary for cardiac development [37] . F) Genes known to inhibit teratoma formation are expressed in hgPSC-derived cardiomyocytes. They are also expressed in hGPSCs (data not shown).

Quick, efficient expansion of hgPSCs using hgPSC expansion medium (GEM)
The data preceding figure 4 were acquired from primary hgPSCs and cardiac cells differentiated directly from primary hgPSC colonies. However, we found that while hgPSC-derived cardiac cells showed much promise for eventual cardiac repair, it was somewhat painstaking to accumulate the amount of cells that would be needed for injection into a patient. Months were needed to generate enough biological material from 1g biopsy-sized pieces of testes. As a result, we modified a cellular reprogramming technique by [22] and applied it to hgPSCs to determine if we could expand the hgPSC population more quickly (and using no animal components) when compared to conventional expansion protocols. Figure 4 illustrates the process of germ cell expansion and subsequent re-establishment of hgPSCs, followed by their differentiation into cardiac lineages. By day 10 in GEM, the cells within colonies took on a cobble-stone appearance typical of previous reports [22] . While in GEM, these cells could be expanded indefinitely and quickly; that is, one colony placed in a 96 well plate would typically be split into two wells of a 96 well plate within 7 days. Comparing expansion rates of primary hgPSCs to expanded hgPSCs grown in GEM from two patients, we found that upon re-establishment in hESC medium, ~4X more colonies were obtained by the fourth passage when compared to primary hgPSCs grown in conventional culture medium. More importantly, it took 30 days less time to expand hgPSCs in GEM compared to primary hgPSCs grown in conventional medium ( Figure  5; Table 3). We have expanded hgPSCs in GEM at least 20X. Reestablishing hgPSC colonies is accomplished by replacing GEM with hESC medium. Usually within ten days of culture, gene expression patterns match primary hgPSCs including the expression of all Yamanaka factors, nano( [34] ; Figure 4F). Confocal microscopy of colonies provided further evidence that these cells retained pluripotent markers ( Figure 4G).
These re-established hgPSC colonies could then be differentiated down the cardiac pathway resulting in an expression pattern of paracrine factors similar to that observed in cardiac cells differentiated from primary hgPSCs ( Figure 4H, I). As a result, we conclude that re-established cardiac colonies are virtually identical to cardiac colonies generated from fresh primary hgPSCs.   While it was important to show that cardiomyocytes derived from hgPSCs expressed cardiac and paracrine factor genes (see Table 1 for paracrine factor list and functions), it was further necessary to identify their physiological ability to secrete those paracrine factors. A consensus of at least six paracrine 'effects' categories has been endorsed within the literature [1,11,39,40] (see Table 1). These categories include (Survival, proliferation, immune cells, remodeling, vascularization, and CPC activation). Secretion of paracrine factors would mean they can affect surrounding tissue.

Germline stem cells for cardiac repair
To detect secretion, two different methods were employed. First, after culturing day 21-differentiated cardiomyocytes for 0hrs, 12hrs, 24hrs and 48hrs, the entire compliment of medium (1.0 ml) from each time point was directly tested by immunoprecipitation (IP) using antibodies to specific paracrine factors, followed by antibody isolation using magnetic protein G-coated beads. The pull-down products were subjected to SDS-PAGE, followed by silver stain of the gels. This highly sensitive approach revealed VEGFA, CTGF, IGF-1, and TGFβ all increasing in concentration over a 48hr period of incubation ( Figure 6).  Figure 5. 500 hgPSC colonies were expanded by traditional conditions (i.e., by trypsinization and passaging 1:2) or they were expanded 1:2 in GEM. Comparing two different patients, no marked difference was observed until the second and third passages where GEM-grown hgPSCs grew ~2X faster than conventional growth. By the fourth passage, GEM-grown colonies re-generated close to 4X more colonies when compared to conventional growth and passaging. More importantly, those ~4x more colonies were obtained in almost half the time.
The second method we used to detect secreted proteins involved precipitation of proteins from the medium using 4 volumes of ice cold ethanol (as published in [28][29][30] ), then separating them by SDS-PAGE, and probing by Western blot using antibodies directed against each paracrine factor ( Figure 6E and F). This Western blot approach was employed primarily where the antibodies had not been tested for applications such as I.P. Here, we found that CTGF and NRG-1 were secreted into the medium by 24-48hrs. When combined, these data provided strong evidence that hgPSC-derived cardiomyocytes secrete paracrine factors known to be proangiogenic and procardiogeneic [1,6,41] .
It is difficult to quantify these data because secreted "housekeeping" genes are not well characterized for these cells; however, we do know that the lower level of protein detection for silver stained gels is in the ~ 1-10ng range. In an attempt to confirm and quantify what we observed from our silver stain data, we analyzed media by ELISA ( Figure 6G). Since, sensitivity of both ELISA and silver stain are similar (0.1-10ng of protein), the ELISA data confirmed what we observed in the silver stains; TGFβ is secreted within the 1-10ng/ml range within 48 hrs of culture. More specifically, our silver stain results showed that from 10µl of sample media loaded into each well, ~5.0 x 10 5 cells secrete paracrine factors in the 1-10ng/ml range by 48 hrs of culture. Figure 6G illustrates ELISA data showing a gradual increase of TGFβ from 0h to 24h, with very little change in secretion between 24h and 48h of incubation. According to Beck et al [42] , the active form of TGFβ can last for over a day after which it is subjected to various types of degradation. This observation by Beck et al [42] may be the reason for the slight decrease we see at 48 hours compared to 24 hours as some of the TGFβ might have started to degrade by 48 hours. Most importantly, however, is that these TGFβ ELISA data confirm the silver stain results. The concentration of each paracrine factor is well within range for effecting of surrounding tissues ( Figure  6). These data are the first to show that hgPSC-derived cardiac cells not only express, but also secrete cardio-protective paracrine effectors.

hgPSC-derived cardiomyocytes can integrate to beating cardiac tissue in vivo
We next wanted to determine if hgPSC-derived cardiomyocytes could fuse with beating heart tissue. Many different animal models have been utilized, with rodent hearts being the most prevalently used model system, to demonstrate infiltration and integration-ability of various types of stem cells within cardiac tissue [14,43] . One problem with rodent model systems for human stem cell research, however, is that the normal beat-rate for mouse hearts can vary between 400-600 beats/minute. The dramatic difference between human cardiac cells and rodent hearts has been reported to create incompatible physiological variances that can skew results [14] . As a result, we chose the mouse fetal heart as a 'better', more physiologically compatible model system for transplanting cardiac-specific, GFPlabeled, hgPSC-derived cardiomyocytes (Figure 2 and 7). This model system is better than adult cardiac model systems because the mouse fetal heart beats at ~60-70 beats/min, which is very similar physiologically to the human heart. cMHC-GFP-positive hgPSC-derived cardiac colonies were physically isolated by mouth pipette using a Leica Fluorescent stereoscope. GFP+ Colonies were then pipetted into tight crevices within the heart tube so that they could not fall away from the beating heart tube. Each well of the 96 well plate contained 10-15 fetal hearts, which forced virtually all loose cardiac colonies to remain in close contact with cardiac tissue. After 48hrs of incubation, fetal hearts were observed live using the same fluorescent stereoscope to identify GFP(+) areas. Upon detection, fetal hearts were fixed in 3.0% formaldehyde for 2hrs, followed by processing for immunofluorescence using an antibody for the cardiac gap junction protein Connexin43 (CNX43). Multiple areas of GFP labeled cells were visibly co-stained with the gap junction protein CNX43 providing evidence that the hgPSCderived cardiac cells directly fused with the mouse cardiac tissue. Fusion of GFP labeled cells was observed in 8 out of the 10 fetal hearts.

Discussion
Although optimism remains for investigating stem cells for treating or curing cardiovascular disease, much disparity remains among the various published approaches and results. Here, we show three novel outcomes that could lead to the production and delivery of paracrine factors known to induce and improve cardiac function in ischemically injured heart muscle [1,8,9,13] . First, we show that hgPSC-derived cardiac cells express and secrete pro-cardiac regenerative paracrine factors. Secondly, the expansion of hgPSCs can be markedly amplified when cultured in GEM using human serum (no animal products); and thirdly, the data provide strong evidence that hgPSCderived cardiomyocytes can physically incorporate into the cardiac niche.
Since the initial findings that a germ-line stem cell could 'revert' to a state that resembled hESCs/hiPSCs both genotypically and phenotypically, understanding their biology and therapeutic potentials has been consistently investigated by various laboratories [1,18,20,31,44] . Defining the 'stemness' of germ-line stem cells has been debated [45] ; however, the most recent study from the Skutella laboratory has provided the best evidence yet as to their true identity [31] . They reported that hgPSCs are adult stem cells, but they share a gene expression profile that is related to, but not identical to true pluripotent stem cells. Perhaps the most important finding in [31] was the confirmation of hgPSC colony plasticity as they were shown to attain the ability to differentiate into cells of all three germ layers [17,18,31,32,46] . From a swath of previously identified 'potential' SSC markers, the elegant studies from two independent laboratories pinpointed SSEA4 as a strong candidate marker for the testicular stem cells that might be giving rise to hgPSCs. Interestingly, these data somewhat contradicted data from other laboratories who had previously suggested Gfr1α and Gpr125 as candidate SSC markers [publications found in [32] ]. In light of this controversy, research on human testes using various methodologies including magnetic-activated cell sorting [MACS [47] , indirect immunofluorescence [48] , and FACS [49] ] provided the evidence that SSEA4+ cells were most likely the best candidates that, when cultured in hESC medium, could progressively expressed pluripotent factors and subsequently generate cells from all three germ layers. Closer inspection of those studied showed that they did not clearly identify the ability of SSEA4+ cells to grow clonally into hgPSC colonies.
Here, we confirmed that hgPSCs do endogenously express all of the pluripotent Yamanaka factors plus newly identified others (Nanog and CD73) known to be important for stemness (This study [18] ) and that they are readily down-regulated once differentiation commences. Thus, once differentiated, hgPSCs show virtually no risk of teratoma formation in vivo. In fact, closer inspection of reports where teratoma formation was specifically analyzed after injection of naked hgPSCs into nude mice revealed that tens of millions of hgPSCs over a month were needed to generate a tiny nodule containing multiple cell types [18] . This result was in stark contrast to the teratomas that formed within three weeks from mere thousands of ESCs injected into the opposing hind flank of the same nude mice. It is these types of observations, when taken together, that keep hgPSC research relevant.
Upon differentiation of hgPSCs, similar to hESC and hiPSC-derived cardiac cells [15,50,51] , we found that hgPSC-derived cardiac cells varied with respect to atrial/ventricle gene expression representing a mixed population of cardiac cells; however, one aspect that was not shared was spontaneous beating. Cardiac cells derived from human hESCs and hiPSCs consistently show rhythmic beating; however, hgPSC-derived cardiac cells did not. Initially, we were discouraged by this non-beating phenotype; however, recent revelations within the literature led us to pursue the alternative, paracrine effector, and regenerative pathway.
For years, numerous investigations have tapped into the 'beating' phenotype for potentially repairing ischemically damaged heart muscle. In fact, some adult stem cell (ASC) work has shown promise as ejection fractions slightly improve after beating ASC-derived cardiomyocytes are injected into infarcted heart muscle. However, one major detriment with many ASC-cardiac repair studies is the potential for arrhythmias as stem cell-derived cardiomyocytes, in many cases, can beat at their own pace. As a result, instead of matching stem cell/cardiac tissue electrophysiology, a paradigm shift towards cardiac tissue regeneration mediated either directly or indirectly by paracrine factors has been proposed as more beneficial. We originally showed that hgPSCs could express a small cohort of cardiac genes representing differentiation [18] ; however, in depth analyses of these cells for potential cardiac repair was not assessed in [18] . There have been hundreds of studies investigating if and how ASCs as well as ESCs and iPSCs function to repair ischemic cardiac tissue (reviewed by [2,7] ). Some of those studies have resulted in clinical trials; however, the varying outcomes of those trials have not led to their routine clinical use because the best mechanism for translating their potential in vivo remains debated. On one hand, generating cardiomyocytes with the ability to electrically couple within the ischemic region of the heart has resulted in some, albeit good, evidence showing attenuation of left ventricular (LV) remodeling and improved LV systolic function [40,52,53] . In addition, data from non-human primates and large mammals have also shown neovascularization upon stem cell injection [50,54] . On the other hand, much of the data where matching stem cell/heart electrophysiology was a focus of the study, stem cells were found to generate a diverse set of mature and immature atrial, ventricle, and even sinusoidal cells periodically resulting in arrhythmic islands of beating cells [50,54] . As a result, although the electrophysiology avenue of research remains ardently investigated among many different types of stem cells, it was exciting to find that the differentiation potential of hgPSCderived cardiomyocytes included expression and secretion of key paracrine factors known to be pro-growth and pro-differentiation [1,9] .
A consensus of at least six paracrine 'effects' categories has been endorsed within the literature [1,11,39,40] . Analyzing the SSEA+, medium-size hgPSC-derived cardiomyocytes revealed that they expressed and secreted paracrine factors from all six paracrine 'effects' categories. Genes including VEGFA (representing vascularization and remodeling), NRG1 (proliferation and vascularization), IGF1 (survival, proliferation, and suppression of immune response), TGFβ (Immune regulation, remodeling), and SDF1 (vascularization and CPC activation) all were not only expressed, but also secreted into the media at or near nanogram/ml levels. As a result, the data presented here provide evidence that hgPSC-derived cardiomyocytes can either directly or indirectly influence cardiac tissue via paracrine effects.
Although identifying expression and secretion of paracrine factors was an important step in determining the ability of hgPSC-derived cardiac cells for potential use in vivo, it was also important to identify their ability to integrate into cardiac tissue. Without integration into cardiac tissue, delivery of paracrine effectors would be impaired or highly unlikely. The model system we employed to identify integration took into consideration both successes and failures identified by many investigations [11,39,51,55] . For example, injecting undifferentiated adult stem cells or differentiated hESCs/hiPSCs into infarcted rodent hearts has resulted in widely varying reports from improved cardiac function to little or no attenuation [7,15,55] . Efficacy of integration into the cardiac niche was one characteristic dictating their successes or failures. Shared gap junctions characteristically represent good integration [56,57] ; however, even if gap junctions are readily observed between endogenous and exogenous cells, problems can still arise including arrhythmias. Arrhythmic beating can result from injected cells 'beating' on their own and/or because they cannot match the beat-rate of the injected heart [7] . Rodent hearts can beat >400 beats/min. A human heart beating that fast would be considered undergoing tachycardia. On the contrary, because embryonic hearts beat at ~70 beats/min, we determined they would serve as a better model system for identifying the integration ability of hgPSC-derived cardiac cells into a cardiac niche. hgPSC-derived cardiac cells expressing a cardiac promoter-driven GFP reporter gene along with immunofluorescent staining of the endogenously expressed cardiomyocyte gap junction protein CNX43 clearly showed multiple areas of integration. As a result, these data provide good evidence justifying further, more detailed investigations of hgPSC-derived cardiac cells for use in cardiac repair, perhaps within a non-human primate model.
None of these data would have been possible, though, if we had not had enough hgPSC-derived cardiac cells to work with. Soon after we began work on the data presented here, a novel technology was published for growing primary epithelial cells indefinitely [22] . That work astonishingly, yet elegantly showed that a small tracheal biopsy could be grown indefinitely in a specifically defined medium (>40 passages), followed by complete re-establishment of all cell types within the trachea when transferred back to a media promoting tracheal development. Their medium was composed of a DMEM/Hams F12 mix (1:1) containing a ROCK inhibitor and 25% J2 cell conditioned medium. We found that hgPSCs grew very quickly in the medium defined by [22] . About 100 colonies would grow to confluency in a 6 well dish within ~5 days and could then be passaged indefinitely; however, we did encounter a problem in about 50% of our experiments. The cells would fill with small, dark vacuoles that would eventually kill the cells. As a result, we tried growing cells in medium using various combinations of the components specified in [22] . We found that growing hgPSCs in medium containing a ROCK inhibitor without the J2 component did not alter their ability to expand indefinitely. We also substituted animal serum with human serum to remove all animal products. Consequently, these two changes were substantial enough to warrant renaming our medium to 'hgPSC expansion medium' (GEM) to quell any confusion. So far we have reached ~20 passages of hgPSCs using GEM while not losing the ability to re-establish colonies positive for all Yamanaka factors, CD73 and Nanog.

Conclusion
With this technique in hand, generating cells for experimental analysis has become much more efficient. We also believe this technique could speed up the time needed for treating patients from months to weeks. The resulting re-established hgPSCs also re-attained their differentiation potential as they not only expressed the same cardiac marker set tested in primary hgPSC-derived cardiac cells, but also the same set of paracrine factors. We believe the data presented here could eventually, truly help patients with heart disease. An important fact to note is that the work presented here is not male-centric.
Evidence from the Tilly laboratory has revealed the existence of gPSCs within the ovary [58,59] . Patients suffering from ischemic heart injury could be subjected to a testes/ovary biopsy (could be done in an IVF clinic). The biopsy is then given to the scientist for isolating SSCs/OSCs (isolated from the biopsy) and subsequent dedifferentiation into hgPSCs. Indefinite expansion of hgPSCs in GEM would be followed by re-establishment into hgPSCs, and then differentiation into paracrine factor yielding cardiomyocytes. These cells would be sent back to the medical doctor who would then inject them back into the patient. Some of the expanded hgPSCs in GEM could be frozen down for future treatment (in case the same patient comes back for treatment again).
While we are planning to expand this work into the ovary to identify the potential of female hgPSCs for treating or curing heart disease in women, we also do not want to rule out the research being pursued on very small embryonic-like stem cells (VSELs). It was recently suggested that VSELs might share similarities with gPSCs with respect to their capacity to restore spermatogenesis in mice [60] . Whether VSELs and gPSCs are one in the same remains to be determined. If they are based on the data presented here, both should be pursued with vigor to help treat or cure heart disease.