Running title: Directed differentiation via vesicle delivery

Extracellular vesicles derived from preosteoblasts influence embryonic stem cell differentiation

Rekha Nair1,2,3, Lívia Santos1,2,3, Siddhant Awasthi1,2, Thomas von Erlach1,2,3, Lesley W. Chow1,2,3, Sergio Bertazzo1,3, Molly M. Stevens1,2,3*

1Department of Materials, Imperial College London SW7 2AZ, UK2Department of Bioengineering, Imperial College London SW7 2AZ, UK3Institute of Biomedical Engineering, Imperial College London SW7 2AZ, UK*Corresponding author

Author information:

Address: Royal School of Mines, South Kensington Campus, London SW7 2AZFax: +44 (0)20 7594 6757Phone: +44 (0)20 7594 6804Email:

Rekha Nair: rekha.nair29@gmail.comLívia Santos: l.rocha-dos-santos@imperial.ac.ukSiddhant Awasthi: siddhant.vit@gmail.comThomas von Erlach: t.von-erlach10@imperial.ac.ukLesley W. Chow: l.chow@imperial.ac.ukSergio Bertazzo: s.bertazzo@imperial.ac.uk

Corresponding author information:

Molly M. StevensAddress: 208, Royal School of Mines, South Kensington Campus, London SW7 2AZTelephone: +44 (0)20 7594 6804Fax: +44 (0)20 7594 6757Email: m.stevens@imperial.ac.ukWebsite: www.stevensgroup.org

1

ABSTRACT

Embryonic stem cells (ESCs) can differentiate into all cell types of the body and therefore hold tremendous

promise for cell-based regenerative medicine therapies. One significant challenge that must be addressed prior

to using ESCs in the clinic is to improve methods of efficiently and effectively directing the differentiation of

this heterogeneous cell population. The work presented here examines the potential of harnessing naturally

derived extracellular vesicles to deliver genetic material from mature cells to undifferentiated ESCs for the

purpose of manipulating stem cell fate. Vesicles were isolated from preosteoblast cells and were found to be

~170 nm in diameter and express the CD40 surface marker. Multiple interactions were visualized between

vesicles and ESCs using confocal microscopy, and no significant difference in cell viability was noted.

Incubation with vesicles caused significant changes in ESC gene expression, including persistence of

pluripotent gene levels as well as increased neurectoderm differentiation. Genetic cargo of the vesicles as well

as the cells from which they were derived were examined using a small microRNA (miRNA) gene array.

Interestingly, ~20% of the examined miRNAs were increased more than two-fold in the vesicles compared to

preosteoblast cells. Together, these results suggest that extracellular vesicles may be utilized as a novel method

of directing stem cell differentiation. Future work examining methods for controlled delivery of vesicles may

improve the clinical potential of these physiological liposomes for therapeutic applications.

Key words: Embryonic stem cells, directed differentiation, extracellular vesicles, microvesicles, exosomes, microRNA

2

INTRODUCTION

Embryonic stem cells are derived from the inner cell mass of the pre-implantation blastocyst and can be

differentiated into any somatic cell of the body [1-3]. These pluripotent cells are well studied for their potential

for cell-based therapies in regenerative medicine and hold tremendous promise for applications where

endogenous cells are lost due to injury or disease. Despite great progress in the field, however, there is still a

relative lack of directed differentiation schemes that can reproducibly, effectively, and efficiently control stem

cell fate. These schemes are critical to develop if pluripotent stem cells, which are a naturally heterogeneous

cell population, are to be used for clinical therapies. Current methods of directing stem cell fate tend to rely on

soluble factors, culture conditions, or small molecules that are timed specifically to recapitulate developmental

processes [4,5]. These methods, however, tend to be inefficient in that each step of the process leads to

intermediate progenitor cell states until the final differentiated product is reached. Since differentiation

efficiency tends to decrease at each state, the resulting population is still relatively heterogeneous [4,6].

Another method of directing stem cell differentiation involves manipulation at the genetic level via ectopic

expression of genes of interest [7,8] or suppression using RNA interference [9-11] to influence the direction of

cell fate [12,13]. However, such genetic manipulation can be less than ideal for clinical applications since

exposure to these added genes or expression vectors may have long term negative consequences [13]. In recent

years, another method of altering recipient cell behavior at the transcript level has emerged that involves

horizontal gene transfer using particles naturally shed from cells. These extracellular vesicles have been

implicated in cell to cell communication in a variety of mammalian cell systems, including cancer cell lines

[14,15], stem cells [16-18], and primary cells [19,20].[21] Subclasses of these vesicles can be categorized based

on their biogenesis pathways and can sometimes be identified based on enrichment of specific surface markers

as well as particle size. For example, exosomes are formed via the endolysosomal pathway, are typically 40-

120 nm, and are enriched in tetraspanins, while microvesicles are formed through the outward budding of the

cell membrane, are 50-1,000 nm, and can be identified by expression of the CD40 ligand [21]. Both types of

vesicles have been shown to contain a variety of genetic information, including RNA, miRNA, and other non-

3

coding RNA [22]. A wide array of literature supports the presence of these vesicles in cancer cells [23,24], but

there has also been evidence to indicate their presence in a variety of other cell systems as well, including ES

cells [17,18,25], mesenchymal stem cells [26-31], and dendritic cells [32,33]. In several of these cases, research

has also shown that cells shedding these vesicles are able to transfer their own genetic material to distant cells

and that this genetic material can be incorporated and translated to protein in the recipient cell population

[22,34]. Such a transfer of genetic material holds tremendous promise for the devised manipulation of a targeted

cell population, and indeed, is already known to play roles in human physiology, including in stem cell

maintenance [17] and tissue repair [35].[21]

The aim of the work presented here was to determine whether vesicles derived from a non-embryonic source

could be used to manipulate cell fate in a pluripotent stem cell system. It was hypothesized that a preosteoblast

cell line would be of interest as a candidate for the vesicle source since its own potential for differentiation

suggests an active signaling base that could result in vesicle-bound paracrine cues, potentially leading to

quantifiable changes in the ES cells. In this study, vesicles were derived from a murine preosteoblast cell line

and characterized using several techniques, including nanoparticle tracking analysis, FACS, and immunogold

transmission electron microscopy. The ability of these cells to alter the behavior of differentiating ES cells was

then examined using confocal microscopy, viability studies, and quantitative PCR. Finally, the genetic content

transferred from the donor cell to its vesicles was examined in depth using a miRNA array. Results from this

work suggest that preosteoblasts can reproducibly produce vesicles containing genetic information that can

effectively alter ES cell fate. This novel means of manipulating stem cell differentiation may signify future

possibilities of altering cell behavior, thereby improving the potential of using heterogeneous pluripotent stem

cell populations for cell-based clinical therapies.

4

MATERIALS AND METHODS

Cell culture

Mouse preosteoblasts derived from bone/calvaria (MC3T3-E1 subclone 4) were purchased from ATCC and

cultured using minimum essential medium (MEM) alpha without ascorbic acid (Invitrogen) supplemented with

10% (v/v) fetal bovine serum. Cells were trypsinized (0.05% (v/v) Trypsin EDTA) routinely once they reached

~90% confluence and were fed every 2-3 days. The preosteoblasts were grown to confluence and were switched

to serum free MEM alpha 48 hours prior to particle collection.

Undifferentiated mouse embryonic stem cells (D3 line, purchased from ATCC) were cultured feeder-free on

tissue culture treated polystyrene plates coated with 0.1% (w/v) gelatin in Dulbecco’s Modified Eagle’s

Medium (DMEM, Invitrogen) supplemented with 10% (v/v) fetal bovine serum, 0.1 mM non-essential amino

acid solution, 2 mM L-glutamine, 0.1 mM 2-mercaptoethanol, and 1000 U/mL leukemia inhibitory factor (LIF).

ESCs were fed every other day and passaged using 0.05% (v/v) Trypsin EDTA when cells reached 70%

confluence. For studies investigating ESC behavior and differentiation in the presence of preosteoblast particles,

cells were plated on 48-well plates at a density of ~5500 cells/cm2 and were cultured without LIF. Both the

MC3T3-E1 and D3 cell lines have been well studied and characterized by numerous research groups.

Researchers probing the behavior of mouse embryonic stem cells in particular have extensively examined the

D3 cell line.

Particle isolation

Conditioned media from preosteoblasts grown in serum-free conditions for 48 hours was collected and spun

down at 4000 rpm (3220 rcf) for 30 minutes at 4oC. The supernatant was then transferred to 25 mL sterile

Konical™ ultracentrifuge tubes (Beckman) and spun at 100,000 x g for two hours at 4oC in a Beckman Coulter

Optima L-100XP ultracentrifuge. Particles were concentrated at the bottom of the conical ultracentrifuge tubes

and could then be resuspended in phosphate-buffered saline (PBS) or ESC medium depending on the

application.

5

Transmission electron microscopy (TEM)

Particles were fixed in paraformaldehyde directly after isolation and incubated on a TEM grid (copper grid with

300 mesh, carbon coated; Electron Microscopy Sciences) for 10 minutes. For samples to be immunostained,

excess sample solution was gently wicked off using filter paper, and the grid containing the particles was

blocked in a 1% (v/v) bovine serum albumin (BSA)/PBS solution for 10 minutes. The sample was then

incubated with anti-CD40 primary antibody (diluted 1:100 in BSA/PBS; Abcam) for one hour and washed five

times (one minute each, in 1% (v/v) BSA/PBS). Particles were then incubated with immunogold solution

(diluted 1:1 in PBS; 7.5-10 nm particles synthesized according to [36,37]) for one hour, washed five times (one

minute each, in PBS), fixed in 4% (v/v) gluteraldehyde for five minutes, and finally washed again five times

(one minute each, in water). Samples were imaged at 80 kV using a JEOL 2010 TEM.

Nanoparticle tracking analysis (NTA)

Particles were resuspended in 10 mL PBS immediately after isolation. After ensuring even distribution of the

particles within PBS through repeated pipetting, 1 mL of the solution was drawn into a syringe and inserted into

the NanoSight LM10. Particle movement was tracked for 60 seconds where particle size distribution was

monitored using Nanoparticle Tracking Analysis software. Within one batch of isolated particles, this process

was repeated an additional two times, and all readings were averaged to obtain n = 1. NTA readings were taken

from four separate batches of isolated particles (n = 4).

Flow activated cell sorting (FACS)

Isolated particles were suspended in 50 uL PBS/0.1% (w/v) BSA. Samples were then labeled with CD40 (5 L;

FITC hamster anti-mouse CD40, BD Pharmingen™) or isotype control (FITC hamster IgM, 1 isotype control,

BD Pharmingen™) and incubated for 20 minutes in the dark. Particle populations were analyzed in a three

laser, 11 color Becton Dickinson Fortessa instrument equipped with DiVa software; ten thousand events were

collected for each sample.

6

Confocal microscopy

ESCs were plated at a density of ~9000 cells/cm2 onto u-Slide 8-well ibiTreat chamber slides (Ibidi) and were

allowed to adhere overnight at 37oC and 5% CO2. Immediately following isolation, particles were suspended in

a solution containing CellMask™ Orange (1:1000 dilution in PBS; Invitrogen) in order to label their

membranes. Undifferentiated ESCs that had been allowed to attach overnight were switched to differentiation

media (ESC medium without LIF) containing the stained vesicles and were then returned to growth conditions

(37oC and 5% CO2). After six and 24 hours, wells were rinsed with PBS to remove traces of medium and non-

adherent particles, fixed for 30 minutes at room temperature in 4% (v/v) paraformaldehyde and blocked for 1h

with 4% (w/v) BSA. Wells from each time point were then incubated for three minutes with 300 uL 3 mM

DAPI solution (1:1000 in water), rinsed in PBS, incubated with 300 uL 488 Phalloidin (1:350 in PBS;

Invitrogen) for 20 minutes, and finally rinsed three times in PBS prior to coverslipping. The samples were

analyzed with a Leica SP5 inverted confocal microscope with 63x (1.4n/a) oil immersion objective under

standard imaging conditions.

Viability (alamarBlue® and LIVE/DEAD®)

ESCs were plated onto 24-well plates at a density of ~5500 cells/cm2 and allowed to adhere overnight at 37oC

and 5% CO2. Immediately after isolation, particles were suspended in ESC medium without LIF and distributed

to wells; each experimental group was given vesicles derived from 6 mL of conditioned MC3T3 medium. Cells

were then returned to growth conditions (37oC and 5% CO2). At 24 and 48 hours, cells were rinsed in PBS and

replenished with a solution of ESC medium without LIF containing a fresh batch of isolated vesicles. After four

and six days, wells were rinsed with PBS and incubated in 10% (v/v) alamarBlue® (Invitrogen) solution for 2.5

hours at 37oC and 5% CO2). Fluorescent readings of the wells containing cells alone and cells incubated with

particles were taken in a Spectramax M5 plate reader (Molecular Devices) (excitation 570 nm, emission 585

nm) to determine the relative levels of metabolic activity of the two populations.

7

In parallel, separate wells were incubated with LIVE/DEAD® solution (Invitrogen; 3 uL 1 mg/mL calcein + 3

uL ethidium homodimer-1 in 1.5 mL PBS) for 20 minutes at 37oC and 5% CO2. Cells were then rinsed in PBS

and imaged using a 1X51 inverted microscope (Olympus) and DP70-BSW software (Olympus). Cells

fluorescing green are considered live since the calcein (excitation 495 nm, emission 515 nm) is well retained in

live cells while those fluorescing red are considered dead since the ethidium homodimer-1 (excitation 528 nm,

emission 617 nm) was able to penetrate the membrane.

RNA and DNA extraction and quantification

ESCs were plated onto 24-well plates at a density of ~5500 cells/cm2 and allowed to adhere overnight at 37oC

and 5% CO2. Immediately after isolation, particles were suspended in ESC medium without LIF and distributed

to wells; each experimental group was given vesicles derived from 6 mL of conditioned MC3T3 medium. Cells

were then returned to growth conditions (37oC and 5% CO2). At 24 and 48 hours, cells were rinsed in PBS and

replenished with a solution of ESC medium without LIF containing a fresh batch of isolated vesicles. After 0

(the day of initial particle distribution), 3, and 7 days, wells containing treated and untreated cells were rinsed

with PBS, trypsinized (0.05% Trypsin EDTA), and spun down (300 x g, 5 minutes). The resulting pellet was

rinsed in PBS and resuspended in 350 uL RNeasy lysis buffer (Qiagen). RNA was then extracted according to

the manufacture’s protocol. Briefly, the lysed cells were homogenized using a Qiashredder and incubated with

an equal volume of 70% ethanol. Samples were then transferred to the RNeasy spin column and rinsed with

Buffer RWI and Buffer RPE prior to final elution in water. RNA was then quantified using a Nanodrop 2000c

(Thermo Scientific).

Isolated particles were examined for the presence of DNA and RNA (see protocol above). For DNA analysis,

particles were dissolved in 6M guanidine hydrochloride immediately after isolation. DNA was then quantified

using the Quant-iT™ PicoGreen® assay (Invitrogen) according to manufacture instructions. Briefly, samples as

well as calf thymus DNA standards were incubated with PicoGreen® buffer TE and solution containing

PicoGreen® dsDNA reagent, a fluorescent nucleic acid stain, and read on a Spectramax M5 plate reader

8

(excitation: 485 nm, emission: 528 nm). A linear equation was generated from DNA standards and used to

determine DNA concentrations in the samples.

Quantitative real time polymerase chain reaction (qRT-PCR)

To explore alterations at the gene expression level of ESCs incubated with the particles, cDNA was first

synthesized from isolated RNA using the QuantiTect Reverse Transcription kit (Qiagen) according to

manufacture’s protocol. Briefly, RNA diluted to 40 ng/uL using RNase-free water was incubated with 2 uL

gDNA Wipeout Buffer for 2 minutes at 42oC. Samples were then incubated with Reverse Transcription (RT)

Master Mix (at a ratio of 1 volume Reverse Transcriptase: 4 volumes RT Buffer: 1 volume RT Primer Mix) at

42oC for 15 minutes, followed by 3 minutes at 95oC. Freshly synthesize cDNA samples were then combined

with QuantiTect SYBR Green and PCR primers and tested under the following cycle conditions: hold at 95oC

for 15 minutes, cycle (40 cycles) at 94oC for 15 seconds, 58oC for 30 seconds, and 72oC for 30 seconds. A melt

curve was then run between 70-95oC. For annealing temperatures and primer details, see Supplementary Table

1. Gene expression of oct4, nanog, sox2, sox17, gata6, foxa2, nestin, sox1, and fgf5 were calculated relative to

undifferentiated ESC samples using the Pfaffl method (reference), with glyceraldehyde-3-phosphate

dehydrogenase (gapdh) used as the housekeeping gene.

miRNA array and analysis

To further explore genetic content contained within the vesicles, a miScript™ miRNA PCR array (Qiagen) was

run on both particles and the cells from which they were derived. Samples were isolated through

ultracentrifugation (particles) or trypsinization (cells) and analyzed according to manufacture’s protocols.

Briefly, total RNA (including small RNAs) was isolated using the miRNeasy Micro Kit (Qiagen) wherein

samples were lysed using QIAzol Lysis Reagent and homogenized via vortexing. Samples were then incubated

at room temperature for 5 minutes, and chloroform was added to induce phase separation. Ethanol was then

added to the aqueous phase, and total RNA including small RNAs were bound, washed, and eluted in 14 uL

RNase-free water. To synthesize cDNA, 250 ng RNA from each sample was incubated with miScript™ 5x

9

HiSpec buffer, 10x Nucleics Mix, and Reverse Transcriptase Mix for 60 minutes at 37oC, then 5 minutes at

95oC. To run the array, cDNA samples were combined with 2x QuantiTect SYBR Green PCR Master Mix, 10x

miScript™ Universal Primer, and RNase-free water and dispensed equally (10 uL per well) into Rotor-Disc®

100 Mouse Cell Differentiation & Development arrays. Arrays were then sealed using a Rotor-disc heat sealer

(Qiagen), placed into the RotorgeneQ PCR machine (Qiagen), and run using the following settings: initial

activation step of 95oC for 15 minutes, 3-step cycling (40 cycles) of 94oC for 15 seconds, 55oC for 30 seconds,

and 70oC for 30 seconds (fluorescence data collection was performed at this step).

Using the miScript miRNA PCR Array Data Analysis Web Portal provided by SABiosciences, a normalization

gene was selected among the examined miRNAs since the housekeeping genes typically used for quantitation

were dissimilar between the cells and vesicles. Software analysis assessed the most tightly regulated gene

amongst those examined and selected miR-1a-3p as the normalization gene. Fold changes in gene expression

were subsequently analyzed using the Ct method of quantitation, whereby vesicles were compared relative to

the cells from which they were shed after individual samples were normalized to internal miR-1a-3p levels. The

Data Analysis Web Portal was also utilized to generate a two-dimensional hierarchical clustering of the data

using average linkage clustering. Clustering results were represented visually by a heat map, with green

indicating decreased expression and red indicating increased expression. The degree of color intensity

corresponds to the magnitude of fold change (either an increase or a decrease). A scatter plot examining

log10(2^Delta Ct) of the vesicles v. cells was also generated to highlight fold regulation, with the red dots and

text indicating an up-regulation in gene expression in vesicles compared to cells and green indicating a decrease

in fold regulation.

Statistics

10

Significance testing was conducted using SYSTAT software. For individual genes in PCR analysis, expression

fold change comparisons within and across time points were evaluated using analysis of variance (ANOVA)

with subsequent post-hoc Tukey analysis to determine significance (p < 0.05).

11

RESULTS

Vesicle characteristics

Particles isolated from mouse preosteoblasts were characterized to investigate morphology, size, and surface

markers (Figure 1). Isolated particles were visualized using immunogold labeling of CD40 and transmission

electron microscopy (TEM). Positive labeling indicating the presence of CD40 on the surface of the particles

was observed throughout. Particles were irregularly shaped, and TEM-based diameter analysis of the images

indicated particle sizes of approximately 200 nm. Additional images of stained and unstained vesicles are

provided in Supplementary Figure 1. Further, nanoparticle tracking analysis (NTA) was used to more accurately

assess particle size. The average size from four different batches of isolated particles was determined to be

174.2 ± 22 nm, confirming data observed from TEM analyses. All NTA assessments displayed similar size

distributions (data from a representative batch is displayed in Figure 1), with only one distinct population seen.

Particle sizing determined by both TEM and NTA suggest that the particles isolated from the preosteoblasts fall

within the size range of microvesicles and not exosomes [21]. Surface markers are often used to further

distinguish between these types of particles, with the presence of CD40 again identifying a microvesicle

population. As noted above, immunogold labeling of the particles revealed the presence of CD40. In addition,

FACS analysis also identified a strong CD40-positive population within the isolated particles. Forward v. side

scatter (FSC-W v. SSC-W) as well as count assessments indicated a distinct positive shift in fluorescent signal

compared to isotype controls.

12

Cellular uptake efficiency of vesicles

The interaction and uptake efficiency of preosteoblast-derived particles incubated with undifferentiated

embryonic stem cells was analyzed qualitatively by high resolution confocal scanning microscopy at multiple

time points (Figure 2). The vesicle membrane was fluorescently labeled with a cell membrane dye prior to

incubation with adherent stem cells to allow for cellular tracking. Within six hours, a large number of vesicles

(red) were detected within all analyzed ESC clusters, indicating high cellular uptake efficiency of the

microvesicles. The particles were widely dispersed within the ESC cluster and were primarily co-localized

within the nuclei, the site of action of the vesicle cargo, which indicates endosomal escape and cargo delivery of

the vesicles within the cell. By 24 hours, some evidence of the particles still remained, though fewer were

visible through confocal microscopy, possibly due to lysosomal degradation of the fluorescently labeled vesicle

membrane within the cells.

The viability of differentiating ESCs was not affected by the presence of the particles. alamarBlue® readings of

cells incubated with preosteoblast-derived vesicles did not change significantly compared to untreated cells at

either day 4 or day 6 after incubation. Further, cells at each time point that were incubated with LIVE/DEAD®

stain primarily stained green (live) with very few visible red (dead) cells, supporting the alamarBlue® results

and indicating that the vesicles did not negatively affect ESC viability. These results were consistent across

vesicles derived from separate batches of MC3T3s (data not shown).

Alteration of recipient cell behavior

In addition to examining the viability of recipient cells, alterations in the population’s gene expression were

investigated (Figure 3). ESCs grown under differentiation conditions (no LIF) were either left untreated or were

incubated with vesicles, and RNA was collected after three or seven days. Levels of gene expression of

pluripotent (oct4, nanog, sox2) and differentiation (mes/endoderm: sox17, gata6, foxa2; ectoderm: nestin, sox1,

fgf5) markers were analyzed using RT-PCR and expressed as fold change values compared to undifferentiated

ESCs (day 0). ESCs incubated with the preosteoblast-derived vesicles expressed higher levels of all pluripotent

13

markers at almost all time points examined. At day 3, oct4 levels were significantly higher in vesicle-treated

cells (p < 0.01), as were nanog levels (p < 0.05) at day 7. In all cases, the fold change was less than 1, with the

exception of oct4 (1.39 ± 0.3) and sox2 (1.07 ± 0.5) at day 3 in the vesicle-treated population. Taken together,

these results suggest that both populations exhibited some extent of differentiation by day 7, but the presence of

the vesicles in the ESC culture delayed the onset of differentiation.

Analysis of lineage markers also indicated a modification in the differentiation trajectory of the treated cell

population. This panel of markers was specifically chosen to represent each of the three germ lineages, with an

up- or down-regulation in any group suggesting that the ES cells were moving toward or away from those

lineages, respectively. No significant differences were noted in the mesendoderm markers, although the vesicle-

treated cells exhibited higher fold changes at all time points, with the exception of gata6 at day 7 (Figure 3).

However, neurectoderm markers nestin and sox1 displayed significant increases in expression at day 7 in the

vesicle-treated cells compared to untreated cells, with nestin fold changes increasing from 1.34 ± 0.8 to 13.39 ±

4.8 and sox1 shifting from 0.18 ± 0.1 to 11.18 ± 6.6 (p ≤ 0.001 in both cases). The gene expression level of fgf5

similarly increased at day 7 from 3.34 ± 2.6 to 41.47 ± 20.3, though the change was not statistically significant.

These results were consistent across vesicles derived from separate batches of preosteoblasts (data not shown)

and suggest that the presence of the vesicles resulted in a shift towards the neurectoderm lineage in the

differentiating ESCs.

Identification of cargo contained within particles

In order to probe why the presence of the vesicles in differentiating ESC cultures resulted in gene expression

changes compared to untreated cells, the content of the particles was examined. For initial assessments on the

cargo contained within the preosteoblast-derived particles, total DNA, RNA, and miRNA was isolated from

multiple batches. On average, the particles isolated from a monolayer of preosteoblasts grown to confluence on

a T225 flask contained 7.06 ± 1.6 ng DNA and 1.17 ± 0.3 mg RNA. To further explore the relevance of this

14

content, a miRNA array was utilized to identify which miRNA molecules were most concentrated in the

vesicles compared to the preosteoblast population from which they were derived (Figure 4).

In total, 84 different miRNA genes affiliated with development and differentiation were analyzed in cell and

isolated vesicle populations (n = 3 in both cases). Heat map analysis clustered individual cell and vesicle groups

together, as expected, with the latter demonstrating overall higher expression levels of the miRNAs examined

(indicated in red) (Figure 4A). Cluster analysis suggested that miRNA contained within the cell population was

generally expressed at lower levels (green) than their vesicle counterparts, indicating that cells undergo a

process that concentrates genetic material within the particles. This trend is corroborated by examining specific

Ct values from each group: 44% of the isolated vesicles exhibited miRNA Ct values under 20 (indicating

relatively high gene expression), while 37% fell under a Ct of 20 in the cell group. Log scatter plots of the

vesicle versus cell populations highlight these differences between the groups (Figure 4B). Points falling along

the diagonal line crossing the center of the graph indicate miRNA expression levels that are equivalent between

groups. As corroborated by the heat map cluster analysis, most points (81%; 68/84 genes) fall above this

diagonal (fold change > 1), suggesting a general increasing expression trend in vesicle miRNAs compared to

the cells from which they were derived. The two outer diagonal lines delineate a two-fold expression increase or

decrease in vesicle as compared to cells. Approximately 20% (17/84 genes) of the analyzed miRNA genes in

isolated vesicles exhibited a greater than two-fold increase compared to cells (red dots). In contrast, only 6% of

genes (5/84) showed greater than two-fold decreased expression (green dots). Of the miRNAs exhibiting an

increase in fold regulation in vesicles compared to cells, five displayed a ten-fold increase or greater: miR-122-

5p (12.1 fold), miR-451a (312.3 fold), miR-183-5p (17.5 fold), miR-144-3p (67.5 fold), and miR-142-5p (14.4

fold). In contrast, only one miRNA (miR-541-5p) was down-regulated more than 10-fold. Figure 4C highlights

these differences, showing the differing percentages of total up- and down-regulated genes. In addition, the

breakdown of up- or down-regulation is displayed, indicating the percentage of genes within those that are up-

regulated (red bar) that exhibit a 2-10 fold increase versus a greater than 10 fold increase. Similarly, the

percentages of genes within those that are down-regulated (green bar) are differentiated based on either a 2-10

15

fold down-regulation or a greater than 10 fold down-regulation (for a list of which genes fall into these

categories, see Supplementary Table 2).

16

DISCUSSION

Particles naturally shed from a variety of cell types have been previously categorized most commonly as

apoptotic bodies, microvesicles, exosomes, as well others [21]. These “physiological liposomes” have been

shown in several cell systems to contain genetic material originating from the host cell and in many instances to

contain a combination of RNA and miRNA [21]. The extracellular vesicles presented in this study were

harbored within the preosteoblast microenvironment and were isolated using ultracentrifugation techniques.

They were found to be ~170 nm in diameter, and immunogold and FACS analyses indicated the presence of

CD40 in the population, suggesting that at least a portion of the isolated particles may fall into the

“microvesicle” category established by recent literature [21]. Preosteoblast vesicles appeared to impact the

maintenance of pluripotency as well as promoted neurectoderm differentiation in treated ESCs. Both mRNA

and miRNA were found in the preosteoblast vesicles, and the definitive alteration in gene expression in the

recipient ESCs suggests that the vesicles may have responded to this cargo. While the role of mRNA in

influencing gene expression is well understood, the impact of miRNA on stem cell pluripotency differentiation

has only recently been investigated and continues to evolve [38-44].

As outlined previously, approximately 20% (17/84) of the miRNAs examined exhibited at least a two-fold

increase in the vesicles compared to the cells from which they were derived. Previous work in other cell

systems has similarly shown an increase in genetic material in vesicles relative to the cells [15,22,45]. Twelve

of the seventeen up-regulated miRNAs in the preosteoblast system are affiliated with the mesoderm lineage,

which may be explained by its derivation from a mesoderm-derived cell source. Of importance to note is that

miRNAs may play roles in multiple lineages, so those associated with the mesoderm may also be important in

other developmental processes. Five of the miRNAs that were up-regulated at least two-fold (miR-182-5p (2.1

fold), miR-183-5p (17.5), miR-320-3p (2.1), miR-205-5p (2.5), miR-133b-3p (8.8)) have been implicated in

pluripotency, which may explain why ESCs incubated with these vesicles exhibited higher levels of pluripotent

genes. In addition, miR-141-3p, which promotes self-renewal in ESCs,[46] was up-regulated 8.2-fold while all

examined miRNAs in the pro-differentiation let-7 family [41,47] exhibited levels around 1 (0.9 – 1.5), with no

17

significant increase relative to preosteoblast cells. Interestingly, three of the miRNAs up-regulated at least two-

fold (miR-182-5p, miR-183-5p, miR-205-5p) have also been reported to influence development of the

ectoderm, specifically in relation to neurological and inner ear development, as well as epidermal differentiation

[48]. Given the up-regulation of these miRNAs in the isolated vesicles, it is possible that one or all of them

played a role in the neurectoderm differentiation observed in the ESCs treated with vesicles. In addition to miR-

183-5p, only two other miRNAs exhibited at least a 15-fold increase in the vesicles: miR-451a (312.3 fold) and

miR-144-3p (67.5 fold), both of which have been implicated in erythropoiesis and megakaryopoiesis [49].

Although research is still relatively young regarding the roles of miRNAs and their corresponding effects in

pluripotent cells, the work presented here has established the importance of understanding what is contained

within naturally shed vesicles since the genetic material enclosed within them may have significant effects on

any recipient cell type. As an example, in other cases where documented protective effects are induced by shed

particles, it is possible that miRNA cargo contained within the particles and transferred to the recipient cell play

critical roles [30,31]. As demonstrated here, one potential application of this transferrable genetic material is as

a means of directing the differentiation of stem cells. It is possible that the stem cell niche in vivo is influenced

by particles shed from cells in that environment, which could direct the fate of localized stem cells. The

technology of a biomaterials construct in which vesicles are incorporated could be one method of delivering

packaged miRNAs to heterogeneous pluripotent populations as a means of controlled modulation at the gene

level. In addition to this potential critical application as a means of controlling differentiation, the transfer of

genetic material between cells has tremendous potential in the development of therapies. It is currently not well

known how cells select the mRNA or miRNA that will be contained within the vesicles, although evidence

suggests a non-random packaging [45], but controlling or understanding that process will be critical for

manipulating vesicle transfer for therapeutic purposes. As an example, consider miR-133b-3p, which was up-

regulated almost nine-fold in the preosteoblast vesicles. This miRNA belongs to the cluster mmiR-133, which

has been described in multiple reports for their potential in the treatment of heart disease [50,51] and promoting

angiogenesis [52,53]. In addition, miR-133b has been shown to inhibit the growth of non-small-cell lung cancer

18

[54] and invasion of prostate cancer cells [55]. If properly harnessed, it is possible that this miRNA, along with

related molecules, may be able to integrate into critical niches to enhance or provide therapeutic efforts towards

the treatment of specific diseases. It is, however, first critical to ensure that the select group of miRNAs are

consistently up-regulated and transferred into the vesicles. Future work coupling controlled delivery methods

along with vesicle biology may also serve to direct the targeting and uptake of the vesicles to the desired

recipient cell population.

The horizontal transfer of genetic material via extracellular vesicle delivery may be a potent mediator for

directing the differentiation of pluripotent populations. The work presented here identifies particles shed from

preosteoblasts that resemble microvesicles and contain both mRNA and miRNA. The presence of these vesicles

in differentiating ESC cultures significantly influenced stem cell fate, prolonging expression levels of

pluripotent markers while increasing later markers of neurectoderm differentiation. Future work investigating

the control of miRNA packaging as well as targeted delivery of the vesicles may significantly enhance the

potential of this technology in cellular reprogramming and directed differentiation.

19

ACKNOWLEDGMENTS

The authors would like to acknowledge Kharissa Nitiputri for her contributions to the immunogold labeling of

the extracellular vesicles. RN was supported by a Whitaker Foundation International Scholarship, and LS was

supported by a Marie Curie Actions Intra-European Fellowship.

20

REFERENCES

1. Evans MJ and MH Kaufman. (1981). Establishment in culture of pluripotential cells from mouse

embryos. Nature 292:154-6.

2. Martin GR. (1981). Isolation of a pluripotent cell line from early mouse embryos cultured in medium

conditioned by teratocarcinoma stem cells. Proc Natl Acad Sci U S A 78:7634-8.

3. Wobus AM, H Holzhausen, P Jakel and J Schoneich. (1984). Characterization of a pluripotent stem cell

line derived from a mouse embryo. Exp Cell Res 152:212-9.

4. Ben-David U, O Kopper and N Benvenisty. (2012). Expanding the boundaries of embryonic stem cells.

Cell Stem Cell 10:666-77.

5. Murry CE and G Keller. (2008). Differentiation of embryonic stem cells to clinically relevant

populations: lessons from embryonic development. Cell 132:661-80.

6. Cohen DE and D Melton. (2011). Turning straw into gold: directing cell fate for regenerative medicine.

Nat Rev Genet 12:243-52.

7. Kim DW, S Chung, M Hwang, A Ferree, HC Tsai, JJ Park, TS Nam, UJ Kang, O Isacson and KS Kim.

(2006). Stromal cell-derived inducing activity, Nurr1, and signaling molecules synergistically induce

dopaminergic neurons from mouse embryonic stem cells. Stem Cells 24:557-67.

8. Klein D, T Schmandt, E Muth-Kohne, A Perez-Bouza, M Segschneider, V Gieselmann and O Brustle.

(2006). Embryonic stem cell-based reduction of central nervous system sulfatide storage in an animal

model of metachromatic leukodystrophy. Gene Ther 13:1686-95.

9. Greber B, H Lehrach and J Adjaye. (2007). Silencing of core transcription factors in human EC cells

highlights the importance of autocrine FGF signaling for self-renewal. BMC Dev Biol 7:46.

10. Liu YP, SV Dambaeva, OV Dovzhenko, MA Garthwaite and TG Golos. (2005). Stable plasmid-based

siRNA silencing of gene expression in human embryonic stem cells. Stem Cells Dev 14:487-92.

11. Zou GM and MC Yoder. (2005). Application of RNA interference to study stem cell function: current

status and future perspectives. Biol Cell 97:211-9.

21

12. Grivennikov IA. (2008). Embryonic stem cells and the problem of directed differentiation. Biochemistry

(Mosc) 73:1438-52.

13. Giudice A and A Trounson. (2008). Genetic modification of human embryonic stem cells for derivation

of target cells. Cell Stem Cell 2:422-33.

14. Al-Nedawi K, B Meehan, J Micallef, V Lhotak, L May, A Guha and J Rak. (2008). Intercellular transfer

of the oncogenic receptor EGFRvIII by microvesicles derived from tumour cells. Nat Cell Biol 10:619-

24.

15. Skog J, T Wurdinger, S van Rijn, DH Meijer, L Gainche, M Sena-Esteves, WT Curry, Jr., BS Carter,

AM Krichevsky and XO Breakefield. (2008). Glioblastoma microvesicles transport RNA and proteins

that promote tumour growth and provide diagnostic biomarkers. Nat Cell Biol 10:1470-6.

16. Lai RC, TS Chen and SK Lim. (2011). Mesenchymal stem cell exosome: a novel stem cell-based

therapy for cardiovascular disease. Regen Med 6:481-92.

17. Ratajczak J, K Miekus, M Kucia, J Zhang, R Reca, P Dvorak and MZ Ratajczak. (2006). Embryonic

stem cell-derived microvesicles reprogram hematopoietic progenitors: evidence for horizontal transfer of

mRNA and protein delivery. Leukemia 20:847-56.

18. Yuan A, EL Farber, AL Rapoport, D Tejada, R Deniskin, NB Akhmedov and DB Farber. (2009).

Transfer of microRNAs by embryonic stem cell microvesicles. PLoS One 4:e4722.

19. Chavez-Munoz C, J Morse, R Kilani and A Ghahary. (2008). Primary human keratinocytes externalize

stratifin protein via exosomes. J Cell Biochem 104:2165-73.

20. Faure J, G Lachenal, M Court, J Hirrlinger, C Chatellard-Causse, B Blot, J Grange, G Schoehn, Y

Goldberg, V Boyer, F Kirchhoff, G Raposo, J Garin and R Sadoul. (2006). Exosomes are released by

cultured cortical neurones. Mol Cell Neurosci 31:642-8.

21. S ELA, I Mager, XO Breakefield and MJ Wood. (2013). Extracellular vesicles: biology and emerging

therapeutic opportunities. Nat Rev Drug Discov 12:347-57.

22

22. Valadi H, K Ekstrom, A Bossios, M Sjostrand, JJ Lee and JO Lotvall. (2007). Exosome-mediated

transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells. Nat Cell

Biol 9:654-9.

23. Antonyak MA, B Li, LK Boroughs, JL Johnson, JE Druso, KL Bryant, DA Holowka and RA Cerione.

(2011). Cancer cell-derived microvesicles induce transformation by transferring tissue transglutaminase

and fibronectin to recipient cells. Proc Natl Acad Sci U S A 108:4852-7.

24. Muralidharan-Chari V, JW Clancy, A Sedgwick and C D'Souza-Schorey. (2010). Microvesicles:

mediators of extracellular communication during cancer progression. J Cell Sci 123:1603-11.

25. Kubikova I, H Konecna, O Sedo, Z Zdrahal, P Rehulka, H Hribkova, H Rehulkova, A Hampl, J Chmelik

and P Dvorak. (2009). Proteomic profiling of human embryonic stem cell-derived microvesicles reveals

a risk of transfer of proteins of bovine and mouse origin. Cytotherapy 11:330-40, 1 p following 340.

26. Bruno S, C Grange, F Collino, MC Deregibus, V Cantaluppi, L Biancone, C Tetta and G Camussi.

(2012). Microvesicles derived from mesenchymal stem cells enhance survival in a lethal model of acute

kidney injury. PLoS One 7:e33115.

27. Chen TS, RC Lai, MM Lee, AB Choo, CN Lee and SK Lim. (2010). Mesenchymal stem cell secretes

microparticles enriched in pre-microRNAs. Nucleic Acids Res 38:215-24.

28. Collino F, MC Deregibus, S Bruno, L Sterpone, G Aghemo, L Viltono, C Tetta and G Camussi. (2010).

Microvesicles derived from adult human bone marrow and tissue specific mesenchymal stem cells

shuttle selected pattern of miRNAs. PLoS One 5:e11803.

29. Kim HS, DY Choi, SJ Yun, SM Choi, JW Kang, JW Jung, D Hwang, KP Kim and DW Kim. (2012).

Proteomic analysis of microvesicles derived from human mesenchymal stem cells. J Proteome Res

11:839-49.

30. Li T, Y Yan, B Wang, H Qian, X Zhang, L Shen, M Wang, Y Zhou, W Zhu, W Li and W Xu. (2013).

Exosomes derived from human umbilical cord mesenchymal stem cells alleviate liver fibrosis. Stem

Cells Dev 22:845-54.

23

31. Zhang B, Y Yin, RC Lai, SS Tan, AB Choo and SK Lim. (2014). Mesenchymal Stem Cells Secrete

Immunologically Active Exosomes. Stem Cells Dev.

32. Montecalvo A, AT Larregina, WJ Shufesky, DB Stolz, ML Sullivan, JM Karlsson, CJ Baty, GA Gibson,

G Erdos, Z Wang, J Milosevic, OA Tkacheva, SJ Divito, R Jordan, J Lyons-Weiler, SC Watkins and AE

Morelli. (2012). Mechanism of transfer of functional microRNAs between mouse dendritic cells via

exosomes. Blood 119:756-66.

33. Thery C, M Ostrowski and E Segura. (2009). Membrane vesicles as conveyors of immune responses.

Nat Rev Immunol 9:581-93.

34. Aliotta JM, M Pereira, KW Johnson, N de Paz, MS Dooner, N Puente, C Ayala, K Brilliant, D Berz, D

Lee, B Ramratnam, PN McMillan, DC Hixson, D Josic and PJ Quesenberry. (2010). Microvesicle entry

into marrow cells mediates tissue-specific changes in mRNA by direct delivery of mRNA and induction

of transcription. Exp Hematol 38:233-45.

35. Gatti S, S Bruno, MC Deregibus, A Sordi, V Cantaluppi, C Tetta and G Camussi. (2011). Microvesicles

derived from human adult mesenchymal stem cells protect against ischaemia-reperfusion-induced acute

and chronic kidney injury. Nephrol Dial Transplant 26:1474-83.

36. Slot JW and HJ Geuze. (1985). A new method of preparing gold probes for multiple-labeling

cytochemistry. Eur J Cell Biol 38:87-93.

37. Slot JW and HJ Geuze. (1981). Sizing of protein A-colloidal gold probes for immunoelectron

microscopy. J Cell Biol 90:533-6.

38. Houbaviy HB, MF Murray and PA Sharp. (2003). Embryonic stem cell-specific MicroRNAs. Dev Cell

5:351-8.

39. Sun X, X Fu, W Han, Y Zhao and H Liu. (2010). Can controlled cellular reprogramming be achieved

using microRNAs? Ageing Res Rev 9:475-83.

40. Mallanna SK and A Rizzino. (2010). Emerging roles of microRNAs in the control of embryonic stem

cells and the generation of induced pluripotent stem cells. Dev Biol 344:16-25.

24

41. Huo JS and ET Zambidis. (2013). Pivots of pluripotency: the roles of non-coding RNA in regulating

embryonic and induced pluripotent stem cells. Biochim Biophys Acta 1830:2385-94.

42. Melton C and R Blelloch. (2010). MicroRNA Regulation of Embryonic Stem Cell Self-Renewal and

Differentiation. Adv Exp Med Biol 695:105-17.

43. Ivey KN and D Srivastava. (2010). MicroRNAs as regulators of differentiation and cell fate decisions.

Cell Stem Cell 7:36-41.

44. Calabrese JM, AC Seila, GW Yeo and PA Sharp. (2007). RNA sequence analysis defines Dicer's role in

mouse embryonic stem cells. Proc Natl Acad Sci U S A 104:18097-102.

45. Chen X, H Liang, J Zhang, K Zen and CY Zhang. (2012). Secreted microRNAs: a new form of

intercellular communication. Trends Cell Biol 22:125-32.

46. Lin CH, AL Jackson, J Guo, PS Linsley and RN Eisenman. (2009). Myc-regulated microRNAs

attenuate embryonic stem cell differentiation. EMBO J 28:3157-70.

47. Melton C, RL Judson and R Blelloch. (2010). Opposing microRNA families regulate self-renewal in

mouse embryonic stem cells. Nature 463:621-6.

48. Weston MD, ML Pierce, HC Jensen-Smith, B Fritzsch, S Rocha-Sanchez, KW Beisel and GA Soukup.

(2011). MicroRNA-183 family expression in hair cell development and requirement of microRNAs for

hair cell maintenance and survival. Dev Dyn 240:808-19.

49. Rasmussen KD, S Simmini, C Abreu-Goodger, N Bartonicek, M Di Giacomo, D Bilbao-Cortes, R

Horos, M Von Lindern, AJ Enright and D O'Carroll. (2010). The miR-144/451 locus is required for

erythroid homeostasis. J Exp Med 207:1351-8.

50. Abdellatif M. (2010). The role of microRNA-133 in cardiac hypertrophy uncovered. Circ Res 106:16-8.

51. Care A, D Catalucci, F Felicetti, D Bonci, A Addario, P Gallo, ML Bang, P Segnalini, Y Gu, ND

Dalton, L Elia, MV Latronico, M Hoydal, C Autore, MA Russo, GW Dorn, 2nd, O Ellingsen, P Ruiz-

Lozano, KL Peterson, CM Croce, C Peschle and G Condorelli. (2007). MicroRNA-133 controls cardiac

hypertrophy. Nat Med 13:613-8.

25

52. Deregibus MC, V Cantaluppi, R Calogero, M Lo Iacono, C Tetta, L Biancone, S Bruno, B Bussolati and

G Camussi. (2007). Endothelial progenitor cell derived microvesicles activate an angiogenic program in

endothelial cells by a horizontal transfer of mRNA. Blood 110:2440-8.

53. Zhang HC, XB Liu, S Huang, XY Bi, HX Wang, LX Xie, YQ Wang, XF Cao, J Lv, FJ Xiao, Y Yang

and ZK Guo. (2012). Microvesicles derived from human umbilical cord mesenchymal stem cells

stimulated by hypoxia promote angiogenesis both in vitro and in vivo. Stem Cells Dev 21:3289-97.

54. Liu L, X Shao, W Gao, Z Zhang, P Liu, R Wang, P Huang, Y Yin and Y Shu. (2012). MicroRNA-133b

inhibits the growth of non-small-cell lung cancer by targeting the epidermal growth factor receptor.

FEBS J 279:3800-12.

55. Tao J, D Wu, B Xu, W Qian, P Li, Q Lu, C Yin and W Zhang. (2012). microRNA-133 inhibits cell

proliferation, migration and invasion in prostate cancer cells by targeting the epidermal growth factor

receptor. Oncol Rep 27:1967-75.

DISCLOSURE STATEMENT

No competing financial interests exist.

26

FIGURE LEGENDS

Figure 1. Vesicle characterization. Vesicles isolated from preosteoblast cultures stained positive for CD40

using immunogold transmission electron microscopy (immunoTEM; top left). Nanoparticle tracking analysis

indicated that particles were on average 174.2 ± 22 nm (top right). Flow cytometric analysis identified a CD40-

positive population (bottom panel), confirming immunoTEM indications of the surface marker.

Figure 2. Cell-vesicle interactions. Extracellular vesicles were stained with CellMask™ Orange and visualized

on clusters of actin (phalloidin/green)- and nuclei (dapi/blue)-stained embryonic stem cells after 6 (top left) and

24 (top right) hours. The presence of vesicles in ESC culture did not influence viability, as indicated by

alamarBlue® (bottom left) and LIVE/DEAD® (bottom right) assays. Scale bars in LIVE/DEAD® images are

200 m and are both from day 6.

Figure 3. ESC gene expression changes. The influence of extracellular vesicles on differentiation ESCs was

monitored using RT-PCR. Pluripotent genes oct4 and nanog as well as neurectoderm differentiation markers

nestin and sox-1 were significantly increased in ESCs treated with vesicles.

Figure 4. Vesicle cargo characterization. MicroRNA packaged within the preosteoblast-derived vesicles were

analyzed using gene expression arrays for mouse cell differentiation and development. (A) Heat map and (B)

scatterplot results indicate the increase in specific miRNAs in vesicles compared to the preosteoblast cells from

which they were derived. (C) The bar graph displays the breakdown of differential up- and down-regulation of

all miRNAs examined.

27

Figure 1

Vesicle characterization. Vesicles isolated from preosteoblast cultures stained positive for CD40 using

immunogold transmission electron microscopy (immunoTEM; top left). Nanoparticle tracking analysis

indicated that particles were on average 174.2±22 nm (top right). Flow cytometric analysis identified a CD40-

positive population (bottom panel), confirming immunoTEM indications of the surface marker.

28

Figure 2

Cell-vesicle interactions. Extracellular vesicles were stained with CellMask™ Orange and visualized on clusters

of actin (phalloidin/green)- and nuclei (DAPI/blue)-stained embryonic stem cells (ESCS) after 6 h (top left) and

24 h (top right). The presence of vesicles in ESC culture did not influence viability, as indicated by alamarBlue®

(bottom left) and LIVE/DEAD® (bottom right) assays. Scale bars in LIVE/DEAD images are 200 μm and are

both from day 6. Live cells are indicated by green fluorescence while dead cells (insets) are indicated by red.

29

Figure 3

ESC gene expression changes. The influence of extracellular vesicles on differentiation ESCs was monitored

using real-time–polymerase chain reaction. Pluripotent genes oct4 and nanog as well as neurectoderm

differentiation markers nestin and sox-1 were significantly increased in ESCs treated with vesicles.

30

Figure 4

31

32

Vesicle cargo characterization. microRNA (miRNA) packaged within the preosteoblast-derived vesicles were

analyzed using gene expression arrays for mouse cell differentiation and development. (A) Heat map and (B)

scatterplot results indicate the increase in specific miRNAs in vesicles compared with the preosteoblast cells

from which they were derived. (C) The bar graph displays the breakdown of differential up- and down-

regulation of all miRNAs examined.

33

34